Abstract Body

Background:

Predicting cause-specific mortality among persons with HIV (PWH) could facilitate targeted care to improve survival. We assessed discrimination of the Veterans Aging Cohort Study (VACS) Index 2.0 in predicting cause-specific mortality among PWH on antiretroviral therapy (ART) in Europe and North America.

Methods:

The VACS Index 2.0 consists of thirteen variables: age, sex at birth, body mass index (BMI), CD4 count, creatinine, HIV-1 RNA viral load, haemoglobin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), platelet count, white blood cell count, albumin, and hepatitis C infection. Using data from 12 cohorts contributing to the Antiretroviral Therapy Cohort Collaboration, VACS 2.0 was calculated for PWH who initiated ART between 2000 and 2018, around a randomly selected visit date within the period 1 year after ART initiation to last visit. Missingness in VACS 2.0 variables was addressed through multiple imputation. We estimated associations between VACS 2.0 and specific causes of death using Cox models, with discrimination evaluated using Harrell’s C-statistic. Absolute mortality risks were modelled with flexible parametric survival models.

Results:

Among 59,741 PWH, 80% were men and at follow-up start mean age was 43 and mean VACS 2.0 score was 41.0. VACS 2.0 values were measured a median of 3.2 years after ART initiation and ranged from 0 to 129: higher values indicate worse prognosis. There were 3,117 deaths during 217,257 person-years (median follow-up was 2.6 years). Non-AIDS-defining cancers were the most common cause of death (n=569), followed by AIDS (n=527). Discrimination for five-year all-cause mortality was C=0.83). Discrimination for specific causes of death was highest for deaths due to AIDS (0.91), liver-related causes (0.90) and respiratory causes (0.88). Discrimination was lowest for suicides/accidental deaths (0.65), unclassifiable causes (0.76), cardiovascular deaths (0.77) and ‘other’ causes (0.78). Predicted probabilities of 5-year mortality were low for except for PWH with very high (>90) VACS 2.0 values e.g. the 5-year probability of AIDS death for those with VACS 2.0 score 40 was 0.1%, increasing to 7.6% for scores of 90.

Conclusions:

To improve discrimination for causes for which discrimination is lower, future VACS Index analyses could incorporate additional measures, such as psychopathological risk factors, extra biomarkers (for example, lipid profiles and inflammatory markers), or conditions included in the Charlson Comorbidity Index.