To the Editor:
We would like to express our thanks to Alberto Morales Salinas for his letter. In our response, we wish to discuss his comments and our manuscript without entering into a debate on the specifics of "reverse epidemiology" described in different pathologic situations. Clearly, in our series incidence of mortality is higher among low weight patients. This is not surprising and there may well be no single factor to justify it. While therapeutic intervals can have a direct, proportionate association with body mass index (BMI) and, in turn, be influenced by multiple factors such as age, gender or kidney function—as Morales Salinas affirms—we are unaware of any study of mortality in heart failure with beta-blockers or ACE inhibitors—the drugs that have most influenced these patients' survival—that has shown an association between the benefits obtained and dosage adjustment for patient BMI or body surface area. Furthermore, we know that mortality in patients with heart failure has been seen to be related to greater blood concentrations of some drugs.1 Consequently, although we cannot affirm that some patients may not have exhibited this susceptibility, we do not feel we can consider the greater mortality among low-weight patients in our series may have been favored by a relative excess of the treatment received ("the vulnerability of the low-weight patient"). On the other hand, in no case could we consider this phenomenon a confounding factor to the finding that overweight and obese patients had a better prognosis than normal weight patients did. At no point in our manuscript do we affirm the existence of a causal relation between obesity and better prognosis. We simply state that, in a strict 2-year follow-up, overweight and obese patients presented lower mortality. Nor do we have an explanation for this. Although some series report3 finding a U shaped mortality curve—ie, greater mortality in patients with outlying weights (low- or over-weight)—and that, therefore, severe obesity did not provide patients with greater protection than being overweight or slightly obese, we did not observe this phenomenon. In fact, as we comment in our study, no patients with morbid obesity (BMI >40) died in the 2-year follow-up although, as we are dealing with a small number of patients, we cannot generalize from our conclusions.
We agree that it is difficult to accept the obesity paradox in patients with heart failure, even though it has been identified in studies with thousands of patients.3-5 It would be easier to do so if we could find an explanation as to why this paradox arises that did not need to depend—exclusively—on pharmacologic variables.