Incidence of adverse events at 4 weeks correlates with adherence and virologic failure at 4 months

Simon Collins, HIV i-Base

Over one hundred presentations focused on adherence, inevitably overlapping with conclusions already reported in previous research.

Nevertheless, all of these presentations were important and clinicians and investigators need to address the issue of integrating adherence support within every clinical care and trial setting.

An interesting presentation from the French APROCO study found that incidence of side-effects four weeks into a new therapy directly correlated with successful adherence 3 months later. 336 patients who received triple combination protease-based regimens (1 x PI + 2 x RTIs), and remained on the same treatment for more the first four months, self-completed a detailed questionnaire at months 1 and 4. This included a table listing 13 symptoms with space to include frequency, severity and related distress, as well as a section with more general questions. Adherence was defined as taking all prescribed pills over the four days prior to the clinic visit.

This careful approach to adverse effects produced a very accurate picture of the reality experienced by patients – 94% reported at least one symptom at M1, and 88% at M4. Fatigue and diarrhoea were the most frequently reported, 41% registered as mild and 7% as severe. A high median number of symptoms was 4 (IQR 2-6) at M1 and 3 (IQR 3-5) at M4 (the decrease over time was statistically significant with p < 0.001). There was also a general decrease in incidence as well as severity in all side-effects (except rash) between M1 and M4. Respectively, 81% and 75% were adherent at M1 and M4 (p=0.03). 66% patients were assessed as adherent at both visits and 10% were non-adherent at both visits.

The percentage of patients with undetectable viral load at M4 (< 500 copies/ml) was also significantly lower among non-adherent patients (81% vs 70%; p = 0.03). Median decrease in viral load titres before and after initiation of HAART (M0 vs M4) was significantly lower among M4 non-adherent (1.6 log vs 1.1 log, p = 0.008). The number of self-reported symptoms was significantly associated with non-adherence at M4 – patients who were non-adherent at M4 had previously declared a higher number of symptoms at M1 (p = 0.006).

The relationship between patients reporting > 4 symptoms at M1 and non-adherence at M4 was also highly significant (OR IC95% = 1.93, 1.11-3.36). Adjustment was made in this analysis for other related factors such as history of previous treatment, younger age, unstable housing, poor social support, alcohol use etc. The decrease in the number of symptoms was only significant among the 252 adherent patients, p (ADH) <0.001 ; p (non-ADH) = 0.35.

Early identification of any patients experiencing side-effects, provides an opportunity to actively help manage these events, as well as incidentally to address adherence issues. Early identification of patients whose treatment is particularly difficult (>4 symptoms) and individual management of symptoms whether by concomitant medicines (anti-diarrhoeal, anti nausea etc) or the opportunity to switch to more tolerable ARVs, provides an opportunity to both improve levels of adherence and treatment outcome.


  1. Spire B, Duran S, Raffi F et al – A high number of self-reported symptoms following HAART initiation is predictive of poor adherence at 4 months of treatment in HIV-infected patients. XIII International AIDS Conference, Durban, July 9-14, 2000. Abstract TuOrD332.

Links to other websites are current at date of posting but not maintained.