A research report submitted to the Faculty of Health Sciences,
University of the Witwatersrand, Johannesburg, in partial fulfillment of
the requirements for the degree of Masters of Medicine.
Johannesburg 2019 / Background: The high level of co-morbid TB/HIV cases with severe organ failure on
presentation in South Africa, results in an increased number of ICU admissions often with
a poor prognosis at presentation. In this study, the aim was to identify patients admitted
with HIV/TB co-infection and calculate the APACHE II scores and SOFA scores for each
patient. Predicted percentage mortality was compared with actual mortality. Predictors of
mortality were further identified, as well as the benefit of initiating ARV treatment in
patients who are ARV naive upon admission to ICU.
Methods: A retrospective audit of consecutive cases over a 24 month period was
completed. Patient demographics; CD 4 count; ARV treatment status; ICU and 30 day
mortality; the APACHE II Score; SOFA scores and correlating predicted percentage
mortality were documented. The survival of patients was assessed using Kaplan Meier
survival curves, and a univariate analysis was performed to identify risk factors for
mortality. Calculated predicted mortality was compared with actual mortality to validate
each scoring system and infer which was the better tool.
Results: Of 75 patients admitted with pulmonary (43 cases) or extra-pulmonary (32 cases)
TB, 23 died in the ICU (mortality 30,7%), and a further 10 died in the first 30 days of
hospitalisation (30 day mortality 44%). A survival analysis established ARV treatment and
CD 4 counts greater than 50 cells/mm3 were associated with a higher survival rate at any
point of the analysis. In the entire study period, only 2 patients were initiated on ARV
therapy during their ICU stay, 1 survived to discharge and 1 died in ICU. The APACHE II
Predicted Mortality was within the 95% Confidence Intervals for all groups while the SOFA
score was outside the upper bound limit of the 95% confidence intervals of actual mortality
for those patients taking ARV treatment (52%, 95% CI 43,1% - 59,5% vs actual mortality
30%, 95% CI 17,7% - 46,1%), those with a CD 4 count of more than 50 (53,5% 95% CI
45,4% - 60,6% vs actual mortality 34%, 95% CI 22,1% - 48,4%) and female patients
(51,2%, 95% CI 41,6% - 58,1% vs actual mortality 35,1%, 95% CI 21,4% - 50,4%).
Conclusion: The study found that both the APACHE II and SOFA scoring systems were
both statistically significant in prognosticating mortality in the study population. The
APACHE II scoring system however showed a slightly improved prognostication in specific
cohorts who had improved survival. It was also confirmed that patients with a CD 4 count
of more than 50 cells/mm3, and those on ARV therapy had a statistically significant
improved mortality. Further studies reviewing survival benefit of ARV initiation in ICU are
warranted.
ACKNOWLEDGEMENTS
Supervisor: Prof GA Richards
Co-Supervisor: Dr SHH Mohamadali
Statistician: Mr MH Zondi
Assistant - Data Collection: Ms S Madanlall / E.K. 2019
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:wits/oai:wiredspace.wits.ac.za:10539/28161 |
| Date | January 2019 |
| Creators | Singh, Avani |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | Online resource (37 leaves), application/pdf |
Page generated in 0.002 seconds