Predictors of early death in a cohort of Ethiopian patients treated with HAART

Abstract

Background: HAART has improved the survival of HIV infected patients. However, compared to patients in high-income countries, patients in resource-poor countries have higher mortality rates. Our objective was to identify independent risk factors for death in Ethiopian patients treated with HAART.

Methods: In a district hospital in Ethiopia, we treated adult HIV infected patients with HAART based on clinical and total lymphocyte count (TLC) criteria. We measured body weight and complete blood cell count at baseline, 4 weeks later, then repeated weight every month and complete blood cell count every 12 weeks. Time to death was the main outcome variable. We used the Kaplan Meier and Cox regression survival analyses to identify prognostic markers. Also, we calculated mortality rates for the different phases of the follow-up.

Results: Out of 162 recruited, 152 treatment-naïve patients contributed 144.1 person-years of observation (PYO). 86 (57%) of them were men and their median age was 32 years. 24 patients died, making the overall mortality rate 16.7 per 100 PYO. The highest death rate occurred in the first month of treatment. Compared to the first month, mortality declined by 9-fold after the 18th week of follow-up. Being in WHO clinical stage IV and having TLC< or = 750/mcL were independent predictors of death. Haemoglobin (HGB) < or = 10 g/dl and TLC<= 1200/mcL at baseline were not associated with increased mortality. Body mass index (BMI) < or = 18.5 kg/m2 at baseline was associated with death in univariate analysis. Weight loss was seen in about a third of patients who survived up to the fourth week, and it was associated with increased death. Decline in TLC, HGB and BMI was associated with death in univariate analysis only.

Conclusion: The high mortality rate seen in this cohort was associated with advanced disease stage and very low TLC at presentation. Patients should be identified and treated before they progress to advanced stages. The underlying causes for early death in patients presenting at late stages should be investigated.

Figure 1

Cohort profile. 34 patients had no follow-up laboratory records at 4 weeks. In 12 of the 34 patients, end point events including eight deaths occurred before the planned first visit.

Figure 2

Mortality according to disease stage, weight loss and decline in total lymphocyte count. A – Mortality according to WHO clinical staging. Note that mortality was highest in patients with stage IV disease particularly in the first 12 weeks of treatment. B- Kaplan-Meier survival curve showing higher mortality among patients with weight loss. Since weight was measured at about 4 weeks of treatment, it does not show earlier deaths. Rather it shows deaths that occurred after 12 weeks of treatment. Note that the time-axis represents the time from baseline to death. C- Change in total lymphocyte count and mortality. This figure shows the higher mortality among patients with no change or decrease in total lymphocyte count from baseline to week 18.