Author + information
- R. Thomas Collins II, MD⁎ (, )
- Jeffrey M. Gossett, MS and
- Christopher J. Swearingen, PhD
- ↵⁎Arkansas Children's Hospital, 1 Children's Way, Mail Slot 512-3, Little Rock, Arkansas 72202
We read with interest the recent paper by Ball et al. (1). Hypertrophic cardiomyopathy (HCM) is an important topic; we commend the authors for analyzing the largest cohort with resting obstructive HCM. However, we question some key points.
First, significant differences exist between the conservative group (CG) and the invasive group (IG). The CG was older (CG age 57 ± 16 years vs. IG age 47 ± 15 years; p < 0.0001), had more women (50% vs. 40%, respectively; p = 0.0006), and had a higher burden of major comorbidities than the IG (18.3% vs. 8.4%, respectively; p = 0.0002). Although age and sex were covariates in the multivariable analysis (MVA), the interaction of age and sex was not evaluated. Therefore, it cannot be determined whether the difference in follow-up year survival results from differences in age, sex, or a combination thereof. Moreover, for unspecified reasons, the MVA dichotomized age at 50 years. Based on the age distributions, it appears that age older than 50 years and treatment classification will be correlated, as approximately 70% of the CG was older than 50 years. Conversely, disease burden was not included in the MVA. In a group with a higher burden of diseases that shorten survival, shorter survival is expected. With 81 deaths (or equivalents of death) observed, the MVA could have, by convention, included up to 8 explanatory variables. The disease burden should have been accounted for in the MVA because disease burden could greatly affect the survival results.
Because of the group differences and the time-varying covariate of follow-up years in the MVA, the use of Kaplan-Meier analysis is inappropriate, even as a secondary analysis. We recommend using plot-predicted values based on multivariable models. Additionally, when using Kaplan-Meier to analyze death for any reason, as opposed to death due to HCM, it is more appropriate to analyze overall survival as a function of age and not follow-up years. The Kaplan-Meier analysis suggests that survival length was approximately 5 years less in the CG. The authors' conclusion and the overarching premise of the paper are that “patients treated invasively have an overall survival advantage compared with conservatively treated patients.” If age is considered, as opposed to follow-up years, the authors' conclusion may not be valid; Table 1 and Figure 1 in the paper by Ball et al. (1) indicate that, on average, the CG actually lived to an older age (approximately 5 years older than the IG).
Although the authors' findings do demonstrate that younger patients with resting obstructive HCM in this cohort study lived longer during follow-up, their other conclusions are in question. The results of the MVA, details on the methods used to select the variables in the model, and lack of appropriate disease severity accounting lead to weaknesses too significant to overlook. As a result, we are not convinced that, based on the results at hand, there is any scientific evidence that invasive therapy in the setting of resting obstructive HCM is associated with any benefit.
- American College of Cardiology Foundation