LAUSR

To the Editor: We thank Dr. Kunisaki [1] for his thoughtful comments on our study regarding the occurrence of cardiovascular disease (CVD) events in continuous positive airway pressure (CPAP) treated and control patients [2]. In amendment to the letter of Dr. Kunisaki, the pairwise matched control patients had used CPAP for amedian of 4.0months and discontinued the treatment prior to the commencement of the 6.2-year study follow-up. We agree that Bhealthy adherer^ effect is an important issue and should be acknowledged during the evaluation of the effectiveness of any treatment. As noted in our paper, due to the retrospective study design, there can be underlying patient characteristics including different personality types causing bias between CPAP-treated and control patients. These potential, underlying biases are, however, extremely difficult to control in retrospective study designs except by matching cases and controls as carefully as possible, which we did. Compared to many randomized controlled trials (RCT) our study cohort was large (N = 2060) and follow-up time exceptionally long (median follow-up time of 8.7 years for CPAP-treated and 6.2 years for control patients). Also the data on usage hours was recorded by the device and thus exact and reliable. Cox regression model was adjusted for all wellknown confounding factors including age, male gender, body mass index, impaired fasting glucose/type 2 diabetes, total plasma cholesterol levels, smoking, chronic obstructive pulmonary disease, hypertension, and previous CVD. Using this approach the result was statistically significant. To be cautious and to put the result in the context, the retrospective study design was mentioned in title, abstract, aims, and discussion. In addition, the limitations of this approach were discussed. It is believed that the beneficial effect of CPAP on CVD outcomes is due to its antihypertensive response [3]. A recent review by Schwarz et al. consisted of three RCTs including only patients with optimal adherence to CPAP (N = 149). The patients were randomized to continue (N = 65) or to withdraw CPAP (N = 84) for 2 weeks. A clinically significant increase in blood pressure was registered during CPAP withdrawal compared to patients who continued the treatment. In addition, greater obstructive sleep apnea severity measured as apnea-hypopnea index was associated with a higher increase in blood pressure in response to CPAP withdrawal [4]. Further, a review by Montesi et al. included 28 RCTs (N = 1948 patients) comparing therapeutic positive airway pressure (PAP) to sham-PAP, pill placebo, or standard care. Compared to controls, PAP treatment was associated with modest but significant reductions in blood pressure [5]. The Bhealthy adherer^ effect hardly explains the shown beneficial effect of CPAP.