Combinación de agonista betaCombination inhaled steroid and long‐acting beta2de acción prolongada y esteroide inhalado además del tiotropio versus tiotropio o la combinación sola para la enfermedad pulmonar obstructiva crónica
Resumen
Antecedentes
El broncodilatador de acción prolongada tiotropio y el tratamiento de combinación con inhalador único con corticosteroides inhalados y agonistas 2 de acción prolongada (CSI/ABAP) se utilizan habitualmente para el tratamiento de mantenimiento de los pacientes con enfermedad pulmonar obstructiva crónica (EPOC). La combinación de estos tratamientos, con diferentes mecanismos de acción, puede ser más efectiva que los componentes individuales.
Objetivos
Evaluar los efectos relativos de los siguientes tratamientos sobre los marcadores de las exacerbaciones, los síntomas, la calidad de vida y la función pulmonar en los pacientes con EPOC.
• Tiotropio más ABAP/CSI versus tiotropio.
• Tiotropio más ABAP/CSI versus ABAP/CSI.
Métodos de búsqueda
Se hicieron búsquedas en el Registro Especializado de Ensayos Controlados del Grupo Cochrane de Vías Respiratorias (Cochrane Airways Group) (abril 2015), ClinicalTrials.gov (www.ClinicalTrials.gov), el portal de ensayos de la Organización Mundial de la Salud (OMS) y las listas de referencias de los artículos pertinentes.
Criterios de selección
Se incluyeron los ensayos controlados aleatorizados (ECA) paralelos, con una duración de tres meses o más, realizados para comparar el tratamiento de combinación con CSI y ABAP además del tiotropio inhalado versus tratamiento con tiotropio solo o la combinación sola.
Obtención y análisis de los datos
La inclusión de los ensayos se evaluó de forma independiente y luego se extrajeron los datos sobre la calidad de los ensayos y los resultados. Se contactó con los autores de los estudios para solicitar información adicional. Se recopiló información de los ensayos sobre los efectos adversos.
Resultados principales
Tiotropio más ABAP/CSI versus tiotropio
Se incluyeron seis estudios (1902 participantes) con bajo riesgo de sesgo que compararon el tiotropio además del tratamiento de combinación con agonista beta 2 de acción prolongada y corticosteroide inhalado versus tiotropio solo. No se encontraron diferencias estadísticamente significativas en la mortalidad entre los tratamientos (odds‐ratio [OR] 1,80; intervalo de confianza [IC] del 95%: 0,55 a 5,91; dos estudios; 961 participantes), así como en las hospitalizaciones por todas las causas (OR 0,84; IC del 95%: 0,53 a 1,33; dos estudios; 961 participantes). El efecto sobre las exacerbaciones fue heterogéneo entre los ensayos y no se realizó un metanálisis. La calidad de vida relacionada con la salud medida con el St. George’s Respiratory Questionnaire (SGRQ) mostró una mejoría estadísticamente significativa en las puntuaciones totales con la administración del tiotropio + ABAP/CSI en comparación con tiotropio solo (diferencia de medias [DM] ‐3,46; IC del 95%: ‐5,05 a ‐1,87; cuatro estudios; 1 446 participantes). La función pulmonar fue significativamente diferente en el grupo de tratamiento combinado (tiotropio + ABAP/CSI), aunque el efecto beneficioso promedio con este tratamiento fue pequeño. Ninguno de los estudios incluidos examinó la tolerancia al ejercicio como un resultado.
Una estimación combinada de estos estudios no mostró una diferencia estadísticamente significativa en los eventos adversos (OR 1,16; IC del 95%: 0,92 a 1,47; cuatro estudios; 1 363 participantes), los eventos adversos graves (OR 0,86; IC del 95%: 0,57 a 1,30; cuatro estudios; 1 758 participantes) y la neumonía (OR de Peto 1,62; IC del 95%: 0,54 a 4,82; cuatro estudios; 1 758 participantes).
Tiotropio más ABAP/CSI versus ABAP/CSI
Uno de los seis estudios (60 participantes) también comparó el tratamiento combinado (tiotropio + ABAP/CSI) versus el tratamiento con ABAP/CSI solo. Este estudio no tuvo poder estadístico suficiente; por lo tanto, los resultados no nos permitieron sacar conclusiones para esta comparación.
Conclusiones de los autores
Esta actualización de la revisión incluye tres estudios adicionales y proporciona nueva evidencia de calidad baja que apoya el hallazgo de que el tiotropio + el tratamiento basado en ABAP/CSI mejora la calidad de vida específica de la enfermedad. La evidencia actual es insuficiente para apoyar el beneficio del tiotropio + tratamiento con ABAP/CSI para la mortalidad, el ingreso hospitalario o las exacerbaciones (evidencia de calidad moderada y baja). En comparación con el uso de tiotropio solo, el tiotropio + el tratamiento con ABAP/CSI no parece aumentar los efectos indeseables ni los eventos adversos no fatales mortales.
PICO
Resumen en términos sencillos
¿El tiotropio más los inhaladores de combinación es mejor que el tiotropio o los inhaladores de combinación solos para el tratamiento de la EPOC?
Antecedentes
La enfermedad pulmonar obstructiva crónica (EPOC) es una enfermedad pulmonar que incluye afecciones como la bronquitis crónica o el enfisema. La EPOC se caracteriza por el estrechamiento de las vías respiratorias y la destrucción del tejido pulmonar. Los síntomas incluyen disnea y tos a largo plazo. Los síntomas de la EPOC son tratables, pero la condición no se puede revertir ni curar. Generalmente de debe a irritantes de las vías respiratorias, como el hábito de fumar o el polvo inhalado.
Habitualmente se utilizan inhaladores con broncodilatadores (que permiten que las vías respiratorias de los pulmones se relajen y amplíen) o agentes antiinflamatorios para aliviar los síntomas y reducir la disminución a largo plazo en la salud provocada por la EPOC. Los ejemplos de estos tratamientos son el tiotropio, que es un broncodilatador, y los inhaladores combinados que contienen otro tipo de broncodilatador (agonistas beta de acción prolongada) junto con agentes antiinflamatorios (esteroides). Estos tratamientos funcionan de diferentes maneras y por lo tanto quizá sean más beneficiosos si se los usa juntos.
Características de los estudios
Esta revisión encontró seis estudios con 1902 participantes que compararon la eficacia a largo plazo y los efectos secundarios del tiotropio combinado con inhaladores de combinación para el tratamiento de los pacientes con EPOC. No todos los pacientes incluidos en estos estudios presentaban EPOC que fuera lo suficientemente grave como para recomendarles el tratamiento combinado según las guías actuales.
Resultados clave
La evidencia actual muestra beneficios potenciales del tratamiento con tiotropio además del tratamiento combinado con corticosteroides inhalados y agonistas beta2 de acción prolongada a través de una mayor calidad de vida relacionada con la salud y una pequeña mejoría en la función pulmonar en los pacientes que reciben este tratamiento combinado. Sin embargo, esta evidecia no permite establecer conclusiones acerca de los efectos de estos tratamientos sobre la mortalidad, la hospitalización para todas las causas y las exacerbaciones. La frecuencia de eventos adversos graves y no graves no aumentó los dos grupos.
Calidad de la evidencia
En general, la evidencia presentada en esta revisión se consideró de calidad moderada o baja, lo que significa que existe una confianza razonable en algunos de los resultados, pero menos confianza en otros.
Authors' conclusions
Summary of findings
| Tiotropium + LABA/ICS combination compared with tiotropium for chronic obstructive pulmonary disease | ||||||
| Patient or population: patients with chronic obstructive pulmonary disease Settings: ambulatory clinics Intervention:tiotropium + LABA/ICS combination Comparison: tiotropium | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments | |
| Assumed risk Tiotropium | Corresponding risk Tiotropium + LABA/ICS combination | |||||
| Mortality (all‐cause) | 8 per 1000 | 15 per 1000 | OR 1.80 | 961 | ⊕⊕⊕⊝ | |
| Hospital admission (all causes) | 156 per 1000 | 101 per 1000 | OR 0.84 | 961 | ⊕⊕⊝ ⊝ | |
| Exacerbation ‐ at 12‐month follow‐up | 628 per 1000 | 601 per 1000 | OR 0.89 | 301 | ⊕⊕⊝⊝ | |
| Serious adverse events (non‐fatal) | 60 per 1000 | 52 per 1000 | OR 0.86 | 1758 | ⊕⊕⊝⊝ a,c | |
| Quality of life up to 6 months (SGRQ) | Mean SGRQ up to 6 months in the intervention groups was 3.46 lower (5.05 to 1.87 lower) | ‐ | (4 studies) | ⊕⊕⊝⊝ d | A lower score indicates better quality of life | |
| FEV1 pre‐dose ‐ FEV1 3‐6 months mean difference | Mean FEV1 pre‐dose ‐ FEV1 3‐6 months mean difference in the intervention groups was 0.06 (0.04 to 0.08 ) | ‐ | (4 studies) | ⊕⊕⊕⊝ | ||
| FEV1 pre‐dose ‐ FEV1 1 year | Mean FEV1 pre‐dose ‐ FEV1 1 year mean difference in the intervention groups was 0.06 (0 to 0.12 ) | ‐ | (1 study) | ⊕⊕⊕⊝ | ||
| *The basis for 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 | ||||||
| aDowngraded one level because of imprecision (95% confidence interval includes both no effect and appreciable harm) bDowngraded one level because of study limitations (incomplete outcome assessment in Aaron 2007) cDowngraded once because of study limitations (incomplete outcome assessment in Aaron 2007 and Hanania 2011; unclear risk of selection bias in Hanania 2011; possible detection bias in Jung 2012) dDowngraded two levels because of study limitations (unclear risk of selection bias and detection bias and incomplete outcome assessment in Hoshino 2011; unclear risk of detection bias in Jung 2012; incomplete outcome assessment in Aaron 2007) eDowngraded one level because of study limitations (unclear risk of selection and detection bias in Cazzola 2007; unclear risk of detection bias in Jung 2012; incomplete outcome assessment in Aaron 2007) | ||||||
Background
Description of the condition
Chronic obstructive pulmonary disease (COPD) is a general term that refers to chronic bronchitis or emphysema, or both. COPD occurs when airflow to the lungs is restricted by narrowing of the airways. Symptoms include cough, breathlessness and reduced exercise capacity. The Global Initiative for Chronic Obstructive Lung Disease (GOLD 2015) guidelines describe COPD as a preventable and treatable condition that is not fully reversible. Worldwide, the main cause of COPD is tobacco smoking, but air pollution, burning of biomass and occupational exposure are also risk factors (GOLD 2015). The prevalence, morbidity and mortality of the disease vary across populations, and the disorder causes a substantial economic and social burden.
Various pharmacological treatments are commonly used in COPD management to relieve symptoms, improve exercise tolerance and quality of life, reduce mortality and prevent and treat exacerbations. Exacerbations of COPD impair patients' quality of life, and a large part of the economic burden of COPD is attributed to the cost of managing exacerbations, particularly those resulting in the use of acute care services or hospitalisations (Hutchinson 2010). Appropriate pharmacological management of the disease is therefore important to reduce and prevent exacerbations. Management of COPD tends to begin with one treatment, and additional therapies are introduced as necessary to control symptoms (GOLD 2015). Self‐management, education, vaccination and rehabilitation can accompany these pharmacological interventions (Effing 2007; Lacasse 2006; Sehatzadeh 2012).
Description of the intervention
The first pharmacological step in treating patients with COPD consists of the use of short‐acting bronchodilators for symptom control when needed. These include short‐acting beta2‐agonists (SABA) and the short‐acting anticholinergic agent ipratropium. To manage persistent COPD symptoms, long‐acting bronchodilators can be introduced (GOLD 2015). Regular treatment with long‐acting bronchodilators is more efficient and convenient than treatment with regular short‐acting bronchodilators (Beeh 2010). Long‐acting bronchodilators include long‐acting beta2‐agonists (LABA) and the long‐acting anticholinergic agent tiotropium. Tiotropium bromide has gained widespread acceptance as once daily maintenance therapy in COPD (Barr 2005; GOLD 2015). Tiotropium reduces COPD exacerbations and related hospitalisations compared with ipratropium (Barr 2005). Most LABA are taken twice daily. They improve lung function compared with ipratropium, but little difference is shown in improving COPD symptoms and exercise tolerance (Appleton 2006). For symptomatic patients with severe or very severe COPD (forced expiratory volume in one second (FEV1) < 50% predicted) and with repeated exacerbations, GOLD 2015 recommends the addition of inhaled corticosteroids (ICS) to bronchodilator treatment. Inhaled corticosteroids are licensed as combination inhalers with LABA. The most common combinations of ICS and LABA in combination inhalers are fluticasone and salmeterol; budesonide and formoterol; and mometasone and formoterol. Combination therapy reduces exacerbation rates and mortality compared with ICS alone (Nannini 2013). Also compared with LABA alone, combination therapy is more effective in reducing exacerbation rates, but with no significant difference in mortality (Nannini 2007b). For patients who continue to have symptoms and are at high risk of experiencing exacerbations, triple therapy with LABA, long‐acting muscarinic antagonists (LAMA, e.g. tiotropium) and ICS is recommended. Such patients are referred to as 'Group D' in the GOLD guidelines; this group typically includes patients classified as GOLD 3 and 4, i.e. FEV1 < 50% of predicted value (GOLD 2015). Benefits of combination inhalers should be viewed against the possible increased risk of pneumonia (Nannini 2007b; Nannini 2013). Potential risks and benefits of treatment with combination inhaler compared with tiotropium are uncertain (Welsh 2010), as are risks and benefits of treatment with combination inhaler in addition to tiotropium, which will be explored in this review.
How the intervention might work
Tiotropium
Tiotropium (TIO) is a long‐acting anticholinergic agent that targets bronchospasm in COPD by relaxing the smooth muscle of the airways. Tiotropium is structurally related to ipratropium, a short‐acting anticholinergic agent that binds to M1, M2 and M3 muscarinic receptors, which in turn open the bronchi (Barr 2005). Although tiotropium binds to the same receptors as ipratropium, it has different kinetic selectivity. Tiotropium dissociates slowly from M1 and M3 receptors, giving a bronchodilator effect lasting over 24 hours, but rapidly from M2 receptors. It appears that M2 receptors are feedback inhibitory receptors, and blocking them (as is the case for ipratropium) releases acetylcholine rather than reducing it as desired (Barr 2005). Benefits of tiotropium, in comparison with placebo, include reduced COPD exacerbations and exacerbation‐related hospitalisations, and improved health‐related quality of life and symptom scores among patients with moderate and severe disease (Barr 2005). Anticholinergic side effects can occur with tiotropium and include dry mouth, constipation and tachycardia.
Inhaled beta2‐agonist plus inhaled corticosteroids
Inhaled beta2‐agonists activate beta2‐receptors in the smooth muscle of the airways, releasing adenylate cyclase and increasing intracellular cyclic adenosine monophosphate (cAMP), which leads to a cascade of reactions resulting in bronchodilation. Beta2‐agonists may act through other mechanisms such as respiratory muscle function or mucociliary clearance; patients have shown improvement in symptoms whilst showing no improvement in lung function tests. Beta2‐agonists are particularly useful because they reverse bronchoconstriction regardless of its initial cause. Side effects include muscle tremors, nervousness and occasional insomnia, but, as with all inhaled medications, systemic side effects are minimised by a comparatively low dose administered directly to the lungs. Inhaled corticosteroids are anti‐inflammatory drugs that have been associated with reduced risk of exacerbation in patients with COPD and with better quality of life outcomes when compared with placebo, with no effect on overall mortality or long‐term FEV1 (GOLD 2015; Yang 2012). Combination inhalers including ICS and LABA reduce exacerbation rates and all‐cause mortality and improve lung function and quality of life compared with placebo (Nannini 2007a). These effects are thought to be greater for combination inhalers than for the component preparations (GOLD 2015). Use of inhaled corticosteroids, alone or in combination with beta2‐agonists, potentially increases the risk of pneumonia (GOLD 2015; Yang 2012).
The combination inhalers currently available are fluticasone/salmeterol (FSC); budesonide/formoterol (BUD/F); and beclomethasone/formoterol (DPB/F).
Combination therapy
The nature of the interaction between the two systems is not yet fully understood, but combining beta2‐adrenergic receptor agonists and muscarinic acetylcholine receptor antagonists is pharmacologically reasonable, given that airway tone is regulated by the parasympathetic and sympathetic nervous systems. The synergistic effect of these therapies can be explained in several ways. One explanation is that the addition of a beta2‐adrenergic receptor agonist decreases release of acetylcholine (ACh) and amplifies bronchial smooth muscle relaxation; another is that the addition of a muscarinic acetylcholine receptor antagonist can reduce bronchoconstrictor effects of ACh and amplify bronchodilation through direct stimulation of the smooth muscle beta2‐adrenergic receptor (Cazzola 2010). An animal model showed activation of calcium‐activated potassium (KCa) channels thought to hyperpolarise the cell membrane, causing reductions in the concentration of intracellular calcium (Ca) and ACh release in prejunctional cholinergic nerves (Brichetto 2003).
Why it is important to do this review
The previous version of this review showed a significant effect of combination therapy tiotropium + LABA/ICS on FEV1 in participants with stable COPD, in comparison with tiotropium therapy alone. However, sparse evidence was found to support similar beneficial effects on other important outcomes, such as all‐cause hospitalisations, exacerbations and mortality. New published trials have been conducted with the aim of comparing these therapies; therefore it is necessary to include their results as part of this review to obtain more precise estimations of treatment effects on outcomes for which combination therapy effects remain unclear.
Objectives
To assess relative effects of the following treatments on markers of exacerbations, symptoms, quality of life and lung function in patients with COPD.
-
Tiotropium plus LABA/ICS versus tiotropium.
-
Tiotropium plus LABA/ICS versus LABA/ICS.
Methods
Criteria for considering studies for this review
Types of studies
For effectiveness and safety objectives, we included randomised controlled trials (RCTs) of parallel design conducted in patients with stable COPD who received the trial treatment for at least 12 weeks.
For economic objectives, we included economic evaluation studies such as cost‐effectiveness analyses and cost‐utility analyses addressing the same interventions in the population of interest for this review. We considered for inclusion the economic evaluation conducted alongside the RCT or economic evaluation modelling studies based on a comprehensive systematic review of the literature. We excluded partial economic evaluation studies reporting cost analysis or cost‐outcome descriptions.
Types of participants
Populations with a diagnosis of COPD. We included only studies that used an external set of criteria to screen participants for this condition (e.g. ATS; BTS; GOLD 2015; TSANZ).
Types of interventions
Inhaled combination corticosteroid and long‐acting beta2‐agonist (such as fluticasone/salmeterol, budesonide/formoterol, beclomethasone/formoterol) and tiotropium bromide versus:
-
inhaled tiotropium bromide alone; or
-
inhaled corticosteroid and long‐acting beta2‐agonist combination.
Types of outcome measures
Primary outcomes
-
Mortality (all‐cause).
-
Exercise tolerance.
-
Hospital admissions: all‐cause and due to exacerbations.
-
Exacerbations: all‐cause, requiring short burst oral corticosteroids or antibiotics as defined by agreed criteria.
-
Health‐related quality of life (measured with a validated scale for COPD, e.g. St George's Respiratory Questionnaire (SGRQ), Chronic Respiratory Disease Questionnaire (CRQ)).
-
Serious adverse events non‐fatal.
-
Pneumonia.
Secondary outcomes
-
Symptoms.
-
Forced expiratory volume in one second (FEV1).
-
Adverse events.
-
Side effects.
-
Cost‐effectiveness of interventions.
Search methods for identification of studies
Electronic searches
The previously published version of this review (Karner 2011) included searches up to July 2010. The search period for this update is July 2010 to April 2015.
For this update, we identified trials from the Cochrane Airways Group Specialised Register (CAGR), which is maintained by the Information Specialist 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 (please 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 the present, with no restriction on the language of publication. We conducted the latest search in April 2015.
Searching other resources
We reviewed reference lists of all primary studies and review articles for additional references. We contacted authors of identified trials and asked them to identify other published or unpublished studies.
Data collection and analysis
Selection of studies
Two review authors (OMG and RJD) screened the titles and abstracts of citations retrieved through literature searches and obtained those deemed to be potentially relevant. We assigned each reference to a study identifier and assessed all references against the inclusion criteria of the protocol.
Two review authors (OMG and RJD) independently examined titles and abstracts for the selection of health economics studies to be included in the critical review of economic data. We removed records that did not report on cost‐effectiveness or cost‐utility analysis. Two review authors (MXR and RJD) independently examined full‐text reports to determine which studies met the eligibility criteria of this review. We resolved disagreements by discussion between review authors. We included only full economic evaluations of high methodological and reporting quality.
Data extraction and management
We extracted the following characteristic information from each study.
-
Design (design, total duration of study and run‐in, number of study centres and locations, withdrawals, date of study).
-
Participants (N, mean age, age range, gender, COPD severity, diagnostic criteria, baseline lung function, smoking history, inclusion criteria, exclusion criteria).
-
Interventions (run‐in, intervention treatment and inhaler type, control treatment and inhaler type).
-
Outcomes (primary and secondary outcomes specified and collected, time points reported).
Two review authors (MXR and OMG) extracted data from the studies onto data collection forms. Review authors discussed discrepancies in the data and resolved them and transferred data from data collection forms into RevMan (RevMan 2014).
Data obtained by authors from the previous version of this review regarding all cause hospital admissions that were supplied by Aaron 2007 and by AstraZeneca (for Welte 2009) on request, were kept for this update without changes.
Assessment of risk of bias in included studies
We assessed all included studies for risk of bias according to the recommendations outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) for the following items.
-
Allocation sequence generation.
-
Concealment of allocation.
-
Blinding of participants and investigators.
-
Incomplete outcome data.
-
Selective outcome reporting.
We noted other sources of bias and graded each potential source of bias as having high, low or unclear risk.
We assessed the methodological quality of economic evaluations by using the Drummond checklist (Drummond 1996), which addresses the following methodological and reporting aspects.
-
Was a well‐defined question posed?
-
Was a comprehensive description of competing alternatives given?
-
Does the paper provide evidence that the programme would be effective (i.e. would the programme do more harm than good)?
-
Were all important and relevant resource uses (costs) for each alternative identified?
-
Were all important and relevant health outcome consequences for each alternative identified?
-
Were costs measured accurately in appropriate units before evaluation and valued credibly?
-
Were health outcome consequences measured credibly?
-
Were costs and health outcome consequences adjusted for the different times at which they occurred (i.e. was discounting applied)?
-
Was an incremental analysis of the consequences and costs of alternatives performed?
-
Was an adequate sensitivity analysis performed?
Quality of the body of evidence for each outcome
We assessed the quality of evidence for the main comparison at the outcome level using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (Guyatt 2011). This methodological approach considers RCTs as providing high‐quality evidence that may be rated down by limitations in any of five areas: design (risk of bias), consistency across studies, directness of the evidence, precision of estimates and presence of publication bias (Guyatt 2011). The GRADE approach results in an assessment of the quality of a body of evidence in one of four grades: (1) high: We are very confident that the true effect lies close to that of the estimate of effect; (2) moderate: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of effect but may be substantially different; (3) low: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of effect; (4) very low: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect (GRADE 2013).
Two review authors (OMG and MXR) independently assessed the quality of the body of evidence found for each of the outcomes identified as critical or important for clinical decision making: mortality, hospital admission (all causes), exacerbation at 12 months, improvement in FEV1, serious adverse events and quality of life. In the case that the study authors did not take measures to ensure concealment of allocation, randomised assignment, completion to follow‐up or blinded outcome assessment, we downgraded the quality of evidence because of design limitations (GRADE 2013). We evaluated consistency by similarity of point estimates, extent of overlap of confidence intervals (CIs) and application of statistical criteria including testing for heterogeneity (I2). We planned to downgrade the quality of evidence if we detected substantial unexplained heterogeneity across study results (i.e. some studies suggest important benefit and others no effect or harm without a clinical explanation) (GRADE 2013). We assessed precision according to the 95% CI around the pooled estimate (GRADE 2013). When studies were conducted in populations other than the target population, the GRADE framework suggests that the quality of evidence should be downgraded because of indirectness (GRADE 2013).
We entered data (i.e. pooled estimates of effects and corresponding 95% CIs) and explicit judgements that were made for each of the above aspects into the GRADEprofiler (GDT), the software used to create Summary of findings (SoF) tables. We explained in the SoF table footnotes all judgements involved in assessment of the aspects of the evidence described above.
Measures of treatment effect
We performed all statistical analyses using RevMan software (RevMan 2014). We analysed dichotomous data (such as mortality, hospital admission, number of participants with one or more exacerbations) using the Mantel‐Haenszel odds ratio (OR) and risk difference (RD), unless events were rare, in which case we employed the Peto OR (as this does not require a continuity correction for zero cells). For statistically significant results of categorical variables, we reported the number needed to treat for an additional beneficial outcome (NNTB).
We analysed continuous outcome data (such as quality of life (score) and FEV1) using the mean difference (MD). We reported the 95% CI on all estimates as fixed‐effect mean differences with 95% CI. When treatment effects were reported as a mean difference with 95% CI, we entered the MD and standard errors calculated from the 95% CI and analysed data using the generic inverse variance (GIV) tool.
Unit of analysis issues
We analysed dichotomous data by using participants as the unit of analysis (rather than events) to avoid counting the same participant more than once.
Dealing with missing data
We contacted investigators and study sponsors to verify key study characteristics and to obtain missing numerical outcome data.
Assessment of heterogeneity
We assessed the amount of statistical variation between study results by using the I2 measurement.
Assessment of reporting biases
We minimised reporting bias from non‐publication of studies or selective outcome reporting by using a broad search strategy, by contacting study authors directly and by checking references of included studies. We planned to assess reporting bias by visual inspection of funnel plots.
Data synthesis
We combined dichotomous data using the Mantel‐Haenszel OR with 95% CIs by using a fixed‐effect model. We combined rate ratios and hazard ratios using GIV in a fixed‐effect model and compared them with the random‐effects model. We planned to calculate the NNTB outcome from the pooled OR and its CI, and to apply appropriate levels of baseline risk. We have presented the findings of our primary outcomes in summary of findings Table for the main comparison, which we generated by using GradePro software.
We did not perform pooled calculations of economic data. Rather, we presented the characteristics and results of included economic studies in a descriptive way in the additional tables (Table 1; Table 2), including the final incremental cost‐effectiveness ratios (ICERs) reported by study authors in Euros (EUR). We did not adjust the values of ICERs provided by study authors because most identified studies were conducted in similar settings and during a similar time period (2009 to 2010) using the same information resource, as all are based on the same clinical trial data.
| Study ID | Country | Study design | Population (N participants, severity indicators, smoking history | Economic outcomes | Interventiona (doses) | Perspective | Price year | Time horizon | ICER reported and adjusted to Euros 2014 |
| Canada | Cost utility | Based on the Aaron 2007 study: N = 449, patients with ≥ 1 exacerbation within 12 months, moderate and severe obstruction defined as FEV1 < 65% post‐bronchodilator, ≥ 10 pack‐years | Incremental cost per exacerbation avoided and incremental cost per QALY with tiotropium + LABA/ICS relative to tiotropium | • Tiotropium 18 mcg once daily + placebo twice daily • Tiotropium 18 mcg once daily + FS 250/25 mg/puff, 2 puffs twice daily | Healthcare system perspective | 2006 | 1 year | Per exacerbation avoid CAN$6510 Per QALY CAN$243180 | |
| Australia, Canada and Sweden | CEA | Based on the Welte 2009 study: N = 659, aged ≥ 40 years, symptoms for ≥ 2 years, ≥ 1 exacerbation within 12 months requiring systemic steroids and/or antibiotics, FEV1 ≤ 50% of predicted normal, FEV1 /FVC < 70% pre‐dose, ≥ 10 pack‐years | Incremental cost‐effectiveness ratio for exacerbation avoided with tiotropium + LABA/ICS relative to tiotropium | • Tiotropium (Handihaler) 18 mcg once daily + budesonide/formoterol (Symbicort Turbuhaler) 320/9 mcg one inhalation twice daily • Tiotropium 18 mcg once daily + placebo (identical Turbuhaler) twice daily | Healthcare system payer perspective | 2009 | 3 months | Per avoiding severe exacerbation: Australia: tiotropium + LABA/ICS dominant Canadian: tiotropium + LABA/ICS dominant Sweden: 244,36 EUR | |
| Denmark, Finland, Norway and Sweden | CEA | Based on the Welte 2009 study: N = 659, aged ≥ 40 years, clinical diagnosis of COPD and symptoms ≥ 2 years, ≥ 1 exacerbation in the previous 12 months requiring systemic steroids and/or antibiotics, FEV1 ≤ 50% predicted normal value, FEV1 /FVC < 70% pre‐dose, ≥10 pack‐years | Incremental cost‐effectiveness ratio for exacerbation avoided with tiotropium + LABA/ICS relative to tiotropium | • Tiotropium (Handihaler) 18 mcg once daily + budesonide/formoterol (Symbicort Turbuhaler) 320/9 mcg one inhalation twice daily • Tiotropium 18 mcg once daily + placebo (identical Turbuhaler) twice daily | Healthcare system payer perspective | 2010 | 3 months | ICER excluding antibiotics Denmark: 212 EUR Finland: 307 EUR Norway: tiotropium + LABA/ICS dominant Sweden: 165 EUR |
aTiotropium was compared with tiotropium + budesonide/formoterol in all included economic evaluations.
CEA: cost‐effectiveness analysis
FEV1: forced expiratory volume in one second
FVC: forced vital capacity
ICER: incremental cost‐effectiveness ratio
ICS: inhaled corticosteroids
LABA: long‐acting beta‐agonists
QALY: quality‐adjusted life‐year
| Study ID | Well‐defined question? | Competing alternatives described? | Effectiveness established? | Relevant costs and consequences identified? | Costs and consequences measured accurately? | Costs and consequences valued credibly? | Discounting performed? | Incremental analysis of costs and consequences performed? | Sensitivity analysis performed? |
| Yes | Yes | Yes | Yes | Yes | Yes | No | Yes | Yes | |
| Yes | Yes | Yes | Yes | Yes | Yes | No | Yes | Yes | |
| Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Subgroup analysis and investigation of heterogeneity
We planned to perform subgroup analyses for effectiveness and safety data on the basis of types of combination therapy and differences in baseline risk (severity of disease at baseline), provided at least three studies per subgroup were included in a specific comparison. However, included studies did not provide data for these subgroup analyses. Included studies reported outcomes at different follow‐up periods, and different follow‐periods may be associated with different treatment effects; therefore, we decided to include three subgroup analyses: at three‐month follow‐up; at six‐month follow‐up; and at 12‐month follow‐up.
Sensitivity analysis
The sensitivity analysis takes into account biases that could significantly impact the outcomes of included studies. We planned to perform a sensitivity analysis to assess how results of the meta‐analysis would be affected by excluding studies determined to be at a high risk of bias. Two studies (Hoshino 2011; Jung 2012) were open‐label studies; therefore we performed a sensitivity analysis for outcomes of quality of life (QoL) and for all severe adverse events (non‐fatal).
Results
Description of studies
Results of the search
The initial search carried out in July 2010 yielded 101 references, from which only three studies (Aaron 2007; Cazzola 2007; Welte 2009) were included and one RCT was classified as awaiting assessment (Fang 2008). Details of the search results from the previous review are described in Appendix 3. We updated these searches in April 2015 and identified 250 new references from July 2010. Of these, 13 references were selected as potentially relevant and underwent full‐text review. Three new studies (Hanania 2011; Hoshino 2011; Jung 2012) and three economic analyses (Mittmann 2011; Najafzadeh 2008; Nielsen 2013) met the criteria for inclusion, two RCTs were classified as ongoing studies (Betsuyaku 2013; Cohuet 2013) and one RCT was classified as awaiting assessment (Lee 2014) because only results from the abstract presented at a scientific meeting were available. We excluded four studies (see the Excluded studies section). We did not find published results of the Fang 2008 study for this review; therefore, it is still awaiting assessment.
In summary, for the clinical effectiveness objective we included a total of six studies (Aaron 2007; Cazzola 2007; Hanania 2011; Hoshino 2011; Jung 2012; Welte 2009), randomising 1902 participants to comparisons of interest for the review. We included three studies (Mittmann 2011; Najafzadeh 2008; Nielsen 2013) for the economic evaluation objective related to cost‐effectiveness (Figure 1).
Study flow diagram.
Included studies
Randomised controlled trials (RCTs)
We included a total of six RCTs in this review. All included studies (Aaron 2007; Cazzola 2007; Hanania 2011; Hoshino 2011; Jung 2012; Welte 2009) assessed the effectiveness of tiotropium plus LABA/ICS in comparison with tiotropium, and just one study (Cazzola 2007) also compared the effectiveness of tiotropium plus LABA/ICS versus LABA/ICS (see the Characteristics of included studies table and Table 3). Not all of the participants enrolled in the included studies were eligible for triple therapy according to the current guidance (GOLD 2015).
| Study ID | Baseline COPD severity | Baseline FEV1% predicted | Exacerbation status preceding enrolment |
| Moderate or severe | Inclusion criteria < 65% predicted Mean FEV1% predicted tiotropium + LABA/ICS group 42.2% and in tiotropium group 42.1% | Required to have had ≥ 1 exacerbation in previous year | |
| Severe and very severe | Inclusion criteria ≤ 50% predicted Mean FEV1% predicted LABA/ICS group = 36.9%, tiotropium group = 38.5%, tiotropium + LABA/ICS group = 39% | No information on exacerbation status before enrolment | |
| Moderate: tiotropium + LABA/ICS = 63% and tiotropium = 72% Severe: tiotropium + LABA/ICS = 37% and tiotropium = 28% | Inclusion criteria ≥ 40 to ≤ 80% predicted Mean FEV1% predicted tiotropium + LABA/ICS group = 56% and tiotropium group = 57.4% | Exacerbations in past 12 months requiring antibiotics/corticosteroid: tiotropium + LABA/ICS: 1 exacerbation = 37%, ≥ 2 exacerbations = 6% Tiotropium: 1 exacerbation = 27%, ≥ 2 exacerbations = 6% | |
| Mild to very severe Overall: mild = 10%, moderate = 33%, severe = 36.6%, very severe = 20%. | Mean FEV1% predicted tiotropium + LABA/ICS group = 64.6% and tiotropium group = 57.1% | Participants were excluded if they had experienced an exacerbation in the previous 12 weeks before commencement of the study. No other information on exacerbation status | |
| Moderate: tiotropium + LABA/ICS = 56.6% and tiotropium = 60.3% Severe: tiotropium + LABA/ICS = 40.8% and tiotropium = 35.5% Very severe: tiotropium + LABA/ICS = 2.7% and tiotropium = 3.5% | Inclusion criteria < 65% predicted Mean FEV1% predicted tiotropium + LABA/ICS group = 47.4% and tiotropium group = 47.5% | No information on exacerbation before enrolment | |
| Severe and very severe | Inclusion criteria ≤ 50% predicted Mean FEV1% predicted tiotropium + LABA/ICS group = 38.1% and tiotropium group = 37.7% | Required to have had ≥ 1 exacerbation in previous year Mean exacerbations last year: 1.4 for both groups, range 1‐7 |
COPD: chronic obstructive pulmonary disease
FEV1: forced expiratory volume in one second
LABA: long‐acting beta‐agonists
ICS: inhaled corticosteroids
Aaron 2007 is a randomised, double‐blind, placebo‐controlled study that was conducted in Canada with the aim of comparing the safety and effectiveness of tiotropium plus placebo versus tiotropium plus fluticasone plus salmeterol as one‐year maintenance therapy in participants with moderate or severe COPD. This study included 449 participants (301 in a comparison of interest in this review) older than 35 years of age, with a forced expiratory volume in one second (FEV1)–to‐forced vital capacity (FVC) ratio of less than 0.70 and a post‐bronchodilator FEV1 of less than 65% of the predicted value. Participants had to have at least one exacerbation of COPD that required treatment with systemic steroids or antibiotics within the 12 months before randomisation. Participants were assigned to receive tiotropium plus placebo or tiotropium plus fluticasone‐salmeterol. Measures of efficacy included the proportion of participants who experienced an exacerbation of COPD that required treatment with systemic steroids or antibiotics, lung function, disease‐specific quality of life, number of hospitalisations for COPD exacerbations and all‐cause hospitalisations. It is likely that most participants in this study would be eligible for triple therapy according to current guidance (GOLD 2015).
Cazzola 2007 is a randomised, double‐blind, double‐dummy, parallel‐group study that was conducted in Italy to compare the efficacy and safety of three treatments for 12 weeks: (1) fluticasone/salmeterol (FSC) 500/50 mg Diskus, one inhalation twice daily + placebo Handihaler, one inhalation once daily; (2) tiotropium 18 mg Handihaler, one inhalation once daily + placebo Diskus, one inhalation twice daily; (3) FSC 500/50 mg Diskus, one inhalation twice daily + tiotropium 18 mg Handihaler, one inhalation once daily. This study included 90 participants 50 years of age or older with well‐controlled severe or very severe COPD (FEV1% predicted ≤ 50%) who were current or former smokers (20 or more pack‐years) and were randomised to receive FSC, tiotropium or their combination. Study authors provided no information on the exacerbation status of participants during the year before enrolment. The primary efficacy measure was the mean change from baseline in pre‐dose FEV1 after three‐months of treatment. Secondary efficacy measures included change from baseline in the validity assessment score (VAS) assessing dyspnoea and supplemental salbutamol. It is likely that most participants in this study would be eligible for triple therapy according to current guidance (GOLD 2015).
Hanania 2011 is a randomised, double‐blind, parallel‐group, multi‐centre study of 24 weeks, conducted at 33 centres in the USA to compare the efficacy and safety of FSC (250/50 mcg twice daily) when added to tiotropium (18 mcg once daily) in participants with symptomatic moderate to severe COPD. The study included 342 participants who were 40 years of age or older with a cigarette smoking history ≥ 10 pack‐years and with a diagnosis of COPD and post‐bronchodilator FEV1 ≥ 40% to ≤ 80% of predicted normal and FEV1/FVC of 0.70. In the year before enrolment, 43% of participants in the tiotropium plus FSC group had experienced at least one exacerbation, and 33% in the tiotropium alone group. Participants were randomised in a 1:1 double‐blind fashion to open‐label tiotropium 18 mcg once daily plus FSC 250/50 mcg twice daily or open‐label tiotropium 18 mcg once daily plus placebo twice daily. Measures of efficacy included evaluation of lung function (pre‐dose FEV1, post‐dose FEV1, pre‐dose FVC and post‐dose FVC), use of rescue medication, healthcare utilisation for COPD exacerbations, health status evaluated with domain scores on the Chronic Respiratory Disease Questionnaire‐Self Administered Standardised (CRQ‐SAS) and safety. It is likely that most participants in this study would not be recommended triple therapy according to current guidance (GOLD 2015).
Hoshino 2011 is a randomised, open‐label, parallel‐group study conducted in Japan with the aim of comparing the efficacy and tolerability of salmeterol/fluticasone propionate added to tiotropium for 12 weeks. This study included 30 participants with an FEV1/FVC ratio less than 0.70, a smoking history > 10 pack‐years and no history of asthma or atopy. Eligible participants had mild to very severe COPD and were newly diagnosed or had not been treated previously with LAMA, LABA or ICS. Investigators provided no information on the exacerbation status of participants in the year before enrolment. They were randomised to receive inhaled tiotropium once daily or inhaled SFC twice daily, in combination with tiotropium once daily, for 12 weeks. Measures of efficacy included changes in airway dimensions on computed tomography (CT), pulmonary function testing and assessments of health‐related quality of life using the SGRQ. It is unclear what proportion of participants in this study would be eligible for triple therapy according to current guidance (GOLD 2015).
Jung 2012 is a randomised, open‐label, multi‐centre two‐arm parallel‐group study conducted in 30 academic hospital‐based pulmonary clinics in Korea with the aim of comparing the efficacy of tiotropium (18 mg once daily) plus FSC (250/50 mg twice daily) versus tiotropium monotherapy. This study included 479 participants diagnosed with moderate to very severe COPD, who had a post‐bronchodilator FEV1/FVC ratio less than 0.70 and FEV1 less than 65% of predicted value; eligible participants were 40 to 80 years of age and had a smoking history of at least 10 pack‐years. Investigators provided no information on the exacerbation status of participants in the year before enrolment. Participants were randomised to one of two treatment groups for 24‐week treatment: tiotropium 18 mg once daily; or tiotropium 18 mg once daily plus FSC, 250/50 mg/puff, one puff twice daily. Measures of efficacy included evaluation of lung function (change in pre‐bronchodilator FEV1 (L); changes in pre‐bronchodilator inspiratory capacity (IC); FVC and percent predicted (% pred) values for FEV1); mean changes in health‐related quality of life; frequency of COPD exacerbations; exacerbations requiring hospitalisation, emergency room visits or outpatient clinic visits; and hospitalisation rates for all causes. It is unclear what proportion of participants in this study would be recommended triple therapy according to current guidance (GOLD 2015), but likely it would be less than half.
Welte 2009 is a randomised, double‐blind, parallel‐group, multi‐centre study conducted to compare the efficacy and tolerability of budesonide/formoterol added to tiotropium for 12 weeks. This study included 660 participants with severe or very severe COPD, with a pre‐bronchodilator FEV1 not exceeding 50% of predicted normal value and a history of at least one exacerbation requiring systemic steroids and/or antibiotics in the previous year. Participants were randomised to receive tiotropium 18 mg once daily plus budesonide/formoterol 320/9 mg one inhalation twice daily or placebo twice daily. Measures of efficacy included clinic assessment of lung function and health status (change in pre‐dose FEV1, pre‐dose and post‐dose spirometry measurements and SGRQ for COPD), morning lung function assessments, COPD symptoms and morning activities, use of reliever medication, exacerbations and tolerability. It is likely that most of the participants in this study would be eligible for triple therapy according to current guidance (GOLD 2015).
Economic evaluation analysis
Of the three economic analyses included, two (Mittmann 2011; Nielsen 2013) reported on the economic evaluation conducted alongside the Welte 2009 clinical trial (the CLIMB trial) in six of the nine participant countries; Nielsen 2013 conducted the economic evaluation in four Nordic countries (Sweden, Denmark, Finland, Norway) and Mittmann 2011 in three countries (Canada, Australia, Sweden); both study authors reported on the incremental cost‐effectiveness ratio for exacerbation avoided with tiotropium + LABA/ICS relative to tiotropium from the healthcare system perspective. Najafzadeh 2008 conducted the economic evaluation from the Canadian healthcare system perspective alongside the Aaron 2007 study (OPTIMAL trial) and reported on the incremental cost‐effectiveness ratio per exacerbation avoided and the incremental cost‐effectiveness ratio per quality‐adjusted life‐year (QALY). Investigators calculated the utilities used for the cost‐utility analysis from the results of SGRQ as applied to trial participants.
We have presented detailed characteristics of these economic evaluations in Table 1.
Excluded studies
The initial search carried out in July 2010 revealed eight studies that failed to meet eligibility criteria for the review (see the Characteristics of excluded studies table). Four of these compared tiotropium alone with combination therapy (Ando 2008; Bateman 2008; Golabi 2006; Hara 2007), and one study compared tiotropium with LABA alone (Petroianni 2008). The remaining three studies were shorter than three months in duration (Biscione 2009; Perng 2006), and one used a cross‐over design (Singh 2008).
Searches updated to April 2015 identified seven studies that failed to meet eligibility criteria for the review (see the Characteristics of excluded studies table). Two of these evaluated tiotropium versus placebo (Tashkin 2008; Troosters 2008), and one study compared tiotropium alone versus the LABA/ICS combination (Sarac 2013). One study (Maltais 2013) was shorter than three months, and four were added to Studies awaiting classification and Ongoing studies.
Risk of bias in included studies
We have presented the assessment of risk of bias in the Characteristics of included studies table, and an overview of the findings in Figure 2.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
We have summarised results of the quality assessment of economic evaluation studies in Table 2. We assessed all included economic evaluations according to their full‐text publications. In general, the economic evaluations included met the methodological and reporting aspects evaluated by the Drummond checklist, and their results can be considered valid (Drummond 1996). In the economic evaluations conducted by Mittmann 2011, Nielsen 2013 and Najafzadeh 2008, discounting was not applied to costs and consequences. However, this was considered to be methodologically correct because the time horizon used in these analyses was three months and one year, respectively, making discounting unnecessary.
Allocation
Aaron 2007, Welte 2009 and Jung 2012 reported adequate sequence generation and allocation concealment. Details for Welte 2009 were supplied on request. For these three studies, randomisation was computer‐generated through central allocation, and both research staff and participants were blinded to the treatment assignment until the end of the study. For Cazzola 2007, Hanania 2011 and Hoshino 2011, sequence generation and/or allocation concealment is unclear because study authors did not report full details and did not respond to personal communication.
Blinding
Aaron 2007, Hanania 2011 and Welte 2009 performed a blinded outcome assessment. In the trial arms of Aaron 2007, inhalers containing placebo and fluticasone/salmeterol were identical in taste and appearance, and they were enclosed in identical tamper‐proof blinding devices. Medication canisters within the blinding devices were stripped of identifying labelling. Clinical data for suspected exacerbations were reviewed by a blinded committee to judge whether data met the study definition of COPD exacerbation. Blinding of participants was not broken for participants who prematurely discontinued treatment with study medications, and the statistician who performed the analysis was initially blinded to participant group assignments. In Hanania 2011, the DISKUS inhalers containing placebo and fluticasone‐salmeterol were identical in taste and appearance. In Welte 2009, treatment assignment was concealed, as active and placebo inhalers were of identical appearance and both clinicians and participants were blinded to treatment until completion of the study. Hoshino 2011 was an open study; however, for evaluation of airway dimensions, a single observer, who was blind to all participant data, measured the outcome. Cazzola 2007 and Jung 2012 did not report details of the outcome assessment; therefore blinding is unclear for these studies.
Incomplete outcome data
Cazzola 2007, Jung 2012 and Welte 2009 reported comparable attrition rates (< 14%) for both intervention and control groups; reasons for attrition were provided in all cases, making the risk of bias low. In Aaron 2007, withdrawal rates were different between intervention groups (74 participants (47%) withdrew from the tiotropium + placebo group, and 37 participants (26%) from the tiotropium + LABA/ICS group); however, mortality data were obtained for all participants, apart from six (2/145 on tiotropium + LABA/ICS and 4/156 on tiotropium + placebo) who withdrew and declined to be involved further in the study; therefore, we rated risk of bias as unclear. Hanania 2011 had high withdrawal rates in both groups (21% in fluticasone/salmeterol (FSC) + tiotropium group and 25% in tiotropium + placebo group); reasons for attrition were provided and were similar among groups; therefore, we rated this study as having an unclear risk of bias. In Hoshino 2011, a total of 36 participants were enrolled in the study, but only 30 were included in the analysis (16 participants on FSC + tiotropium and 14 on tiotropium + placebo); therefore the withdrawal rate was 20% because of loss to follow‐up, making risk of attrition bias unclear.
Selective reporting
All six studies adequately reported outcome data for primary and secondary outcomes that were pre‐specified in the study record.
Effects of interventions
Because of the small number of eligible studies for the two comparisons (tiotropium + LABA/ICS vs tiotropium alone and vs LABA/ICS alone), no subgroup analysis by disease severity or by type of combination therapy was possible.
Comparison 1. Tiotropium plus LABA/ICS versus tiotropium
We identified six RCTs addressing the comparison of tiotropium + LABA/ICS versus tiotropium + placebo (Aaron 2007; Cazzola 2007; Hanania 2011; Hoshino 2011; Jung 2012; Welte 2009).
Primary outcomes
All‐cause mortality
Two studies (Aaron 2007; Welte 2009) reported mortality at three months and 12 months of follow‐up, respectively. Both studies recruited participants who, on average, were likely to have fulfilled current GOLD criteria for triple therapy (GOLD 2015). These two studies did not find a significant effect on mortality with the use of tiotropium + LABA/ICS compared with tiotropium + placebo. Meta‐analysis of these studies showed a non‐statistically significant trend towards reduced risk of mortality with the use of tiotropium + LABA/ICS (two studies; 961 participants; OR 1.80, 95% CI 0.55 to 5.91; I2 = 0%). The quality of evidence for this outcome is moderate because of imprecision in estimates of effect (summary of findings Table for the main comparison).
Hospital admission (all causes)
The same two studies (Aaron 2007; Welte 2009) reported on all causes of hospital admission at three months and 12 months of follow‐up, respectively. Welte 2009 did not find a significant difference in hospital admissions at three months of follow‐up. Aaron 2007 found a statistically significant reduction in hospital admission at 12 months of follow‐up with the use of tiotropium + LABA/ICS. Meta‐analysis of these studies (Figure 3) showed a non statistically significant trend towards to decreased risk of hospital admission associated with the use of tiotropium + LABA/ICS (two studies; 961 participants; OR 0.84, 95% CI 0.53 to 1.33; I2 = 0%); the quality of evidence for this outcome is low because of the risk of bias in included studies and imprecision of the estimate (summary of findings Table for the main comparison).
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.2 Hospital admission (all causes).
Exacerbations
Three studies (Aaron 2007; Jung 2012; Welte 2009) reported on exacerbations at three, six and 12 months of follow‐up, respectively. Welte 2009 found a significant difference in exacerbations at three months of follow‐up with the use of combined therapy tiotropium + LABA/ICS (one study; 660 participants; OR 0.36, 95% CI 0.22 to 0.60). Jung 2012 did not find a significant difference in exacerbations at six months of follow‐up (one study; 479 participants; OR 0.83, 95%CI 0.52 to 1.34). Aaron 2007 did not find a statistically significant reduction in exacerbations at 12 months of follow‐up with the use of combined therapy tiotropium + LABA/ICS (one study; 301 participants; OR 0.89, 95% CI 0.56 to 1.41). We did not pool study results because statistical heterogeneity across studies was considerable. Of note, Jung 2012 may have recruited a population of participants with less severe COPD, not all of whom would be recommended triple therapy according to current guidelines. The quality of evidence for this outcome is low because of the risk of bias and imprecision in estimates of effect (summary of findings Table for the main comparison).
Quality of life
Four studies (Aaron 2007; Hoshino 2011; Jung 2012; Welte 2009) reported on quality of life using the SGRQ (Meguro 2007) at six months. The meta‐analysis of these studies showed a statistically significant difference in quality of life (SGRQ total score) in favour of combined therapy of tiotropium + LABA/ICS compared with tiotropium + placebo (Figure 4) (four studies; 1446 participants; MD ‐3.46, 95% CI ‐5.05 to ‐1.87; I2 = 16%). Only one study reported on the percentage of participants who were responders to treatment. Welte 2009 reported the percentage of participants with improvement in SGRQ score greater than four units, which was significantly higher in the tiotropium + LABA/ICS group (49.5%) than in the tiotropium + placebo group (40.0%) (P value = 0.016). The percentage of participants who showed deterioration in SGRQ score greater than four units was similar in the two groups (tiotropium + LABA/ICS 27.6%, tiotropium + placebo 29.7%).
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.4 Quality of life up to 6 months (SGRQ).
The sensitivity analysis excluding two open‐label studies, which also included participants with less severe COPD (Hoshino 2011; Jung 2012), revealed no changes in the direction of treatment effect (two studies; 961 participants; MD ‐2.5, 95% CI ‐4.16 to ‐0.84; I2 = 0%).
The quality of evidence for this outcome is low because of very serious risk of bias in the trial design (summary of findings Table for the main comparison).
Exercise tolerance
None of the included studies reported exercise tolerance as an outcome.
Serious adverse events non‐fatal (all reported)
Four studies (Aaron 2007; Hanania 2011; Jung 2012; Welte 2009) reported on serious adverse events (non‐fatal). Aaron 2007 reported no differences in serious adverse events between intervention groups; a total of 19 serious adverse events not related to COPD (respiratory failure, cancer and myocardial infarction or acute arrhythmia) were reported in both intervention groups, and one case of pneumonia in the combined therapy group. Jung 2012 reported no differences in serious adverse event rates between trial arms; the event most commonly reported in the combined therapy group was productive cough, whereas dyspnoea was the most common event in the tiotropium group. Two cases of pneumonia were reported in each intervention group. Hanania 2011 reported no statistically significant differences in serious adverse event rates between therapy groups; two cases of pneumonia were reported in the combined therapy group, and nobody in the tiotropium presented with pneumonia. In Welte 2009, six cases of pneumonia were reported in each trial arm as serious adverse events, representing < 1% of the total adverse events reported by the trial.
Meta‐analysis for all non‐fatal serious adverse events reported in these studies showed no statistically significant differences (four studies; 1758 participants; OR 0.86, 95% CI 0.57 to 1.30; I2 = 9%) (Figure 5). The sensitivity analysis excluding the open‐label study (Jung 2012) revealed no differences in the treatment effect estimation (three studies; 1303 participants; OR 0.67, 95% CI 0.4 to 1.13; I2 = 0%). Exclusion of both Hanania 2011 and Jung 2012 from the meta‐analysis on the basis of the less severe population recruited widened the CIs but had little impact on size and direction of the effect.
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.7 Serious adverse events all reported (non‐fatal).
The quality of evidence for this outcome is low because of risk of bias and imprecision in estimates of effect (summary of findings Table for the main comparison).
Independent meta‐analysis for pneumonia including four studies (Aaron 2007; Hanania 2011; Jung 2012; Welte 2009) revealed no statistically significant differences in effects on pneumonia between treatments (four studies; 1758 participants; Peto OR 1.62, 95% CI 0.54 to 4.82; I2 = 0%).
Secondary outcomes
Symptoms
Welte 2009 was the only included study that reported changes in COPD symptom scores for breathlessness (MD ‐0.142, 95% CI ‐0.214 to ‐0.069), night awakening (MD ‐0.157, 95% CI ‐0.222 to ‐0.092), chest tightness (MD ‐0.142, 95% CI ‐0.212 to ‐0.072) and cough (MD ‐0.161, 95% CI ‐0.238 to ‐0.084) among 660 participants. Scores for all symptoms favoured the tiotropium + LABA/ICS group compared with the tiotropium + placebo group.
Forced expiratory volume in one second (FEV1)
Six studies (Aaron 2007; Cazzola 2007; Hanania 2011; Hoshino 2011; Jung 2012; Welte 2009) reported mean change in FEV1 at three to six months. These studies found statistically significant changes in FEV1 with the use of tiotropium + LABA/ICS compared with tiotropium + placebo (four studies; 1678 participants; MD 0.06, 95% CI 0.04 to 0.08; I2 = 0%; Figure 6); however, these changes are not clinically significant. The quality of evidence for this outcome is moderate as a result of the risk of bias.
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.4 FEV1 pre‐dose.
Hoshino 2011 was excluded from the analysis because all data were not available. Hanania 2011 was not included in the pooled estimates of effect because this study generated significant heterogeneity (I2 > 90%). We considered the statistical heterogeneity to result from differences in baseline risk, as participants in Hanania 2011 had on average a greater degree of dyspnoea (modified Medical Research Council scale (MRCm) > 2), which may have resulted in a greater response to pharmacological management.
Exclusion of Jung 2012 due to concerns about the relevance of the recruited population to current guidelines on triple therapy had a minimal impact on size and direction of the effect estimate (GOLD 2015).
Adverse events (not serious)
Four studies (Aaron 2007; Cazzola 2007; Hanania 2011; Welte 2009) reported adverse events. These studies did not find statistically significant differences with the use of tiotropium + LABA/ICS compared to tiotropium + placebo. Meta‐analysis of these studies did not show a statistically significant difference (four studies; 1363 participants OR 1.16, 95% CI 0.92 to 1.47, I2 = 0%). Removing Hanania 2011, a study in which the participants may not, on average, have fulfilled current GOLD criteria for receiving triple therapy, had little impact on the outcome (GOLD 2015).
Cost‐effectiveness of interventions
In the economic evaluation conducted by Mittmann 2011; Nielsen 2013 Tiotropium+ LABA/ICS was the dominant strategy (less costly and more effective) in three of the countries where it was assessed, in comparison with Tiotropium (Canada, Australia and Norway) (Table 1). In all other countries including Sweden, Finland, Denmark and Norway the ICER per exacerbation avoided was under the established willingness to pay threshold (between 600 to 1000 EUR); therefore the Tiotropium + LABA/ICS therapy resulted in a cost‐effective alternative. Sensitivity analyses in both studies indicated that the variables with the largest effect on the ICER were hospitalisation costs, the incidence of exacerbations and hospital admission‐related costs.
In the study conducted by Najafzadeh 2008 Tiotropium + LABA/ICS showed significantly better quality of life and less hospital admissions than Tiotropium alone, but this improvement in health outcomes was associated with increased costs. The Tiotropium therapy showed the highest probability of being cost‐effective when the "willingness to pay" is CAN $6.000°°; when QALY was used as the effectiveness outcome, treatment with Tiotropium had the highest probability of being the best option compared to Tiotropium + LABA/ICS.
Comparison 2. Tiotropium plus LABA/ICS versus LABA/ICS
Cazzola 2007 was the only eligible study identified that compared tiotropium + LABA/ICS versus LABA/ICS + placebo (60 participants) and recruited those with severe or very severe COPD at baseline, most of whom were likely to have met current GOLD criteria for triple therapy (GOLD 2015). This study reported results for the following outcomes of interest for this review.
Primary outcome
Mortality (all‐cause)
Cazzola 2007 reported zero serious adverse events; therefore, we assumed that no deaths occurred during the study.
Secondary outcomes
Forced expiratory volume in one second (FEV1)
Tiotropium in combination with LABA/ICS improves FEV1 significantly compared with LABA/ICS + placebo (MD 0.05, 95% CI 0.00 to 0.09), but MD and CI were below the minimally clinically important difference of 100 to 140 mL.
Serious adverse events (non‐fatal)
No serious adverse events were reported in either intervention group.
Adverse events
More adverse events were reported in the tiotropium + LABA/ICS group (15/30) than in the tiotropium + placebo group (8/30), but the CI was wide because of the small numbers of participants (OR 2.75, 95% CI 0.93 to 8.10).
Withdrawal
Fewer withdrawals were reported in the tiotropium + LABA/ICS group (1/30) than in the tiotropium + placebo group (4/30), but the number of events was small and was not statistically significant (OR 0.22, 95% CI 0.02 to 2.14).
Discussion
Summary of main results
This systematic review set out to investigate the long‐term (≥ three months) effects of tiotropium in combination with long‐acting beta‐agonists/inhaled corticosteroids (LABA/ICS) compared with either LABA/ICS alone or tiotropium alone for the treatment of chronic obstructive pulmonary disease (COPD). We identified six randomised controlled trials (RCTs). All six studies looked at the effects of combination therapy (tiotropium + LABA/ICS) compared with tiotropium alone, whereas only one of these studies (Cazzola 2007) compared triple therapy versus LABA/ICS alone. Additionally, we included three cost‐effectiveness analyses based on data from two of these studies (Aaron 2007; Welte 2009) for evaluation of economic outcomes.
Tiotropium + LABA/ICS versus tiotropium
Since the first version of this systematic review was published, three clinical trials comparing tiotropium + LABA/ICS versus tiotropium alone have been published (Hanania 2011; Hoshino 2011; Jung 2012). These three trials reported on quality of life; exacerbations, FEV1 and non serious adverse events but do not report on hospital admission and mortality; therefore the evidence for the last two outcomes remain the same that failed to show a statistically significant difference in mortality and hospital admission between tiotropium + LABA/ICS versus tiotropium alone (moderate and low quality of evidence). Participants recruited to the two studies included in the analyses of hospitalisations and mortality (Aaron 2007; Welte 2009) are likely to have been candidates for tiotropium + LABA/ICS therapy according to current guidance GOLD 2015 (i.e. forced expiratory volume in one second (FEV1) < 50% predicted and frequent exacerbations ), suggesting that these findings are clinically applicable.
Even though investigators found statistically significant differences between treatment arms for other important outcomes such as FEV1 and quality of life, these results must be interpreted with caution, as the differences found may not be clinically significant. According to Jones 2005, the minimal clinically important difference (MCID) for FEV1 may vary but is accepted to be within the range of 100 to 140 mL (American Thoracic Society/European Respiratory Society Task Force). In this review, the difference in treatment effect on FEV1 was 60 mL; this difference did not reach the MCID needed to have a beneficial impact on participants' quality of life (Jones 2005). Similarly, the MCID in quality of life scores evaluated with St. George's Respiratory Questionnaire (SGRQ) has been proposed to be four units of improvement; the meta‐analysis for this outcome showed a difference smaller than a four‐unit change that could be reached just as part of a Hawthorne effect. Westwood et al described that in participants with COPD, a Hawthorne effect influences SGRQ scores in COPD trials; typically, this results in improvement of two to three points on the SGRQ with placebo (Westwood 2011). Welte 2009 reported a percentage of participants with improvement in SGRQ score greater than four units, which was significantly higher in the tiotropium + LABA/ICS group (49.5%) than in the tiotropium + placebo group (40.0%) (P value = 0.016). The percentage of participants with a decrease in SGRQ score greater than four units was similar in the two groups (tiotropium + LABA/ICS 27.6%, tiotropium + placebo 29.7%).
Of note, Hoshino 2011 and Jung 2012 contributed data to the quality of life analysis, and Jung 2012 to the FEV1 analysis, but on average, participants included in these studies may not have been candidates for tiotropium + LABA/ICS therapy according to current guidelines (GOLD 2015). However, removing these studies from the analyses had little impact on the size or direction of the effect estimate.
We did not pool data on exacerbations reported in these studies, as we considered that several sources of variation among the studies resulted in important heterogeneity (I2 > 80%) that could not be removed by subgroup analyses based on length of follow‐up (three, six and 12 months) (Aaron 2007; Jung 2012; Welte 2009) nor by the definition of exacerbation used (Aaron 2007 defined exacerbation as worsening of COPD leading to treatment with systemic steroids and/or antibiotics; Jung 2012 cited the definition of Rodriguez‐Roisin: sustained worsening of the patient's condition from the stable state and beyond normal day‐to‐day variations that is acute in onset and necessitates a regular change in medication in a patient with underlying COPD; and Welte 2009 defined an exacerbation as worsening of COPD leading to treatment with systemic steroids and/or hospitalisation/emergency room visits). All these individual studies failed to show significant differences in exacerbations between the two treatment arms at six and 12 months of follow‐up. The study by Welte 2009 showed a significant reduction in the risk of exacerbation at three‐month follow‐up associated with triple therapy in comparison with tiotropium alone. However, this finding may be considered clinically irrelevant because the follow‐up period needed to define the real effect that any COPD treatment could have on exacerbation is 52 months (Cazzola 2008; Miravitlles 2004).
The effect of tiotropium + LABA/ICS combination treatment on mortality remains uncertain because of the small number of events. The difference in serious adverse event rates between intervention groups was not statistically significant. For pneumonia, the number of cases in each study was small compared with the number of withdrawals and the number of participants lost to follow‐up. Withdrawals did not seem to be linked to adverse events but rather to the efficacy of treatment. Even though use of inhaled corticosteroids has been associated with pneumonia, our findings suggest no safety concerns related to use of the tiotropium + LABA/ICS combination in the treatment of patients with COPD when compared with tiotropium alone.
Economic evaluation results show a high probability that tiotropium + LABA/ICS combination treatment could be a cost‐effective alternative in various settings, as it was associated with fewer hospital admissions and better quality of life, which may drive most of the long‐term costs associated with this condition. The two economic evaluations conducted in Canada (Mittmann 2011; Najafzadeh 2008) differ in the final conclusions presented regarding cost‐effectiveness of combined therapy (tiotropium + LABA/ICS); these differences may be explained by differences in willingness to pay thresholds used in the sensitivity analyses, making comparison of the conclusions of these two studies impossible.
LABA/ICS + tiotropium versus LABA/ICS + placebo
The one pilot study (Cazzola 2007; no publications after 2007) that looked at the effect of LABA/ICS + tiotropium versus LABA/ICS + placebo showed significantly greater improvement in FEV1 with tiotropium + LABA/ICS compared with LABA/ICS; however, the mean difference in FEV1 was not clinically significant. All other outcomes of interest were not studied, revealed no events or did not achieve a statistically significant difference.
Overall completeness and applicability of evidence
For the comparison of benefits and risks of treatment with tiotropium + LABA/ICS versus LABA/ICS, we identified just one small eligible study (Cazzola 2007), which did not look at, or report on, any of the primary outcomes specified in this review, except for mortality. Therefore, this review found little applicable evidence for this comparison.
Current international guidance suggests that triple therapy with long‐acting muscarinic agonists (LAMA), LABA and ICS should be reserved for patients who continue to have symptoms despite receiving dual therapy with either LABA/ICS or LABA + LAMA, have an FEV1 < 50% predicted and are at a high risk of experiencing exacerbations (i.e. ≥ two exacerbations in the preceding year) (GOLD 2015). Lack of detailed reporting of baseline characteristics has somewhat limited our ability to assess to what extent the studies included in this review recruited participants who would meet these criteria. This could potentially limit the generalisability of our findings to a clinical setting. Table 3 summarises the available information, and, when relevant, we performed a sensitivity analysis that excluded the study or studies in which investigators raised concerns about the relevance of the recruited population. This did not have a substantial impact on any of the effect estimates. We have also included in the individual analyses footnotes that detail the baseline characteristics of participants included in the analysis.
Quality of the evidence
Methods used for randomisation and outcome assessment in some of the included studies were not clearly described and in some cases explanations were missing, thus presenting a source of potential bias.
The quality of the evidence, according to the GRADE (Grades of Recommendation, Assessment, Development and Evaluation Working Group) framework, was moderate for mortality but low for hospital admissions and other important outcomes such as exacerbations, adverse events and quality of life (GRADE 2013).
Potential biases in the review process
The issue of large and/or uneven numbers of withdrawals, as mentioned above (Quality of the evidence), will, even if addressed, have a high likelihood of introducing selection bias, as no consensus has been reached on how best to handle participants for whom no data are available. The high drop‐off rates observed in these types of studies may have been a consequence of the long‐term follow‐up required to measure effectiveness outcomes.
We analysed available data as specified in the protocol. However, we expanded the review question from the protocol to include the comparison of tiotropium + LABA/ICS versus LABA/ICS + placebo. We also highlighted the percentage of participants with a clinically significant change in health‐related quality of life as reported by study authors, although this was not specified in Measures of treatment effect.
Agreements and disagreements with other studies or reviews
Liu 2014 and Rodrigo 2012 published reviews that evaluated the long‐term efficacy and adverse effects of tiotropium + LABA/ICS treatment compared with tiotropium. Both reviews revealed benefits for lung function, quality of life and exacerbation risk. Our review validates these findings. However, with respect to this last outcome, we believe it is not advisable to combine the results, given that length of follow‐up differed among the studies, that the only study that showed length of follow‐up to 52 weeks is Aaron 2007 and that results should be presented independently, as has been done in the present review. The present review also describes a beneficial impact on hospitalisation risk, another clinically relevant outcome that was not previously considered. Additionally, the current review presents results reflecting the GRADE method, which allows the reader to consider the quality of the evidence for each outcome ‐ a critical piece of information on which to base clinical decisions.
A systematic review looking at LABA/ICS combination treatment compared with placebo has shown that combination treatment significantly reduces mortality and exacerbation rates and improves lung function (Nannini 2013). LABA/ICS also increases the risk of pneumonia compared with placebo. A systematic review comparing tiotropium versus placebo showed that tiotropium treatment was associated with a significant improvement in participants' quality of life and reduced the risk of exacerbations, with a number needed to treat for an additional beneficial outcome (NNTB) of 16 to prevent one exacerbation. Tiotropium also reduced exacerbations leading to hospitalisation when compared with placebo (Karner 2014).
Study flow diagram.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.2 Hospital admission (all causes).
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.4 Quality of life up to 6 months (SGRQ).
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.7 Serious adverse events all reported (non‐fatal).
Forest plot of comparison: 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, outcome: 1.4 FEV1 pre‐dose.
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 1 Mortality (all‐cause).
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 2 Hospital admission (all causes).
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 3 Exacerbation.
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 4 Quality of life up to 6 months (SGRQ).
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 5 Sensitivity analysis ‐ QoL up to 6 months (SGRQ).
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 6 FEV1 pre‐dose.
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 7 Serious adverse events all reported (non‐fatal).
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 8 Pneumonia.
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 9 Sensitivity analysis ‐ SAE all reported (non‐fatal).
Comparison 1 Tiotropium + LABA/ICS combination versus tiotropium + placebo, Outcome 10 Adverse event.
Comparison 2 Tiotropium + LABA/ICS combination vs LABA/ICS combination + placebo, Outcome 1 FEV1 GIV.
Comparison 2 Tiotropium + LABA/ICS combination vs LABA/ICS combination + placebo, Outcome 2 Adverse event.
Comparison 2 Tiotropium + LABA/ICS combination vs LABA/ICS combination + placebo, Outcome 3 FEV1.
| Tiotropium + LABA/ICS combination compared with tiotropium for chronic obstructive pulmonary disease | ||||||
| Patient or population: patients with chronic obstructive pulmonary disease Settings: ambulatory clinics Intervention:tiotropium + LABA/ICS combination Comparison: tiotropium | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments | |
| Assumed risk Tiotropium | Corresponding risk Tiotropium + LABA/ICS combination | |||||
| Mortality (all‐cause) | 8 per 1000 | 15 per 1000 | OR 1.80 | 961 | ⊕⊕⊕⊝ | |
| Hospital admission (all causes) | 156 per 1000 | 101 per 1000 | OR 0.84 | 961 | ⊕⊕⊝ ⊝ | |
| Exacerbation ‐ at 12‐month follow‐up | 628 per 1000 | 601 per 1000 | OR 0.89 | 301 | ⊕⊕⊝⊝ | |
| Serious adverse events (non‐fatal) | 60 per 1000 | 52 per 1000 | OR 0.86 | 1758 | ⊕⊕⊝⊝ a,c | |
| Quality of life up to 6 months (SGRQ) | Mean SGRQ up to 6 months in the intervention groups was 3.46 lower (5.05 to 1.87 lower) | ‐ | (4 studies) | ⊕⊕⊝⊝ d | A lower score indicates better quality of life | |
| FEV1 pre‐dose ‐ FEV1 3‐6 months mean difference | Mean FEV1 pre‐dose ‐ FEV1 3‐6 months mean difference in the intervention groups was 0.06 (0.04 to 0.08 ) | ‐ | (4 studies) | ⊕⊕⊕⊝ | ||
| FEV1 pre‐dose ‐ FEV1 1 year | Mean FEV1 pre‐dose ‐ FEV1 1 year mean difference in the intervention groups was 0.06 (0 to 0.12 ) | ‐ | (1 study) | ⊕⊕⊕⊝ | ||
| *The basis for 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 | ||||||
| aDowngraded one level because of imprecision (95% confidence interval includes both no effect and appreciable harm) bDowngraded one level because of study limitations (incomplete outcome assessment in Aaron 2007) cDowngraded once because of study limitations (incomplete outcome assessment in Aaron 2007 and Hanania 2011; unclear risk of selection bias in Hanania 2011; possible detection bias in Jung 2012) dDowngraded two levels because of study limitations (unclear risk of selection bias and detection bias and incomplete outcome assessment in Hoshino 2011; unclear risk of detection bias in Jung 2012; incomplete outcome assessment in Aaron 2007) eDowngraded one level because of study limitations (unclear risk of selection and detection bias in Cazzola 2007; unclear risk of detection bias in Jung 2012; incomplete outcome assessment in Aaron 2007) | ||||||
| Study ID | Country | Study design | Population (N participants, severity indicators, smoking history | Economic outcomes | Interventiona (doses) | Perspective | Price year | Time horizon | ICER reported and adjusted to Euros 2014 |
| Canada | Cost utility | Based on the Aaron 2007 study: N = 449, patients with ≥ 1 exacerbation within 12 months, moderate and severe obstruction defined as FEV1 < 65% post‐bronchodilator, ≥ 10 pack‐years | Incremental cost per exacerbation avoided and incremental cost per QALY with tiotropium + LABA/ICS relative to tiotropium | • Tiotropium 18 mcg once daily + placebo twice daily • Tiotropium 18 mcg once daily + FS 250/25 mg/puff, 2 puffs twice daily | Healthcare system perspective | 2006 | 1 year | Per exacerbation avoid CAN$6510 Per QALY CAN$243180 | |
| Australia, Canada and Sweden | CEA | Based on the Welte 2009 study: N = 659, aged ≥ 40 years, symptoms for ≥ 2 years, ≥ 1 exacerbation within 12 months requiring systemic steroids and/or antibiotics, FEV1 ≤ 50% of predicted normal, FEV1 /FVC < 70% pre‐dose, ≥ 10 pack‐years | Incremental cost‐effectiveness ratio for exacerbation avoided with tiotropium + LABA/ICS relative to tiotropium | • Tiotropium (Handihaler) 18 mcg once daily + budesonide/formoterol (Symbicort Turbuhaler) 320/9 mcg one inhalation twice daily • Tiotropium 18 mcg once daily + placebo (identical Turbuhaler) twice daily | Healthcare system payer perspective | 2009 | 3 months | Per avoiding severe exacerbation: Australia: tiotropium + LABA/ICS dominant Canadian: tiotropium + LABA/ICS dominant Sweden: 244,36 EUR | |
| Denmark, Finland, Norway and Sweden | CEA | Based on the Welte 2009 study: N = 659, aged ≥ 40 years, clinical diagnosis of COPD and symptoms ≥ 2 years, ≥ 1 exacerbation in the previous 12 months requiring systemic steroids and/or antibiotics, FEV1 ≤ 50% predicted normal value, FEV1 /FVC < 70% pre‐dose, ≥10 pack‐years | Incremental cost‐effectiveness ratio for exacerbation avoided with tiotropium + LABA/ICS relative to tiotropium | • Tiotropium (Handihaler) 18 mcg once daily + budesonide/formoterol (Symbicort Turbuhaler) 320/9 mcg one inhalation twice daily • Tiotropium 18 mcg once daily + placebo (identical Turbuhaler) twice daily | Healthcare system payer perspective | 2010 | 3 months | ICER excluding antibiotics Denmark: 212 EUR Finland: 307 EUR Norway: tiotropium + LABA/ICS dominant Sweden: 165 EUR | |
| aTiotropium was compared with tiotropium + budesonide/formoterol in all included economic evaluations. CEA: cost‐effectiveness analysis FEV1: forced expiratory volume in one second FVC: forced vital capacity ICER: incremental cost‐effectiveness ratio ICS: inhaled corticosteroids LABA: long‐acting beta‐agonists QALY: quality‐adjusted life‐year | |||||||||
| Study ID | Well‐defined question? | Competing alternatives described? | Effectiveness established? | Relevant costs and consequences identified? | Costs and consequences measured accurately? | Costs and consequences valued credibly? | Discounting performed? | Incremental analysis of costs and consequences performed? | Sensitivity analysis performed? |
| Yes | Yes | Yes | Yes | Yes | Yes | No | Yes | Yes | |
| Yes | Yes | Yes | Yes | Yes | Yes | No | Yes | Yes | |
| Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Study ID | Baseline COPD severity | Baseline FEV1% predicted | Exacerbation status preceding enrolment |
| Moderate or severe | Inclusion criteria < 65% predicted Mean FEV1% predicted tiotropium + LABA/ICS group 42.2% and in tiotropium group 42.1% | Required to have had ≥ 1 exacerbation in previous year | |
| Severe and very severe | Inclusion criteria ≤ 50% predicted Mean FEV1% predicted LABA/ICS group = 36.9%, tiotropium group = 38.5%, tiotropium + LABA/ICS group = 39% | No information on exacerbation status before enrolment | |
| Moderate: tiotropium + LABA/ICS = 63% and tiotropium = 72% Severe: tiotropium + LABA/ICS = 37% and tiotropium = 28% | Inclusion criteria ≥ 40 to ≤ 80% predicted Mean FEV1% predicted tiotropium + LABA/ICS group = 56% and tiotropium group = 57.4% | Exacerbations in past 12 months requiring antibiotics/corticosteroid: tiotropium + LABA/ICS: 1 exacerbation = 37%, ≥ 2 exacerbations = 6% Tiotropium: 1 exacerbation = 27%, ≥ 2 exacerbations = 6% | |
| Mild to very severe Overall: mild = 10%, moderate = 33%, severe = 36.6%, very severe = 20%. | Mean FEV1% predicted tiotropium + LABA/ICS group = 64.6% and tiotropium group = 57.1% | Participants were excluded if they had experienced an exacerbation in the previous 12 weeks before commencement of the study. No other information on exacerbation status | |
| Moderate: tiotropium + LABA/ICS = 56.6% and tiotropium = 60.3% Severe: tiotropium + LABA/ICS = 40.8% and tiotropium = 35.5% Very severe: tiotropium + LABA/ICS = 2.7% and tiotropium = 3.5% | Inclusion criteria < 65% predicted Mean FEV1% predicted tiotropium + LABA/ICS group = 47.4% and tiotropium group = 47.5% | No information on exacerbation before enrolment | |
| Severe and very severe | Inclusion criteria ≤ 50% predicted Mean FEV1% predicted tiotropium + LABA/ICS group = 38.1% and tiotropium group = 37.7% | Required to have had ≥ 1 exacerbation in previous year Mean exacerbations last year: 1.4 for both groups, range 1‐7 | |
| COPD: chronic obstructive pulmonary disease FEV1: forced expiratory volume in one second LABA: long‐acting beta‐agonists ICS: inhaled corticosteroids | |||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mortality (all‐cause) Show forest plot | 2 | 961 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.80 [0.55, 5.91] |
| 1.1 At 3‐month follow‐up | 1 | 660 | Odds Ratio (M‐H, Fixed, 95% CI) | 3.03 [0.12, 74.59] |
| 1.2 At 12‐month follow‐up | 1 | 301 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.64 [0.45, 5.93] |
| 2 Hospital admission (all causes) Show forest plot | 2 | 961 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.84 [0.53, 1.33] |
| 2.1 At 3‐month follow‐up | 1 | 660 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.64 [0.27, 1.49] |
| 2.2 At 12‐month follow‐up | 1 | 301 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.94 [0.55, 1.62] |
| 3 Exacerbation Show forest plot | 3 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3.1 At 3‐month follow‐up | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 At 6‐month follow‐up | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.3 At 12‐month follow‐up | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Quality of life up to 6 months (SGRQ) Show forest plot | 4 | Mean Difference (Fixed, 95% CI) | ‐3.33 [‐4.72, ‐1.94] | |
| 5 Sensitivity analysis ‐ QoL up to 6 months (SGRQ) Show forest plot | 2 | Mean Difference (Random, 95% CI) | ‐2.50 [‐4.16, ‐0.84] | |
| 6 FEV1 pre‐dose Show forest plot | 4 | Mean Difference (Random, 95% CI) | Subtotals only | |
| 6.1 FEV1 3‐6 months mean difference | 4 | Mean Difference (Random, 95% CI) | 0.06 [0.04, 0.08] | |
| 6.2 FEV1 1 year | 1 | Mean Difference (Random, 95% CI) | 0.06 [0.00, 0.12] | |
| 7 Serious adverse events all reported (non‐fatal) Show forest plot | 4 | 1758 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.86 [0.57, 1.30] |
| 8 Pneumonia Show forest plot | 4 | 1758 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 1.62 [0.54, 4.82] |
| 9 Sensitivity analysis ‐ SAE all reported (non‐fatal) Show forest plot | 3 | 1303 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.67 [0.40, 1.13] |
| 10 Adverse event Show forest plot | 4 | 1363 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.16 [0.92, 1.47] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 FEV1 GIV Show forest plot | 1 | Mean Difference (Fixed, 95% CI) | Totals not selected | |
| 2 Adverse event Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3 FEV1 Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
