ACTG 300 Leaves Many Questions Unanswered

On June 18, the NIH-sponsored pediatric trial ACTG 300 was brought to an early conclusion. Its Data and Safety Monitoring Board (DSMB) had determined that the data already collected amply demonstrated that AZT/3TC, or AZT/ddI for that matter, were superior to ddI monotherapy for treating children with HIV. ACTG 300 started enrolling in July 1995. It recruited children aged 6 weeks to 15 years who had symptomatic HIV infection and little to no prior anti-HIV treatment.

The original trial design included three regimens: ddI alone, AZT plus ddI, and AZT plus 3TC. Its purpose was to compare differences in clinical disease progression (i.e. the occurrence of new opportunistic infections, inadequate growth, neurologic deterioration or death) among those receiving the various treatments. But enrollment in the AZT/ddI arm of the trial was halted in June of 1996, on the news that ACTG 152 had found no significant difference between AZT/ddI and ddI alone (see Treatment Issues, March, 1996).

ACTG 300 has been hailed as proving that "two drugs are better than one" (NIH press release) or more simply that AZT/3TC "showed a 70 percent reduction in the risk of disease progression compared to ddI alone" (press release from BioChem Pharma, owner of patent rights to 3TC). As usual, the results have to be interpreted within the narrow confines of the trial. Of the 596 participants whose experience was analyzed, 471 were on either AZT/3TC or ddI. Since a much smaller number were assigned to AZT/ddI (from the first 11 months of trial enrollment only), most of the analysis performed so far compares only AZT/3TC to ddI alone. When AZT/ddI was compared with AZT/3TC, the results were similar, as shown in the table, and better than the ddI monotherapy arm. Also, participants' average time in the trial was only 11 months, considerably shorter than ACTG 152's 32 months. The final results might have been different if follow up had been continued as long in ACTG 300 as in the earlier trial. Results for both viral load (see table) and CD4 count from the different ACTG 300 trial arms seem to converge for the participants monitored 48 weeks or longer. Such convergence could presage a convergence in clinical disease progression.

Finally, virtually all the disease progression events occurred in the 53% of participants who were less than three years old. It is unknown, again because of the short follow-up time, how well the improvements seen in the younger children is mirrored in the older ones. The trial enrollees older than three obviously survived at least three years of HIV infection without any antiviral treatment. For these relatively slow progressors, ddI monotherapy might indeed have comparable benefits to double nucleoside analog combinations. Science, in any case, has moved on -- to triple drug therapies containing protease inhibitors and NNRTIs. By the time it was prematurely ended, ACTG 300 was already an anachronism.

ACTG 300 Results

	ddI	AZT+ddI	AZT+3TC
Disease progression and death, all participants	38 (n=235)	14 (n=125)	15 (n=236)
Disease progression and death, participants who began before 5/16/96	31 (n=124)	14 (n=125)	13 (n=123)
Deaths, all participants	15	2	3
Viral load (HIV RNA), participants who began before 5/16/96
week 12 (n=248)	-0.21 log	-0.57 log	-0.90 log
week 36 (n=216)	-0.31 log	-0.55 log	-0.72 log
week 48 (n=198)	-0.58 log	-0.70 log	-0.71 log
Notes: 50% of participants were observed for less than 9.4 months (the median follow up time); number followed for at least 48 weeks was 200. Both combination arms, including AZT/ddI, resulted in a statistically significant reduction in disease progression and increased survival when compared to ddI monotherapy (contrary to ACTG 152, which found an equivalence between AZT/ddI and ddI alone). Progression was mostly due to growth failure and neurologic deterioration (and neurologic deterioration was measured using a more sensitive scale than in ACTG 152). 83% of endpoints and all deaths occurred in the group that was less than 3 years of age. Only four cases of disease progression occurred in those greater than 3 years of age. Baseline viral load was about 5 logs, or 100,000 HIV RNA copies per ml of plasma.