Study population and setting
This large prospective cohort study included all patients aged 20-99 years (n=8,275,949; median age 47; 50% female) who were registered to the QResearch database including 1,205 medical practices in England on January 1, 2020, with follow-up through April 27, 2020. The authors considered patients’ history of using ACE inhibitors and angiotensin receptor blockers (ARBs) to test whether using these medications was associated with 1) COVID-19 diagnosis confirmed via RT-PCR or 2) severe COVID-19 disease resulting in ICU admission. Outcomes were assessed via linking two national databases (one for COVID-19 diagnosis, and the other for COVID-19 ICU admission) to the QResearch database. Exposure variables and covariates were extracted from medical records. Patients were defined as current users of ACE inhibitors or ARBs if they had both: 1) at least three prescriptions in their medical history; and 2) at least one prescription in the 90 days before cohort enrollment. Multiple imputation was used to replace missing data. Multivariable Cox proportional hazards models were used to estimate hazard ratios (HRs) for ACE inhibitors and ARBs with respect to each outcome.
Summary of Main Findings
A total of 7.8% of patients were current users of ACE inhibitors, and 3.7% were current users of ARBs. During the study period, 19,486 patients (0.24%) were diagnosed with COVID-19, and 1,284 (0.02%) were admitted to the ICU with COVID-19. ACE inhibitor use was associated with lower hazards of COVID-19 diagnosis (HR: 0.71, 95% CI: 0.67-0.74) but was not significantly associated with COVID-19 ICU admission (HR: 0.89, 0.75-1.06). Similarly, ARB use was associated with lower hazards of COVID-19 diagnosis (HR: 0.63, 95% CI: 0.59-0.67) but was not significantly associated with COVID-19 ICU admission (HR: 1.02, 0.83-1.25). Models were adjusted for a wide range of covariates, including age, ethnicity, deprivation quintile, BMI, smoking, comorbidities, and concurrent medications. When restricted to patients with hypertension or congestive heart failure, inferences were unchanged. There was significant heterogeneity by ethnic group in both 1) the risks of COVID-19 diagnosis, and 2) the HRs for ACE inhibitors and ARBs for COVID-19 diagnosis. Black, Caribbean, and Asian patients had higher risks of COVID-19 diagnosis and higher HRs for ACE inhibitors and ARBs than white patients in adjusted models.
This study considered an extremely large number of patients who were free from COVID-19 at baseline. This allowed calculation of risk for COVID-19 diagnosis, as opposed to prior studies that only included hospitalized patients or those with a positive SARS-CoV-2 test result. The analysis relied on the linking of large validated databases with wide coverage; thus, the study population is likely to be broadly representative of the English population, though mild or asymptomatic cases may be underrepresented.
Users of ACE inhibitors and ARBs may have differed from non-users in ways that were not accounted for; for example, if they were healthier than non-users with similar measured covariates, hazard ratios may have been biased downward. Including a large number of covariates in the model increases the chance of introducing bias by conditioning on common shared effects. There was no measurement of actual medication use; rather, prescriptions were used to define exposure as binary. Incomplete testing coverage means that if the probability of being tested were associated with ACE inhibitors / ARB use in some way (e.g., through socioeconomic status), results would be biased. ICU admission is not a perfect proxy for disease severity, since limited capacity meant admitted patients were those who were deemed most likely to benefit from admission. This may have resulted in selection bias, though the direction is not obvious. During the study period, almost all (98.6%) of COVID-19 PCR tests were performed at the hospital among symptomatic patients with indications for testing and hospital admission, so the results may not be generalizable to mild or asymptomatic cases.
This is the largest study to date of ACE inhibitors and ARBs with respect to risks of COVID-19 diagnosis and severe disease.
This review was posted on: 9 October 2020