Mean Corpuscular Volume as a Marker for Adherence to Zidovudine-Containing Therapy in HIV-Infected Adults

OBJECTIVES To assess whether mean corpuscular volume (MCV) is useful in detecting non-adherence to AZTcontaining therapy. DESIGN Observational study within randomised controlled trial. METHODS We combined data from two treatment arms in SPARTAC, an RCT of short-course cART in primary HIV infection, classifying participants as responders (HIV-RNA decrease ≥1 log(10) or reaching <400copies/ml) or nonresponders following cART initiation. We assessed the sensitivity and specificity of using different percentage increases in MCV for accurately differentiating between responders and non-responders. We further examined changes in MCV levels up to 24 weeks after protocol-indicated cART cessation. RESULTS Of 119 participants included in this analysis, 73 (61%) were women, 71 of whom were randomised in Africa. Ninety-eight (88%) and 84 (85%) were classified as responders at 4 and 12 weeks respectively following cART initiation. MCV increased by a mean 3% and 1% at week 4, and 14% and <1% at 12 weeks for responders and non-responders. A 2% MCV increase at 4 weeks had 62% sensitivity and specificity for identifying virological response. At 12 weeks, an 8% increase had 89% sensitivity and specificity. In responders, MCV remained lower for individuals in African compared to non-African sites throughout and rose from 85 vs 90 fL at cART start to 96 vs 103 fL at 12 weeks post-initiation then fell to 88 vs 93 fL and 86 vs 89 fL at 12 and 48 weeks post-cessation. CONCLUSION In low-income countries, where HIV RNA may be unavailable, 12-weekly MCV measurements may be useful in monitoring adherence to AZT-containing regimens.