Prolonged antibiotics for non-cystic fibrosis bronchiectasis in children and adults

  • Review
  • Intervention

Authors


Abstract

Background

The vicious cycle hypothesis for bronchiectasis predicts that bacterial colonisation of the respiratory tract perpetuates inflammatory change. This damages the mucociliary escalator, preventing bacterial clearance and allowing persistence of pro-inflammatory mediators. Conventional treatment with physiotherapy and intermittent antibiotics is believed to improve the condition of people with bronchiectasis, although no conclusive data show that these interventions influence the natural history of the condition. Various strategies have been tried to interrupt this cycle of infection and inflammation, including prolonging antibiotic treatment with the goal of allowing the airway mucosa to heal.

Objectives

To determine the benefits of prolonged antibiotic therapy in the treatment of patients with bronchiectasis.

Search methods

We searched the Cochrane Airways Group Trials Register and reference lists of identified articles. Searches were current as of February 2014.

Selection criteria

Randomised trials examining the use of prolonged antibiotic therapy (for four or more weeks) in the treatment of bronchiectasis compared with placebo or usual care.

Data collection and analysis

Two review authors independently assessed trial quality and extracted data. We contacted study authors to ask for missing information.

Main results

Eighteen trials met the inclusion criteria, randomly assigning a total of 1157 participants. Antibiotics were given for between four weeks and 83 weeks. Limited meta-analysis was possible because of the diversity of outcomes reported in these trials. Based on the number of participants with at least one exacerbation, the meta-analysis showed significant effects in favour of the intervention (odds ratio (OR) 0.31, 95% confidence interval (CI) 0.19 to 0.52; P value < 0.00001), with events occurring in 271 per 1000 people in the intervention arm (95% CI 126 to 385) and in 546 per 1000 in the control population, based on evidence of moderate quality. A non-statistically significant reduction in hospitalisation favoured the use of prolonged antibiotics with a moderate quality grade of supporting evidence (37 per 1000 in the intervention arm (95% CI 13 to 96) and 87 per 1000 in control (OR 0.40, 95% CI 0.14 to 1.11; P value = 0.08). Drug resistance developed in 36 of 220 participants taking antibiotics compared with 10 of 211 participants given placebo or standard therapy (OR 3.48, 95% CI 1.20 to 10.07; P value = 0.02), translating to natural frequencies of 155 per 1000 in the intervention arm (95% CI 59 to 346) and 50 per 1000 in the control arm. The intervention was well tolerated with no overall significant difference in withdrawal between treatment and placebo groups (OR 0.91, 95% CI 0.56 to 1.49). Diarrhoea was commonly reported as an adverse event, particularly with an oral intervention.

Authors' conclusions

Available evidence shows benefit associated with use of prolonged antibiotics in the treatment of patients with bronchiectasis, at least halving the odds of exacerbation (with 275 fewer exacerbations per every 1000 people treated in the antibiotic arm compared with the control arm) and hospitalisation (50 fewer hospitalisations per 1000 people in the antibiotic arm compared with the control arm). However, the risk of emerging drug resistance is increased more than threefold. This review is limited by diversity of trials and by evidence of moderate to low quality. Further randomised controlled trials with adequate power and standardised end points are required.

Resumen

Antibióticos prolongados para la bronquiectasia de la fibrosis no quística en niños y adultos

Antecedentes

La hipótesis del ciclo vicioso para la bronquiectasia anticipa que la contaminación bacteriana del tracto respiratorio prolonga el cambio inflamatorio. Lo anterior provoca daños en la escalera mucociliar que impide la eliminación de las bacterias y da lugar a la persistencia de mediadores proinflamatorios. Se cree que el tratamiento convencional con fisioterapia y antibióticos intermitentes mejora el estado de los pacientes con bronquiectasia, a pesar de que no existen datos concluyentes que demuestren que estas intervenciones afectan la historia natural de la enfermedad. Se han probado diversas estrategias para interrumpir este ciclo de infección e inflamación, incluida la prolongación del tratamiento con antibióticos con el objetivo de permitir que sane la mucosa de las vías respiratorias.

Objetivos

Determinar los beneficios del tratamiento prolongado con antibióticos en dosis altas en pacientes con bronquiectasia.

Métodos de búsqueda

Se realizaron búsquedas en el Registro de ensayos del Grupo Cochrane de Vías Respiratorias (Cochrane Airways Group) y en las listas de referencias de los artículos identificados. Las búsquedas se actualizaron hasta febrero de 2014.

Criterios de selección

Ensayos aleatorios que examinaran la administración de tratamiento prolongado con antibióticos (durante cuatro semanas o más) para la bronquiectasia en comparación con placebo o atención habitual.

Obtención y análisis de los datos

Dos autores de la revisión, de forma independiente, evaluaron la calidad de los ensayos y extrajeron los datos. Se contactó con los autores de los estudios para solicitar la información faltante.

Resultados principales

Dieciocho ensayos cumplieron los criterios de inclusión, asignando al azar a un total de 1157 participantes. Los antibióticos se administraron por un período entre cuatro y 83 semanas. Fue posible realizar un metanálisis limitado debido a la diversidad de resultados informados en estos ensayos. Sobre la base del número de participantes con al menos una exacerbación, el metanálisis mostró efectos significativos a favor de la intervención (odds ratio [OR] 0,31; intervalo de confianza [IC] del 95%: 0,19 a 0,52; valor de p < 0,00001), y se observaron eventos en 271 por 1000 pacientes en el brazo de intervención (IC del 95%: 126 a 385) y en 546 por 1000 en la población control, según pruebas de calidad moderada. Hubo una reducción estadísticamente no significativa en la hospitalización a favor de la administración prolongada de antibióticos con pruebas de un grado de calidad moderado que apoyaron este hecho (37 por 1000 en el brazo de intervención [IC del 95%: 13 a 96] y 87 por 1000 en el control [OR 0,40; IC del 95%: 0,14 a 1,11; valor de p = 0,08]). Se desarrolló resistencia a los fármacos en 36 de 220 participantes que recibieron antibióticos en comparación con diez de 211 participantes que recibieron placebo o tratamiento estándar (OR 3,48; IC del 95%: 1,20 a 10,07; valor de p = 0,02), lo que se traduce en frecuencias naturales de 155 por 1000 en el brazo de intervención (IC del 95%: 59 a 346) y de 50 por 1000 en el brazo control. La intervención presentó buena tolerabilidad, sin diferencias significativas en general en los retiros entre los grupos de tratamiento y placebo (OR 0,91; IC del 95%: 0,56 a 1,49). La diarrea se informó habitualmente como un evento adverso, en particular con una intervención oral.

Conclusiones de los autores

Las pruebas disponibles muestran efectos beneficiosos asociados con el uso prolongado de antibióticos para el tratamiento de los pacientes con bronquiectasia, ya que se reducen al menos a la mitad las probabilidades de exacerbaciones (275 exacerbaciones menos por cada 1000 pacientes tratados en el brazo de antibiótico en comparación con el brazo control) y de hospitalización (50 hospitalizaciones menos por 1000 pacientes en el brazo de antibiótico en comparación con el brazo control). Sin embargo, el riesgo de que surja resistencia a los fármacos aumenta más del triple. Esta revisión está limitada por la diversidad de ensayos y por las pruebas de calidad moderada a baja. Se necesitan más ensayos controlados aleatorios con poder estadístico adecuado y variables principales de evaluación estandarizadas.

Plain language summary

Prolonged antibiotics for purulent bronchiectasis in children and adults

Does prolonged antibiotic therapy provide benefit in treatment of patients with purulent bronchiectasis?

Why is this question important?

Non-cystic fibrosis (CF) bronchiectasis is a chronic respiratory condition characterised by abnormal dilatation of the airways. Although its global prevalence is largely unknown, available data from Australia, New Zealand, the United States and England show that bronchiectasis is now diagnosed with increasing frequency. The lungs of patients with bronchiectasis have excessive secretions, which tend to consist of different types of micro-organisms. Long-term antibiotic therapy was proposed to halt persistent and ongoing damage to the lung due to insult from micro-organisms. Therefore, we seek to assess the effects of prolonged antibiotic therapy on patients with bronchiectasis.

How did we answer the question?

We looked for all studies comparing prolonged antibiotic therapy versus usual care and/or a dummy medication (placebo).

What did we find?

We found 18 studies including 1157 people with non-cystic fibrosis bronchiectasis; most were adults. Twelve studies used a tablet form of antibiotics (e.g. azithromycin, erythromycin, roxithromycin, amoxycillin, clarithromycin, penicillin, oxytetracycline, ciprofloxacin). The remaining six studies reported use of inhaled medications. Antibiotics were given for between four weeks and 83 weeks. In seven studies, similar disease severity at baseline was supported by the similar history of previous hospitalisation and severe attacks.

Overall quality of evidence was rated as moderate. When an outcome is rated as high quality, further research is very unlikely to change our confidence in the estimate of effect, but moderate ratings reflect some uncertainty in the findings.

Conclusion

Prolonged antibiotic therapy in bronchiectasis provides benefit, especially in reducing the risk of future exacerbations and hospitalisations. Antibiotics are well tolerated by participants without significant differences in overall adverse effects (e.g. intolerance, chest symptoms, fatigue, fever, palpitations).

However, antibiotic resistance is a matter of concern, particularly for patients with drug allergies, which further limit their future treatment.

Ringkasan bahasa mudah

Antibiotik berpanjangan untuk bronkiektasis bernanah dalam kalangan kanak-kanak dan orang dewasa

Adakah terapi antibiotik berpanjangan memberi manfaat dalam rawatan pesakit yang mempunyai bronkiektasis bernanah?

Mengapakah soalan ini penting?

Bronkiektasis fibrosis bukan sistik adalah suatu keadaan kronik yang bercirikan dilatasi saluran pernafasan yang abnormal. Walaupun prevalen global penyakit ini tidak diketahui, data sedia ada dari Australia, New Zealand, Amerika Syarikat dan England menunjukkan bronkiektasis kini didiagnos dengan lebih kerap. Paru-paru pesakit bronkiektasis mempunyai rembesan yang berlebihan dan cenderung untuk mendapat jangkitan dari pelbagai jenis mikroorganisma. Terapi antibiotik jangka panjang telah dicadangkan untuk menghentikan kerosakan paru-paru berterusan akibat serangan mikroorganisme. Oleh itu, kami berusaha untuk menilai keberkesanan terapi antibiotik yang berpanjangan ke atas pesakit bronkiektasis.

Bagaimanakah kami menjawab soalan tersebut?

Kami mencari semua kajian yang membandingkan terapi antibiotik berpanjangan dengan rawatan biasa dan/atau ubat plasebo.

Apa yang kami dapati?

Kami mendapat 18 kajian, melibatkan 1157 orang yang mempunyai bronkiektasis fibrosisbukan sistik; kebanyakannya adalah orang dewasa. Dua belas kajian menggunakan tablet antibiotik (contohnya azithromycin, erythromycin, roxithromycin, amoxicillin, clarithromycin, penisilin, oxytetracycline, ciprofloxacin). Manakala, baki enam kajian melaporkan penggunaan ubat-ubatan yang dihidu. Antibiotik diberikan selama empat minggu hingga 83 minggu. Dalam tujuh kajian, keterukan penyakit yang serupa sebelum kajian bermula telah disokong oleh bilangan sejarah kemasukkan hospital dan serangan penyakit yang teruk.

Kualiti keseluruhan bukti adalah sederhana. Apabila sesuatu hasil dinilai sebagai berkualiti tinggi, penyelidikan lanjut tidak akan mengubah keyakinan kami dalam anggaran kesan rawatan, tetapi penilaian sederhana mencerminkan beberapa ketidakpastian dalam penemuan-penemuan kajian.

Kesimpulan

Terapi antibiotik yang berpanjangan dalam bronkiektasis memberi manfaat, terutamanya dalam mengurangkan risiko serangan penyakit dan kemasukkan hospital pada masa depan. Antibiotik juga mudah diterima oleh peserta tanpa perbezaan yang ketara dengan kesan buruk (contohnya ketaktoleranan, simptom dada, keletihan, demam, dan jantung berdebar-debar).

Walau bagaimanapun, ketahanan antibiotik adalah satu perkara yang membimbangkan, terutamanya bagi pesakit yang mempunyai alahan ubat, yang akan menghadkan rawatan mereka pada masa depan.

Catatan terjemahan

Diterjemahkan oleh Raymond Chieng Siang Ching (Melaka Manipal Medical College). Disunting oleh Noorliza Mastura Ismail (Kolej Perubatan Melaka-Manipal). Untuk sebarang pertanyaan mengenai terjemahan ini sila hubungi raymondchieng@gmail.com.

Resumen en términos sencillos

Antibióticos prolongados para la bronquiectasia purulenta en niños y adultos

¿La administración prolongada de antibióticos tiene efectos beneficiosos en el tratamiento de los pacientes con bronquiectasia purulenta?

¿Por qué es importante esta pregunta?

La bronquiectasia de la fibrosis no quística es una afección respiratoria crónica caracterizada por la dilatación anormal de las vías respiratorias. Aunque su prevalencia global es en gran parte desconocida, los datos disponibles de Australia, Nueva Zelandia, los Estados Unidos e Inglaterra indican que la bronquiectasia actualmente se diagnostica con una frecuencia cada vez mayor. Los pulmones de los pacientes con bronquiectasia presentan secreciones excesivas, que tienden a incluir diferentes tipos de microorganismos. Se ha propuesto el tratamiento con antibióticos a largo plazo para detener el daño persistente y continuo a los pulmones debido al deterioro provocado por los microorganismos. Por lo tanto, se intenta evaluar los efectos del tratamiento prolongado con antibióticos en los pacientes con bronquiectasia.

¿Cómo se respondió a la pregunta?

Se realizaron búsquedas de todos los estudios que compararon el tratamiento prolongado con antibióticos versus atención habitual o una medicación simulada (placebo).

¿Qué se encontró?

Se encontraron 18 estudios que incluyeron a 1157 pacientes con bronquiectasia de la fibrosis no quística, en su mayoría adultos. Doce estudios utilizaron antibióticos en forma de comprimidos (p.ej. azitromicina, eritromicina, roxitromicina, amoxicilina, claritromicina, penicilina, oxitetraciclina, ciprofloxacino). Los seis estudios restantes informaron el uso de fármacos inhalados. Los antibióticos se administraron por un período entre cuatro y 83 semanas. En siete estudios, la gravedad similar de la enfermedad al inicio fue apoyada por antecedentes similares de hospitalización anterior y crisis graves.

La calidad global de las pruebas se calificó como moderada. Cuando un resultado se considera de alta calidad, es muy poco probable que los estudios de investigación adicionales cambien la confianza en la estimación del efecto, aunque las calificaciones moderadas reflejan algunas dudas en los hallazgos.

Conclusión

El tratamiento prolongado con antibióticos en la bronquiectasia proporciona efectos beneficiosos, especialmente en cuanto a la reducción del riesgo de futuras exacerbaciones y hospitalizaciones. Los antibióticos son bien tolerados por los participantes, sin diferencias significativas en los efectos adversos generales (p.ej. intolerancia, síntomas respiratorios, fatiga, fiebre, palpitaciones).

Sin embargo, la resistencia a los antibióticos es un tema de interés, en particular en los pacientes con alergias a los fármacos, lo que limita aún más el tratamiento futuro.

Notas de traducción

La traducción y edición de las revisiones Cochrane han sido realizadas bajo la responsabilidad del Centro Cochrane Iberoamericano, gracias a la suscripción efectuada por el Ministerio de Sanidad, Servicios Sociales e Igualdad del Gobierno español. Si detecta algún problema con la traducción, por favor, contacte con Infoglobal Suport, cochrane@infoglobal-suport.com.

Summary of findings(Explanation)

Summary of findings for the main comparison. Prolonged antibiotics for purulent bronchiectasis in children and adults
  1. aPresence of heterogeneity in 30% to 60% range, representing possibility of moderate heterogeneity.
    bSmall number of events producing wide confidence intervals around estimate of effect.
    cPresence of heterogeneity in 75% to 100% range, representing possibility of considerable heterogeneity.

Prolonged antibiotics for purulent bronchiectasis in children and adults
Patient or population: adults and children with purulent bronchiectasis
Settings:
Intervention: prolonged antibiotics
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
Number of participants
(studies)
Quality of evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Control Prolonged antibiotics
Exacerbations
Number of events
Follow-up: 1.5 to 24 months
546 per 1000 271 per 1000
(126 to 385)
OR 0.31
(0.12 to 0.52)
654
(10 studies)
⊕⊕⊕⊝
Moderate a
15 studies reported data on hospitalisations; 10 dichotomous studies are reported here. Generic inverse variance combining dichotomous and continuous data produced statistical significance in favour of antibiotic arm (OR 0.31, 95% CI 0.19 to 0.52; I² = 51%; P value < 0.00001)
Hospitalisations
Number of events
Follow-up: 1.5 to 24 months
87 per 1000 37 per 1000
(13 to 96)
OR 0.40
(0.14 to 1.11)
643
(7 studies)
⊕⊕⊕⊝
Moderate a
Cross-over study Drobnic 2005 (n = 30 participants) also reported on hospitalisations with mean (± SD) improvement of 0.15 ± 0.37 observed in antibiotic arm compared with 0.75 ± 1.16 in placebo arm at 13 month follow-up
Emergence of resistance
Number of events
Follow-up: 1.5 to 24 months
47 per 1000 148 per 1000
(56 to 334)
OR 3.48
(1.20 to 10.07)
431
(6 studies)
⊕⊕⊕⊝
Moderate b
Nine studies reported on emergence of resistance; however, because of variability in reporting of outcome data, only 6 could be meta-analysed. No difference was observed between groups for the remaining 3 studies
St George Respiratory Questionnaire (SGRQ) total
Mean and standard deviation. Scale from 0 to 100
Follow-up: 6 to 16 months
Mean SGRQ (total) score ranged across control groups from
-6.4 to 4.1 points
Mean SGRQ (total) score in intervention groups was
2.75 lower
(7.08 lower to 1.57 higher)
 315
(5 studies)
⊕⊕⊝⊝
Low c
Quality of life was reported in 10 studies, with meta-analysis possible in 5. Substantial heterogeneity was observed between studies. Although sensitivity analysis using a fixed-effect model produced statistically significant results, primary analysis using a random-effects model was not statistically significant
*The basis for the assumed risk (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; OR: Odds ratio.
GRADE Working Group grades of evidence.
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Description of the condition

Non-cystic fibrosis (CF) bronchiectasis is a chronic respiratory condition characterised by abnormal dilatation of the bronchial lumen (Loebinger 2009; O'Donnell 2008). Global prevalence of this disorder is largely unknown (Kwak 2010; Maguire 2012), primarily because bronchiectasis may be noted as a secondary diagnosis for other conditions (AIHW 2010). Formal diagnosis requires special investigation and therefore usually occurs only among patients referred to a respiratory specialist (AIHW 2010). However, available data on prevalence show that bronchiectasis is now being diagnosed with increasing frequency (Martínez-García 2011; McShane 2013; O'Donnell 2008). In the United States of America (USA), Seitz 2012 found an increasing yearly prevalence from 2000 to 2007, with an annual percentage change of 8.74%. Prevalence estimates in the USA range from 4.2 per 100,000 persons 18 to 34 years of age to 271.8 per 100,000 among those 75 years of age and older (Weycker 2005). In Finland, the incidence of non-CF bronchiectasis among children younger than 15 years has been estimated at 0.5 per 100,000 per year and at 3.9 per 100,000 for the overall population (Saynajakangas 1998). Higher prevalence estimates have been reported in children living in Northern England with 17.2 per 100,000 per year diagnosed, and in New Zealand with estimates of 3.7 per 100,000 (Twiss 2005). In Australia, however, bronchiectasis among Aboriginal and Torres Strait Islander children is estimated to occur in as many as 14 per 1000 Indigenous youth (Chang 2002), with similar statistics reported when New Zealand children of Pacific and Maōri origin were compared with children of European ancestry (Twiss 2005). Statistics indicate that prevalence increases with age and peaks at around 80 years, and that higher rates have been observed among women (AIHW 2010; Lee 2011; Seitz 2012).

The underlying pathogenesis of non-CF bronchiectasis is not entirely understood (Wilson 2013b). In most patients, the airways become chronically infected with a variety of pathogens involved in the aetiology of this condition (Chalmers 2012; Murray 2011; Tunney 2013), including Haemophilus influenzae (which dominates across all ages),Streptococcus pneumoniae, Moraxella catarrhalis and Pseudomonas aeruginosa (Grimwood 2011). The landmark study by Reid 1950 first refined pathological phenotypes as cylindrical, varicose and saccular, and Cole 1986 further explained the evolution of bronchiectasis through the vicious cycle model. This model continues to be adapted but generally describes development of a vicious cycle of infection and inflammation whereby symptoms persist, with frequent exacerbations causing further airway damage (Cole 1986; King 2009; Murray 2011; O'Donnell 2008). Idiopathic bronchiectasis is the most common diagnosis and tends to be bilateral, cylindrical and lower lobe predominant (Wilson 2013b). Severe infections of any type can damage the bronchial wall sufficiently to induce bronchiectasis localised to the site of infection (Wilson 2013b). The next most common aetiologies are allergic bronchopulmonary aspergillosis, common variable immunodeficiency and primary ciliary dyskinesia (Wilson 2013b).

Bronchiectasis causes significant morbidity and increased risks for mortality. Typically patients will suffer with persistent cough, chronic daily sputum expectoration, recurrent chest infection and poor health-related quality of life (Murray 2011). Mortality was examined in one recent Belgium cohort analysis of 245 participants newly diagnosed between 2006 and 2012; analysis revealed increased risk of death among participants with co-morbid chronic obstructive pulmonary disease (COPD) compared with participants with other bronchiectasis (Goeminne 2014). However, national statistics in Australia show that only a small number of reported deaths are directly attributable to bronchiectasis (AIHW 2010). A long-term survival study that provided 13 years of follow-up among 91 individuals with non-CF bronchiectasis in the United Kingdom found that mortality, as was reported in 29.7% of participants, was associated with a degree of restrictive and obstructive disease, poor gas transfer and chronic Pseudomonas infection (Loebinger 2009). Additional risk factors observed to lower the likelihood of survival included increasing age and greater numbers of lobes affected (Goeminne 2014).

Description of the intervention

Recent evidence has helped improve our understanding of the role of mucus stasis in bacterial colonisation and has led to emphasis on therapies that enhance airway clearance (McShane 2013). Low-dose, long-term macrolide therapy has been shown to decrease exacerbation frequency and airway inflammation (McShane 2013). Long-term use of antibiotics may provide benefit by reducing exacerbations but currently is not recommended as part of routine treatment (Valery 2012; Wu 2014). However, for patients with three or more exacerbations per year requiring antibiotic therapy, or for those with fewer exacerbations causing significant morbidity, prolonged antibiotic therapy should be considered (Pasteur 2010). Long-term use of inhaled antibiotics is believed to be safe and effective in reducing the sputum bacterial load, as these agents deliver a high concentration of drug to the airway along with reduced systemic absorption, thereby reducing the risk of systemic adverse effects (McShane 2013). However, most data supporting the use of tobramycin, gentamycin and colistin compounded or re-constituted into a nebulised form for the management of bronchiectasis have come from studies on the CF population (McShane 2013), thus limiting the generalisability of findings to the non-CF bronchiectasis population.

How the intervention might work

Various strategies have been developed to interrupt the vicious cycle model of Cole 1986, whereby a repetitious syndrome of cough, sputum production and recurrent exacerbations is said to be the primary cause of bronchiectasis. If bacteria are the primary cause of airway inflammation, then according to this hypothesis, bacterial clearance through the use of short- or long-term antibiotics should reduce airway inflammation (Chalmers 2012; Downey 2007; Ordonez 2003; Sagel 2011). Reducing inflammation would effectively allow time for airway healing and ultimately would modify the long-term course of the disease (Evans 2007; Wu 2014). One study of 49 stable participants with non-CF bronchiectasis reported a link between bacterial load and airway inflammation through bronchoalveolar lavage (Angrill 2001). Another study in patients with COPD that included 43 participants with bronchiectasis also observed the correlation (Hill 2000). A strong evidence base from studies of patients with cystic fibrosis bronchiectasis supports this theory by reporting that antibiotic treatment during both stable and exacerbation periods resulted in reduced markers of inflammation. The disease pathophysiology is different; therefore direct comparison between these conditions is precarious (Chalmers 2012).

Why it is important to do this review

To date, clinical research trials have described mixed results on the effectiveness of antibiotics for reducing airway inflammation and subsequent bronchiectasis (Chalmers 2012). Some studies have suggested that antibiotic therapy had little or no effect on airway inflammation (Tsang 1999), but others examining long-term use of antibiotics have revealed reduced levels of myeloperoxidase (MPO) and neutrophil elastase activity in sputum (Hill 1988; Lin 1997), indicating reduced airway inflammation (Chalmers 2012). The last Cochrane review update on this topic produced some evidence to suggest a small benefit can be achieved with the use of prolonged antibiotics in the treatment of bronchiectasis (Evans 2007), although recommendations were limited by a paucity of data. Since 2007, several trials using different types of antibiotic therapy have been published. Therefore, an update of the existing Cochrane review undertaken to examine newly published evidence on the role of prolonged antibiotic therapy in the management of non-CF bronchiectasis is required.

Objectives

To determine the benefits of prolonged antibiotic therapy in the treatment of patients with bronchiectasis.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled parallel and cross-over trials.

Types of participants

Adult and paediatric participants diagnosed with bronchiectasis by plain film chest radiograph, bronchography or high-resolution computed tomography who reported daily sputum expectoration for at least three months were included. Studies were excluded if patients had been receiving continuous or high-dose antibiotics immediately before the study, or if they had received a diagnosis of cystic fibrosis (CF), sarcoidosis or allergic bronchopulmonary aspergillosis.

Types of interventions

Any dose of prolonged antibiotic therapy of four or more weeks versus placebo or as required treatment.

Types of outcome measures

Primary outcomes
  1. Exacerbations.

  2. Hospitalisations.

Secondary outcomes
  1. Response rates.

  2. Sputum volume and purulence.

  3. Measures of lung function (e.g. forced expiratory volume in one second (FEV1)).

  4. Systemic markers of infection (e.g. leucocyte count, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR)).

  5. Adverse events (e.g. cardiac arrhythmias, GI symptoms, hearing impairment).

  6. Deaths.

  7. Emergence of resistance to antibiotics.

  8. Exercise capacity (e.g. Six-Minute Walk Distance (6MWD)).

  9. Quality of life (e.g. St George Respiratory Questionnaire (SGRQ)).

Reporting by the trial of one of more of the outcomes listed here was not an inclusion criterion for the review.

Search methods for identification of studies

Electronic searches

We identified trials from the Cochrane Airways Group Specialised Register (CAGR), which is maintained by the Trials Search Co-ordinator for the Group. The Register contains trial reports identified through systematic searches of bibliographic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Allied and Complementary Medicine Database (AMED) and PsycINFO, and by handsearching of respiratory journals and meeting abstracts (see Appendix 1 for further details). We searched all records in the CAGR using the search strategy provided in Appendix 2.

We also conducted a search of ClinicalTrials.gov (www.ClinicalTrials.gov) and the World Health Organization (WHO) trials portal (www.who.int/ictrp/en/). We searched all databases from their inception to February 2014, with no restriction on language of publication.

Searching other resources

In addition, we checked the reference lists of all available primary studies and review articles to identify potentially relevant citations and made inquiries to the authors of primary studies regarding other published or unpublished trials known to them.

Data collection and analysis

Selection of studies

Two review authors (KH and CN) independently examined the output generated by the literature search. We obtained all potentially relevant articles and selected trials from identified studies on the basis of previously agreed inclusion criteria. Review authors described study characteristics and outcomes.

Data extraction and management

Two review authors (KH and CN) independently extracted data and risk of bias data from included trials and resolved conflicts by discussion with a third review author (KVC).

Assessment of risk of bias in included studies

Two review authors (KH and CN) assessed each study for risk of bias for random sequence generation, allocation concealment, blinding of participants and outcome assessors, handling of missing data, selective outcome reporting and other threats to validity in the studies, in line with recommendations provided in theCochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We conducted a retrospective risk of bias assessment by applying the above method to all original studies included in the previous version of this review.

Measures of treatment effect

We extracted and analysed continuous and dichotomous outcome data using standard statistical techniques with a fixed-effect model for all studies deemed similar enough to be pooled. In the presence of significant heterogeneity, we employed a random-effects model.

For continuous outcomes, we calculated mean differences (MDs) with 95% confidence intervals (CIs) and pooled MDs or standardised mean differences (SMDs). For dichotomous outcomes, we calculated risk ratios (RRs) with 95% CIs.

We performed a narrative synthesis for each of the included studies and combined all data using Review Manager software.

Unit of analysis issues

We included in the review a mixture of cross-over and parallel studies with the potential for unit of analysis issues to occur. We used the generic inverse variance (GIV) method (by entering effect estimates and their standard errors) to adjust for unit of analysis errors when meta-analysing the data, as per Section 7.7.7 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Dealing with missing data

We evaluated missing information regarding participants on an available case analysis basis, as described in Chapter 16.2.2 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). When statistics essential for analysis were missing (e.g. when group means and standard deviations for both groups were not reported) and could not be calculated from other data, we attempted to contact study authors to obtain missing data. Loss of participants that occurred before baseline measurements were obtained was assumed to have no effect on the eventual outcome data of the study. We assessed and discussed losses after baseline measurements were taken by using an intention-to-treat approach.

Assessment of heterogeneity

We assessed statistical heterogeneity by using a combination of tests including I2 statistic ≥ 50% and visual inspection of data; had 10 or more studies been included, we would have used funnel plots as well. We considered the Der-Simonian and Laird method of analysis presented with a P value less than 0.05 as statistically significant.

In the presence of significant heterogeneity (as per the criteria above), we re-analysed data using both fixed-effect and random-effects models.

Assessment of reporting biases

We planned to explore potential reporting biases by using a funnel plot if we were able to meta-analyse 10 or more studies. Instead, we extrapolated on this possible risk of bias within the other bias section in the risk of bias tables.

Data synthesis

We combined data from all trials using Review Manager 5.2 software. We reported studies by using intention-to-treat analysis when all participants who were randomly assigned during the study were assessed, regardless of whether they received the intervention/study treatment to which they were allocated.

Subgroup analysis and investigation of heterogeneity

Planned subgroups included adults versus children.

Sensitivity analysis

We planned to conduct a sensitivity based on risk of bias. However, we did not do this as none of the included studies had high risk of bias for sequence generation or allocation concealment.

Results

Description of studies

See Characteristics of included studies, Characteristics of excluded studies and Characteristics of studies awaiting classification for additional details on 18, 202 and four studies, respectively.

Results of the search

We have provided in Table 1 details of previous searches up to January 2007. Update searches from 2007 to February 2014 retrieved 249 citations. After exclusions based on title and abstract, we assessed 221 full-text papers for eligibility. We determined that nine new studies were eligible for inclusion and hence included a total of 18 studies in this review. See Figure 1 for the study flow diagram.

Figure 1.

Study flow diagram.

Table 1. Search history
Search yearsDetail
All years to January 2003N citations retrieved: 447
N studies included: 6
January 2003 to January 2004No new studies identified
January 2004 to January 2005No new studies identified
January 2005 to January 2007N citations retrieved: 11
N retrieved: 4
N included: 3

For details on each included study, see Characteristics of included studies and the description of characteristics of included studies provided in Table 2.

Table 2. Description of characteristics of included studies
  1. * = median; † = mean and range; ‡ = median and IQR; ¶ = mean + SE.

Study ID Design Country Age (mean ± SD) Participants (N) Follow-up (weeks)
Int Cont Int Cont
Valery 2013RCTAustralia/NZ3.99 ± 2.144.22 ± 2.30454496
De Diego 2013RCTSpain57 ± 1161 ± 121614NA
Serisier 2013aRCTAustralia61.1 ± 10.563.5 ± 9.5595852
Serisier 2013bRCTAustralia/NZ70 ± 5.659.5 ± 13.2202224
Altenburg 2013RCTNetherlands59.9 ± 12.364.6 ± 9.1464452
Wilson 2013aRCTAust; Germany; Spain; Sweden; UK; USA64.7 ± 11.861.4 ± 11.960648
Liu 2012RCTChinaNANA2422NA
Wong 2012RCTNZ60.9 ± 13.659.0 ± 13.3717052
Murray 2011RCTScotland58 (53-67)‡64 (55.7-69)‡323364
Yalçin 2006RCTTurkey13.1 ± 2.711.9 ± 2.9171752
Cymbala 2005COTUSA70.8 ± 9.770.8 ± 9.7111152 to 56
Drobnic 2005COTSpain64.5 (38-75)†64.5 (38-75)†303052
Barker 2000RCTUSA66.663.237376
Orriols 1999RCTSpain62.0 ± 8.5¶61.4 ± 10.3¶8952
Tsang 1999RCTChina50 ± 1559 ± 16141052
Koh 1997RCTKorea13.3 ± 2.512.9 ± 2.6131212
Currie 1990RCTUK54*51*191952
MRC 1957RCTUKInt 1: 34.3; Int 2: 32.332.6Int 1: 36; Int 2: 444052

Included studies

Study design

Sixteen of the included trials were parallel-group studies, and two used a cross-over design (Cymbala 2005; Drobnic 2005). These 18 studies including international multi-centred studies randomly assigned a total of 1157 participants and were published between 1957 and 2013. Two of these studies were conducted in the USA (Barker 2000; Cymbala 2005), three in the United Kingdom (Currie 1990; MRC 1957; Murray 2011), three in Spain (De Diego 2013; Drobnic 2005; Orriols 1999), and one each in one Australia (Serisier 2013a BLESS), China (Liu 2012), Hong Kong (Tsang 1999), the Netherlands (Altenburg 2013), New Zealand (Wong 2012), South Korea (Koh 1997) and Turkey (Yalçin 2006). Three studies were conducted across multiple countries being Serisier 2013b ORBIT (Australia and New Zealand), Valery 2013 (Australia and New Zealand) and the Wilson 2013a study (Australia, Germany, Spain, Sweden, UK and USA).

Participants

A total of 1157 participants (from n = 1627 eligible participants) were randomly assigned across the 18 studies; most were adults (n = 1009) with a mean age of 59 years. In all, 148 children participated in the three paediatric studies (Koh 1997; Valery 2013; Yalçin 2006). Participants from all of the included studies, except Currie 1990, MRC 1957 and Valery 2013, were diagnosed with non-CF bronchiectasis, confirmed by computed tomography scan. Exclusion criteria were mentioned by all of the included studies and were explicitly described by some.

Baseline lung function

Fourteen studies reported baseline lung function (Altenburg 2013; Barker 2000; Currie 1990; Cymbala 2005; De Diego 2013; Drobnic 2005; Koh 1997; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Wong 2012).

Baseline exacerbation rates or episodes

Seven studies adjusted baseline exacerbation rates or episodes (Altenburg 2013; De Diego 2013; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Wong 2012). In the Barker 2000 study, the intervention group had more frequent exacerbations (intervention n = 5 vs control n = 1), but the Currie 1990 study reported more frequent exacerbations in the control arm (intervention n = 11 vs control n = 15). Eight studies (Cymbala 2005; Drobnic 2005; Koh 1997; Liu 2012; MRC 1957; Murray 2011; Orriols 1999; Valery 2013) did not report baseline exacerbations.

Pseudomonas aeruginosa isolate

Altenburg 2013, Barker 2000, Currie 1990, De Diego 2013, Drobnic 2005, Murray 2011, Orriols 1999, Serisier 2013a BLESS, Serisier 2013b ORBIT, Wilson 2013a and Wong 2012 reported Pseudomonas aeruginosa. All participants as per trial protocols from Barker 2000, Drobnic 2005, Orriols 1999 and Wilson 2013a had positive P. aeruginosa growth at baseline. Currie 1990 cultured 8 of 17 participants taking amoxycillin (intervention) and 7 of 19 given placebo but also reported having 21 participants with Pseudomnas at baseline or during treatment. In Serisier 2013a BLESS, 23 of 59 (39%) participants taking erythromycin (intervention) versus 18 of 58 (31%) given placebo had P. aeruginosa. Altenburg 2013 reported growth of P. aeruginosa in 6 of 43 (14%) participants taking azithromycin(intervention) and in 6 of 40 (15%) given placebo, and De Diego 2013 had P. aeruginosa present in 7 of 16 participants taking azithromycin (intervention) and in 5 of 14 given placebo at baseline. Study authors from Wong 2012 reported that 9 of 71(13%) participants taking azithromycin (intervention) and 8 of 70 (11%) given placebo were positive for P. aeruginosa. Table 3 provides a summary showing types of interventions and status of P. aeruginosa.

Table 3. Types of interventions and status of Pseudomonas aeruginosa
  1. PA = Pseudomonas aeruginosa. *Eradicated. †At day 28. ‡Negative bacterial culture at end of intervention. §Transient disappearance of Pseudomonas in culture. #Actual follow-up period was not reported.

Study IDPatients with PA (baseline)Patients with PA (end of intervention)AntibioticsControlDelivery modeInt duration (weeks)Follow-up (weeks)
Altenburg 20136/43 on intervention, 6/40 on placebo5/43 on intervention, 4/40 on placeboAzithromycinPlaceboOral5252
De Diego 20137/16 on intervention, 5/14 on control0/16 on interventionAzithromycinUsual careOral1212#
Serisier 2013 BLESS23/59 on intervention, 18/58 on placebo17/59 on intervention, 17/58 on placeboErythromycinPlaceboOral5252
Valery 20130/45 on intervention, 0/44 on placebo0/45 on intervention, 0/44 on placeboAzithromycinPlaceboOral82.8 (mean)96
Liu 2012NANARoxithromycinPlaceboOral2424
Wong 2012 EMBRACE9/71 on intervention, 8/70 on placebo5/46 on intervention, 5/45 on placebo*AzithromycinPlaceboOral2452
Yalçin 2006NANAClarithromycinUsual careOral1252
Cymbala 2005NANAAzithromycinUsual careOral2652 to 56
Tsang 1999NA10/11 on intervention, 6/10 on placeboErthromycinPlaceboOral852
Koh 1997NANARoxithromycinPlaceboOral1212
Currie 19908/17 on intervention, 7/19 on placebo5/17 on intervention, 4/19 on placeboAmoxycillinPlaceboOral3252
MRC 1957NANAInt 1: penicillinPlaceboOral5252
Int 2: oxytetracycline
Serisier 2013 ORBIT20/20 on intervention, 22/22 on placebo8/20 on intervention, 19/22 on placebo†CiprofloxacinPlaceboInhaled2424
Wilson 2013a32/60 on intervention, 35/64 on placebo14/40 on intervention, 4/49 on placebo‡CiprofloxacinPlaceboInhaled48
Drobnic 200530/30 on intervention, 30/30 on placebo26/30 on intervention§TobramycinPlaceboNebulised2652
Barker 200037/37 on intervention, 37/37 on placebo13/37 on intervention, 0/37 on placebo*TobramycinPlaceboNebulised46
Orriols 19997/7 on intervention, 8/8 on control7/7 on intervention, 8/8 on controlCeftazidime/tobramycinSymptomatic treatmentNebulised5252
Murray 201113/27 on intervention, 11/30 on placebo4/13 on intervention*GentamicinPlaceboNebulised5264
Smoking history

No information on smoking was available in 11 studies (Barker 2000; Currie 1990; De Diego 2013; Drobnic 2005; Koh 1997; Liu 2012; MRC 1957; Orriols 1999; Valery 2013; Wilson 2013a; Yalçin 2006). Smoking was described variably in Tsang 1999 (1 of 11 taking erythromycin and 2 of 10 given placebo were ex-smokers), Cymbala 2005 (8 of 11 participants were past or present smokers), Murray 2011 (8 of 27 taking azithromycin and 8 of 30 given placebo were ex-smokers), Wong 2012 (1% given intervention vs 6% given placebo were smokers), Serisier 2013a BLESS (10 of 59 (16.9%) and 15 of 58 (25.9%) were ex-smokers), Serisier 2013b ORBIT (1 of 20 taking ciprofloxacin was an ex-smoker) and Altenburg 2013 (20 of 43 taking azithromycin and 18 of 40 given placebo were past or present smokers).

Radiological extent of bronchiectasis

The radiological extent of bronchiectasis was reported in terms of the number of lobes involved by Orriols 1999 (n = 15) and Tsang 1999 (n = 22) (at least three lobes were involved in all participants) and Currie 1990 (12 of 17 taking amoxycillin and 10 of 19 given placebo had three or more lobes).

Interventions

The length of studies varied with intervention duration and ranged from 4 to 82.8 (mean value from Valery 2013) weeks. Antibiotics (and route of administration) assessed in the studies were tobramycin (Barker 2000; Drobnic 2005; Orriols 1999: nebulised), ceftazidime (Orriols 1999: nebulised), amoxycillin (Currie 1990: oral), roxithromycin (Koh 1997; Liu 2012: oral), penicillin (MRC 1957: oral), oxytetracycline (MRC 1957: oral), erythromycin (Serisier 2013a BLESS; Tsang 1999: oral), clarithromycin (Yalçin 2006: oral) and azithromycin (Altenburg 2013; Cymbala 2005; De Diego 2013; Valery 2013; Wong 2012: oral). All studies were placebo-controlled with the exception of Cymbala 2005, De Diego 2013, Orriols 1999 and Yalçin 2006, in which control groups were given usual medical care.

Excluded studies

We recorded reasons for exclusion of 202 studies in Characteristics of excluded studies. The most common reasons for exclusion in order of prevalence include the following: not a randomised controlled trial (n = 67), no prolonged antibiotics (n = 51), not all bronchiectasis patients (n = 37), full review not possible and drug not used in clinical practice (n = 18), not non-CF bronchiectasis (n = 17), other reasons (n = 10) and no control group (n = 2).

Risk of bias in included studies

Full details of our risk of bias judgements can be found under the "Risk of bias" section at the end of each Characteristics of included studies table and in Figure 2. Overall, the methodological quality of included studies was good. Two independent review authors (KH and CN) reached agreement when assessing study quality.

Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Generation of the randomisation sequence was considered adequate in five studies (Altenburg 2013; Cymbala 2005; Serisier 2013a BLESS; Valery 2013; Wong 2012) and unclear in 13 studies (Barker 2000; Currie 1990; De Diego 2013; Drobnic 2005; Koh 1997; Liu 2012; MRC 1957; Murray 2011; Orriols 1999; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Wong 2012).

The method of allocation concealment was determined to be adequate in four studies (Altenburg 2013; MRC 1957; Valery 2013; Wong 2012) and unclear in 13 studies (Barker 2000; Currie 1990; De Diego 2013; Drobnic 2005; Koh 1997; Liu 2012; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Yalçin 2006). However, Cymbala 2005 was assessed to have high risk of bias for allocation concealment.

Blinding

Risk associated with participant blinding was deemed to be low when the study provided identical tablets or visually identical liquid and packaging and the same follow-up for participants. Adequate blinding of assessors was reported on the basis of information provided by study authors, such as a clear and well-defined study protocol, central or third party allocation of assessors and treating doctors and the same follow-up, measurements and tests provided for participants from both arms of the study.

Risk of detection bias due to inadequate blinding of outcome assessors has been judged as high in five studies (Cymbala 2005; De Diego 2013; Murray 2011; Orriols 1999; Yalçin 2006) and unclear in five trials (Koh 1997; Liu 2012; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a), with no specific details provided in the trial reports. Although Murray 2011 was assessed to have high risk of bias for blinding of participants, risk was unclear for blinding of outcome assessors. Altenburg 2013, Barker 2000, Currie 1990, Drobnic 2005, MRC 1957, Serisier 2013a BLESS, Valery 2013 and Wong 2012 were deemed to have low risk of bias.

Incomplete outcome data

Low risk of bias was assessed in eight studies (Currie 1990; De Diego 2013; Koh 1997; Murray 2011; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Valery 2013). We report high risk of attrition bias in Wilson 2013a and Drobnic 2005 and unclear risk in eight studies (Altenburg 2013; Barker 2000; Cymbala 2005; Liu 2012; MRC 1957; Orriols 1999; Wong 2012; Yalçin 2006).

Selective reporting

Selective reporting (reporting bias) was considered to introduce low risk in eight studies (Altenburg 2013; Currie 1990; De Diego 2013; Koh 1997; Murray 2011; Serisier 2013a BLESS; Tsang 1999; Wong 2012); however it was determined to bring high risk in five studies (Cymbala 2005; Drobnic 2005; MRC 1957; Serisier 2013b ORBIT; Yalçin 2006). The remaining five studies (Barker 2000; Liu 2012; Orriols 1999; Valery 2013; Wilson 2013a) were identified to have unclear risk for reporting bias.

Other potential sources of bias

Other potential sources of bias were not identified in nine studies (Barker 2000; Currie 1990; Drobnic 2005; Koh 1997; MRC 1957; Murray 2011; Serisier 2013a BLESS; Tsang 1999; Valery 2013), but this could not be adequately assessed in eight studies (Altenburg 2013; Cymbala 2005; De Diego 2013; Liu 2012; Orriols 1999; Wilson 2013a; Wong 2012; Yalçin 2006). High risk of the other potential sources of bias was identified in Serisier 2013b ORBIT.

Effects of interventions

See: Summary of findings for the main comparison Prolonged antibiotics for purulent bronchiectasis in children and adults

Primary outcomes

1. Exacerbations

Fifteen of 18 included studies with 956 participants reported data on exacerbations. A total of 13 studies (with 884 participants) could be pooled in a meta-analysis using generic inverse variance, with 10 reporting dichotomous data (Barker 2000; Currie 1990; Koh 1997; Liu 2012; Murray 2011; Serisier 2013b ORBIT; Tsang 1999; Valery 2013; Wilson 2013a; Wong 2012) and three continuous data (Altenburg 2013; De Diego 2013; Serisier 2013a BLESS), producing statistically significant benefits in favour of the antibiotic arm (odds ratio (OR) 0.31; 95% confidence interval (CI) 0.19 to 0.52; I² = 51%; P value < 0.00001; Figure 3) and reporting similar results for the sensitivity analysis using a fixed-effect model (P value < 0.0001; Analysis 2.1). Data from two cross-over studies (Cymbala 2005; Drobnic 2005) involving 12 and 60 participants, respectively, were limited to individual studies. Drobnic 2005 did not detect a significant difference between treatments in terms of mean exacerbations per participant (mean difference (MD) -0.4, P value = 0.32) but found a significant difference in the mean number of admissions in favour of tobramycin (MD -0.6, P value = 0.03). Cymbala 2005, which included 12 participants, reported fewer exacerbations requiring antibacterial therapy in the azithromycin phase than in the control phase (5 events vs 16, P value = 0.019). Exacerbation was not reported by three trials (MRC 1957; Orriols 1999; Yalçin 2006). For absolute treatment differences, 55 of 100 people in the control group had one or more exacerbations over 1.5 to 24 months compared with 32 (95% CI 25 to 41) of 100 given prolonged antibiotics (Figure 4).

Figure 3.

Forest plot of comparison: 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), outcome: 1.1 Exacerbations.

Figure 4.

Absolute treatment differences for exacerbations with events occurring in 546 per 1000 in the control arm and in 271 per 1000 (95% CI 126 to 385) people on the intervention.

Number of courses of antibiotics

Supplemental antibiotic use for pulmonary or non-pulmonary exacerbations was reported in 14 studies including 1021 participants (Altenburg 2013; Barker 2000; Currie 1990; Cymbala 2005; Drobnic 2005; MRC 1957; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Valery 2013; Wilson 2013a; Wong 2012). Three of the 18 studies analysed did not report on supplementary antibiotic courses during the study treatment phase (Koh 1997; Liu 2012; Yalçin 2006).

Supplemental antibiotic prescription and use correlated with identified exacerbations; however all but two studies failed to indicate duration or number of courses (Altenburg 2013; Barker 2000; Currie 1990; Cymbala 2005; Drobnic 2005; MRC 1957; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Wilson 2013a; Wong 2012). Tsang 1999 recorded that two participants required treatment with sparfloxacin (200 mg daily for 10 days), and Valery 2013 mentioned use of supplemental non-macrolide antibiotics for two weeks following pulmonary exacerbations.

The two cross-over trials (Cymbala 2005; Drobnic 2005) showed that use of supplemental antibiotics correlated with infective pulmonary exacerbations, with less use reported in the treatment arm than in the placebo arm. These trials did not report the number or duration of courses. The number of supplementary antibiotic courses taken during the study treatment phase of the Currie 1990 study was reported to be related to the number of courses taken during the previous year rather than to the allocated treatment regimen.

2. Hospitalisations

Seven parallel studies (Altenburg 2013; Barker 2000; Orriols 1999; Serisier 2013a BLESS; Valery 2013; Wilson 2013a; Wong 2012) with 13 of 322 intervention participants and 28 of 321 placebo participants reported hospitalisations, with a non-statistically significant result in favour of the intervention (OR 0.40, 95% CI 0.14 to 1.11; P value = 0.08; I² = 36%; Analysis 1.2). The sensitivity analysis using the fixed-effect model did however produce statistical significance (P value = 0.02; Analysis 2.2). Drobnic 2005, a cross-over trial with 30 participants, reported improvement in the number of admissions per participant using mean ± SD (0.15 ± 0.37 taking intervention vs 0.75 ± 1.16 given placebo). In terms of absolute treatment differences, 9 of 100 people in the control group had one or more hospitalisations over 1.5 to 24 months compared with 4 (95% CI 2 to 8) of 100 given prolonged antibiotics (Figure 5).

Figure 5.

Absolute treatment differences for hospitalisations with events occurring in 87 per 1000 in the control arm and in 37 per 1000 (95% CI 13 to 96) on the intervention.

Secondary outcomes

1. Response rates

A significant treatment effect was noted following meta-analysis for clinical response rates in two studies (OR 3.37, 95% CI 1.60 to 7.09; n = 110; P value = 0.001; Analysis 1.3). In spite of slightly differing definitions of response between studies (Currie 1990 reported physician assessment of diary cards; Barker 2000 reported physician assessment of overall medical condition), no heterogeneity was observed. Currie 1990 reported no significant differences in disease progression (3/17 vs 4/19 participants in the antibiotic and control groups, respectively). Participant-reported symptoms were measured by the Lower Respiratory Tract Infection Visual Analogue Scale (LRTI VAS) in the Altenburg 2013 study, which showed a larger decrease in total score (indicating fewer symptoms) among participants receiving the intervention compared with control participants at the end of the treatment period (six months; P value = 0.047).

2. Sputum volume and purulence

Sputum volume and purulence were reported in 15 studies including 975 participants, one of which was a cross-over study (Currie 1990; Barker 2000; Cymbala 2005; Drobnic 2005; De Diego 2013; Koh 1997; MRC 1957; Murray 2011; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Valery 2013; Wilson 2013a; Wong 2012; Yalçin 2006). As these results were variably reported, a meta-analysis was not performed. Of these studies, nine recorded data for both sputum volume and purulence (MRC 1957; Currie 1990; Tsang 1999; Yalçin 2006; Murray 2011; Wong 2012; De Diego 2013; Serisier 2013a BLESS; Wilson 2013a). Four of the 15 studies reported on sputum purulence alone (Koh 1997; Barker 2000; Drobnic 2005; Serisier 2013b ORBIT) and one study reported solely on sputum volume (Cymbala 2005).

Of studies reviewing populations of children, Valery 2013 reported on sputum purulence via a sputum colour chart (Bronkotest). However, sputum characteristics were reported as a definition of an exacerbation rather than as a distinct secondary outcome in itself, and no follow-up results were mentioned thereafter. Yalçin 2006 on the other hand reported a significant difference in sputum volume with improvement in the treatment arm of the study. Sputum purulence was measured via total cell number and neutrophil ratios in bronchoalveolar lavage, and also showed a significant improvement in the treatment arm of the study.Both parallel and cross-over trials showed a greater reduction in sputum volume and purulence in the long-term antibiotic group than in the placebo group.

Sputum diary cards were used in one study on 38 participants (Currie 1990). Independent assessment of diary cards in the Currie 1990 study revealed improvement in sputum colour and/or reduction in reported sputum volume, which persisted throughout the post-treatment phase in three intervention participants. Symptomatically, the prolonged antibiotic course was considered successful, as judged by independent assessment of diary card data and in the opinion of participants (Currie 1990).

Koh 1997 reported significantly lower sputum purulence scores in the antibiotic group compared with the placebo group (1.39 ± 0.6 vs 2.17 ± 0.72; P value < 0.01). Yalçin 2006 reported that sputum production was significantly less in the antibiotic group (P value = 0.0001). Cymbala 2005 reported no significant differences in median sputum volume production between treatments (25 mL vs 24 mL). A greater reduction in 24-hour purulent sputum volume was noted in the Currie 1990 study between exacerbations during the study treatment phase in the amoxycillin group (median 20% of pre-treatment volume) compared with the placebo phase (88% of pre-treatment volume; P value = 0.008), although concentrations of Haemophilus spp. in sputum between exacerbations were similar in the two groups. However, no differences between groups were observed in the post-treatment phase. The reduction in 24-hour total sputum volume between acute exacerbations during the study treatment phase was also significantly greater in the amoxycillin group (42% of pre-treatment values) than in the placebo group (81%; P value = 0.04), although in the post-treatment phase no differences were observed between groups. Tsang 1999 reported that participants in the erythromycin group (n = 11) but not in the placebo group (n = 10) had significantly improved 24-hour sputum volume after eight weeks (P value < 0.05).

3. Measures of lung function

Lung function was reported in 16 studies with a total number of 990 randomly assigned participants (Altenburg 2013; Barker 2000; Currie 1990; Cymbala 2005; De Diego 2013; Drobnic 2005; Koh 1997; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Valery 2013; Wilson 2013a; Wong 2012; Yalçin 2006). Most studies reported a combination of forced expiratory volume in one second (FEV1) and forced vital capacity (FVC); however some studies reported FEV1 alone. As lung function was variably reported, only limited data were available for meta-analysis. The only available pooled effect estimates showed no evidence of any effect for FEV1% predicted (standardised mean difference (SMD) -0.12, 95% CI -0.34 to 0.10), change in FEV1 (SMD -0.07, 95% CI -0.84 to 0.70) and absolute FEV1 (SMD -0.36, 95% CI -1.03 to 0.30; Analysis 1.4). A sensitivity analysis based on the fixed-effect model also found no evidence of any effect (Analysis 2.3).

Among the adult population, 10 parallel studies showed no evidence of any effect on lung function measurements in intervention and placebo groups at the end of the trial (Barker 2000; Currie 1990; De Diego 2013; Koh 1997; Murray 2011; Orriols 1999; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Wong 2012). Altenburg 2013 showed statistically significant improvement in FEV1 and FVC at the end of the trial period in the intervention group versus the placebo group (P value = 0.047 and P value = 0.02, respectively). However, these results were not maintained over the six months after cessation of the intervention. Serisier 2013a BLESS showed a significant arrest in the decline of FEV1 among participants receiving erythromycin during the treatment period (P value = 0.04).

Two cross-over trials including 72 participants (Cymbala 2005; Drobnic 2005) reported no evidence of any effect on lung function between intervention and placebo groups at the end of the treatment period. Long-term azithromycin administration during Cymbala 2005 demonstrated no significant changes in FEV1 (1.28 ± 0.55 L vs 1.20 ± 0.46 L (azithromycin) vs 1.18 ± 0.51 L (control); P value = 0.153) or FVC (2.25 ± 0.72 L vs 2.22 ± 0.60 L (azithromycin) vs 2.12 ± 0.62 L (control); P value = 0.440). Non-significant changes (%) in these measures between treatment groups (tobramycin vs placebo) were also noted in Drobnic 2005 (FEV1 –3.50, 95% CI –5.95 to –1.02 vs –1.20, 95% CI –4.38 to –1.98; P value = 0.240; FVC –5.45, 95% CI –8.03 to –2.87 vs –1.30, 95% CI –4.88 to –2.28; P value = 0.056).

The two studies involving children (Valery 2013; Yalçin 2006) reported no evidence of any effect on lung function between antibiotic and placebo groups at the end of the trials (total n = 126). Yalçin 2006 showed no evidence of an effect in FEV1% between treatment groups (clarithromycin vs placebo); however, investigators did report improvement in forced expiratory flow at 25% to 75% (FEF25-75%) in the clarithromycin group at the beginning of the study and at month three (P value = 0.015). Similarly, Valery 2013 showed no evidence of any effect on mean predicted FEV1% between antibiotic and placebo groups (84.7 (12.9) vs 81.0 (18.3); P value = 0.38). However, FEV1 data were incomplete, as spirometry was conducted only in participants older than six years.

4. Systemic markers of infection
Leucocyte count

A total of 402 participants - 201 participants taking intervention versus 201 given placebo - were measured for leucocyte counts in six studies (Altenburg 2013; Currie 1990; Drobnic 2005; Koh 1997; Murray 2011; Wong 2012). Koh 1997 (n = 13 taking roxithromycin, n = 12 given placebo), Currie 1990 (n = 17 taking amoxycillin, n = 19 given placebo), Drobnic 2005 (n = 30 on both arms, cross-over trial), Murray 2011 (n = 27 taking intervention, n = 30 given placebo) and Altenburg 2013 (n = 43 taking azithromycin, n = 40 given placebo) reported no changes in both arms. However, Wong 2012 with 141 participants (n = 71 taking azithromycin, n = 70 given placebo) favoured the intervention (P value = 0.013). A meta-analysis of three studies with 165 participants was possible but produced no evidence of any effect for the primary random-effect analysis (MD -0.39, 95% CI -0.87 to 0.09; Analysis 1.5) and the sensitivity analysis (Analysis 2.4).

C-reactive protein (CRP)

Six studies (Altenburg 2013; De Diego 2013; Murray 2011; Serisier 2013a BLESS; Wilson 2013a; Wong 2012) with 552 participants (276 in the intervention arm vs 276 in the placebo arm) reported on CRP. As variable reporting methods were used, we were unable to perform a meta-analysis of all data. Altenburg 2013 measured the change in serum CRP levels and leucocyte counts seen in 83 participants (43 taking azithromycin vs 40 given placebo) and showed no significant differences between treatment groups (intervention median 5.0, interquartile range (IQR) 2 to 11.3 mg/dL; control median 4.5, IQR 2 to 15.3 mg/dL). Murray 2011 favoured intervention with 57 participants (27 given intervention and 30 placebo). Wong 2012 reported on 141 participants (71 taking azithromycin and 70 given placebo) and favoured intervention (P value = 0.006). De Diego 2013 reported improvement in the change from baseline for 20 participants (16 taking azithromycin and 14 given placebo). In the Wilson 2013a study, 124 participants (60 given intervention and 64 placebo) showed benefit in favour of the intervention, but this finding was not statistically significant (P value = 0.173). Serisier 2013a BLESS reported no effect on CRP among 54 participants taking erythromycin and 50 given placebo.

Erythrocyte sedimentation rate (ESR)

Three studies with 128 participants (Currie 1990; De Diego 2013; Drobnic 2005) reported using ESR as a marker for infection. Two of these studies could be pooled in a meta-analysis, producing no evidence of any effect in the primary analysis (MD -2.47, 95% CI -9.76 to 4.82; Analysis 1.6) or in the sensitivity analysis (Analysis 2.5). Drobnic 2005, a cross-over trial, found no differences between intervention and control groups (n = 30), and the De Diego 2013 study, with 16 participants given intervention and 14 acting as controls, reported a change from baseline (intervention -14 (4) vs control -9 (12)). The ESR in the amoxycillin group of the Currie 1990 study (n = 13 to 17) decreased slightly during treatment but did not fall in the placebo group (n = 11 to 18; P value < 0.05).

5. Adverse events

Adverse events were reported variably in 15 studies with 683 participants (Altenburg 2013; Barker 2000; Currie 1990; Cymbala 2005; De Diego 2013; Drobnic 2005; MRC 1957; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Valery 2013; Wilson 2013a; Yalçin 2006), and when possible, meta-analysis was performed for each of the reported adverse events (Analysis 1.7). Eleven studies (Altenburg 2013; Barker 2000; Currie 1990; MRC 1957; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Yalçin 2006) reported that 39 of 358 participants in the intervention arm and 43 of 325 participants in the placebo arm withdrew from the trials because of intolerable or severe adverse effects. No significant difference was observed when these events were pooled (OR 0.91, 95% CI 0.56 to 1.49; N = 683; Analysis 1.7).

Four studies (Altenburg 2013; Currie 1990; Cymbala 2005; MRC 1957) reported on diarrhoea. Three parallel studies reported that the intervention increased risk of diarrhoea in 24 of 136 participants given antibiotic versus 6 of 95 given placebo (n = 231) (OR 3.33, 95% CI 1.50 to 7.37; P value = 0.003; Analysis 1.7).

Bronchospasm was reported in three studies (Drobnic 2005; Murray 2011; Wilson 2013a) and wheeze in two studies (Barker 2000; Drobnic 2005). Drobnic 2005 had 3 of 30 participants reporting bronchospasm whilst nebulised with tobramycin. Murray 2011 and Wilson 2013a reported that 10 of 92 participants given intervention and 5 of 97 given placebo had bronchospasm (OR 2.19, 95% CI 0.76 to 6.33; P value = 0.15, I² = 21%; Analysis 1.7). Barker 2000 reported that 6 of 37 participants taking tobramycin and 0 of 37 given placebo had wheeze (OR 8.56, 95% CI 1.63 to 44.93; P value = 0.01), and 12 of 37 participants taking tobramycin and 3 of 37 given placebo had worsening of dyspnoea. However, De Diego 2013 reported statistically significant improvement in dyspnoea with 16 participants taking azithromycin and 14 given placebo (P value < 0.05).

Chest pains/palpitations were reported by Altenburg 2013 using oral azithromycin, Barker 2000 using nebulised tobramycin and Serisier 2013a BLESS using oral erythromycin (8 of 139 participants given intervention vs 1 of 135 given placebo; OR 5.21, 95% CI 1.35 to 20.14; P value = 0.02). Four studies (Altenburg 2013 using oral azithromycin, Currie 1990 using oral amoxycillin, Serisier 2013b ORBIT using inhaled ciprofloxacin, Serisier 2013a BLESS using oral erythromycin) with 15 of 138 participants given intervention and 13 of 137 placebo reported nausea (OR 1.20, 95% CI 0.54 to 2.67; P value = 0.66, I² = 37%).

No significant differences between treatments were reported for fatigue (Altenburg 2013; Barker 2000), haemoptysis (Barker 2000; MRC 1957; Wilson 2013a), fever (Barker 2000) or increased sputum (Barker 2000). In addition to diarrhoea, which was the most common adverse event reported in the MRC 1957 study, anorexia, nausea, vomiting, flatulence and epigastric discomfort were also reported in order of frequency. Other recorded adverse events in this study included swelling of the foot accompanied by a rash that lasted a week, which developed eight weeks after the start of treatment with penicillin; paraesthesiae of both hands lasting a few weeks, which occurred 16 weeks after the start of oxytetracycline; and joint pain, which was reported to occur the day after administration of lactose capsules (MRC 1957). No participants withdrew from the Yalçin 2006 study. Valery 2013 reported that the most common adverse events in this trial were non-pulmonary infections (71 of 112 events in the azithromycin group vs 132 of 209 in the placebo group) and bronchiectasis-related events (episodes or investigations; 22 of 112 events in the azithromycin group vs 48 of 209 in the placebo group). However, study drugs were well tolerated and no serious adverse events were attributed to the intervention. Authors of the Tsang 1999 study reported that none of the participants given erythromycin experienced haemoptysis during the study period.

Diarrhoea was the most common treatment-related adverse event reported in the Cymbala 2005 cross-over study, with 3 of 12 participants (25%) reporting it during the intervention period. Study authors did not mention any episodes of diarrhoea occurring during the usual care period. One participant's dose of azithromycin was changed to the alternative thrice-weekly regimen, which was reported to improve diarrhoea. The single participant who dropped out of the study had reported diarrhoea during a previous visit. No other serious or severe treatment-related adverse events were reported (Cymbala 2005).

6. Deaths

No significant difference in mortality (OR 1.48, 95% CI 0.28 to 7.85; n = 595) was noted amongst the seven studies that reported on this outcome (Altenburg 2013; MRC 1957; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Valery 2013; Wilson 2013a); n = 595; Analysis 1.8). The two deaths reported in the intervention arm from Murray 2011 were unexpected and were due to previously undiagnosed metastatic colorectal cancer and myocardial infarction. In Drobnic 2005, a cross-over trial, 5 of 30 participants died from respiratory failure during study periods. Investigators did not report the phase of the study in which these deaths occurred.

7. Emergence of resistance to antibiotics

Nine studies (Altenburg 2013; Barker 2000; Drobnic 2005; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Valery 2013; Wilson 2013a; Wong 2012) with a total of 571 participants reported this outcome. As a result of the different types of results presented, data from only six studies (Barker 2000; Drobnic 2005; Orriols 1999; Valery 2013; Wilson 2013a; Wong 2012) with 431 participants were used for the meta-analysis. Drug resistance in 36 of 220 participants taking antibiotic and in 10 of 211 participants given placebo or standard therapy favoured the placebo or standard therapy group (OR 3.48, 95% CI 1.20 to 10.07; P value = 0.02; Analysis 1.9) with modest heterogeneity (I² = 29%); similar but more significant results were obtained with the sensitivity analysis (P value = 0.006; Analysis 2.6). Murray 2011 detected no resistance at the end of the study.

Resistance patterns were comparable between groups at baseline for the Altenburg 2013 study (35% macrolide resistance in eight participants in the azithromycin group vs 27.5% in nine participants in the placebo group). During treatment, 53 of 60 pathogens (88%) tested for sensitivity in 20 participants in the azithromycin group became macrolide resistant, compared with 29 of 112 pathogens (26%) in 22 participants in the placebo group (P value = 0.001). Macrolide-resistant commensal oropharyngeal streptococci were reported to be significantly increased in the Serisier 2013a BLESS study with erythromycin therapy (median change 27.7%, IQR 0.04% to 41.1%) compared with placebo (median change 0.04%, IQR -1.6% to 1.5%; difference 25.5%, IQR 15.0% to 33.7%; P value < 0.001). In the Barker 2000 study, tobramycin-resistant P. aeruginosa strains developed resistance in 11% of intervention participants compared with 3% of placebo participants, although this difference was not significant. Two participants developed resistance during the tobramycin period of the Drobnic 2005 study. Bacteria persisted in both participants until months two and three of the placebo period, and two additional participants had P. aeruginosa strains isolated during the second period of placebo treatment. The frequency and emergence of resistance did not differ significantly during the second period of placebo treatment. Other organisms were also monitored, with Alcaligenes xylosoxidans and Streptococcus pneumoniae identified in sputum cultures of two participants during the intervention treatment period, and with Acinetobacter spp and Stenotrophomona maltophila identified in two participants during the treatment period (Drobnic 2005).

The Currie 1990 study also reported on emergence of resistance to gentamicin, which occurred with similar frequency in the two study arms (three amoxycillin and two placebo participants). No significant increase in resistance of P. aeruginosa isolates to ciprofloxacin was observed in the Serisier 2013b ORBIT study, and likewise in the Liu 2012 study, with no differences between groups observed.

In terms of absolute treatment differences, 5 of 100 people in the control group had emergent antibiotic resistance over 1.5 to 24 months, compared with 15 (95% CI 8 to 27) of 100 in the prolonged antibiotics arm (Figure 6).

Figure 6.

Absolute treatment differences for emergence of antibiotic resistance with events occurring in 47 per 1000 in the control arm and in 148 per 1000 (95% CI 56 to 334) on the intervention.

Bacterial colonisation

A total of 15 studies with 809 participants (403 given intervention and 406 placebo) reported on this outcome (Altenburg 2013; Barker 2000; Currie 1990; Cymbala 2005; De Diego 2013; Drobnic 2005; Murray 2011; Orriols 1999; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Valery 2013; Wilson 2013a; Wong 2012; Yalçin 2006). Given the expected nature of a trial including participants with multiple organisms from recurrence/regrowth/colonisation/active infection, a meta-analysis was not performed. In the Altenburg 2013 study, the microbiological profile did not differ significantly between groups at baseline or after 52 weeks of treatment. A total of 437 sputum samples were cultured for microbiology; this yielded one or more pathogens on 339 occasions. Haemophilus influenzae, Streptococcus pneumoniae, Staphylococcus aureus, Moraxella catarrhalis and Haemophilus parainfluenzae were most frequently encountered, together accounting for 87% of the total number of pathogens. All participants with Haemophilus influenzae, Pseudomonas aeruginosa and Aspergillus at baseline from the De Diego 2013 study became culture negative after taking azithromycin for three months; however, study authors did not report the change in bacterial isolates in the placebo arm. Serisier 2013a BLESS reported no differences between the two groups for emergence of new sputum pathogens at any time after the start of erythromycin; however, eradication of sputum pathogens occurred in more erythromycin-treated participants (n = 17 (30.4%) vs n = 6 placebo participants (10.9%); OR 3.6, 95% CI 1.3 to 10.6; P value = 0.01). Serisier 2013b ORBIT also reported on new pathogens isolated from 12 participants given placebo and 9 taking ciprofloxacin. Wong 2012 reported that 6 of 46 participants taking azithromycin cultured new organisms, along with 10 of 45 participants given placebo. Disappearance of P. aeruginosa in sputum was observed in four participants taking tobramycin and in four participants given placebo in Drobnic 2005;however, this cross-over study was transient and P. aeruginosa re-grew after a mean time of three months. Examination of individual microbiological responses among participants in the Barker 2000 study showed that one-third (13 of 37) of tobramycin-treated individuals had P. aeruginosa eradicated from their sputum. Twelve of these 13 participants were assessed as having an improved medical condition at week 6, and an additional third (12 of 37) showed a reduction of at least 2 log10 in P. aeruginosa density at week 4. Nine of these 12 participants were assessed as improved. The last third (12 of 37) of tobramycin-treated participants had no microbiological response, and 10 of these individuals were not improved. In contrast, 33 of 35 (94%) participants given placebo had no microbiological response, and only 2 of 35 (6%) had a greater than 2 log10 decrease in P. aeruginosa density. Orriols 1999 reported that five of seven individuals taking ceftazidime and tobramycin and six of eight given placebo cultured P. aeruginosa. In the Tsang 1999 study, Pseudomonas aeruginosa and Haemophilus influenzae were isolated from the sputum in 10 participants and 1 participant taking erythromycin, respectively, and Pseudomonas aeruginosa, Haemophilus influenzae, Klebsiella pneumoniae and Escherichia coli were isolated in the sputum of six, two, one and one participants given placebo, respectively. Currie 1990 with 38 participants reported Haemophilus influenzae, P. aeruginosa and Staphylococcus aureus at baseline and subsequently several gram-negative bacilli but concluded that no important changes in bacterial flora were observed in either group.

8. Exercise capacity

Exercise capacity was reported in four studies (Murray 2011; Serisier 2013a BLESS; Serisier 2013b ORBIT; Wong 2012) and was objectively assessed in 365 participants using the six-minute walk test (6MWT). Three of these studies (Serisier 2013a BLESS; Serisier 2013b ORBIT; Wong 2012) reported no statistically significant differences in the 6MWT between intervention and placebo groups at the end of the intervention period. Murray 2011 reported no statistically significant differences in either group until month 12 of the trial, when improvement in exercise capacity was noted (mean 510 m, 95% CI 350 to 690 vs mean 415 m, 95% CI 267.5 to 530; P value = 0.03). These effects were not sustained at follow-up. A meta-analysis of two studies produced no evidence of any effect during primary analysis (SMD 0.18, 95% CI -0.22 to 0.58; Analysis 1.10) or sensitivity analysis (Analysis 2.7).

9. Quality of life

Quality of life measure as measured on the St George Respiratory Questionnaire (SGRQ) was reported in 10 studies involving 667 participants (Altenburg 2013; Cymbala 2005; De Diego 2013; Drobnic 2005; Liu 2012; Murray 2011; Serisier 2013a BLESS; Serisier 2013b ORBIT; Wilson 2013a; Wong 2012). Six of these studies (Altenburg 2013; Cymbala 2005; De Diego 2013; Liu 2012; Murray 2011; Wilson 2013a) demonstrated significant improvement in quality of life among participants with bronchiectasis who were given long-term antibiotic treatment when compared with those given placebo. De Diego 2013 reported clinically and statistically significant changes in quality of life in the antibiotic group (-7.9 ± 3.1 vs 4.1 ± 3.8; P value < 0.05) but no difference in SGRQ scores at six months. A meta-analysis on total SGRQ scores from five parallel trials with 161 participants given intervention and 154 controls found a non-statistically significant benefit favouring the intervention (MD -2.75, 95% CI -7.08 to -1.57; I² = 67%). On sensitivity analysis and re-evaluation based on a fixed-effect model, results remained non-statistically significant (P value = 0.06; Analysis 2.8). Following re-analysis based on the random-effects model, results were no longer statistically significant (MD -3.71, 95% CI -9.84 to 2.41; I2 = 90%; Analysis 1.11).

Among the cross-over studies, Cymbala 2005 reported that 82% of participants receiving long-term antibiotics had improved quality of life; however, formal objective measures were not employed. On the other hand, Drobnic 2005 showed no significant differences in quality of life (via SGRQ) between treatment and placebo arms at the end of the trial (mean –0.90 ± 3.93 vs –0.83 ± 6.89).

Discussion

Summary of main results

The primary aim of this review was to assess the effects of prolonged antibiotic therapy in adults and children with non-cystic fibrosis bronchiectasis. We reviewed the evidence from 18 trials (1162 participants) between 1957 and 2013. Our primary outcome, exacerbation, was reported differently by 15 studies involving 925 participants and was significantly reduced by the intervention. However, the baseline exacerbation rate was adjusted (by individual study authors) in 7 (Altenburg 2013; De Diego 2013; Serisier 2013a BLESS; Serisier 2013b ORBIT; Tsang 1999; Wilson 2013a; Wong 2012) out of 18 trials. Other influential factors such as Pseudomonas colonisation, smoking history and lobar involvement were variably reported and adjusted. All participants from five studies (Barker 2000; Drobnic 2005; Orriols 1999; Serisier 2013b ORBIT; Wilson 2013a) had Pseudomonas aeruginosa at baseline. The effect of prolonged antibiotic therapy on hospitalisation was favourable across six of seven studies with 643 participants (Altenburg 2013; Serisier 2013a BLESS; Valery 2013; Wilson 2013a; Wong 2012; Orriols 1999) (Analysis 1.2), although results were not statistically significant. The baseline rate of hospitalisation for these studies was not described adequately, limiting the reliability of change scores from baseline. Participants included in this review were drawn from diverse populations. It does not necessarily follow that patients with differing anatomical distribution or aetiologically distinct underlying diseases or baseline characteristics will respond in a homogenous way to the same intervention. No significant effect on mortality, exercise capacity, lung function or quality of life was reported. However, follow-up plans of the included studies ranging from 4 to 96 weeks suggest that the effect on mortality may not be conclusive.

Overall completeness and applicability of evidence

Despite an extensive literature search and relatively high burden of disease in the community, only 18 studies met the inclusion criteria of this review. The primary aim of this review was to address the possible benefit of prolonged antibiotic therapy for individuals with non-cystic fibrosis bronchiectasis.

The trials were of differing duration. We arbitrarily chose a cutoff of four weeks of treatment for eligibility because this duration was longer than the standard 7 to 10 day course of treatment traditionally provided for uncomplicated exacerbations. Furthermore, long-term use of antibiotics was defined as use of this treatment with a prophylactic intent in patients at a chronic stable phase (i.e. not for exacerbations). The studies reported by Wilson 2013a (four weeks), Barker 2000 (four weeks), De Diego 2013 (12 weeks), Koh 1997 (12 weeks) and Tsang 1999 (eight weeks) were probably too brief to allow firm conclusions about disease-modifying effects.

Inclusion of six trials using inhaled antibiotics (Barker 2000; Drobnic 2005; Murray 2011; Orriols 1999; Serisier 2013b ORBIT; Wilson 2013a) is also worthy of consideration. The mechanism of action of inhaled antibiotics may differ from that of antibiotics given by the oral route; however, increased withdrawals due to severe adverse events or intolerable effects were noted in both study arms when compared with oral interventions.

In conclusion, 15 trials in this review demonstrate a trend supporting use of prolonged antibiotics in the management of bronchiectasis with significant risk of emergence of drug resistance. Therefore, appropriate patient selection is essential. The intervention was safe and tolerated by study participants, particularly when the oral route was used. Trials exhibited significant shortcomings, particularly in terms of study duration, baseline exacerbation, hospitalisation and bacterial colonisation/isolation, hence further research is required. These new studies should incorporate methods allowing separate study of specific issues such as Pseudomonas aeruginosa colonisation and drug resistance at individual and community levels.

Quality of the evidence

The overall quality of studies included in this review was moderate. Most included studies reported that they blinded assessors and participants from allocation; however not enough information was provided for this assessment. In addition, information was insufficient to assess attrition bias, especially with reporting bias (Cymbala 2005; Drobnic 2005; MRC 1957; Serisier 2013b ORBIT; Yalçin 2006). Small sample sizes further limited the ability of study authors to draw reliable conclusions regarding efficacy, tolerability and safety.

Attempts to outline a mechanism for clinical effects (other than sputum volume, purulence, cultures and resistance patterns) were limited to the studies reported by Tsang 1999 and Orriols 1999. In the former investigation, no change was seen in sputum cytokines or leukotrienes despite improvements in lung function and sputum volume. Orriols 1999 showed a reduction in airway responsiveness and suggested that this might have occurred through indirect effects on airway inflammation.

In general, the trials suggest that long-term antibiotics in bronchiectasis are an effective intervention for reducing sputum volume and purulence, but they have a very limited impact on the natural history of the condition. However, several important issues related to this review weaken the power of review authors to draw final conclusions; these will be addressed in turn.

Potential biases in the review process

No significant biases were anticipated or were found to occur during the review process. Criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions were strictly followed to limit the potential for bias during screening, data extraction and analyses of included studies. Risk of bias was assessed independently by two review authors (KH and CN), who resolved conflicts by discussion with a third review author (KVC). We contacted the corresponding author of two included studies (Altenburg 2013; Serisier 2013a BLESS; Serisier 2013b ORBIT) and the authors of the three initially unclassified citations to ask for raw data and clarification of methodological techniques. The review authors of this meta-analysis have reported no conflicts of interest.

Agreements and disagreements with other studies or reviews

Emergence of drug resistance was reported by Barker 2000 and was repeatedly reproduced by newer studies with follow-up longer than six months (Analysis 1.9). Widespread use of antibiotics poses a high likelihood of inducing substantial population-level antibiotic resistance in a range of micro-organisms (Serisier 2014). In particular, the long-acting agent azithromycin rapidly induces sustained macrolide resistance, and its introduction in the USA in 1992 was temporally associated with a substantial increase in rates of macrolide resistance in pneumococcal isolates (Serisier 2013). This review was limited to emergence of drug resistance among participants; however, available evidence from this review supports the imperative importance of patient selection for intervention, especially with long-acting antibiotics. The process of patient selection can be assisted by the bronchiectasis severity index (BSI), which is based on prior hospital admission(s), Medical Research Council dyspnoea score greater than or equal to 4, forced expiratory volume in one second (FEV1) < 30% predicted, Pseudomonas aeruginosa colonisation, colonisation with other pathogenic organisms and three or more lobes involved as seen on high-resolution computed tomography (Chalmers 2013), which was validated across five cohorts (1310 participants) with four-year follow-up. Similar clinical domains were based on another score called "FACED" (F for forced expiratory volume in one second; A for Age; C for Pseudomonas colonisation; E for extent of bronchiectasis counting lobar involvement on CT scan; and D for dyspnoea, using the Medical Research Council dyspnoea score proposed by Martinez-Garcia et al, with an intent to predict five-year mortality in mild, moderate and severe disease (defined as scores of 0 to 2, 3 to 4 and 5 to 7, respectively) of 4%, 25% and 56%, respectively (Martinez-Garcia 2014; Saleh 2014). Moreover, the development of multi-drug-resistant organisms among patients with drug allergy will add extra burden. Therefore, in clinical practice, conscientious selection of an appropriate intervention for each patient is of utmost importance.

Another recent review of long-term macrolides for non-cystic fibrosis bronchiectasis found that macrolides were a reliable treatment option for individuals with stable bronchiectasis, as revealed by available data (Wu 2014). However, similar to our review, for outcomes such as macrolide resistance, meta-analysis of all included studies was not possible because outcome data were incompletely reported. Authors of the Wu 2014 review conclude that their results justify further investigation of macrolides in combination with usual treatment regimens for bronchiectasis.

Authors' conclusions

Implications for practice

This review has demonstrated a positive effect for prolonged antibiotics in bronchiectasis at the cost of emergence of resistance. Available data support the use of prolonged courses of antibiotics in selected patients.

Implications for research

Further trials are required to examine this question in relation to issues surrounding the management of Pseudomonas aeruginosa colonisation, which should be addressed separately, as should the question of route of administration. A non-antibacterial effect of macrolide antibiotics should also be sought. Possible adverse effects intervention based on route of administration of the intervention, particularly patient tolerance, should be monitored in future studies. These new trials should be appropriately powered and of adequate duration. They should have standardised endpoints and should report baseline characteristics of participants, particularly extent of disease (lung function, radiological extent/lobes involved), disease burden (baseline exacerbation, hospitalisation, sputum volume and purulence), smoking status and co-morbidities.

Acknowledgements

The authors would like to acknowledge all the staff at Cochrane Airways Group, in particular Emma Welsh, Elizabeth Stovold, Emma Jackson and Chris Cates for their ongoing help and support. We also wish to thank peer reviewers for their time and feedback. Thank you also to Alfredo De Diego and David Serisier for supplying additional information about the De Diego 2013 and Serisier 2013a BLESS and Serisier 2013b ORBIT studies.

The Background and Methods sections of this review are based on a standard template used by the Cochrane Airways Group.

Data and analyses

Download statistical data

Comparison 1. Continuous antibiotics versus standard treatment with or without added placebo (parallel groups)
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Exacerbations13 Odds Ratio (Random, 95% CI)0.31 [0.19, 0.52]
2 Hospitalisations7643Odds Ratio (M-H, Random, 95% CI)0.40 [0.14, 1.11]
3 Response rates2110Peto Odds Ratio (Peto, Fixed, 95% CI)0.30 [0.14, 0.63]
4 Lung function8 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
4.1 FEV1 % predicted5338Std. Mean Difference (IV, Random, 95% CI)-0.12 [-0.34, 0.10]
4.2 Change in FEV1384Std. Mean Difference (IV, Random, 95% CI)-0.07 [-0.84, 0.70]
4.3 FEV1236Std. Mean Difference (IV, Random, 95% CI)-0.36 [-1.03, 0.30]
5 Sputum leucocytes3165Mean Difference (IV, Random, 95% CI)-0.39 [-0.87, 0.09]
6 Erythrocyte sedimentation rate (ESR)287Mean Difference (IV, Random, 95% CI)-2.47 [-9.76, 4.82]
7 Adverse events10 Peto Odds Ratio (Peto, Fixed, 95% CI)Subtotals only
7.1 Withdrawals (intolerable side effects)10683Peto Odds Ratio (Peto, Fixed, 95% CI)0.91 [0.56, 1.49]
7.2 Diarrhoea3231Peto Odds Ratio (Peto, Fixed, 95% CI)3.33 [1.50, 7.37]
7.3 Rash3140Peto Odds Ratio (Peto, Fixed, 95% CI)1.98 [0.67, 5.80]
7.4 Wheeze174Peto Odds Ratio (Peto, Fixed, 95% CI)8.56 [1.63, 44.93]
7.5 Dyspnoea174Peto Odds Ratio (Peto, Fixed, 95% CI)4.41 [1.43, 13.61]
7.6 Chest pain/palpitations3274Peto Odds Ratio (Peto, Fixed, 95% CI)5.21 [1.35, 20.14]
7.7 Increased cough2198Peto Odds Ratio (Peto, Fixed, 95% CI)1.12 [0.48, 2.61]
7.8 Nausea4275Peto Odds Ratio (Peto, Fixed, 95% CI)1.01 [0.45, 2.26]
7.9 Fatigue2157Peto Odds Ratio (Peto, Fixed, 95% CI)0.99 [0.30, 3.33]
7.10 Hemoptysis3310Peto Odds Ratio (Peto, Fixed, 95% CI)0.87 [0.45, 1.70]
7.11 Fever174Peto Odds Ratio (Peto, Fixed, 95% CI)0.63 [0.17, 2.38]
7.12 Increased sputum174Peto Odds Ratio (Peto, Fixed, 95% CI)1.74 [0.53, 5.71]
7.13 Abdominal pain2195Peto Odds Ratio (Peto, Fixed, 95% CI)5.10 [1.38, 18.86]
7.14 Headache2122Peto Odds Ratio (Peto, Fixed, 95% CI)0.22 [0.05, 1.01]
7.15 Bronchospasm2189Peto Odds Ratio (Peto, Fixed, 95% CI)2.19 [0.76, 6.33]
7.16 Vomiting1112Peto Odds Ratio (Peto, Fixed, 95% CI)4.37 [0.07, 290.15]
7.17 Acute sinusitis1117Peto Odds Ratio (Peto, Fixed, 95% CI)0.50 [0.05, 4.89]
8 Deaths7595Peto Odds Ratio (Peto, Fixed, 95% CI)1.48 [0.28, 7.85]
9 Emergence of resistance6431Odds Ratio (M-H, Random, 95% CI)3.48 [1.20, 10.07]
10 Exercise capacity (6MWD)296Std. Mean Difference (IV, Random, 95% CI)0.18 [-0.22, 0.58]
11 Change in St George Respiratory Questionnaire5 Mean Difference (IV, Random, 95% CI)Subtotals only
11.1 Symptoms2147Mean Difference (IV, Random, 95% CI)-14.16 [-40.22, 11.90]
11.2 Total5315Mean Difference (IV, Random, 95% CI)-2.75 [-7.08, 1.57]
12 Number of participants with exacerbations10654Odds Ratio (M-H, Random, 95% CI)0.36 [0.19, 0.69]
13 Exacerbation rates - continuous3230Std. Mean Difference (IV, Random, 95% CI)-0.80 [-1.32, -0.29]
Analysis 1.1.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 1 Exacerbations.

Analysis 1.2.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 2 Hospitalisations.

Analysis 1.3.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 3 Response rates.

Analysis 1.4.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 4 Lung function.

Analysis 1.5.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 5 Sputum leucocytes.

Analysis 1.6.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 6 Erythrocyte sedimentation rate (ESR).

Analysis 1.7.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 7 Adverse events.

Analysis 1.8.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 8 Deaths.

Analysis 1.9.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 9 Emergence of resistance.

Analysis 1.10.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 10 Exercise capacity (6MWD).

Analysis 1.11.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 11 Change in St George Respiratory Questionnaire.

Analysis 1.12.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 12 Number of participants with exacerbations.

Analysis 1.13.

Comparison 1 Continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 13 Exacerbation rates - continuous.

Comparison 2. Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups)
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Exacerbations13 Odds Ratio (Fixed, 95% CI)0.34 [0.25, 0.46]
2 Hospitalisations7643Odds Ratio (M-H, Fixed, 95% CI)0.44 [0.23, 0.87]
3 Lung function8 Std. Mean Difference (IV, Fixed, 95% CI)Subtotals only
3.1 FEV1 % predicted5338Std. Mean Difference (IV, Fixed, 95% CI)-0.12 [-0.34, 0.09]
3.2 Change in FEV1384Std. Mean Difference (IV, Fixed, 95% CI)-0.03 [-0.47, 0.40]
3.3 FEV1236Std. Mean Difference (IV, Fixed, 95% CI)-0.36 [-1.03, 0.30]
4 Sputum leucocytes3165Mean Difference (IV, Fixed, 95% CI)-0.39 [-0.87, 0.09]
5 Erythrocyte sedimentation rate (ESR)287Mean Difference (IV, Fixed, 95% CI)-3.04 [-8.76, 2.68]
6 Emergence of resistance6431Odds Ratio (M-H, Fixed, 95% CI)3.65 [1.74, 7.64]
7 Exercise capacity (6MWD)296Std. Mean Difference (IV, Fixed, 95% CI)0.18 [-0.22, 0.58]
8 Change in St George Respiratory Questionnaire5 Mean Difference (IV, Fixed, 95% CI)Subtotals only
8.1 Symptoms2147Mean Difference (IV, Fixed, 95% CI)-6.19 [-14.32, 1.95]
8.2 Total5315Mean Difference (IV, Fixed, 95% CI)-2.29 [-4.69, 0.10]
Analysis 2.1.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 1 Exacerbations.

Analysis 2.2.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 2 Hospitalisations.

Analysis 2.3.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 3 Lung function.

Analysis 2.4.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 4 Sputum leucocytes.

Analysis 2.5.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 5 Erythrocyte sedimentation rate (ESR).

Analysis 2.6.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 6 Emergence of resistance.

Analysis 2.7.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 7 Exercise capacity (6MWD).

Analysis 2.8.

Comparison 2 Sensitivity analysis (fixed effect): continuous antibiotics versus standard treatment with or without added placebo (parallel groups), Outcome 8 Change in St George Respiratory Questionnaire.

Appendices

Appendix 1. Sources and search methods for the Cochrane Airways Group Specialised Register (CAGR)

Electronic searches: core databases

Database Frequency of search
CENTRALMonthly
MEDLINE (Ovid)Weekly
EMBASE (Ovid)Weekly
PsycINFO (Ovid)Monthly
CINAHL (EBSCO)Monthly
AMED (EBSCO)Monthly

Handsearches: core respiratory conference abstracts

Conference Years searched
American Academy of Allergy, Asthma and Immunology (AAAAI)2001 onwards
American Thoracic Society (ATS)2001 onwards
Asia Pacific Society of Respirology (APSR)2004 onwards
British Thoracic Society Winter Meeting (BTS)2000 onwards
Chest Meeting2003 onwards
European Respiratory Society (ERS)1992, 1994, 2000 onwards
International Primary Care Respiratory Group Congress (IPCRG)2002 onwards
Thoracic Society of Australia and New Zealand (TSANZ)1999 onwards

MEDLINE search strategy used to identify trials for the CAGR

Bronchiectasis search

1. exp Bronchiectasis/

2. bronchiect$.mp.

3. bronchoect$.mp.

4. kartagener$.mp.

5. (ciliary adj3 dyskinesia).mp.

6. (bronchial$ adj3 dilat$).mp.

7. or/1-6

Filter to identify RCTs

1. exp "clinical trial [publication type]"/

2. (randomized or randomised).ab,ti.

3. placebo.ab,ti.

4. dt.fs.

5. randomly.ab,ti.

6. trial.ab,ti.

7. groups.ab,ti.

8. or/1-7

9. Animals/

10. Humans/

11. 9 not (9 and 10)

12. 8 not 11

The MEDLINE strategy and RCT filter are adapted to identify trials in other electronic databases.

Appendix 2. Search strategy for the Cochrane Airways Group Register

2013 search

#1 BRONCH:MISC1

#2 MeSH DESCRIPTOR Bronchiectasis Explode All

#3 bronchiect*

#4 #1 or #2 or #3

#5 MeSH DESCRIPTOR Anti-Bacterial Agents Explode 1

#6 antibiotic* or anti-biotic*

#7 anti-bacteri* or antibacteri*

#8 *cillin

#9 *mycin

#10 *oxacin

#11 *tetracycline

#12 macrolide*

#13 quinolone*

#14 trimethoprim

#15 ceph*

#16 sulpha*

#17 #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16

#18 #4 and #17

[In search line #1, MISC1 denotes the field where the reference has been coded for condition, in this case, bronchiectasis]

Previous search (used until 2011)

*cillin OR antibiotic* OR *tetracycline OR *mycin OR macrolide* OR quinolone* OR *oxacin OR trimethoprim OR *sulpha OR *ceph or anti-bacteri* or "anti bacteri*"

[Limited to bronchiectasis records]

What's new

DateEventDescription
9 August 2017AmendedNew literature search run to assess the need to update this review. Twenty eight potentially eligible studies identified and added to Studies awaiting classification.

History

Protocol first published: Issue 2, 2000
Review first published: Issue 4, 2003

DateEventDescription
21 February 2014New search has been performedLiterature search run
21 February 2014New citation required and conclusions have changed9 additional studies identified (total now: n = 18); stronger evidence base available with clear benefit in terms of reduced hospitalisation and exacerbations
14 January 2011New search has been performedNew literature search run. No new studies identified. Minor copyedits made to 'Risk of bias' table and layout of review
18 August 2008AmendedConverted to new review format
4 January 2008New search has been performedLiterature search re-run. New excluded study: Bilton 2006
20 February 2007New citation required and conclusions have changed

New studies: Cymballa 2005; Drobnic 2005; Yalcin 2006

These 3 studies used a cross-over design and brought additional data on quality of life, lung function, exacerbation rates and adverse events

As differences in the design of these studies prevented us from pooling their findings with those of the other trials included in the review, they have not had a significant impact on the conclusions of the review

Additional trials needed in children

Contributions of authors

KH updated the protocol, selected RCTs for inclusion, assessed the quality of papers and risk of bias, extracted data, described the studies, conducted data analysis and assisted in manuscript preparation.

CN selected RCTs for inclusion, assessed the quality of papers, performed risk of bias assessments and extracted data.

KVC updated the protocol, performed data analysis and assisted in manuscript preparation.

DE updated the previous version of this review and were involved in manuscript preparation for this update.

BS updated the protocol and assisted in manuscript preparation.

Declarations of interest

None known.

Sources of support

Internal sources

  • Department of Respiratory and Sleep Service, Flinders Medical Centre, Australia.

  • Department of Respiratory Medicine, The Queen Elizabeth Hospital, Australia.

External sources

  • No sources of support supplied

Differences between protocol and review

The presence of clinical heterogeneity between treatments used in this review precluded the assumption of a fixed-effect model (with a single treatment effect for all studies). As such, the primary analyses were changed to a random-effects model, with a fixed-effect model included as a sensitivity analysis.

The small number of paediatric studies and the presence of high heterogeneity between them led to the decision that pre-specified subgroup analyses for paediatric versus adult populations would not be performed.

The order of the outcomes was changed. Acute exacerbations and hospitalisations was added as a primary outcome. Methods were updated to the latest Cochrane methodological expectations, and a 'Summary of findings' table, a 'Characteristics of included studies' table and a table reporting types of interventions and status of Pseudomonas aeruginosa were added.

Notes

We selected outcomes for Summary of findings for the main comparison on the basis of what respiratory clinicians (KH, CN and BJS) believed to be the most relevant clinical outcomes that would likely be used to inform patient care.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Altenburg 2013

Methods

Country: the Netherlands

Design: randomised double-blind placebo-controlled trial

Objective/aim: to determine efficacy of macrolide maintenance treatment for adults with non-cystic fibrosis bronchiectasis

Study site: outpatient

Methods of analysis:

  1. Intention-to-treat

Participants

Eligible for study: 360

Randomly assigned: 89

Completed: intervention 42, control 39

Age: intervention 59.9 ± 12.3, control 64.6 ± 9.1

Gender (M/F): intervention 18/25, control 12/28

Bronchiectasis diagnosis criteria:

  1. 18 years or older and non-CF bronchiectasis diagnosed by plain bronchography or HRCT

  2. Minimum of 3 lower respiratory tract infections treated with oral or intravenous antibiotics in preceding year with at least 1 sputum culture yielding 1 or more bacterial respiratory pathogens in the year before study entry

Recruitment: outpatient clinics

Co-morbidities: not mentioned

Reasons for participantexclusion:

  1. Excluded if patients received prolonged (> 4 weeks) macrolide therapy during previous 3 months, oral or IV course of corticosteroids within 30 days of screening or any antimicrobial treatment for an LRTI in the past 2 weeks

  2. Known allergy or intolerance to macrolides

  3. Women with childbearing potential avoiding contraceptives, as well as lactating women

  4. Patients with liver disease or with elevated transaminase levels

Baseline FEV: intervention 77.7 (24.4% predicted), control 82.7 (27.2% predicted)

Baseline exacerbations: intervention 4 (3-9), control 5.0 (3-12)

Interventions

Setting: outpatient

Duration of intervention: 52 weeks

Type of antibiotic(s): azithromycin (250 mg daily)

Type of control: placebo

Outcomes

Pre-specified outcomes:

Primary:

  1. Number of infectious exacerbations during 12 months of treatment

Secondary:

  1. Lung function

  2. Sputum bacteriology

  3. Inflammatory markers

  4. Adverse effects

  5. Symptom scores

  6. Quality of life

Follow-up period: 90 days after 12 months of treatment

NotesThe study was supported by an unrestricted research grant from GlaxoSmithKline. Teva Netherlands supplied the azithromycin tablets used in this study
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated allocation sequence was used
Allocation concealment (selection bias)Low riskCentrally controlled allocation was used
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIdentical process for both placebo and intervention. Clearly defined instructions and protocol for participants and personnel
Blinding of outcome assessment (detection bias)
All outcomes
Low risk
  1. Unlikely to have blinding broken

  2. Clear protocol available to differentiate exacerbation

  3. Treating physicians during exacerbation were not investigators and were blinded

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskInsufficient information
Selective reporting (reporting bias)Low riskReported pre-specified outcomes
Other biasUnclear riskNo other biases identified

Barker 2000

Methods

Country: United States of America

Design: placebo-controlled double-blind randomised study

Objective/aim: to evaluate microbiological efficacy and safety of inhaled tobramycin for treatment of patients with bronchiectasis and Pseudomonas aeruginosa

Study site: homes and clinics

Methods of analysis:

  1. Logistic regression

  2. Wilcoxon signed rank test (treatment group for assessing airway reactivity)

Participants

Eligible for study: 78

Randomly assigned: 74

Completed: intervention 31, control 29

Age: intervention 66.6 ± 13, control 63.2 ± 13.5

Gender (M/F): intervention 14/23, placebo 15/22

Bronchiectasis diagnosis criteria: confirmed by high-resolution computer tomography

Inclusion criteria: grossly purulent sputum containing ≥ 104 cfu/g P. aeruginosa

Recruitment: 16 sites

Co-morbidities: nil

Reasons for participant exclusion:

  1. Sputum contained fewer than 104 cfu/g P. aeruginosa

  2. Diagnosis of cystic fibrosis, allergic bronchopulmonary aspergillosis, acute pulmonary process requiring medical intervention indicated by new infiltrate on CXR, significant recent haemoptysis or received antibiotics within 2 weeks of screening visit

Baseline FEV: intervention 56.2 (21.2), control 53.3 (22.1)

Baseline exacerbations: intervention 5, control 1

Interventions

Setting: hospital clinics

Duration of intervention: 4 weeks

Type of antibiotic(s): 300 mg nebulised tobramycin

Type of control: matched placebo (1.25 mg quinine sulfate)

Outcomes

Pre-specified outcomes:

Primary endpoints:

  1. Change in P. aeruginosa density from baseline to 4 weeks

Secondary endpoints:

  1. Change in P. aeruginosa density from baseline value to week 2 and week 6

  2. Investigators' subjective assessment of change in participant's general medical condition

  3. Per cent change in FEV1 percent predicted and in FVC per cent predicted from week 0 to week 4

  4. Safety endpoints including incidence of adverse events, change in serum chemistry and haematological measures and airway reactivity

Follow-up: 6 weeks

NotesHowever, paper also states that this was an 8-week study that was sponsored by PathoGenesis Corporation, Seattle, WA, USA
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskParticipants assigned in blocks of 2 parallel groups
Allocation concealment (selection bias)Unclear riskInsufficient information available to reveal adequacy of allocation concealment
Blinding of participants and personnel (performance bias)
All outcomes
Low risk

Both investigators and participants blinded

Drugs chosen (similar taste, pH)

Self-administered via same neb

Blinding of outcome assessment (detection bias)
All outcomes
Low risk

Both investigators and participants blinded

Drugs chosen (similar taste, pH)

Self administered via same neb

Although double-blinded study with same follow-up protocol, source of serum drug level not mentioned clearly (e.g. from both groups vs from intervention)

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo information regarding missing data but missing data/withdrawn numbers apparently balanced between the 2 groups
Selective reporting (reporting bias)Unclear riskReported pre-specified outcomes but some numerical data not mentioned
Other biasLow riskNo other biases identified

Currie 1990

Methods

Country: United Kingdom

Design: randomised double-blind placebo-controlled parallel-group study

Objective/aim: to determine the value of prolonged course of higher-dose oral amoxycillin in a double-blind randomised placebo-controlled study

Study site: clinics and homes

Methods of analysis:

  1. Wilcoxon rank sum test, χ2 test and Kendall’s rank correlation

  2. Group results were expressed as medians (because of the distribution of results)

Participants

Eligible for study: 69

Randomly assigned: intervention 19, control 19

Completed: intervention 15, control 14

Age: intervention 54 (median), control 51 (median)

Gender (M/F): intervention 9/6, control 10/4

Bronchiectasis diagnosis criteria: daily purulent sputum expectoration for at least 3 months and definitive evidence of bronchiectasis on bronchogram or plain chest radiograph

Recruitment: via hospital admissions and clinics

Co-morbidities: not included

Reasons for participant exclusion: < 16 y/o, pregnant or planning pregnancy, on OCP, on continuous higher-dose oral antibiotics or nebulised antibiotics during previous month, on continuous oral higher-dose amoxycillin (6 g OD) for > 1 month during previous 3 months, major alterations in other therapy in previous 8 weeks, evidence of allergic bronchopulmonary aspergillosis, primary ciliary dyskinesia, CF, sarcoidosis, NIDDM, cardiac failure, low daily sputum volume, haemoptysis, too ill, likely to be poor complier

Baseline FEV: intervention 1.9 (0.9), 54% (median), control 1.6 (0.9), 51% (median)

Baseline exacerbations (3 or more in the past year): intervention 11, control 15

Interventions

Setting: outpatient (specialised bronchiectasis clinic in tertiary referral centre)

Duration of intervention: 32 weeks

Type of antibiotic(s): amoxycillin (6 g daily)

Type of control: placebo

Outcomes

Pre-specified outcomes:

  1. Sputum bacteriology

  2. Host inflammatory activity

  3. Extent and severity of lung damage (radiology, pulmonary function tests)

  4. Screening for adverse effects

Follow-up period: 12 months

NotesStudy was supported by the Chest, Heart and Stroke Association
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation not mentioned
Allocation concealment (selection bias)Unclear riskNot mentioned
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blinded; packaging of drugs matched and same protocol used for both groups
Blinding of outcome assessment (detection bias)
All outcomes
Low risk

Blinding unlikely to be broken

Same protocol

Incomplete outcome data (attrition bias)
All outcomes
Low riskReported missing/withdrawal data
Selective reporting (reporting bias)Low riskReported pre-specified outcomes
Other biasLow riskNo other biases identified

Cymbala 2005

Methods

Country: United States of America

Design: open-label, cross-over study

Objective/aim: to determine whether long-term low-dose azithromycin would decrease the number of exacerbations and improve pulmonary function

Study site: clinics and homes

Methods of analysis:

  1. Mann-Whitney rank sum test to compare number of exacerbations between study phases

  2. Bonferroni t-test

  3. Paired t-test

Participants

Eligible for study: 12

Randomly assigned: intervention 9 receiving azithromycin first, then followed by 1 month washout period, then normal treatment regimen

Control 3 started with normal treatment regimen first, then received azithromycin

Completed: intervention 8, control 3

Age: 70.8 ± 9.7 years

Gender: M/F 5/6

Bronchiectasis diagnosis criteria: high-resolution CT scan demonstrating airways larger than accompanying vessels

Recruitment: hospital (teaching, rural), regional health centres

Co-morbidities: not mentioned

Reasons for participant exclusion:

  1. < 18 y/o

  2. Serious intolerance, allergy or sensitivity to azithromycin and macrolides

  3. Deemed unable to follow instructions

Baseline FEV: 48.5 ± 18.9

Baseline exacerbations: not mentioned

Interventions

Setting: homes

Duration of intervention: 6 months

Type of antibiotic(s): azithromycin (500 mg twice weekly)

Type of control: usual treatment regimen

Outcomes

Pre-specified outcomes: not mentioned

Follow-up period: 13 months for azithromycin group, 12 months for regular regimen group

NotesFirst year of study was unfunded, although investigators were fortunate enough to receive donations of study medication from local sales representatives. During second year of study, a small unrestricted stipend was received from manufacturer of azithromycin that covered participant incidentals (i.e. travel expenses, extra pulmonary function tests) only
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskStudy authors reported to use randomisation schedule via random numbers table
Allocation concealment (selection bias)High riskCross-over design without blinding
Blinding of participants and personnel (performance bias)
All outcomes
High risk

No blinding as per study authors (no matching placebo)

No placebo used

Phone calls made to increase compliance during intervention

Blinding of outcome assessment (detection bias)
All outcomes
High risk

No blinding as per study authors

No matching placebo and not clear regarding control arm and its execution

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

Lack of pre-specified outcomes

Diary cards not reported (e.g. QOL mentioned but no data available)

Selective reporting (reporting bias)High risk

Reported only sputum volume and pulmonary function

Pre-specified outcomes not mentioned

Lack of report regarding exacerbation and management of exacerbation

Other biasUnclear riskUnclear regarding other biases due to lack of published protocol

De Diego 2013

Methods

Country: Spain

Design: randomised open-label prospective study

Objective/aim: to explore the effect of long-term therapy with azithromycin with regards to airway oxidative stress markers in exhaled breath condensate (EBC) of adult patients with stable non-cystic fibrosis (CF) bronchiectasis

Study site: homes

Methods of analysis:

  1. Wilcoxon test for quantitative variables

  2. Chi2 test for qualitative variables

Participants

Eligible for study: 40

Randomly assigned: 36

Completed: intervention 16, control 14

Age: intervention 57 ± 11, control 61 ± 12

Gender (M/F): intervention 7/9, control 7/7

Bronchiectasis diagnosis criteria: clinical data and high-resolution computed tomography lung scan

Recruitment: outpatient clinics

Co-morbidities: not mentioned

Reasons for participant exclusion:

  1. Patients with bronchiectasis secondary to CF

  2. Pulmonary surgical processes

  3. Immunodeficiency secondary to human immunodeficiency virus, malignancy, common variable immunodeficiency

  4. Emphysema

  5. Allergic bronchopulmonary aspergillosis or diffuse interstitial pulmonary disease

  6. Known intolerance to macrolides or severe liver disease

Baseline FEV: intervention 1.48 (0.8), 56% (6), control 1.63 (0.7), 68% (7)

Baseline exacerbations: intervention 3.5 (1.8)/last year, control 3.1 (1.3)/last year

Interventions

Setting: outpatient setting

Duration of intervention: 3 months

Type of antibiotic(s): azithromycin (oral 250 mg 3 times a week)

Type of control: not mentioned

Outcomes

Pre-specified outcomes:

Primary outcomes:

  1. Changes in nitric oxide, 8-isoprostane, pH, nitrites and nitrates in exhaled breath condensate

Secondary outcomes:

  1. Changes in exacerbation rates, dyspnoea (Borg scale), sputum volume (mL), sputum colour (15-point scale), bacterial infection, health-related quality of life (St George Respiratory Questionnaire), lung function and radiological extension

Follow-up period: not mentioned

NotesThis work was supported by a grant from Fundacion Valenciana de Neumologia
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation not mentioned
Allocation concealment (selection bias)Unclear riskNo further information available
Blinding of participants and personnel (performance bias)
All outcomes
High riskOpen label without a placebo, no further information available
Blinding of outcome assessment (detection bias)
All outcomes
High riskOpen label intervention and no information on the blinding of an outcome assessment
Incomplete outcome data (attrition bias)
All outcomes
Low risk 
Selective reporting (reporting bias)Low riskReported pre-specified outcomes
Other biasUnclear riskNo other biases identified

Drobnic 2005

Methods

Country: Spain

Design: double-blind placebo-controlled cross-over trial

Objective/aim: to determine whether direct aerosol delivery of tobramycin to lower airways may control infection and produce only low systemic toxicity

Study site: homes and clinics

Methods of analysis:

  1. Student’s t-test

  2. Mann-Whitney U test

  3. χ2test

  4. Fisher’s exact test

Participants

Eligible for study: 60

Randomly assigned: intervention 30, control 30

Completed: intervention 20, control 20

Age: intervention 64.5 (38-75), control 64.5 (38-75)

Gender: not mentioned

Bronchiectasis diagnosis criteria: diagnosed with bronchiectasis confirmed by HRCT at a tertiary referral centre

Recruitment: via clinics, hospital admissions

Co-morbidities: not mentioned

Reasons for participant exclusion:

  1. Tobramycin hypersensitivity

  2. Pseudomonas aeruginosa in sputum resistant to tobramycin

  3. Auditory threshold in either ear > 20 dB at frequencies between 500 and 8000 Hz

  4. Serum creatinine ≥ 1.5 mg/dL

  5. Cystic fibrosis

Baseline FEV1 (predicted): 51.78 ± 16.45

Baseline exacerbation: not mentioned but defined exacerbation

Interventions

Setting: outpatient

Duration of intervention: 6 months

Type of antibiotic(s): tobramycin (nebulised) 100 mg/2 mL was diluted in NaCl 0.9% to make 8 mL

Type of control: nebulised 8 mL of 0.9% NaCl

Outcomes

Pre-specified outcomes: nil mentioned

Follow-up period: 12 months for those who commenced with placebo; 13 months for those who commenced with tobramycin (allowing 1 month washout period)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation not mentioned
Allocation concealment (selection bias)Unclear riskInsufficient information available to determine adequacy of allocation concealment
Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble-blind placebo-controlled cross-over study; although trial was double-blinded, apparent that participants receiving placebo used only 1 type of agent (NaCl). However, 2 agents mixed for participants given intervention (i.e. drug + NaCl)
Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble-blind placebo-controlled cross-over study
Incomplete outcome data (attrition bias)
All outcomes
High risk

Withdrawal and attrition described according to protocol

3 participants with adverse effects excluded from analysis

  1. Unclear whether 10 with missing/incomplete data affected results/analysis

(10 missing/incomplete vs final 20 participants)

Selective reporting (reporting bias)High risk
  1. No data on PA density except for a figure (PA density one of the key areas of interest for the trial)

  2. Report on toxicity, which is one of the key areas, not clear

  3. 3 participants with adverse effects excluded from the analysis

  4. Supplementary use of oral antibiotics not reported

Other biasLow riskNo other biases identified

Koh 1997

Methods

Country: Korea

Design: randomised double-blind parallel placebo-controlled study

Objective/aim: to see whether roxithromycin could reduce the degree of airway responsiveness in bronchiectasis

Study site: clinics

Methods of analysis:

  1. Paired t-test

  2. Wilcoxon signed rank test

  3. ANOVA to analyse variance

Participants

Eligible for study: 27

Randomly assigned: 25

Completed: intervention 13, control 12

Age: intervention 13.3 ± 2.5, control 12.9 ± 2.6

Gender(M/F): intervention 7/6, control 7/5

Bronchiectasis diagnosis criteria: based on clinical features and confirmed by CT with bronchography when necessary

Recruitment: outpatient clinics

Co-morbidities: atopy, asthma (treated with prn beta-agonists)

Reasons for participant exclusion:

  1. Could not perform PFT

  2. PC20 < 25 mg/mL

  3. Clinically unstable bronchiectasis

  4. Antibioic use within 1 month of study

  5. Upper respiratory tract infection within 4 weeks before study

  6. Failed to attain maximal response plateau both before and after treatment

  7. Normal or reduced airway responsiveness

  8. Cystic fibrosis, humoral immune deficiency, bronchopulmonary aspergillosis

Baseline FEV:

  1. Intervention FEV1 max: 42.2 ± 8.3, FEV %: 83 ± 6

  2. Control FEV1 max 39.6 ± 9.3, FEV %: 84 ± 7

Baseline exacerbations: not mentioned

Interventions

Setting: outpatient

Duration of intervention: 12 weeks

Type of antibiotic(s): roxithromycin (4 mg/kg BD)

Type of control: placebo (BD)

Outcomes

Pre-specified outcomes:

  1. Changes in FEV1

  2. Sputum purulence

  3. Sputum leucocyte scores

Follow-up period: 12 weeks

NotesStudy was supported by grant No. 2- 94-112 from the Seoul National University Hospital Research Fund
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo further information available
Allocation concealment (selection bias)Unclear riskNo further information available although stated to be double-blinded
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo further information available
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskAssessors not involved in dividing participants into 2 separate groups
Incomplete outcome data (attrition bias)
All outcomes
Low risk2 patients had exacerbations and given antibiotics. Graphs show reduced number of participants at 12 week assessment.
Selective reporting (reporting bias)Low riskReported pre-specified outcomes
Other biasLow riskNo other biases identified

Liu 2012

Methods

Country: China

Design: randomised controlled trial

Objective/aim:

to investigate impact of treatment with low-dose roxithromycin on clinical symptoms and CT scores in patients with stable bronchiectasis

Study site: not mentioned

Methods of analysis: not mentioned

Participants

Eligible for study: 50

Randomly assigned: not mentioned

Completed: intervention 24, control 22

Age: not mentioned

Gender (M/F): not mentioned

Bronchiectasis diagnosis criteria: not mentioned

Recruitment: not mentioned

Co-morbidities: not mentioned

Reasons for participantexclusion: not mentioned

Baseline FEV: not mentioned

Baseline exacerbations: not mentioned

Interventions

Setting: not mentioned

Duration of intervention: 6 months

Type of antibiotic(s): oral roxithromycin 150 mg daily

Type of control: ambroxol hydrochloride tablet 90 mg 3 times a day

Outcomes

Pre-specified outcomes:

  1. Quality of life (SGRQ)

  2. Degree of dyspnoea (British MRC scale)

  3. Score for CT evaluation of thorax

Follow-up period: not mentioned

NotesAbstract and data from Wu 2014 (long-term macrolides for non-cystic fibrosis bronchiectasis: a systematic review and meta-analysis)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskBased on the abstract only, no specific details available
Allocation concealment (selection bias)Unclear riskBased on the abstract, no specific details available
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskBased on the abstract, no specific details available
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskBased on the abstract, no specific details available
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskBased on the abstract, no specific details available
Selective reporting (reporting bias)Unclear riskBased on the abstract, no specific details available
Other biasUnclear riskBased on the abstract, no specific details available

MRC 1957

Methods

Country: United Kingdom

Design: randomised double-blind placebo-controlled parallel-group study (with 3 arms)

Objective/aim: not mentioned

Study site: not mentioned

Methods of analysis: not mentioned

Participants

Eligible for study: 122

Randomised: 36 (penicillin), 44 (oxytetracycline), 40 (lactose)

Completed: 36 (penicillin), 40 (oxytetracycline), 36 (lactose)

Age: penicillin 34.3, oxytetracycline 32.3, control 32.6.

Gender (M/F): 76/56

Bronchiectasis diagnosis criteria: symptoms for 3 months, frank bronchiectasis affecting at least 2 lung segments on bronchogram

Recruitment: outpatient clinics

Co-morbidities: adjusted but no detail mentioned

Reasons for participant exclusion: tuberculosis, patients receiving antibiotics or sulphonamides in preceding 4 weeks

Baseline FEV: not mentioned

Baseline exacerbations: not mentioned

Interventions

Setting: outpatient

Duration of intervention: 52 weeks

Type of antibiotic(s): oral penicillin or oral oxytetracycline (4 g/wk)

Type of control: oral lactose

Outcomes

Pre-specified outcomes:

  1. 24-Hour specimen of sputum

  2. Observation on severity of cough, dyspnoea, haemoptysis, disability

  3. Final assessment: clubbing, weight, general assessment of progress, participant's own assessment of response to treatment

Follow-up period: 52 weeks

NotesReport by a subcommittee of antibiotics clinical trials (non-tuberculosis), committee of Medical Research Council
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod not mentioned
Allocation concealment (selection bias)Low riskAllocated randomly according to centrally held list
Blinding of participants and personnel (performance bias)
All outcomes
Low riskSame capsules (packaging) used and issued in bottles to participants
Blinding of outcome assessment (detection bias)
All outcomes
Low riskCentrally controlled list of cohort; same protocol for assessments, visits and measurements
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskMissing data not explained/addressed
Selective reporting (reporting bias)High riskThe primary and secondary outcomes to the trial were not clearly defined and reported
Other biasLow riskNo other biases identified

Murray 2011

Methods

Country: Scotland

Design: randomised controlled trial

Objective/aim:

To assess the efficacy of continuous nebulised gentamicin over 1 year in non-cystic fibrosis bronchiectasis

To assess whether treatment effects are sustained over a 3 month treatment-free period

Study site: bronchiectasis clinic

Methods of analysis:

  1. Mann-Whitney U test for unpaired numerical data

  2. Wilcoxon for paired numerical data

  3. Fisher's exact test to compare categorical variables

Participants

Eligible for study: 65

Randomly assigned: intervention 32, control 33

Completed: intervention 27, control 30

Age: intervention 58 (53-67), control 64 (55.7-69)

Gender (M/F): intervention 9/18, control 15/15

Bronchiectasis diagnosis criteria: clinically significant and radiologically proved bronchiectasis (HRCT)

Recruitment: regional bronchiectasis services

Co-morbidities: not mentioned

Reasons for participant exclusion:

Current smokers, cystic fibrosis, active pulmonary mycobacteria, active sarcoidosis; active allergic bronchopulmonary aspergillosis; chronic obstructive pulmonary disease; poorly controlled asthma (> 20% diurnal variation in peak expiratory flows despite treatment); creatinine clearance < 30 mL/min; vestibular instability; previously documented intolerance to aminoglycosides (ototoxicity or nephrotoxicity)

Baseline FEV: intervention 1.8 (1.34-2.31), control 1.61 (1.22-3.41)

Baseline exacerbations: not mentioned

Interventions

Setting: bronchiectasis clinics

Duration of intervention: 12 months

Type of antibiotic(s): nebulised gentamicin 80 mg BD

Type of control: nebulised saline BD

Outcomes

Pre-specified outcomes:

Primary:

  1. ≥ 1 log unit reduction in sputum bacterial density

Secondary:

  1. Qualitative sputum bacteriology

  2. Emergence of gentamicin-resistant Pseudomonas strains

  3. Sputum myeloperoxidase and free neutrophil elastase

  4. 24-Hour sputum volume

  5. Sputum purulence

  6. FEV1, FVC, FEF, mid-expiratory phase

  7. Exercise capacity

  8. Leicesters Cough Questionnaire

  9. St George Respiratory Questionnaire

  10. Exacerbation

  11. Adverse effects

Follow-up period: 3 month treatment-free period (15 months)

NotesAuthor disclosure: M.P.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. J.R.W.G. was a consultant for Transave Corp ($5,001–$10,000) and was on the Board or Advisory Board for Bayer (up to $1,000). He received grant support from Transave Corp and the Cystic Fibrosis Trust (more than $100,001). C.J.D. was a consultant for Transave Corp ($10,001–$50,000). A.J.S. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. T.S.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. J.D.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.P.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.T.H. was on the Board or Advisory Board for Bayer ($1,001–$5,000) and received lecture fees from GlaxoSmithKline (up to $1,000)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskUnclear how random sequence generation happened
Allocation concealment (selection bias)Unclear riskMethods not described; randomisation done by study pharmacist
Blinding of participants and personnel (performance bias)
All outcomes
High riskParticipants unmasked because of funding limitations
Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk

Although all study investigators were masked and bronchiectasis specialist nurse was the contact person, participants were not blinded

Unclear who was responsible for regular assessments and therefore unable to determine likelihood of breaking blinding

Incomplete outcome data (attrition bias)
All outcomes
Low risk

Clear explanation regarding withdrawal from study and missing data

Reasonable extent of missing data

Selective reporting (reporting bias)Low riskReported pre-specified outcomes and addressed missing data
Other biasLow riskNo other biases identified

Orriols 1999

Methods

Country: Spain

Design: prospective randomised non-blinded pilot

Objective/aim: to investigate long-term effectiveness and safety of inhaled antibiotic treatment in non-cystic fibrosis patients with bronchiectasis and chronic infection by Pseudomonas aeruginosa, after standard endovenous and oral therapy for long-term control of infection

Study site: university centres

Methods of analysis:

  1. Fisher's exact test

  2. Mann-Whitney U test or Wilcoxon test

Participants

Eligible for study: 17

Randomly assigned: intervention 8, control 9

Completed: intervention 7, control 8

Age: intervention 62.0 ± 8.5, control 61.4 ± 10.3

Gender (M/F): intervention 6/1, control 4/4

Bronchiectasis diagnosis criteria: bronchography, thoracic computed axial tomography or both

Recruitment: not mentioned

Co-morbidities: not mentioned

Reasons for participant exclusion:

  1. Beta-lactam or aminoglycoside sensitivity

  2. Bacterial resistance on antigram

  3. Kidney failure

Baseline FEV: intervention FEV1: 1037 mL (386), control FEV1: 866 (225)

Baseline exacerbations: not mentioned

Interventions

Setting: outpatient

Duration of intervention: 12 months

Type of antibiotic(s): inhalation tobramycin and ceftazidime BD

Type of control: symptomatic treatment

Outcomes

Pre-specified outcomes:

  1. Number of hospital admissions

  2. Number of hospital days

  3. Oral antibiotic use

  4. FVC, FEV1

  5. PaO2, PaCO2

  6. Drug toxicity

  7. Bacterial resistance

Follow-up period: 12 months

NotesAntibiotic treatment changed to piperacillin and amikacin if clinical impairment and bacterial resistance
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskRandomisation mentioned; however, methods not described
Allocation concealment (selection bias)Unclear riskNot reported
Blinding of participants and personnel (performance bias)
All outcomes
High riskStudy reported to be non-blinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskStudy reported to be non-blinded
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskInsufficient information to permit a judgement
Selective reporting (reporting bias)Unclear riskInsufficient information to permit a judgement
Other biasUnclear riskInsufficient information to permit a judgement

Serisier 2013a BLESS

Methods 
Participants

Eligible for study: 117

Randomly assigned: intervention 59, control 58

Completed: intervention 54, control 53

Age: intervention 61.1 ± 10.5, control 63.5 ± 9.5

Gender (M/F): intervention 21/38, control 25/33

Bronchiectasis diagnosis criteria: bronchiectasis documented by HRCT

Recruitment: regional adult cystic fibrosis centre, other respiratory centres and 2 regional public radio advertisement campaigns

Co-morbidities:

Reasons for participant exclusion:

  1. Cystic fibrosis

  2. Current mycobacterial disease or bronchopulmonary aspergillosis

  3. Any reversible cause for exacerbation

  4. Maintenance oral antibiotic prophylaxis

  5. Prior macrolide use except short term

  6. Changes to medications in preceding 4 weeks

  7. Cigarette smoking within 6 months

  8. Medications or co-morbidities with potential for important interactions with erythromycin

Baseline FEV: intervention: FEV1 (pre-BD) 1.82 (0.65); %pred 66.9 (17.2); FEV1 (post-BD) 1.93 (0.65); %pred 70.2 (17.2); control: FEV1 (pre-BD) 1.83 (0.77); %pred 70.1 (20.3); FEV1 (post-BD) 1.93 (0.79); %pred 73.6 (20.6)

Baseline exacerbations (> 5/y): intervention: 22 (37.3%); control: 20 (34.5%)

Interventions 
Outcomes 
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated randomisation sequence
Allocation concealment (selection bias)Unclear riskCentral allocation; methods of allocation not described
Blinding of participants and personnel (performance bias)
All outcomes
Low riskUnlikely to have blinding broken; identical shape, appearance and taste of erythromycin and placebo tablets
Blinding of outcome assessment (detection bias)
All outcomes
Low riskProtocol created prospectively and all scripts provided centrally
Incomplete outcome data (attrition bias)
All outcomes
Low riskIncomplete data explained including missing data
Selective reporting (reporting bias)Low riskAddressed all pre-specified endpoints from both primary and secondary outcomes
Other biasLow riskNo other biases identified

Serisier 2013b ORBIT

Methods

Country: Australia, New Zealand

Design: Phase II multi-centred randomised double-blind placebo-controlled trial

Objective/aim: to evaluate microbial efficacy of 28 day inhaled dual release ciprofloxacin for inhalation

Study site: not mentioned

Methods of analysis:

  1. Modified intention-to-treat

  2. Kaplan-Meier survival analysis

Participants

Eligible for study: 42

Randomised: intervention 20, control 22

Completed: intervention 18, control 19

Age: intervention 70 ± 5.6, control 59.5 ± 13.2

Gender (M/F): intervention 10/10, control 9/13

Bronchiectasis diagnosis criteria: clinically stable adults with CT scan proven bronchiectasis

Recruitment: not mentioned.

Co-morbidities: IHD, hypertension, diabetes , cerebrovascular disease

Reasons for participant exclusion:

1. Known local or systemic hypersensitivity to fluoroquinolone or quinolone antibiotics

2. Pulmonary exacerbation during screening phase defined as requiring treatment with inhaled, oral or IV antibiotics before first dose of study drugs

3. Diagnosis of CF

4. Diagnosis of allergic bronchopulmonary aspergillosis

5. Received any IV, oral or inhaled antipseudomonal antibiotic within 28 days before visit 1

6. Used tizanidine within 28 days before visit 1

7. Initiated supplemental oxygen within 28 days before visit 1

8. Used any intravenous or intramuscular corticosteroid or had used oral corticosteroid > 10 mg/d or > 20 mg every other day within 28 days of visit 1

9. Changes in treatment regimen or initiation of treatment with any of the following medications within 28 days before visit 1

a. Azithromycin

b. Hypertonic saline

c. Mucolytics

d. Bronchodilator medications

e. Oral corticosteroid

10. Changes in physiotherapy technique or schedule within 28 days before visit 1

11. History of solid organ (e.g. lung) transplantation

12. History of non-tuberculosis mycobacteria requiring treatment within 12 months before visit 1

13. Serum creatinine levels ≥ 1.5 × upper limit of normal (ULN) at screening visit

14. Serum transaminase levels > 3 × ULN at screening visit

15. Febrile illness within 1 week before visit 1

16. Massive haemoptysis (≥ 300 mL or requiring blood transfusion) within 6 months before visit 1

17. Used any over-the-counter product, herbal product, diet aid, hormone supplement, etc., within 7 days before dosing unless approved by both investigator and sponsor

18. Received investigational drug or device within 28 days before visit 1

19. Any serious or active medical or psychiatric illness, which in the opinion of the investigator, would have interfered with participant's treatment assessment, or compliance with protocol

20. History or suspicion of unreliability, poor co-operation or non-compliance with medical treatment

21. Unable to use nebuliser

22. Unable to understand instruction for use of study drugs or to complete QoL questionnaire at visit 1

23. Previously enrolled in this study

24. Pregnant, planned to become pregnant during study, were nursing mothers or were unwilling to use an acceptable method of contraception for duration of the study

Baseline FEV: intervention 1.57 (0.77), control 1.47 (0.53)

Baseline exacerbations: Both groups were stratified by 2-3 or ≥ 4 exacerbations

Interventions

Setting: outpatient clinic

Duration of intervention: 24 weeks

Type of antibiotic(s): dual release ciprofloxacin for inhalation (liposomal ciprofloxacin (150 mg in 3 mL) and free ciprofloxacin (60 mg in 3 mL))

Type of control: placebo/matched placebo: control liposomes (15 mg in 3 mL) and normal saline (0.9% in 3 mL)

Outcomes

Pre-specified outcomes:

Primary endpoints:

  1. Mean change in sputum P. aeruginosa bacterial density

Secondary endpoints:

  1. Safety

  2. Time to first pulmonary exacerbation

  3. FEV

  4. 6MWT

  5. SGRQ

  6. Tolerability

Follow-up period: 24 weeks

NotesThis study was funded by Aradigm Corp.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskStratified by number of exacerbations as per protocol
Allocation concealment (selection bias)Unclear riskStratified by number of exacerbations as per protocol
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskMatched placebo
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskDid not state
Incomplete outcome data (attrition bias)
All outcomes
Low riskIntention-to-treat analysis
Selective reporting (reporting bias)High risk

SGRQ not reported

6MWT not reported

Other biasHigh risk

Number of participants required met

Single participant randomly assigned but not culturing P. aeruginosa

Limitations stated

Moderate sample size: type 1 error

Time to exacerbation only conventionally statistically significant

Age difference between groups

Tsang 1999

Methods

Country: Hong Kong

Design: double-blind randomised placebo-controlled study

Objective/aim: to evaluate effect of low-dose erythromycin on sputum volume and lung function indices in steady state bronchiectasis

Study site: outpatient

Methods of analysis:

  1. ANOVA

  2. Paired Student’s t-test

  3. Wilcoxon rank sum test

Participants

Eligible for study: 24

Randomly assigned: intervention 14, control 10

Completed: intervention 11, control 10

Age: intervention 50 ± 15, control 59 ± 16

Gender (M/F): intervention 3/8, control 2/8

Bronchiectasis diagnosis criteria: Individuals with proven bronchiectasis, diagnosis by high-resolution computed tomography (HRCT). Origin of bronchiectasis was determined after history taking, examination and investigations including nasal respiratory ciliary beat frequency assessment

Recruitment: outpatient clinics of University of Hong Kong

Co-morbidities:

  1. Thrombocytopaenia

  2. Bone marrow transplant

  3. Biliary calculi

  4. Renal transplant

  5. Hypertension

  6. Hepatitis B

Reasons for participant exclusion:

  1. Unreliable clinic attendance

  2. Adverse reaction to macrolides

  3. Females who were lactating

Baseline FEV: intervention 1.06 (0.77 ± 1.47), control 0.99 (0.71 ± 1.39)

Baseline exacerbations (number of exacerbation in previous 12 months): intervention 2.1 ± 1.68, control 2.91 ± 0.2

Interventions

Setting: outpatient

Duration of intervention: 8 weeks

Type of antibiotic(s): erythromycin (500 mg BD)

Type of control: placebo (BD)

Outcomes

Pre-specified outcomes:

  1. Clinical and laboratory assessment

  2. Sputum collection and assessment of physical character

  3. Sputum microbiology

  4. Measurement of sputum pro-inflammatory cytokines and LTB4 concentrations

Follow-up period: 12 months

NotesStudy supported by CRCG grant from University of Hong Kong
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot specified
Allocation concealment (selection bias)Unclear riskNot specified
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskDouble-blind but not specified
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskDouble-blind but not specified
Incomplete outcome data (attrition bias)
All outcomes
Low riskPrimary and secondary endpoints not specified directly but do address aims
Selective reporting (reporting bias)Low riskPrimary and secondary endpoints not specified directly but do address aims; however sputum volume not reported
Other biasLow riskInhaled steroids and bronchodilators will change results of PFT; no other biases identified

Valery 2013

Methods

Country: Australia and New Zealand

Design: multi-centre double-blind randomised parallel-group placebo-controlled trial

Objective: Among indigenous children with bronchiectasis, can long-term (12-24 months) azithromycin treatment reduce the frequency of pulmonary exacerbations compared with placebo?

Secondary questions: Does long-term azithromycin treatment reduce or lengthen hospitalised pulmonary exacerbations and severity of pulmonary exacerbations, improve growth, decrease school absenteeism, improve respiratory symptoms, improve pulmonary function as measured by FEV1 and sputum characteristics?

Is azithromycin associated with any serious adverse events or with increased antibiotic resistance in respiratory bacterial pathogens in the nasopharynx?

Methods of analysis:

  1. Intention-to-treat

  2. Cox proportional hazards regression, Kaplan-Meier estimates

  3. Chi2 tests

  4. Fisher’s exact test

Participants

Eligible for study: 92

Randomly assigned: intervention 45, control 44

Completed: intervention 40, control 35

Age: intervention 3.99 ± 2.14, control 4.22 ± 2.30

Gender: intervention M/F 26/19, control M/F 21/23

Bronchiectasis diagnosis criteria: with confirmed HRCT scan diagnosis of bronchiectasis or clinical diagnosis of bronchiectasis (probable bronchiectasis, but no HRCT scan available) after clinical review and other appropriate investigations have been completed (including full blood count, immune function test, sweat test, CXR and, when indicated, contrast videofluoroscopy

Recruitment: outpatient

Co-morbidities:

Reasons for participant exclusion:

  1. Children receiving chemotherapy

  2. Immunosuppressive treatment

  3. Long-term antibiotic use

  4. Identified underlying cause of chronic suppuratives lung disease or bronchiectasis (e.g. CF, primary immunodeficiency, ciliary dyskinesia, primary aspiration)

  5. Other disorders (e.g. renal or hepatic failure, diabetes, central nervous system or neuromuscular disorder, congenital cardiac abnormality)

  6. History of macrolide hypersensitivity

Baseline FEV: not mentioned

Baseline exacerbations: not mentioned

Interventions

Setting: outpatient

Duration of intervention: 12 to 24 months

Type of antibiotic(s): azithromycin (oral 30 mg/kg once a week)

Type of control: placebo

Outcomes

Pre-specified outcomes:

Primary outcome

  1. Pulmonary exacerbation rate as established by medical record review (number of episodes per child/over study period)

Secondary outcomes

  1. Time to first pulmonary exacerbation

  2. Duration of exacerbation episode (discharge date minus admission date plus 1 day)

  3. Severity (admission to hospital, oxygen supplementation)

  4. Weight-for-age Z-scores (Z-score at last study clinic minus its value at baseline)

  5. Respiratory signs and symptoms (assessed by study personnel on history and physical examination)

  6. Sputum characteristics (using a validated sputum colour chart)

  7. Bronkotest (Bronkotest, Middlesex, UK)

  8. School absenteeism

  9. FEV1 percentage in those 6 years of age and older

  10. Serious adverse events, antibiotic resistance in bacterial pathogens cultured from deep nasal swabs

Follow-up period: 24 months from study entry

NotesThis work is supported by the National Health and Medical Research Council (NHMRC) of Australia (project grant numbers 389837 (clinical component) and 545223 (microbiology component)), the Telstra Foundation (seeding grant – Telstra Community Development Grant, 2004), the Health Research Council of New Zealand (grant no. 08/158) and the Auckland Medical Research Foundation (grant no. 81542). P.C. Valery is supported by an Australian Research Council Future Fellowship (No. 100100511). A.C. is supported by NHMRC fellowship 545216. This work is produced as part of the In-Kind activities of the Lowitja Institute incorporating the Cooperative Research Centre for Aboriginal and Torres Strait Islander Health
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskUsed computer-generated, permuted-block design
Allocation concealment (selection bias)Low riskUsed sequentially numbered, double-sealed, opaque envelopes
Blinding of participants and personnel (performance bias)
All outcomes
Low riskUsed identical packaging and placebo that was the same in appearance, taste and smell to azithromycin
Blinding of outcome assessment (detection bias)
All outcomes
Low riskParticipants, families, health professionals and study personnel unaware of treatment assignment until data analysis was completed
Incomplete outcome data (attrition bias)
All outcomes
Low riskClear explanation regarding missing data and reason for early termination of study
Selective reporting (reporting bias)Unclear riskReported pre-specified outcomes
Other biasLow riskNo other bias identified

Wilson 2013a

Methods

Country: Australia, Germany, Spain, Sweden, UK, USA

Design: multi-centre randomised placebo-controlled double-blind study

Objective/aim: to assess safety and efficacy of ciprofloxacin DPI treatment for 28 days, with 56 days of follow-up in individuals with non-CF bronchiectasis, by examining changes in bacterial load and other important clinical outcomes, as well as tolerability

Study site: outpatient

Methods of analysis:

  1. Frequency analysis

  2. Descriptive analysis

  3. 2-Way ANCOVA

Participants

Eligible for study: 277

Randomly assigned: intervention 60, control 64

Completed: 103

Age: intervention 64.7 ± 11.8, control 61.4 ± 11.9

Gender (M/F): intervention 21/39, control 21/43

Bronchiectasis diagnosis criteria:

adults with proven and documented diagnosis of idiopathic or post-infective non-CF bronchiectasis confirmed by HRCT

Recruitment: not mentioned

Co-morbidities: not mentioned

Reasons for participant exclusion:

  1. Active non-tuberculous mycobacterial infection

  2. Recent significant haemoptysis (volume requiring clinical intervention, within previous 4 weeks)

  3. Use of nebulised antibacterials as maintenance treatment or systemic antibacterials for an exacerbation, within 4 weeks before randomisation

Baseline FEV: intervention 57.2 ± 13.7, control 54.6 ± 14.8

Baseline exacerbations: intervention 1 to 2 exacerbations = 30 (50%), 3 to 4 exacerbations = 25 (41.7%), ≥ 5 exacerbations = 5 (8.3%); control 1 to 2 exacerbations = 35 (54.7%), 3 to 4 exacerbations = 26 (40.6%), ≥ 5 exacerbations = 3 (4.7%)

Interventions

Setting: not mentioned

Duration of intervention: 28 days

Type of antibiotic(s): ciprofloxacin (inhaled 32.5 mg BD)

Type of control: placebo (BD)

Outcomes

Pre-specified outcomes:

Primary endpoints:

  1. To determine effects of ciprofloxacin DPI on total bacterial density of pre-defined potential respiratory pathogens in sputum

Secondary endpoints:

  1. Time to exacerbation

  2. Emergence of new potential respiratory pathogens - changes in inflammatory biomarkers

  3. Emergence of resistance among baseline pathogens

  4. Change in 24 hour sputum volume and colour from baseline

  5. Changes in FEV1 and FVC

  6. Health status (SGRQ)

  7. Adverse effects

Follow-up period: 84 days

NotesStudy was sponsored by Bayer Pharma AG, Germany. R. Wilson and A. De Soyza wish to acknowledge the support of the National Institute for Health Research (NIHR) infrastructure in the form of NIHR Biomedical Research Unit funding (Royal Brompton Hospital, London, UK) and the Comprehensive Local Research Networks funding (other UK centres)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of allocation sequence not mentioned
Allocation concealment (selection bias)Unclear riskNot mentioned
Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk

Delivery of intervention vs control not clear

Appearance/packaging, taste of intervention drug and placebo, etc., not mentioned

Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk

Execution of assessment not mentioned

Unclear regarding management/protocol for exacerbation while patients were on trial

Incomplete outcome data (attrition bias)
All outcomes
High risk

Total 74 completed as per Figure 1, but numbers varied as per Tables 2 and 3

“N” values from different groups at different stages not available (see Figure 1)

Selective reporting (reporting bias)Unclear riskUnclear if per protocol or intention-to-treat analysis
Other biasUnclear riskMany endpoints/outcomes reported with just mean values without SD

Wong 2012

Methods

Country: New Zealand

Design: multi-centre double-blind placebo-controlled randomised controlled trial

Objective/aim: to test whether azithromycin decreases frequency of exacerbations, increases lung function and improves health-related quality of life in patients with non-cystic fibrosis bronchiectasis

Study site: not mentioned

Methods of analysis:

  1. Intention-to-treat analysis

  2. Poisson regression model (number of exacerbations)

  3. Cox proportional hazard regression model (time to first exacerbation)

Participants

Eligible for study: 141

Randomly assigned: intervention 71, control 70

Completed: intervention 67, control 60

Age: intervention 60.9 ± 13.6, control 59.0 ± 13.3

Gender (M/F): intervention 23/48, control 20/50

Bronchiectasis diagnosis criteria: HRCT

Recruitment: 3 centres in New Zealand

Co-morbidities: asthma

Reasons for participant exclusion:

  1. History of cystic fibrosis

  2. Hypogammaglobulinaemia

  3. Allergic bronchopulmonary aspergillosis

  4. Positive culture of non-tuberculous mycobacteria in past 2 years or at screening

  5. Macrolide treatment for longer than 3 months in past 6 months

  6. Unstable arrhythmia

Baseline FEV: intervention FEV1 pre-BD: 1.87 (0.74) FEV1 post-BD: 1.94 (0.74), control FEV1 pre-BD: 1.88 (0.69) FEV1 post-BD: 1.95 (0.71)

Baseline exacerbations (in the past year): intervention 3.34 (2.61), control 3.93 (2.49)

Interventions

Setting: outpatient

Duration of intervention: 6 months

Type of antibiotic(s): azithromycin (oral 500 mg, 3 days a week (M/W/F))

Type of control: placebo

Outcomes

Pre-specified outcomes:

Primary endpoints:

  1. Rate of event-based exacerbations in the first 6 months

  2. Forced expiratory volume in 1 second (FEV1) before bronchodilation

  3. SGRQ total score at end of treatment period

Secondary endpoints:

  1. Time to first exacerbation

  2. Rate of symptom-based exacerbations

  3. Pre-bronchodilator and post-bronchodilator forced vital capacity (FVC)

  4. Post-bronchodilator FEV1

  5. Exercise capacity (as measured by 6MWT)

  6. SGRQ total score at 12 months

  7. Concentration of C-reactive protein (assessed only at 6 months)

  8. Sputum cell counts and microbiology

  9. Adverse events.

Follow-up period: 12 months

NotesStudy was funded by Health Research Council of New Zealand and Auckland District Health Board Charitable Trust
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskParticipants randomly assigned in a 1:1 ratio, with permuted block size of 6 and sequential assignment
Allocation concealment (selection bias)Low riskEligible participants randomly assigned to receive azithromycin or placebo by a statistician independent of the reporting statistician with a computer-generated random number list
Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants, research assistants and investigators masked to treatment allocation; same protocol applied
Blinding of outcome assessment (detection bias)
All outcomes
Low riskNo biases identified
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskInsufficient information
Selective reporting (reporting bias)Low riskNo biases identified
Other biasUnclear riskOne of the secondary outcomes(CRP) was not repeated at the end of the trial

Yalçin 2006

  1. a

    BAL: Bronchoalveolar lavage; CF: Cystic fibrosis; HRCT: High-resolution computed tomography; PA: Pseudomonas aeruginosa.

Methods

Country: Turkey

Design: randomised controlled trial

Objective/aim: to evaluate effects of macrolide antibiotics on process of inflammation (by measuring IL-8, TNF-a, IL-10 levels and cell profiles in BAL fluid), pulmonary function and sputum production in children with steady-state bronchiectasis, secondary to causes other than CF or primary immunodeficiencies

Study site: Department of Paediatric Chest Diseases at Hacettepe University Faculty of Medicine

Methods of analysis:

  1. Wilcoxon signed ranks test

  2. Mann–Whitney U test was used for comparing per cent changes between the 2 groups

Participants

Eligible for study: 34

Randomly assigned: intervention 17, control 17

Completed: intervention 17, control 17

Age: intervention 13.1 ± 2.7, control 11.9 ± 2.9

Gender (M/F): intervention 8/9, control 11/6

Bronchiectasis diagnosis criteria: bronchiectasis, which developed as the result of causes other than CF or primary immunodeficiencies, based on clinical and high-resolution computed tomography (HRCT) findings

Recruitment: outpatient clinics

Co-morbidities: not mentioned

Reasons for participantexclusion:

  1. Antibiotic therapy in preceding 4 months

Baseline FEV (% of predicted): intervention 74% predicted, control 79% predicted

Baseline exacerbations: not mentioned

Interventions

Setting: not mentioned

Duration of intervention: 3 months

Type of antibiotic(s): clarithromycin (15 mg/kg/d)

Type of control: supportive therapy (mucolytics, expectorants, postural drainage)

Outcomes

Pre-specified outcomes:

  1. Not specifically stated as outcomes but are measured

    1. Effects on process of inflammation through measurement of IL-8, TNF-a, IL-10, BAL cell profile, PFT, sputum production

Follow-up period: 3 months

NotesSANOVEL Pharmaceuticals Inc. supplied cytokine kits
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo further information available
Allocation concealment (selection bias)Unclear riskNo further information available
Blinding of participants and personnel (performance bias)
All outcomes
High riskPlacebo was not mentioned
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo further information available and no pre-specified outcomes for assessment
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo further information available
Selective reporting (reporting bias)High risk

Majority of results were reported without numerical data.

No pre-specified outcomes

Other biasUnclear riskInsufficient information

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    For additional details on excluded studies, see additional tables.

Adachi 1988Not all patients with bronchiectasis
Al Mobeireek 1998Full review not possible; however no particular intervention was used
Alberto 1968Not all patients with bronchiectasis
Alekseenko 1978No prolonged antibiotics
Allen 1988No prolonged antibiotics, 1 week duration
Allen 1990No prolonged antibiotics
Alora 1978No bronchiectasis
Anwar 2008Not an RCT
Barach 1952Not an RCT
Begg 2000No prolonged antibiotics
Bergogne 1982No prolonged antibiotics
Bergogne 1985No antibiotics
Bevilacqua 1973Not all patients with bronchiectasis
Bilton 2006Study recruited participants when they experienced an exacerbation
Bilton 2013No antibiotics
Bradley 2011No antibiotics
Brambilla 1992Not all patients with bronchiectasis
Bruninx 1973Not all patients with bronchiectasis
Chan 1996No prolonged antibiotics
Chang 2010Not all patients with bronchiectasis
Chang 2012No prolonged antibiotics
Cherniack 1959Not all patients with bronchiectasis
Cherniack 1966Not an RCT
Choi 1996No bronchiectasis
Chonabayashi 1988No bronchiectasis
Chonabayashi 1988aNo bronchiectasis
Chuchalin 1997Not all patients with bronchiectasis
Clarke 1981Not an RCT
Crowther 1999No patients with non-CF bronchiectasis
Cseri 1975Not all patients with bronchiectasis
Currie 1987Not an RCT
Czako 1968No antibiotics
Danesi 1985Not an RCT
Davies 1983No prolonged antibiotics
Dobrev1990No prolonged antibiotics
Douglas 1957Not an RCT
Doutsu 1988Full review not possible; however drug not used in clinical practice
Finelgold 1981Not all patients with bronchiectasis
Foweraker 1993Not all patients with bronchiectasis
Franklin 1953Not an RCT
Garcia 1984Not all patients with bronchiectasis
Grau 1992Not an RCT
Haidl 2013No non-CF bronchiectasis
Hara 1997Not an RCT
Harada 1988No prolonged antibiotics and not all patients with bronchiectasis
Hayashi 1989Not an RCT
Helm 1956Not an RCT
Hernandez 2002No prolonged antibiotics
Hill 1986Not an RCT
Hill 1988Not an RCT
Hill 1992Full review not possible; however drug not used in clinical practice
Hill 1994No prolonged antibiotics
Hossain 2010Full review of abstract not possible
Howie 1976Not an RCT
Hughes 1973No bronchiectasis
Hughes 1973aNo bronchiectasis
Ida 1988Not all patients with bronchiectasis
Iino 1993Not an RCT
Inoue 1989Not an RCT
Ip 1993Not an RCT
Ip 1998No prolonged antibiotics
Irabu 1989Not an RCT
Irabu 1992Not an RCT
Izumi 1994Not an RCT
Izumikawa 1988Not an RCT
Jia 2010No bronchiectasis
Kaji 1988Not an RCT
Kanaij, 1963Not an RCT
Karachalios 1985Not an RCT
Kawai 1988Not an RCT
Kawai 1996Not an RCT
Kawano 1974Not an RCT
Kawashima 1989Not all patients with bronchiectasis
Kellett 2011No antibiotics
Kikuchi 1991Not an RCT
Knox 1955Not an RCT
Kobayashi 1984Not an RCT
Kobayashi 1995Not an RCT
Koh 2000No antibiotics
Koyama 1992Not an RCT
Krawczyk 1981Not an RCT
Krofton 1971No bronchiectasis
Kudo 1988Full review not possible; however drug not used in clinical practice
Kudo 1992Not an RCT
Kulpati 1990No bronchiectasis and not an RCT
Kurasawa 1988No bronchiectasis
Kurishima 1992Not an RCT
Kuze 1988Not an RCT
Lam 1986aNo prolonged antibiotics
Lam 1989No prolonged antibiotics, trial of 10 days duration
Lam 2006No prolonged antibiotics
Lamotte 1981Not an RCT
Ledson 2000Not an RCT
Lin 1996No control group; comparison of erithromycin vs amoxicillin
Lin 1997No prolonged antibiotics, trial of 3 days duration
Lin 2006No antibiotics for bronchiectasis
Lioberes 1990Does not fulfil inclusion criteria
Lunacharskaia 1968Not all patients with bronchiectasis
Maekawa 1967No bronchiectasis
Mahashur 1993Not an RCT
Mahashur 2002Not all patients with bronchiectasis
Masekela 2013Not all patients with bronchiectasis
Masuno 1992Not an RCT
Matsumoto 1983Not an RCT
Matsumoto 1984Not an RCT
Matsumoto 1984aFull review not possible; however drug not used in clinical practice
Matsumoto 1986Full review not possible; however drug not used in clinical practice
Matsumoto 1992Full review not possible; however drug not used in clinical practice
Matsushima 1988Full review not possible; however drug not used in clinical practice
Matsuura 1993Not an RCT
Matzeu 1981Not an RCT
May 1972Not all patients with bronchiectasis
May 1974Not all patients with bronchiectasis
Mazzei 1993Not all patients with bronchiectasis
McVay 1953Not an RCT
Mehta 1991No prolonged antibiotics
Messens 1973Not all patients with bronchiectasis
Mijuskovic 1972No prolonged antibiotics
Milcev 1985Not all patients with bronchiectasis
Min 1988Not an RCT
Minami 1996Not an RCT
Ming 2005Not all prolonged antibiotics
Miyamoto 1984Full review not possible; however drug not used in clinical practice
Molodtsova 1980Full review not possible; however drug not used in clinical practice
Morrone 1989Not all patients with bronchiectasis
Murayama 1992Not all patients with bronchiectasis
Music 1989No bronchiectasis
Nagy 1968Not an RCT
Nakagawa 1993Not an RCT
Nakagawa 1995No bronchiectasis and no prolonged antibiotics
Nakamori 1992Not an RCT
Nakamura 2007No prolonged antibiotics
Nakatani 1993Not an RCT
Nasu 1988Not an RCT
Neumayr 1963Not an RCT
Nobuoka 1985Not an RCT
O'Donovan, 1987No prolonged antibiotics
Oishi 1994Not an RCT
Oizumi 1978Not an RCT
Oki 1993Not an RCT
Ovcharenko 1991No prolonged antibiotics
Patterson 2005No antibiotics; this study randomly assigned participants between airway clearance techniques with no antibiotic regimens
Patterson 2007No prolonged antibiotics
Pezza 1983No bronchiectasis
Pines 1967Not all patients with bronchiectasis
Pines 1979No prolonged antibiotics
Pines 1981Not all patients with bronchiectasis
Piovano 1997No prolonged antibiotics
Popov 1988Not an RCT
Prigogine 1988No prolonged antibiotics
Ramer 1981No prolonged antibiotics
Rayner 1994Not an RCT
Rimoldi 1987Not all patients with bronchiectasis
Saito 1993Full review not possible; however drug not used in clinical practice
Saito 1999Full review not possible; however drug not used in clinical practice
Saito 2002Not all patients with bronchiectasis
Santiveri 1995No prolonged antibiotics
Sawae 1991Not all patients with bronchiectasis
Sawae 1995No prolonged antibiotics
Sawaki 1986Not an RCT
Schulz 1972Not all patients with bronchiectasis
Serisier 2011aNot an RCT
Serisier 2011bNo antibiotics
Shewsbury 2004Published rationale and methodology
Shigeno 1984Full review not possible; however drug not used in clinical practice
Shimada 1996No prolonged antibiotics
Shimada K 1988Full review not possible; however drug not used in clinical practice
Shimokata 1990Not an RCT
Shimokata 1992Exclusion criteria (panbronchiolitis)
Shirai 1995Exclusion criteria (panbronchiolitis)
Shishido 1988Not all patients with bronchiectasis
Shishido 1995Exclusion criteria (panbronchiolitis)
Singleton 2007Full review not possible; intervention not available to be determined based on abstract
Singleton 2008Not an RCT
Soejima 1988No prolonged antibiotics
Stass 2013Not all patients with bronchiectasis
Stockley 1984Not an RCT
Stockley 1985Not an RCT
Suga 1988Full review not possible; however drug not used in clinical practice
Suyama 1988No prolonged antibiotics
Tagaya 2002No prolonged antibiotics
Tageldin 1994Not an RCT
Takamoto 1994Not all patients with bronchiectasis
Tamura 1992Full review not possible; however drug not used in clinical practice
Tanimoto 1992No prolonged antibiotics
Tanimoto 1993No prolonged antibiotics
Taskar 1992Review of bromhexine, not antibiotics
Tsang 1993No prolonged antibiotics
Tsang 1994No prolonged antibiotics
Tsang 1999aNo prolonged antibiotics
Unoura 1988Exclusion criteria (panbronchiolitis)
Vergnon 1985No prolonged antibiotics
Watanabe 1982No prolonged antibiotics
Watanabe 1989Exclusion criteria (panbronchiolitis)
Watanabe 1991Exclusion criteria (panbronchiolitis)
Watanabe 1994No control group; comparison of 2 antibiotics and no prolonged antibiotics
Watanabe 1995Exclusion criteria (panbronchiolitis)
Watanabe 2001Not all patients with bronchiectasis
Yamada 1993Not all patients with bronchiectasis
Yamaki 1988Not all patients with bronchiectasis
Young Whan 1993Effect of long-term erythromycin on diffuse panbronchiolitis, not non-CF bronchiectasis
Zhang 2003No prolonged antibiotics

Characteristics of studies awaiting assessment [ordered by study ID]

Aksamit 2017

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Alder 2012

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Altenburg 2016

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Altenburg 2016a

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Altenburg 2016b

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Barker 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Burr 2015

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Burr 2016

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Chalmers 2012

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Chang 2013

MethodsRandomised double-blind double-dummy placebo-controlled parallel-group trial
Participantsn = 170 children from 6 Australian and New Zealand centres
InterventionsGroup 1 will receive azithromycin (5 mg/kg daily) with placebo amoxycillin-clavulanate, and group 2 will receive amoxycillin-clavulanate (22.5 mg/kg twice daily) with placebo azithromycin for 21 days as treatment for non-severe respiratory exacerbations
Outcomes

Exacerbations

Parent proxy cough-specific quality of life (PC-QOL) questionnaire

Time to next exacerbation

Hospitalisation

Duration of exacerbation

Spirometry data

Descriptive viral and bacteriological data from nasal samples and blood inflammatory markers

NotesBased on published protocol

Chen 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

De Soyza 2016

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Dimakou 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Hare 2015

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Haworth 2013

MethodsRandomised controlled trial
Participantsn = 54 (weeks 0 and 12); n = 31 (week 26), for each arm
InterventionsNebulised colistimethate sodium 1 MIU/mL for ≤ 6 months
OutcomesSt George’s Respiratory Questionnaire (SGRQ) across 3 domains (symptoms, activity, impact)
NotesBased on abstract

Haworth 2013a

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Haworth 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Juthong 2011

MethodsRandomised double-blind placebo-controlled trial
Participants20 bronchiectatic patients, mean age 56 years, participated in this study. Nine participants were randomly assigned to roxithromycin 300 mg, and 11 participants received placebo once daily
InterventionsRoxithromycin 300 mg once daily for 8 weeks
Outcomes

Symptom score

Quality of life (SGRQ)

Pulmonary function test

Sputum culture

Adverse events

NotesContacted study author for further information, but not available at time of completion of this review

Kobbernagel 2016

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Liu 2012a

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Liu 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Lourdesamy 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

New Study

Methods 
Participants 
Interventions 
Outcomes 
Notes 

O'Donnell 2016

Methods 
Participants 
Interventions 
Outcomes 
Notes 

O'Donnell 2016a

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Orriols 2015

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Rogers 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Tabernero 2014

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Tabernero 2015

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Terpstra 2016

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Twiss 2008

MethodsRandomised double-blind placebo-controlled cross-over (2 × 3 month treatment period)
Participants15 participants 5 to 15 years of age
InterventionsTrial of 80 mg gentamicin nebulised twice daily
Outcomes

Monthly spirometry

Symptom scores

Hospitalisation

Antibiotic history

Sputum samples for microbiology

NotesBased on abstract

Yalcin 2006

Methods 
Participants 
Interventions 
Outcomes 
Notes 

Ancillary