Cotrimoxazole use is associated with lower total and non-high density lipoprotein cholesterol among people with HIV: a cross-sectional study

Abstract

AbstractSection Background

See More

While cotrimoxazole has anti-inflammatory effects among people with Human Immunodeficiency Virus (PWH), its effect on lipid profiles is not well-characterized. We compared the lipid profiles of PWH on cotrimoxazole to those not on it and assessed its association with individual lipid parameters.

AbstractSection Methods

PWH on antiretroviral therapy (ART) were randomly selected from an urban HIV clinic at Kiruddu National Referral Hospital in Kampala, Uganda. Participants were administered an interviewer-administered study questionnaire, underwent anthropometric measurements, and had blood samples evaluated for total cholesterol, low-density lipoprotein cholesterol (LDL-c), high-density lipoprotein cholesterol (HDL-c), and triglycerides. Non-HDL-c was the difference between total cholesterol and HDL-c and was considered elevated if > 4.0 mmol/l. Dyslipidemia was characterized as elevated total cholesterol (> 5.0 mmol/L), high LDL-c (> 4.14 mmol/L), high triglycerides (≥ 1.7 mmol/L), or low HDL-c (< 1.03 mmol/L for males and < 1.29 mmol/L for females. We used linear, logistic and modified Poisson regression analyses to determine if cotrimoxazole use was independently associated with the individual lipid parameters when controlling for age, smoking, dolutegravir, anthropometrics, alcohol use, HIV clinical stage and glycated hemoglobin.

AbstractSection Results

Among 395 PWH enrolled, 100 (25.3%) were on cotrimoxazole prophylaxis for a median (IQR) of 1.0 (0.6–2.5) year, and 364 (92.2%) were on dolutegravir-based ART with a median (IQR) ART duration of 4 (1.8–10.0) years. PWH on cotrimoxazole had significantly lower total cholesterol (4.2 vs. 4.6 mmol/l, p < 0.001), non-HDL-c (3.09 vs. 3.62 mmol/l, p < 0.001), and LDL-c (2.7 vs. 3.0 mmol/l, p = 0.002) compared to those not on the drug. Additionally, a higher proportion of those not on cotrimoxazole had low HDL-c (75.3% vs. 61.6%, p = 0.009) and elevated non-HDL-c (33.3% vs. 11.1%, p < 0.001). Compared to PWH not on cotrimoxazole, a significantly lower proportion of PWH on cotrimoxazole had dyslipidaemia (83.8% (n = 83) vs. 91.1% (n = 265), p = 0.045). After adjusting for confounders, cotrimoxazole use was independently associated with lower total cholesterol levels (adjusted beta coefficient (β) = -0.18, 95% confidence interval (CI) -0.35 to -0.01), p = 0.040), non-HDL-c (β = -0.34, 95% CI -0.58 to -0.11, p = 0.005) and lower odds for elevated non-HDL-c (adjusted prevalence ratio = 0.45, 95% CI 0.26–0.77, p = 0.003).

AbstractSection Conclusion

Cotrimoxazole use was independently associated with lower total cholesterol and non-HDL-c levels. These findings suggest a potential lipid-lowering effect of cotrimoxazole, warranting further longitudinal investigation into its potential long-term cardioprotective benefits among PWH.

Background

Globally, nearly 40 million people are living with Human Immunodeficiency Virus (HIV), with two-thirds residing in sub-Saharan Africa [1]. The introduction of antiretroviral therapy (ART) has significantly improved life expectancy among people with HIV (PWH), yet they continue to experience a higher burden of comorbidities and fewer years of disease-free survival compared to HIV-negative individuals [2, 3]. Non-communicable diseases (NCDs), particularly cardiovascular disease (CVD), are emerging as a leading cause of morbidity and mortality among the aging population of PWH [4, 5]. The pathogenesis of CVD in PWH is multifactorial, driven by chronic immune activation, persistent inflammation, dyslipidemia, and traditional risk factors such as smoking and hypertension, which increase the risk of atherosclerosis, thrombosis, myocardial infarction, heart failure, and sudden cardiac death [6]. Given the inflammatory and metabolic derangements in PWH, identifying cost-effective and widely accessible strategies to reduce CVD risk is crucial. Statin therapy with Pitavastatin was recently shown to reduce cholesterol levels and CVD events among PWH but its availability and affordability in low-income settings is a challenge [7].

Cotrimoxazole, a fixed-dose combination of sulfamethoxazole and trimethoprim, has been historically recommended for prophylaxis against bacterial infections, diarrheal diseases, malaria, toxoplasmosis, and Pneumocystis jirovecii pneumonia among PWH [8, 9]. Emerging evidence suggests that cotrimoxazole may exert anti-inflammatory effects in PWH by reducing systemic immune activation, a major driver of morbidity and mortality [10]. Studies in both adults and children have demonstrated that continued cotrimoxazole use lowers levels of inflammatory markers, particularly C-reactive protein (CRP) and interleukin-6 (IL-6), which are strongly linked to lipid abnormalities and CVD risk [11, 12]. The proposed mechanisms underlying the anti-inflammatory effects of cotrimoxazole include reductions in gut bacterial translocation, reflected by reduction in levels of lipopolysaccharide (LPS) and soluble CD14 (sCD14), markers of microbial translocation and monocyte activation, respectively [10]. Therefore, reducing gut bacterial translocation results in lower systemic inflammation, which would otherwise impair reverse cholesterol transport and cholesterol flux through the liver to the bile and feces [13]. The study by McGillicuddy and colleagues demonstrates that inflammation, induced via endotoxemia, downregulates key hepatic cholesterol transporters such as ABCG5, ABCG8, and ABCB11, which are essential for exporting cholesterol into bile. This suppression reduces hepatic secretion of cholesterol and bile acids, leading to markedly lower cholesterol content in bile and feces. Even low-dose inflammation was sufficient to decrease biliary transporter expression and cholesterol excretion without altering plasma cholesterol or hepatic cholesterol uptake, highlighting a direct hepatic effect [13]. Additionally, animal studies suggest that cotrimoxazole may directly modulate lipid metabolism by inhibiting 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) reductase, a key enzyme in cholesterol biosynthesis, and enhancing lecithin-cholesterol acyltransferase (LCAT) activity, which promotes the conversion of free cholesterol into cholesteryl esters [14, 15]. These effects are hypothesized to contribute to lower plasma cholesterol and triglyceride levels and increased high-density lipoprotein cholesterol (HDL-c) levels [14,15,16].

Despite these potential lipid-lowering effects, data on cotrimoxazole’s impact on lipid metabolism in PWH remains limited and conflicting. An analysis of electronic health records at the University of California, San Francisco, found that PWH prescribed trimethoprim—a component of cotrimoxazole—had significantly higher levels of triglycerides, low-density lipoprotein cholesterol (LDL-c), and total cholesterol compared to matched PWH not on the drug [17]. However, a small, randomized controlled trial reported no significant differences in blood cholesterol levels between individuals receiving trimethoprim and those in the placebo group, highlighting inconsistencies in the available evidence [18].

Recent guidelines advise cotrimoxazole prophylaxis to be discontinued if the individual is clinically stable, has achieved immune recovery (CD4 > 350 cells/mm³), and has suppressed viral load on ART [19]. Therefore, if cotrimoxazole is found to lower blood lipids, it could provide a compelling rationale for its prolonged use in PWH who have achieved immune reconstitution. Given that cotrimoxazole is inexpensive, widely available, and already integrated into HIV care programs, its potential role in mitigating CVD risk would represent an accessible and scalable intervention for PWH in resource-limited settings. This study aimed to compare lipid profiles between PWH on cotrimoxazole prophylaxis and those not receiving the drug and to determine the association of cotrimoxazole use with the individual lipid parameters.

Materials and methods

Study design, setting, and participants

This study was a cross-sectional secondary analysis utilizing data from a prior investigation conducted at Kiruddu National Referral Hospital (KNRH) in Kampala, Uganda [20]. The original study aimed to compare cardiometabolic characteristics among PWH with and without a history of tuberculosis (TB) – including blood lipids as a primary outcome. The HIV clinic at KNRH provides ongoing care to over 2,500 PWH. According to national treatment guidelines, all adult PWH attending this clinic are initiated on standard first-line ART predominantly consisting of tenofovir, lamivudine and dolutegravir [21]. This analysis focused on PWH aged 18 years and older who were previously enrolled in the primary study and had recorded lipid profile data. In the original study, participants were randomly sampled from the clinic’s HIV care database and invited to take part following an overnight fast. There were no participants on lipid lowering agents in the primary study. All data and measurements were conducted between October 2023 to May 2024.

Study variables and data collection

In original study, a structured questionnaire was employed to collect socio-demographic and clinical information, including any history of cigarette smoking, alcohol use, and family history of CVD risk factors. Smoking intensity was quantified in pack-years, while alcohol consumption was evaluated using the Cut-down, Anger, Guilt and Eye opener (CAGE) questionnaire, with a score of two or more indicating hazardous alcohol use [22]. Clinical data on ART regimens, the WHO HIV clinical stage [23], viral load status, prior opportunistic infections, CD4 counts, and cotrimoxazole usage and its duration were extracted from the KNRH HIV treatment database. Cotrimoxazole use was defined as documented prescription and self-reported use at the time of study enrollment. The weight, height, mid-upper arm, neck, waist, and hip circumference were obtained using a weighing scale (Seca 760^®), a stadiometer (Seca 213^®), and a measuring tape respectively. Neck circumference was measured using a flexible, non-elastic measuring tape with the participant in a standing position and the head in the Frankfort horizontal plane. The measurement was taken at the level of the thyroid cartilage, ensuring the tape was positioned horizontally and did not compress the skin. Participants were instructed to breathe normally, and the measurement was recorded at the end of a gentle exhalation to minimize respiratory variation. Body mass index (BMI) was computed as weight (kg) divided by height² (m²). The waist-to-height ratio was calculated as a ratio of the waist circumference to the height of the individual. Office blood pressure (BP) was measured twice at 20-minute intervals using an Omron^® Hem 7120 digital BP machine, with the average recorded. Blood samples were analyzed for fasting plasma glucose (FPG), glycated hemoglobin (HbA1c), and serum uric acid. Lipid profiles, including total cholesterol, LDL-c, HDL-c, and triglycerides, were assessed using the Cobas^® 6000 analyzer (Roche Diagnostics, USA).

Study outcomes and operational definitions

The primary outcome was the comparison of lipid parameters, including total cholesterol, LDL-c, HDL-c, non- HDL-c, triglycerides, very low density lipoprotein cholesterol (VLDLc), and atherogenic index of plasma (AIP), between PWH on and off cotrimoxazole. Secondary outcomes involved determining and comparing the prevalence of dyslipidemia, elevated LDL-c, HDL-c, non-HDL-c, triglycerides and AIP among cotrimoxazole users versus non-users. Dyslipidemia was characterized as elevated total cholesterol (> 5.0 mmol/L), high LDL-c (> 4.14 mmol/L), high triglycerides (≥ 1.7 mmol/L), or low HDL-c (< 1.03 mmol/L for males and < 1.29 mmol/L for females) [24]. Non-HDL-c was calculated as total cholesterol – HDL-c and was considered elevated if > 4.0 mmol/l [25]. The AIP was calculated as the log (triglycerides/HDL-c) and was considered elevated if ≥ 0.11 [26]. Very low density lipoprotein was estimated as triglyceride level/5 [27]. Diabetes mellitus was defined as an FPG level of ≥ 7.0 mmol/L, HbA1c of ≥ 6.5%, or documented use of antidiabetic medication [28]. Hypertension was defined as a systolic BP ≥ 140 mmHg, diastolic BP of ≥ 90 mmHg, or current use of antihypertensive medication [29]. Overweight was classified as a BMI between 25.0 and 29.9 kg/m², while obesity was defined as a BMI of ≥ 30.0 kg/m². Central obesity was determined based on a waist circumference of ≥ 102 cm in males and ≥ 88 cm in females or a waist-to-hip ratio of ≥ 0.90 in males and ≥ 0.85 in females [30]. Elevated BP was a systolic BP ≥ 140 mmHg and/or diastolic BP ≥ 90 mmHg.

Post hoc study power calculation

Considering a two-sided 95% confidence interval and the observed mean difference of −0.4mmol/l in the total cholesterol between PWH on cotrimoxazole and those not on cotrimoxazole, the study power to detect this difference was 93.28%. Similarly, for a mean difference of −0.3 mmol/l in the LDL-c, the study power to detect this difference was 82.14%.

Statistical analysis

Data analysis was performed using Stata version 18.0 (StataCorp LLC, College Station, Texas, USA). Categorical variables were compared between cotrimoxazole users and non-users using Pearson’s chi-square test or Fisher’s exact test where appropriate. Continuous variables were analyzed using the independent t-test or Mann-Whitney U test, depending on the distribution. To examine the independent relationship between cotrimoxazole use and lipid levels, as continous variables, multivariable linear regression models were developed for total cholesterol, non-HDL-c, HDL-c, LDL-c, AIP, and triglycerides. We adjusted for potential confounders such as age, sex, BMI, smoking status, alcohol use, ART regimen (dolutegravir), and CD4 count. Variables included in the final model were those that showed a p < 0.05 at bivariable analysis. A p-value of less than 0.05 was considered statistically significant. We further constructed modified Poisson regression models with robust standard errors for factors associated with dyslipidemia, elevated cholesterol, non-HDL-c, AIP and low HDL-c because of the high prevalence of these outcomes. For elevated LDL-c, we constructed logistic regression models because of the low prevalence of high LDL-c.

Results

Among the 396 PWH in the prior primary study, one had no lipid measurement and was not included in this analysis.

Of 395 PWH enrolled, more than half (53.4%, n = 211) were female, and the median (IQR) age was 41.0 (32.0–50.0) years. A total of 308 (94.8%) PWH were virally suppressed (viral load < 1,000 copies/ml), 88 (22.3%) were overweight, 64 (16.2%) were obese, and 196 (49.9%) had central obesity. Further, 127 (32.2%) had hypertension, 54 (13.7%) had diabetes mellitus, and 348 (89.2%) had dyslipidemia. The blood pressure was elevated in 24 (24.0%) of PWH on cotrimoxazole vs. 102 (25.9%) of those not on cotrimoxazole, p = 0.62. Other characteristics are shown in Table 1. Overall, 25.3% (n = 100) were on cotrimoxazole prophylaxis for a median of 1.0 (0.6–2.5) years.

Table 1 Characteristics of study participants

Comparison of lipid profiles of PWH with and without cotrimoxazole use

PWH on cotrimoxazole had significantly lower total cholesterol (4.2 vs. 4.6 mmol/l, p < 0.001), non-HDL-c (3.09 vs. 3.62 mmol/l, p < 0.001), and LDL-c (2.7 vs. 3.0 mmol/l, p = 0.002) compared to those not on the drug. Additionally, a higher proportion of those not on cotrimoxazole had low HDL-c (75.3% vs. 61.6%, p = 0.009) and elevated non-HDL-c (33.3% vs. 11.1%, p < 0.001). Overall, dyslipidaemia was less prevalent among PWH on cotrimoxazole than those not on the drug (83 (83.8%) vs. 265 (91.1%), p = 0.045) (Table 2). Participants who had been on cotrimoxazole for at least one year were less likely to have low HDL-c (52.9% vs. 70.5%, p = 0.081), elevated total cholesterol (19.6% vs. 25.0%, p = 0.528), or elevated non-HDL-c (9.8% vs. 13.6%, p = 0.560) compared to those who were on the drug for less than a year but the differences were not statistically significant.

Table 2 Comparison of lipid parameters between PWH on currentcotrimoxazole prophylaxis and those who are not

Association between cotrimoxazole use and total cholesterol and LDL-c at linear regression

After adjusting for confounders, cotrimoxazole use was independently associated with lower total cholesterol levels (adjusted beta coefficient (β) = −0.18, 95% confidence interval (CI) −0.35 to −0.01), p = 0.040), non-HDL-c (β = −0.34, 95% CI −0.58 to −0.11, p = 0.005) and lower odds for elevated non-HDL-c (adjusted prevalence ratio = 0.45, 95% CI 0.26–0.77, p = 0.003) (Tables 3, 4 and 5). Additional analyses for the association of cotrimoxazole use and other lipid parameters are shown in Supplementary Tables 1–6.

Table 3 Linear regression analysis for correlates of serum total cholesterol levels among people with HIV

Table 4 Linear regression analysis for correlates of serum non-HDL-c levels among people with HIV

Table 5 Modified poisson regression analysis for factors associated with elevated serum non-HDL-c levels among people with HIV

Discussion

The study objective was to compare lipid parameters between PWH with and without cotrimoxazole use and to assess if cotrimoxazole use is independently associated with lipid parameters. We found that cotrimoxazole prophylaxis was independently associated with modest but statistically significant reductions in total cholesterol and non-HDL-c, as well as a 55% lower prevalence of elevated non-HDL-c, highlighting its potential lipid-modulating effects. These findings suggest that cotrimoxazole prophylaxis, beyond its known antimicrobial and anti-inflammatory properties [8], may contribute to improved total cholesterol in PWH.

The multivariable analysis indicates that PWH on cotrimoxazole had total cholesterol and non-HDL-c levels that were 0.18 mmol/l (7 mg/dl) and 0.34 mmol/l (13 mg/dl) lower, on average, compared to those not taking the medication, respectively. This represents a respective 3.9% and 11% lower total cholesterol and HDL-c level compared to those not taking cotrimoxazole. While this appears to be a modest reduction, it is important to note that a 1% reduction in total cholesterol could potentially reduce the incidence of coronary heart disease (CHD) by 2% [31]. Similarly, a 1 mmol/l reduction in non-HDL-c levels results into a 20% risk reduction for major adverse cardiovascular events – similar to lowering LDL-c [32]. Moreover, PWH on cotrimoxazole had lower LDL-c and less frequent low HDL-c, both of which are important in reducing the risk for all-cause mortality, CHD-related mortality, and any CHD event [33].

It was interesting to observe that cotrimoxazole use was associated with significantly lower non-HDL-c at multivariable analysis, but not lower LDL-c. Non-HDL-c reflects the total burden of all atherogenic, apoB-containing lipoproteins—including intermediate-density lipoprotein (IDL), lipoprotein(a), and remnant particles—in addition to LDL-c and VLDL-c [34]. Therefore, even modest decreases in these less abundant fractions can drive down non-HDL-c without producing statistically significant shifts in isolated LDL-c or calculated VLDL-c as was observed in our study. Moreover, LDL-c and VLDL-c are typically derived by the Friedewald formula (LDL-c = TC – HDL-c – TG/5; VLDL-c = TG/5), which may obscure small but biologically relevant changes in particle composition when triglyceride levels are within the normal or mildly elevated range [27]. In contrast, non-HDL-c is calculated directly as total cholesterol minus HDL-c and thus more sensitively captures incremental declines across the entire spectrum of atherogenic lipoproteins. Taken together, these methodological and compositional factors likely explain why cotrimoxazole prophylaxis produced a significant reduction in non-HDL-c even though LDL-c and VLDL-c remained unchanged. Moreover, in exploratory analyses, we observed a trend toward greater reductions in total and non–HDL-c with longer duration of cotrimoxazole prophylaxis, suggesting a potential cumulative lipid-modulating effect over time although the sample size for this analysis was small. If cotrimoxazole indeed modulates lipid metabolism, it could have profound clinical implications for PWH. Given the increasing burden of CVD in this population, strategies to mitigate CVD risk factors such as dyslipidemia are critical [35].

The association of cotrimoxazole with lower lipid levels may be attributed to its potential direct influence on cholesterol metabolism. Animal studies have suggested that cotrimoxazole inhibits HMG-CoA reductase, a key enzyme in cholesterol biosynthesis, and increases LCAT activity, which enhances the conversion of free cholesterol into cholesteryl esters [14, 15]. This mechanism would be expected to lower plasma cholesterol and triglyceride levels while increasing HDL-c, which is partially supported by our findings. However, we did not observe significant differences in triglyceride levels between the two groups, warranting further investigation into whether cotrimoxazole affects triglyceride metabolism differently from cholesterol metabolism. The other mechanism by which cotrimoxazole lowers blood lipids is through its anti-inflammatory effects. Chronic inflammation contributes to dyslipidemia by impairing reverse cholesterol transport, increasing lipoprotein oxidation, and promoting endothelial dysfunction [13, 36]. Cotrimoxazole reduces systemic immune activation by lowering CRP and IL-6, both of which are linked to cardiovascular disease risk [11, 12]. Additionally, it decreases microbial translocation markers such as LPS and sCD14, further mitigating inflammation-driven metabolic disturbances [10].

Similar to our findings, Yazie and colleagues found cotrimoxazole use among PWH in Ethiopia to be associated with lower LDL-c levels at bivariable analysis [37]. Further, the prevalence of dyslipidemia and hypercholesterolemia was found to be significantly lower among people exposed to trimethoprim, a component of cotrimoxazole, albeit among elderly HIV negative Chinese individuals [38]. While our results contrast with findings from an electronic health record analysis at the University of California, San Francisco, which reported higher triglyceride, LDL-c, and total cholesterol levels in PWH prescribed trimethoprim, a component of cotrimoxazole, the authors do not provide sufficient study population characteristics to compare to ours [17]. Trimethoprim has been hypothesized to inhibit OCT1, an organic cation transporter involved in hepatic thiamine uptake, which may disrupt lipid metabolism [17]. However, this effect remains speculative, and a randomized controlled trial among HIV negative individuals reported no significant differences in total cholesterol levels between individuals receiving trimethoprim and those on placebo [18].

Study strenghts and limitations

Several limitations of our study must be acknowledged. First, the cross-sectional nature of our study prevents us from establishing causality between cotrimoxazole use and lipid levels. Baseline lipid levels (before initiating cotrimoxazole) would be desirable to establishing the effect of cotrimoxazole. However, these were unavailable. Second, we did not assess dietary patterns, physical activity, or genetic predisposition to dyslipidemia, which could influence lipid profiles. Lastly, our study was conducted in an urban Ugandan population, and the generalizability of our findings to other settings requires further validation. The study possesses several strengths. It rigorously adjusted for a wide range of potential confounding factors, including age, sex, BMI, smoking status, alcohol use, ART regimen, and CD4 count, when examining the association between cotrimoxazole use and lipid levels. Furthermore, post hoc power calculations demonstrated that the study had high statistical power (93.28% for total cholesterol and 82.14% for LDL-c) to detect the observed differences in key lipid parameters, lending robustness to the findings. The study also benefited from a relatively large sample size of 395 PWH, enhancing the reliability of its conclusions. Importantly, the association between cotrimoxazole and lower lipid levels was found to be independent of antiretroviral therapy and viral suppression, as a high percentage of the participants were virally suppressed and had been on ART for a median of four years.

Conclusions

In this cross sectional study, cotrimoxazole prophylaxis was independently associated with a respective 3.9% and 11% lower total cholesterol and HDL-c level. These novel findings suggest that beyond its well-established role in infection prophylaxis, cotrimoxazole may confer ancillary cardiovascular benefits by lowering atherogenic lipoprotein levels. Given the low cost and widespread availability of cotrimoxazole in resource-limited settings, its potential lipid-modulating effects merit further exploration. Prospective studies and randomized trials are needed to confirm these observations, elucidate underlying mechanisms, and determine whether long-term cotrimoxazole use can meaningfully reduce cardiovascular events among PWH.