Statin usage and cardiovascular risk among people living with HIV in the U.S. Military HIV Natural History Study

Abstract

Objectives:

Using the American College of Cardiology/American Heart Association 2013 atherosclerotic cardiovascular disease (ASCVD) management guidelines, we conducted a retrospective cross-sectional analysis of people living with HIV in the U.S. Military HIV Natural History Study to determine whether individuals were receiving statins when indicated.

Methods:

Prescription data was taken from Military Health System data. Statin eligibility was defined by ASCVD guidelines. We used the 10-year ASCVD pooled cohorts’ equation to evaluate risk for each participant.

Results:

Across all categories, 31.9% (n=390) of individuals met criteria for statin use, and when adding these subjects to the number of those already receiving statins (n=96), 62.1% of all eligible subjects (n=302/486) were actually receiving statin therapy. In multivariable analysis, individuals of African American race (OR 0.48 [95% CI 0.31–0.73]) or Hispanic ethnicity (OR 0.42 [95% CI 0.19–0.94]) were less likely to receive statin prescriptions than White individuals. Individuals with a higher CD4 count (OR 1.12 [95% 1.05–1.20 per 100 cells/uL]) were significantly more likely to receive a statin prescription.

Conclusions:

These data highlight discrepancies between ASCVD guidelines and primary care management of PLWH in the military health system, along with important racial differences. Targeted interventions are critical to identify and treat appropriate candidates for statin therapy among PLWH in the military and other settings.

Introduction:

The implementation of combination antiretroviral therapy (ART) for human immunodeficiency virus (HIV) infection has increased the lifespan of persons living with HIV (PLWH). As a result, the leading causes of mortality have shifted from complications of acquired immunodeficiency syndrome (AIDS) to non-AIDS-related illnesses, such as cardiovascular disease.(1–3)As PLWH continue to live longer, the delivery of primary care services has shifted to management of co-morbid conditions and delivery of preventive care services, to include cardiovascular risk mitigation such as with the use of statins.

The baseline risk of cardiovascular death in PLWH is estimated to be at least 50% greater than that of HIV-uninfected persons,(4) which may occur for a variety of reasons, including chronic inflammation, unsuppressed HIV viremia, and T-helper cell disturbances; these have been shown to increase risk for cardiovascular and related events.(5, 6) Antiretroviral medications, including protease inhibitors (PIs) and nucleoside/nucleotide reverse-transcriptase inhibitors (NRTIs) have also been linked to increased cardiovascular risk, though these effects are sometimes debated.(7–11) Last, traditional risk factors, such as smoking, hypertension, dyslipidemia, obesity, and insulin resistance are more prevalent among PLWH and may now account for the majority of CVD risk.(12) Taken together, these factors synergistically place PLWH at a higher risk of myocardial infarction (MI), stroke, and cardiovascular-related death.(13–16)

Effective strategies are therefore needed to mitigate cardiovascular risk. Statin-based therapy is the standard of care for all individuals at risk and is notably included in the Infectious Disease Society of America (IDSA) guidelines for primary care of PLWH(17), further supported by a scientific statement from the American Heart Association (AHA).(18) In major studies involving HIV-uninfected individuals, targeted statin therapy demonstrated a 20% reduction in cardiovascular events and 10% reduction in overall mortality over five years.(19, 20) However, adherence to these guidelines is poor, both among HIV-negative individuals and PLWH. In several population-based studies, anywhere between 30–60% of individuals for whom statins are recommended actually take them.(21–24) Statin use is generally low among HIV-positive men, even among men who have already had previous myocardial infarctions.(25) Similarly, only half of eligible women living with HIV are prescribed statins.(26) While a number of clinical and systemic factors can impair practice, patient-related factors including higher pill burden, lower educational level, or membership in disadvantaged racial, sexual, or gender groups have also been described among PLWH that may contribute to guideline non-adherence.(27)

Using the recommendations for statin therapy based on IDSA and American Heart Association/American College of Cardiology (AHA/ACC) 2013 guidelines, the primary objective of our study was to determine the percentage of PLWH who were eligible to receive statin-based therapy, evaluate the proportion who were receiving such therapy, and identify individual characteristics associated with appropriate statin prescription. As this study takes place in the context of the United States Department of Defense (DoD) and its Military Health System (MHS), a single-payer system providing health care to active-duty personnel, military retirees, and their families, we have a unique opportunity to examine practices where access to care is generally universal. We therefore seek to contribute to the general knowledge of CVD risk factor prevalence among an HIV-positive cohort, in whom the prevalence of cardiovascular disease may be underestimated.(6, 13)

Methods:

Study Design and Population:

The United States Military HIV Natural History Study (NHS) is an ongoing cohort study comprised of PLWH within the MHS and has been prospectively collecting data since 1986. For this analysis, we included NHS participants who were between the age of 21–75 years old and who had a study visit during the period of analysis (October 1, 2015 - September 30, 2016); subjects <40 years old were included in the analysis design as the military HIV-positive population is younger than many cohorts and we were not certain as to the cardiovascular risk profile in this group. Since 2002, all prescription data for NHS participants have been available through MHS medical records including the Pharmacy Data Transaction Service (PDTS), a central database that records medications provided by military treatment facilities or paid for by MHS Tricare insurance. To account for prescriptions obtained outside the MHS, medication use including statins were collected through coordinator interview and medical record review. For this analysis, we classified individuals as receiving statin therapy if they had a prescription filled within the study period or up to 6 months prior to their most recent study visit.

Statin eligibility was defined using the ACC/AHA 2013 atherosclerotic cardiovascular disease (ASCVD) management guidelines, whose criteria included the following: 1) known ASCVD (cerebral vascular accident or transient ischemic attack, acute coronary syndrome or angina, arterial revascularization, and peripheral arterial disease); 2) low-density lipoprotein (LDL) cholesterol > 190 mg/dL; 3) diagnosis of diabetes for individuals aged 40–75 with concurrent LDL > 70 mg/dL, or ASCVD risk calculated to be ≥ 7.5% over 10 years via the pooled cohorts equation (also known as risk estimator). Inputs into the ASCVD risk estimator include age, sex, race, blood pressure and use of anti-hypertensive medication, tobacco use, and cholesterol levels. For the purpose of this analysis, individuals already receiving statins were presumed to have an indication. The ACC/AHA 2013 guidelines were chosen as a single set of criteria as they were generally well-accepted regarding management of statins and had adequate time for implementation in the 2014 VA/DoD CVD prevention guidelines.(28) This study was approved by the Institutional Review Board of the Uniformed Services University (Bethesda, Maryland, USA) and at all NHS study sites.

Statistical Analysis:

Descriptive statistics were presented as medians (interquartile ranges, IQR) or frequencies (percentages), depending on whether the variables were continuous or categorical. Differences in covariates of interest between statin and non-statin users were assessed using Kruskal-Wallis test for continuous variables and Chi-square/Fisher’s exact test for categorical variables. In order to estimate a 10-year ASCVD risk score for each subject, we utilized the 10-year ASCVD risk estimator using the pooled cohort equation.(29) If input values for the ASCVD risk estimator were outside the required range, these values were replaced with the upper or lower threshold so as to calculate a subject’s 10-year ASCVD risk score. Unadjusted and adjusted logistic regression models were fitted to investigate the associations between statin use and covariates identified as potential risk factors by the investigators and through literature review; these factors included demographic (sex, race/ethnicity, age); medical (current smoking, SBP, anti-hypertensive drug use); and HIV/ART-related (HIV diagnosis era, h/o AIDS defining diagnosis, ART era, time from HIV diagnosis to ART initiation, use of mono/dual-NRTI before combination ART [“ARV before ART”], CD4, VL), and initial ART regimen [NNRTI, PI, boosted PI, INSTI, other; includes tenofovir y/n]. As the number of categories introduced by the initial ART regimen combinations was large, a separate model of ART regimens was evaluated (data not shown) and only the presence of tenofovir in the initial regimen was significantly associated with statin use, thus it was included in the model, but other ART categories were not. The set of covariates in the adjusted model were selected by a stepwise method with p-value < 0.05. All p-values were two-sided and were considered statistically significant for a p-value less than 0.05. Because the ASCVD calculator provides risk estimates for individuals between the ages of 40–75, we conducted a sensitivity analysis excluding individuals less than 40 years old to assess any potential impact on our findings. Analyses were performed with SAS 9.4. (SAS Institute, NC, USA)

Results:

A total of 1,223 subjects met inclusion criteria for the study and had all data needed to complete the ASCVD 10-year risk estimator. Selected demographic characteristics of statin recipients and non-recipients are presented in Table 1. The median age of all subjects was 47 years old (Interquartile Range (IQR)[33–55]); 94.8% (n=1,160) of subjects were male and 46% (n=557) were African American. The median age at HIV diagnosis was 29 years old (IQR: 25–36) and all subjects were receiving ART; median ART adherence was 96% (IQR 92–100%). Individuals receiving statin therapy were significantly older at the time of their visit and at the times of HIV diagnosis and ART initiation (p<0.001 for all comparisons). Individuals receiving statins also had lower CD4 counts at the time of ART initiation (p=0.013) and their most recent study visit (p=0.003). Significant differences were not observed in viral load between statin recipients and non-recipients.

Table 1 –

Baseline Characteristics by Statin Use

	Statin Use No	Statin Use Yes	Total	P-value
No.Subjects	921	302	1223
Demographics
Sexa				0.89
Male	874 (94.9)	286 (94.7)	1160 (94.8)
Female	47 (5.1)	16 (5.3)	63 (5.2)
Racea				<.0001
White	316 (34.3)	160 (53.0)	476 (38.9)
African American	436 (47.3)	121 (40.1)	557 (45.5)
Hispanic/Other	169 (18.3)	21 (7.0)	190 (15.5)
ASCVD Estimator Inputs
Total Cholesterol at Last visitb	174.0 [154.0 – 198.0]	166.0 [143.0 – 196.0]	173.0 [151.0 – 198.0]	0.0006
HDL at Last visitb	47.0 [39.0 – 57.0]	43.0 [35.0 – 53.0]	46.0 [38.0 – 56.0]	<.0001
LDL at Last visitb	106.0 [86.0 – 124.0]	97.0 [75.0 – 118.0]	104.0 [82.0 – 123.0]	<.0001
SBP at Last visitb	125.0 [116.0 – 134.0]	127.5 [119.0 – 137.0]	126.0 [117.0 – 135.0]	0.002
Age at Last visit (years)b	41.1 [31.1 – 51.2]	55.4 [50.3 – 60.9]	46.5 [33.4 – 54.6]	<.0001
Anti-hypertensive drug usea	233 (25.3)	202 (66.9)	435 (35.6)	<.0001
Smokinga	154 (16.7)	44 (14.6)	198 (16.2)	0.38
Age at HIV Diagnosis (years)b	28.1 [24.1 – 34.2]	34.2 [28.6 – 39.1]	29.2 [24.8 – 36.0]	<.0001
Age at ART Initiation (years)b	31.1 [25.9 – 37.8]	40.1 [34.9 – 45.0]	33.6 [27.5 – 40.2]	<.0001
Laboratory
CD4 at HIV Diagnosis (cells/ul)b	463.0 [333.0 – 613.0]	496.0 [344.0 – 641.0]	470.0 [335.0 – 621.0]	0.18
CD4 at HIV Diagnosis (cells/ul)a,c				0.48
0–200	53 (6.2)	15 (5.9)	68 (6.1)
201–350	183 (21.5)	50 (19.8)	233 (21.1)
351–500	246 (28.8)	64 (25.3)	310 (28.0)
500+	371 (43.5)	124 (49.0)	495 (44.8)
Missing	68 (7.4)	49 (16.2)	117 (9.6)
CD4 at ART Initiation (cells/ul)b,c	374.5 [270.0 – 512.0]	344.0 [223.0 – 480.0]	370.0 [264.0 – 499.0]	0.013
CD4 at Last visit (cells/ul)b,c	691.5 [525.5 – 865.0]	741.0 [581.0 – 956.0]	703.0 [539.0 – 887.0]	0.003
nadir CD4 (cells/ul)b	330.0 [237.5 – 440.0]	290.0 [192.0 – 374.5]	318.0 [226.0 – 424.5]	<.0001
log10(VL) at Last visitb	1.3 [1.3 – 1.3]	1.3 [1.3 – 1.3]	1.3 [1.3 – 1.3]	0.35
VL at Last visit (copies/mL)a				0.78
0–200	881 (95.7)	290 (96.0)	1171 (95.7)
200+	40 (4.3)	12 (4.0)	52 (4.3)

Notes: Data were presented as

^an (%) or

^bmedian (1^st quartile -3^rd quartile).

^cSubjects with missing values were not included in the computation of percentages and statistics. P-values were calculated by Kruskal-Wallis test and Chi-squared/Fisher’s exact test for continuous and categorical variables, respectively. ASCVD: atherosclerotic cardiovascular disease; HDL: high-density lipoproteins; LDL: low-density lipoproteins; SBP: systolic blood pressure

In terms of statin eligibility (Table 2), 6.7% (n=82) of individuals had history of a prior cardiovascular event, 10.9% (n=133) had diabetes, and only 0.5% (n=6) of subjects had LDL>190. Across all categories, 31.9% (n=390) individuals met at least one criterion for statin use according to the ACC/AHA cholesterol management guidelines. When adding these subjects to the number of those already receiving statins (n=96), 62.1% of all eligible subjects (n=302/486) were actually receiving statin therapy.

Table 2 –

Distribution of Conditions and Statin Use

	Statin Use - Noa	Statin Use - Yesa	Totalb	P-value
No. Subjects	921 (%)	302 (%)	1223 (%)
Conditions 1
C1a - Ever had ASCVD	20 (2.2)	62 (20.5)	82 (6.7)	<0.001
C1b - LDL > 190 mg/dL	5 (0.5)	1 (0.3)	6 (0.5)	----
C1c - DM w LDL > 70 mg/dL and age 40–75	43 (4.7)	60 (19.9)	103 (8.4)	<0.001
Ever had diabetes	49 (5.3)	84 (27.8)	133 (10.9)	<0.001
LDL > 70 mg/dL	817 (88.7)	245 (81.1)	1062 (86.1)	<0.001
Age b/n 40 and 75 years	476 (51.7)	289 (95.7)	765 (62.5)	<0.001
C1d - ASCVD Risk ≥ 7.5%	170 (18.5)	190 (62.9)	360 (29.4)
Summary
--Met none of Conditions C1a – C1d	737 (80.0)	96 (31.8)	833 (68.1)	<.0001
--At least met one of conditions C1a – C1d	184 (20.0)	206 (68.2)	390 (31.9)
--Met one of the Conditions and/or was receiving a statin	184 (20.0)	302 (100.0 )	486 (39.7)

Note: Data were presented as n with

^arow or

^bcolumn percentage. P-values were calculated by Chi-squared/Fisher’s exact test. ASCVD: atherosclerotic cardiovascular disease; LDL: low-density lipoproteins; DM: diabetes mellitus.

¹Individuals may be counted more than once if meeting multiple criteria.

In univariate analysis, HIV-related characteristics did not differ between those who were and were not prescribed statins when eligible (Table 3), though individuals receiving statins reported higher ART adherence (95.0% [IQR 90.0%−100%] vs 92.0% [IQR 75.0%−98.0%]). Individuals who were prescribed statins based on ACC/AHA recommendations had significantly lower systolic blood pressure (mmHg) when compared to those who were not prescribed statins (127.5 [IQR 119.0 – 137.0] vs. 133.0 [IQR 121.5–141.5], p=0.001). In contrast, individuals who were not prescribed statins were more likely to report tobacco use when compared to those who were prescribed a statin (30.4% vs 14.6%, p<0.001). Last, African Americans (52.8%) were less likely than Hispanics (58.3%) or White individuals (70.5%) to receive a statin prescription when indicated (p<0.001 for group comparison).

Table 3 –

Baseline Characteristics by Prescription of Statin per ACC/AHA Recommended Therapy

	Statin Prescription No	Statin Prescription Yes	Total	P-value
Number of Subjects	184	302	486
ASCVD Estimator Inputs
Total Cholesterol at Last visitb	181.5 [161.0 – 207.5]	166.0 [143.0 – 196.0]	174.0 [150.0 – 200.0]	<.0001
HDL at Last visitb	43.0 [35.0 – 51.0]	43.0 [35.0 – 53.0]	43.0 [35.0 – 52.0]	0.76
LDL at Last visitb	115.0 [93.5 – 134.5]	97.0 [75.0 – 118.0]	105.0 [80.0 – 125.0]	<.0001
SBP at Last visitb	133.0 [121.5 – 141.5]	127.5 [119.0 – 137.0]	129.0 [120.0 – 139.0]	0.0014
Age at Last visit (years)b	55.8 [51.0 – 60.4]	55.4 [50.3 – 60.9]	55.7 [50.4 – 60.7]	0.85
Anti-hypertensive drug usea	120 (65.2)	202 (66.9)	322 (66.3)	0.71
Smokinga	56 (30.4)	44 (14.6)	100 (20.6)	<.0001
Demographics
Male	176 (95.7)	286 (94.7)	462 (95.1)
Racea				<.0001
White	59 (32.1)	160 (53.0)	219 (45.1)
African American	110 (59.8)	121 (40.1)	231 (47.5)
Hispanic/Other	15 (8.2)	21 (7.0)	36 (7.4)
Age at HIV Diagnosis (years)b	32.4 [27.6 – 38.6]	34.2 [28.6 – 39.1]	33.4 [28.1 – 38.9]	0.24
Age at ART Initiation (years)b	39.3 [34.7 – 43.7]	40.1 [34.9 – 45.0]	39.7 [34.8 – 45.0]	0.53
HIV+/ART related
Use of tenofovir at ART Initiationa,c	40 (22.2)	79 (26.2)	119 (24.7)	0.32
Year of ART Initiation (Calendar year)a,c	1998 [1997 – 2004]	1998 [1997 – 2006]	1998 [1997 – 2005]	0.91
Minimum Self-reported adherence from HIV diagnosis to Last Visit (%)b	92.0 [75.0 – 98.0]	95.0 [90.0 – 100.0]	95.0 [85.0 – 99.0]	<.0001
Laboratory
CD4 at HIV Diagnosis (cells/ul)b	533.0 [343.0 – 679.0]	496.0 [344.0 – 641.0]	506.5 [343.5 – 652.5]	0.36
CD4 at ART Initiation (cells/ul)b,c	353.0 [203.5 – 465.5]	344.0 [223.0 – 480.0]	350.0 [214.0 – 477.0]	0.69
CD4 at Last visit (cells/ul)b,c	654.0 [486.0 – 828.0]	741.0 [581.0 – 956.0]	708.0 [537.0 – 926.0]	0.0006
nadir CD4 (cells/ul)b	280.0 [182.0 – 373.0]	290.0 [192.0 – 374.5]	284.0 [188.0 – 374.0]	0.68
log10(VL) at Last visitb	1.3 [1.3 – 1.3]	1.3 [1.3 – 1.3]	1.3 [1.3 – 1.3]	0.32
VL at Last visit (copies/mL)b				0.05
0–200 (%)	169 (91.8)	290 (96.0)	459 (94.4)
>200+(%)	15 (8.2)	12 (4.0)	27 (5.6)

Notes: Data were presented as

^an (%) or

^bmedian (1^st quartile −3^rd quartile).

^cSubjects with missing values were not included in the computation of percentages and statistics. P-values were calculated by Kruskal-Wallis test and Chi-squared/Fisher’s exact test for continuous and categorical variables, respectively. ASCVD: atherosclerotic cardiovascular disease; HDL: high-density lipoproteins; LDL: low-density lipoproteins; SBP: systolic blood pressure; DM: diabetes mellitus

Our multivariable model (Table 4) was adjusted for total, HDL and LDL cholesterol levels as these both impact and are impacted by statin use. Individuals of African American race (OR 0.48 [95% CI 0.31–0.73]) or Hispanic ethnicity (OR 0.42 [95% CI 0.19–0.94]) were less likely to receive statin prescriptions than White individuals. Individuals with a higher CD4 count (OR 1.12 [95% 1.05–1.20 per 100 cells/uL]) or who received tenofovir disoproxil fumarate at ART initiation (OR 1.66 [95% 1.01–2.74]) were significantly more likely to receive a statin prescription. Sensitivity analysis which excluded individuals younger than 40 (n=12) did not appreciably change our results.

Table 4 –

Unadjusted/Adjusted Odds Ratios for Statin Prescription

	Unadjusted Model				Adjusted Model
	Odds Ratio (95% CI)			P-value	Odds Ratio (95% CI)			P-value
Sex
Female vs. Male	1.22	0.51	2.92	0.6501
Race
African American vs. White	0.43	0.29	0.64	<.0001	0.48	0.31	0.73	0.0008
Hispanic/Other vs. White	0.50	0.24	1.05	0.0661	0.42	0.19	0.94	0.03
Age at last visit (year)	1.00	0.98	1.02	0.9737
HIV Diagnosis era
2000 - Current vs. Prior to 2000	1.14	0.76	1.71	0.5297
ART era
2001 – 2015 vs. 1996–2000	1.22	0.83	1.78	0.3083
On tenofovir disoproxil fumarate at ART Initiation †	1.29	0.83	2.00	0.2605	1.66	1.01	2.74	0.048
ARV before ART Initiation †	0.76	0.52	1.10	0.1488
Time from HIV Diagnosis to ART Initiation (year)	0.99	0.96	1.02	0.5661
AIDS Def. Dx prior to ART Initiation †	0.58	0.30	1.10	0.0969
CD4 at last visit (per 100 cells)	1.11	1.04	1.18	0.0015	1.12	1.05	1.20	0.0015
Log 10 (VL) at last visit (copies/mL)	0.66	0.46	0.96	0.0303
SBP at last visit (per 10 mmHg)	0.84	0.75	0.95	0.0058	0.86	0.75	0.98	0.03
Anti-hypertensive drug use †	1.07	0.72	1.58	0.7383
Smoking status †	0.39	0.25	0.62	<.0001	0.35	0.22	0.58	<.0001

Notes

^*Outcome = Statin adherence (i.e. On Statin as the event)

^**For the adjusted model, Stepwise method was used to select the subset of covariates.

^†For binary covariates, the response ‘No’ was considered as the reference category. SBP: systolic blood pressure

Discussion:

In our cohort of PLWH who have universal access to health care and prescription medication, we found insufficient adherence to statin-focused guidelines, even as our overall percentage (62.1%) is higher than reported by some other cohorts.(23, 24, 30, 31) In addition, we observed racial and ethnic disparities despite open access to care and medications. Given the important role of statins in cardiovascular risk reduction, our study underscores an important area for improvement among all PLWH and racial/ethnic minorities in particular.

The relatively higher adherence to ACC/AHA recommendations of the U.S. Military Health System (MHS) compared with other medical systems may derive from several unique factors. Both active-duty military personnel and retirees receive universal access to care; the former are obligated to receive routine medical care, and the latter may enjoy smoother transitions into routine medical care than is true of PLWH in other cohorts. Prior studies regarding the provision of statin therapy have found access to care to be an important barrier to appropriate preventive care,(32) but our study revealed a gap that may involve additional factors such as physician awareness or time,(33) patient adherence,(34, 35) existing medication count,(36) and/or drug-drug interactions (35) may also impact prescribing practices. Gender may also play a significant factor (37) but our study predominantly involved male subjects and was unable to address this disparity. Further study is needed to understand which of these are most relevant in the MHS and other contexts.

Even with the benefit of universal care, we nevertheless found important racial disparities with respect to statin utilization, as African Americans and Hispanics had more than 50% lower odds of being prescribed a statin despite meeting criteria per ACC/AHA guidelines. A recent paper from our cohort likewise found that African Americans had poorer blood pressure control than White individuals, further demonstrating the impact of race on health outcomes, even in a universal health care system.(38) As African Americans in our cohort experience a higher prevalence of cardiovascular disease, the issue of statin-based primary and secondary prevention is even more urgent.(39) These findings correlate with disparities that are consistently seen in other cohorts that appear to be amplified in PLWH.(40, 41) Given the complexity of these issues and the design of this project, projects beyond the scope of this retrospective descriptive study will be needed to address these findings. Clearly, barriers to optimal care exist even where access to care is not problematic. Further research into the causes of these disparities and more so into meaningful interventions to correct them is essential within both the MHS and health systems in general.(42)

While our analysis was based on the ASCVD risk estimator, newer tools such as the Data Collection on Adverse events of Anti-HIV Drugs (D:A:D) risk calculator have attempted to incorporate HIV-specific risk factors into their recommendations. While the association between any specific drug and subsequent use of a statin can be a reflection of the predominant drugs in use during the period of study, our finding that appropriate statin prescription is associated with the use of tenofovir-based regimens at ART initiation nevertheless demonstrates the importance that HIV-related management can have on cardiovascular disease prevention. Statin therapy may be further complicated by interactions with many individual drugs in antiretroviral regimens,(43) again arguing for a more HIV-focused approach to overall risk assessment. Nevertheless, while these new risk assessment tools are promising, they have yet to be widely accepted in practice.(44) Several studies have compared the D:A:D with other tools, including the Systematic Coronary Risk Evaluation adjusted for national data (SCORE-NL); Framingham CVD Risk Score (FRS); and American College of Cardiology and American Heart Association Pooled Cohort Equations (PCE), and found that D:A:D performed comparably or better than its counterparts.(45, 46) More broadly, attempts to add HIV-specific risk factors to standard formulas have yielded mixed results.(13, 26)

This study had several limitations. As a cross-sectional analysis, a causal relationship between statin use and population characteristics cannot be inferred. In terms of our analysis, we did not attempt to distinguish high- and low- intensity statin therapy, as is recommended in the ACC/AHA guidelines, because interactions with antiretroviral drug regimens can lead to physician practices that vary more widely.(47) Our study assumed than any statin prescription dispensed from a pharmacy was actually taken by subjects, and thus rates of actual patient usage may be lower. Moreover, it is possible that individuals obtained statins outside of the MHS Tricare network, and these prescriptions were not identified through coordinator interview and external record review, though the percentage of these individuals is likely small given this easily available benefit to patients who utilize the MHS pharmacy network. A potential limitation is that the participants’ level of engagement in care (frequency of healthcare visits) was not included in the identification of statin-eligibility or in the statistical models, however this is unlikely to be a significant confounder because the MHS provides beneficiaries with open access to healthcare and medications, HIV-positive active duty have mandatory periodic healthcare visits at least one to two times annually, and the NHS subjects have biannual study visits that are linked to expert HIV care, so subjects have many opportunities for statin requirements to be identified and addressed. Last, without direct queries of physicians or patient subjects, we may be overlooking justifiable reasons for not prescribing a statin, such as patient refusal or previous intolerance. That data is not otherwise readily available.

In conclusion, cardiovascular disease mitigation is becoming an essential aspect of the modern care of PLWH, as morbidity and mortality among PLWH continue to shift from AIDS-related to non-AIDS related causes. Our study has demonstrated important gaps and disparities in risk modification of cardiovascular disease that is available through use of statin-based therapy. Ongoing study and targeted interventions are important to promote improved cardiovascular disease prevention among PLWH both in and out of the military.

Data Availability Statement:

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.