Cardiovascular Risk Calculators Fail Some Patients With HIV

A cardiovascular risk calculator routinely used by physicians underestimates the risk for women and Black people living with HIV in high-income countries, while it overestimates risk for people living with HIV in low- and middle-income countries.

Risk scores are a key factor in guideline recommendations for prescription of statins and determine how many providers approach counseling with a specific patient.

For a Black or African American woman living with HIV in a high-income country, the risk score should be multiplied by 2.8 to more accurately estimate risk, researchers argue in The Lancet HIV. The scores should be multiplied by 2.6 for a woman of another ethnicity and by 1.25 for a Black or African American man. These figures are based on the pooled cohort equations for atherosclerotic cardiovascular disease (ASCVD), a risk calculator produced by the American College of Cardiology and the American Heart Association.

“The underprediction in certain groups strongly suggests that exclusive reliance on the pooled cohort equation will lead to undertreatment with a now proven cardiovascular prevention strategy in many instances,” Steven Grinspoon, MD, professor of medicine at Harvard Medical School, Boston, told Medscape Medical News.

For this analysis, Grinspoon and colleagues leveraged data from REPRIEVE, the landmark randomized controlled trial he led which demonstrated that daily statin therapy for people with HIV reduced major cardiovascular events by 35%. Eligible participants were assessed as having low to moderate ASCVD risk at study entry, so would not have been offered a statin under general population guidelines.

The findings led to guideline changes across the world. For example, US federal guidelines now strongly recommend statins for people living with HIV aged 40-75 years who have an estimated ASCVD risk above 5%. Below this threshold, statins are favored but with a lower grade of evidence, given the lower anticipated risk reduction.

The new study focused on those who received a placebo in REPRIEVE. It examined predicted and observed incidence of myocardial infarction, stroke, and cardiovascular death over 5 years. The analysis included 3893 people living with HIV aged over 40 years in 12 countries. While the cohort was diverse in many respects (65% non-White individuals and 31% women), a study limitation is that most had well-controlled HIV.

“Discrimination” of the ASCVD risk score (ie, its ability to differentiate people who go on to have a major cardiovascular event from those who do not) was moderate, while its “calibration” (ie, the extent to which it accurately reflected observed risk) was judged to be good.

Results were similar for the D:A:D risk score, a calculator developed with data from HIV cohorts that includes CD4 cell count alongside traditional risk factors. And both calculators significantly underpredicted the risk for women and Black people in high-income countries.

“If we look at the overall results, the risk estimators perform best if you are a white male living in a high-income country,” Madeleine Durand, MD told Medscape Medical News. She is an associate clinical professor at the Université de Montréal, Montreal, Canada, and was not involved in the study.

“Knowing your personal risk is highly privileged information,” Durand said. “The thing every person would want is a perfect estimation of their personal risk.”

Accurate estimates would help providers and patients in shared decision making. A patient whose absolute risk is low may decide that a daily medication for many years is not worth it, whereas a person with higher risk may be motivated to take the statin.

“If we look at it through an equity lens, everybody should be granted the same privileged information about their cardiovascular risk,” she said.

While the study’s correction factors for women and Black people could benefit from external validation, Durand felt that these figures would be useful to clinicians.

“They’re easy to present to patients,” she said. “It does help you give numbers to people to be able to best advise them on statin therapy.”

However, adjusting the score for Black people goes against the trend in recent years to remove race from clinical risk calculators, with indicators of socioeconomic status sometimes incorporated instead. PREVENT and SCORE-2 are two widely used cardiovascular risk calculators which do not consider race.

Grinspoon said that he is working on an exploratory analysis to assess the performance of these more modern calculators in the REPRIEVE cohort but anticipates that underestimation may be even more of an issue.

“What really needs to be done is to use the REPRIEVE data to build a new score, in an agnostic way, using virological, inflammatory and traditional risk factors, perhaps using artificial intelligence or other strategies,” he said. Artificial intelligence might help identify additional or unanticipated risk factors, he hopes.

“It’s going to take a little while, but we’re working on it.”

Another priority for future research is to develop risk scores for people living with HIV in low- and middle-income countries, whose risk was overestimated by the ASCVD and the D:A:D risk calculators — both created with data sets from high-income countries.

The study was funded by the US National Institutes of Health, Kowa Pharmaceuticals America, Gilead Sciences, and ViiV Healthcare. Grinspoon reported personal fees from ViiV Healthcare and Theratechnologies. Durand had no relevant financial disclosures.