Background: Feasibility trials are preliminary trials that assess the viability and acceptability of intervention studies and the effects of the intervention on intermediate endpoints. Due to their short duration, they are unable to establish the effects of the intervention on long-term clinical outcomes. We propose a novel method that could transform the interpretation of feasibility trials using modified two-stage randomisation analyses. Methods In this two-stage process, we explored the effects of a 6-month feasibility factorial randomised controlled trial (RCT) of lycopene and green tea dietary interventions (ProDiet) on 159 serum metabolic traits in 133 men with raised PSA levels but prostate cancer (PCA) free. In the first stage, we conducted an intention-to-treat analysis, using linear regression to examine the effects of the interventions on metabolic traits, compared to the placebo group and instrumental variable analysis to assess the causal effect of the intervention on the outcomes. In the second stage, we used a two-sample Mendelian Randomization (MR) approach to assess the causal effect of metabolic traits altered by the interventions, on PCA risk, using summary statistics data from an international PCA consortium of 44,825 cancer cases and 27,904 controls. ResultsThe systemic effects of lycopene and green tea supplementation on serum metabolic profile were comparable to the effects of the respective dietary advice interventions (R2= 0.65 and 0.76 for lycopene and green tea respectively). Metabolites which were altered in response to lycopene supplementation were acetate (standard deviation difference versus placebo (β)): 0.69; 95% CI= 0.24, 1.15; p=0.003), valine (β: -0.62; -1.03, -0.02; p=0.004), pyruvate (β: -0.56; -0.95, -0.16; p=0.006), and docosahexaenoic acid (β: -0.50; -085, -0.14; p=0.006). The instrumental variable analysis showed there was no evidence that green tea altered the metabolome, but lycopene was associated with an increase in acetate (β=2.13; p=0.006) and decreases in pyruvate (β=-1.90; p=0.009), valine (β=-1.79; p=0.023), diacylglycerol (β=-1.81; p=0.026), alanine (β=-1.55; p=0.015) and DHA (p=0.097), where the regression coefficient represents the standard deviation (SD) difference in metabolite measures per unit change in lycopene (µmol/L) or EGCG (nM).Using MR, a genetically instrumented SD increase in pyruvate increased the odds of PCA by 1.29 (1.03, 1.62; p=0.027). Conclusion Using a two-stage randomization analysis in a feasibility RCT, we found that lycopene lowered levels of pyruvate, which our Mendelian randomization analysis suggests may be causally related to reduced PCA risk.
Bias in estimating the causal hazard ratio when using two-stage instrumental variable methods