Is d4T a viable option for children in low-income countries?

Logo 6th Paediatric WorkshopPolly Clayden, HIV i-Base

Two African studies presented at the 6th International Workshop on HIV Paediatrics looked at whether or not stavudine (d4T) is a viable option for children in low-income countries – with different conclusions.

Victor Musiime from the Joint Clinical Research Centre Kampala presented data from CHAPAS-3 on behalf of researchers from Uganda, Zambia and the Medical Research Council at UCL, London. [1]

In this trial, children aged 1 month to 13 years were randomly assigned (1:1:1) to receive paediatric co-formulated dual NRTI plus single NNRTI or fixed dose combination (FDC) pills in regimens of NNRTI plus lamivudine (3TC) plus either d4T, zidovudine (AZT) or abacavir (ABC). The primary endpoint was clinical grade 2/3 adverse event or laboratory adverse event confirmed grade 3 or any grade 4.

There had been no previous randomised comparisons of the three NRTIs in children. Dr Musiime noted that although d4T-associated lipodystrophy is well documented in adults and adolescents there are few data for younger children receiving lower World Health Organisation (WHO) recommended doses. Of the alternative NRTIs, some cohort data has questioned the efficacy of ABC and AZT-related anaemia is likely in malnourished children in settings where malaria is endemic.

CHAPAS-3 included both treatment-naive (n=365) and -experienced (n=113) children. At baseline children were well matched between arms. Overall half were girls; the median age of the naive and experienced children was respectively 2.6 years (IQR 1.6-4.00) and 6.2 years (IQR 5.5-7.2). Just over half (57%) of the treatment-naive children were less than 3 years old. The experienced children had been on d4T for a median of 3.5 years. All the experienced children had viral load <50 copies/mL at baseline, and the naive children had a median of 53,768 copies/mL (IQR 23,060-146,132). Median CD4 percentage was 20% cells/mm3 (IQR 13-25%) and 35% cells/mm3 (30-39%) in the naive and experienced children respectively.

Overall 353 (73%) children received nevirapine (NVP) and the remaining 125 (26%) efavirenz (EFV). All children less than 3 years old received NVP and the proportion receiving this NNRTI was similar within each arm.

The median follow up in the trial was 2.3 years (range 1.8-3.1). Loss to follow up was very low: 91% of children remained at the end of the trial and 98% of follow up visits were completed. There were 19 deaths (all treatment-naive children and 9 within 12 weeks of starting treatment), 17 children were lost to follow up and 8 were withdrawn from the trial.

Only 30 (6%) children substituted their assigned NRTIs: 8 switched from AZT for haematological toxicity; 2 from d4T for lipodystrophy; 9 because of TB treatment. Only 5 (1%) switched to second-line antiretroviral treatment.

The investigators reported that 312 children had 917 primary endpoints with no difference between arms: AZT vs d4T HR 0.99 (0.75-1.29); ABC vs d4T HR 0.88 (0.67-1.15), p=0.63.

Grade 3/4 neutropenia occurred more frequently in children in the AZT arm, p=0.04.

Lipodystrophy occurred in 2 children receiving d4T, p=0.11 – both cases were in experienced children aged 6 and 8 years who had received d4T for 2.5 and 5 years respectively.

Change in sum-of-four skinfold thickness and waist-hip ratio z-scores were similar across all arms: respectively p=0.33 and p=0.49.

At 96 weeks viral load was <100 copies/mL in 76%, 76% and 84% treatment-naive children receiving d4T, AZT and ABC respectively, p=0.32. For experienced children the respective proportions were 97%, 100% and 97%, p=0.51.

Changes in CD4 percentage were similar across arms (p=0.15), increasing from 20% to 36% in naive children, and remaining stable in experienced. There were 14 new WHO 3/4 events: 3, 4 and 7 in the d4T, AZT and ABC arms respectively. Of the 19 naive children that died 7, 3, 9 were in the d4T, AZT and ABC arms respectively, p= 0.5 for progression to WHO stage 3/4 and death.

Dr Musiime concluded: “Priority should be to identify children early and start ART, whichever NRTI is available”. He suggested that younger children could receive d4T.

In a related presentation, Renate Strehlau showed data from the NEVEREST 3 trial on behalf of colleagues from University of Witwatersrand, Johannesburg and Columbia University, New York. NEVEREST-3 was a randomised clinical trial investigating the virological efficacy of an EFV-containing regimen as long-term maintenance therapy in NVP-exposed children, conducted at Rahima Moosa Mother and Child Hospital, Johannesburg. Within the main study, children were randomised to switch to an ABC-containing regimen or remain on one containing d4T.

Dr Stehlau explained that in South Africa 3TC and d4T were used at the start of the ART programme – in 2004 – in first line treatment for children. In 2010 the national guidelines recommended ABC instead of d4T for children starting ART but those already receiving d4T with no adverse events should continue that NRTI. The 2013 guidelines recommended an ABC-containing regimen for children starting treatment and to change d4T to ABC if the viral load is undetectable. At the time that NEVEREST-3 was conducted the guidelines did not yet recommend switching children with no d4T-related adverse events.

The aim of the sub-study was to look at changes in CD4 and viral load; prevalence of lipodystrophy; changes in lipid concentrations and occurrence of ABC-related hypersensitivity reactions.

Of 300 screened, 213 children were randomised to remain on d4T (n=106) or switch to ABC (n=107). The majority of the 87 who were not eligible for randomisation had features that suggested lipodystrophy. Children were similar in both treatment arms: just over half were girls; at treatment initiation they were a mean age of 9.7 months and 8.5 in the d4T and ABC arms respectively and at screening children in both arms were a mean age of 4.2 years. They had been on ART for a mean of 3.4 years and 94% had viral load <50 copies/mL.

Unblinded clinician assessment identified more children in the d4T arm displaying features consistent with lipodystrophy through 48 weeks. But the differences between arms were only significant at 12 weeks (d4T vs ABC, 10.4 vs 2.9%, p=0.03) and 40 weeks (15.7 vs 4.9%, p=0.01) post randomisation.

At screening, fasting lipogram results did not show a significant overall difference in total cholesterol values between the two arms. At 8 weeks after switching to ABC, fasting lipogram results showed mean total cholesterol to be higher in children who had switched to ABC (4.4 vs 4.7 mmol/L, p=0.02). The difference was no longer evident at 48 weeks post randomisation.

The proportion of children with elevated LDL greater than 3.4mmol/L, or 131mg/dL, was also higher in children switched to ABC (11 vs 25%, p=0.01) at 8 weeks after the switch. But, at 48 weeks post-randomisation there was no significant difference in the proportion of children with raised LDL levels.

Although the data were not shown in the presentation, Dr Strehlau noted that after 48 weeks of follow-up, the d4T vs ABC groups also did not show significant differences in mean triglyceride levels.

Viral load, CD4, and anthropometric results did not differ between randomisation groups 56 weeks post-randomisation. There were no differences in occurrence of adverse skin manifestations between arms and no cases of ABC hypersensitivity reaction.

Dr Strelau concluded: “Switching virally suppressed children who are tolerating d4T to ABC, appears to be safe and may provide benefit with respect to reduction in the prevalence of lipodystrophy”.


It was good to see such recent data from CHAPAS-3 presented as a late breaker and comment from the audience included congratulations to the investigators for such a low rate of loss to follow up.

Another comment was that two years might not be enough time to see d4T side effects in children but no one could disagree with Dr Musiime’s conclusion that priority should be to identify children early and start ART.

The follow up time in the NEVEREST-3 substudy was not long either but this group did observe some feature consistent with lipodystrophy at two time points.

The younger age of some of the children in CHAPAS-3 might account for this.


  1. Musiime V et al. CHAPAS 3: A randomised trial comparing stavudine vs zidovudine vs abacavir as NRTI backbone in NNRTI- based first-line ART in 478 HIV- infected children in Uganda and Zambia. 6th International Workshop on HIV Pediatrics. Melbourne. 18-19 July 2014. Oral abstract O_21. (PDF)
  2. Strehlau R et al. Stavudine: a viable drug option for children in resource limited settings? 6th International Workshop on HIV Pediatrics. Melbourne. 18-19 July 2014. Oral abstract O_04.

Links to other websites are current at date of posting but not maintained.