STRENGTHS AND LIMITATIONS OF THIS STUDY
The use of administrative claims data enables longitudinal tracking of patients and reliable capturing of mortality in any setting.
Targeting Medicaid beneficiaries allows the study to investigate the effect of statins on a vulnerable population prone to lower socioeconomic status and greater racial disparity.
This study investigated the effect of the intensity of statin treatment on mortality post COVID diagnosis, for which there is less evidence.
This study is subject to unmeasured confounders and potential non-accurate information due to the observational nature and the use of administrative claims data.
The study findings may not be generalised to other population.
The COVID-19 pandemic started in the USA in January 2020 and has resulted in 7.5 million hospitalisations and 921 000 deaths as estimated by the Centers for Disease Control and Prevention between February 2020 and September 2021.1 COVID-19 also disrupted healthcare systems by creating additional barriers to timely care, worsening the healthcare crisis. Vulnerable populations were disproportionately affected by the pandemic.2 This inequity was most exacerbated among those belonging to the low-income population, racial and ethnic minorities and other vulnerable groups of society.3 Medicaid is often the primary source of health insurance for such populations.
Manifestations of the COVID-19 infection include acute respiratory distress syndrome (ARDS), myocardial injury and thrombosis.4 In terms of comorbidities, people with heart disease or diabetes, who typically require periodic access to care, were reported to have worse COVID-19-related outcomes than the general population.5 People with dyslipidaemia have been found to have a greater risk of developing severe symptoms of COVID-19.6 Considering the anti-inflammatory, antithrombotic and cardioprotective effects of statins, researchers have proposed it as a promising drug class for the treatment of COVID-19 as they can possibly attenuate the ARDS, myocardial, lipidaemic or thrombotic crises resulting from infection with COVID-19.7 8
States like Mississippi bear a disproportionate burden of cardiovascular or metabolic comorbidities. Mississippi ranks fourth among US states in terms of proportion of patients with cardiovascular diseases and third in terms of proportion of patients with diabetes.9 10 Given that statins are commonly prescribed to treat these comorbidities, Mississippi serves as a crucial setting to evaluate the association between statin use and COVID-19 outcomes. An examination of the potential protective benefits of statin use on COVID-19 outcomes, in such a medically vulnerable population, can help develop strategies to mitigate COVID-19-related severe health outcomes in this population.
Certain aspects on the role of statin use on COVID-19 outcomes are unclear in the current literature. The existing evidence regarding the impact of statin use on mortality is not consistent. Based on prior literature evidence from retrospective studies, statin use has been reported to be associated with a reduction in mortality rate.8 11–18 However, certain studies have also reported the association between antecedent statin use and higher short-term death risk.19–21 Still other studies and meta-analyses have reported that statin use was not associated with inhospital mortality.14 22–28 Additionally, a majority of existing observational studies have used electronic medical records, which limits the generalisable population to hospitalised patients with severe cases of COVID-19 and also limits the study outcome before discharge.4 7 29 30 Administrative claims data have advantages, including the ability to longitudinally track patients and capture healthcare utilisation reliably. Therefore, it is important to evaluate this association using administrative claims data to better understand this association in the real world. The evidence regarding the effect of the intensity of statin treatment on improved COVID-19-related outcomes is scarce. Lastly, current literature lacks information about how statin use affects mortality in the Medicaid population who generally have poor baseline health and have a greater likelihood of being severely affected by the COVID-19 pandemic.3
Therefore, the aim of this cohort study is to address the gaps in the literature by assessing the association between antecedent statin use and all-cause mortality among Mississippi Medicaid-enrolled patients between June 2019 and September 2021. Results of this study will help understand how antecedent statin use and its intensity may affect mortality and inform the development of effective treatment strategies to improve outcomes.
Data source and patient cohort
Mississippi Medicaid administrative claims data from 2019 to 2021 was used in this cohort study. The study was approved as exempt by the Institutional Review Board at the University of Mississippi, and a waiver of informed consent was granted due to the de-identified nature of the data (Protocol #20 x-336). This report adhered to the guidelines for cohort studies as outlined by the Strengthening the Reporting of Observational Studies in Epidemiology.31
The study sample included Mississippi Medicaid beneficiaries who had a diagnosis of COVID-19 between March 2020 and June 2021 in any setting. Patients with COVID-19 were included if they were aged 19–64 years and their first COVID-19 diagnosis in the data set was considered as the index date. The International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes (U07.1, B34.2) were used to identify confirmed COVID-19 cases. The sample was restricted to those with continuous enrolment in Mississippi Medicaid for 9 months before and 3 months after the index date (ie, the first COVID-19 diagnosis date between March 2020 and June 2021) of each patient. For patients who died within 3 months post the index date, the continuous enrolment was required from 9 months before index date to the date of death. Due to a lack of complete reimbursement billing records, those who were dual enrolled in Medicare at any time during the study period were excluded from the study sample. Medicare is the primary payer for these individuals, and thus, these individuals were excluded to avoid any potential bias arising due to our lack of access to information on Medicare-covered medical services and prescription fills.
Patient and public involvement
No patient involved.
The primary exposure variable was antecedent statin use in the 3-month period before index date. Statin use was extracted from the prescription fill records using national drug codes for each study patient. Patients who initiated statins after COVID-19 diagnosis in the non-statin user group were excluded from the study sample to mitigate potential bias in the estimate of statins’ effect.
Propensity score matching
Given the fact that the higher use of statins in patients with a greater cardiovascular disease burden might confound the true effect of statin use compared with patient without statin user, the propensity score matching approach was employed to match each statin with one non-statin user.32 A logistic regression model was used to predict the propensity of statin administration, adjusted for age, sex, race, a history of comorbidities and prescription medication use. Clinical comorbidities, including chronic lung disease, hypertension, diabetes, acute myocardial infarction, hyperlipidaemia, heart failure, stroke, ischaemic heart disease, atrial fibrillation, chronic liver disease and kidney disease, were identified in medical claims using ICD-10-CM codes in the 6 months before statin use identification period. The records of prescription drug use, including beta-blockers, angiotensin receptor blockers, oral anticoagulants, angiotensin-converting enzyme inhibitors and antiplatelet P2Y12 receptor inhibitors, were tracked from the pharmacy claims in the same identification period as the comorbidities. The matching was based on the logit of propensity score implementing a nearest-neighbour strategy with a calliper width equal to 0.2.33 In the ‘nearest neighbour matching within a specified calliper distance’ approach, each antecedent statin user was matched to one non-statin user whose propensity score was closest to that of the corresponding antecedent statin user, and the difference in their propensity scores was below 0.2 of the SD of the logit of the propensity score. This method was employed as it has been shown to produce less biased estimates as compared with other propensity score matching algorithms since the matched pairs are restricted to be within a specified distance of each other.34 Additionally, the study used a calliper width of 0.2 of the SD of the logit of the propensity score as it has been shown to produce the most precise estimates of the treatment effect in several simulation studies comparing various calliper widths.33 35 Standardised mean differences were used to assess the balance of covariates between the matched groups.36
The outcomes of interest in this study were 30-day, 60-day and 90-day all-cause death following index COVID-19 diagnosis.
Mean and SD as well as counts and percentage were used to depict baseline patient characteristics and study outcomes, as appropriate. McNemar’s tests were used to make the unadjusted comparisons between statin users and matched non-statin users on all-cause death outcomes.
Conditional logistic regression models were used to estimate the effect of any antecedent statin use on study outcomes. Variables that were not used in the matching, including a history of cancer, the month of COVID-19 diagnosis and long-term care residency at diagnosis, were included as covariates in the regression models. To further explore the effect of the intensity of statin use, conditional logistic regression models were also performed to evaluate the relationship between the intensity of statin use and our study outcomes. The intensity of statin use was determined by the dose and statin type on statin prescriptions based on clinical guidance and classified to low/moderate and high.37 Statins in the low-intensity, medium-intensity and high-intensity categories have been listed in online supplemental table 1.
Sensitivity analysis was conducted in a subgroup of patients with cardiovascular diseases (ie, hypertension, stroke, coronary artery disease and transient ischaemic attack), for whom statins are usually prescribed7 and are more likely to experience severe COVID-19 outcomes.4 38 Since the subgroup of patients was no longer matched, unconditional logistic regression models adjusted for all variables were used.
Statistical significance was determined based on a two-sided level of α=0.05. All statistical analyses were conducted using SAS, V.9.4 (SAS Institute Inc., Cary, NC, USA).
A total of 10 792 patients with COVID-19 met the inclusion and exclusion criteria (figure 1), with 1415 statin users (13.1%) identified. After applying propensity score matching, we included 1107 matched pairs of statin users and non-statin users for analysis. The mean (SD) age of statin users was 51.0 (9.6) years, with 507 patients (45.8%) aged >55 years. Among statin users, 739 (66.8%) were women, and 672 (60.7%) were African Americans. The demographic characteristics of the non-statin user group were similar. Most patients had hypertension (statin users 74.1% vs non-statin users 77.7%), approximately half of patients had hyperlipidaemia (statin users 49.2% vs non-statin users 46.0%) and diabetes (statin users 48.6% vs non-statin users 49.8%), and more than 30% of patients had used beta-blockers (statin users 33.9% vs non-statin users 34.2%). The standardised mean difference values indicate the two groups were well balanced (table 1). Most patients (statin users 72.8% vs non-statin users 69.4%) were first diagnosed with COVID-19 in 2020. Of the statin users, 75 (6.8%) patients had cancer, while 95 (8.6%) non-statin users had cancer. Other baseline characteristics of Medicaid beneficiaries with COVID-19 and statin use and matched non-statin use beneficiaries are presented in table 1.
McNemar’s tests (table 2) showed a statistically significant lower proportion of statin users experiencing 30-day (39 statin users (3.5%) vs 60 non-statin users (5.4%), p=0.03), 60-day (49 statin users (4.4%) vs 73 non-statin users (6.6%), p=0.03) and 90-day (53 statin users (4.8%) vs 82 non-statin users (7.4%), p=0.01) mortality compared with non-statin users. Results of the multivariable analysis for estimating the association between statin use and all-cause death are presented in table 3. Antecedent statin use was significantly associated with lower odds of mortality in the follow-up period: adjusted OR (aOR), 0.51 (95% CI: 0.31 to 0.83) in 30 days post index period; aOR, 0.56 (95% CI: 0.37 to 0.85) in 60 days post index period; and aOR, 0.55 (95% CI: 0.37 to 0.82) in 90 days post index period.
Among patients using statins, 707 (63.9%) patients were classified as using a low-intensity or moderate-intensity statin, and 400 (36.1%) were classified as using a high-intensity statin. Table 4 presents the results of the adjusted analysis to assess the association of intensity of statin use and study outcomes. Those with low-intensity/moderate-intensity statin use had lower odds of mortality in the 60-day follow-up period (aOR: 0.56 (95% CI, 0.33 to 0.95)) and in the 90-day follow-up period (aOR: 0.51 (95% CI, 0.31 to 0.85)) compared with non-statin users. Although not statistically significant, there was a trend suggesting that patients with low-intensity/moderate-intensity statin use had lower odds of mortality within 30 days of the index date (aOR: 0.55 (0.30, 1.01)) compared with those who did not use statins. High-intensity statin use was associated with a lower risk of mortality within 30 days (aOR: 0.45 (95% CI, 0.21 to 0.97)) compared with no statin use.
Online supplemental figures 1–3 present the results of the sensitivity analysis, which included 1917 beneficiaries with cardiovascular diseases and COVID-19. Among those patients, statin use was associated with decreased odds of mortality within the 60-day and 90-day periods following COVID-19 diagnosis (aOR of 60-day death, 0.65 (95% CI, 0.44 to 0.99) and aOR of 90-day death, 0.62 (95% CI, 0.42 to 0.91)).
This study found that antecedent statin use was associated with a lower likelihood of all-cause death in the 90-day period post diagnosis, even after adjusting for sociodemographic and clinical characteristics. This study also looked at the effect of intensity of statin use on mortality and found that both low/moderate and high intensity of statin use is associated with a decreased risk of mortality post COVID-19 diagnosis. Similar study findings are found in our sensitivity analysis of patients with cardiovascular conditions. To the authors’ knowledge, this is one of the very few studies in published literature that assesses the impact of statin use and also intensity of statin use on COVID-19 outcomes using administrative claims data. Utilising claims data to assess this relationship provides the advantage of being able to follow a real-world patient population longitudinally from exposure to outcomes. Moreover, assessing this relationship among the Medicaid beneficiaries of Mississippi who might be disproportionately affected by COVID-19 adds to the uniqueness of this study. Therefore, this study adds important evidence to the published literature on the impact of antecedent statin use and intensity on COVID-19 outcomes.
The literature reports mixed findings regarding the impacts of antecedent statin use among COVID-19 patients. Several meta-analyses demonstrated that statin use prior to COVID-19 infection or COVID-related hospitalisation was associated with a reduced risk of mortality in the short term.14–17 39 Recent large cohort studies also demonstrated the beneficial effect of historical statin use on COVID-19-related death.30 40 However, some other studies reported that the significant association was detected only in post diagnosis/hospitalisation statin use and improved outcomes, but not in antecedent statin use.41–43 As with prior analyses, we provide further evidence of the statin’s protective effect against death post COVID-19 infection for 90 days in Medicaid patients treated in a routine clinical setting. As a result of statins’ ability to prevent severe COVID-19 outcomes (eg, lung injury, lung fibrosis, respiratory failure and death), statins can increase the activity of angiotensin-converting enzyme 2 (ACE2), which significantly declines when SARS-CoV-2 enters the host cell.44 45 The enzyme ACE2 is an essential enzyme in the renin-angiotensin-aldosterone system, inhibiting the activity of angiotensin II that promotes cardiovascular disease.46 In addition, statins have pleiotropic effects that might contribute to their benefits in COVID-19 patients, including enhancing endothelial dysfunction, antioxidant properties, immunomodulatory properties and antithrombotic properties.47 48 Based on our findings, statin use has a protective effect against death, which may be due to the severe COVID-19-related symptoms and conditions, particularly for patients with chronic conditions which require statins prior to COVID-19 infection.
Statin intensity and COVID-19 outcomes have been studied sparsely. Choi et al conducted a retrospective study of patients of Mount Sinai Health System hospitals.49 Compared with patients with no statin use, both low-intensity to moderate-intensity and high-intensity statin users had lower rates of death.49 Using French National Healthcare Data System database, Bouillon et al found a lower risk of inhospital death from COVID-19 for low-intensity and moderate-intensity statin users as compared with non-users, but not for high-intensity statin users.18 Our study revealed that low-moderate intensity statin use was significantly associated with a decreased risk of death at 60 and 90 days, as well as a lower mortality rate at 30 days. The small sample size of COVID-19 patients who died within 30 days may explain the lack of statistical significance of this estimate. A larger sample size would be needed to confirm the effect on short-term outcomes. Further, we found that high statin intensity was only associated with a reduced risk of mortality in 30 days post COVID-19 diagnosis. This finding may be explained by the fact that patients prescribed higher-intensity statins often have more severe comorbid cardiovascular conditions, leading to a higher risk of COVID-19 infections and exacerbations, which can potentially counteract the protective effects of statins. This may also be due to suboptimal adherence to high doses of statins prior to infection with COVID-19. A further study accounting for statin adherence is required to explore the association between antecedent statin use and COVID outcomes.
The COVID-19 pandemic resulted in a public health crisis in the USA and caused a disruption of healthcare access across the country. Vulnerable populations such as Black, Asian and other minority communities and Medicaid enrollees were disproportionately affected.50 51 According to a recent report from the American College of Cardiology, patients with cardiovascular disease also suffered from the indirect impact of the COVID-19 pandemic, caused by the substantial disruption in care.52 Recent studies have reported the severe impact of the COVID-19 pandemic on medication adherence among patients with chronic diseases, which eventually resulted in worsening of disease symptoms.53 54 COVID-19 still persists, and as a healthcare system, we need to be better prepared for facing such disruptions in the future. Telehealth is one such tool that can be leveraged to ensure continued access to healthcare among people with chronic diseases. A systematic review reported that eHealth and telehealth interventions targeting medication adherence among patients with hypertension, diabetes or dyslipidaemia resulted in improved adherence.55 Although many states had expanded Medicaid coverage for telehealth during the COVID-19 pandemic, it’s imperative for state Medicaid programmes to ensure continued access to telehealth provisions so that the healthcare system is better prepared for handing such disruptions in the future.56
Our study is subject to certain limitations associated with the use of administrative claims data for healthcare research. Administrative claims only have information about whether the medication has been dispensed and cannot indicate if medication has been consumed by the patient. Use of administrative claims data limited our ability to assess disease severity in the study sample. This study uses claims for the Mississippi Medicaid programme; therefore, the results may not be generalisable to the national population or the population with commercial insurance. It is important to note that this study did not account for COVID-19 vaccination status as a covariate due to potential limitations in accurately capturing vaccination records from various sources, leading to unreliable information in the Medicaid claims data. Future studies should incorporate vaccine registry data to assess the impact of vaccination status on the relationship between statin use and COVID-19 outcomes. Additionally, this study did not account for use of approved COVID-19 treatments (eg, Paxlovid) as many of these medications were approved for use in the USA in December 2021 or later, and our study period (until September 2021) pre-dates their approval. Future research should account for use of approved COVID-19 treatments while examining the impact of statin use on COVID-19 outcomes. For statin users, the effect of different statin intensity was not evaluated in our study. Several retrospective studies have found no difference in mortality rates for COVID-19 patients whose statin therapy was administered at varying intensities.57–59 Therefore, our study focused on comparing the mortality rates between statin users and non-users, without evaluating the effect of different statin intensities. The effect of statin compliance and dosage on mortality was not assessed in this study. Based on the results of this study, statins’ clinical benefits on COVID-19 outcomes by compliance and dosing range should be investigated in the future.
Statins have the potential to prevent mortality among patients diagnosed with COVID-19, in particular for those with high-risk comorbidities. Mississippi Medicaid beneficiaries with statin use were found to have over 40% lower risk of death within 90 days post COVID-19 diagnosis. Our study highlights the significance of continuity of care for patients with chronic conditions during the pandemic. Findings of this study can inform clinical decisions for healthcare providers and can also assist policy makers in preparing for future interruptions of healthcare. Further exploration is needed to investigate the potential mechanisms as well as optimal use and dosage of statins in COVID-19 treatment. Further studies of other vulnerable populations are also necessary due to the limited generalisability of Mississippi Medicaid data.
Data availability statement
Data may be obtained from a third party and are not publicly available. The data for this study were obtained via a data use agreement with the Mississippi Division of Medicaid and cannot be shared publicly. Individuals interested in accessing the data can request access through CMS (resdac.org).
Patient consent for publication
This study involves human participants. The study was approved as exempt by the Institutional Review Board at the University of Mississippi, and a waiver of informed consent was granted due to the de-identified nature of the data (Protocol #20x-336). The consent for participation was not necessary due to the de-identified nature of the data.
This post was originally published on https://bmjopen.bmj.com