Table 2.
Primary and secondary endpoints (time to event)
| Stavudine (n=156); N (%) |
Zidovudine (n=157) |
Abacavir (n=164) |
Abacavir vs zidovudine; HR (95% CI) | |||||
|---|---|---|---|---|---|---|---|---|
| N (%) | HR vs stavudine (95% CI) | N (%) | HR vs stavudine (95% CI) | p value* | ||||
| Primary endpoint adverse event† | 104 (67%) | 103 (65%) | 0·99 (0·75–1·29) | 105 (64%) | 0·88 (0·67–1·15) | 0·63 | 0·89 (0·68–1·17) | |
| Specific subsets of primary endpoint adverse events | ||||||||
| Anaemia, grade 3/4 | 5 (3%) | 9 (6%) | 1·93 (0·64–5·76) | 6 (4%) | 1·15 (0·35–3·78) | 0·42 | 0·60 (0·21–1·69) | |
| Anaemia, grade 4 | 5 (3%) | 7 (4%) | 1·45 (0·46–4·57) | 3 (2%) | 0·57 (0·14–2·38) | 0·38 | 0·39 (0·10–1·52) | |
| Neutropenia, grade 3/4 | 4 (3%) | 12 (8%) | 3·21 (1·03–9·98) | 5 (3%) | 1·21 (0·32–4·49) | 0·04 | 0·38 (0·13–1·07) | |
| Neutropenia, grade 4 | 3 (2%) | 10 (6%) | 3·55 (0·97–12·9) | 4 (2%) | 1·29 (0·29–5·75) | 0·06 | 0·36 (0·11–1·16) | |
| Hypersensitivity reaction‡ | 5 (3%) | 1 (0·6%) | 0·22 (0·03–1·86) | 2 (1%) | 0·38 (0·07–1·95) | 0·21 | 1·75 (0·16–19·3) | |
| Lipodystrophy/lipoatrophy | 2 (1%) | 0 | .. | 0 | .. | 0·08 | .. | |
| Mitochondrial disease§ | 1 (0·6%) | 0 | .. | 1 (0·6%) | .. | 0·65 | .. | |
| Grade 3/4 adverse events¶ | 46 (29%) | 53 (34%) | 1·24 (0·83–1·84) | 51 (31%) | 1·01 (0·68–1·50) | 0·48 | 0·82 (0·56–1·20) | |
| Grade 3/4 adverse events adjudicated as NRTI related‖ | 6 (4%) | 12 (8%) | 2·12 (0·79–5·66) | 5 (3%) | 0·80 (0·25–2·63) | 0·10 | 0·38 (0·13–1·08) | |
| Serious adverse events | 46 (29%) | 44 (28%) | 0·98 (0·65–1·48) | 42 (26%) | 0·78 (0·51–1·19) | 0·46 | 0·80 (0·52–1·22) | |
| Serious adverse events adjudicated as NRTI related‖ | 8 (5%) | 12 (12%) | 1·50 (0·61–3·67) | 6 (6%) | 0·64 (0·22–1·85) | 0·22 | 0·43 (0·16–1·14) | |
| Toxicity causing ART modification** | 4 (3%) | 9 (6%) | 2·24 (0·69–7·32) | 1 (1%) | 0·23 (0·03–2·09) | 0·03 | 0·10 (0·01–0·83) | |
| New WHO stage 3 or 4 event or death | 9 (6%) | 7 (4%) | 0·84 (0·31–2·26) | 13 (8%) | 1·37 (0·58–2·30) | 0·55 | 1·62 (0·65–4·07) | |
| Death†† | 7 (4%) | 3 (2%) | 0·48 (0·12–1·85) | 9 (5%) | 1·23 (0·46–3·29) | 0·35 | 2·56 (0·69–9·45) | |
All HRs were stratified for randomisation stratification factors. No evidence of interaction between naive versus experienced strata on any outcome in table 2 (p>0·1; 21 tests), except for serious adverse events (p=0·02; naive children with serious adverse events: 46 allocated stavudine, 40 allocated zidovudine, and 39 allocated abacavir; experienced children: none allocated stavudine, four allocated zidovudine, and three allocated abacavir; serious adverse events were most commonly lower respiratory tract infections or other specific infections). HR=hazard ratio. NRTI=nucleoside reverse-transcriptase inhibitors. ART=antiretroviral treatment.
Stratified log-rank test.
Clinical grade 2 or greater, laboratory grade 3 (confirmed), laboratory grade 4 (all).
Includes grade 4 Stevens-Johnson syndrome from appendix p 7; adjudicated blind to actual NRTI received. Both children in the abacavir group continued abacavir without adverse effects. See appendix p 11 for details of adjudicated relation to antiretrovirals. One additional grade 1 hypersensitivity reaction (not a primary endpoint) also occurred in the abacavir group: again the child continued abacavir without adverse effects.
One cardiomyopathy (stavudine group) and one myopathy (abacavir group).
Clinical grade 3 or greater, laboratory grade 3 (confirmed), laboratory grade 4 (all).
See appendix p 4 for details of grade 3/4 adverse events and serious adverse events judged by the endpoint review committee as possibly having some relation to any of stavudine or zidovudine or abacavir (all events adjudicated for each drug blind to NRTI actually received, so toxicity from any of the three NRTIs could be attributed to each event). No grade 3/4 adverse events or serious adverse events were judged probably or definitely related.
Not including changes for tuberculosis treatment (see main text); one toxicity substitution in child randomised to abacavir was actually from zidovudine to nevirapine for anaemia, following previous substitution of nevirapine to zidovudine (triple NRTI regimen) for tuberculosis treatment. Two stavudine and one zidovudine substitutions from nevirapine to lopinavir/ritonavir for rash/hypersensitivity reactions; all other substitutions were from stavudine or zidovudine.
See appendix p 15 for relation between NRTIs and deaths.