We thank Bangalore and colleagues and Patel for their interest in our article. Bangalore and colleagues essentially indicate that the diagnosis of coronary death based on death certificates has suboptimal validity and point out that the fraction of coronary deaths (n = 133) of all deaths (n = 792) (16.8%) is too low. In the Greek population at large, the corresponding fraction is 13.4%,1 possibly because coronary mortality remains relatively low in Greece. Still, we agree with Bangalore and colleagues that some misclassification is unavoidable. Nondifferential misclassification, however, in general tends to attenuate associations. Concern for bias-generated association would have been justified in a case-control study, but in this prospective investigation, selection or information biases are unlikely. With respect to noncoronary mortality, the mortality ratio for those taking midday naps on any frequency or duration was 1.06 (95% confidence interval, 0.87-1.28), whereas for overall mortality it was 0.97 (95% confidence interval, 0.82-1.15).
Naska A, Oikonomou E, Trichopoulou A, Psaltopoulou T, Trichopoulos D. Is Siesta More Beneficial Than Nocturnal Sleep?—Reply. Arch Intern Med. 2007;167(19):2144. doi:10.1001/archinte.167.19.2144-a