Background:
With an increased focus on cardiovascular disease (CVD) as a leading cause of death in people with HIV, monitoring patterns in CVD incidence and its influencing factors in real-life settings is crucial for informed clinical decision-making.
Methods:
We followed participants from the D:A:D and RESPOND cohort collaborations from baseline (D:A:D: latest of study entry or 1 Jan 2001; RESPOND: latest of local cohort enrolment or 1 Jan 2012) until the earliest of first CVD event (myocardial infarction [MI], stroke, invasive cardiovascular procedure [ICP]), final follow-up, or 1 Feb 2016 (D:A:D)/31 Dec 2021 (RESPOND). We calculated age-standardised CVD incidence rates (IRs) two-yearly from 2001–2021 and assessed temporal trends by Poisson regression, adjusting for time-updated potential confounders and cohort.
Results:
Of 66,680 included individuals, 18% were age >50 (median 40, interquartile range [IQR] 33–47) at baseline, 74% were male, 38% current smokers, 45% had dyslipidemia, 8% hypertension, 3% diabetes, and 1% prior CVD. Median CD4 cell count was 437 (270–630). Over a median of 8.8 (4.5–13.1) years (586,510 person-years, PY), there were 2,811 CVD events (IR 4.79/1000 PY [95% confidence interval (CI) 4.62–4.97]; 1363 MIs, 768 strokes, 680 ICPs). While the crude CVD incidence remained relatively stable over time, age-standardised IRs decreased from 8.65/1000 PY in 2001–2002 to 3.74/1000 PY in 2019–2021, IR ratio (IRR) 0.30 (95% CI 0.24–0.38, p<0.0001), with a steeper decline up to 2009 (Fig. A). The prevalence of most CVD risk factors was similar or decreased over time except for hypertension, which increased (Fig. B), possibly partly due to increased monitoring. Adjusting for hypertension accentuated the temporal CVD trends (IRR 0.26 [95% CI 0.20–0.32], p<0.0001) while changes in demographics (gender, ethnicity, mode of HIV acquisition), other known CVD risk factors (smoking, chronic kidney disease, body mass index, diabetes, dyslipidemia) or stage of HIV disease (CD4 nadir, prior AIDS) did not influence the decline in CVD IR.
Conclusions:
Combining data from two large, international collaborations, we have shown a decline in age-standardised CVD incidence in people with HIV from 2001 to 2021, most pronounced from 2001 to 2009. While causes of the decline in CVD incidence need to be investigated further, hypertension may have contributed to a slower decline over time. The CVD decline did not appear to be affected by changes in demographics, HIV disease stage and most known CVD risk factors.
