A prospectively measured genotype does not automatically become a clinically useful treatment-selection test. The design must distinguish a general prognostic difference from a difference in the effect of the treatment itself.
What comparison did the pilot actually make?
Pal and colleagues’ 2015 open-label pilot enrolled 11 methamphetamine-dependent participants with the OPRM1 A118G variant and 11 with A118A. All received one 380-mg intramuscular sustained-release naltrexone injection and weekly psychotherapy during a four-week treatment phase. Staff with participant contact and participants were blinded to genotype.
Mean methamphetamine-negative urine counts were 1.7 versus 1.8, respectively (p=0.97); reported abstinent days were 11.5 versus 14.8 (p=0.51). The urine screen cutoff was 1,000 ng/mL. Other abstinence comparisons also were nonsignificant. The authors emphasized the small sample and absence of placebo groups.
Selection was not random assignment to genotype or drug
The investigators prospectively screened genotypes and selected a comparison group balanced on sex and recent-use frequency. That procedure helped form comparable groups on selected characteristics. It did not randomly assign the genetic variant, and everyone received the same active medication.
To establish that a variant selects treatment benefit, the important question is whether the treatment-versus-comparator effect differs by genotype. Comparing two genotype groups after both receive treatment does not supply that interaction. Even a positive difference could reflect prognosis, other group characteristics or the treatment response; the design would need to distinguish those explanations.
The null result does not validate interchangeability either. With 11 participants per group, a failure to find a difference is not the same as demonstrating sufficiently similar effects within a prespecified margin.
Why another substance’s result cannot fill the gap
A pharmacogenetic finding from treatment of alcohol problems would motivate a methamphetamine study, rather than establish its answer. The relevant outcome, population and treatment context must be tested. A shared medication name does not make two disorders the same experimental setting.
For a client assessing a proposed genetic opinion, the practical question is whether the claimed prediction has been validated for the intended decision. That includes the alternative treatment, the target population and the outcome being predicted. A small exploratory group comparison cannot supply all three by implication.
This historical pilot therefore leaves genotype-guided naltrexone treatment for methamphetamine dependence unresolved. It supports neither a validated individual forecast nor a conclusion about medication efficacy relative to placebo, which was not part of the experiment.
Source
Pal R, et al. Impact of prospectively determined A118G polymorphism on treatment response to injectable naltrexone among methamphetamine-dependent patients: an open-label, pilot study. J Addict Med. 2015;9:130–135. doi:10.1097/ADM.0000000000000107. Full text.