Addendum To: Study protocol of the HD-MED study aiming to personalize drug treatment in Huntington’s disease: a longitudinal,observational study to assess medication use and efficacy in relation to pharmacogenetics

From the authors: As the HD-med study progressed, we realized that the current statistical methods in our article ¹ were outdated. In the analysis, we will no longer combine the extremes of the pharmacogenetic profile (poor metabolizers (PMs) and ultra-rapid metabolizers (UMs)). Furthermore, we will no longer determine negative medication effects and definite prescription changes using the primary outcome definitions given in the original manuscript ¹ . Instead, we decided to proceed with a simplified data preprocessing approach, as described below.

With this addendum paper, we provide transparency on our updated methodology. Currently, recruitment and data collection are ongoing. We have not yet seen the final data. No changes in data collection procedures have been made. The updated approach has no relevant impact on the required sample size.

PMs are at risk of experiencing more side effects, while UMs are at risk of experiencing no or less effect of a given drug. By grouping PMs and UMs in the same category of “divergent phenotype,” the true effect might be averaged out. Therefore, we thought it would be better not to combine these groups and just record the CYP2C19 and CYP2D6 genotype-predicted phenotype (PM, intermediate metabolizer (IM), normal metabolizer (NM) or UM). Depending on the actual number of subjects in each group, we may combine IMs and NMs in the analysis to increase power.

In addition, actionable drugs are clustered in pharmacogenetic groups, either CYP2C19 or CYP2D6. Prodrugs are examined separately. We will count how often adverse drug reactions and lack of efficacy occur, compared to the control group (=no adverse drug reactions nor lack of efficacy) per genotype-predicted phenotype for each pharmacogenetic group. Results will be analyzed with ANalysis Of VAriance (ANOVA) or its non-parametric counterpart, depending on the distribution of the data. Causality of adverse events is rated by the Liverpool Adverse Drug Reaction Causality Assessment Tool ² by two independent investigators. Only possible, probable, and definite causality assessments are included in the analysis. Disagreements are resolved by consensus. Lack of efficacy is reported according to Figure 2 of the original manuscript, ¹ where each “negative medication effects” is replaced by “lack of efficacy.” The “Drug change due to side effects?” box in Figure 2 of the original manuscript ¹ has therefore become redundant. The new standardized operating procedure to determine the lack of efficacy is included in this addendum (Figure 1). Drug–drug interactions and corresponding co-medication-induced phenoconversion are taken into account when present, according to the previously reported predictions in Table 2 of the original manuscript. ¹

Figure 1.

Standardized operating procedure to determine lack of efficacy. To complete separately for actionable drug–gene interactions in CYP2C19 and CYP2D6.

The primary objective and main outcome are redefined to reflect these changes.

Primary objective: To classify the effect of the pharmacogenetic profile of CYP2C19 and CYP2D6 in Huntington’s disease gene expansion carriers.

Main outcome: Number of adverse events and lack of efficacy per CYP2C19 and CYP2D6 genotype-predicted phenotype.