Integrating Polygenic Risk Scores (PRS) for Personalized Diabetes Care: Advancing Clinical Practice with Tailored Pharmacological Approaches

Abstract

The rising global prevalence of diabetes poses a serious threat to public health, national economies, and the healthcare system. Despite a high degree of disease heterogeneity and advancing techniques, there is still an unclear diagnosis of patients with diabetes compounded by the array of long-term microvascular and macrovascular complications associated with the disease. In addition to environmental variables, diabetes susceptibility is significantly influenced by genetic components. The risk stratification of genetically predisposed individuals may play an important role in disease diagnosis and management. Precision medicine methods are crucial to reducing this global burden by delivering a more personalised and patient-centric approach. Compared to the European population, genetic susceptibility variants of type 2 diabetes mellitus (T2DM) are still not fully understood in other major populations, including South Asians, Latinos, and people of African descent. Polygenic risk scores (PRS) can be used to identify individuals who are more susceptible to complex diseases such as diabetes. PRS is selective and effective in developing novel diagnostic interventions. This comprehensive predictive approach facilitates the understanding of distinct response profiles, resulting in the development of more effective management strategies. The targeted implementation of PRS is especially advantageous for people who fall into a higher-risk category for diabetes. Through early risk assessment and the creation of individualised diabetes treatment plans, the integration of PRS in clinical practice shows potential for reducing the prevalence of diabetes and its complications. Diabetes self-management depends significantly on patient empowerment, with behavioural monitoring emerging as a vital facilitator. The main aim of this review article is to formulate a more structured intervention strategy by advocating for increased awareness of the clinical utility of PRS and counseling among healthcare practitioners, patients, and individuals at risk of diabetes.

Plain Language Summary

The expanding global prevalence of multiple subtypes of diabetes alarmingly threatens the development of more effective methods of management. Integrating risk stratification measures like polygenic risk scores (PRS) into clinical practice can significantly improve patient outcomes. It should be accompanied by patient counseling, which has already been demonstrated to provide extra benefits for this course of action.

The utilisation of PRS becomes crucial in tackling this problem. Through the use of genetic data to identify people according to risk, PRS can improve the accuracy of diagnosis and tailor treatment plans. This strategy is in line with the larger push for precision medicine, which aims to personalise treatment to satisfy the distinctive requirements associated with each patient.

There are a few obstacles that must be overcome for PRS to be successfully incorporated into standard clinical practice. The absence of thorough population-based research, high costs, and complicated interpretation are important challenges that need to be overcome. For PRS to be used more widely, efforts must be made to improve its usability and accessibility.

Effective diabetes therapy also requires behavioural monitoring and patient empowerment in addition to hereditary variables. Improving patients’ and healthcare professionals’ understanding of PRS can promote improved self-care and treatment plan compliance. Genetic knowledge and behavioural techniques can be combined to create more complex and successful interventions.

To sum up, the incorporation of PRS into the treatment of diabetes signifies a noteworthy progression in the direction of personalised medicine. Although there are still obstacles to overcome, there are significant potential advantages to better risk classification and customised treatment plans. Through sustained research advancement and addressing implementation challenges, we may make progress towards a future with more accurate, efficient, and accessible diabetes management.

Key Summary Points

Polygenic risk scores (PRS) represent a promising approach to precision medicine by providing information about an individual’s genetic susceptibility to diabetes, thereby increasing treatment efficacy and patient outcomes through personalised care.

By identifying those who are more likely to develop diabetes, PRS can help physicians in creating individualised plans for diagnosis and treatment that will enhance patient outcomes.

More targeted studies are needed to investigate the genetic vulnerability to type 2 diabetes mellitus (T2DM) among large populations other than Europeans, including South Asians, Latinos, and people of African descent.

Despite its potential, PRS has several drawbacks, including steep costs, challenging interpretation, and a lack of population-based research. Adjustments to accessibility, ease of use, and cost are required for widespread clinical adoption.

Behavioural monitoring and patient empowerment are essential components of effective diabetes care. Improved customised treatment strategies and a decrease in the prevalence of diabetes may result from greater knowledge of the PRS and its integration into clinical practice.

Introduction

The rising global prevalence of metabolic noncommunicable diseases, such as diabetes, places a significant strain on public health, national economies, and healthcare delivery systems. Diagnosing different subtypes of diabetes having overlapping symptoms may become challenging for clinicians [1–3]. According to a recent cross-sectional study conducted by the Indian Council of Medical Research–India Diabetes (ICMR–INDIAB), the overall weighted prevalence of diabetes and prediabetes among people in India is 11.4% and 15.3%, respectively [4]. These statistics highlight the necessity for advanced and improved management strategies in economically growing nations. However, the problem of diabetes screening continues, with approximately half of the affected individuals remaining undiagnosed [5]. Studies have reported that as a result of greater ethnic diversity, South Asian populations have more early onset of type 2 diabetes than the European population [6]. Among them, Asian Indians show more insulin resistance when compared to other ethnic groups [7]. At the same time, a patient with diabetes often has several microvascular and macrovascular complications caused by the disease such as renal failure, retinopathy, neuropathy, foot amputation, etc. [8].

Risk stratification is vital in estimating a person’s lifetime probability of developing a disease or disease-associated complications. In the vision of the future of personalised medicine where genomics plays a big role in patients’ disease risk evaluation, one approach is the integration of polygenic risk score (PRS) into clinical practice. The clinical overlap between different forms of diabetes contributes to misdiagnoses, worsened by a lack of awareness and incomplete understanding within the healthcare systems [9]. Genetic risk scores are a major advancement in the field of personalised genomic medicine, which could be implemented with consideration of the different clinical and financial conditions of each disease. Through careful risk score calculations, it is possible to prognosticate type 2 diabetes, enabling timely interventions and improving clinical outcomes [10]. PRS can develop a preventive care approach when combined with clinical risk scores. Studies have reported that it is helpful in gestational risk prediction as well [11]. Estimating individuals’ susceptibility towards a disease using PRS is a game changer but its clinical utility is yet to be established. Recognising the significant role of genetic information in analysing clinical outcomes, there is an urgent need to unfold the genetic architecture of diabetes traits to develop personalised medicine and early diagnostic interventions [12]. The main aim of this review article is to formulate a more structured intervention strategy for increased awareness of the clinical utility of PRS and counselling among healthcare practitioners, patients, and individuals at risk of diabetes.

This article does not contain any new studies with human participants or animals performed by any of the authors.

Genetic Landscape of Diabetes

Diabetes is a multifactorial, complex polygenic disorder arising from the intricate interplay of several genetic factors in conjugation with other factors such as lifestyle, environment, and demographics. Comparing genetic and non-genetic determinants plays a significant role in forecasting the disease burden within a population. The aetiology of these conditions predominantly grows from the cumulative effect of hundreds to thousands of substantial and minute genetic variants. Identifying mutations within actionable genes makes it easier to conduct more targeted surveillance and therapeutic interventions, thereby promoting the use of precision medicine strategies for carriers [13, 14]. Heritability and genetic liability are the two important factors in determining how genetics affects a person’s likelihood of developing a health condition. In the case of type 1 diabetes (T1D), the primary predictive determinant for individuals at increased risk lies in the concurrent presence of multiple islet autoantibodies (IA) within the serum [15]. Studies on genetic associations T1D have historically focused on the human leukocyte antigen (HLA) region; however, current genome-wide association studies (GWAS) have found over 50 unique regions that significantly contribute to the genetic architecture of T1D risk [16]. Assessing the variability of HLA haplotypes proves to be a crucial contributor to early diagnostic considerations. When it comes to insulin-dependent diabetes, an individual having HLA B8 exhibits a 2.6-fold increased risk of developing the disease compared to an HLA B8-negative individual. Similarly, the relative risk related to HLA B15 is 1.8, and for HLA B18 it is 2.0, highlighting the significance of HLA haplotypes in disease predisposition [17]. Apart from HLA genes, there are several non-HLA-associated loci responsible for causing T2D, like INS, PTPN22, IL2RA, IL10, CTLA4, IL2, BACH2, etc. [18]. Several studies show an important contribution of lncRNA in autoimmune disease development [19]. An important example is the p.E508K missense mutation within the HNF1A gene, which occurs in approximately 0.1% of the general population and is associated with a five times increased risk of type 2 diabetes (T2D) [20]. Multiple studies have consistently demonstrated that individuals with an affected first-degree relative exhibit a significantly elevated risk of developing T2D [21].

Using GWAS, researchers have rigorously uncovered more than 300 genetic loci associated with T2D [8]. The combination of these strong GWAS findings explains more than 19% of the phenotypic diversity in T2D susceptibility [22]. Another major step in comprehensive coverage of new susceptibility variants for various ethnic groups suffering from diabetes is contributed via the development of databases such as UK Biobank and GenomegaDB [23, 24]. Research demonstrates a novel susceptibility locus associated with T2D among Indians at 2q21, with the lead signal being rs9552911 single nucleotide polymorphism (SNP) [25]. A recent GWAS expansion conducted on 898,130 individuals of European ancestry resulted in an extended inventory of 135 newly implicated T2D risk predisposed variants [26, 27].

Many SNPs linked to important genes, including IL6, AGTR1, NOS3, and TNFA, are reported to impact diabetes progression and associated complications significantly [28]. Diabetes progresses not only through beta cell destruction but also is accompanied by a spectrum of macrovascular and microvascular complications, like cardiovascular disease, retinopathy, neuropathy, hypertension, and kidney diseases [29]. Figure 1 illustrates the spectrum of genes associated with different forms of diabetes and their macrovascular and microvascular complications.

Fig. 1.

List of genes involved in different forms of diabetes and their complications

Understanding Polygenic Risk Scores

Polygenic risk scores (PRS) offer an advanced method of estimating an individual’s susceptibility to diabetes based on their genetic characteristics [2]. The calculation involves profiling the patient’s genetic composition and comparing it with disease-specific GWAS data [30]. PRS mainly depends on the weighted sum of all the risk alleles and the degree to which each allele contributes to risk. This additive effect of each small variant contributing to disease causation for population-based GWAS can be calculated via different approaches like genome-wide complex trait analysis and a few other machine learning methods for several common complex diseases [31–34]. It can be used as a population screening tool to segregate people at greater risk for diabetes. Incorporation of PRS has shown a greater positive result in improving accuracy in many other complicated disorders like coronary artery disease and prostate cancer [35–39]. The analysis of PRS is primarily focused on the calculation of relative risk, absolute risk, or percentile rank [40]. Each metric has an important role, especially in the case of T2D. Percentile rank calculation helps recommend lifestyle changes to individuals who are at elevated risk of T2D. On the other hand, absolute risk calculation offers a strong way of assessing a person’s risk over a particular period of their lifetime, establishing itself as the most reliable among the three types of PRS [41]. Relative risk can be utilized by clinicians to measure the magnitude of change in risk of an individual.

Compared to the European population, genetic susceptibility variants of T2DM are still not fully understood in other major populations, including South Asians, Latinos, and people of African descent [42–46]. To calculate PRS, an individual’s genetic ancestry, which is similar to large GWAS data, the reference effect size is taken and may require access to an ancestry-matched genotype-level reference panel. Differential genetic drift may result in unanticipated bias when scores are inferred from one population to another [41]. It has been reported that impaired fasting glucose is a prominent hallmark of both increased hepatic insulin resistance and inadequate first-phase insulin production in many Indians. In addition, research has shown that, in contrast to other ancestral groups, South Asians have additional factors beyond insulin resistance that can account for reduced insulin secretion, with an unhealthy lifestyle being a significant one [47]. This gap poses a challenge for use in non-European ancestry populations, as PRS are derived from European populations [3, 48, 49]. Replicating the relationships of candidate SNPs with T2D in several ethnic groups is critical to separate variants that are common and reproducible susceptibility genes. Therefore, there is an urgent need to uncover more new genetic variants in the context of T2DM, especially for the South Asian population. Using GWAS data, we were able to identify 1136 shared genetic variants amongst the European and Asian populations (Fig. 2). This could be useful in finding additional risk alleles that could serve as pharmacological targets for the creation of more customized pharmaceutical treatments.

Fig. 2.

Shared genetic variants between European and Asian populations by GWAS studies

Clinical Relevance of PRS in Diabetes Risk Assessment

Interestingly, PRS has the tendency to discriminatively identify several subtypes of diabetes, including type 1, type 2, maturity-onset diabetes of the young (MODY), and other forms [3, 50]. Additionally, PRS plays an important role in prognosticating complications and evaluating their response to stringent interventions targeting blood pressure and glucose control [51]. This overall predictive ability helps to increase the understanding of each individual’s unique response profiles, contributing to formulating more efficient management approaches. Incorporating PRS into clinical practice holds promise for altering diabetic disease progression and serious complications, such as foot amputations and blindness, by enabling early risk assessment [52]. Diabetic polyvascular disease (DPD) appears to be a significant clinical concern, exacerbated by complications of diabetes that selectively impact major organs, including the heart, brain, peripheral vasculature, retina, and kidney of an individual [53]. The given spectrum of complications shows a compelling need for increased therapeutic vigilance and the implementation of targeted self-management strategies to address this multifaceted disease. Studies show that calculating PRS at birth for common noncommunicable diseases could enhance clinicians’ ability to adopt a more accurate and proactive-thinking approach, by recognizing that a person’s risk profile may change over their lifespan [54]. PRS proves to be particularly beneficial for neonates as a lifetime risk predictor, resulting in primordial preventive means. It acts as a secondary prevention strategy for diagnosing patients with or without family history. A family history depicts mainly the strong pattern of disease inheritance among close relatives; however, PRS offers additional benefits in genetic risk assessment when the disease condition is mainly influenced by the environment and family history is not well understood. Additionally, it is a tertiary preventive measure for adults who are newly diagnosed with diabetes to estimate their risk of complications. Since PRS provides diabetes risk stratification and early management, it may be possible to improve the accuracy and efficacy of preventative interventions that lower the likelihood of complications from the disease when integrated with good glycemic control [55]. In cases where individuals have a decent glycaemic index while having a higher risk of experiencing complications from their condition, by providing additional information about an individual’s genetic predisposition to bottlenecks, PRS can provide insights beyond traditional clinical examinations; nevertheless, further study is needed to confirm this evidence. A recent study has reported that glucocorticoid-induced hyperglycemia is more common in people with a higher genetic susceptibility to T2D. This discovery provides a method for risk assessment as part of a precision medicine strategy stating the relevance of PRS [56]. The ADVANCE (The Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation) trial produced a multi-PRS comprising 598 SNPs and 10 PRS, including obesity, diabetes, and albuminuria [57, 58]. These studies that take into account multiple PRS are useful in the field of medicine.

Calculating PRS and Its Cost

Estimating a PRS requires familiarity with certain data sources used in PRS calculation. GWAS comprise large-scale genetic information including SNPs that serve as a leading source in selecting relevant SNPs and their weights in PRS calculation [59]. Specifically for diabetes, it is important to choose relevant reference datasets that help in deciding which genetic variant to include and how to weigh them according to their association strengths. Other sources for large-scale genetic and phenotypic datasets from different ethnic groups include UK Biobank, dbGaP, Polygenic Score Catalog, and other biobanks. This culmination of different data sources ensures the integrity and authenticity of the genetic data to make it more reliable for researchers to calculate PRS. The selection of the reference dataset depends on many factors including the disease of interest, a larger sample size, and genetic diversity along with broad genomic coverage. This is followed by the preprocessing of data such as data cleaning and quality control to ensure the reliability of PRS. GWAS studies pinpoint genetic variants that exhibit statistically significant associations with the phenotype under investigation. The SNP selection criteria are mainly dependent on the significance threshold (often based on p values or adjusted significance levels) to determine which SNPs from GWAS results will be included in the PRS calculation. To enhance the accuracy and independence of each SNP, linkage disequilibrium (LD) pruning is performed.

The standard equation for the calculation of PRS is ${\text{PRS}}_j={\mathrm{\Sigma }}_i^N\beta i_{}^{\ast }{\text{dosage}}_{\mathit{ij}}$ where N = number of SNPs in the score, βi = is the effect size (or beta) of variant I, and dosage_ij = the number of copies of SNP i in the genotype of individual j.

The total cost for calculating PRS varies a lot depending on the data availability and the technology used. Since several steps are involved in the calculation including genotyping or sequencing, data processing, validation, and clinical integration. A cost–utility analysis based on a Markov model suggests that the implementation of the PRS for early prediction of diabetic nephropathy (DN) in patients with T2D compared to the traditional screening approaches indicating that PRS was the dominant strategy for both healthcare and society [60]. The AoUPRS (All of Us Program Polygenic Risk Score) tool was designed in a recent groundbreaking study to address the difficulties researchers faced when calculating PRS using AoU data, particularly the high costs and inefficiencies involved in data extraction and analysis [61].

An important step in integrating PRS with phenotypic data enables assessing how genetic risk influences phenotypic outcomes, with potential applications in risk prediction and personalised medicine. Figure 3 summarises the key steps involved in PRS calculation.

Fig. 3.

Steps involved in polygenic risk score (PRS) calculation

Personalised Precision Approach via PRS in Pharmacogenomics

Metformin is the most often prescribed medicine for diabetes and is estimated to reduce diabetes-related mortality by 42% [62]. Despite its widespread use and recognised cost-effectiveness, the utility of metformin remains subject to debate within the scientific community. Although it is usually well tolerated, it does not come without side effects. Previous studies have reported that around 30% of individuals suffer from gastrointestinal complications due to the use of metformin [63]. These can possibly be due to any disruption in genes such as SLC22A1 and SLC9A4 encoding the transporters like organic cation transporter (OCT) 1, plasma membrane monoamine transporter (PMAT), carnitine/cation transporter 1, OCT3, and serotonin reuptake transporter (SERT) that are involved in metformin absorption. Rare renal impairment has been recognised as another possible side effect of metformin, particularly in individuals with prior renal dysfunction or those who are predisposed to factors such as advanced age or concurrent nephrotoxic medicines [64]. Additionally, the use of metformin in pregnant women with gestational diabetes has been a subject of concern, as an emerging report suggests a possible association between medicine and adverse outcomes for both mother and fetus [65].

Some individuals with certain genetic variants are more likely to develop adverse drug reactions caused by certain drugs (Table 1). Pharmacogenomic interpretation of these variants using PRS is advantageous in developing a more personalised approach [26]. Studies have shown that variability in genes encoding drug receptors, transporters, and metabolizing enzymes affects an individual’s response toward oral antidiabetic drugs (OADs) [65]. Identifying responsible variants via genetic testing and altering therapies can result in a drastic shift in the recovery of patients. Jahnavi et al. [1] reported that after identification of the mutation in causal variants like KCNJ11 and ABCC8, affected children showed positive responses to the switch of therapy from insulin to oral sulfonylurea. Considering all these factors, incorporating PRS into clinical practices offers a promising avenue for patient categorization based on their susceptibility to a particular complication and response to a particular drug.

Table 1.

List of the genetic variants that contribute to the different drugs used to treat T2DM

Drugs for patient with diabetes	Family/class of drug	Mechanism of action	Pathway impacted	Gene involved	Shared genes between European and Asian
Metformin	Biguanides	Insulin sensitization	AMPK	SLC22A1 SLC2A2 SLC47A1	SLC22A1 SLC2A2
Glibenclamide	Sulfonylureas	Inhibits the ATP-sensitive potassium channels (KATP) on the pancreatic β-cell membrane	AMPK	KCNJ11 ABCC8	KCNJ11
Glipizide		Partial block of the potassium channels in the beta cells of the pancreatic islets	Metabolic amplifying pathway	ABCC8	ABCC8
Glimepiride		Stimulates insulin release	KATP channel pathway	ABCC8 KCNJ11	ABCC8 KCNJ11
Gliclazide		Closure of KATP channels	Metabolic amplifying pathway	ABCC8	ABCC8
Repaglinide	Meglitinides	Stimulation of insulin by closing KATP channels in pancreatic beta cells	KATP channel pathway	ABCC8 KCNJ11	ABCC8 KCNJ11
Voglibose	Alpha-glucosidase inhibitors	Inhibits alpha-glucosidase enzymes	Glycolysis	Not involved
Acarbose		Inhibits alpha-glucosidase enzymes	Carbohydrate digestion and absorption pathway	Not involved
Sitagliptin	DPP4 inhibitors	Impacts the activity of the DPP4 enzyme	Incretin pathway	Not involved
Vildagliptin		Inhibits DPP4 enzyme	Incretin pathway	Not involved
Linagliptin		Inhibits DPP4 enzyme	Incretin pathway	Not involved
Pioglitazone	Thiazolidinediones	Activates PPARγ receptors	ERK/MAPK pathway	PPARG ADIPOQ	PPARG
Rosiglitazone		Activates PPARγ receptors	ERK/MAPK pathway	PPARG PPAR ADIPOQ	PPARG
Exenatide	GLP-1 receptor analogues	Stimulates GLP-1 receptors	GLP-1 signaling pathway	GLP1R	GLP1R
Liraglutide		Activates GLP-1 receptors	GLP-1 receptor pathway	INS, GCG, PDX1, GLP1R, TCF7L2, SLC2A2	GLP1R SLC2A2 TCF7L
Semaglutide		Activates GLP-1 receptors	GLP-1 receptor pathway	INS, GCG, PDX1, GLP1R, TCF7L2, SLC2A2	GLP1R SLC2A2 TCF7L
Dulaglutide		Activates GLP-1 receptors	GLP-1 receptor pathway	INS, GCG, PDX1, GLP1R, TCF7L2, SLC2A2	GLP1R SLC2A2 TCF7L

Translation of PRS from Discovery to Clinic

Recent advancements in the field of PRS have made it a little bit easier to analyse data specifically for different ethnic groups. Earlier, it was difficult because of the lack of effect sizes in terms of summary statistics, so we had to rely on European data only. Recently a study [26] described the DIAMANTE (DIAbetes Meta-ANalysis of Trans-Ethnic Association Studies) consortium, which provides summary statistics for all the diverse ethnic groups. This can be used to calculate PRS for the Indian population as well. Clinicians can prescribe PRS directly from consumers, or they can buy it as a test. Some genetic testing companies offer PRS at a reasonable price; however, it is less popular as a result of the limited evidence of the utility of PRS (Fig. 4).

Fig. 4.

Clinical utility of polygenic risk scores (PRS) in conjunction with lifestyle risk factors, facilitating the management of diabetes and reducing the likelihood of associated complications

The first phase is to develop and validate PRS in large population cohorts using statistical genetics, which may also help determine areas with already higher probability of disease. After the computation of GWAS data comes a laboratory-level phase in which the collected blood or saliva sample is used to isolate the serological or genetic marker of that individual and is further processed for PRS calculation with publicly available loci and weights [66]. After principal components adjustment and standardisation, the PRS pipeline is followed. Further, these values are compared to the PRS threshold for odds ratio (OR) > 2 [30, 66, 67]. This is followed by risk categorization and report creation. This report is then contextualized by physicians; by comparing it with other risk factors, comorbidities, and patient preferences, it is utilized for patient care. In case any new variant is discovered while genotyping, it is filtered and annotated. It is further processed to ACMG-AMP (American College of Medical Genetics and Genomics–Association for Molecular Pathology) classification, followed by Sanger sequencing for confirmation.

Strategies and Solutions for Addressing Clinician Concerns and Enhancing Patient Understanding in PRS Implementation

The way reported risk is understood and communicated to patients is the main obstacle to increasing the reliability and popularity of PRS among doctors and patients. There are no appropriate guidelines supporting PRS use currently. Interpreting monogenic risk allele results, for which appropriate criteria have already been published, is much simpler than this. Clinicians need to have extensive expertise and experience to effectively communicate early assessment results to patients, as they can be challenging to interpret [68]. Although primary care providers mostly understand the information, patients mostly struggle to understand the information. Patient counselling plays a crucial role in enhancing patient understanding of PRS by explaining how multiple genetic variants contribute to overall risk, interpreting the results in the context of family history and lifestyle factors, and discussing potential implications for healthcare decision-making and risk management strategies [69]. Patient empowerment is an important determinant of diabetes self-management, along with behavioural monitoring that emerges as a fundamental facilitator. Many studies have reported that increased genetic awareness among patients increases the changes in health behaviour attitude [70]. In addition to feeling better, patients have demonstrated a considerable increase in awareness of their condition, indicating a favourable overall result of patient counselling. According to a recent study, patients and healthcare professionals are optimistic about PRS if clinical reports are carefully prepared. PRS was even perceived by practitioners as a logical progression of their current practice. Through the use of this tailored approach, individuals can make more informed decisions about their health and better understand the complexity of their genetic risk profile. Table 2 lists the possible concerns of clinicians in implementing PRS in their clinical practice. It also provides the best possible solutions to overcome their concerns. Similarly, Table 3 depicts the major studies showing the utility of PRS and discusses relevant studies considering both pros and cons for the domains including prediction of diabetes, diabetes-related complications, adverse effects, response to pharmaceutical agents, and response to lifestyle measures for prevention of diabetes or management of diabetes.

Table 2.

List of possible concerns of clinicians in implementing PRS and their possible solutions

Possible concerns of clinicians in implementing PRS	Best possible solution
Lack of understanding by physicians	Development of proper evidence-based guidelines for their use
Perceived lack of cost benefit	Making PRS cost-effective to be available to diverse socioeconomic groups
Perceived time constraints	Development of educational aids
Lack of ability to narrate the results in a person-friendly language	Organizing training camps and workshops for clinicians to make it easy and less time-consuming

Table 3.

In-depth analysis of diabetes prediction models utilizing polygenic risk scores (PRS)

Study references	Study design	Population characteristics	PRS methodology	Strength of association	Pros	Cons
Ashenhurst et al. [51]	Observational cohort study	Hispanic/Latino populations	11,999 SNPs, LD pruning, PRS calculation	European descent (AUC = 0.656), and Hispanic/Latino individuals (AUC = 0.635)	Clinical relevance, generalized to another populations	Limited data availability, excluding rare variants with large effects in the PGS, not accounting for all risk factors in the evaluation
Hahn et al. [30]	Prospective cohort study	Korean population	239,062 SNPs, LD-based clumping	p < 5 × 10−8, minor allele frequency (< 1%), false-positive error threshold = 0.05	Diverse ethnic validation, comprehensive study design, longitudinal analysis, machine learning techniques	Limited data accessibility, limited generalizability, data filtering
Song et al. [75]	Observational cohort study	Korean population	1004 SNPS, AUC of 0.658	OR 1.964, 95% CI 1.901–2.028, p ≈ < 0	Incorporation of PRS for risk prediction, genetic and environmental interaction analysis, large sample size and independent validation cohort	Moderate predictive accuracy, limited clinical risk factors, exclusively Korean population, complexity of interaction analyses
Liu et al. [73]	Prospective cohort study	Chinese population	80 SNPs	Hazard ratio 2.06, 95% CI 1.42–2.97	Incorporation of a PRS to refine risk stratification, practical implications, long-term follow-up, holistic approach	Data limitations, generalizability, ethical considerations, resource intensive
Shitomi-Jones et al. [65]	Case–control	Jat Sikhs North India	239,062 SNPs, ROC curves	t value = − 12.2, and p < 0.001	Genetic isolation, robust statistical analysis, clinical relevance	Limited sample size, population specificity, focused only on selected genetic loci, ethical and clinical considerations
PRS in prediction of diabetes-related complications
Shitomi-Jones et al. [65]	Literature review	European and Asian population	A weighted sum of risk allele counts from variants, integrating effect sizes from GWAS	NIL	Comprehensive approach, Inclusion of diverse populations Utilization of PRS increases predictability of disease onset and therapeutic responses	Variants with low allele frequency and rare variants may require larger sample sizes for robust analysis, need for ethical considerations and communication
Tremblay et al. [58]	Cohort study	Multiple populations	598 SNP	AUC of 0.65, 95% CI	Longitudinal observation, validation in diverse cohorts, ethical approval and participant consent	Limited diversity, small external cohorts, age and diagnosis criteria, potential bias and confounding factors
O’Sullivan et al. [74]	Systematic review or a meta-analysis	UK population	NIL	OR, hazard ratios, (AUC-ROC) curve	Clinical utility, public availability, enhanced risk assessment, expansion of biobanks	Limited diversity, dependency on specific datasets, need for improved clinical phenotyping
Response to lifestyle measures for the prevention of diabetes or management of diabetes
Almalki et al. [71]	Observational or cohort study	Arabian population	NIL	NIL	Guide for clinicians on diagnosing and managing central diabetes insipidus in adults	Focus on adult patients might limit its applicability to pediatric cases, and the content could be complex for non-specialists
Karachaliou et al. [72]	Observational study	Mixed population	NIL	NIL	Critical public health issues by focusing on diabetes prevention and care in low-income settings, offering valuable insights for policymakers and healthcare providers	Limited generalizability to other regions, and the article might lack novel solutions or practical applicability in resource-limited environments
Response to pharmaceutical agents
Pollin et al. [76]	Observational study	Caucasian, African American, American Indian, Hispanic, and Asian/Pacific Islander populations	32 lipid-associated SNPs	Correlation coefficients ranging from 0.07 to 0.17, p values 5 × 10−5 to 1 × 10−19	Insight into genetic influence, clinical implications, diverse population	Complexity of genetic factors, limitations of interventions, additional risk factors
Varga et al. [77]	Cohort study	Multiethnic composition	234 SNPs	β = − 0.11 µmol/L per genetic risk scores risk allele; 95% CI, − 0.188 to − 0.033; P = 5 × 10–3; Pinter action = 1 × 10–3 for lifestyle versus placebo)	Genetic, lifestyle, and phenotypic data, offers a holistic view of factors contributing to lipid and lipoprotein variations Enhances the generalizability of the study’s findings, potentially benefiting diverse populations	Individuals predisposed to diabetes, may limit the generalizability of its findings to broader populations Usage of PRS may not capture the full spectrum of genetic variations, potentially limiting the depth of genetic insights
Kim et al. [78]	Prospective longitudinal study	Korean	1004 SNPs	AUC = 0.658 (95% CI 0.651–0.666)	Longitudinal design, ethnic relevance, genetic relevance, interaction analysis	Limited generalizability, predictive performance, lack of follow-up data, potential biases

AUC area under the curve, CI confidence interval, LD linkage disequilibrium, OR odds ratio, ROC receiver operating characteristic curve, SNP single nucleotide polymorphism

Conclusions

Polygenic risk scores (PRS) hold immense promise in revolutionising the landscape of precision medicine for diabetes. Analysing the familial background of an individual gives valuable insights into the genetic spectrum of diabetes. Hence the estimation of heritability plays a significant role in determining the degree to which genetic factors contribute to diabetes susceptibility within families apart from other environmental and behavioural factors. By tailoring genetic information gathered with the help of PRS in clinics, one can easily enhance the treatment efficacy and improve the patient’s outcome towards T2DM. Instead of treating all patients based solely on symptoms, it may be more practical to concentrate on those who are at the highest risk or most likely to respond to therapy, given factors including cost, side effects, resistance development, and patient preferences. In consideration of the concerns of clinicians, PRS still requires some adjustments before physicians fully embrace it. Improving PRS’s accessibility and ease of use at a reasonable cost will make composite diagnoses more readily accepted by patients and medical professionals. A sizable healthcare system will need to get involved, but if PRS successfully starts patients on therapy early, it will enhance healthcare quality and extend cost savings. Because PRS requires more time and is more challenging to interpret, many doctors hesitate to employ it. This may be fixed with the right policies in place and training sessions conducted appropriately. More focused population-based GWAS research is required to prevent bias when determining scores from one group to another. This will aid in the identification of other novel and population-specific variants. A significant disadvantage of PRS, aside from ethnicity bias, is that it alone assesses an individual’s genetic liability. Not all genetic variants causing complicated illnesses are captured by the polygenic variants found using GWAS. Despite being more expensive, trickier to interpret, and with less population research accessible on PRS, it still has a remarkable ability to detect early risk stratification and disease management. More research is needed to evaluate the utility and cost-effectiveness of PRS in order to fully reap its benefits. Furthermore, patient counseling aids patients in understanding PRS by outlining how different genetic variations affect overall risk, interpreting the results in the context of family history and lifestyle factors, and discussing potential ramifications for healthcare decision-making and risk management techniques.

Author Contributions

Omna Singh, Madhur Verma, and Sanjay Kalra contributed to the concept and design of the study; Omna Singh and Nikita Dahiya were involved in the data analysis, interpretation of the results, and writing the first draft of the study; Madhur Verma, Sabyasachi Senapati, Rakesh Kakkar, and Sanjay Kalra contributed to developing and revising the manuscript extensively. All the authors have read and approved the final version of the manuscript. Omna Singh is the guarantor of this work and, as such, has full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Funding

No funding or sponsorship was received for the publication of this article.

Data Availability

Raw data are available from the lead author on reasonable request.

Declarations

Conflict of Interest

Sanjay Kalra is an Editorial Board member of Diabetes Therapy. Sanjay Kalra was not involved in selecting peer reviewers for the manuscript or any subsequent editorial decisions. Omna Singh, Madhur Verma, Nikita Dahiya, Sabyasachi Senapati, and Rakesh Kakkar have nothing to disclose.

Ethical Approval

This article does not contain any new studies with human participants or animals performed by any of the authors.