Skip navigation

HIV protease inhibitors: ACTG 320 trial

Bandolier 41 discussed the importance of the concentration of HIV virus in plasma or serum (viral load) as a surrogate marker for beating the disease. We gave some exciting results of the ACTG 320 trial taken from an Internet source. The full paper has now been published [1], and the pukka results are in the Table.

ACTG 320 trial

Briefly, this was a randomised, double-blind, placebo-controlled trial in which people with HIV-1 infection and CD4 cell counts of no more than 200/µL received either a two nucleoside regimen plus protease inhibitor (indinavir), or the same regimen plus placebo. Stratification was by CD4 count more or less than 50 cells/µL. Follow up was at 4, 8, and 16 weeks, and every 8 weeks thereafter up to 40 weeks.


1156 people were randomised, and the median duration was 38 weeks. Patients lost to follow-up were few, at 5%. There were 227 (20%) premature discontinuations, of which a major portion was patients seeking open-label treatment with protease inhibitor.

Protease inhibitor and HIV: results from ACTG320 trial [1]
Outcome Nucleosides plus indinavir Nucleosides plus placebo Relative risk Absolute risk reduction NNT
All patients
Aids or death 33/577 63/579 0.53 (0.35 - 0.79) 0.05 19 (12 - 50)
Death 8/577 18/579 0.45 (0.20 - 1.03)* 0.02 1000)
Plasma viral load <500 molecules/mL 52/577 347/579 6.7 (5.1 - 8.7) 0.51 2.0 (1.8 - 2.2)
Neutropenia 29/577 87/579 0.33 (0.22 - 0.49) 0.10 10 (7.5 - 15)
Hyperbilirubinaemia 35/577 6/579 5.9 (2.5 - 14) 0.05 20 (14 - 34)
Renal colic 21/577 5/579 4.2 (1.6 - 11) 0.03 36 (22 - 93)
Aids or death CD4 <=50/cu mm 23/219 44/220 0.53 (0.33 - 0.85) 0.09 11 (6.2 - 35)
Aids or death CD4 51-200/cu mm 10/358 19/359 0.53 (0.25 - 1.12) 0.02 40 (18 - inf)
Plasma viral load <500 molecules/mL From [2] 24/28 5/28 2.0 (1.3 - 3.2) 0.43 2.3 (1.5 - 4.9)
* Statistically significant by hazard ratio calculation in original article
AIDS or death
Addition of protease inhibitor halved the rate of progression to AIDS or death. Overall the NNT was 19, but in those patients with the lowest CD4 cell counts of <=50/µL (that is, the most advanced disease), the same halving of the rate produced an NNT of 11 because of the higher progression rate.

Plasma viral load
Protease inhibitor was very effective in reducing the plasma viral load to levels below the sensitivity limit of the assay (500 molecules/mL). Sixty percent of patients treated with protease inhibitor had such low values, compared with 9% in the placebo group. The NNT was 2. A similar NNT of 2 for reducing plasma viral load to such low levels was seen in an accompanying article [2].
CD4 cell count
CD4 cell counts rose substantially in patients given protease inhibitor, by nearly 150/µL over 40 weeks. Placebo-treated patients had rises which were below 50/µL.

Adverse effects
Neutropenia was much less common in patients on protease inhibitor, but both hyperbilirubinaemia and renal colic were more common with protease inhibitor.


Bandolier has already said that this is exciting stuff. The one fly in this particular ointment may be the need to generate more certainty about the use of plasma viral load as a prognostic factor. It's fine to chase it as an end point of treatment, but the ACTG 320 trial, for instance, fails to couple data from viral load to outcome. Trialists and their commercial sponsors, who have the original data, could do us all a favour by doing a bit more on this.


  1. SM Hammer, KE Squires, MD Hughes et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. New England Journal of Medicine 1997 337: 725-33.
  2. RM Gulick, JW Mellors, D Havlir et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral infection. New England Journal of Medicine 1997 337: 734-9.

previous or next story in this issue