Premium
In reference to Admission after sleep surgery is unnecessary in patients without cardiovascular disease
Author(s) -
Jones Philip M,
Martin Janet
Publication year - 2015
Publication title -
the laryngoscope
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.181
H-Index - 148
eISSN - 1531-4995
pISSN - 0023-852X
DOI - 10.1002/lary.25036
Subject(s) - biostatistics , perioperative , epidemiology , medicine , perioperative medicine , citation , anesthesiology , family medicine , library science , classics , surgery , history , psychiatry , computer science
In the article “Admission after sleep surgery is unnecessary in patients without cardiovascular disease,” Rotenberg et al. have made an important contribution regarding the risk of same-day discharge following sleep surgery for obstructive sleep apnea (OSA). The conclusions were reached after finding a 0% incidence of readmissions or adverse events among same-day discharge patients. We urge caution, however, in generalizing the conclusions of their study to prematurely prompt changes in postoperative OSA monitoring policies until further prospective studies of adequate size and quality provide definitive data to support or refute this preliminary conclusion. A major objective of the study was to determine the complication rate postdischarge. We contend that the authors should provide a measure of the statistical imprecision for their 0% point estimate of adverse events by using a confidence interval (CI), as recommended by current guidelines. If zero events are observed in a dataset, how precisely do we know this 0% incidence? For dichotomous outcomes such as incidence of oxygen desaturation or death, the binomial distribution can be used to calculate CIs. However, a simple, easy-to-use approximation can be substituted (provided the sample size is 30). When zero events occur in a sample, the upper 95% confidence limit of the true population risk of the event (as a proportion) is 3/n, where n is the sample size. Although the authors appear to have powered their study using this concept of uncertainty, no upper confidence limit of the 0% incidence of adverse events was reported. Using the above rule, the upper 95% confidence limit is 3/39, or 7.7%. Therefore, given the sample size, the plausible range of adverse events, although observed to be 0% in this particular patient set, could plausibly be as high as 7.7% in future studies of similar patients. A risk of 7.7% would likely be unacceptable to most patients or caregivers for most serious adverse events. Therefore, this study is underpowered to conclude that “Most patients with OSA undergoing surgery can be safely discharged home without any subsequent respiratory complications.” Larger studies are required to determine with adequate confidence whether it is safe to allow same-day postoperative discharge for patients with OSA following sleep surgery. A study of 500 patients with zero adverse events would have an upper 95% confidence limit of 0.6%, which would likely be acceptable to patients and caregivers. Larger studies would also more definitively conclude whether transient nocturnal oxygen desaturations in these patients are benign or linked to more serious adverse events. We encourage further efforts to expand the evidence base on this topic through larger studies and metaanalyses.