, Introduction
Polypharmacy (PP), defined as the regular and simultaneous use of multiple medications, is a major public health burden [1]. PP is associated with an increased risk of adverse drug events, drug–drug interactions, more frequent hospitalization, poor adherence, and greater health care costs [2,3]. Furthermore, several studies have reported that PP is associated with an increased risk of all-cause and cardiovascular mortality, morbidity, and quality of life [4–6].
Patients with chronic kidney disease (CKD), as well as those with end-stage kidney disease (ESKD), not only have numerous comorbidities but also experience various CKD-related complications that require medications. Therefore, PP is common across all stages of CKD [7–12]. The prevalence of PP in patients with CKD has been reported to range from 27 to 91% and increases with the progression of CKD [7–9,13].
PP is common in kidney transplant recipients (KTRs). The prevalence of PP is higher in KTRs than in patients with CKD stages 3 to 5 [14]. KTRs require long-term immunosuppressive therapy along with other medications to manage comorbidities, such as hypertension, diabetes mellitus, and infections. The use of immunosuppressive medications, although crucial for the success of kidney transplantation (KT), introduces additional complexity to the pharmacological regimen [15]. These drugs may interact with other medications, affecting their efficacy or causing harmful side effects. Moreover, nonadherence to prescribed regimens or errors in medication management can lead to adverse clinical outcomes, including graft rejection, infections, and even graft loss [16]. Although PP is an inevitable aspect of KTR management, its impact on patient outcomes has not been sufficiently investigated. In this study, we aimed to investigate the influence of PP on KTRs, focusing on PP’s effects on patient health, transplant outcomes, and the management of potential complications.
Methods
Study design and participants
The Korean Cohort Study for Outcomes in Patients with Kidney Transplantation (KNOW-KT) is a multicenter, prospective, observational cohort study conducted at eight centers in the Republic of Korea [17]. The detailed design and methods of the KNOW-KT were previously published (NCT02042963 at http://www.clinicaltrials.gov). A total of 1,080 KT recipients were enrolled in the KNOW-KT between July 2012 and August 2016. We excluded 46 patients who underwent baseline examinations without subsequent visits, and 62 patients with no available data at 1 year after KT. Finally, 972 patients were enrolled (Fig. 1).
The study was conducted in accordance with the principles of the Declaration of Helsinki and the Declaration of Istanbul, and the study protocol was approved by the Institutional Review Board of the participating centers (No. 4-2014-0290, 2014AN0158, 2013AN005). Informed consent was obtained from all participants.
Data collection and measurements
Baseline demographic and clinical characteristics were collected at the time of KT as follows: recipient information (age, sex, weight, height, smoking history, cause of ESKD, dialysis modality, duration of dialysis, comorbidities, and history of coronary artery disease [CAD]), donor information (age, sex, height, weight, comorbidities, expanded criteria donor and serum creatinine level at procurement), and transplantation-related information (date of KT, prior KT history, and donor-recipient relationship). Laboratory data (white blood cell count, hemoglobin, triglyceride, high-density lipoprotein cholesterol, and urine protein/creatinine ratio [UPCR]), resting office blood pressure, prescribed medications, and events (death, graft failure, and cardiovascular events) were measured at each annual visit. Body mass index (BMI) was calculated as weight (kg) divided by height in meters squared (m2), and estimated glomerular filtration rate (eGFR) was calculated using the 2021 CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) equation [18].
Exposure
We collected annual data on the prescribed medications including immunosuppressants (tacrolimus, cyclosporin, mycophenolate mofetil [MMF], mycophenolic acid [MPA], mizoribine, azathioprine, sirolimus, everolimus, or steroid), antihypertensive drugs (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, diuretics, beta-blocker, direct renin inhibitor, nitrate, calcium channel blocker, alpha-blocker, or minoxidil), anticoagulants (aspirin, warfarin, ticlopidine, or clopidogrel), lipid-lowering agents (statins, ezetimibe, fibrates, or nicotinic acid), vitamin D or calcium supplements (vitamin D replacement, vitamin D analog, calcimimetics, oral phosphate, or bisphosphonate), and antidiabetic medications (insulin, sulfonylurea, alpha-glucosidase, biguanides, meglitinides, thiazolidinediones, incretin mimetics, or DPP4 inhibitors). Medication data were collected annually at the time of each scheduled follow-up visit, beginning with study enrollment. Only medications prescribed at the outpatient clinic during the follow-up visits were included in the analysis. Each active ingredient in a combination drug was counted as a separate medication. Medications obtained through self-administration or prescribed at external institutions were not captured; thus, they were not included in the medication counts. PP is generally defined as the use of five or more medications and excessive PP as 10 or more. Given the limited number of patients taking 0–4 medications in our cohort, we classified patients into two groups at 1-year posttransplantation: non-excessive PP (non-ePP; <10 medications) and excessive PP (ePP; ≥10 medications).
Outcomes
The primary outcome was all-cause mortality, which was defined as death from any cause occurring after KT. We examined three secondary outcomes: graft failure, death-censored graft failure, and cardiovascular events. Graft failure was defined as return to dialysis, retransplantation, or death after KT. Death-censored graft failure was defined as return to dialysis or retransplantation censored for death with a functional graft. Cardiovascular events were defined as the occurrence of CAD, congestive heart failure, cerebrovascular disease (including ischemic and hemorrhagic stroke), or peripheral vascular disease after KT. CAD included myocardial infarction or unstable angina requiring coronary revascularization such as percutaneous coronary intervention or coronary artery bypass graft surgery. The outcomes were assessed annually during the follow-up period. Each patient was followed up until the day that the outcome developed (death or graft loss), withdrawal from the study, loss to follow-up, or December 31, 2022.
Statistical analysis
Continuous variables are expressed as mean ± standard deviation or median (interquartile range [IQR]), and categorical variables are expressed as percentages or frequencies. Continuous variables were compared using the Student t test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables. Categorical variables were compared using the chi-square test or Fisher exact test.
To reduce potential confounding factors when comparing outcomes between the groups, we used the inverse probability of treatment weighting (IPTW) method based on propensity scores (PSs). For IPTW, the weight was calculated as 1/PS for the treated patients (ePP group) and 1/1-PS for untreated patients (non-ePP group). PSs were estimated using a multivariable logistic regression model in which the group (non-ePP or ePP) was regressed on selected covariates (all covariates are listed in Table 1). The balance of covariates between the groups after IPTW was checked using the absolute value of the standardized difference between the groups, where a value <0.1 was considered a negligible difference and a value range of 0.1 to 0.2 was considered a small difference [19].
For the outcomes, we calculated the unweighted and weighted incidence rates (defined as the number of events per 1,000 person-years of follow-up). For time-to-event outcomes, a Kaplan-Meier survival analysis was conducted, and the differences between the groups were assessed using the log-rank test. Additionally, Cox proportional hazards regression models were used to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) for each outcome. The assumption of proportional hazards in IPTW analysis was evaluated using the Schoenfeld residual method. A two-tailed p-value of <0.05 was considered statistically significant. For the sensitivity analysis, winsorization was applied at the 1st and 99th percentiles to stabilize extreme IPTW weights, and Cox proportional hazards regression was reanalyzed to confirm the robustness of the findings. For the competing risk analysis, a cause-specific Cox proportional hazards model was used, considering death as a competing risk factor for cardiovascular events. This approach allowed us to estimate the risk of cardiovascular events while accounting for the influence of mortality as a competing event. All statistical analyses were performed using the R software version 4.3.1 (R Foundation for Statistical Computing).
Results
Baseline characteristics
A total of 972 patients were included in this study. The median age of patients was 47 years; 37.4% of the patients were male, 91.4% had hypertension, 25.2% had diabetes mellitus, 5.9% had a history of CAD, 61 (6.3%) had previously received a kidney transplant, and 205 (21.1%) had preemptive transplantation. The causes of ESKD were diabetic nephropathy, hypertensive nephropathy, glomerulonephritis, and other causes in 20.4%, 22.7%, 31.2%, and 25.7% of patients, respectively. The median dialysis duration before transplantation was 26 months. The mean eGFR at 1 year after KT was 62.7 mL/min/1.73 m2, and the median UPCR was 0.1 g/g creatinine. The mean medication count at 1 year after KT was 9.8, and PP was noted in 478 patients (49.2%). At 1-year posttransplantation, most KTRs were prescribed between seven and 13 medications, reflecting a substantial medication burden (Fig. 2). This pattern supports the use of a 10 or more medications threshold to define ePP, as only a small portion of KTRs were prescribed fewer than five medications.
The baseline characteristics according to ePP status are shown in Table 1. Before weighting, patients in the ePP group had a higher BMI, were more likely to receive a preemptive kidney transplant, were less likely to receive an expanded criteria deceased donor kidney transplant, and had a shorter duration of dialysis than those in the non-ePP group. The prevalence of diabetes mellitus, dyslipidemia, and a history of CAD was more frequently observed in patients in the ePP group than in the non-ePP group. Overall laboratory data 1 year after KT were comparable between the groups, except for fasting glucose levels. After weighting, no significant differences were observed across all baseline characteristics between the groups, with almost all standardized difference values <0.1 (Table 1).
The distribution of prescribed medication classes at 1 year after KT was compared between the non-ePP and ePP groups (Supplementary Table 1, available online). Patients in the ePP group received a significantly higher number of agents across all drug classes, including immunosuppressants, antihypertensive agents, antidiabetic agents, lipid-lowering agents, and vitamin D or calcium supplements (all p < 0.05). The distribution of immunosuppressants at 1 year after KT was compared between the non-ePP and ePP groups (Supplementary Table 2, available online). A higher proportion of patients in the ePP group were prescribed mycophenolate, including both MMF and MPA, as well as steroids, whereas tacrolimus use was similar between groups.
Outcomes
The weighted median follow-up for death was 6.9 years (IQR, 5.9–7.9 years). During follow-up, 36 KTRs died, with an unweighted incidence rate of 4.8 and 6.5 per 1,000 person-years in the non-ePP and ePP groups, respectively (Table 2). Among the patients who died during follow-up, the most common causes of death were infection (n = 11) and malignancy (n = 4). There were no significant differences in the distribution of causes of death between the non-ePP and ePP groups (p = 0.80).
Graft function was compared between the non-ePP and ePP groups after KT. Before IPTW, the graft function tended to be slightly lower in the ePP group throughout the follow-up period. After IPTW adjustment, a similar pattern was observed, with no significant differences in graft function between the groups at any time point (Table 3).
During follow-up, 97 KTRs experienced graft failure, with an unweighted incidence rate of 15.1 and 16.0 per 1,000 person-years in the non-ePP and ePP groups, respectively (Table 2). Similar results were found for the other outcomes (death and cardiovascular events), with the ePP group having a higher incidence rate than the non-ePP group (Table 2). No significant differences were observed in the incidence rates of all-cause mortality, graft failure, or death-censored graft failure after weighting. However, there was a significant difference in the incidence rate of cardiovascular events between the groups (incidence rate ratio, 1.7; 95% CI, 1.14–2.41; p = 0.009) (Table 2). In addition, no significant associations were observed between all-cause mortality, graft failure, death-censored graft failure, or cardiovascular events and immunosuppressant use, number of immunosuppressants, or total number of medications (data not shown).
Patients in the ePP group had lower survival rates than those in the non-ePP group (p = 0.001). Significant differences in the cumulative probability of graft failure, death-censored graft failure, and cardiovascular events were observed between the two groups (weighted log-rank p = 0.03, p = 0.03, and p < 0.001, respectively) (Fig. 3).
After weighted Cox proportional hazard analysis, no significant differences were observed in the rates of all-cause mortality, graft failure, or death-censored graft failure between the non-ePP and ePP groups (HR, 1.30 [95% CI, 0.78–2.02] for all-cause mortality; HR, 1.00 [95% CI, 0.73–1.30] for graft failure; HR, 0.90 [95% CI, 0.61–1.21] for death-censored graft failure) (Table 2). However, the risk of cardiovascular event was significantly higher in the ePP group than in the non-ePP group (HR, 1.80; 95% CI, 1.07–2.96) (Table 2). A sensitivity analysis using winsorization to address the potential impact of extreme IPTW weights showed that the results remained consistent. The increased risk of cardiovascular events in the ePP group remained significant (HR, 1.78; 95% CI, 1.07–2.96; p = 0.03).
After applying IPTW and considering competing risks, the risk of cardiovascular events was 74% higher in the ePP group than in the non-ePP group (HR, 1.74; 95% CI, 1.05–2.90; p = 0.03). Additionally, when death was accounted for as a competing risk, it significantly influenced the occurrence of cardiovascular events, increasing the risk by 3.23 times (HR, 3.23; 95% CI, 1.39–7.51; p = 0.006). These results indicated a statistically significant association between ePP and the risk of cardiovascular events.
Subgroup-specific associations between excessive polypharmacy and cardiovascular events
The association between ePP and cardiovascular events was evaluated according to baseline characteristics and medication use. ePP was significantly associated with an increased risk of cardiovascular events among male recipients and lipid-lowering agent users (Supplementary Table 3, available online). In contrast, ePP was not significantly associated with cardiovascular events in recipients receiving antidiabetic agents, antihypertensive drugs, vitamin D or calcium supplements, or anticoagulants, although the direction of the effect was generally consistent across subgroups.
After adjusting for the use of major immunosuppressive agents (tacrolimus, MMF or MPA, and steroids), ePP remained significantly associated with an increased risk of cardiovascular events (HR, 1.91; 95% CI, 1.12–3.25; p = 0.02), indicating that the relationship was independent of immunosuppressive burden.
Discussion
In this study, the incidence of ePP was 49.2%, with an average of 9.8 medications at 1-year posttransplantation. Before IPTW, the ePP group had a higher prevalence of diabetes mellitus, dyslipidemia, and a history of CAD; however, after weighting, the baseline characteristics were well-balanced among the groups. No significant differences were observed between the groups in all-cause mortality, graft failure, or death-censored graft failure. However, the ePP group had a significantly higher risk of cardiovascular events than the non-ePP group. This finding remained robust after adjusting for potential confounders through IPTW, even after accounting for death as a competing risk.
PP is a well-recognized issue in KTRs owing to the necessity of lifelong immunosuppressive therapy and management of comorbid conditions such as hypertension, diabetes mellitus, and dyslipidemia. The prevalence of PP in KT populations varies across studies, with reported rates ranging from 40% to 60% [14,20–22] depending on the definition and threshold used. In this study, 49% of KTRs met the criteria for ePP, which is consistent with the results of previous studies.
Despite its potential risks, the direct association between PP and clinical outcomes such as mortality, graft failure, and cardiovascular disease remains uncertain. In patients with CKD including ESKD, several studies have reported that PP was associated with clinical outcomes, including mortality, cardiovascular outcomes, CKD progression, and lower quality of life [8,11–14]. However, in KTRs, studies investigating the association between PP and clinical outcomes are often limited by small sample sizes and a lack of investigation on graft outcomes, which are crucial for KTRs [5,14,21]. In this study, no significant relationship was found between ePP and all-cause mortality, graft failure, or death-censored graft failure after applying IPTW to adjust for baseline differences between groups. Importantly, a significant association was noted between ePP and cardiovascular events, with an 80% increased risk (HR, 1.8; p = 0.03) in the ePP group. Competing risk analysis indicated that the association between PP and cardiovascular events remained statistically significant after accounting for death as a competing event (HR, 3.23; p = 0.006). Although this finding suggests a potential link between PP and increased cardiovascular risk, it does not establish a causal relationship. PP may serve as a surrogate marker for overall disease burden, rather than functioning as an independent risk factor. Several pathophysiological mechanisms have been proposed to explain the contribution of PP to cardiovascular risk. The use of multiple medications increases the probability of drug-drug interactions and cumulative adverse effects, including electrolyte imbalance, electrocardiographic abnormalities, and dysregulation of glucose and lipid metabolism, which may predispose to cardiovascular complications [23-26]. Some immunosuppressants and antihypertensive agents also have potential proatherogenic effects [25,26]. Additionally, PP is associated with arterial stiffness, endothelial dysfunction, and chronic inflammation [27,28]. A high medication burden may reduce adherence, potentially resulting in suboptimal control of conventional cardiovascular risk factors, such as hypertension, diabetes mellitus, and dyslipidemia [29]. The phenomenon of the prescribing cascade, whereby adverse drug effects are misinterpreted as new clinical conditions leading to the initiation of additional medications, may exacerbate medication-related harm. Age-related pharmacokinetic and pharmacodynamic changes, as well as reduced graft function, may alter drug metabolism, increase systemic exposure to medications with cardiovascular effects, and amplify adverse cardiovascular outcomes [30]. Therefore, the observed association between PP and cardiovascular events in KTRs should be interpreted with caution. Nonetheless, the findings emphasize the importance of regular medication reviews, individualized pharmacological assessments, and strategies for comprehensive cardiovascular risk management in KTRs.
To further clarify whether immunosuppressive regimens contributed to the association between PP and cardiovascular risk, we examined the distribution of immunosuppressants across groups. A significantly higher proportion of patients in the ePP group received mycophenolate-based agents and steroids, whereas tacrolimus use was comparable between groups. These findings suggest that the observed cardiovascular risk may be partially influenced by the immunosuppressive burden. However, the widespread use of triple therapy in KTRs and the absence of large differences in core immunosuppressant types, it is more likely that ePP itself—rather than any specific immunosuppressive regimen—accounts for the increased cardiovascular risk.
Furthermore, a subgroup analysis was conducted to investigate the association between PP and cardiovascular risk. ePP was significantly associated with cardiovascular events, particularly in male recipients and those taking lipid-lowering agents. Although no statistically significant associations were observed in patients receiving antihypertensive, antidiabetic, or anticoagulant agents, the direction of the effects was generally consistent. These findings are consistent with established cardiovascular risk factors, such as sex and dyslipidemia [31], suggesting that the adverse impact of PP may be more pronounced in individuals with an already elevated cardiovascular risk. This may reflect a potential interaction between specific medication classes and cardiovascular outcomes; however, a causal relationship could not be definitively established. While immunosuppressive burden was considered a potential contributor, multivariable Cox analysis showed that ePP remained significantly associated with cardiovascular events, whereas only tacrolimus use showed a protective association. These results suggest that the relationship between PP and cardiovascular outcomes cannot be solely attributed to the type or number of core immunosuppressive agents.
To address potential confounding factors, we applied IPTW, a PS-based method that helps to create a pseudo-population in which the treatment assignment is independent of the measured baseline characteristics. IPTW is widely used in observational studies to mimic the effects of randomization, thereby improving causal inference [32]. However, IPTW has several limitations. Extreme weights can introduce instability into the estimates, and this method does not eliminate unmeasured confounding factors. Additionally, the effective sample size may have been reduced due to the reweighting process. Despite these limitations, IPTW was deemed appropriate for this study because it successfully balanced the key baseline characteristics between the ePP and non-ePP groups, as demonstrated by the standardized mean differences approaching zero. This approach allowed a more reliable comparison of clinical outcomes, thereby reducing the risk of confounding biases that may have influenced the results of previous studies.
Our study had several limitations. First, as this was an observational cohort study, not all prescribed medications, over-the-counter medications, or herbal medicines were included. Second, we did not consider medication adherence, dose adjustments, or changes in therapy over time, which may have influenced the observed associations. Third, while IPTW effectively balanced the baseline characteristics, residual confounding due to unmeasured variables could not be entirely excluded. Finally, although a competing risk analysis was performed, additional validation is needed to confirm our findings and further explore the causal pathways linking PP and cardiovascular risk in KTRs. However, we collected data on almost all major classes of medications, and the increased medication burden in patients was largely affected by the increased medication counts in the major classes [7]. Furthermore, this is the first study to investigate the association between PP and clinical outcomes in KTRs using the IPTW methods to achieve balance between the two comparison groups, strengthening the validity of these findings.
In conclusion, ePP is highly prevalent in KTRs and is associated with an increased risk of cardiovascular events but not with all-cause mortality or graft failure. By applying IPTW and competing risk analysis, we demonstrated that cardiovascular risk remained significant even after adjusting for potential confounders. Identifying patients at higher risk of medication-related adverse outcomes could facilitate timely interventions and potentially improve cardiovascular outcomes. Further prospective studies are warranted to determine the causal pathways, assess medication adherence, and evaluate the clinical effectiveness of structured PP management interventions in this population.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement table 1
Print
