P15. The Framingham score – a useful predictor of death (of all causes) in HIV-patients above 50 years of age on effective antiretroviral treatment (ART)

Göran A Bratt1, Cina Missalidis1, Julia Ekblom2 and Anders Blaxhult1
Affiliates: 1Venhälsan, Department of Infectious Diseases/Venhälsan, Södersjukhuset, Stockholm 2Student at Tekniska Högskolan, Stockholm

View posterpresentation

Methods
In the original retrospective study of the cohort in 2013 the median age was 56 years. The patients had been known HIV-positive for a median of 18 years, including 4 years without ART. 54% had had a CD4 nadir 100 cop/ml) was present only in 1% (5/547). Median CD4 was 600 x 106/L. 32% were smokers, 44% had hypertension and 43% hyperlipidemia. In 1% diabetes was diagnosed prior to and in 10% after the HIV diagnosis. Fasting metabolic parameters from the most recent yearly control were evaluated. The median 10 years general CVD risk (CHD, stroke, peripheral arterial disease or heart failure) calculated using the Framingham algorithm was 15.6%; 37.5% had a risk >20%. This cohort has now been re-examined for mortality over a 3year period. Logistic regression was used to calculate OR for death in relation to various dichotomized baseline characteristics.

Preliminary Results
4% (25/547) died during the three year observational period. Causes of death were: malignancies 48%, cardiovascular disease 28% and other 24%. Factors related to death were: diabetes diagnosed after HIV-diagnosis (OR 3,0; p 0,025); hypertension (BP >140/90 or on antihypertensive treatment) (OR 1,9; p 0,16); smoking at time of 2013 study (OR 1,5; p 0,33) and hsCRP >median (OR 1,6; p 0,31).A Framingham score >20% correlated to death of any cause (OR 2,9; p 0,013) and to non-cardiovascular death (trend; OR 2,4; p 0,075). Factors related to the HIV-infection (eg time with known HIV infection, time without ART, CD4 nadir <200) or ART (eg prior D4t or DDI treatment) had no impact on mortality. However, there was a weak trend for subjects on abacavir to have higher risk of dying as compared to patients on tenofovir (OR 1,5; p 0,3 vs OR 0,5; p 0,14). Thus 6% (11/197) of those on abacavir have died as compared to 3% (9/288) of those on tenofovir (p 0,076; Fisher´s exact test). Median hsCRP was higher among abacavir than tenofovir patients (1,6 vs 1,2 mg/L; p 0,01 (non-parametric test)).

Conclusion
Co-morbidities were frequent in this cohort. Although an increased mortality as compared to the normal population was documented, the number of deaths was small and our conclusions are only preliminary. However, classic risk factors like diabetes, hypertension and smoking, but not HIV-related factors seem to predict death in a 3 year period. The weak trend to difference between mortality related to nucleoside analogues probably has a complex background. The difference in hsCRP might indicate higher low-grade inflammation related to abacavir. In modern HIV-care prevention and treatment of co-morbidities should have high priority. The Framingham score (includes: age, sex, blood pressure, diabetes, current smoking, total cholesterol, and HD, and used to predict cardiovascular morbidity), seems useful in predicting all cause as well as non-cardiovascular mortality in this population.