Statins Are Underused in Women With NAFLD After Cardiovascular Events Compared With Matched Control Subjects

Abstract

Patients with nonalcoholic fatty liver disease (NAFLD) often have hyperlipidemia and are at a higher risk of cardiovascular disease (CVD).1Targher G. et al.Gut. 2020; 69: 1691-1705Crossref PubMed Scopus (135) Google Scholar The first-line lipid-lowering drugs, statins are widely used for primary and secondary CVD prevention.2Cai T. et al.BMJ. 2021; 374: n1537Crossref PubMed Scopus (28) Google Scholar However, there have been concerns about statins underprescription for patients with NAFLD because of fear of hepatoxicity.3Rzouq F.S. et al.Am J Med Sci. 2010; 340: 89-93Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar We aimed to describe and compare the pattern of statins use before and after CVD events in patients with NAFLD to the general population. We used data from the Decoding the Epidemiology of Liver Disease in Sweden (DELIVER) cohort from 2006 to 2016.4Hagstrom H, et al. Scand J Gastroenterol https://doi.org/10.1080/00365521.2022.2051202Google Scholar Methods are shown in the Supplementary Material. Briefly, patients diagnosed with NAFLD and matched control subjects (on age, sex, and municipality at the first NAFLD diagnosis) from the general population were followed from the NAFLD diagnosis until the date of a first nonfatal CVD occurrence, considered the index date. The Swedish Prescribed Drug Register was used to identify dispensed drugs with a 6-month lookback and lookforward period from the index date. Patients were considered to be on primary CVD prophylaxis if they had ≥1 dispensed statin in the lookback period. Likewise, patients were on secondary CVD prophylaxis if they had ≥1 dispensed statin in the lookforward period. Therefore, four groups of patients were defined: (1) no statins before nor after CVD, (2) statins before and after CVD, (3) statins before CVD only, and (4) statins after CVD only. “Statins after CVD” refers to the sum of patients with statins before and after CVD, plus those with statins after CVD only. We also defined statins “ever dispensed before CVD” as any statins dispensation at any time before the index CVD. In total, 623 patients with NAFLD (6.7%) and 1572 (1.7%) control subjects developed nonfatal CVD events. The mean age at CVD diagnosis was 68.9 years (standard deviation, 11.0) and 52.1% were men. At index CVD, type 2 diabetes, hypertension, obesity, and dyslipidemia were more common in patients with NAFLD than the control subjects (data not shown). The proportion of patients with statins ever dispensed before CVD was 48.1% for NAFLD and 26.7% for control subjects (P < .001) (Table 1). Within six months of index CVD, patients with NAFLD were more likely to have statins dispensed than control subjects throughout the period (29.5% vs 15.6%; P < .001). This imbalance was offset by more new statins users in the control subjects (36.1% in NAFLD vs 47.1% in control subjects), leading to an overall nonsignificant difference in statins after index CVD between NAFLD (65.6%) and control subjects (62.7%) in a crude analysis (P = .20). There was no difference in statins discontinuation between the two groups. We observed a similar pattern of statin dispensations for stroke or myocardial infarction separately; however, patients with NAFLD were dispensed statins more frequently than control subjects before and after myocardial infarction, whereas no such difference was found for patients with stroke.Table 1Statins and Other Drugs Dispensations Before and After Cardiovascular Events by NAFLD StatusStatin dispensationsNAFLDMatched controlP valueCVD Number of patients with CVD6231572 Ever dispensed before CVD300 (48.1)420 (26.7)< .001 Within 6 mo before and after CVDNo statins before nor after200 (32.1)553 (35.2).171Statins before and after184 (29.5)245 (15.6)< .001Statins before only14 (2.3)33 (2.1).829Statins after only225 (36.1)741 (47.1)< .001Statins after CVD (all)aStatin after event refers to the sum of statin before and after CVD, plus statin after only.409 (65.6)986 (62.7).199Stroke Number of patients2201428 Ever dispensed before stroke79 (35.9)363 (25.4)< .001 Within 6 mo before and after strokeNo statins before nor after89 (40.4)510 (35.7).174Statins before and after39 (17.7)214 (15.0).294Statins before only6 (2.7)27 (1.9).410Statins after only86 (39.1)677 (47.4).021Statins after stroke (all)aStatin after event refers to the sum of statin before and after CVD, plus statin after only.125 (56.8)891 (62.4).113MI Number of patients403144 Ever dispensed before MI221 (54.8)57 (39.6).002 Within 6 mo before and after MINo statins before nor after111 (27.5)43 (29.9).596Statins before and after145 (36.0)31 (21.5).001Statins before only8 (2.0)6 (4.2).155Statins after only139 (34.5)64 (44.4).034Statins after MI (all)aStatin after event refers to the sum of statin before and after CVD, plus statin after only.284 (70.5)95 (66.0).315NOTE. Values are n (%).CVD, cardiovascular disease; MI, myocardial infarction; NAFLD, nonalcoholic fatty liver disease.a Statin after event refers to the sum of statin before and after CVD, plus statin after only. Open table in a new tab NOTE. Values are n (%). CVD, cardiovascular disease; MI, myocardial infarction; NAFLD, nonalcoholic fatty liver disease. We performed multivariable logistic regression to investigate the factors associated with statins after CVD events (include patients with statins before and after CVD, plus those with statins after CVD only). The multiadjusted odds ratio (aOR) and 95% confidence interval (CI) was 0.77 (95% CI, 0.60–0.96) for patients with NAFLD compared with the matched control subjects after adjusting for demographic factors and comorbidities (Supplementary Table 1). Women were less likely to have statins dispensed (aOR, 0.77; 95% CI, 0.64–0.94), whereas patients with type 2 diabetes (aOR, 1.34; 95% CI, 1.06–1.71), dyslipidemia (aOR, 4.63; 95% CI, 3.39–6.30), hypertension (aOR, 1.35; 95% CI, 1.08–1.68), or aspirin users (aOR, 1.48; 95% CI, 1.22–1.81) were more likely to have statins dispensed after CVD. Importantly, we found a multiplicative interaction between gender and NAFLD (P for interaction < .001), in that statin dispensations were significantly lower in women with NAFLD compared with matched control subjects (aOR, 0.56; 95% CI, 0.40–0.78) but not in their male counterparts (aOR, 1.03; 95% CI, 0.76–1.42). Thus, underdispensations of statins in NAFLD were confined to women. Dispensations of other drugs frequently administered after CVD events (eg, β-blockers) was higher in patients with NAFLD than control subjects (Supplementary Table 2). However, aspirin was less frequently dispensed for patients with NAFLD after stroke. Although statins were commonly dispensed in patients with NAFLD before a CVD event, a large proportion of these patients who had an indication for primary CVD prevention, did not have a dispensation within six months of their first CVD events. This could signal a poor compliance to statin therapy and should be addressed in future studies. After the index CVD, patients with NAFLD were dispensed statins to a lower extent than control subjects after considering confounders, and this underdispensation was confined to women with NAFLD. The more frequent dispensations of other drugs to NAFLD suggests that a more severe risk profile in this group than patients with CVD from the general population without NAFLD diagnosis. Because of potential hepatoxicity, there have been concerns among clinicians in initiation and continuation of statins among patients with underlying liver disease.3Rzouq F.S. et al.Am J Med Sci. 2010; 340: 89-93Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar A previous study showed that statins were underprescribed for patients with indications, and that this was particularly common among those with NAFLD.5del Ben M. et al.Nutr Metab Cardiovasc Dis. 2017; 27: 161-167Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar It is now known that the cardiovascular benefit from statins outweighs the small risk for hepatoxicity.6Riley P. et al.Cardiovasc Ther. 2009; 27: 216-220Crossref PubMed Scopus (24) Google Scholar Statins are considered safe in NAFLD,7Pose E. et al.J Hepatol. 2019; 70: 194-202Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar and we found that statins were dispensed more commonly to patients with NAFLD than the general population for primary CVD prevention. However, for secondary CVD prevention, statins underdispensation was observed in patients with NAFLD, especially in women with NAFLD. Underprescription of statins to women in general was reported in other studies.8Zhang H. et al.PLoS One. 2016; 11e0155228Google Scholar,9Nanna M.G. et al.Circ Cardiovasc Qual Outcomes. 2019; 12e005562Google Scholar There might be a gender disparity in the response to CVD investigations and treatments, such that women are less likely to receive treatments for myocardial infarction than men.10D’Onofrio G. et al.Circulation. 2015; 131: 1324-1332Crossref PubMed Scopus (155) Google Scholar We confirm that this gender bias seems to be present also in patients with NAFLD. Reasons for this can be underprescription of statins to women with NAFLD because of gender bias, fear of hepatotoxicity, or other factors. Another explanation would be that women with NAFLD are less likely to be compliant to statin therapy. Both explanations suggest a need for improved education on the importance of statin therapy. The strengths of the study include the use of national registers to identify patients and drug dispensations from all pharmacies in Sweden. Limitations include that diagnoses only made in primary care are not recorded; however, patients with CVD events are likely to be followed in higher tiers of care and thus be captured by our methodology. We acknowledge some missing data on disease severity and possible confounders (eg, lipoprotein and aminotransferase levels). In conclusion, statins were underused in women with NAFLD after a CVD event compared with general population control subjects also with a CVD event. Appropriate prescription and investigation of treatment compliance should be encouraged in women with NAFLD for improved care after CVD. This cross-sectional study is based on data from the Decoding the Epidemiology of Liver Disease in Sweden (DELIVER).1Hagstrom H, et al. https://doi.org/10.1080/00365521.2022.2051202 (in press)Google Scholar All patients diagnosed with nonalcoholic fatty liver disease (NAFLD) according to the International Code of Disease Classification from the Swedish National Patient Register (NPR) between January 1987 and December 2016, was identified and matched up to 10 individuals on age, sex, and municipality at the year of diagnosis from the general population without a diagnosis of NAFLD (matched control) randomly selected from the Total Population Register.2Ludvigsson J.F. et al.Eur J Epidemiol. 2016; 31: 125-136Crossref PubMed Scopus (652) Google Scholar We excluded participants aged <18, who had cardiovascular disease (CVD) before NAFLD diagnosis, and had other contaminant liver diseases. Among all participants, those who survived a CVD event of a stroke or myocardial infarction diagnosed from NPR were identified and linked to the Swedish Prescribed Drug Register to track drug dispensed from all Swedish pharmacies using the codes from the Anatomical Therapeutic Chemical Classification System.3The Swedish National Board for Health and WelfareNational Prescribed Drug Register.https://www.socialstyrelsen.se/en/statistics-and-data/registers/national-prescribed-drug-registerDate accessed: December 17, 2021Google Scholar Because the Prescribed Drug Register was established in July 2005, patients who developed a nonfatal CVD event from July 2006 to December 2016 were enrolled in the study. The date of first CVD event was used as the index date. A 6-month lookback period from the index date was used to define drug dispensation before CVD. Likewise, to determine dispensation after CVD, we identified dispensation records from the index date forward 6 months. Patients were considered exposed if they had at least 1 dispensation for statin in the 6 months before CVD. Similarly, patients were considered exposed after the index CVD if they had at least 1 dispensation for statins and other drugs of interest in the 6 months following the index date. Therefore, 4 groups of patients were categorized based on the exposure to statin within 6 months: (1) no statins before and after CVD, (2) statins before and after CVD, (3) statins before CVD only, and (4) statins after CVD only. “Statins after CVD” refers to the total number of statins dispensations before and after CVD, plus statins only after CVD. Furthermore, because statins might be dispensed for periods longer than 6 months, we define statin “ever dispensed before CVD” as at least 1 statin dispensation at any time before index CVD (ie, also before the 6-month window that was used to investigate statin use in conjunction to the first CVD event). We used diagnosis from NPR and drug dispensations from Swedish Prescribed Drug Register to identify patients with type 2 diabetes, hypertension, and dyslipidemia. Obesity was defined from NPR. Aspirin user was defined as at least 1 aspirin dispensation at any time before index CVD from Swedish Prescribed Drug Register. We used chi-square tests or Fisher exact tests to compare statins and other drugs dispensations before and after CVD event by NAFLD status. Univariate logistic regression models were used to calculate odds ratio of factors associated with statins dispensations after CVD events. Adjusted ORs were estimated after adjusting for age groups, sex, type 2 diabetes, hypertension, obesity, and dyslipidemia, and previous aspirin use in the multivariate logistic regression models. Supplementary Table 1OR and 95% CI Associated With Statins After CVD EventaStatins after CVD include patients with statins before and after CVD, plus those with statins after CVD only.VariableCrude ORbUnivariate model. (95% CI)Adjusted ORcMultivariable model mutually adjusted for variables listed above. (95% CI)NAFLD1.13 (0.93–1.37)0.77 (0.60–0.96)Age groups <40ReferenceReference 40–644.68 (1.73–12.6)3.31 (1.17–9.29) 65–853.58 (1.33–9.62)2.68 (0.95–7.52) >850.91 (0.32–2.55)0.67 (0.22–2.00)Female sex0.75 (0.63–0.89)0.77 (0.64–0.94)Type 2 diabetes1.67 (1.35–2.05)1.34 (1.06–1.71)Dyslipidemia3.45 (2.24–5.31)4.63 (3.39–6.30)Hypertension1.62 (1.34–1.97)1.35 (1.08–1.68)Obesity1.17 (0.82–1.66)0.74 (0.49–1.10)Previous aspirin user1.71 (1.43–2.05)1.48 (1.22–1.81)Number of total drugs prescription active at index date1.06 (1.02–1.09)1.01 (0.97–1.05)CI, confidence interval; CVD, cardiovascular disease; NAFLD, nonalcoholic fatty liver disease; OR odds ratio.a Statins after CVD include patients with statins before and after CVD, plus those with statins after CVD only.b Univariate model.c Multivariable model mutually adjusted for variables listed above. Open table in a new tab Supplementary Table 2Dispensation of Other Drugs After Cardiovascular Events, Stratified on Presence of NAFLD Before First CVD EventDrug typesNAFLDMatched cohortP valueDispensation after CVD Number of patients6231572 Aspirin380 (61.0)817 (52.0)< .001 Other antiplatelets1 (0.1)3 (0.2).881aFisher exact test. Anticoagulants64 (10.3)157 (10.0).841 Ezetimibe14 (2.3)14 (0.89).011 β-Blockers391 (62.7)625 (39.8)< .001 Calcium channel blockers162 (26.0)525 (33.4).001 ACE inhibitors379 (60.8)817 (52.0)< .001 ARBs121 (19.4)248 (15.8).039 Oral antidiabetic drugs157 (25.2)176 (11.2)< .001 Insulin139 (22.3)115 (7.3)< .001 GLP1-RA13 (2.1)1 (0)< .001aFisher exact test.Dispensation after stroke Number of patients2201428 Aspirin87 (39.6)711 (49.8).005 Other antiplatelets1 (0.5)2 (0.1).308aFisher exact test. Anticoagulants20 (9.1)140 (9.8).740 Ezetimibe2 (0.9)13 (0.9)1.000aFisher exact test. β-Blockers83 (37.7)516 (36.1).648 Calcium channel blockers61 (27.7)481 (33.7).090 ACE inhibitors107 (48.6)722 (50.6).595 ARBs33 (15.0)205 (14.4).800 Oral antidiabetic drugs53 (24.1)152 (10.6)< .001 Insulin38 (17.3)96 (6.7)< .001 GLP1-RA2 (0.9)1 (0.1).049aFisher exact test.Dispensation after MI Number of patients403144 Aspirin293 (72.7)106 (73.6).834 Other antiplatelets0 (0)1 (0.7).263 Anticoagulants44 (10.9)17 (11.8).772 Ezetimibe12 (2.9)1 (0.7).200aFisher exact test. β-Blockers308 (76.4)109 (75.7).859 Calcium channel blockers101 (25.1)44 (30.6).226 ACE inhibitors272 (67.5)95 (66.0).739 ARBs88 (21.8)43 (30.0).053 Oral antidiabetic drugs104 (25.8)24 (16.7).026 Insulin101 (25.1)19 (13.2)< .001 GLP1-RA11 (2.7)0 (0)< .001NOTE. Values are n (%).ACE, angiotensin-converting-enzyme; ARBs, angiotensin II receptor blockers; CVD, cardiovascular disease; GLP1-RA, glucagon-like peptide-1 receptor agonist; MI, myocardial infarction; NAFLD, nonalcoholic fatty liver disease.a Fisher exact test. Open table in a new tab CI, confidence interval; CVD, cardiovascular disease; NAFLD, nonalcoholic fatty liver disease; OR odds ratio. NOTE. Values are n (%). ACE, angiotensin-converting-enzyme; ARBs, angiotensin II receptor blockers; CVD, cardiovascular disease; GLP1-RA, glucagon-like peptide-1 receptor agonist; MI, myocardial infarction; NAFLD, nonalcoholic fatty liver disease.

Topics & Concepts