Cardiovascular diseases remain the leading cause of mortality worldwide, and pharmacological treatment strategies are aimed at reducing the risk of these conditions. However, there is considerable interindividual variability in drug response, both in terms of efficacy and safety [1].

Cardiac surgery patients are characterized by inherently high cardiovascular risk, largely attributable to multimorbidity and polypharmacy [2–10]. In the perioperative period, they receive multicomponent pharmacotherapy, including antithrombotic agents (anticoagulants and antiplatelet drugs), statins, antiarrhythmic agents, proton pump inhibitors, analgesics, and often antibacterial therapy.

Several of these drug classes are notable for the high clinical significance of precise dosing and monitoring due to narrow therapeutic ranges, pronounced variability in clinical effect, and substantial potential for drug–drug interactions. Consequently, even minor deviations in therapy efficacy or safety can disrupt the hemostasis–thrombosis balance and be associated with critical perioperative outcomes, primarily bleeding and thrombotic complications (myocardial infarction, stroke, thrombosis) [3, 5–11].

Additional sources of variability arise directly during the surgical intervention and in the early postoperative period: systemic inflammatory response (SIR), hemodilution, alterations in plasma protein composition, and organ dysfunction. Cardiopulmonary bypass has been shown to substantially alter drug pharmacokinetics and pharmacodynamics, potentially leading to either subtherapeutic or supratherapeutic drug exposure [11].

Against this background, standardized dosing regimens do not consistently ensure predictable clinical effects or acceptable safety profiles, justifying the need for personalized approaches to therapy selection and adjustment, including monitoring of efficacy and safety, and, where available, therapeutic drug monitoring.

A narrative review by Kayani et al. (2025) emphasizes that genetic variability can influence metabolism, efficacy, and adverse reaction risk for major cardiovascular drug classes (antiplatelet agents, anticoagulants, statins, calcium channel blockers, etc.), explaining part of the variability in treatment outcomes and supporting personalized therapeutic approaches in high-risk patients, including those under surgical stress and polypharmacy conditions [12].

Clinical determinants of interindividual variability in drug response are known to include age, body weight, liver and kidney function, severity of inflammatory response, nutritional status and vitamin K intake (for warfarin), comorbidities, drug interactions, and treatment adherence [12–15]. Of particular importance in cardiac surgery practice is polypharmacy and the use of drugs with narrow therapeutic indices, where even moderate changes in pharmacokinetics and pharmacodynamics can lead to clinically significant complications. Organizational parameters also play a substantial role: the frequency of monitoring efficacy and safety indicators (e.g., INR during warfarin therapy), availability of laboratory tests, dose adjustment protocols, and continuity between inpatient and outpatient phases of patient management.

Table 1 summarizes the major groups of factors contributing to high interindividual variability in treatment response in cardiac surgery patients.

Table 1. Factors influencing response variability in cardiac surgery patients

Factor groupSpecific determinantsImpact on therapyClinical-demographicAge, body weight, sex, body surface areaChanges in volume of distribution, metabolic clearance, receptor sensitivityComorbidityDiabetes mellitus, chronic kidney disease, liver disease, hypothyroidismAltered pharmacokinetics, increased toxicity risk (statins), reduced efficacy (clopidogrel)Perioperative stressCardiopulmonary bypass, systemic inflammatory response, hemodilution, organ dysfunctionChanges in protein binding, clearance, drug exposureConcomitant therapyCYP inhibitors/inducers (amiodarone, proton pump inhibitors, rifampicin)Altered active metabolite concentrations, phenoconversionGenetic factorsCYP2C9, VKORC1, CYP2C19, SLCO1B1 polymorphismsChanges in metabolism, transport, and drug sensitivity

Thus, response variability results from complex interactions between clinical and genetic factors. Additional complexity is introduced by the phenomenon of phenoconversion — a situation in which genotypically "extensive" metabolizers phenotypically exhibit characteristics of slow metabolism under the influence of external factors, primarily concomitant therapy [16]. This underscores the need for comprehensive assessment in therapy personalization.

Despite the significant contribution of clinical factors, they explain only part of the interindividual variability in drug response. In several cases, after accounting for age, body weight, organ function, and concomitant therapy, substantial residual variability persists. This has prompted active investigation into the genetic determinants of drug response.

In clinical pharmacology, drug–gene interactions are distinguished — situations in which a patient's genetic variants modify the efficacy or safety of standard pharmacotherapy — as well as drug–drug–gene interactions, where drug interactions and patient genotype mutually potentiate their effects on drug exposure and clinical response. An example of such interaction is shown in the study by Sychev D.A. et al. (2023), where the combination of rivaroxaban with verapamil (a CYP3A4/P-gp inhibitor) was associated with higher drug concentrations and increased bleeding frequency in elderly patients [17].

In the work by Litonius et al. (2024), based on real-world drug prescription data analysis, it was shown that a substantial proportion of patients receive drugs whose efficacy and safety are potentially modified by pharmacogenetic variants. The most clinically significant interactions were associated with widely used cardiovascular drugs, including antithrombotic agents, statins, and proton pump inhibitors [18].

In a Russian cohort of patients with high thrombotic risk, it was demonstrated that CYP2C9 (*2/*3), VKORC1 (–1639G>A), and CYP4F2 (rs2108622) variants occur in a significant proportion of patients, with the frequency profile being generally comparable to European populations [19]. For antiplatelet therapy, the carriage rate of CYP2C19 loss-of-function alleles (*2/*3*) in different regions of Russia ranged from 16–27.5%, while the gain-of-function allele CYP2C1917 was found in 15.4–33.3% [20].

In Russia, the practical application of pharmacogenetic testing followed by warfarin initial dose selection based on results has been described in the study by Panchenko E.P. et al. (2020). This prospective multicenter randomized study conducted in the Russian Federation included 263 patients who had not previously received warfarin and had indications for long-term anticoagulation. Patients were randomized into two groups: the pharmacogenetic group (loading warfarin dose calculated using the Gage algorithm based on CYP2C9 and VKORC1 genotyping, with dose adjustment from day 5 according to INR) and the standard group (initial warfarin dose 5 mg with titration from day 3 according to INR). Results of the pharmacogenetic approach demonstrated the absence of major bleeding in the pharmacogenetic group compared to 6 cases in the standard group (p=0.031); reduced time to therapeutic INR: 11 days vs. 17 days (p=0.046); and reduced frequency of INR fluctuations ≥4.0: 11% vs. 30.9% (p=0.002) [21].

Pharmacogenetic dosing proved particularly effective in patients with increased sensitivity to warfarin.

Notably, the "cumulative" effect of drug–gene interactions may have measurable clinical consequences in surgical patients. In a retrospective cohort study of the Million Veteran Program (n=10,098), the presence of ≥2 clinically significant drug–gene interaction events in the perioperative period was associated with increased length of hospitalization, higher 30-day readmission rates, and increased risk of adverse cardiovascular outcomes [22].

Contribution of clinical and genetic factors to warfarin dose variability. Warfarin remains the primary oral anticoagulant for patients following cardiac surgery, including patients with mechanical prosthetic heart valves [3, 4, 22, 23]. However, its use is complicated by a narrow therapeutic window and considerable interindividual dose variability [23, 24].

Warfarin pharmacogenetics is one of the most extensively studied models of personalized therapy. Genetic factors collectively account for up to 50–60% of interindividual dose requirement variability. The greatest contributions are made by VKORC1 polymorphisms (≈25–30%) and CYP2C9 (≈10–15%). In patients after valvular cardiac surgery, addition of CYP2C9 and VKORC1 genotypes increases dose prediction accuracy from 32% (clinical factors only) to 43% [24–29].

Pharmacogenetics of antiplatelet therapy. Clopidogrel is a prodrug requiring CYP2C19-dependent biotransformation for active metabolite formation. Carriage of CYP2C19 loss-of-function alleles (*2, *3*) is associated with reduced antiplatelet effect and increased risk of ischemic complications [28]. CPIC guidelines (2022) recommend considering CYP2C19 phenotype when selecting a P2Y12 inhibitor and, when indicated, considering alternative agents (ticagrelor, prasugrel) in intermediate and poor metabolizers [9, 10, 29].

Pharmacogenetics of lipid-lowering therapy: statins. Statins remain a key component of secondary cardiovascular prevention and are often continued in the perioperative period in cardiac surgery patients [7, 10]. The most studied pharmacogenetic risk factor for statin-associated musculoskeletal symptoms (SAMS) is the SLCO1B1 c.521T>C (p.Val174Ala, rs4149056) variant, which reduces hepatic OATP1B1 transporter activity, decreases hepatic clearance, and increases systemic statin exposure, particularly for simvastatin [30].

CPIC guidelines (2022) for statin-associated musculoskeletal symptoms recommend the following approach [31]:

In cardiac surgery, the clinical significance of such variants may be potentiated by combination with drug interactions (e.g., verapamil, amiodarone) and metabolic stress, necessitating individualized drug selection and dosing. In the Million Veteran Program surgical cohort, the "statin–SLCO1B1" combination was among the most frequent drug–gene interactions [22].

For a concise summary of clinically significant pharmacogenetic markers in cardiac surgery, Table 2 is presented below.

Table 2. Clinically significant pharmacogenetic markers in cardiac surgery

GeneDrugEffect on pharmacokinetics/pharmacodynamicsClinical outcomeRecommendation (CPIC/DPWG)VKORC1WarfarinReduced target enzyme activity↑ warfarin sensitivity, ↓ dose requirementReduce initial and maintenance doseCYP2C9Warfarin↓ S-warfarin metabolism↑ exposure, ↑ bleeding riskReduce initial and maintenance doseCYP4F2WarfarinImpaired vitamin K metabolism↑ dose requirement (minor)Consider in combined algorithmsCYP2C19Clopidogrel↓ Active metabolite formation↑ risk of ischemic events (stent thrombosis, MI)Use alternative agents (ticagrelor, prasugrel) in PM/IMSLCO1B1Simvastatin, atorvastatin↓ Hepatic uptake↑ systemic exposure, ↑ myopathy/rhabdomyolysis riskAvoid high-dose simvastatin; prefer pravastatin, rosuvastatin (with ABCG2 adjustment)ABCG2Rosuvastatin↓ Intestinal and hepatic efflux↑ systemic exposure, ↑ myopathy riskReduce initial rosuvastatin doseCYP2C9Fluvastatin, pitavastatin↓ Metabolism↑ exposure, ↑ myopathy riskConsider dose reduction or alternative statin

Notes: PM — poor metabolizers, IM — intermediate metabolizers.

Scientific and clinical limitations of the evidence base. Despite compelling pharmacogenetic rationale, results from clinical studies remain heterogeneous. Meta-analyses demonstrate variability in the effect of pharmacogenetically guided algorithms on time in therapeutic range (TTR) and complication rates [32, 33]. Genotype-guided strategies predominantly improve early anticoagulation quality metrics (time to therapeutic INR, TTR), whereas their impact on "hard" clinical outcomes (bleeding, thromboembolism) remains less consistent and depends on study design and monitoring intensity [33–41].

Similar limitations are observed for other drug classes. For clopidogrel, the impact of genotype-guided strategies on clinical outcomes remains inconclusive [28, 29], and for statins, the clinical effectiveness of routine pharmacogenetic testing in reducing SAMS frequency continues to be debated [30–32]. Currently, in the Russian population, there are no prospective randomized studies evaluating the effectiveness of genotype-guided initial dosing for antithrombotic agents (clopidogrel) or statins, which particularly underscores the importance of generating local data and integrating pharmacogenetically guided dosing into clinical practice [17, 20, 21, 41].

Population limitations. The effectiveness of pharmacogenetic algorithms varies substantially depending on the ethnic composition of study populations. Most validated algorithms have been developed predominantly in European and Asian cohorts, limiting their generalizability to other populations. In patients of African ancestry, additional variants (e.g., CYP2C95, *6, *8, *11) are clinically significant and are not always included in standard algorithms [37]. Universal application of existing algorithms without consideration of population-specific features may reduce dose prediction accuracy [26].

Implementation barriers. The implementation of pharmacogenetic testing into clinical practice faces several barriers. As detailed in works [38, 39], these include:

A survey of Russian physicians demonstrated generally positive attitudes toward testing alongside low levels of actual implementation, primarily due to these barriers [40]. In this regard, additional studies aimed at developing and implementing practice-oriented algorithms in the domestic context appear warranted.

Cardiac surgery patients are characterized by high medication burden and rapid changes in clinical status; therefore, focusing on a single gene or single drug may be insufficient [1–13]. Data from a large surgical cohort show that the presence of ≥2 drug–gene interactions is associated with worse postoperative outcomes [22], confirming the clinical significance of multilocus pharmacogenetic risk.

In light of the above, the development and evaluation of a comprehensive pharmacogenetic algorithm for cardiac surgery patients appears promising. Unlike existing approaches focused on individual drugs, the proposed algorithm would include:

Such a multiparametric approach would not only improve dosing precision and reduce drug-related complications but also establish a foundation for comprehensive pharmacotherapy personalization, which is most appropriate for the cardiac surgery patient profile as a high pharmacological risk group.

It should be noted that this narrative review has several methodological limitations. First, the literature search was not systematic, which may have led to preferential inclusion of studies supporting the authors' position and incomplete coverage of contradictory data. Second, primarily English- and Russian-language publications were analyzed, creating a potential language bias and limiting extrapolation of findings to studies published in other languages. Third, no formal quality assessment of included studies was performed using standardized tools (e.g., GRADE, Newcastle-Ottawa), which is characteristic of narrative reviews but reduces the reliability of presented conclusions. Furthermore, the absence of a pre-registered search protocol does not allow full replication of the source selection strategy, which is also inherent to this type of review. These limitations should be considered when interpreting the results.

Analysis of the literature indicates that pharmacogenetic approaches can enhance warfarin dosing precision and optimize antiplatelet and lipid-lowering therapy. However, the persisting variability in clinical study results and the dependence of effects on monitoring conditions and concomitant therapy necessitate adaptation of pharmacogenetic strategies to specific clinical environments. For cardiac surgery patients, the most justified strategy is comprehensive assessment of clinical and genetic factors aimed at reducing the risk of drug-related complications in the perioperative period. Currently, there is a need for additional studies focused on the practical implementation and application of pharmacogenetic algorithms, the results of which will enable evaluation of this strategy's effectiveness in cardiac surgery patients and personalization of pharmacotherapy.