May 1, 2012
Investigators analysed the results of 16 clinical trials, using the Medication Event Monitoring System (MEMS) to monitor treatment compliance and to see which pattern of non-adherence had the greatest effect on viral load.
"Patterns of adherence have different impact on the risk of detectable HIV," comment the authors. "Missing days of medication consecutively may have a greater impact on being detectable than missing the same amount of time in a non-consecutive manner."
Adherence is the most important factor affecting the virological success of HIV therapy. The best results are seen in patients who take all or nearly all their doses of medication correctly and patients are encouraged to aim for near-perfect levels of adherence. Newer, more potent antiretrovirals appear to be more forgiving and seem to remain effective at somewhat lower levels of adherence.
However, it is unknown if particular patterns of non-adherence are associated with a greater risk of lack of viral suppression.
MEMS electronically records the opening of medication containers and is a useful tool for assessing patient adherence to therapy. The system is able to identify both individual missed doses and longer treatment interruptions.
Investigators from the US MACH14 study looked at the results of 16 studies using MEMS to see which of these two patterns of non-adherence involved the greater risk of poor suppression of HIV.
A total of 1088 patients were included in the analysis. The studies were conducted between 1997 and 2009.
Using MEMS, the investigators evaluated the impact of dose frequency and timing on viral load.
In the 28 days prior to each viral load measurement, the investigators estimated the proportion of time that drug levels were within therapeutic ranges. If MEMS recorded that a dose was taken three or more hours late this was considered a missed dose, or "non-covered" time. MEMS also enabled the investigators to record the length of treatment interruptions.
The categories for "covered-time" were: zero to 25%; 26 to 50%; 51 to 75%; 76 to 92%; and 93 to 100%. Duration of treatment interruptions were: zero to 48 hours (reference); 2 to 7 days; 7 to 14 days; 14 to 21 days; and 21 days or longer.
Most of the patients (69%) were men and their mean age was 40 years. Approximately a third were taking their first antiretroviral regimen. Half the patients were taking a combination of drugs based on an unboosted protease inhibitor, 14% were taking an NNRTI-based regimen, 8% a boosted protease inhibitor, and 27% were treated with other types of regimens.
A total of 3795 viral load measurements were available for analysis. Some 38% of these were detectable (above 400 copies/ml).
The mean amount of time for which patients had therapeutic levels of medication in their blood was 56%. The mean longest treatment interruption was seven days.
Compared to adherence levels of between 93 and 100%, a "covered time" of between zero and 25% was associated with a threefold increase in having a detectable viral load (OR = 3.22; 95% CI, 2.48-4.19). Adherence of between 26 and 50% was also associated with a significant increase in the risk of viral load being detectable (OR = 1.68; 95% CI 1.28-2.22). Adherence of less than 93% was associated with a non-significant trend towards an increased risk of detectability.
Treatment interruptions were associated with an even greater risk of viral load being detectable. The risk increased with the duration of the interruption and was highest for patients who stopped taking their treatment for 21 days or longer (OR = 3.65; 95% CI, 2.77-4.81). However, an interruption of between 7 and 14 days was associated with a doubling of the risk of a detectable viral load (OR = 2.06; 95% CI, 1.58-2.68).
"In this study there was a clear dose-relationship with each increasing week of interrupted time ... the risk was statistically significant starting at interruptions between 7 and 14 days," observe the authors. "This does not suggest that shorter interruptions are safe. Our data does not suggest a tolerable lower bound."
Occasional missed doses were more likely to lead to viral load becoming detectable for people taking a regimen based on an NNRTI than for those treated with a combination containing a boosted protease inhibitor.
The investigators then conducted further analysis which excluded the most adherent participants (above 95%), as well as those with the poorest adherence (below 5%). This showed that adherence at any level below 93%, as well as treatment interruptions lasting longer than 48 hours, were associated with a significant increase in the risk of a detectable viral load.
"Future research should focus on individual, interpersonal, and structural determinants of consecutive missed doses and the evaluation of interventions designed to improve adherence," suggest the authors. "Patient provider communication should focus on patterns of medication-taking and work towards shortening and eliminating interruptions in treatment."