To the Editor I read with interest the recent article by Cheraghlou and colleagues 1 regarding primary tumor excision techniques in Merkel cell carcinoma (MCC). The article's central observation, an association between the use of Mohs surgery and improved overall survival (OS) compared with wide excision, appears to be accurate. However, the assertion that Mohs surgery provides a "more effective treatment for MCC primary tumors than conventional" 1 wide local excision does not appear to be supported by the presented data.The authors appropriately recognize the limitations of a study using the National Cancer Database (NCDB) to examine this difficult question. A more detailed examination of these and other issues demonstrates how unlikely it is that variations in surgical treatment of the primary MCC were responsible for the observed difference in OS.First, due to the age and comorbidities that characterize the MCC population, OS is a poor surrogate end point for treatment efficacy. Up to half of the deaths in patients with MCC are due to non-MCC causes. 2 Second, MCC-related deaths, when they do occur, almost always occur within about 4 years of diagnosis. 3 The difference in OS observed in the NCDB data only emerges after 4 years, making it very likely that it was due to causes of death other than MCC. In addition, the Mohs and wide excision cohorts were not similar regarding comorbidities or tumor size, both of which were more favorable in the Mohs subgroup. Expanding any conclusions from such a small (4% of the overall study group) and unrepresentative subgroup to apply to MCC in general is quite hazardous. Finally, the magnitude of the effect on OS is probably not biologically feasible. Suggesting that any alteration of local therapy would result in a 41% reduction in mortality seems overly optimis-tic. Even radiation therapy, which has been associated with marked reductions in local recurrence for MCC, has not been found to significantly contribute to distant metastasis-free survival. 4,5 Overall, it appears that the cause and effect of the demonstrated association is the opposite of what was suggested; patients with favorable life expectancies were very selectively referred for Mohs surgery, rather than the surgical approach having any effect on outcomes. Such a bias in referral patterns may have implications for other nonrandomized comparisons of local therapies in skin cancer.