Four-year mortality and ART outcomes among adults with HIV and microbiologically confirmed tuberculosis during early ULAM rollout at a Cape Town district hospital

Abstract

Background. Tuberculosis (TB) is the leading cause of mortality in people living with HIV (PLWH). In 2015, the World Health Organization advocated use of the urine lipoarabinomannan (ULAM) assay to accelerate the diagnosis of TB, and thus its treatment, in selected patients with HIV. Objectives. To describe the demographic and clinical profiles of inpatients with HIV and microbiologically confirmed TB at a district hospital during early ULAM rollout, and to describe their short and longer-term outcomes, including their 4-year all-cause mortality, antiretroviral therapy (ART) uptake, and virological outcomes. Method. We conducted an observational, retrospective cohort medical record review describing a subset of inpatients at Mitchells Plain Hospital for whom ULAMs were performed between 2016 and 2018, during the initial rollout. Results. A total of 107 participants were included. The median (interquartile range (IQR)) age was 35 (29 - 42) years with a slight male preponderance of 55.14% (n=59). A total of 75.7% (n=81) of participants had been previously diagnosed with HIV, but only 35.51% (n=38) were on ART at enrolment. The median CD4 count was 39 (18 - 69.5) cells/mL. A total of 96% (n=96/100) of participants had anaemia – with ULAM-positive participants having lower haemoglobins than ULAM-negative participants (p=0.036). Participants who were ULAM positive also had a lower platelet count than those who were ULAM negative (p=0.019). The only biochemical result for which there was a statistically significant difference between groups was urea: participants who were ULAM positive had higher ureas (p=0.037). The median length of stay was 10 (6 - 20) days, and the 30-day mortality rate was 4.67% (n=5). The 4-year all-cause mortality was 34.58% (n=37). At 4 years, 31.78% (n=34) of participants were on ART and 42.99% (n=46) achieved an undetectable viral load (VL) at some point within their follow-up. Participants who achieved an undetectable HIV VL were significantly less likely to have died by 4 years (odds ratio 0.06 (95% confidence interval 0.02 - 0.17, p<0.001). Conclusion. This cohort of medical inpatients with HIV and microbiologically confirmed TB were profoundly immunocompromised. Despite three-quarters of participants having been diagnosed with HIV previously and all PLWH being eligible for ART, only one- third of participants were on ART at enrolment. The 4-year all-cause mortality was significantly higher than the 30-day mortality. Disengagement from care was common, highlighting the importance of quality primary healthcare programmes to curtail HIV- associated morbidity and mortality.