We determined whether polymorphisms in Fcγ receptor (FcγR) IIa or FcγRIIIa genes were associated with outcomes in Vax004, a trial testing recombinant gp120 vaccination in preventing sexually acquired HIV infection. Male subjects (n = 1725), including infected and uninfected vaccinees and placebo recipients, were genotyped. We observed no association between FcγRIIa genotype and infection rate in vaccinees or placebo recipients. However, FcγRIIIa genotype was associated with infection rate among vaccinees (P = .035). Exploratory analyses revealed that vaccinees homozygous for the FcγRIIIa V allele in the lowest behavioral risk group had a greater rate of infection than low risk vaccinees with at least 1 F allele (hazard ratio [HR] = 3.52; P = .002). No such association was seen among vaccinees with high-risk behaviors or among placebo recipients in either risk stratum. Vaccinated low-risk VV subjects had a greater infection rate than low-risk VV placebo recipients (HR = 4.51; P = .17) or low-risk placebo recipients with any genotype (HR = 4.72; P = .002). Moreover, low-risk VV vaccinees had infection rates similar to individuals with high behavioral risk, irrespective of genotype. Our results generate the hypothesis that recombinant gp120 vaccine may have increased the likelihood of acquiring HIV infection in individuals with the VV genotype (present in ∼ 10% of the population) at low behavioral risk of infection.

Personalized medicine asks if a new treatment will help a particular patient, rather than if it improves the average response in a population. Without a causal model to distinguish these questions, interpretational mistakes arise. These mistakes are seen in an article by Demidenko [2016] that recommends the "D-value," which is the probability that a randomly chosen person from the new-treatment group has a higher value for the outcome than a randomly chosen person from the control-treatment group. The abstract states "The D-value has a clear interpretation as the proportion of patients who get worse after the treatment" with similar assertions appearing later. We show these statements are incorrect because they require assumptions about the potential outcomes which are neither testable in randomized experiments nor plausible in general. The D-value will not equal the proportion of patients who get worse after treatment if (as expected) those outcomes are correlated. Independence of potential outcomes is unrealistic and eliminates any personalized treatment effects; with dependence, the D-value can even imply treatment is better than control even though most patients are harmed by the treatment. Thus, D-values are misleading for personalized medicine. To prevent misunderstandings, we advise incorporating causal models into basic statistics education.