Mortality after Loss-from-HIV-Care (LFHC): Implications for Persons of Older Age, with Late Diagnoses, and Unsuppressed Viral Load at the Time of LFHC from a Pennsylvania HIV Clinic.

Background: The occurrence and predictors of mortality after loss-from-HIV-care (LFHC) are not well elucidated, and were therefore assessed in this study.

Methods: The retrospective cohort study included 389 persons >12 years of age at the time of LFHC from a Pennsylvania clinic during 2003-2012. Predictor characteristics were reviewed, and follow-up outcomes were ascertained through 2014 by way of record-linkage of clinic Careware™ data with disease registries, claims databases of publicly-funded state health insurance, and death registry. Multivariate logistic regression analyses were performed to determine the likelihood of death after LFHC, and how this varies by several predictors.

Results: Approximately 25% of 389 persons studied were deceased, and all others were presumed alive. Multiple logistic regression analyses showed greater likelihood of death for those whose status at the time of LFHC was: a) 45-55 years old (adjusted odds ratio/aOR=3.41;95%Confidence-Interval,CI=1.41—8.25) compared to those 13-34, and even greater for those >=55 (aOR=5.51;95%CI=2.11-14.41); b) late diagnosed(HIV stage-3/AIDS) with unsuppressed viral load/VL (aOR=4.91;95%CI=1.14-21.23) compared to persons diagnosed before HIV stage-3 with suppressed VL, who were comparable to others. The likelihood of death was lower for: a) persons whose probable mode of HIV acquisition was ‘not documented’, other than heterosexual contact or males who have sex with men (aOR=0.28;95%CI=0.10-0.80) compared to those with a history of injection drug use, b) a mostly Latino/Hispanic group including a few others of unknown race/ethnicity (aOR=0.42;95%CI=0.19-0.91), compared to Caucasians/whites, who were comparable to African-Americans/blacks, and c) those with no documented re-engagement/not trackable after LFHC (aOR=0.51;95%CI=0.28-0.92), compared to persons with documented re-engagement at other clinics. There were no differences by several other characteristics.

Conclusion: The higher occurrence of death associated with older age and late diagnoses with unsuppressed VL at the time of LFHC may indicate a frailty effect suggesting a need for timely screening and closer case management support to address retention-in-care and viral suppression needs of such vulnerable populations. Fewer deaths among those not re-engaged require further study, including out-migration.