Accuracy of the sleep-related breathing disorder scale to diagnose obstructive sleep apnea in children: a meta-analysis

Highlights

• An indication of tonsillectomy for obstructive sleep apnea syndrome (OSAS) in children still relieson clinical signs.

• This meta-analysis found the sleep-related breathing disorder scale (SRBDs) to be accurate for the diagnosis of OSAS.

• The SRBDs is potentially useful to identify children appropriate to undergo complementary evaluation before tonsillectomy.

Abstract

Objectives

The main objective of this meta-analysis was to assess the accuracy of the Sleep-Related Breathing Disorder (SRBD) Scale in the diagnosis of obstructive sleep apnea syndrome (OSAS) in children.

Patients/methods

A literature search of studies comparing SRBD to polysomnography for the diagnosis of OSAS in children was performed. Risks of biases were quantified using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool. Analyses determined the summary receiver operator characteristic area under the curve (SROC), the pooled sensitivity (Se), the specificity (Sp), and the positive and negative likelihood ratios (LR+ and LR–). Results were graded and are expressed as means [95% confidence interval]. Post-test probabilities were computed for various populations.

Results

Eleven studies were included; and two were considered to have high risk of bias. The SROC was 0.73 [CI: 0.63; 0.82]. The combined Se, Sp, LR+ and LR– were: 0.72 [CI: 0.68; 0.77], 0.59 [CI: 0.56; 0.63], 1.74 [CI: 1.32; 2.30], 0.53 [CI: 0.39; 0.71], respectively. Sub-group analyses displayed similar results in comparison to overall results. GRADE evidence for the overall analysis was low to moderate. Finally, pre-test to post-test probabilities were estimated to be: 3.5%–1%, 50%–30% and 75%–30%, for the general population, the obese patients and the patients assigned for surgical treatment of OSAS, respectively.

Conclusions

The current meta-analysis indicates that the SRBD scale has acceptable accuracy in detecting patients with OSAS. It may be useful when evaluating patients with suspected OSAS before surgery.

Study registration

PROSPERO database (CRD42018088216).

Introduction

According to recent estimations, the incidence of obstructive sleep apnea syndrome (OSAS) in children ranges from 1% to 3% and increases to 60% in high-risk patients such as overweight and obese children [1], [2], [3]. Consequently, the diagnosis and management of OSAS is a daily challenge for physicians caring for children in hospitals and the community [4]. Sleep Breathing disorders in children include many conditions ranging from primary snoring to OSAS [3], [5], [6], [7]. In addition, obese patients, a high-risk population for developing OSAS, might suffer from the obstructive hypoventilation syndrome, that requires specific management [8], [9].

Many physicians worldwide rely on clinical assessment – physical examination and simplified questionnaires – when diagnosing OSAS in children. This is the case even though clinical assessment has been demonstrated to be significantly inferior to the gold standard, polysomnography, whether in diagnosing OSAS or in assessing severity [1]. Reasons advocated for this practice are reduced or absent availability of polysomnography and/or inconsistency of recommendations regarding the diagnosis of OSAS [1], [10], [11]. Surveys among US surgeons found polysomnography to be performed in 10% of children before tonsillectomy [12], [13]. Inaccurate diagnosis leads to mismatched therapy and potentially serious consequences in OSAS patients undergoing surgery [14].

At the beginning of this century, Chervin and collaborators [15] developed the Sleep-Related Breathing Disorder (SRBD) Scale, deriving their scale from the pediatric sleep questionnaire (a 22-item SRBD questionnaire with a positive classification defined as the presence of more than 33% of positive response). Using an Apnea-Hypopnea Index (AHI) > 5 to define the presence of OSAS during gold standard polysomnography comparison, they found this questionnaire to exhibit a sensitivity and specificity of 85% and 87%, respectively, in the construct cohort; and 81% and 87%, respectively, in the validation cohort. A meta-analysis of clinical studies in this era has found moderate sensitivity and specificity of mixed clinical signs in the diagnosis of OSAS in children in comparison to polysomnography [16]. As an example, Van Someren et al., found the doctor's examination, to exhibit a sensitivity and specificity of 59% and 73%, respectively when compared to polysomnography [17] and A similar result (sensitivity 55%) was also found in another review of the reliability of history and physical examination in the diagnosis of OSAS [18].

The main objective of this meta-analysis was to determine the accuracy of the SRBD Scale compared to polysomnography, in the diagnosis of the OSAS in children.