Mortality in Sleep Apnea

Abstract

To the Editor: I read with great interest the two recent articles and related editorial on sleep apnea and mortality.1He J Kryger MH Zorick FJ Conway W Roth T. Mortality and apnea index in obstructive sleep apnea: experience in 385 male patients.Chest. 1988; 94: 9-14Abstract Full Text Full Text PDF PubMed Scopus (1357) Google Scholar, 2Gonzales-Rothi RJ Foresman GE Block AJ. Do patients with sleep apnea die in their sleep?.Chest. 1988; 94: 531-538Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar, 3Gonzales-Rothi RJ Block AJ. Mortality and sleep apnea.Chest. 1988; 94: 678-679Abstract Full Text Full Text PDF Scopus (3) Google Scholar The increasing healthcare expenditure on diagnosis and treatment of sleep apnea, the morbidity and even mortality associated with treatment, necessitate careful study of the natural history of this disorder. The two reported studies on the mortality of sleep apnea, from Detroit1He J Kryger MH Zorick FJ Conway W Roth T. Mortality and apnea index in obstructive sleep apnea: experience in 385 male patients.Chest. 1988; 94: 9-14Abstract Full Text Full Text PDF PubMed Scopus (1357) Google Scholar and Gainesville,2Gonzales-Rothi RJ Foresman GE Block AJ. Do patients with sleep apnea die in their sleep?.Chest. 1988; 94: 531-538Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar produced conflicting results; however, neither study does justice to the topic under investigation. Methodologies and conclusions demand further explanation. A critical methodologic requirement for any survival study is that the population under investigation be clearly defined, and that a consecutively-seen group of patients be followed for an appropriate period of time, which most commonly is accepted to be five or ten years. Both of these requirements appear to have been poorly met by each study. The editorial, written by the authors of the Gainesville study, notes that only 55 percent of eligible patients in the Detroit study were included in the final analysis.1He J Kryger MH Zorick FJ Conway W Roth T. Mortality and apnea index in obstructive sleep apnea: experience in 385 male patients.Chest. 1988; 94: 9-14Abstract Full Text Full Text PDF PubMed Scopus (1357) Google Scholar Although the Gainesville group state: “Complete follow-up was obtained on all patients in this study …”, they report a mean follow-up of 90 percent of all eligible patients, due to medical records that were not suitable for review.2Gonzales-Rothi RJ Foresman GE Block AJ. Do patients with sleep apnea die in their sleep?.Chest. 1988; 94: 531-538Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar As there were only 13 deaths, if the 14 patients that were not included had all died, the conclusions of the group may well have been very different. The Detroit group did not explain why there were only 22 deaths out of 385 patients of mean age 51 years, who were reportedly followed for up to eight years. Unfortunately, the mean or median follow-up time was not reported by the authors, but it is most likely that the median follow-up time was less than five years; a much shorter time than we are led to believe from the data as presented. The relatively short follow-up time might, in part, explain the relatively small number of deaths reported. Similarly, the mention of five and eight years of follow-up by the Gainesville group is misleading. They stated: “Kaplan-Meier distributions of all patients over the eight-year study period …”, and repeatedly emphasized the five year nature of their study, both in the original article and editorial, but they also do not give the mean or median follow-up time of their patients. Their data suggest that the median follow-up time is likely to have been less than 2.5 years. The Gainesville group state in their abstract: “There were no statistically significant differences in mortality between the two groups.” This statement is also misleading, as it implies that similar groups were being compared, whereas the mean age at death of the controls was 66 years, compared with the mean age at death of 50 years for sleep apnea syndrome patients. In addition, the control group (according to the authors) had significantly more life-threatening disease. Perhaps a more meaningful statement by the authors might have been “There were no statistically significant differences in mortality between a younger, healthier group of sleep apnea patients compared with an older, less healthy control group.” The median age at death for the general population in 1984 was 78.2 years; much higher than the mean age at death of 50 years for the sleep apnea patients.4National Center for Health Statistics Vital statistics of the United States, 1984, vol II, sec., DHHS pub. no. (PHS) 87-1104. DC: US Government Printing Office, Washington1987Google Scholar The mean or median age at death of the Detroit group's sleep apnea patients was also not reported but would be of interest. One interesting point that was not discussed by either group was the fact that the treated sleep apnea patients had a higher death rate than the untreated group in both studies. This could reflect either ineffective treatment or, more likely, might reflect the greater severity of disease. The Detroit group indicated that the high mortality seen in their patients with UPP may be due to a loss to follow-up. Yet the Detroit group specify that 78 of their 98 UPP patients had subsequent follow-up polysomnography, and follow-up of this group of patients has previously been reported.5Conway W Fujita S Zorick F Sicklesteel J Roehrs T Wittig R Roth T. Uvulopalatopharyngoplasty. One-year follow-up.Chest. 1985; 88: 222-226Crossref Scopus (100) Google Scholar A more likely explanation why there were so many deaths in the UPP group would be that this group was followed more closely and, therefore, more deaths were reported. Patients treated by other means were more likely to have been lost to follow-up; therefore, their deaths go unreported. In view of their large group of essentially “untreated” patients who had a lower mortality, it is also hard to accept the explanation that the ineffective treatment of the UPP patients was responsible for the high mortality quoted. It is clear that there are many methodologic and reporting errors in both these papers. The Gainesville group stressed the importance of “multicenter cooperative prospective longitudinal studies”. However, because of the morbidity associated with this disease, it is unlikely that such studies will be able to be performed in untreated patients; therefore, we will continue to be dependent upon retrospective studies in obtaining data on such patients. These two papers illustrate the need for careful collection and analysis of data, and one would hope that the “waters will not be muddied” by poorly performed mortality studies. Both of the reported mortality studies are more similar in their findings than the conclusions of the authors would suggest. As both studies indicate, patients with sleep apnea are liable to early death, whether or not it occurs during sleep. The clinician must not be discouraged from “agressive therapeutic intervention”, as the Gainesville group implied. Additional information is needed to more clearly demonstrate the exact mortality associated with this disease. Despite the deficiencies of these studies, they are both to be commended for raising awareness of the issues regarding the mortality of sleep apnea.