By Liz Highleyman

It is well known that effective antiretroviral therapy dramatically reduces the risk of illness and death over the long term, but it is not clear when these benefits start to take effect.

At the 48th International Conference on Antimicrobial Agents and Chemotherapy (ICAAC) last week in Washington, DC, Stephen Berry and colleagues from Johns Hopkins University presented data on short-term hospitalization after starting HAART.

Between 1997 and 2005, the investigators analyzed 1327 hospitalizations of previously HAART-naive patients at the Johns Hopkins Hospital HIV clinic and affiliated sites in Baltimore and throughout Maryland within 1 year after starting combination therapy. Most patients (n = 965) were responders (defined as >1 log10 decrease in HIV RNA within 6 months), while 362 were non-responders. A majority of cohort participants were African-American, and blacks were significantly less likely than whites to be responders.

Results

• Through 45 days after HAART initiation, the mean hospitalization rate of responders was similar to that of non-responders (82 vs 80 events per 100 person-years [PY]).

• The hospitalization rate through 45 days was similar to the rate during the 6 months prior to starting therapy for both responders and non-responders.

• Between 45 and 90 days, the hospitalization rate for responders decreased significantly, to about 55 events per 100 PY (P < 0.05 vs baseline).

• During this period the hospitalization rate for non-responders also declined, but the decrease did not reach statistical significance.

• Type of HAART regimen -- NNRTI-based or protease inhibitor-based -- did not affect response rates at 45-90 days.

• Between 91 and 180 days, the hospitalization rate for responders settled near 40 events per 100 PY (P < 0.01 both for decrease from baseline and vs non-responders after 90 days).

• During this period, the hospitalization rate for responders was about half the pre-treatment rate, and less that half the rate of non-responders.

• The hospitalization rate for responders remained stable through 1 year; the rate for non-responders fell significantly by 1 year, but remained more than double that of responders.

• The most important predictor of hospitalization was low pre-treatment CD4 cell count (< 50 cells/mm3 or 50-199 cells/mm3).

• Women, blacks, and injection drug users (IDUs) were at higher risk of hospitalization.

• The decrease in hospitalization after 45 days was largely attributable to lower incidence of infections, including both classic opportunistic infections and non-AIDS-defining infections such as cellulitis and endocarditis

Based on these findings, the researchers concluded, "For 90 days after HAART initiation, virologic responders remain at comparable hospitalization risk to non-responders and therefore warrant close clinical surveillance."

"Further studies will need to evaluate causes of hospitalization in this high risk time period," they added.

"While we have known that laboratory numbers like CD4 count and HIV RNA level (viral load) improve within 14-28 days of starting HAART, we have not known the exact pattern of the change in risk of serious illness," Berry noted in a media statement released by the conference organizers. "Our findings suggest that for 45 to 90 days after HAART initiation, patients and HIV physicians should keep particularly close watch over signs and symptoms of illness and of infectious illness in particular."

Johns Hopkins Univ., Baltimore, MD.

11/04/08

Reference
SA Berry, KAGebo, RD Moore, and others. A High Risk of Hospitalization Immediately Follows HAART Initiation. 48th International Conference on Antimicrobial Agents and Chemotherapy (ICAAC 2008). Washington, DC. October 25-28, 2008. Abstract H-2292.

Other source
ICAAC. Risk of Illness Requiring Hospitalization Continues in the Short Term After Starting Medicines for HIV. Press release. October 27, 2008.