Estimation of the benefit from pre‐emptive genotyping based on the nationwide cohort data in South Korea

Abstract

Genetic variants affect drug responses, making pre‐emptive genotyping crucial for averting serious adverse events (SAEs) and treatment failure. However, assessing the benefits of pre‐emptive genotyping based on genetic distribution, drug exposure, and demographics is challenging. This study aimed to estimate the population‐level benefits of pre‐emptive genotyping in the Korean population using nationwide cohort data. We reviewed actionable gene‐drug combinations recommended by both the Clinical Pharmacogenomics Implementation Consortium (CPIC) and the Dutch Pharmacogenetics Working Group (DPWG) as of February 2022, identifying high‐risk phenotypes. We collected reported risk reduction from genotyping and standardized it into population attributable risks. Healthcare reimbursement costs for SAEs and treatment failures were obtained from the Health Insurance Review and Assessment Service Statistics in 2021. The benefits of pre‐emptive genotyping for a specific group were determined by multiplying drug exposure from nationwide cohort data by individual genotyping benefits. We identified 31 gene‐drug‐event pairs, with CYP2D6 and CYP2C19 demonstrating the greatest benefits for both male and female patients. Individuals aged 65–70 years had the highest individual benefit from pre‐emptive genotyping, with $84.40 for men and $100.90 for women. Pre‐emptive genotyping, particularly for CYP2D6 and CYP2C19, can provide substantial benefits.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Genetic variation is a major source of individual differences in drug efficacy and safety. Personalized therapy based on genotypes can optimize drug outcomes by reducing adverse events (AEs) and maximizing efficacy.

• WHAT QUESTION DID THIS STUDY ADDRESS?

This study investigated how much benefit pre‐emptive genotyping can provide by reducing AEs and maximizing efficacy. The optimal strategy for implementing pre‐emptive genotyping was also investigated.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

The results of the study showed that genotyping on CYP2D6 and CYP2C19 variants provided the greatest benefits in both men and women. In addition, genotyping on an age group of 65–70 years had the largest benefit.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

The results of study provided economic rationale on the implementation of pre‐emptive genotyping and suggested a population‐level optimal implementation strategy.

INTRODUCTION

The genetic variant is a major cause of individual differences in drug efficacy and safety. ¹ One example is the relationship between cytochrome P450 enzyme (CYP) 2C19 and clopidogrel, where CYP2C19 loss‐of‐function variants are strongly linked to therapeutic failure. ² Another example involves TPMT and NUDT15 variants, where severe myelosuppression can lead to discontinuation of azathioprine in the treatment of inflammatory bowel disease. ³ To date, ~400 pharmacogenomic variants have been included in the US Food and Drug Administration (FDA) labels. ⁴ Identifying these variants is expected to optimize drug outcomes by reducing adverse events (AEs) and maximizing efficacy. ⁵

However, despite the wealth of gene‐drug information available, implementing genotyping in clinical practice is challenging. One major issue is the selection of actionable genes. A study comparing pharmacogenomic guidance from various committees (including the Clinical Pharmacogenomics Implementation Consortium [CPIC], and Dutch Pharmacogenetics Working Group [DPWG]) found that only 18% of the cases agreed. ⁶ Another study highlighted that genes related to drug metabolism and transporters were less actionable (~30%) compared to molecular targets related to oncology (~70%). ⁴ These issues pose barriers to the implementation of genotyping in clinical settings. ¹

Once actionable genes are selected, the implementation of genotyping becomes another concern. Previously, point‐of‐care genotyping was commonly performed, evaluating a specific gene at the point of prescription. ⁷ However, the decreasing cost of genotyping now enables pre‐emptive genotyping for multiple actionable genes. ⁸ This approach is more cost‐effective compared to single gene assays. ⁹ A model‐based study evaluating the cost‐effectiveness of pharmacogenomic panel testing for cardiovascular diseases (CYP2C19, CYP2C9, VKORC1, and SLCO1B1 genes) demonstrated that the pre‐emptive approach is more cost‐effective than reactive genotyping. ⁹

Genetic distribution, drug exposure, and demographic structure of a population are important for implementing pre‐emptive genotyping. Several pharmacogenomic variants, such as CYP2C19 and HLA, are associated with race and biogeographic ancestry. Asians have a higher frequency of loss‐of‐function alleles of CYP2C19 than other ancestries (55%), which renders a higher possibility in Asians. ¹⁰ In addition, HLA–B*1502 found in Han Chinese was associated with a 2500‐fold higher risk of Stevens‐Johnson syndrome. ¹⁰ The different frequencies of high‐risk alleles according to population require evaluation based on a specific population. Drug exposure is closely related to prescription patterns of a country and is evaluated through electronic medical records ¹¹ or insurance data. ¹² Demographic structure is associated with drug exposure ¹³ and certain subpopulations could benefit more from pre‐emptive genotyping. For example, whereas only 11.2% of patients aged less than 13 years are exposed to drugs recommended for genotyping in CPIC and DPWG (e.g., clopidogrel and warfarin), 50.6% of patients aged greater than 65 years are exposed to these drugs. ¹³

In this study, we aimed to estimate the benefits of pre‐emptive genotyping with a focus on preventing serious AEs (SAEs). We used nationwide cohort data to estimate population‐level benefits in the Korean population. Based on these results, we suggest optimal strategies for implementing pre‐emptive genotyping.

METHODS

Selection of gene‐drug combination and corresponding SAEs

Gene‐drug combinations commonly recommended by the CPIC and the DPWG as of February 2022 were reviewed. ¹⁴ SAEs that could be prevented by genotyping were evaluated and summarized into representative AEs per gene‐drug combination. SAEs for each drug were either treatment failures or toxicities of clinical significance. For each selected AE, high‐risk phenotypes and the relative risk to the reference phenotype were identified. The frequency of the phenotype in the Korean population was calculated using the PharmGKB genotype frequency database. ¹⁵ Genotype frequency in the East Asian population was alternately used when Korean data were not available.

Collection of estimation of risk reduction from genotyping

Relative risks (RRs) of SAEs in a gene‐drug pair were collected from large‐scale randomized controlled trials or meta‐analyses. Combined with the exposure to risk factors (frequency of high‐risk genotype in the Korean population, p), RR was converted into population‐attributable risk (PAR) using Levin's formula:

$$\mathrm{PAR}=\frac{p\left(\mathrm{RR}-1\right)}{p\left(\mathrm{RR}-1\right)+1}$$

($p$: exposure of the risk factor in the population; RR: relative risk) ¹⁶ , ¹⁷ PAR was interpreted as the proportion of SAEs that were attributable to high‐risk genotype (RR > 1) or could be prevented by pre‐emptive genotyping (RR < 1). The prevalence of drug‐specific SAEs was obtained from literature.

Estimation of average healthcare reimbursement costs

SAEs in a gene‐drug pair were matched with the closest diagnosis codes from the International Classification of Diseases (ICD). The average healthcare reimbursement costs were obtained from the Health Insurance Review and Assessment Service Statistics for 2021. If an SAE corresponded to two or more probable ICD codes, the average cost of each ICD code was used.

Estimation of drug exposure

Drug exposure in the Korean population was estimated from the National Health Insurance Sharing Service National Sample Cohort database consisting of 1,108,369 sample patients (2% of the total population) stratified by demographics. ¹⁸ The exposure to a drug was defined as a proportion of patients who were prescribed the drug at least once during the study interval (between 2002 and 2019). Drug exposure was calculated in total and subgroups stratified by sex and age (5‐year intervals). The study was exempt from human subject review by the Institutional Review Board of Seoul National University Bundang Hospital (IRB no. X‐1907‐552‐903). Statistical analyses were conducted using SAS version 9.4 (SAS Institute) and R version 4.2.2 (R Core Team, Vienna, Austria).

Estimation of the population benefit of pre‐emptive genotyping

Individual benefit per genotyping was calculated as the following formula:

$$\begin{array}{c}\text{Benefit}\mathrm{per}\text{genotyping}\hfill \\ =\left|\mathrm{PAR}\right|\times \mathrm{SAE}\text{prevalence}\hfill \\ \times \text{Average healthcare reimbursement cost}.\hfill \end{array}$$

The benefit of preemptive genotyping in a specific group (i.e., sex and age groups) was calculated as the product of drug exposure in a specific group and the individual benefit per genotyping. To compare the benefits with genotyping cost, healthcare reimbursement costs for genotyping each gene were obtained from the Health Insurance Review and Assessment Service reimbursement costs in 2021.

RESULTS

Gene‐drug‐event combinations and the corresponding healthcare reimbursement costs

A total of 95 gene‐drug pairs were recommended in the CPIC guideline, among which 35 gene‐drug pairs were commonly recommended in the DPWG guideline. Single representative SAEs were extracted for 34 gene‐drug pairs with the highest evidence level in the CPIC guideline, except for clopidogrel, where two SAEs (relapse of myocardial infarction and major bleeding) were selected. A total of 36 gene‐drug‐event combinations with recommendations were finally identified. CYP2C19 had the greatest number of gene‐drug‐event combinations (11 combinations), followed by CYP2D6 (10 combinations; Table 1) Excluding gene‐drug‐event combinations where RRs or PARs were not identified, 31 gene‐drug‐event combinations were included in the analysis (Table 1). A summary of the SAEs matched to the closest diagnostic codes and healthcare reimbursement costs are presented in Table S1.

TABLE 1.

Actionable gene and drug combination recommended in both CPIC and DPWG by February 2022.

Gene	High‐risk phenotype	Drug	SAE	Frequency of high‐risk phenotype	Effect of genes a		Overall prevalence of SAE (%) a
					RR	PAR
CYP2B6	IM/PM	Efavirenz	Psychosis	0.37	0.68 35	−0.13	3.3 36
CYP2C19	UM/RM/PM	Citalopram	Relapse of depression	0.16	1.16 37	0.02	24.1 38 , b
	UM/PM	Clomipramine	Relapse of depression	0.13	(−)	(−)	47.9 39 , b
	IM/PM	Clopidogrel	Relapse of myocardial infarction	0.59	0.56 40	−0.35	3.0 40
	IM/PM	Clopidogrel	Major bleeding	0.59	0.81 40	−0.13	6.4 40
	UM/RM/PM	Escitalopram	Relapse of depression	0.16	3.3 41	0.49	24.1 38 , b
	UM/PM	Imipramine	Relapse of depression	0.13	(−)	(−)	47.9 39 , b
	UM/RM/NM	Lansoprazole	Peptic ulcer bleeding	0.41	4.23 42	0.57	2.2 43
	UM/RM/NM	Omeprazole	Peptic ulcer bleeding	0.41	5.13 42	0.63	2.2 43
	UM/RM/NM	Pantoprazole	Peptic ulcer bleeding	0.41	1.74 42	0.23	2.2 43
	PM	Sertraline	Relapse of depression	0.13	1.38 44	0.05	24.1 38 , b
	EM	Voriconazole	Relapse of fungal infection	0.38	1.31 45	0.11	25.8 45 , b
CYP2C9	IM/PM	Phenytoin	SJS/TEN	0.05	20.86 46	0.50	0.045 47 ,c
	IM/PM	Warfarin	Major bleeding	0.05	0.68 48	−0.02	6.36 49
CYP2D6	IM/PM	Amitriptyline	Relapse of depression	0.33	5.77 50	0.61	47.9 39 , b
	UM	Atomoxetine	Relapse of ADHD	0.16	(−)	(−)	(−)
	IM/PM	Clomipramine	Relapse of depression	0.33	5.77 50	0.61	47.9 39 , b
	UM	Codeine	Respiratory depression	0.33	(−)	(−)	(−)
	IM/PM	Doxepin	Relapse of depression	0.33	5.77 50	0.61	47.9 39 , b
	IM/PM	Imipramine	Relapse of depression	0.33	5.77 50	0.61	47.9 39 , b
	IM/PM	Nortriptyline	Relapse of depression	0.33	5.77 50	0.61	47.9 39 , b
	UM	Paroxetine	Relapse of depression	0.16	0.91 50	−0.01	24.1 38 , b
	PM	Tamoxifen	Breast cancer recurrence	0.16	1.25 51	0.04	0.13 52
	UM	Tramadol	Respiratory depression	0.33	(−)	(−)	(−)
CYP3A5	EM/IM	Tacrolimus	Graft rejection	0.44	1.32 53	0.12	19.91 53
DPYD	IM/PM	Fluorouracil	Myelosuppression	0.003	9.76 54	0.03	22.17 54
HLA‐B	*15:02	Abacavir	Hypersensitivity reaction	0.05	23.6 55	0.53	5.4 55
NUDT15	IM/PM	Azathioprine	Myelosuppression	0.18	6.86 56	0.51	7.0 56
	IM/PM	Mercaptopurine	Myelosuppression	0.18	6.86 56	0.51	7.0 56
	IM/PM	Thioguanine	Myelosuppression	0.18	6.86 56	0.51	7.0 56
SLCO1B1	Decreased/poor function	Atorvastatin	Rhabdomyolysis	0.17	1.49 57	0.08	0.0042 58
	Decreased/poor function	Simvastatin	Rhabdomyolysis	0.17	3.39 57	0.29	0.0042 58
TPMT	PM	Azathioprine	Myelosuppression	0.0003	3.72 56	<0.01	7.0 56
	PM	Mercaptopurine	Myelosuppression	0.0003	3.72 56	<0.01	7.0 56
	PM	Thioguanine	Myelosuppression	0.0003	3.72 56	<0.01	7.0 56
VKORC1	c.‐1639 GA/AA	Warfarin	Major bleeding	0.88	0.68 48	−0.39	6.36 49

Abbreviations: ADHD, attention deficit hyperactivity disorder; CPIC, Clinical Pharmacogenomics Implementation Consortium; DPWG, Dutch Pharmacogenetics Working Group; EM, extensive metabolizer; IM, intermediate metabolizer; NM, normal metabolizer; PAR, population attributable risk; PM, poor metabolizer; RM, rapid metabolizer; RR, relative risk; SAE, serious adverse events; SJS, Stevens‐Johnson syndrome; TEN, toxic epidermal necrolysis; UM, ultra‐rapid metabolizer.

^a Positive PARs were interpreted as the proportion of SAEs attributable to high‐risk phenotypes. Negative PARs represented the proportion of SAEs that could be prevented through genotyping. Therefore, absolute values of PAR were used to calculate the preventable proportion of SAEs from genotyping. Missing values for RR or the overall prevalence of SAEs were denoted as (−).

^b Overall treatment discontinuation rates when using tricyclic antidepressants or selective serotonin reuptake inhibitors are used as the overall prevalence of SAEs (i.e., treatment failure).

Drug exposure

Exposure to actionable drugs according to sex and age group is presented in Figure 1 and Table S2. Patients using fluorouracil, pantoprazole, or phenytoin were not included in the sample cohort database due to low prescription rate or obsolete/misclassified codes, and therefore drug exposure was not estimated. Drug exposure was generally higher in female patients than in male counterparts. Drug exposure increased with age until the maximum exposure was in the age group of 65–70 years. Drug exposure decreased in age groups higher than 70 years. Among male patients, tramadol (40.1%), lansoprazole (14.2%), and omeprazole (11.5%) were the most frequently used, whereas, in women, tramadol (44.3%), lansoprazole (16.5%), and amitriptyline (14.1%) were the most frequently used. Tramadol presents the most frequently used drug in the entire population.

FIGURE 1.

Summary of the drug exposure stratified by sex and age: proportion of patients in each age group.

Population benefit for pre‐emptive genotyping

Overall, CYP2D6 and CYP2C19 showed the greatest benefits in both, male and female patients (Table 2). The age group of 65–70 years had the largest benefit for male ($84.40) and female ($100.90) patients from pre‐emptive genotyping (Table 2 and Figure 2). Healthcare reimbursement costs for genotyping are listed in Table S3. Genotyping costs for each genotype ranged from $100.80 to $210.00.

TABLE 2.

Calculated benefits from the pre‐emptive genotyping by sex, age group, and gene.

Genes	Age group
	0–5	5–10	10–15	15–20	20–25	25–30	30–35	35–40	40–45	45–50	50–55	55–60	60–65	65–70	70–75	75–80	80–85
Male
CYP2B6
CYP2C19	0.7	1.1	1.3	1.9	2.5	3.8	5.0	6.9	9.5	13.2	17.3	21.9	25.2	27.9	27.9	23.1	13.4
CYP2C9									0.1	0.1	0.2	0.3	0.4	0.5	0.4	0.4	0.3
CYP2D6	4.5	3.7	3.9	5.1	7.1	8.8	10.6	13.3	17.2	21.8	27.8	35.5	41.8	46.7	44.6	38.7	25.9
CYP3A5							0.1	0.1	0.1	0.1	0.1	0.1	0.2	0.1	0.1
HLA‐B
NUDT15			0.1	0.1	0.1			0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1
SLCO1B1
TPMT
VKORC1	0.1	0.1	0.1	0.1	0.3	0.4	0.6	0.9	1.3	2.2	3.5	5.4	8.2	9.1	8.6	7.2	5.5
Total	5.3	4.9	5.4	7.2	1	13.0	16.3	21.3	28.3	37.5	49.0	63.3	75.9	84.4	81.7	69.4	45.1
Female
CYP2B6
CYP2C19	0.8	1.3	2.0	2.9	3.3	3.9	4.8	6.4	9.0	12.0	15.9	21.0	25.4	27.7	27.5	20.6	10.8
CYP2C9										0.1	0.1	0.2	0.3	0.4	0.4	0.3	0.2
CYP2D6	3.4	3.9	5.5	8.2	10.5	13.3	16.4	23.3	33.1	41.2	49.5	57.6	64.0	65.6	59.5	47.6	28.0
CYP3A5				0.1	0.1	0.1	0.1	0.1	0.2	0.2	0.2	0.1	0.2	0.1	0.1
HLA‐B
NUDT15				0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1
SLCO1B1
TPMT
VKORC1		0.1	0.1	0.1	0.2	0.2	0.4	0.6	0.8	1.3	2.3	3.7	5.4	7.0	7.1	6.1	3.5
Total	4.2	5.3	7.6	11.4	14.2	17.6	21.8	30.5	43.2	54.9	68.1	82.7	95.4	100.9	94.6	74.6	42.5

Note: Benefits were calculated as the product of the absolute value of population‐attributable risk, overall prevalence of serious adverse events (SAEs), drug exposure in a specific group, and average healthcare reimbursement cost for SAEs. Benefits were calculated in US dollars ($1 = 1145.07 KRW in 2021). Estimated benefits of less than $0.10 are noted as blank for clarity.

FIGURE 2.

Calculated benefits from the preemptive genotyping by sex, age group, and gene.

DISCUSSION

The benefits of pre‐emptive genotyping are difficult to estimate because they require various assumptions. For example, it is difficult to estimate the probability of a patient taking a specific drug. In addition, the allele frequency of high‐risk phenotypes and the prevalence of SAEs can vary among different populations. Different healthcare systems affect the cost of SAEs, which requires country‐specific analyses. ¹⁹

We used nationwide cohort data in Korea to provide objective estimates required for a cost–benefit analysis. Because the healthcare system in Korea is centralized by the national insurance system, data can represent the total population in Korea. ¹⁸ It could be advantageous over using a hospital‐based electronic health system data, ⁵ which is generally susceptible to selection bias, ²⁰ to estimate drug exposure. Furthermore, the data were collected longitudinally, providing an estimate of population‐level drug exposure.

Healthcare reimbursement costs can vary among healthcare systems despite the same diagnostic code. The situation is exemplified in the comparison between Korea and the United States. The healthcare system in Korea is established on the National Healthcare Insurance, where most patients were mandatorily enrolled in the system. ²¹ In contrast, the healthcare system in the United States is highly dependent on private insurance. ²¹ The difference has resulted in the remarkable difference for the healthcare reimbursement costs for several common SAEs. For acute myocardial infarction, the average healthcare reimbursement cost was $2739 in Korea (Table S1, 2021), whereas it was $15,000 in the United States. ²² Similarly, for acute hemorrhagic cerebrovascular disease (subarachnoid and intracerebral hemorrhages), the costs were estimated to $6653–8434 in Korea, whereas it was $24,800 in the United States. ²² Although a detailed evaluation of the treatment process is required, these cases support the idea that the estimation of benefits should be performed within the context of each healthcare system.

We found that sex and age were associated with drug exposure, which may have affected the benefits of genotyping. Elderly patients (≥65 years) can be mostly benefited from genotyping, which is consistent with the tendency of high‐cost users of prescription drugs. ²³ The benefit of pre‐emptive genotyping was similarly emphasized in the elderly due to polypharmacy. ²⁴ , ²⁵ Additionally, the frequency of high‐risk phenotypes can be a crucial factor in estimating costs. For example, the low frequency of high‐risk phenotypes in DPYD (0.003) and TPMT (0.0003) in the Korean population limits the overall benefits of pre‐emptive genotyping despite the high cost of SAEs.

Of note, there could be a discrepancy between the estimated drug exposure in our analysis and that of the entire patient population. We found that drug exposure of the several drugs was omitted due to low prescription rate (phenytoin and fluorouracil) in the sample population. In addition, obsolete or misclassified drug codes might result in the unexpected omission of pantoprazole in the analysis. We estimated that the cost saving for CYP2C19‐pantoprazole‐peptic ulcer bleeding was $5.00, whereas the corresponding values were $12.50 for lansoprazole and $13.80 for omeprazole before applying drug exposure. Therefore, considering the expected drug exposure of pantoprazole (almost double the lansoprazole or omeprazole ²⁶ ) in the patient population, the cost savings for pantoprazole would be similar to those of omeprazole or lansoprazole.

In addition, drug exposure can be estimated differently according to the definition and source of data. Representatively, drug exposure in pediatric population is highly variable according to literature. A retrospective study in an academic children's hospital estimated that 49.3% of pediatric patients were diagnosed with genotype‐associated diseases, among whom 30.9% were prescribed actionable drugs. ²⁷ Another study estimated 1.3% of annual exposure for pediatric patients, which would yield different value for cumulative estimates. ²⁸ Therefore, the definition and source of data must be accounted for estimating the potential benefits from genotyping.

Another issue in estimating benefits is the heterogeneous risk reduction when genotyping. A similar issue was addressed previously, as most pharmacoeconomic studies focused on single‐gene genotyping. ⁷ Actually, the amount of risk reduction from genotyping was reported only in a few gene‐drug‐event cases, and the methods used were highly heterogeneous. We attempted to solve this issue using PAR to provide a standardized and quantitative estimate of preventable risk and integrate PAR with the prevalence obtained separately. ¹⁶ The approach enabled more efficient use of reported values to estimate the benefit of genotyping.

It is noteworthy that current healthcare reimbursement costs for genotyping are higher than the calculated benefits from pre‐emptive genotyping. The results should be carefully interpreted with the costing methods for genotyping. Genotyping costs are divided into direct and indirect costs. ²⁹ Direct costs include consumables and reagents, whereas indirect costs are associated with infrastructure costs, such as facilities, administrative costs, and maintenance fees. ²⁹ We suppose that current healthcare reimbursement costs might reflect individual‐level total genotyping costs, which could be reduced when a population‐level multiple genotyping strategy is adopted. Therefore, the results should be cautiously interpreted, and further investigation into the costs according to various genotyping strategies is required.

Our results support the potential benefits of preemptive genotyping. As we restricted the estimation of the benefit only to commonly recommended gene‐drug pairs in two guidelines and only included AEs in which the medical cost is quantifiable, the actual benefit from genotyping would be higher. In addition, the benefit estimated in our study only included direct medical costs for brevity, which would be increased, including the indirect medical cost (e.g., transportation expenses and expenses related to changing jobs ³⁰ ) as recommended in pharmacoeconomic evaluation. ¹⁷ For example, the indirect cost of acute myocardial infarction in Korea accounted for 42.3% of the total costs in 2012. ³¹ Therefore, the calculated benefit in our study would be the minimal estimate but could suggest implementing cost‐effective genotype panels. ³²

The cost‐effectiveness of pre‐emptive genotyping compared to reactive genotyping is still on debate. Reactive genotyping is easier to perform and has already shown cost‐effectiveness. ³³ However, the strategy is restricted to specific gene‐drug pairs and application to other drugs is limited. In contrast, pre‐emptive genotyping has advantage in comprehensive understanding of multiple gene‐drug pairs, and would be more efficient when multiple genes were evaluated in a single panel. ³³

We performed an additional illustrative test with two major genes (CYP2C19 and CYP2D6) in our study. We obtained the combined distribution of CYP2C19 and CYP2D6 from the previous study in 1003 Japanese patients. ³⁴ To evaluate whether the frequency of high‐risk phenotypes for CYP2C19 and CYP2D6 were independent, or separate genotyping would suffice, we performed an independence test (Table S4). We found that the distribution of high‐risk phenotypes of CYP2C19 and CYP2D6 were not independent, which implied testing two genes together could be an efficient way to identify high‐risk phenotypes rather than separately. The preliminary results from the analysis could support the rationale for pre‐emptive genotyping.

Our study had several limitations. We simplified benefits to include only the prevention of SAEs other than mild‐to‐moderate AEs, which could have underestimated the actual benefits. Drug exposure estimates do not consider the indication for a drug and may oversimplify the real‐world situation. In addition, the long‐term benefits of genotyping need to be evaluated in terms of improving the quality of life. The estimation of benefits was based on the reported allele frequency and insurance costs in the Korean population, which can limit the application of the results to other countries. The prevalence of SAEs can vary among populations and requires further investigation. More comprehensive prospective studies are required to investigate the economic value of pre‐emptive genotyping in South Korea.

In conclusion, pre‐emptive genotyping can yield measurable benefits in preventing SAEs within the healthcare system. Considering drug exposure and genotyping distribution, genotyping CYP2D6 and CYP2C19 in the age group of 65–70 years would result in the greatest benefits, estimated at least $84.40–100.90 per individual.

AUTHOR CONTRIBUTIONS

K.Y.H., S.H., J.Y.N., K.‐S.Y., I.‐J.J., J.‐Y.C., and S.Y. wrote the manuscript. K.Y.H., S.H., J.Y.N., and S.Y. designed the research. K.Y.H., S.H., and J.Y.N. performed the research. K.Y.H., S.H., J.Y.N., and S.Y. analyzed the data.

FUNDING INFORMATION

This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF‐2019R1C1C1006688) and the Seoul National University Bundang Hospital Research Fund (14‐2020‐0040).

CONFLICT OF INTEREST STATEMENT

The authors declared no competing interests for this work.

Supporting information

Data S1

Huh KY, Hwang S, Na JY, et al. Estimation of the benefit from pre‐emptive genotyping based on the nationwide cohort data in South Korea. Clin Transl Sci. 2024;17:e13772. doi: 10.1111/cts.13772