Polysomnography Should Not Be Required Both Before and After Adenotonsillectomy for Childhood Sleep Disordered Breathing

In April 2002, the American Academy of Pediatrics (AAP) published a Clinical Practice Guideline for the diagnosis and management of uncomplicated childhood obstructive sleep apnea (OSA) syndrome. OSA is one component of a spectrum of sleep disorders known as sleep disordered breathing (SDB). SDB ranges from primary snoring (isolated snoring without obstructive apnea, arousals, or gas exchange abnormalities) to upper airway resistance syndrome (UARS) to OSA. In the past, primary snoring (PS) was considered benign and therefore not requiring surgical therapy. The AAP guideline emphasizes that pediatricians should screen all children for snoring, since snoring is a presenting symptom of more severe SDB that may require an adenotonsillectomy as the first line of treatment. Complex high-risk patients should be referred to a specialist.¹ Since clinical evaluation is known to be poor at distinguishing OSA from primary snoring,² the diagnostic test of choice is full-night polysomnography (PSG). One can extrapolate from this consensus based guideline that to diagnose OSA one requires PSG. However, PSG is expensive, time consuming, and often unavailable in many institutions that offer pediatric surgical care.

A recent survey of pediatric otolaryngologists revealed that most diagnose SDB using clinical criteria alone. Seventy-five percent of respondents requested PSG in less than 10% of children who present with a sleep disturbance. Doubt about the diagnosis of SDB was the primary reason (60%) to request PSG. Interestingly, the majority (64%) of respondents would proceed with an adenotonsillectomy when the clinical picture strongly suggested SDB irrespective of whether or not the diagnosis was confirmed by PSG.³ For a multitude of reasons, PSG is not indicated for every child with SDB prior to an adenotonsillectomy. Firstly, recent research suggests that habitual snoring may not be benign.^4,5 An Australian study of 26 children between 5 and 10 years of age underwent PSG. They were divided into two groups: (1) no snoring, and (2) snoring but with an apnea-hypopnea index (AHI) of <1 event an hour. Their definition of PS was rigid and would exclude children with OSA at any institution. Neurocognitive testing demonstrated significant deficits for verbal IQ, global IQ, selective attention, sustained attention, and memory index.⁶ Another study divided children into 3 groups: (1) PSG confirmed OSA and adenotonsillectomy; (2) No OSA on PSG but adenotonsillectomy; (3) No OSA on PSG and no surgery. They utilized a clinical assessment tool preoperatively and postoperatively. Group 2 had significant symptom improvement compared to Group 3, indicating that clinical benefits are not limited to children with OSA.⁵ Furthermore, research to date has not shown a correlation between severity of SDB and severity of behavioral/quality of life problems.^7,8 It appears that children with mild SDB or even PS still benefit from surgical therapy. The majority of these children would not be considered for surgical therapy by PSG criteria alone. As more studies support the behavioral morbidity of mild SDB, it will become less important to distinguish PS from OSA.

Another confounding issue is the lack of standardization in sleep acquisitions and clinical reports between centers. For example, two sensors exist to detect airflow abnormalties: flow thermistor (a qualitative sensor that detects flow by temperature fluctuations) and a nasal pressure sensor (a quantitative sensor which is more sensitive in detecting hypopneas). PSG with a nasal pressure sensor will have a higher obstructive index. Subsequently a child who is tested without a nasal pressure sensor may be diagnosed with PS but may have OSA. There is also no consensus on how to define a hypopnea. Some laboratories require a 50% reduction in airflow with either an associated arousal or desaturation of 3% to 4%, while others require a 30% reduction in airflow with either an arousal or desaturation. Furthermore, many adult laboratories that perform pediatric PSG do not utilize end-tidal CO₂ monitoring which is essential to either confirm or exclude the diagnosis of partial obstructive hypoventilation. Besides the technical limitations, there is no consensus on what constitutes an abnormal study. An abnormal study can be defined as an AHI greater than 1^9–12 or greater than 5.^13–15

Although the definition of PS states that PS is isolated snoring without obstructive apnea, arousals or gas exchange, some investigators allow a few respiratory events: OAI <1 and AHI <5 with normal gas exchange,⁴ AHI >1 but <5 with normal gas exchange and no EDS.^14,16 One institution's criteria for PS may qualify as mild SDB in a different institution. Furthermore, some centers include all respiratory events and report a respiratory disturbance index (RDI); others use an AHI; and a few use an obstructive apnea-hypopnea index (OAHI). This lack of uniformity leads to difficulties in comparing results between centers. The lack of consensus within the sleep community undermines the value of PSG as a useful tool to make clinical decisions.

The second edition of the International Classification of Sleep Disorders states that for children, more than one apnea or hypopnea an hour with duration of ≥2 respiratory cycles is abnormal. However, the committee qualified its recommendation stating that the criteria may be modified once more comprehensive data become available.¹⁷ Of interest, most PSG transcriptions report an AHI which makes no distinction between central events when no effort to breathe is made and from obstructive respirations. From the 1995 American Thoracic Society paper, a central apnea index of up to 3 events an hour is within the normal range.¹⁸ Potentially one could have a predominance of central respiratory events rather than obstructive respiratory events, and unnecessary surgery could be performed. To even further complicate the discussion, gas exchange appears to affect the threshold for clinical relevance. The TuCASA group demonstrated that those children with a RDI >5 were more likely to have parental reported snoring, excessive daytime sleepiness, and learning problems. A sub-analysis of the data, which evaluated gas exchange problems, revealed the clinical relevance threshold dropped from 5 events to 1 event an hour if the respiratory events were associated with a 3% desaturation.¹⁹ These findings suggest that gas exchange abnormalities contribute to the morbidity of SDB. A follow-up prospective study using formal cognitive testing supports TuCASA group's initial findings.²⁰ Clinically, a child with an AHI of 3 events an hour with each event having an associated oxygen desaturation may have more severe SDB than a child with a higher index but normal gas exchange. Thus, not only is there lack of uniformity in reporting results; we may not be paying enough attention to those factors in the PSG that may be more closely related to morbidity.

Another argument to support the role of routine preoperative PSG is that it can diagnose and quantify OSA. The severity of OSA can then be used to predict postoperative morbidity. However, studies to date have not always supported this association. For example, a Wisconsin study stratified 38 patients with preoperative PSG into two groups: OSA and UARS; the study did not find any difference in the length of stay (LOS) between the two groups. One would expect that patients with the more severe form of SDB would have a longer average LOS, but the UARS group remained in the hospital for 0.8 more days.²¹ Another study documented that children with severe OSA were more likely to have respiratory complications than those with mild OSA (31% compared to 6%). However, severe OSA was defined as ≥5 respiratory events an hour.²² A study which evaluated the safety of outpatient tonsillectomy demonstrated that a brief period of postoperative observation (<3 hr) was sufficient to identify those children who required overnight observation. Although, preoperative PSG was not performed, 86% (n=134) of the patients underwent surgery for obstructive breathing patterns. Twenty-nine patients were less than 3 years of age. Only 11 of the 134 planned outpatients required inpatient admission; of these, 8 admissions were for pulmonary issues. Only 4 patients who were initially discharged were readmitted, and none of these patients had respiratory problems.²³ Furthermore, since REM sleep is diminished on the first postoperative night, the actual preoperative AHI may not be a good predictor of morbidity as obstructive breathing is more common in REM. Postoperative monitoring rather than PSG is essential to identify children who experience complications following sleep apnea surgery.

A comprehensive history and physical examination is the initial step in establishing the probability that SDB exists. For a healthy child who presents to my clinic for a SDB evaluation, I adopt the baseball approach to determine the role of surgery without preoperative testing. If I can establish that the child has three strikes—(1) Nighttime symptoms: habitual snoring, gasping, pauses or struggling to breathe; (2) Daytime symptoms: un-refreshed sleep, attention deficit, hyperactivity, emotional lability, temperamental, poor weight gain; and (3) enlarged tonsils—I will proceed with surgery without PSG. If I cannot establish three strikes, but the child has other indications for an adenotonsillectomy—recurrent tonsillitis, markedly enlarged (4+) tonsils with dysphagia—I will recommend surgery.

In the following circumstances, preoperative PSG has a role: (1) Surgery is unlikely to be curative (children with morbid obesity, craniofacial anomalies, or neuromuscular disorders); (2) The physical exam does not match the history (i.e., small tonsils and adenoid); (3) The parents need objective information; (4) Increased risk for surgery (heart anomalies, bleeding disorder).

For an otherwise healthy child with who has night time symptoms, daytime symptoms, and enlarged tonsils, do not waste health care dollars and delay treatment for a serious health condition by requesting a PSG. As for postoperative PSG, it is unnecessary in otherwise healthy children unless they have severe OSA ( AHI >20) or the parents report persistent symptoms such as snoring or witnessed apneas.²⁴