Reply to "In-Hospital Mortality Risk of Older Patients With COVID-19 Infection".

To the Editor:

We are delighted by the interest in our research by Dr Kawada and we welcome the opportunity to better clarify the results of our recent study on COVID-19 patients aged ≥80 years old. In the study, we explored the overlapping effect between clinical covariates and baseline frailty status, in older patients hospitalized for COVID-19. Our results underlined the independent effect of severe frailty on the risk for in-hospital death, as well as male sex, the severity of disease at presentation, and overall comorbidities ≥3.

As noted by Dr Kawada, the group of patients with no or mild frailty [Clinical Frailty Scale (CFS) score = 1-3] was reduced (61 patients overall), with only 2 (3.3%) deaths in this group. Conversely, in the mild frailty group (CFS score = 4-6) there were 417 patients (128 deaths, 30.7%), and in the severely frail group (CFS score = 7-9) there were 251 patients (157 deaths, 62.5%). This has led to some uncertainty in the estimation of the exact hazard ratio for death using the first category as a reference and dividing the frailty scale into 3 groups. Nevertheless, it should be noted that the sample size is adequate for the overall estimation. Assuming the well-known rule of 10 cases needed for each liberty degree at multivariate analysis, we would have needed at least 100 events for the multivariate analysis (consistently, in our study we had 287 events). Moreover, evaluating the CFS variable as a linear scale in a Cox regression model, we obtained an estimated hazard ratio of 1.4 for each point of CFS score, which gives an about 10-fold risk of death for a patient with CFS score = 8¸ which is pretty similar to the 12-fold estimated risk reported in the article for the group CFS score = 7-9. Indeed, apart from the overall estimated HR, the crude figures of mortality clearly indicate that frail patients have an increased risk of death, which is the major finding conveyed by our research.

Because life expectancy is lower in males, it could be speculated that in the older cohorts, such as the one evaluated in our study, the risk of death for males could be increased. However, given the short length of our follow-up (the end point was in-hospital mortality), we do not think that this could justify the observed excess in male mortality. Similarly, several authors evidenced the increased risk for mortality among male patients with COVID-19, independent of the age group.

Finally, as noted by Dr Kawada, we did not consider the BMI among the evaluated variables. The overlapping relationship among low BMI, malnutrition, and sarcopenia is a key determinant of the frailty status. Although a low BMI may be a component of sarcopenia, the 2 conditions are not the same, neither is malnutrition. Moreover, particularly in older adults, limiting the analysis to BMI could not reflect the nutritional and sarcopenic condition of the patients.

We look forward to improving our analysis, including these suggestions, to better clarify the reasons underlying the excess mortality observed in older COVID-19 patients.