To evaluate the relationship between genetic markers of placebo response and treatment response to percutaneous tibial nerve stimulation (PTNS) or sham for treatment of fecal incontinence (FI). The hypothesis was that there will be a linear relationship between  the number of Catechol-O-Methyltransferase (COMT) met alleles and change from baseline in St Mark’s score.

Blood specimens were collected from a subset of women (N=96) with FI in the Neuromodulation for Accidental Bowel Leakage trial (NOTABLe, NCT 032786). Samples of participants who completed at least 10/12 active PTNS or sham treatments were analyzed. DNA was extracted from whole blood and genotyping was performed on target single nucleotide polymorphisms (SNPs) previously associated with placebo response: rs4680 (Val158Met) in catechol-O-methyltransferase (COMT); rs4570625 (G-703 T) upstream of tryptophan hydroxylase-2 (TPH2); rs6265 (Val66Met) in brain derived neurotropic factor (BDNF); rs324420 (Pro129Thr) in fatty acid amide hydrolase (FAAH); rs510769 (intronic) and rs1799971 (Asn40Asp) in mu-opiod receptors (OPRM1). The primary outcome was improvement in St. Mark’s score after 10-12 weeks of either active or sham treatment. Secondary outcomes were improvement in number of fecal incontinence episodes (FIEs), dichotomized Patient Global Impression of Improvement (PGI-I) (1-2 vs >2), and responder status (dichotomized by a St. Mark’s improvement threshold of 4). An additive linear regression interaction model adjusted for BMI, race, baseline FIEs or St. Mark’s score was used to identify SNP by treatment effects significantly associated with primary and secondary accounts. In the absence of treatment interaction, SNP main effects were tested.

There were no differences in age, BMI, FIEs or St. Mark’s score at baseline or follow up between PTNS (n=64) and sham (n=32) treated patients. Lower BMI was significantly associated with improvement in St. Mark’s score in the PTNS group (p=0.01). There was a significant interaction between the COMT SNP (Val158Met) and treatment group (PTNS vs sham) for change in St. Mark’s score (interaction p=0.02), change in FIEs (interaction p=0.01) and near-significant for PGI-I (interaction p=0.06), (Table). Val158Met was associated with better response to treatment in PTNS treated patients (St. Mark’s score improvement: beta=1.72; p=0.06, FIE improvement: beta=2.36; p=0.001, and PGI-I: OR=2.00; p=0.06) but was not associated with treatment response in sham (St. Mark’s score improvement: beta=-1.37; p=0.24, FIE improvement: beta=0.09; p=0.95, and PGI-I: OR=0.64; p=0.48). PTNS participants homozygous for the Met allele (Met/Met) had the greatest improvement of response, Val/Met heterozygotes had an intermediate response, and Val allele homozygotes (Val/Val) had the lowest improvement in St. Mark’s score and FIEs (Figure). No significant interaction or main effects were observed for TPH2, BDNF, FAAH, or OPRM1 SNPs.

As this secondary analysis was designed, it was thought that responders to sham PTNS treatment may have a higher association with genetic markers of placebo response. In this trial, women with FI homozygous for the COMT Met allele (Met/Met) were more likely to have a better response to active PTNS treatment, followed by Val/Met heterozygotes. These results suggest there may be an unexplored physiologic relationship between the increased dopamine availability conferred by the COMT Met allele and PTNS treatment response.

Although it was thought that the homozygous COMT Met allele may be associated with response in the sham treatment group, women with fecal incontinence homozygous for the COMT Met allele had a greater response to active PTNS treatment. These data should be validated in women undergoing other neuromodulation modalities for FI, as well as PTNS and sacral neuromodulation in women with overactive bladder symptoms.