Analyses suggested that overall, baseline blood glucose levels were significantly and positively associated with PHDD during treatment [F (1, 1225) = 5.21, p = .023], after controlling for baseline PHDD [F (1, 1225) = 36.25, p < .0001], gender [F (1, 1225) = 3.33, p = .07], and BMI [F (1, 1225) = 0.31, p = .58]. As expected, there was also a main effect of time in treatment, such that PHDD increased across the trial [F (3, 1225) = 2.90, p = .049] suggesting decreased treatment effects at longer follow-ups, however, there was no significant Glucose ×Time interaction [F (3, 1225) = 2.15, p = .09]. The associations between baseline blood glucose level and PHDD across treatment and follow-up, captured by Pearson Product Moment Correlations, were as follows: r = .082 (p = .002) at week 4, r = .067 (p = 0.014) at week 8, r = .042 (p = .13) at week 12, and r = .034 (p = .22) at week 16. Overall, these results suggested that blood glucose level at baseline was positively associated with PHDD during the course of treatment and that this relationship reached statistical significance at weeks 4 and 8.
