Micronutrient deficiency and treatment adherence in a randomized controlled trial of micronutrient supplementation in ART-naïve persons with HIV

Abstract

Introduction: The MAINTAIN study is an on-going RCT comparing high-dose micronutrient and anti-oxidant supplementation versus recommended daily allowance (RDA) vitamins in slowing HIV immune deficiency progression in ART-naïve people with HIV infection.

Objective: We planned analysis of the first 127 participants to determine the baseline prevalence of serum micronutrient deficiencies and correlates, as well as tolerance and adherence to study interventions.

Methods: Participants receive eight capsules twice daily of 1) high-dose or 2) RDA supplements for two years and are followed-up quarterly for measures of immune deficiency progression, safety and tolerability. Regression analysis was used to identify correlates of micronutrient levels at baseline. Adherence was measured by residual pill count, self-report using the General Treatment Scale (GTS) and short-term recall HIV Adherence Treatment Scale (HATS).

Results: Prior micronutrient supplementation (within 30 days) was 27% at screening and 10% of study population, and was not correlated with baseline micronutrient levels. Low levels were frequent for carotene (24%<1 nmol/L), vitamin D (24%<40 nmol/L) and serum folate (20%<15 nmol/L). The proportion with B12 deficiency (<133 pmol/L) was 2.4%. Lower baseline levels of B12 correlated lower baseline CD4 count (r = 0.21, p = 0.02) with a 21 pmol/L reduction in B12 per 100 cells/µL CD4. Vitamin D levels were higher in men (p<0.001). After a median follow-up of 1.63 years, there were 19 (15%) early withdrawals from the study treatment. Mean treatment adherence using pill count was 88%. Subjective adherence by the GTS was 81% and was moderately but significantly correlated with pill count (r = 0.29, p<0.001). Adherence based on short-term recall (HATS) was >80% in 75% of participants.

Conclusion: Micronutrient levels in asymptomatic HIV+ persons are in keeping with population norms, but micronutrient deficiencies are frequent. Adherence levels are high, and will permit a valid evaluation of treatment effects.

Trial registration: ClinicalTrials.gov NCT00798772.

Figure 1. Average treatment adherence by returned capsule count, and with multiple imputation for non-returns.

Adherence to study medication over time was assessed by counting capsule packets returned at weeks 12, 24, 36, 48, 60, 72 and 84. Treatment adherence was calculated and compared using different methods. When only those who attended visits and returned capsule packets were included in analysis, average adherence was 88%. Intention-to-treat analysis, with 26% imputation for missing returns, yielded three estimates: 66% (0% imputed for missing values), 91% (100% imputed for missing values), and 88% (average adherence of returns imputed for missing values).

Figure 2

A. CD4 count (cells/mL) vs. serum micronutrients. The B₁₂ threshold of 133 pmol/L demonstrates that 3 (2.4%) of baseline measures are below the lower limit of the normal range. The slope (21 pmol/L per 100 cells/µL) indicates the relationship between CD4 lymphopenia and serum B₁₂ levels, which are largely within the normal range for B₁₂. B. Plasma virus load (log10 copies/mL) vs. serum micronutrients. The 25-OH D thresholds indicate that 67%, 24% and 3.5% of baseline measures indicate insufficiency (<75 nmol/L), deficiency (<40 nmol/L) and severe deficiency (<20 nmol/L), respectively. Higher baseline levels of vitamin D were more common in men (p<0.001) and correlated higher baseline plasma HIV-RNA level (p = 0.02). All other demographic and baseline clinical factors were not significantly associated with baseline micronutrient levels.

Figure 3. Treatment adherence by residual pill count over time.

The data points and bars indicate mean percentage and standard deviation of the expected unreturned medication capsules on the Y-axis. The X-axis indicates the study follow-up visit week. The number of subjects included in analysis is included under the X-axis. Imputation of average adherence was used for missing values.