Integrated management of cryptococcal and opportunistic infections to improve outcomes in advanced HIV disease (IMPROVE)

Cryptococcus remains the most common cause of HIV-associated meningitis, accounting for 19% of all AIDS-related deaths globally. Despite antifungal therapy, 10-week mortality in sub-Saharan Africa remains between 24-45%, even in the context of clinical trials. The key drivers of this persistently high mortality are not known. The overarching hypothesis for my PhD is that co-prevalent opportunistic infections (OIs) including tuberculosis (TB), Human Herpes virus (HHV) infections, and blood stream infections (BSIs) are common amongst patients with cryptococcal meningitis, and contribute to poor patient outcomes.

The overall aim of my PhD research was to synthesise data on the prevalence of key OIs amongst adults with HIV-associated cryptococcal meningitis, and to generate safety and feasibility data to inform TB preventive therapy (TPT) strategies in advanced HIV disease (AHD).

In my first PhD work package I conducted three retrospective cohort studies to investigate the prevalence, and prognostic significance of baseline EBV and CMV infections amongst patients presenting with meningitis across five African countries. These three cohort studies constitute my first three PhD research papers presented in chapters three to five. By utilising clinical data and cryopreserved baseline clinical samples for EBV and CMV quantitative polymerase chain reactions (qPCR), from two prospective observational studies, and three randomised trials, I demonstrated that EBV and CMV co-infections occur frequently in both the plasma and CSF of adults presenting with AHD and cryptococcal meningitis. As presented in chapters three to five, CMV viraemia occurred in 37-49%, CMV central nervous system infection in 5%, EBV viraemia in 73%, and EBV central nervous system infection in 18-27% of participants with cryptococcal meningitis. Whilst both CMV- and EBV co-infections were common however, utilising data from 811 participants from the AMBITION-cm trial recruited from five African countries, I demonstrated that CMV and EBV co-infections were associated with contrasting clinical phenotypes. EBV co-infections were associated with higher median CD4 counts and a pro-inflammatory cryptococcal profile known to be associated with improved survival and were not associated with poor outcomes; suggesting that EBV in the context of HIV-associated meningitis is probably a bystander phenomenon. Conversely, CMV co-infections were associated with a pauci-inflammatory response, higher fungal burdens and increased mortality as compared to participants without CMV co-infections. High-level CMV viraemia (≥1000 copies/ml) was significantly and independently associated with greater than double the odds of 2-, and 10-week mortality (adjusted odds ratio (aOR) 2.44; 95%CI 1.33-4.45), after adjustment for potential confounders. In addition, in the cohort study of 497 Ugandan adults diagnosed with cryptococcal meningitis presented in chapter five, I demonstrated that high level CMV viraemia (≥1000 IU/ml) was associated with double the hazard of incident TB disease during follow-up (subdistribution adjusted hazard ratio 2.18; 95%CI 1.11–4.27). These data add to the growing body of evidence that CMV viraemia in the context of AHD is associated with worse short- and medium-term outcomes.

Thereafter for my second PhD work package, I present data from the IMPROVE cohort study, a prospective cohort study of 450 adults with HIV-associated cryptococcal meningitis conducted in Uganda to describe the co-prevalence of TB, CMV and BSI through delivery of an enhanced opportunistic screening package. Through design and delivery of this prospective cohort study, I demonstrated that active TB disease (45% prevalence) and CMV viraemia (34% prevalence) are common amongst hospitalised adults diagnosed with cryptococcal meningitis, with community acquired (1% prevalence) and nosocomial BSI (3% prevalence) occurring less frequently. Co-prevalent TB disease (aOR 1.65; 95%CI 1.06-2.57) and CMV viraemia (aOR 2.12; 95%CI 1.32-3.40) were associated with increased odds of 18-week mortality within the IMPROVE prospective cohort; and participants with both TB disease and CMV viraemia had the highest overall mortality risk, 52% at 18-weeks.

Finally for my third PhD work package, I present the results of the nested IMPROVE randomised controlled trial (RCT), a multi-site open-label, non-inferiority strategy trial to evaluate the safety and feasibility of two strategies for the delivery of 1HP (one month of daily rifapentine plus isoniazid) in adults with AHD following initiation of treatment for cryptococcal meningitis. We randomised adults with cryptococcal meningitis and no evidence of active TB disease in a 1:1 ratio to inpatient initiation of 1HP or outpatient initiation of 1HP at week-6 after initiation of anti-fungal therapy (standard of care). The primary end point was TB-disease free 1HP treatment completion at 18-weeks, powered for a 15% noninferiority margin. A total of 205 participants underwent randomisation. In the primary intention-to-treat analysis, 69.9% (72/103) with inpatient 1HP initiation had TB-disease free survival and 1HP treatment completion at 18-weeks vs. 61.8% (63/102) in the outpatient 1HP arm (adjusted risk difference 7.1% 95%CI -5.9% to 20.0%). No significant differences existed in any secondary endpoints when comparing inpatient vs. outpatient 1HP initiation. The IMPROVE trial demonstrated that inpatient initiation of 1HP TPT was non-inferior to outpatient initiation amongst adults with AHD and cryptococcal meningitis. Our data suggest that following exclusion of active TB disease, inpatient initiation of 1HP TPT is feasible and safe.

In conclusion though my PhD research, I have demonstrated that co-infections amongst adults diagnosed with HIV-associated cryptococcal meningitis are common. In particular TB disease and CMV viraemia are highly prevalent, occurring in over half of study participants. Both CMV viraemia and TB disease were strongly associated with increased 18-week mortality and represent potentially modifiable risk factors to improve survival amongst adults with AHD and cryptococcal meningitis. Optimised pre-emptive treatment strategies for TB disease and CMV viraemia may improve outcomes for this extremely high risk population, and interventional trials are now required. Finally, results from the IMPROVE RCT suggest that – following exclusion of active TB disease - 1HP TPT can be safely and feasibly initiated prior to discharge and should be considered as part of an AHD integrated package of care for hospitalised adults treated for cryptococcal meningitis.