Antiinflamatorios para la apnea obstructiva del sueño en niños
Resumen
Antecedentes
La apnea obstructiva del sueño (AOS) es un trastorno caracterizado por la obstrucción parcial o completa de las vías respiratorias superiores durante el sueño. Aproximadamente entre un 1% y un 4% de los niños son afectados por la AOS, y la hipertrofia adenoamigdalar es el factor de riesgo subyacente más frecuente. La extirpación quirúrgica de las adenoides o las amígdalas agrandadas es el tratamiento de primera línea recomendado actualmente para la AOS causada por la hipertrofia adenoamigdalar. Debido al riesgo perioperatorio y a la tasa de recurrencia estimada de hasta el 20% después de la cirugía, recientemente ha aumentado el interés por alternativas menos invasivas a la adenoamigdalectomía. Como las adenoides y las amígdalas agrandadas están formadas por tejido linfoide hipertrófico, se propusieron los antiinflamatorios como una posible opción de tratamiento no quirúrgico para los niños con AOS.
Objetivos
Evaluar la eficacia y la seguridad de los antiinflamatorios para el tratamiento de la AOS en niños.
Métodos de búsqueda
Se identificaron los ensayos a partir de las búsquedas en el Registro Especializado del Grupo Cochrane de Vías Respiratorias, CENTRAL y MEDLINE (1950 a 2019). Para la identificación de los ensayos clínicos en curso, se buscó en ClinicalTrials.gov y en el portal de ensayos de la Organización Mundial de la Salud (OMS).
Criterios de selección
Ensayos controlados aleatorizados (ECA) que compararan fármacos antiinflamatorios con placebo en niños de entre uno y 16 años de edad con AOS diagnosticada de forma objetiva (índice de apnea/hipopnea [IAH] ≥ 1 por hora).
Obtención y análisis de los datos
Ambos autores realizaron de forma independiente el cribado, la extracción de datos y la evaluación de la calidad. Los resultados de las comparaciones de los «corticosteroides intranasales» y el «montelukast» frente al placebo se agruparon por separado mediante el uso de modelos de efectos aleatorios. Los resultados primarios de esta revisión fueron el IAH y los eventos adversos graves. Los resultados secundarios incluyeron el índice de dificultad respiratoria, el índice de desaturación, el índice de excitación respiratoria, la saturación de oxígeno arterial nadir, la saturación media de oxígeno arterial, la posibilidad de evitar el tratamiento quirúrgico de la AOS, la puntuación de los síntomas clínicos, el tamaño de las amígdalas y los eventos adversos.
Resultados principales
Se incluyeron cinco ensayos con un total de 240 niños de uno a 18 años con AOS de leve a moderada (IAH 1 a 30 por hora). Todos los ensayos se realizaron en clínicas especializadas en medicina del sueño en centros de atención terciaria. El periodo de seguimiento osciló entre seis semanas y cuatro meses.
Tres ECA (n = 137) compararon corticosteroides intranasales frente al placebo; dos ECA compararon montelukast oral frente al placebo (n = 103). Se excluyó un ensayo del metanálisis ya que los pacientes no fueron analizados como asignados al azar. También hubo preocupación en cuanto al informe selectivo en otro ensayo.
No se sabe con certeza cuál es la diferencia en el IAH (DM ‐3,18; IC del 95%: ‐8,70 a 2,35) entre los niños que reciben corticosteroides intranasales en comparación con placebo (2 estudios, 75 participantes; evidencia de certeza baja). Por el contrario, los niños que recibieron montelukast oral tuvieron un IAH más bajo (MD ‐3,41; IC del 95%: ‐5,36 a ‐1,45) en comparación con los del grupo de placebo (2 estudios, 103 participantes; evidencia de certeza moderada).
No se sabe con certeza si los resultados secundarios son diferentes entre los niños que reciben corticosteroides intranasales en comparación con placebo: índice de desaturación (DM ‐2,12; IC del 95%: ‐4,27 a 0,04; 2 estudios, 75 participantes; evidencia de certeza moderada), índice de excitación respiratoria (DM ‐0,71, IC del 95%: ‐6,25 a 4,83; 2 estudios, 75 participantes; evidencia de certeza baja), y saturación de oxígeno nadir (DM 0,59%, IC del 95%: ‐1,09 a 2,27; 2 estudios, 75 participantes; evidencia de certeza moderada).
Los niños que recibieron montelukast oral tuvieron un índice de excitación respiratoria inferior (DM ‐2,89; IC del 95%: ‐4,68 a ‐1,10; 2 estudios, 103 participantes; evidencia de certeza moderada) y un nadir de la saturación de oxígeno inferior (DM 4,07; IC del 95%: 2,27 a 5,88; 2 estudios, 103 participantes; evidencia de certeza alta) en comparación con los del grupo de placebo. Sin embargo, no existe seguridad en cuanto a la diferencia en el índice de desaturación (DM ‐2,50; IC del 95%: ‐5,53 a 0,54; 2 estudios, 103 participantes; evidencia de certeza baja) entre el grupo de montelukast y el de placebo.
En todos los ensayos se evaluaron y presentaron los eventos adversos, que fueron poco frecuentes, y de naturaleza leve (por ejemplo, hemorragia nasal) y se distribuyeron de manera uniforme entre los grupos de estudio.
Ningún estudio examinó la posibilidad de evitar el tratamiento quirúrgico para la AOS como resultado.
Conclusiones de los autores
No hay evidencia suficiente de la eficacia de los corticosteroides intranasales para el tratamiento de la AOS en los niños; pueden tener efectos beneficiosos a corto plazo sobre el índice de desaturación y la saturación de oxígeno en los niños con AOS leve a moderada, aunque la certeza del beneficio sobre el resultado primario del IAH, así como el índice de excitación respiratoria, fue baja debido a la imprecisión de las estimaciones y la heterogeneidad entre los estudios.
El montelukast tiene efectos beneficiosos a corto plazo en el tratamiento de la AOS en niños sanos, no obesos y no tratados quirúrgicamente (certeza moderada para el resultado primario y certeza moderada y alta, respectivamente, para dos resultados secundarios) en cuanto a la reducción significativa del número de apneas, hipopneas y de excitación respiratoria durante el sueño. Además, el montelukast fue bien tolerado en los niños estudiados. Sin embargo, la pertinencia clínica de los efectos observados del tratamiento sigue sin estar clara, debido a que todavía no se han establecido las diferencias mínimas clínicamente importantes para los resultados basados en la polisomnografía en los niños.
Todavía no se dispone de datos sobre la eficacia y la seguridad a largo plazo del uso de medicamentos antiinflamatorios para el tratamiento de la AOS en la infancia. Además, aún no se han investigado los resultados centrados en el paciente, como la capacidad de concentración, la vigilancia o el rendimiento escolar.
Actualmente no hay ningún ECA sobre el uso de otros tipos de medicamentos antiinflamatorios para el tratamiento de la AOS en niños.
En los ECA futuros se debería investigar la sostenibilidad de los efectos del tratamiento, la posibilidad de evitar el tratamiento quirúrgico para la AOS y la seguridad a largo plazo de los medicamentos antiinflamatorios para el tratamiento de la AOS en los niños, e incluir resultados centrados en el paciente.
PICO
Resumen en términos sencillos
Antiinflamatorios para el tratamiento de la apnea obstructiva del sueño en niños
Antecedentes
La apnea obstructiva del sueño (AOS) es la obstrucción parcial o completa de las vías respiratorias superiores durante el sueño. Afecta a entre el 1% y el 4% de los niños. La razón subyacente más común de la AOS en la infancia es el agrandamiento de las adenoides o las amígdalas. El tratamiento recomendado actualmente es la extirpación quirúrgica de las adenoides y las amígdalas. El tratamiento con medicamentos antiinflamatorios es una alternativa, un tratamiento no quirúrgico en los casos más leves de AOS.
Pregunta de la revisión
El objetivo de esta revisión fue evaluar la evidencia sobre los beneficios y la seguridad de los medicamentos antiinflamatorios para el tratamiento de la AOS en niños de entre uno y 16 años de edad. Los resultados primarios de la revisión fueron el número de pausas en la respiración y los episodios de respiración superficial por hora de sueño (índice de apnea/hipopnea, IAH) y los efectos indeseables graves de la medicación.
Características de los estudios
Se incluyeron cinco estudios en la revisión. Cada uno de estos estudios incluyó de 25 a 62 niños de entre uno y 11 años de edad con AOS de leve a moderada, tratados en clínicas ambulatorias especializadas en el sueño. En los estudios incluidos se utilizaron dos tipos diferentes de medicamentos antiinflamatorios. Setenta y cinco niños fueron asignados al azar para recibir un spray nasal de corticosteroides intranasales o un placebo. Ciento tres niños fueron asignados al azar para recibir comprimidos de montelukast o placebo.
Fuentes de financiación de los estudios
Tres estudios fueron apoyados por los fabricantes de los fármacos.
Resultados clave
No se sabe con certeza la diferencia en el número de pausas respiratorias, episodios de respiración superficial, episodios con falta de oxígeno en la sangre o interrupción del sueño entre los niños que reciben el spray nasal con corticosteroides en comparación con placebo (2 estudios con 75 niños).
Los niños que recibieron comprimidos de montelukast tuvieron menos pausas respiratorias, menos episodios de respiración superficial y menos interrupciones del sueño en comparación con los que fueron tratados con placebo (2 estudios con 103 niños). Sin embargo, no existe seguridad en cuanto a la diferencia en el número de episodios con falta de oxígeno en la sangre entre el grupo de montelukast y el de placebo.
En todos los ensayos se evaluaron los efectos no intencionales, aunque los mismos fueron poco frecuentes y de naturaleza leve (p.ej. hemorragias nasales). No se informó de efectos graves no intencionales.
Certeza de la evidencia
La certeza de la evidencia en cuanto al spray nasal con corticosteroides para el tratamiento de la AOS fue baja para el resultado primario debido a las amplias manifestaciones de los efectos beneficiosos, con alguna inconsistencia del efecto entre los dos estudios incluidos. Se excluyó un estudio del análisis debido a las preocupaciones graves sobre la calidad de los resultados del estudio.
La evidencia en cuanto al uso de comprimidos de montelukast en el tratamiento de la AOS fue de certeza moderada para el resultado primario. Hubo preocupaciones acerca del diseño y la realización de un estudio y alguna inconsistencia del efecto entre los dos estudios incluidos.
La evidencia se actualizó hasta octubre 2019.
Authors' conclusions
Summary of findings
| Corticosteroids compared to placebo for obstructive sleep apnoea in children | ||||||
| Patient or population: obstructive sleep apnoea in children | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
| Risk with placebo | Risk with corticosteroids | |||||
| AHI (number of breathing pauses and episodes of shallow breathing per hour of sleep time) | The mean AHI was 5.18 per hour | MD 3.18 per hour lower | ‐ | 75 | ⊕⊕⊝⊝ | |
| Desaturation index (number of times per hour of sleep time that the blood's O₂ level drops by a certain proportion from baseline) | The mean desaturation index was 3.0 per hour | MD 2.12 per hour lower | ‐ | 75 | ⊕⊕⊕⊝ | |
| Respiratory arousal index (number of times per hour of sleep time that an increased respiratory effort without a decrease in oxygen saturation is recorded) | The mean respiratory arousal index was 7.38 per hour | MD 0.71 per hour lower | ‐ | 75 | ⊕⊕⊝⊝ | |
| Nadir SpO₂ (lowest SpO₂ during sleep) | The mean nadir SpO₂ was 88.24% | MD 0.59% higher | ‐ | 75 | ⊕⊕⊕⊝ | |
| Serious adverse events (N experiencing events) | No serious adverse events reported. The following unintended effects were reported: nasal bleeding (4), nasal discomfort (1), diarrhoea (1) | No serious adverse events reported. The following unintended effects were reported: nasal bleeding (3), nasal discomfort (1), throat discomfort (1), vomiting (1) | ||||
| *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Point estimates vary considerably across studies, very little overlap between confidence intervals, P = 0.10 for heterogeneity, I² = 63%. Downgraded once for inconsistency of results. 2 Upper boundary of 95% confidence interval suggests potential for negative effect of the treatment on the outcome. Downgraded once for imprecision. 3 Point estimates vary considerably across studies, very little overlap between confidence intervals, P = 0.06 for heterogeneity, I² = 72%. Downgraded once for imprecision. 4 Point estimate indicates no clinically significant effect with wide confidence intervals. Downgraded once for imprecision. | ||||||
| Montelukast compared to placebo for obstructive sleep apnoea in children | ||||||
| Patient or population: obstructive sleep apnoea in children | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
| Risk with placebo | Risk with montelukast | |||||
| AHI (number of breathing pauses and episodes of shallow breathing per hour of sleep time) | The mean AHI was 7.55 per hour | MD 3.41 per hour lower | ‐ | 103 | ⊕⊕⊕⊝ | |
| Desaturation index (number of times per hour of sleep time that the blood's O₂ level drops by a certain proportion from baseline) | The mean desaturation index was 5.16 per hour | MD 2.5 per hour lower | ‐ | 103 | ⊕⊕⊝⊝ | |
| Respiratory arousal index (number of times per hour of sleep time that an increased respiratory effort without a decrease in oxygen saturation is recorded) | The mean respiratory arousal index was 9.70 per hour | MD 2.89 per hour lower | ‐ | 103 | ⊕⊕⊕⊝ | |
| Nadir SpO₂ (lowest SpO₂ during sleep) | The mean nadir SpO₂ was 87.38% | MD 4.07% higher | ‐ | 103 | ⊕⊕⊕⊕ | |
| Serious adverse events (N experiencing event) | No serious adverse events reported. The following unintended effects were reported: headache (1), nausea (2) | No serious adverse events reported. The following unintended effects were reported: headache (1), nausea (1) | ||||
| *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Point estimates vary across studies, P = 0.10 for heterogeneity, I² = 54%. Downgraded once for inconsistency of results. 2 Point estimates vary greatly across studies, P = 0.004 for heterogeneity, I² = 88%. Downgraded once for inconsistency of results. 3 Upper boundary of 95% confidence interval suggests potential for negative effect of the treatment on the outcome. Downgraded once for imprecision. 4 Point estimates vary across studies, P = 0.12 for heterogeneity, I² = 59%. Downgraded once for imprecision. | ||||||
Background
Description of the condition
Sleep‐disordered breathing is one of the most common sleep disorders in childhood. Affected children may exhibit habitual snoring, laboured breathing, sleep disruption, and/or gas exchange abnormalities. The most severe form of sleep‐disordered breathing, obstructive sleep apnoea (OSA), is defined as a “disorder of breathing during sleep characterised by prolonged partial upper airway obstruction and/or intermittent complete obstruction (obstructive apnoea) that disrupts normal ventilation during sleep and normal sleep patterns, accompanied by symptoms or signs” (AAP 2012).
Approximately 1% to 6% of children are affected by OSA (Lumeng 2008; AAP 2012). Symptoms include habitual snoring, laboured breathing, witnessed apnoea, disturbed sleep, daytime neurobehavioral problems, and sleepiness (AASM 2006; AAP 2012). The most common underlying risk factor for the development of OSA is adenotonsillar hyperplasia (Marcus 2001; AAP 2012). The relative enlargement of the tonsils during the pre‐school years reduces pharyngeal space and promotes airway collapse during sleep, resulting in OSA (Marcus 2001). Other risk factors include male sex, overweight and obesity, prematurity, African American ethnicity, craniofacial anomalies, and neuromuscular disorders (Rosen 2003; Lumeng 2008; AAP 2012).
A child with suspected OSA should either undergo polysomnography or be referred to a sleep medicine specialist or otolaryngologist for a more extensive evaluation (AAP 2012). A definitive diagnosis of OSA is made by attended overnight polysomnography in a sleep laboratory (AAP 2012). Other methods, such as nocturnal video recordings, pulse oximetry, ambulatory polysomnography, or daytime nap polysomnography may be performed if standard polysomnography is not available (AAP 2012). The most commonly used polysomnographic parameter to identify children with OSA is the apnoea/hypopnoea index (AHI), which corresponds to the number of breathing pauses (apnoeas) or episodes of shallow breathing (hypopnoeas) per hour of sleep (AASM 2006). An AHI of one per hour or more is usually considered an abnormal finding in children (AASM 2006). As the sensitivity of polysomnographic equipment for the detection of apnoeas and hypopnoeas differs, the AHI may vary across sleep laboratories. Despite this shortcoming, the AHI is considered the mainstay for the diagnosis of OSA in children (AASM 2006).
Description of the intervention
If a child with OSA exhibits adenotonsillar hyperplasia and does not have a contraindication to surgery, surgical removal of enlarged adenoids and tonsils (adenotonsillectomy) is recommended as the first‐line treatment for OSA (AAP 2012), despite there being only one randomised controlled trial (RCT) that examined and demonstrated the benefits of adenotonsillectomy compared to watchful waiting (Marcus 2013). However, adenotonsillectomy is painful, requires hospitalisation and may result in anaesthetic complications, acute upper airway obstruction, haemorrhage, postoperative respiratory compromise, or velopharyngeal incompetence (Crysdale 1986; McColley 1992; Lee 1996; Ruboyianes 1996; AAP 2012). The frequency of complications from adenotonsillectomy range between 5% and 34% (AAP 2002). There is also evidence that OSA may persist after surgery. In a large follow‐up study, only 27% of children showed complete resolution of OSA after adenotonsillectomy (Bhattacharjee 2010). By contrast, the RCT by Marcus and colleagues reported normalisation of OSA in 79% of patients in the 'early adenotonsillectomy' group after seven months (Marcus 2013).
Given these findings, there is an increased interest in painless, safe, and effective non‐surgical treatment modalities for children with surgically untreated mild to moderate OSA or as a second‐line treatment of persisting or recurrent OSA after surgery. One promising alternative is the intranasal application of anti‐inflammatory medications such as corticosteroids (Kheirandish‐Gozal 2013).
How the intervention might work
As the enlarged adenoids and tonsils consist of hypertrophied lymphoid tissue, anti‐inflammatory medications may be effective in reducing the size of adenoids and tonsils and have therefore been proposed in the past as a useful non‐invasive treatment option in children with residual OSA or contraindications to surgery (Alexopoulos 2004; Goldbart 2005; Kheirandish 2006). This approach was further supported by the observation that a substantial proportion of children with OSA exhibit symptoms of allergic rhinitis (McColley 1997) or asthma (Brockmann 2014). Therefore, anti‐inflammatory drugs such as corticosteroids, leukotriene receptor antagonists (e.g. montelukast), and cromoglicic acid may be effective treatment options for childhood OSA.
Why it is important to do this review
Adenotonsillectomy as first‐line treatment for OSA has several disadvantages with regard to acceptance, safety, efficacy, sustainability of treatment effects, and costs and may not be readily available in regions with limited healthcare resources. Given the emerging evidence on non‐surgical alternatives to adenotonsillectomy, the role of anti‐inflammatory medications as first‐line treatment for OSA in childhood should be investigated in a systematic review.
Objectives
To assess the efficacy and safety of anti‐inflammatory drugs for the treatment of OSA in children.
Methods
Criteria for considering studies for this review
Types of studies
We considered all types of individual RCTs, including cross‐over trials, for inclusion in the review. We did not consider cluster randomised trials.
Types of participants
We included studies of children between one and 16 years of age with objectively diagnosed OSA (according to polysomnography; AHI ≥ 1/h, where AHI is the sum of obstructive and mixed apnoeas and obstructive hypopnoeas). We accepted studies with up to 20% of 17‐ to 18‐year‐old participants.
We excluded trials that only included participants with the following co‐morbidities: malformation syndromes, neuromuscular disorders, and morbid obesity (corresponding to a body mass index of 40 kg/m² or higher). OSA in children with these conditions is mainly driven by excess fatty tissue and anatomical and neurological factors and is therefore unlikely to be affected by anti‐inflammatory medications. Adenotonsillectomy, weight loss, and treatment with continuous positive airway pressure may be more appropriate treatments for OSA in these patient groups, particularly in children with obesity (Verhulst 2008).
Types of interventions
We included studies randomising participants to receive any anti‐inflammatory drugs used for the treatment of OSA in children. Eligible drugs with anti‐inflammatory properties included steroids, leukotriene receptor antagonists, ketotifen and chromones, administered either systemically or locally for any duration. Eligible comparison was against placebo.
Types of outcome measures
The diagnosis of OSA and indication for treatment are mainly based on clinical symptoms and the AHI as the primary objective indicator of OSA. In addition to the AHI, there are several polysomnography‐based indices such as arousal and desaturation indices that reflect the severity of OSA in children. As treatment is only indicated if complaints or subjective symptoms (or both) are reported, secondary outcomes include symptom scores and abnormalities in the physical examination.
Primary outcomes
-
AHI as measured by polysomnography
-
Serious adverse events
Secondary outcomes
-
Respiratory disturbance index as measured by polysomnography
-
Desaturation index as measured by polysomnography
-
Respiratory arousal index as measured by polysomnography
-
Nadir arterial oxygen saturation (SpO₂) as measured by polysomnography
-
Mean SpO₂ as measured by polysomnography
-
Avoidance of surgical treatment for OSA (in adenotonsillectomy‐naive children)
-
Clinical symptom score (based on parent report of, for example, snoring, witnessed apnoea, daytime sleepiness etc.)
-
Tonsillar size (on an ordinal scale from 0 (not visible) to +4 (tonsils touch))
-
Adverse events
We considered outcomes at all measured time points.
Search methods for identification of studies
Electronic searches
The first published version of this review was based on searches up to March 2010. For the current update, the last search was performed on 10 October 2019.
We identified trials from searches of the following databases.
-
The Cochrane Airways Trials Register (Cochrane Airways 2019), searched through the Cochrane Register of Studies (CRS), all years to 17 October 2019
-
Cochrane Central Register of Controlled Trials (CENTRAL), searched through the Cochrane Register of Studies (CRS), all years to 17 October 2019
-
MEDLINE (Ovid SP) 1946 to 17 October 2019
-
US National Institutes of Health Ongoing Trials Register (www.ClinicalTrials.gov)
-
World Health Organization International Clinical Trials Registry Platform (www.apps.who.int/trialsearch)
The search strategies are shown in Appendix 1. We did not restrict our search by language or type of publication.
Searching other resources
We searched the CENTRAL database and Cochrane Airways Trials Register for conference abstracts and grey literature. We reviewed the reference lists of all primary studies and review articles for additional references. We contacted investigators of ongoing and unpublished completed trials that were identified in the trial register searches.
Data collection and analysis
Selection of studies
Two review authors (SK, MSU) independently screened titles and abstracts of the records retrieved by the electronic and handsearches. We based inclusion/exclusion of articles on the eligibility criteria outlined above. For potentially relevant articles, both review authors obtained and reviewed the full text for possible inclusion in the study. We resolved disagreement by discussion; no third party adjudication was necessary.
Data extraction and management
Two review authors (SK, DH) independently performed data extraction using an electronic form. We checked accuracy of entered data manually. We assigned each study a unique study identifier and re‐checked for eligibility before data extraction.
The following data were extracted from the studies.
-
General: bibliographic details; funding source.
-
Methods: design; blinding; randomisation; dropouts/withdrawals; study quality assessment ('Risk of bias' tool).
-
Participants: country; target population; setting; number enrolled; diagnostic criteria used; inclusion/exclusion criteria; baseline demographic and clinical data (age, sex, AHI, respiratory disturbance index, desaturation index, respiratory arousal index, clinical symptom score, tonsillar size).
-
Interventions: intervention (type/dose/frequency/duration); comparison (type/dose/frequency/duration); compliance.
-
Outcome: timing of outcome assessment(s); reported outcomes (continuous: mean, and standard deviation or whichever measure of variability is reported; dichotomous: number of events) by group. For continuous outcomes, we recorded the number of participants, mean and standard deviation (or whichever measure of variability was reported) for each group and assessment. For dichotomous outcomes, we recorded the number of participants and the count of outcomes for each group and assessment.
Assessment of risk of bias in included studies
Two investigators (SK, MSU) independently assessed the quality of the trials. We evaluated each trial's quality using the 'Risk of bias' tool (Higgins 2011). The 'Risk of bias' tool appraises the study's adequacy in relation to the following six domains.
-
Sequence generation
-
Allocation concealment
-
Blinding of participants, personnel and outcome assessors
-
Incomplete outcome data
-
Selective outcome reporting
-
Other sources of bias
We assigned a judgement of high, low or unclear risk of bias.
Measures of treatment effect
For continuous outcomes (AHI, desaturation index, and respiratory arousal index), we calculated the mean difference and 95% confidence interval (CI). For dichotomous outcomes (avoidance of surgical treatment for OSA), we had planned to calculate the relative risk and 95% CI. We had planned to treat ordinal outcomes (tonsillar size) as continuous.
Where only the median and interquartile range were reported for continuous outcomes, we planned to use the median instead of the mean and to calculate the standard deviation as interquartile range divided by 1.35, provided the sample size was large (Higgins 2011).
Unit of analysis issues
The unit of analysis was the patient.
Dealing with missing data
We had planned to conduct a sensitivity analysis to explore the influence of studies where missing outcome data may have introduced a serious bias.
Assessment of heterogeneity
We assessed heterogeneity using the Q‐test and the I² statistic. We considered values of more than 25% for the I² statistic to represent levels of statistical heterogeneity which would merit further investigation.
Assessment of reporting biases
We had planned to assess publication bias statistically using a linear regression approach and graphically using a funnel plot (Egger 1997).
Data synthesis
We estimated treatment effects, separately pooled for corticosteroids and montelukast, using a random‐effects model.
Subgroup analysis and investigation of heterogeneity
Allergies/asthma/atopy are risk factors for OSA and can be effectively treated with anti‐inflammatory drugs. We had therefore planned to perform a subgroup analysis for children with and without allergies/asthma/atopy in order to assess if an observed treatment effect on OSA is only due to the anti‐allergic effect of the drug on the allergy/asthma/atopy. We had further planned a subgroup analysis for the type and route of administration of steroid used, and for trials with more than one drug in a treatment arm.
Sensitivity analysis
If heterogeneity had been detected, we would have performed sensitivity analyses to assess the effect of the following items on pooled estimates.
-
Low‐quality studies as assessed by the 'Risk of bias' tool
-
Random versus fixed‐effect modelling
-
Publication status
-
Funding source (drug manufacturer versus independent)
Summary of findings and assessment of the certainty of the evidence
We created a 'Summary of findings' table using the outcomes AHI, desaturation index, respiratory arousal index, nadir SpO₂, and serious adverse events as these outcomes are considered to be clinically relevant according to current guidelines (Iber 2007). We used the five GRADE considerations (risk of bias, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of a body of evidence as it relates to the studies that contribute data for the prespecified outcomes. We used the methods and recommendations described in Section 8.5 and Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), using GRADEpro software (GRADEpro GDT). We justified all decisions to downgrade the certainty of the evidence for each outcome using footnotes and made comments to aid the reader's understanding of the review where necessary. The inclusion of a 'Summary of findings' table was a post hoc addition in the current review update.
Results
Description of studies
Results of the search
See Figure 1 for the study flow diagram. The original review identified three eligible studies (Kuhle 2011). The electronic database search run on 17 October 2019 retrieved 224 studies; a handsearch of conference proceedings did not identify any further studies. Screening of the titles and abstracts yielded 19 potentially relevant studies, two of which were eligible for inclusion in this review in addition to the three previously identified studies.
Study flow diagram: review update
Abbreviations: RCT: randomised controlled trial; OSA: obstructive sleep apnoea
We found two ongoing studies (Aihuan 2012; Marcus 2014). Aihuan 2012 compares intranasal budesonide to oral montelukast; the trial registry record was last updated May 2015, but the lead investigator did not respond to our email requests. The Steroids for Pediatric Apnea Research in Kids (SPARK) trial compares a 3‐month course of nasal fluticasone against placebo in children with mild to moderate OSA (Marcus 2014). The trial is still recruiting and scheduled to be completed in June 2020.
We found two unpublished studies (Li 2007; Bernardi 2013). Li 2007 performed a randomised controlled trial of four months of intranasal mometasone furoate versus placebo in 52 children aged six to 18 years in Hong Kong. The registry entry was last updated in April 2017. Bernardi 2013 compared oral montelukast to intranasal budesonide. The study was completed in June 2010 but remains unpublished. We classified both studies as 'awaiting classification' because results were not available and the respective lead investigators have not responded to our email requests.
Included studies
We included five studies in this review (Brouillette 2001; Kheirandish‐Gozal 2008; Goldbart 2012; Chan 2015; Kheirandish‐Gozal 2016). Kheirandish‐Gozal 2008 included some children with an AHI of less than one per hour, while our inclusion criteria specified an AHI of one or more per hour. However, as only four out of the enrolled 62 children (6%) actually had an AHI less than one per hour at baseline, we decided to include the study and to perform a sensitivity analysis if necessary. All studies were single‐centre studies. Three trials were conducted in North America, one in Israel, one in Sweden, and one in China. Three studies were funded in part or fully by the drug manufacturer (Brouillette 2001; Kheirandish‐Gozal 2008; Kheirandish‐Gozal 2016). For two studies, only the active drug and placebo were provided by the drug manufacturer (Goldbart 2012; Chan 2015).
Population
Brouillette 2001 included 25 children at a tertiary care centre between one and 10 years of age with mild to moderate OSA who had been referred from otolaryngologists for assessment before adenotomy/tonsillectomy. Kheirandish‐Gozal 2008 included 62 children between the ages of six and 11 years who were evaluated for mild to moderate OSA at a tertiary care centre. Goldbart 2012 enrolled 46 children between two and 10 years with non‐severe OSA (patients with AHI between 1 to < 10 per hour), who had been referred to a tertiary care centre for evaluation of their snoring. Chan 2015 included 62 children between six and 18 years of age with habitual snoring (≥ 3 nights per week) and mild OSA (obstructive AHI ≥ 1 to 5). Kheirandish‐Gozal 2016 included 57 children aged 2 to 10 years with symptomatic snoring with an of AHI of two or more but less than 30.
Brouillette 2001, Kheirandish‐Gozal 2008, Goldbart 2012, and Kheirandish‐Gozal 2016 excluded severe cases of OSA and children with recent or current steroid treatment. All studies excluded children with neuromuscular disorders. Brouillette 2001, Kheirandish‐Gozal 2008, Chan 2015 and Kheirandish‐Gozal 2016 excluded children with craniofacial anomalies. Goldbart 2012 also excluded children with obesity (BMI z‐score > 1.645), syndromic or defined genetic abnormalities, current or previous use of montelukast, acute upper respiratory tract infection, or use of corticosteroids or antibiotics. Chan 2015 also excluded children with genetic syndromes, congenital or acquired neurological diseases, previous upper airway surgery, or known fixed nasal obstruction such as previous nasal fracture. Kheirandish‐Gozal 2016 also excluded children under current treatment with corticosteroids or oral montelukast.
Study design
Four trials were parallel group designs with treatment for six weeks (Brouillette 2001), 12 weeks (Goldbart 2012), 16 weeks (Kheirandish‐Gozal 2016), and four months (Chan 2015), respectively. Kheirandish‐Gozal 2008 used a cross‐over design with a 2‐week washout period between the two 6‐week cycles.
Interventions
Three studies used nasal steroids: Brouillette 2001 used fluticasone 50 μg per nostril twice daily for the first week, followed by 50 μg per nostril once daily for the remaining five weeks. Kheirandish‐Gozal 2008 used budesonide 32 μg per nostril once daily for six weeks. Chan 2015 used mometasone furoate 100 µg per nostril once daily for four months. Goldbart 2012 and Kheirandish‐Gozal 2016 used montelukast 4 mg or 5 mg (depending on the child's age) orally once daily. The comparison group received placebo in all studies.
Outcomes assessed
Three of the included studies used the AHI as their primary outcome (Brouillette 2001; Kheirandish‐Gozal 2008; Chan 2015). One study did not specify a primary endpoint (Kheirandish‐Gozal 2016). The trial registration record for the latter study had been withdrawn prior to enrolment, but listed the AHI as the primary endpoint (Kheirandish 2017). One study listed multiple primary endpoints (obstructive apnoea index, AHI, desaturation index, respiratory arousal index) (Goldbart 2012).
Other outcomes included the desaturation index (Brouillette 2001; Kheirandish‐Gozal 2008; Goldbart 2012; Chan 2015; Kheirandish‐Gozal 2016), respiratory or total arousal index (Brouillette 2001; Kheirandish‐Gozal 2008; Goldbart 2012; Chan 2015; Kheirandish‐Gozal 2016), nadir SpO₂ (Brouillette 2001; Kheirandish‐Gozal 2008; Goldbart 2012; Chan 2015; Kheirandish‐Gozal 2016), mean SpO₂ (Brouillette 2001), tonsillar size (Brouillette 2001; Kheirandish‐Gozal 2008; Chan 2015; Kheirandish‐Gozal 2016), and clinical symptom scores (Brouillette 2001, Chan 2015).
No study reported the avoidance of surgical treatment for OSA as an outcome, although it was listed as a primary outcome in Kheirandish‐Gozal 2016.
Excluded studies
We excluded 35 studies that were not RCTs, eight studies that did not use polysomnography to diagnose OSA or assess the outcome, and two studies that used adenotonsillectomy as the comparison group.
Risk of bias in included studies
An overview of our risk of bias judgements can be found in Figure 2.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
Adequacy of allocation concealment was unclear for two studies, because the research co‐ordinator was involved in allocation and patient management (Kheirandish‐Gozal 2008; Kheirandish‐Gozal 2016). All other studies were at low risk of selection bias.
Blinding
For two studies it was unclear whether the placebo was indistinguishable from the intervention, which may have compromised blinding of patients and personnel (Kheirandish‐Gozal 2008; Kheirandish‐Gozal 2016).
Incomplete outcome data
Brouillette 2001 and Goldbart 2012 reported complete data for all participants. The dropout rates in the trials by Chan 2015 (12/62) and Kheirandish‐Gozal 2016 (7/64) were similar for the intervention and control groups. Kheirandish‐Gozal 2008 had a significantly higher dropout rate in the group of children who were initially treated with placebo compared to those who initially received the intervention (14/32 vs. 5/30; relative risk 2.63, 95% CI 1.08 to 6.40), and we therefore judged it to be at high risk of bias. In addition, the reasons for withdrawal were not equally distributed between treatment arms.
Selective reporting
One trial did not report on the primary outcome (need for adenotonsillectomy) specified in the trial registration (Kheirandish‐Gozal 2016; Kheirandish 2017). The remaining four studies appeared to be free of selective outcome reporting, but trial registration information was not available for three of these studies (Brouillette 2001; Goldbart 2012; Chan 2015).
Other potential sources of bias
The cross‐over study, Kheirandish‐Gozal 2008, suffered from a number of potential biases. Firstly, study groups were imbalanced at baseline with the placebo‐first group having a mean AHI that was two standard errors lower than in the intervention group. Four out of 32 (13%) and none out of 30 participants (0%) had an AHI less than one per hour in the placebo and treatment groups, respectively. This baseline imbalance was also not accounted for in the analysis. Secondly, the authors presented and compared outcomes from children receiving intranasal budesonide in either arm with those who received placebo in the first arm, thereby omitting data from 25 children that received placebo in the second arm. Thirdly, the authors reported that those children who received budesonide in the first arm and placebo in the second arm had comparable sleep outcomes after the treatment phase compared with the placebo phase. This observation indicates that there might have been a carry‐over effect and that the cross‐over design or washout period was inadequate.
Goldbart 2012 performed an interim analysis after 30 patients had been enrolled. The impact of this analysis on blinding, trial continuation, and outcome selection is unclear.
Effects of interventions
See: Summary of findings for the main comparison Corticosteroids compared to placebo for obstructive sleep apnoea in children; Summary of findings 2 Montelukast compared to placebo for obstructive sleep apnoea in children
Given that the authors of the intranasal budesonide trial did not analyse the data according to the randomisation and were unable to provide us with their results for the first phase of the study, we did not consider this trial in the analysis (Kheirandish‐Gozal 2008). We pooled in separate meta‐analyses results from two trials on corticosteroids (n = 75; Brouillette 2001; Chan 2015), and two trials on montelukast (n = 103; Goldbart 2012; Kheirandish‐Gozal 2016).
Corticosteroids versus placebo
Primary outcomes
The mean change in the AHI was −3.18 per hour (95% CI −8.70 to 2.35; 2 studies, 75 participants; Analysis 1.1) in children receiving intranasal corticosteroids compared to placebo. There was heterogeneity in the estimates between the studies (P = 0.10 for heterogeneity, I² = 63%). Due to imprecision of the estimate and considerable heterogeneity between studies, we rated the overall certainty of the evidence as low. No serious adverse events were reported.
Secondary outcomes
The desaturation index changed by −2.12 per hour (95% CI −4.27 to 0.04; 2 studies, 75 participants; Analysis 1.2) in the intervention group compared to the control group. We rated the overall certainty of the evidence as moderate due to imprecision.
The respiratory arousal index changed by −0.71 (95% CI −6.25 to 4.83; 2 studies, 75 participants; Analysis 1.3) in the intervention group compared to the control group. Due to the small effect and considerable heterogeneity between studies (P = 0.06 for heterogeneity, I² = 72%), we rated the overall certainty of the evidence as low.
The nadir SpO₂ increased by 0.59% (95% CI −1.09 to 2.27; 2 studies, 75 participants; Analysis 1.4) in the intervention group compared to the control group. Due to imprecision, we rated the overall certainty of the evidence as moderate.
The difference in change in mean SpO₂, tonsillar size, and symptom score was uncertain between the fluticasone group and the control group in Brouillette 2001 (n = 25). We are uncertain about any difference in nasal symptoms between the mometasone and placebo groups (Chan 2015, n = 50). Adverse events reported were minor and infrequent and appeared similar between the intervention and placebo groups (Table 1).
| Study | Intervention (N experiencing event) | Control (N) | Comments |
| 0 | nose bleed (1) | ||
| 0 | 0 | ||
| nasal bleeding (3), nasal discomfort (1), throat discomfort (1), vomiting (1) | nasal bleeding (3), nasal discomfort (1), diarrhoea (1) | Total number of adverse events reported (10) does not match sum of individual adverse events (11) | |
| headache (1), nausea (1) | headache (1), nausea (2) |
Since Kheirandish‐Gozal 2008 did not analyse and report their data as randomised, we did not consider this trial in the analysis.
Montelukast versus placebo
Primary outcomes
The mean change in the AHI was −3.41 per hour (95% CI −5.36 to −1.45; 2 studies, 103 participants; Analysis 2.1) in children receiving montelukast compared to placebo. There was considerable heterogeneity between the studies (P = 0.14 for heterogeneity, I² = 54%), and we therefore rated the certainty of the evidence as moderate. No serious adverse events were reported.
Secondary outcomes
The desaturation index changed by −2.50 per hour (95% CI −5.53 to 0.54; 2 studies, 103 participants; Analysis 2.2) in the intervention group compared to the control group. Point estimates also varied greatly across studies (P = 0.004 for heterogeneity, I² = 88%). Due to imprecision of the estimate and considerable heterogeneity between studies, we rated the overall certainty of the evidence as low.
The respiratory arousal index changed by −2.89 (95% CI −4.68 to −1.10; 2 studies, 103 participants; Analysis 2.3) in the intervention group compared to the control group. Due to considerable heterogeneity between studies (P = 0.12 for heterogeneity, I² = 59%), we rated the overall certainty of the evidence as moderate.
The nadir SpO₂ increased by 4.07% (95% CI 2.27% to 5.88%; 2 studies, 103 participants; Analysis 2.4) in the intervention group compared to the control group. We rated the overall certainty of the evidence as high.
The difference in change in mean SpO₂ was uncertain between the intervention group and the control group (Goldbart 2012, n = 46). There was also uncertainty over the difference in tonsillar size between the intervention and control group (Kheirandish‐Gozal 2016, n = 57). One study did report no adverse events (Goldbart 2012); while the other reported minor adverse events in two and three children in the intervention and control group, respectively (Kheirandish‐Gozal 2016) (Table 1).
Discussion
Summary of main results
In this review, we aimed to evaluate the available evidence on anti‐inflammatory medications as treatment for OSA in children. We included five RCTs evaluating four different drugs in the review; and we included four studies evaluating three drugs in the meta‐analysis. Sample sizes of the included studies were small, with the largest one enrolling 62 children. Three trials investigated intranasal corticosteroids (fluticasone, budesonide, and mometasone) and two assessed the oral leukotriene receptor antagonist montelukast. We did not find any studies that examined other anti‐inflammatory medications for the treatment of OSA.
In both trials that examined intranasal corticosteroids, the AHI in the intervention group decreased compared to placebo. However, based on the pooled effect estimate from the meta‐analysis, we are uncertain about the reduction of the AHI. We are also uncertain about the effects of corticosteroids on the desaturation index, the respiratory arousal index, the nadir SpO₂, the mean SpO₂, tonsillar size, or the symptom score. We rated the overall certainty of the evidence as low for the AHI and respiratory arousal index and moderate for the desaturation index and the nadir SpO₂.
Meta‐analysis of the results from the two montelukast trials demonstrated a beneficial effect of montelukast on the AHI, the respiratory arousal index, and the nadir SpO₂. The overall certainty of the evidence was low for the desaturation index, moderate for the AHI and respiratory arousal index, and high for the nadir SpO₂. We are uncertain about the difference in the mean SpO₂ and the tonsillar size. The clinical relevance of the observed intervention effects on AHI, respiratory arousal index, and nadir SpO₂ remains unclear, however, because minimal clinically important differences for these outcomes have not been established yet in paediatric sleep medicine. The effect sizes (Cohen's d) for the AHI in the two studies were 0.86 and 1.12 (Goldbart 2012 and Kheirandish‐Gozal 2016 respectively), which would be considered a large effect based on the cut‐off (> 0.8) proposed by Cohen 1988. Hence, the overall treatment effect of montelukast may be deemed clinically relevant.
Adverse events were rare and of minor nature for both medications, and their frequency did not differ between the intervention and control groups. Hence, the short‐term use of intranasal steroids and montelukast appears to be well tolerated.
Overall completeness and applicability of evidence
The five included studies were limited to less severe forms of OSA because adenotonsillectomy is considered the treatment of choice for children with moderate to severe OSA. All studies excluded children with neuromuscular disorders and most studies excluded children with genetic and craniofacial anomalies. Only one study included adolescents and all studies excluded newborns and infants below two years of age. All studies measured the AHI and SpO₂ parameters, which are accurate markers of the presence and severity of OSA in children. Hence, the evidence from the studies primarily applies to otherwise healthy preschool and school‐aged children with mild to moderate forms of OSA and without other concomitant severe chronic diseases. The evidence cannot be applied to newborns, infants, and adolescents or to children with more severe or complex forms of OSA. The latter limitation largely reflects current practice, however, as infants and children with OSA related to an underlying genetic disorder should undergo a thorough diagnostic evaluation and receive individualised treatment combinations such as surgery, respiratory support, and/or intraoral appliances. Efficacy and safety data for anti‐inflammatory medications for the treatment of OSA in childhood are currently only available for up to four months; no study has examined long‐term efficacy and safety. Similarly, studies have almost exclusively used clinical outcomes instead of more patient‐centred outcomes such as concentration ability, executive functioning, vigilance, behavior, school performance, or health‐related quality of life.
Certainty of the evidence
The certainty of the evidence for the effect of intranasal corticosteroids on the AHI was low. We could not include one of the three studies in the meta‐analysis due to high risk of bias, leaving only two studies with 75 patients. There was heterogeneity in the estimates between the studies (P = 0.10 for heterogeneity, I² = 63%). As a result, even though both studies individually showed statistically significant reductions in the AHI in the intervention group compared to the control group, the confidence interval of the pooled estimate of the treatment effect included the null when we analysed study results using a random‐effects model. The certainty of the evidence for the secondary outcomes ranged from low (respiratory arousal index) to moderate (desaturation index and nadir SpO₂). We downgraded the rating for the outcome 'respiratory arousal index' because of the small treatment effect and considerable heterogeneity between studies.
The certainty of the evidence for a beneficial effect of montelukast on the AHI was moderate. The effect estimate was fairly precise, but the two included studies were of unclear and high risk of bias, respectively, and there was considerable heterogeneity between the studies (P = 0.14 for heterogeneity, I² = 54%). The certainty of the evidence for the secondary outcomes ranged from low (desaturation index) and moderate (respiratory arousal index) to high (nadir SpO₂). We downgraded the certainty of the evidence for desaturation index and respiratory arousal index for the considerable heterogeneity between the two studies. We further reduced the rating for the desaturation index for imprecision.
The specific methodological issues of the included studies are as follows. The trial comparing budesonide to placebo had a number of methodological issues that precluded its results from being used in the meta‐analysis (Kheirandish‐Gozal 2008). There was a higher number of withdrawals in the group of children who started on placebo compared with those who started on budesonide. In addition, there appeared to be a carry‐over effect and this effect was not appropriately dealt with in the analysis. Finally, the authors did not include the data of those children who received placebo in the second phase of the study, thereby effectively breaking the randomisation. We contacted the investigators to obtain the data for the first phase of the study, but they were unable to provide us with the data. Goldbart 2012 performed an unplanned interim analysis, and it is unclear what influence this analysis had on the further conduct of the trial with regard to blinding and outcome selection. We deemed the other trial that compared montelukast with placebo to be at high risk of bias (Kheirandish‐Gozal 2016). The primary outcome stated in the trial registration (need for adenotonsillectomy) was not examined in the published paper; instead, a sample size calculation based on the outcome AHI is presented (Kheirandish 2017). It is unclear whether the AHI was in fact the primary outcome of the study.
The small number of participants in the trials limited our ability to detect serious or rare adverse events. All trials included were of short‐term nature with follow‐up time between six and 16 weeks post treatment. Hence the sustainability of the treatment effects and the occurrence of late or long‐term adverse events remain unclear.
Out of the five included studies, three trials had not been registered a priori (Brouillette 2001; Goldbart 2012; Chan 2015); and for one trial the registration was withdrawn before enrolment (Kheirandish‐Gozal 2016). Given this high proportion of unregistered trials, we cannot exclude reporting bias due to unregistered trials with negative findings.
Potential biases in the review process
We conducted the review in accordance with the Methodological Expectations of Cochrane Intervention Reviews guidelines (MECIR 2018). Except for some minor deviations (see Differences between protocol and review), we followed our protocol. We used a broad literature search developed by a Cochrane information specialist. All steps in the review process were performed independently by two reviewers. The review underwent editorial review with the Cochrane Airways Group as well as peer review. Taken together, we are confident that the conclusions from this review represent an objective, fair, and accurate assessment of the current state of the evidence.
Agreements and disagreements with other studies or reviews
There are no other systematic reviews that have specifically examined the association between anti‐inflammatory medications and OSA in children. A network meta‐analysis in 2017 compared the effects of fluticasone, budesonide, montelukast, mometasone, and caffeine on the AHI in children with OSA (Zhang 2017). The RCTs on anti‐inflammatory medications that were included in that review were also identified in the present review (Brouillette 2001; Kheirandish‐Gozal 2008; Goldbart 2012; Chan 2015). The authors concluded that fluticasone and budesonide have "good effects" in the treatment of OSA. This conclusion should, however, be interpreted with great caution as the authors included the study by Kheirandish‐Gozal 2008 in their analysis, which we excluded due to serious risk of bias.
A Cochrane Review examined the effect of intranasal corticosteroids for nasal airway obstruction in children with moderate to severe adenoidal hyperplasia (Zhang 2008). More recently, a systematic review examined the effect of mometasone on adenoid hyperplasia among children with OSA (Chohan 2015). Both reviews found evidence for a beneficial effect of corticosteroids and mometasone, respectively. Since adenoidal hyperplasia can be considered a risk factor for OSA, the findings lend further support for the use of intranasal corticosteroids in children with OSA.
Lastly, a Cochrane Review of drug therapies for OSA in adults included only one trial (out of 30 in total) that used an anti‐inflammatory drug (Mason 2013), reflecting the different pathophysiology and treatment of OSA in children and adults (Alsubie 2017). The investigators found that fluticasone in patients with allergic rhinitis reduced the severity of OSA compared to placebo and improved subjective daytime alertness (Kiely 2004).
Study flow diagram: review update
Abbreviations: RCT: randomised controlled trial; OSA: obstructive sleep apnoea
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Corticosteroids versus placebo, Outcome 1 Apnoea/hypopnoea index.
Comparison 1 Corticosteroids versus placebo, Outcome 2 Desaturation index.
Comparison 1 Corticosteroids versus placebo, Outcome 3 Respiratory arousal index.
Comparison 1 Corticosteroids versus placebo, Outcome 4 Nadir oxygen saturation.
Comparison 2 Montelukast versus placebo, Outcome 1 Apnoea/hypopnoea index.
Comparison 2 Montelukast versus placebo, Outcome 2 Desaturation index.
Comparison 2 Montelukast versus placebo, Outcome 3 Respiratory arousal index.
Comparison 2 Montelukast versus placebo, Outcome 4 Nadir oxygen saturation.
| Corticosteroids compared to placebo for obstructive sleep apnoea in children | ||||||
| Patient or population: obstructive sleep apnoea in children | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
| Risk with placebo | Risk with corticosteroids | |||||
| AHI (number of breathing pauses and episodes of shallow breathing per hour of sleep time) | The mean AHI was 5.18 per hour | MD 3.18 per hour lower | ‐ | 75 | ⊕⊕⊝⊝ | |
| Desaturation index (number of times per hour of sleep time that the blood's O₂ level drops by a certain proportion from baseline) | The mean desaturation index was 3.0 per hour | MD 2.12 per hour lower | ‐ | 75 | ⊕⊕⊕⊝ | |
| Respiratory arousal index (number of times per hour of sleep time that an increased respiratory effort without a decrease in oxygen saturation is recorded) | The mean respiratory arousal index was 7.38 per hour | MD 0.71 per hour lower | ‐ | 75 | ⊕⊕⊝⊝ | |
| Nadir SpO₂ (lowest SpO₂ during sleep) | The mean nadir SpO₂ was 88.24% | MD 0.59% higher | ‐ | 75 | ⊕⊕⊕⊝ | |
| Serious adverse events (N experiencing events) | No serious adverse events reported. The following unintended effects were reported: nasal bleeding (4), nasal discomfort (1), diarrhoea (1) | No serious adverse events reported. The following unintended effects were reported: nasal bleeding (3), nasal discomfort (1), throat discomfort (1), vomiting (1) | ||||
| *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Point estimates vary considerably across studies, very little overlap between confidence intervals, P = 0.10 for heterogeneity, I² = 63%. Downgraded once for inconsistency of results. 2 Upper boundary of 95% confidence interval suggests potential for negative effect of the treatment on the outcome. Downgraded once for imprecision. 3 Point estimates vary considerably across studies, very little overlap between confidence intervals, P = 0.06 for heterogeneity, I² = 72%. Downgraded once for imprecision. 4 Point estimate indicates no clinically significant effect with wide confidence intervals. Downgraded once for imprecision. | ||||||
| Montelukast compared to placebo for obstructive sleep apnoea in children | ||||||
| Patient or population: obstructive sleep apnoea in children | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
| Risk with placebo | Risk with montelukast | |||||
| AHI (number of breathing pauses and episodes of shallow breathing per hour of sleep time) | The mean AHI was 7.55 per hour | MD 3.41 per hour lower | ‐ | 103 | ⊕⊕⊕⊝ | |
| Desaturation index (number of times per hour of sleep time that the blood's O₂ level drops by a certain proportion from baseline) | The mean desaturation index was 5.16 per hour | MD 2.5 per hour lower | ‐ | 103 | ⊕⊕⊝⊝ | |
| Respiratory arousal index (number of times per hour of sleep time that an increased respiratory effort without a decrease in oxygen saturation is recorded) | The mean respiratory arousal index was 9.70 per hour | MD 2.89 per hour lower | ‐ | 103 | ⊕⊕⊕⊝ | |
| Nadir SpO₂ (lowest SpO₂ during sleep) | The mean nadir SpO₂ was 87.38% | MD 4.07% higher | ‐ | 103 | ⊕⊕⊕⊕ | |
| Serious adverse events (N experiencing event) | No serious adverse events reported. The following unintended effects were reported: headache (1), nausea (2) | No serious adverse events reported. The following unintended effects were reported: headache (1), nausea (1) | ||||
| *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Point estimates vary across studies, P = 0.10 for heterogeneity, I² = 54%. Downgraded once for inconsistency of results. 2 Point estimates vary greatly across studies, P = 0.004 for heterogeneity, I² = 88%. Downgraded once for inconsistency of results. 3 Upper boundary of 95% confidence interval suggests potential for negative effect of the treatment on the outcome. Downgraded once for imprecision. 4 Point estimates vary across studies, P = 0.12 for heterogeneity, I² = 59%. Downgraded once for imprecision. | ||||||
| Study | Intervention (N experiencing event) | Control (N) | Comments |
| 0 | nose bleed (1) | ||
| 0 | 0 | ||
| nasal bleeding (3), nasal discomfort (1), throat discomfort (1), vomiting (1) | nasal bleeding (3), nasal discomfort (1), diarrhoea (1) | Total number of adverse events reported (10) does not match sum of individual adverse events (11) | |
| headache (1), nausea (1) | headache (1), nausea (2) | ||
| Since Kheirandish‐Gozal 2008 did not analyse and report their data as randomised, we did not consider this trial in the analysis. | |||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Apnoea/hypopnoea index Show forest plot | 2 | 75 | Mean Difference (IV, Random, 95% CI) | ‐3.18 [‐8.70, 2.35] |
| 2 Desaturation index Show forest plot | 2 | 75 | Mean Difference (IV, Random, 95% CI) | ‐2.12 [‐4.27, 0.04] |
| 3 Respiratory arousal index Show forest plot | 2 | 75 | Mean Difference (IV, Random, 95% CI) | ‐0.71 [‐6.25, 4.83] |
| 4 Nadir oxygen saturation Show forest plot | 2 | 75 | Mean Difference (IV, Random, 95% CI) | 0.59 [‐1.09, 2.27] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Apnoea/hypopnoea index Show forest plot | 2 | 103 | Mean Difference (IV, Random, 95% CI) | ‐3.41 [‐5.36, ‐1.45] |
| 2 Desaturation index Show forest plot | 2 | 103 | Mean Difference (IV, Random, 95% CI) | ‐2.50 [‐5.53, 0.54] |
| 3 Respiratory arousal index Show forest plot | 2 | 103 | Mean Difference (IV, Random, 95% CI) | ‐2.89 [‐4.68, ‐1.10] |
| 4 Nadir oxygen saturation Show forest plot | 2 | 103 | Mean Difference (IV, Random, 95% CI) | 4.07 [2.27, 5.88] |
