Patients and follow up

A total of 1253 patients with the minimum of three months of follow up were included in the final analysis. Median duration of follow up was 18 months. The mean age of the patients at baseline was 36.3 years (median 34.5; range 18-70). Other characteristics of the patients at baseline are summarized in Table 2. The smaller number of women taking ARV therapy is a reflection of gender discrimination prevalent in the society. About 80% of patients decided to take d4T/3TC as their backbone nucleoside because it was much cheaper than the AZT/3TC-based regimen.

Table 2. Characteristics of patients at baseline

Almost 78% patients initiated therapy when their CD4 dropped to less than 200/mm³, an indication of patients presenting late to our clinics. Additionally, we routinely use a CD4 count of less than 200/mm³ as an indication to offer ARV therapy.

Adverse events

Most of the adverse events occurred within 1-8 weeks of initiation of therapy. Rash was documented in 6.9% (n=86, 95% CI 5.5-8.3), clinical hepatitis in 3.2% (n=41, CI 2.3-4.8) and gastrointestinal disturbances such as nausea, vomiting and diarrhoea in 15.5% (n=237, CI 13.1-16.9) of patients. Eight patients developed Grade 4 rash and 2 died due to the same. Most of the rashes occurred when the patient switched from the lead-in to full dose of NVP, while Grade 4 rashes occurred within 2-3 days of initiation of NVP. None of the patients had fulminant hepatitis.

Female gender was associated with significantly higher risk of development of any adverse event (see Table 3). Hepatitis virus status was not assessed for all patients; hence the contribution of these viruses to development of hepatotoxicity could not be evaluated.

Table 3. Risk factors for development of adverse events

Immunological improvement

The mean CD4 count at baseline was 130.5/mm³ (SD 97.8; median 115; range 2-814). Twelve percent patients were lost to follow up. Patients who showed an improvement in CD4 counts reported more than 95% adherence to their regimen. Ninety-three patients had a significant decline in their CD4 counts warranting change in the treatment regimen; only seven of these reported more than 95% adherence to their initial regimen.

Mean increases in CD4 counts during 24 months of follow up are shown in Table 4. There was a rapid improvement in the mean CD4 count in the first 3-6 months, which later reached a plateau (Figure 1). The difference in the mean CD4 counts at baseline and at 12 and 24 months was significant (p<0.001).

Table 4. Improvement in CD4 counts during 24 months of treatment

Figure 1: Box plot of CD4 response over time

Clinical outcomes

The incidence of deaths amongst patients was 5.2 per 100 person years of follow up. The causes of HIV-attributable death were disseminated TB including tuberculous meningitis (21), cryptococcal meningitis (4), severe bacterial pneumonia with sepsis (3), progressive multifocal leukoencephalopathy (3), lymphoma (2), toxoplasmosis (1), Pneumocystis carinii pneumonia (1), cryptosporidial diarrhoea with renal failure (1), cytomegalovirus pneumonia (1), herpes simplex encephalitis (1) and not determined (2). Eight deaths were attributed to drug toxicity. The causes of death included lactic acidosis (4), severe skin rash (2) and pancreatitis (2). The median baseline CD4 count of patients who died was 93/mm³.

The incidence of development of clinical events was 28.1 per 100 person years of follow up. The frequency of occurrence of various clinical events is shown in Table 5. Sixty-seven percent occurred within 2-12 weeks of initiation of therapy and were defined as IRDs. Only baseline CD4 counts were associated with increased risk of development of clinical events or death (OR 0.994, CI 0.990-0.997, p<0.001).

Table 5. Type and frequency of occurrence of clinical events in patients receiving HAART