Statistics from Altmetric.com
We thank Ben Mulhearn1 for his correspondence on our article.2 The correspondence raises the important question of whether moderate-dose to high-dose glucocorticoid (GC) therapy is an independent risk factor for COVID-19-related death or whether this association is rather due to confounding by indication.
First of all, we would like to re-emphasise that causal interpretation of results based on data from a cross-sectional observational study like the one we report is not possible. The only study design that rises to the level of demonstrating cause and effect is the randomised controlled trial. Confounding by indication is a particularly important and very common type of bias in observational studies, and it occurs when an exposure appears to be associated with an outcome (distorted association), but the outcome (eg, ‘death’) is, in fact, caused by the indication for which the exposure was used, that is, the indication (eg, ‘disease activity’) prompted the exposure (eg, ‘GC therapy’).3
For these reasons, in our manuscript we explicitly caution against interpreting our estimates causally. Specifically, we use non-directional and non-causal language, reporting an ‘association’ of GC prednisolone-equivalent dosage >10 mg/day with COVID-19-related death and avoiding language that could imply a causal relationship.
Due to space constraints, in our discussion we focused on other aspects of the results. However, this correspondence provides an opportunity for us to extend the discussion about the interplay between disease activity and GC therapy, bearing in mind that this aspect is particularly important for the treating clinician. Indeed, the benefit to risk ratio of GC in the context of rheumatic diseases has been a matter of significant debate for many years.4
In our data set, the vast majority of patients who did not receive GCs or received them in prednisolone-equivalent dosages equal to or less than 10 mg/day were either in remission or had low disease activity (83%), while the majority of patients receiving more than 10 mg/day of prednisolone-equivalent dosages had moderate/high/severe disease activity (53%), suggesting that higher dosages of GC therapy and moderate/high/severe disease activity often occur together, as indicated by Ben Mulhearn.1 Of particular relevance to clinicians is the question of whether patients are at additional risk if they are chronically treated with GC, particularly with moderate to high GC dosages, or if they receive a new GC prescription or their GC dosage is increased in the context of a disease flare.
In one of the sensitivity analyses of our manuscript where patients were stratified by disease activity, GC therapy was not associated with the outcome, independently of the dose (see online supplemental table 15 of the original article2). This might be interpreted as absence of an independent GC effect; however, given that the OR estimates do not decrease in comparison with the main analysis, it is likely that this finding reflects reduced statistical power due to the smaller sample size in the subgroup analysis.
Thus, to disentangle the effects more clearly, we have run our analysis again, now explicitly modelling the interaction between GC therapy and disease activity in an additive sense. To avoid overfitting (due to the increase in the number of exposure variables), this analysis was only performed for the entire group of patients (main model of the original article2), allowing us to consider prednisolone-equivalent dosages in three categories (1–5 mg/day, 6–10 mg/day, >10 mg/day). The results for the corresponding ORs of the different possible combinations of disease activity and GC dosages are presented in figure 1. The ORs of the other effects in the (main) model underwent only minor numerical changes and are not shown.
The results show that the effects of GC therapy in the absence of a moderate/high/severe disease activity are not significant anymore when modelling explicitly all of the subgroups representing the interaction between GC therapy and disease activity. This supports the notion that confounding by indication is likely to have contributed to the significantly increased OR observed with prednisolone-equivalent dosages >10 mg/day in our primary analysis and that GC therapy might not be an independent factor associated with COVID-19-related death.2 However, a residual contribution from GC cannot be definitely ruled out.
All subgroups with moderate/high/severe disease activity show a significant association with COVID-19-related death (figure 1), independently of whether GCs are being taken or not. The fact that the strength of this association increases when higher dosages of GC are being taken could either suggest that the GC dose might be viewed as a proxy for the level of disease activity or reflect a potential residual contribution from GC, or both.
In conclusion, the above additional analyses suggest that GC use in this observational data set could represent confounding by indication and that higher disease activity itself is the main exposure independently associated with COVID-19-related death.
This further highlights the importance of appropriate patient management and controlling rheumatic disease activity in the context of a pandemic. If GCs are required to manage disease activity (potentially as a bridging therapy), adherence to general recommendations (and common sense) is advised, and the requirement for continuing GC treatment should be continuously re-evaluated. The dose chosen has to take into account both the desired therapeutic response and the risk of undertreatment and development of GC-related adverse events (ie, give the GC for the shortest period of time and keep the dose as low as the level sufficient in order to achieve the required therapeutic effect in each individual patient).5 Finally, it should also be noted that patients on long-term GC therapy are at risk of GC-induced adrenal suppression and may therefore require GC supplementation in the context of major trauma, surgery or significant intercurrent infection, including COVID-19.6
Handling editor Josef S Smolen
Twitter @carmona_loreto, @philipcrobinson, @pedrommcmachado
Contributors MS, AS and PMM drafted the first version of the manuscript. All authors revised the manuscript and approved the final version.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Disclaimer The views expressed here are those of the authors and participating members of the COVID-19 Global Rheumatology Alliance and do not necessarily represent the views of the American College of Rheumatology, the European Alliance of Associations for Rheumatology, the UK National Health Service, the National Institute for Health Research or the UK Department of Health, or any other organisation.
Competing interests MS has nothing to disclose. AS reports personal fees from lectures for AbbVie, MSD, Roche, BMS and Pfizer, all outside the submitted work. KLH reports she has received non-personal speaker’s fees from AbbVie and grant income from BMS, UCB and Pfizer, all unrelated to this manuscript, and is supported by the NIHR Manchester Biomedical Research Centre. LC has not received fees or personal grants from any laboratory, but her institute works by contract for laboratories among other institutions, such as AbbVie Spain, Eisai, Gebro Pharma, Merck Sharp & Dohme España, Novartis Farmaceutica, Pfizer, Roche Farma, Sanofi Aventis, Astellas Pharma, Actelion Pharmaceuticals España, Grünenthal and UCB Pharma. MG reports grants from the National Institutes of Health, NIAMS, outside the submitted work. SL-T has nothing to disclose. EFM reports that LPCDR received support for specific activities: grants from AbbVie, Novartis, Janssen-Cilag, Lilly Portugal, Sanofi, Grünenthal, MSD, Celgene, Medac, Pharma Kern and GAfPA; grants and non-financial support from Pfizer; and non-financial support from Grünenthal, outside the submitted work. LG reports personal consultant fees from AbbVie, Amgen, BMS, Biogen, Celgene, Gilead, Janssen, Lilly, Novartis, Pfizer, Samsung Bioepis, Sanofi-Aventis and UCB; and grants from Amgen, Lilly, Janssen, Pfizer, Sandoz, Sanofi and Galapagos, all unrelated to this manuscript. PCR reports personal fees from AbbVie, Eli Lilly, Gilead, Janssen, Novartis, Pfizer, Roche and UCB; non-financial support from BMS; and research funding from Janssen, Novartis, Pfizer and UCB, all outside the submitted work. JY reports consulting fees from AstraZeneca and Eli Lilly; and grants from Pfizer, outside the submitted work. PMM has received consulting/speaker’s fees from AbbVie, BMS, Celgene, Eli Lilly, Janssen, MSD, Novartis, Orphazyme, Pfizer, Roche and UCB, all unrelated to this manuscript, and is supported by the National Institute for Health Research (NIHR), University College London Hospitals (UCLH), Biomedical Research Centre (BRC).
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Patient consent for publication Not required.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.