We read with great interest the article by Stamer et al.1 in a recent issue of the European Journal of Anaesthesiology concerning the role of genetic variants and the incidence of postoperative nausea and vomiting (PONV). Using a hypothesis-driven approach, 13 genetic variants as well as clinical factors were investigated. As a result, the 5-HTTLPR promoter polymorphism in the serotonin transport gene (rs 4795541) was, in addition to clinical factors, independently associated with PONV. We congratulate the authors for this study, which sheds light on the complex pathophysiology and genetics regarding the incidence of PONV.
Previous studies investigating genetic contribution to PONV, in particular, those with a hypothesis-driven candidate gene approach, have often constituted small patient cohorts. The rationale of this latter approach is based on the assumption that a genetic variation may affect gene function, which in turn might be relevant in the pathophysiology of PONV. The drawback of this study type is the bias evoked by multiple confounders,2 which hampers replication of candidate gene association studies in independent cohorts.
Using yet another approach, a genome-wide association study, Janicki et al.3 searched for variants underlying common disorders in scanning the whole genome, that is, free from assumptions for biological or positional candidate loci, genes, and variants. In doing so, they identified four candidate markers mediating susceptibility to PONV: a rs2165870 SNP in the M3 muscarinic acetylcholine receptor (CHRM3), a rs349358 single nucleoide polymorphism in the potassium voltage-gated channel subfamily B member 2 (KCNB2), a rs3218315 SNP in the interleukin-2 receptor subunit beta (IL2RB) and an intergenic rs703363 single nucleoide polymorphisms.
We performed a prospective, controlled study with 454 patients undergoing elective surgery (DRKS00005664, University of Duisburg-Essen Medical Faculty Ethics Committee no. 12-5065-BO), who had been genotyped for the latter four single nucleoide polymorphisms as well as for single nucleoide polymorphisms selected from CGAS.4,5 The CHRM3 rs2165870 single nucleoide polymorphisms and the KCNB rs349358 SNP were the only single nucleoide polymorphisms, which could be confirmed as independent PONV predictors. Due to limited data concerning the 5-HTTLPR promoter polymorphism (rs4795541) at that time, we did not investigate the single nucleoide polymorphisms. Encouraged by the results by Stamer et al.,1 and given the scientific need to always validate genetic association studies in an independent cohort, we decided to reanalyse our data set.4,5 Our study cohort included 466 patients (see flow chart in Klenke et al. 20184). Genotyping the rs4795541 single nucleoide polymorphism resulting in a 44 bp insertion/deletion was performed with PCR primers 5′-GGCGTTGCCGCTCTGAATGC-3′ and 5′-GAGGGACTGAGCTGGACAACCAC-3′ resulting in two PCR products representing the long (L) or short (S) alleles, respectively. In eight patients, genotyping was unsuccessful, so 458 patients were eventually genotyped. No post hoc power analysis was performed and no evidence for a significant deviation from the Hardy--Weinberg expectations was detected. Minor L-allele frequency was 0.43, in accordance with the literature and Stamer et al.1 We analysed by χ2-tests the association of the HTTLPR variant with nausea, retching/vomiting or nausea, and retching/vomiting (PONV) at different time intervals (Table 1).
Table 1 -
Association of rs4795541
with nausea, retching/vomiting, or nausea and retching/vomiting at different time intervals
||Nausea [n (%)]
||Retching/vomiting [n (%)]
||PONV [n (%)]
|0 to 2 h
|2 to 6 h
|6 to 24 h
|0 to 24 h
LL, LL genotype; SS, SS genotype; SL, SL genotype.
We failed to observe a significant association of the HTTLPR variant with PONV at any time interval and the P values obtained were far from the limits of statistical significance. Moreover, no differences in PONV incidences were detected in a recessive genetic model or for the application of the rescue medication granisetron. Accordingly, we were unable to replicate the findings by Stamer et al.1
Thus, these divergent results demonstrate that the genetic contribution to PONV risk is much more complicated than expected and even small differences in study design may result in different findings, which apply only for specific conditions. In Table 2, we note some differences between the respective studies,1,4,5 which may or may not explain divergent findings.
Table 2 -
Differences in design in studies by Stamer et al.
and Klenke et al
||Stamer et al.1
||Klenke et al.4,5
||Propofol, fentanyl, cis-atracurium
||Propofol or thiopentone or etomidate, fentanyl, rocuronium/succinylcholine
||Isoflurane in air, fentanyl, fentanyl and remifentanil, or remifentanil
||Isoflurane in nitrous oxide, fentanyl
||One to three antiemetics depending on patients’ risk(cohort A: dimenhydrinate, ondansetron and/or metoclopramide; cohort B: dexamethasone, ondansetron and/or metoclopramide
||Randomized to either no prophylaxis, dexamethasone, acustimulation, or dexamethasone and acustimulation
||Ondansetron, dimenhydrinate, droperidol, metoclopramide
|Definition of PONV
||Severe, intermediate or no PONV
|Time interval of significant association of PONV and genetic single nucleoide polymorphism
||0 to 24 h
||2 to 6 h
The studies by Stamer et al.1
and Klenke et al.4,5
have been compared. PONV, postoperative nausea and vomiting.
We would also like to point out that, conversely, a positive association of the CHRM3rs2165870 single nucleoide polymorphism3,4 was not seen in the study by Stamer et al.1
Iravani et al.6 pointed out that precision medicine is on the way to anaesthesia too. We believe that PONV is a good example for the possibility of incorporating precision medicine into anaesthesia. However, Iravani et al.6 also speculated that the immense complexity of the genotype/phenotype relationship and the multifaceted nature of confounding variables is to be elucidated before the predictive power of genomic medicine can be translated into clinical anaesthesia practice. The divergent findings of these studies underline the complexity of precision medicine and, therefore, also the complexity of genetic contribution to PONV risk.
Acknowledgements relating to this article
Assistance with the letter: none.
Financial support and sponsorship: none.
Conflicts of interest: none.
1. Stamer UM, Schmutz M, Wen T, et al. A serotonin transporter polymorphism is associated with postoperative nausea and vomiting: an observational study in two different patient cohorts. Eur J Anaesthesiol
2. Jorgensen TJ, Ruczinski I, Kessing B, et al. Hypothesis-driven candidate gene association studies: practical design and analytical considerations. Am J Epidemiol
3. Janicki PK, Vealey R, Liu J, et al. Genome-wide Association study using pooled DNA to identify candidate markers mediating susceptibility to postoperative nausea and vomiting. Anesthesiology
4. Klenke S, de Vries GJ, Schiefer L, et al. CHRM3 rs2165870 polymorphism is independently associated with postoperative nausea and vomiting, but combined prophylaxis is effective. Br J Anaesth
5. Klenke S, de Vries GJ, Schiefer L, et al. Genetic contribution to PONV risk. Anaesth Crit Care Pain Med
6. Iravani M, Lee LK, Cannesson M. Standardized care versus precision medicine in the perioperative setting: can point-of-care testing help bridge the gap? Anesth Analg