icon-folder.gif   Conference Reports for NATAP  
  52nd ICAAC Interscience Conference on
Antimicrobial Agents and Chemotherapy
September 9-12, 2012, San Francisco
Back grey_arrow_rt.gif
Partial or No Antiretroviral Adherence Raises
Chance of Hospital Stay 50% in US Study

  52nd ICAAC, September 9-12, 2012, San Francisco

Mark Mascolini

Nonadherence or partial adherence (taking only part of an antiretroviral regimen) raised by half the chance of hospital admission in a 4500-person US study [1]. Adherence was consistently better with single-tablet regimens .

Combining three antiretrovirals in a single pill improves overall adherence and eliminates partial adherence. Three single-tablet regimens are now licensed in the United States-Atripla (efavirenz with tenofovir/emtricitabine, TDF/FTC), Complera (rilpivirine plus TDF/FTC), and Stribild (elvitegravir/cobicistat plus TDF/FTC). Many studies assess overall antiretroviral adherence, but none assessed partial adherence with single-tablet or multipill regimens until this analysis. The researchers defined multipill regimens as 2-tablet-daily regimens nucleosides plus a boosted PI, a nonnucleoside, or the integrase inhibitor raltegravir.

To assess the impact of single-tablet regimens on adherence and hospital admission, researchers tapped LifeLink healthcare claims from a privately insured US population with HIV infection between June 2009 and December 2011. Using pharmacy refill records, the investigators assessed adherence as the percent of time taking (1) a complete regimen, (2) a partial regimen (defined as taking some but not all components of a regimen), or (3) no antiretrovirals (complete nonadherence). Follow-up began when antiretroviral therapy began and continued until (1) the entire regimen was stopped, (2) the key component of the regimen (a protease inhibitor [PI], nonnucleoside, or integrase inhibitor) changed from one class to another, or (3) the last point in the record.

Overall 7825 people received a complete antiretroviral regimen during the study period and 4549 had continuous insurance benefit eligibility and stayed on their regimen for at least 90 days. There were 1751 people taking a single-tablet regimen, 522 taking a multipill raltegravir combination, 1601 taking a multipill boosted PI combination, and 675 taking a multipill nonnucleoside combination.

Average age ranged from 44.4 in the single-tablet group to 48.0 in the nonnucleoside group. About 80% of study participants were men. A somewhat higher proportion of people taking a single-tablet regimen (26.7%) were antiretroviral naive when they started that combination, compared with 16.3% starting raltegravir, 13.9% starting a boosted PI, and 7.6% starting a nonnucleoside. Mental disorder rates for the groups taking a single tablet, raltegravir, a PI, or a nonnucleoside were 17.8%, 25.1%, 19.4%, and 16.7%. Respective rates of drug or alcohol abuse were 7.3%, 10.2%, 10.1%, and 5.0%.

People taking a single-tablet regimen received a complete regimen during 90% of study days, versus 78.9% for raltegravir regimens, 80.4% for PI regimens, and 85% for nonnucleoside regimens (all P < 0.001 versus single-tablet regimens) . Percentage of days with no antiretrovirals available was similar for each regimen: 10% for single-tablet regimens, 10% for raltegravir regimens, 11.8% for PI regimens, and 8.3% for nonnucleoside regimens. Partial adherence rates were 11.1% for raltegravir regimens, 7.9% for PI regimens, 6.6% for nonnucleoside regimens, and of course 0% for single-tablet regimens.

Both partial and complete nonadherence raised the odds of hospital admission, and the longer the poor adherence, the greater the odds:

Hospital admission odds with partial adherence (versus under 20 days):

20 to 40 days: odds ratio (OR) 1.497, P = 0.0183

40 to 60 days: OR 1.687, P = 0.0215

More than 60 days: OR 1.574, P = 0.0009

Hospital admission odds with complete nonadherence (versus under 20 days):

20 to 40 days: OR 1.481, P = 0.0025

40 to 60 days: OR 1.492, P = 0.0078

More than 60 days: OR 1.946, P < 0.0001

For all three types of multipill regimen, complete plus partial nonadherence independently raised the odds of hospital admission compared with complete nonadherence alone. Being antiretroviral naive when entering the study period raised the odds of hospital admission by half (OR 1.512, P = 0.0009) .

The researchers believe their findings "support the use of single-tablet regimens to prevent the occurrence of partial adherence, and suggest a potential approach to prevent the adverse consequences associated with partial adherence." The investigators issued the usual caveat that a nonrandomized comparison cannot eliminate the possibility that unmeasured variables may have affected outcomes. They cautioned that they had no lab results to assess the impact of different adherence patterns on treatment response.


1. Cohen C, Davis K, Meyers J. Association between selective adherence to antiretroviral therapy and hospitalization risk in an HIV population. 52nd Interscience Conference on Antimicrobials and Chemotherapy (ICAAC). September 9-12, 2012. San Francisco. Abstract H-211.