Reply to Huang et al

Clinical Infectious Diseases, vol. 72, pp. e1157–e1158

Abstract

TO THE EDITOR—We thank Huang and colleagues for their letter [1] on our article [2] and we appreciate the opportunity to respond. Huang et al provide pooled analyses of 5 recently published randomized controlled trials (RCTs) that compare outcomes between patients treated with tocilizumab (TCZ) to patients not treated with TCZ [3–7]. They concluded that tocilizumab does not provide a beneficial effect on mortality and that there was no observed difference in serious adverse events [1]. Both findings provide valuable insights into the effect of TCZ in patients with coronavirus disease 2019 (COVID-19). The absence of an effect on mortality in these RCTs seems to contradict the decreased mortality observed in our meta-analysis on observational studies [2]. If this is true, this finding suggests that the difference in effect on mortality between the pooled estimates of observational studies and RCTs is due to bias and confounding in the observational studies, provided that the RCTs are perfect and free of such bias and confounding. However, reality may be more complex. Although randomization may achieve balanced groups at baseline, it is not guaranteed that no differences could occur after randomization. Differences in co-medication use are of interest in the postrandomization period as they could potentially confound the observed effect of TCZ on mortality [2]. In particular, the administration of systemic corticosteroids is associated with a lower all-cause mortality [8]. In 3 of 5 RCTs, the patients in the control group received more corticosteroids than the patients in the TCZ group [4, 6, 7]; in the study by Hermine et al [4], 55% of patients in the control group received corticosteroids compared with 30% of patients in the TCZ group, and 28% of the patients in the control group received dexamethasone compared with 14% of patients in the TCZ group. In the study by Salama et al [7], 88% of patients in the control group received corticosteroids compared with 80% of patients in the TCZ group, and 67% of patients in the control group received dexamethasone compared with 55% of patients in the TCZ group. In the study by Rosas et al [6], 55% of patients in the control group received corticosteroids compared with 36% of patients in the TCZ group. In the study by Salvarani et al [5], the patients in the control group had a lower mean C-reactive protein, lower mean interleukin 6, lower mean ferritin, and lower mean D-dimer and received more antivirals compared with patients in the TCZ group. In the same study, TCZ rescue medication was given in 14 of 66 patients in the control group before the 30-day follow-up. The use of TCZ in the control group could therefore confound the effect of TCZ on 30-day mortality rates. In addition, the overall mortality rate in the control groups was approximately 11% for the RCTs [1] compared with 31% for the observational studies [2], suggesting that the patients in the RCTs are very different from those in the observational studies. We therefore agree with Huang et al that more evidence from ongoing RCTs is needed. Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

Authors 3

1. Jishnu Malgie Aachen Department of Orthopedics, Leiden University Medical Center

   Leiden University Medical Center

   Affiliation as printed

   Department of Orthopedics, Leiden University Medical Center, Leiden, The Netherlands

2. Jan W. Schoones Aachen Directorate of Research Policy

   Leiden University Medical Center

   Affiliation as printed

   Directorate of Research Policy (formerly Walaeus Library), Leiden University Medical Center, Leiden, The Netherlands

3. Bart G. C. W. Pijls corresponding Aachen Department of Orthopedics, Leiden University Medical Center

   Leiden University Medical Center

   Affiliation as printed

   Department of Orthopedics, Leiden University Medical Center, Leiden, The Netherlands