Miriam E. Tucker
November 04, 2020
Two observational studies have yielded opposite findings regarding the effect of routine statin use by people with diabetes who are hospitalized with COVID-19.
One study, which was conducted by Omar Saeed, MD, and colleagues and was published in the Journal of the American Heart Association, aligned with prior studies of COVID-19 patients that found that statin use was associated with reduced in-hospital mortality.
Those findings, by Bertrand Cariou, MD, PhD, of Nantes University Hospital, Nantes, France, were published in Diabetes and Metabolism.
These seemingly contradictory findings occurred despite the use by both teams of a statistical technique called inverse probability of treatment weighting (IPTW), which uses propensity scores to balance baseline variables.
Don't Stop Statins, but They Are Not a Magic Bullet
Asked to comment, Daniel J. Drucker, of Mt. Sinai Hospital, Toronto, Canada, told Medscape Medical News: "The different results, not uncommon in the COVID-19 literature, highlight the challenges of making meaningful causal inferences from retrospective observational data."
They emphasize, Drucker said, "why [randomized controlled trials] will remain the gold standard, as so many observational studies diverge in their conclusions... As people with type 2 diabetes are at increased risk of COVID-19–related mortality, this is precisely the population in whom statin use should be prospectively studied."
Cariou told Medscape: "The fact that our data did not support the previous studies highlights that observational studies per se are not sufficient to claim some causality."
He said that from a clinical perspective, "I would recommend not to change the treatment of diabetic patients who get COVID-19 and thus not to stop statins. But I would mitigate the statement that statin is a magic bullet to reduce COVID-19–related mortality."
US Study: After Adjustments, a 12% Mortality Reduction With Statins
The study by Saeed, of Montefiore Medical Center, New York City, and coworkers involved a total of 4252 patients admitted to their center between March 1 and May 2, 2020, with a confirmed COVID-19 diagnosis. Of those, 53% (2266) had diabetes and 32% (1355) had been treated with a statin (76% with atorvastatin).
Those taking statins were older (69 vs 63 years; P < .01) and had higher scores on the Charlson Comorbidity Index (5 vs 3; P < .01).
Despite those differences, the patients treated with a statin were less likely to die in the hospital (23% vs 27%; P < .01). Those with diabetes who had been taking a statin had significantly lower in-hospital mortality (24% vs 39%; P < 0.01), although the difference was not significant for those without diabetes (20% vs 21%; P = .82).
Propensity score matching and IPTW were adjusted for a long list of characteristics, comorbidities, and laboratory values. Among the patients with diabetes, there was a 12% lower risk for death during hospitalization for statin users in comparison with nonusers (hazard ratio for both, 0.88; P < .001).
Possible mechanisms for statin benefit include known anti-inflammatory properties, as well as reduction of reactive oxygen species and platelet reactivity. Statins can also limit viral endotheliitis, which has been known to occur during SARS-CoV-2 infection, Saeed and colleagues say.
From France, Worse Outcomes With Statins Despite Adjustments
The new French analysis involved 2449 patients from the CORONADO study, all of whom had type 2 diabetes and had been hospitalized with confirmed COVID-19. Of those, 48.7% (1192) were taking statins prior to hospitalization, and 51.3% (1257) were not.
Those taking statins were older (71.7 vs 70.2 years; P = .0014) and had more comorbidities, including hypertension (86.6% vs 74.1%; P < .0001), macrovascular (53.8% vs 26.6%; P < .0001) and microvascular (49.5% vs 41.1%; P = .0005) diabetes complications, and heart failure (14.2% vs 8.1% P = .0014).
The composite primary outcome -- combined tracheal intubation for mechanical ventilation and/or death within 7 days of admission -- was similar among patients who had received statins and those who had not (29.8% vs 27.0%; P = .1338).
However, mortality was significantly higher among statin users at 7 days (12.8% vs 9.8%; P = .02) and 28 days (23.9% vs 18.2%; P < .001).
After applying IPTW, there was also a significant increased risk for the primary outcome with routine statin use within 7 days (odds ratio, 1.38) and for death within 7 days (1.74) and 28 days (1.46).
However, routine statin use was not significantly associated with increased risk for tracheal intubation for mechanical ventilation. Sensitivity analyses that adjusted for glycemic control and renal function produced similar results.
As for possible mechanisms, statins increase cellular expression of angiotensin-converting enzyme 2, the primary receptor for SARS-CoV-2, the authors point out.
What Explains the Difference?
Cariou told Medscape Medical News that because CORONADO is specifically a study of patients with diabetes, whereas the other isn't, "We have some precise and unique data on diabetes characteristics, including duration of diabetes, complications, A1C, and antidiabetic treatments, that help to capture some confounding biases in the propensity score analysis.
"The variables used in the propensity score in the paper from Saeed were not the same. This certainly can explain the differences between the two studies.
"We have to be aware of the limitation of such an approach... Even though many covariates were captured in CORONADO, it remains nonetheless possible that some residual confounding factors were still persistent in the [propensity score] analysis."
He added, "In an ideal world, only [randomized controlled trials] can validate the hypothesis of an impact -- beneficial or deleterious -- on COVID-19 prognosis. I'm not aware of such trials, and it would be difficult to conduct such trials in a shorter term, since you've an incredible number of proposals of interventional trials in the field of COVID-19."
Moreover, Drucker pointed out, "Even [randomized controlled trials] might not resolve the question completely: How far in advance and for how long does one need to start a statin, at what dose, aiming for what target LDL [low-density lipoprotein level] to demonstrate target engagement, to confer possible benefit in the context of a new COVID-19 infection? Does in-hospital statin use confer any additional benefit? Hopefully [randomized controlled trials] will start to provide some answers."
The CORONADO study receives funding from the Fondation Francophone de Recherche sur le Diabete and was supported by Novo Nordisk, MSD, Abbott, AstraZeneca and Eli Lilly; the Federation Française des Diabetiques CORONADO initiative emergency grant; Societe Francophone du Diabete CORONADO initiative emergency grant; Air Liquide Healthcare International; CORONADO initiative emergency grant, Allergan; CORONADO initiative emergency grant, AstraZeneca; CORONADO initiative emergency grant, LifeScan; CORONADO initiative emergency grant, NHC Pharmaceuticals; CORONADO initiative emergency grant, Novo Nordisk; CORONADO initiative emergency grant, Sanofi; and CORONADO emergency grant, PHRC National COVID-19 Hospitalization and Care Organization Division, part of the Hospital Clinical Research Program. Cariou reports grants, non-financial support and/or personal fees from Abbott, Air Liquide Healthcare, Allergan, Amgen, Akcea Therapeutics, AstraZeneca, Eli Lilly, GENFIT, Gilead Sciences, LifeScan, Merck Sharp & Dohme, NHC Pharmaceuticals, Novo Nordisk, Pierre Fabre, Regeneron and Sanofi. Saeed is supported by grants from the National Institute for Health / National Heart, Lung, and Blood Institute and the National Center for Advancing Translational Science Clinical and Translational Science Award at Einstein-Montefiore. Drucker has served as an advisor or consultant or speaker within the past 12 months to Forkhead Biotherapeutics, Intarcia Therapeutics, Merck Research Laboratories, Novo Nordisk Inc, and Pfizer Inc.