DART: high rates of viral suppression after five years and a single CD4 test with a threshold of 250 cells/mm3 could reduce unnecessary switching
Polly Clayden, HIV i-Base
DART was a randomised trial comparing clinically driven monitoring (CDM) to laboratory (CD4, haematology, biochemistry) plus clinical monitoring (LCM) of 3316 HAART-naïve adults conducted in Uganda and Zimbabwe. People in both monitoring arms showed high and similar 5-year survival rate – 90% vs 87% in the LCM and CDM arms respectively – differing by a small percentage that only occurred after two years of follow up. This compared to an historical 5-year survival rate prior to HAART of only 8% in the Uganda cohort. 
First line HAART in this trial was AZT/3TC plus either TDF (74%), ABC (9%) or NVP (16%). Participants needing to switch to second line received LPV/r plus NRTI/s and/or NNRTI. Neither the CDM nor LCM group had real time viral load monitoring.
Ugandan patients who did not participate in one of two, nested second line RCTs had a viral load test when they left the trial and joined the national programme.
Further findings from the DART trial were presented at CROI 2011.
Cissy Kityo and colleagues showed high rates of virological suppression at 5 years after HAART initiation among the Ugandan participants alive and in follow up. 
Both monitoring groups switched to second line therapy following WHO stage 4 or multiple stage 3 events; the LCM group also switched at CD4 <100 cells/mm3.
A viral load measurement was available the end of the trial for the majority of eligible participants: 1207 (80%) and 187 (70%) respectively receiving first and second line at exit. The viral load sample was taken at a median of 5.2 years after initiation of HAART and 2.7 years after start of second line for those who had switched.
Of the participants who remained on first line, 81.9% (95%CI 78.5-84.9%) in LCM and 74.2% (95%CI 70.6-77.6%) in CDM had viral load <200 copies mL, p=0.001. In the LCM group 5.6% (95% CI, 3.9-7.8% had viral loads <10,000 copies, which was lower than the 10.4% (95%CI 8.1-13.1%) of participants in CDM.
Of those who switched, viral loads were similar across the two monitoring groups, p=0.6. Viral load <200 copies/mL was achieved in 88.8% (95%CI 83.3-92.9%) of participants receiving second line.
When the investigators examined the CD4 count nearest to the exit viral load measurement (taken at a maximum of 6 months apart), they found a negative association, r=0.4, as would be expected.
Of 283 (20%) participants with viral load >200 copies/mL, 29% in the LCM group and 42.2% in CDM had CD4 <200 cells/mm3.
The investigators noted that CD4 counts <100 cells/mm3 were rare in either arm; only 2 people in LCM and 7 in CDM.
A related study showed a single CD4 test with a threshold of >250 cells/mm3 could reduce inappropriate switching in clinically monitored patients. 
Charles Gilks and colleagues investigated the relationship between CD4 count at switch and the reason for doing so in all 675 (361 LCM and 314 CDM) DART participants switching to second line.
In the CDM arm, 206 (66%) switched due to WHO stage 4 events and 76 (24%)/32 (10%) participants single or multiple WHO stage 3 events, respectively. In LCM 265 (73%) participants switched because their CD4 count fell below 100 cells/mm3, 43 (12%) for other CD4 reasons, 37 (10%) due to WHO 4 events and 6 (23%)/10 (3%) single or multiple WHO stage 3 events.
In the LCM arm, clinical failure provoked switching in 7 (2%) of patients with CD4 >250 cells/mm3; 3 due to WHO stage 4 events, 1 single WHO stage 3 event and 3 for other CD4 reasons. This compared to 64 (20%) of participants who switched with CD4 > in the CDM arm, p=0.001. The investigators noted, however, that deaths within one year of switching were similar in CDM whether participants switched above or below 250 cells/mm3, 11/64 (17%) vs 33/250 (13%) respectively.
In the CDM group, switching due to a single WHO grade 3 event was significantly more frequent with a CD4 count of >250 cells/mm3 (27/76, 36%) compared to multiple WHO stage 3 events (4/32, 12%) or WHO stage 4 events (33/206, 16%), p=0.001.
Viral load measurements at switch were available for 108 and 113 participants in the LCM and CDM groups respectively. Of these, 15 (14%) vs 32 (28%) respectively were <400 copies/mL, p=0.009.
In the CDM group, 25/31 (81%) with clinical failure and CD4 > 250 cells/mm3 had viral load <400 copies/mL vs 7/82 (9%) with CD4 <250 cells/mm3, p<0.001.
The investigators noted a trend to switching for single WHO stage 3 events compared to multiple WHO stage 3 or stage 4, but this was not significant, p=0.22.
They concluded that among clinically monitored patients, a single CD4 test with a threshold of 250 cells/mm3 could identify up to 80% with viral load <400 copies/mL who are unlikely to benefit from second line therapy. In DART, nearly 40% of participants who failed clinically with a single WHO stage 3 event had CD4 >250 cells/mm3. They wrote: “Targeting this group would be particularly likely to avoid premature, costly switching to second line.”
- DART trial team. Routine versus clinically driven laboratory monitoring of HIV antiretroviral therapy in Africa (DART): a randomised non-inferiority trial. The Lancet, Volume 375, Issue 9709. 9 January 2010.
- Kityo C et al. High rates of virologic suppression among patients not receiving routine virologic monitoring after 5 years of first-line ART. 18th CROI, 27 February–2 March 2011, Boston. Poster abstract 677.
- Gilks C et al. A single CD4 Test with threshold >250 cells/mm3 can markedly reduce switching to second-line ART in African patients managed without CD4 or viral monitoring. 18th CROI, 27 February–2 March 2011, Boston. Poster abstract 676.