Search

Despite its potential to improve metabolic health outcomes, longitudinal physical activity (PA) patterns and their association with cardiometabolic disease among people living with HIV (PLWH) have not been well characterized. We investigated this relationship among PLWH in the Centers for AIDS Research Network of Integrated Clinical Systems with at least one PA self-report between 2008 and 2015. The 4-item Lipid Research Clinics PA instrument was used to categorize habitual PA levels as: Very Low, Low, Moderate, or High. We analyzed demographic differences in PA patterns. Multivariable generalized estimating equation regression models were fit to assess longitudinal associations of PA with blood pressure, lipid, and glucose levels. Logistic regression modeling was used to assess the odds of being diagnosed with obesity, cardiovascular disease (CVD), cerebrovascular disease, hypertension, diabetes, or multimorbidity. A total of 40,462 unique PA assessments were provided by 11,719 participants. Only 13% of PLWH reported High PA, while 68% reported Very Low/Low PA at baseline and did not increase PA levels during the study period. Compared to those reporting High PA, participants with Very Low PA had almost 2-fold increased risk for CVD. Very Low PA was also associated with several risk factors associated with CVD, most notably elevated triglycerides (odds ratio 25.4), obesity (odds ratio 1.9), hypertension (odds ratio 1.4), and diabetes (odds ratio 2.3; all p < 0.01). Low levels of PA over time among PLWH are associated with increased cardiometabolic disease risk.

Background: To obtain optimal health outcomes, persons living with human immunodeficiency virus (HIV) must be retained in clinical care. We examined the relationships between 4 possible combinations of 2 separate retention measures (missed visits and the Institute of Medicine [IOM] indicator) and all-cause mortality. Methods: The sample included 4162 antiretroviral therapy (ART)–naive patients who started ART between January 2000 and July 2010 at any of 5 US sites of the Center for AIDS Research Network of Integrated Clinical Systems. The independent variable of interest was retention, captured over the 12-month period after the initiation of ART. The study outcome, all-cause mortality 1 year after ART initiation, was determined by querying the Social Security Death Index or the National Death Index. We evaluated the associations of the 4 categories of retention with all-cause mortality, using the Cox proportional hazards models. Results: Ten percent of patients did not meet retention standards for either measure (hazard ratio [HR], 2.26; 95% confidence interval [CI], 1.59–3.21). Patients retained by the IOM but not the missed-visits measure (42%) had a higher HR for mortality (1.72; 95% CI, 1.33–2.21) than patients retained by both measures (41%). Patients retained by the missed-visits but not the IOM measure (6%) had the same mortality hazards as patients retained by both measures (HR, 1.01; 95% CI, .54–1.87). Conclusions: Missed visits within the first 12 months of ART initiation are a major risk factor for subsequent death. Incorporating missed visits in clinical and public health retention and viral suppression programming is advised.

Background: Guidelines for initial antiretroviral treatment (ART) regimens have evolved, with integrase strand transfer inhibitors (INSTIs) increasingly prominent. Research on virologic failure (VF) with INSTI therapy is predominantly from clinical trials not care settings, especially for recently approved medications including dolutegravir. We compared outcomes among people living with HIV (PLWH) who initiated recommended regimens in clinical care across the United States. Setting: We examined 2 groups of PLWH at 8 clinics who initiated ART regimens (August 1, 2013-March 31, 2017): those ART treatment-naive at initiation, and those treatment-experienced. Methods: The outcome in this longitudinal cohort study was VF, defined as a viral load of ≥400 copies/mL ≥6 months after ART initiation. We examined the proportion of individuals who remained on, switched, or discontinued the regimen. Associations between regimens and outcomes were examined with adjusted Cox proportional hazards models. Results: Among 5177 PLWH, a lower proportion experienced VF on dolutegravir- versus other INSTI- or darunavir-based regimens for previously treatment-naive (7% vs. 12% vs. 28%) and treatment-experienced PLWH (6% vs. 10% vs. 21%). In adjusted analyses, hazard ratios were similar across regimens for the combined outcome of regimen discontinuation or treatment switch. The hazard ratios for VF comparing dolutegravir- to darunavir-based regimens was 0.30 (95% CI: 0.2 to 0.6) among previously treatment-naive PLWH and was 0.60 (95% CI: 0.4 to 0.8) among treatment-experienced PLWH. Conclusions: The proportion of previously treatment-naive PLWH remaining on recommended ART regimens did not differ by regimen. The likelihood of VF was lower with dolutegravir- than darunavir-based regimens for previously treatment-naive and treatment-experienced PLWH.

Purpose: Substance use is linked with poor outcomes among people living with HIV (PLWH) and is associated with mental health disorders. This analysis examines the impact of decreasing substance use, even without abstinence, on depressive symptoms among PLWH. Methods: Data are from PLWH enrolled in the Centers for AIDS Research Network of Integrated Clinical Sites cohort. Participants completed longitudinal assessments of substance use (modified ASSIST) and depressive symptoms (PHQ-9). Changes in substance use frequency were categorized as abstinence, reduced use, and nondecreasing use. Adjusted linear mixed models with time-updated change in substance use frequency and depressive symptom scores were used to examine associations between changes in the use of individual substances and depressive symptoms. Analyses were repeated using joint longitudinal survival models to examine associations with a high (PHQ-9 ≥10) score. Results: Among 9905 PLWH, 728 used cocaine/crack, 1016 used amphetamine-type substances (ATS), 290 used illicit opiates, and 3277 used marijuana at baseline. Changes in ATS use were associated with the greatest improvements in depressive symptoms: stopping ATS led to a mean decrease of PHQ-9 by 2.2 points (95% CI: 1.8 to 2.7) and a 61% lower odds of PHQ-9 score ≥10 (95% CI: 0.30 to 0.52), and decreasing ATS use led to a mean decrease of 1.7 points (95% CI: 1.2 to 2.3) and a 62% lower odds of PHQ-9 score ≥10 (95% CI: 0.25 to 0.56). Stopping and reducing marijuana and stopping cocaine/crack use were also associated with improvement in depressive symptoms. Conclusions: We demonstrated that both substance use reduction and abstinence are associated with improvements in depressive symptoms over time.

Background: Efavirenz has been a mainstay of antiretroviral therapy (ART) for over 15 years in the US. Its association with neuropsychiatric side effects may influence clinical prescribing and management. Methods: We included HIV-infected adults enrolled in care at seven sites across the US, who initiated combination ART between 1999 and 2015. We examined the proportion initiating and continuing on efavirenz, overall and by mental health status. Log binomial and Cox models were used to estimate associations between mental health, clinical and sociodemographic characteristics and initiating or switching from efavirenz as first-line ART. Results: Of the 8,230 participants included, 3,710 (45%) initiated efavirenz. In multivariable analyses, prior mono- or dual-ART, ART initiation after 2006, being female, intravenous drug use, antidepressant prescription, previous mental health diagnosis and baseline CD4+ T-cell count >350 cells/mm3 were inversely associated with initiating efavirenz. Participants initiating efavirenz had a faster time to a regimen switch, compared with those initiating an efavirenz-free regimen (P-value

Cause-specific mortality is an important outcome in studies of interventions to improve survival, yet causes of death can be misclassified. Here, we present an approach to performing sensitivity analyses formisclassification of cause of death in the parametric g-formula. The g-formula is a useful method to estimate effects of interventions in epidemiologic research because it appropriately accounts for time-varying confounding affected by prior treatment and can estimate risk under dynamic treatment plans.We illustrate our approach using an example comparing acquired immune deficiency syndrome (AIDS)-related mortality under immediate and delayed treatment strategies in a cohort of therapy-naive adults entering care for human immunodeficiency virus infection in the United States. In the standard g-formula approach, 10-year risk of AIDSrelatedmortality under delayed treatment was 1.73 (95% CI: 1.17, 2.54) times the risk under immediate treatment. In a sensitivity analysis assuming that AIDS-related death was measured with sensitivity of 95% and specificity of 90%, the 10-year risk ratio comparing AIDS-related mortality between treatment plans was 1.89 (95% CI: 1.13, 3.14). When sensitivity and specificity are unknown, this approach can be used to estimate the effects of dynamic treatment plans under a range of plausible values of sensitivity and specificity of the recorded event type.

Reducing racial/ethnic disparities in human immunodeficiency virus (HIV) disease is a high priority. Reductions in HIV racial/ethnic disparities can potentially be achieved by intervening on important intermediate factors. The potential population impact of intervening on intermediates can be evaluated using observational data when certain conditions are met. However, using standard stratification-based approaches commonly employed in the observational HIV literature to estimate the potential population impact in this setting may yield results that do not accurately estimate quantities of interest. Here we describe a useful conceptual and methodological framework for using observational data to appropriately evaluate the impact on HIV racial/ethnic disparities of interventions. This framework reframes relevant scientific questions in terms of a controlled direct effect and estimates a corresponding proportion eliminated. We review methods and conditions sufficient for accurate estimation within the proposed framework. We use the framework to analyze data on 2,329 participants in the CFAR [Centers for AIDS Research] Network of Integrated Clinical Systems (2008-2014) to evaluate the potential impact of universal prescription of and ≥95% adherence to antiretroviral therapy on racial disparities in HIV virological suppression. We encourage the use of the described framework to appropriately evaluate the potential impact of targeted interventions in addressing HIV racial/ethnic disparities using observational data.

Background inform policy and practice for broad populations. The average treatment effect (ATE) for a target population, however, may be different from the ATE observed in a trial if there are effect modifiers whose distribution in the target population is different that from that in the trial. Methods exist to use trial data to estimate the target population ATE, provided the distributions of treatment effect modifiers are observed in both the trial and target population—an assumption that may not hold in practice. Methods The proposed sensitivity analyses address the situation where a treatment effect modifier is observed in the trial but not the target population. These methods are based on an outcome model or the combination of such a model and weighting adjustment for observed differences between the trial sample and target population. They accommodate several types of outcome models: linear models (including single time outcome and pre- and post-treatment outcomes) for additive effects, and models with log or logit link for multiplicative effects. We clarify the methods’ assumptions and provide detailed implementation instructions. Illustration We illustrate the methods using an example generalizing the effects of an HIV treatment regimen from a randomized trial to a relevant target population. Conclusion These methods allow researchers and decision-makers to have more appropriate confidence when drawing conclusions about target population effects.

Marginal structural models for time-fixed treatments fit using inverse-probability weighted estimating equations are increasingly popular. Nonetheless, the resulting effect estimates are subject to finite-sample bias when data are sparse, as is typical for large-sample procedures. Here we propose a semi-Bayes estimation approach which penalizes or shrinks the estimated model parameters to improve finite-sample performance. This approach uses simple symmetric data-augmentation priors. Limited simulation experiments indicate that the proposed approach reduces finite-sample bias and improves confidence-interval coverage when the true values lie within the central “hill” of the prior distribution. We illustrate the approach with data from a nonexperimental study of HIV treatments.

Objective: To explore the effectiveness of raltegravir-based antiretroviral therapy (ART) on treatment response among ART-naive patients seeking routine clinical care. Design: Cohort study of adults enrolled in HIV care in the United States. Methods: We compared virologic suppression and CD4 + cell count recovery over a 2.5 year period after initiation of an ART regimen containing raltegravir or efavirenz using observational data from a US clinical cohort, generalized to the US population of people with diagnosed HIV. We accounted for nonrandom treatment assignment, informative censoring, and nonrandom selection from the US target population using inverse probability weights. Results: Of the 2843 patients included in the study, 2476 initiated the efavirenz-containing regimen and 367 initiated the raltegravir-containing regimen. In the weighted intent-To-Treat analysis, patients spent an average of 74 (95% confidence interval: 41, 106) additional days alive with a suppressed viral load on the raltegravir regimen than on the efavirenz regimen over the 2.5-year study period. CD4 + cell count recovery was also superior under the raltegravir regimen. Conclusion: Patients receiving raltegravir spent more time alive and suppressed than patients receiving efavirenz, but the probability of viral suppression by 2.5 years after treatment was similar between groups. Optimizing the amount of time spent in a state of viral suppression is important to improve survival among people living with HIV and to reduce onward transmission.

Heavy drinking is prevalent among people living with HIV. Studies use tools like patient-reported outcomes (PROs) to quantify alcohol use in a detailed, timely manner. However, if alcohol misuse influences PRO completion, selection bias may result. Our study included 14,145 adult HIV patients (133,036 visits) from CNICS who were eligible to complete PROs at an HIV primary care visit. We compared PRO completion proportions between patients with and without a clinical diagnosis of at-risk alcohol use in the prior year. We accounted for confounding by baseline and visit-specific covariates. PROs were completed at 20.8% of assessed visits. The adjusted difference in PRO completion proportions was -3.2% (95% CI -5.6 to -0.8%). The small association between receipt of an at-risk alcohol use diagnosis and decreased PRO completion suggests there could be modest selection bias in studies using the PRO alcohol measure.

Increased scientific attention given to cannabis in the United States has particular relevance for its domestic HIV care population, given that evidence exists for both cannabis as a therapeutic agent and cannabis use disorder (CUD) as a barrier to antiretroviral medication adherence. It is critical to identify relative risk for CUD among demographic subgroups of HIV patients, as this will inform detection and intervention efforts. A Center For AIDS Research Network of Integrated Clinical Systems cohort (N = 10,652) of HIV-positive adults linked to care at seven United State sites was examined for this purpose. Based on a patient-report instrument with validated diagnostic threshold for CUD, the prevalence of recent cannabis use and corresponding conditional probabilities for CUD were calculated for the aggregate sample and demographic subgroups. Generalized estimating equations then tested models directly examining patient demographic indices as predictors of CUD, while controlling for history and geography. Conditional probability of CUD among cannabis-using patients was 49%, with the highest conditional probabilities among demographic subgroups of young adults and those with non-specified sexual orientation (67–69%) and the lowest conditional probability among females and those 50+ years of age (42% apiece). Similarly, youthful age and male gender emerged as robust multivariate model predictors of CUD. In the context of increasingly lenient policies for use of cannabis as a therapeutic agent for chronic conditions like HIV/AIDS, current study findings offer needed direction in terms of specifying targeted patient groups in HIV care on whom resources for enhanced surveillance and intervention efforts will be most impactful.

Background. The long-term effectiveness of human immunodeficiency virus (HIV) treatments containing integrase inhibitors is unknown. Methods. We use observational data from the Centers for AIDS Research Network of Integrated Clinical Systems and the Centers for Disease Control and Prevention to estimate 4-year risk of AIDS and all-cause mortality among 415 patients starting a raltegravir regimen compared to 2646 starting an efavirenz regimen (both regimens include emtricitabine and tenofovir disoproxil fumarate). We account for confounding and selection bias as well as generalizability by standardization for measured variables, and present both observational intent-to-treat and per-protocol estimates. Results. At treatment initiation, 12% of patients were female, 36% black, 13% Hispanic; median age was 37 years, CD4 count 321 cells/µL, and viral load 4.5 log10 copies/mL. Two hundred thirty-five patients incurred an AIDS-defining illness or died, and 741 patients left follow-up. After accounting for measured differences, the 4-year risk was similar among those starting both regimens (ie, intent-to treat hazard ratio [HR], 0.96 [95% confidence interval {CI}, .63–1.45]; risk difference, −0.9 [95% CI, −4.5 to 2.7]), as well as among those remaining on regimens (ie, per-protocol HR, 0.95 [95% CI, .59–1.54]; risk difference, −0.5 [95% CI, −3.8 to 2.9]). Conclusions. Raltegravir and efavirenz-based initial antiretroviral therapy have similar 4-year clinical effects. Vigilance regarding longer-term comparative effectiveness of HIV regimens using observational data is needed because large-scale experimental data are not forthcoming.

Background: Evidence about the effect of initiating efavirenz-containing combination antiretroviral therapy (ART) as the first-line therapy on suicidal thoughts remains conflicting. Methods: Using data from a cohort of HIV-infected adults enrolled in routine care across 5 sites in the United States, we included participants with a baseline patient-reported outcome measure and detectable viral load who initiated ART between 2011 and 2014. Participants were followed until the earliest of the following: first suicidal thoughts, discontinuation of initial ART regimen, death, loss to care (>12 months with no HIV appointments), or administrative censoring (2014-2015). Suicidal thoughts were measured using a Patient Health Questionnaire-9 item. We used weighted marginal structural Cox models to estimate the effect of initiating efavirenz-containing ART, versus efavirenz-free ART, on the hazard of active or passive suicidal thoughts after ART initiation, accounting for confounding by channeling bias. Results: Overall, 597 participants were followed for a median of 19 months (13, 132 total person-months); 147 (25%) initiated efavirenz-containing ART. At ART initiation, 38% of participants reported suicidal thoughts or depressive symptoms. Initiating efavirenz-based ART was associated with a hazard ratio (HR) for suicidal thoughts below the null in the crude analysis [HR, 0.88; 95% confidence interval (CI): 0.53 to 1.45] and above the null in the weighted analysis (HR, 1.21; 95% CI: 0.66 to 2.28). Among those with a prior mental health issue, the weighted HR was 1.76 (95% CI: 0.45 to 6.86). Conclusions: After accounting for measured channeling bias, we observed no strong evidence that initiating efavirenz-containing ART increased the hazard of suicidal thoughts.

ObjectivesAs community viral load (CVL) measurements are associated with the incidence of new HIV-1 infections in a population, we hypothesized that similarly measured community drug resistance (CDR) could predict the prevalence of transmitted drug resistance (TDR).MethodsBetween 2001 and 2011, the prevalences of HIV-1 drug resistance for patients with established infection receiving HIV care (i.e. CDR) and TDR in recently infected patients were determined in San Diego. At each position in HIV-1 reverse transcriptase (RT) and protease (pro), drug resistance was evaluated both as the overall prevalence of resistance-associated mutations and by weighting each resistance position to the concurrent viral load of the patient and its proportion to the total viral load of the clinic (CVL). The weighting was the proportion of the CVL associated with patients identified with resistance at each residue. Spearman ranked correlation coefficients were used to determine associations between CDR and TDR.ResultsWe analysed 1088 resistance tests for 971 clinic patients and baseline resistance tests for 542 recently infected patients. CDR at positions 30, 46, and 88 in pro was associated with TDR between 2001 and 2011. When CDR was weighted by the viral load of patients, CDR was associated with TDR at position 103 in RT. Each of these associations was corroborated at least once using shorter measurement intervals.ConclusionsDespite evaluation of a limited percentage of chronically infected patients in San Diego, CDR correlated with TDR at key resistance positions and therefore may be a useful tool with which to predict the prevalence of TDR.

Background: Persons with human immunodeficiency virus (HIV) have higher risks for myocardial infarction (MI) than the general population. This is driven in part by higher type 2 MI (T2MI, due to coronary supply-demand mismatch) rates among persons with HIV (PWH). In the general population, T2MI has higher mortality than type 1 MI (T1MI, spontaneous and generally due to plaque rupture and thrombosis). PWH have a greater burden of comorbidities and may therefore have an even greater excess risk for complication and death in the setting of T2MI. However, mortality patterns after T1MI and T2MI in HIV are unknown. Methods: We analyzed mortality after MI among PWH enrolled in the multicenter, US-based Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort (N = 28,186). Incident MIs occurring between January 1, 1996, and December 31, 2014, were centrally adjudicated and classified as T1MI or T2MI. We first compared mortality following T1MI vs. T2MI among PWH. Cox survival analyses and Bayesian model averaging were then used to evaluate pre-MI covariates associated with mortality following T1MI and T2MI. Results: Among the 596 out of 28,186 PWH who experienced MI (2.1%; 293 T1MI and 303 T2MI), mortality rates were significantly greater after T2MI (22.2/100 person-years; 1-, 3-, and 5-year mortality 39%, 52%, and 62%) than T1MI (8.2/100 person-years; 1-, 3-, and 5-year mortality 15%, 22%, and 30%). Significant mortality predictors after T1MI were higher HIV viral load, renal dysfunction, and older age. Significant predictors of mortality after T2MI were low body-mass index (BMI) and detectable HIV viral load. Conclusions: Mortality is high following MI for PWH and substantially greater after T2MI than T1MI. Predictors of death after MI differed by type of MI, reinforcing the different clinical scenarios associated with each MI type and the importance of considering MI types separately.

Background: Prior studies suggest that transgender women (TW) with human immunodeficiency virus (HIV) are less likely to be virally suppressed than cisgender women (CW) and cisgender men (CM). However, prior data are limited by small sample sizes and cross-sectional designs. We sought to characterize the HIV care continuum comparing TW to CW and CM in the United States and Canada. Methods: We analyzed annual HIV care continuum outcomes by gender status from January 2001 through December 2015 among adults (aged ≥18 years) in 15 clinical cohorts. Outcomes were retention in care and viral suppression. Results: The study population included TW (n = 396), CW (n = 14 094), and CM (n = 101 667). TW had lower proportions retained in care than CW and CM (P

Background: Anemia is common among people living with HIV infection (PLWH) and is associated with adverse health outcomes. Information on risk factors for anemia incidence in the current antiretroviral therapy (ART) era is lacking. Methods: Within a prospective clinical cohort of adult PLWH receiving care at eight sites across the United States between 1/2010-3/2018, Cox proportional hazards regression analyses were conducted among a) PLWH free of anemia at baseline and b) PLWH free of severe anemia at baseline to determine associations between time-updated patient characteristics and development of anemia (hemoglobin < 10 g/dL), or severe anemia (hemoglobin < 7.5 g/dL). Linear mixed effects models were used to examine relationships between patient characteristics and hemoglobin levels during follow-up. Hemoglobin levels were ascertained using laboratory data from routine clinical care. Potential risk factors included: age, sex, race/ethnicity, body mass index, smoking status, hazardous alcohol use, illicit drug use, hepatitis C virus (HCV) coinfection, estimated glomerular filtration rate (eGFR), CD4 cell count, viral load, ART use and time in care at CNICS site. Results: This retrospective cohort study included 15,126 PLWH. During a median follow-up of 6.6 (interquartile range [IQR] 4.3-7.6) years, 1086 participants developed anemia and 465 participants developed severe anemia. Factors that were associated with incident anemia included: older age, female sex, black race, HCV coinfection, lower CD4 cell counts, VL ≥400 copies/ml and lower eGFR. Conclusion: Because anemia is a treatable condition associated with increased morbidity and mortality among PLWH, hemoglobin levels should be monitored routinely, especially among PLWH who have one or more risk factors for anemia.

Decisions about when to start or switch a therapy often depend on the frequency with which individuals are monitored or tested. For example, the optimal time to switch antiretroviral therapy depends on the frequency with which HIV-positive individuals have HIV RNA measured. This paper describes an approach to use observational data for the comparison of joint monitoring and treatment strategies and applies the method to a clinically relevant question in HIV research: when can monitoring frequency be decreased and when should individuals switch from a first-line treatment regimen to a new regimen?. We outline the target trial that would compare the dynamic strategies of interest and then describe how to emulate it using data from HIV-positive individuals included in the HIV-CAUSAL Collaboration and the Centers for AIDS Research Network of Integrated Clinical Systems. When, as in our example, few individuals follow the dynamic strategies of interest over long periods of follow-up, we describe how to leverage an additional assumption: no direct effect of monitoring on the outcome of interest. We compare our results with and without the “no direct effect” assumption. We found little differences on survival and AIDS-free survival between strategies where monitoring frequency was decreased at a CD4 threshold of 350 cells/μl compared with 500 cells/μl and where treatment was switched at an HIV-RNA threshold of 1000 copies/ml compared with 200 copies/ml. The “no direct effect” assumption resulted in efficiency improvements for the risk difference estimates ranging from an 7- to 53-fold increase in the effective sample size.

OBJECTIVE: Anaemia is common among people living with HIV (PLWH) and has been associated with certain, often older, antiretroviral medications. Information on current antiretroviral therapy (ART) and anaemia is limited. The objective was to compare the associations between anaemia incidence or haemoglobin change with core ART classes in the current ART era. DESIGN: Retrospective cohort study. SETTING: USA-based prospective clinical cohort of PLWH aged 18 and above receiving care at eight sites between January 2010 and March 2018. PARTICIPANTS: 16 505 PLWH were included in this study. MAIN OUTCOME MEASURES: Anaemia risk and haemoglobin change were estimated among PLWH for person-time on a protease inhibitor (PI) or an integrase strand transfer inhibitor (INSTI)-based regimen, relative to a non-nucleoside reverse transcriptase inhibitor (NNRTI)-based reference. We also examined PLWH on regimens containing multiple core classes. Cox proportional hazards regression analyses were conducted to measure the associations between time-updated ART classes and incident anaemia or severe anaemia. Linear mixed effects models were used to examine the relationships between ART classes and haemoglobin change. RESULTS: During a median of 4.9 years of follow-up, 1040 developed anaemia and 488 developed severe anaemia. Compared with NNRTI use, INSTI-based regimens were associated with an increased risk of anaemia (adjusted HR (aHR) 1.26, 95% CI 1.00 to 1.58) and severe anaemia (aHR 1.51, 95% CI 1.07 to 2.11) and a decrease in haemoglobin level. Time on multiple core classes was also associated with increased anaemia risk (aHR 1.39, 95% CI 1.13 to 1.70), while no associations were found for PI use. CONCLUSION: These findings suggest INSTI use may increase the risk of anaemia. If confirmed, screening for anaemia development in users of INSTIs may be beneficial. Further research into the underlying mechanisms is warranted.