Solid organ transplant recipients are at increased risk of developing coronary artery disease (CAD) [1–2]. Cardiovascular risk factors, previous end-stage organ disease, and pharmacological immunosuppressive therapy may contribute to the development of CAD [3]. For example, Fussner et al. reported that 30% of liver transplant (LT) recipients suffer from CAD events at 8 years after liver transplantation [4].
Rates of incident CAD in LT recipients have rarely been examined in large national registries, nor has their rate of incident CAD been compared with the general population, which would be an appropriate control group. In addition, it remains unclear whether cardiovascular risk factors that predict CAD events in the general population are equally predictive in LT recipients [5].
Here, we assess the rate of CAD among LT recipients compared with matched controls drawn from the general population. In addition, we evaluate whether conventional cardiovascular risk factors, as captured through registry data, are predictive of incident CAD in this population.
2Materials and Methods2.1Study designWe conducted a nationwide, registry-based cohort study to compare the rates of new CAD events among all LT recipients in Sweden and matched controls from the general population between 1987 and 2020.
The analysis was based on the Decoding the epidemiology of liver disease in Sweden (DELIVER) cohort, assembled using nationwide health registries in Sweden [6]. The cohort comprises all individuals in Sweden diagnosed with acute or chronic liver disease, along with matched controls from the general population, matched on age, sex, municipality of residence, and calendar year of transplant. Information on the data sources, ICD codes used to define liver transplantation, indications for transplantation, cardiovascular risk factors, and CAD outcomes is provided in Supplementary Appendices 1–2 and Supplementary Tables 1–5.
2.2Study participantsWe included all adults with an ICD code for liver transplantation status in Sweden during the study period. The reference group included up to 10 matched controls for each LT recipient. Exclusion criteria for LT recipients and controls were age < 18, incorrectly coded or reused personal identification number at the time of the LT, or emigration or death at or before baseline. We defined baseline as the date of liver transplantation discharge or the corresponding matching date for controls.
2.3Cardiovascular risk factors and indications for liver transplantationCardiovascular risk factors, including hypertension, prior CAD, stroke, diabetes (type 1 and 2 combined), obesity, hyperlipidaemia, as well as kidney and lung disease, were classified as present or absent based on administrative coding from the national patient register using inpatient and outpatient data up to the liver transplantation discharge date.
The indications for liver transplantation—such as alcohol-related liver disease, viral hepatitis, metabolic dysfunction-associated steatotic liver disease (MASLD), acute liver failure, hepatocellular carcinoma (regardless of aetiology), rare conditions like Wilson’s disease and Budd-Chiari syndrome, autoimmune liver diseases, and multiple aetiologies excluding isolated hepatocellular carcinoma—were classified based on the data up to baseline.
2.4OutcomesThe primary outcome was the composite of new CAD events after liver transplantation, defined as incident myocardial infarction, coronary revascularisation procedures, angina, or CAD-related death.
2.5StatisticsContinuous data were reported as medians with interquartile ranges (IQR) and categorical variables were reported as frequencies and percentages. Wilcoxon rank-sum test, Pearson’s χ2, or two-sample t-test were used to test the differences where appropriate. Incidence rates were reported per 1,000 person-years. The primary outcome of CAD and the prespecified subgroup analyses based on the indication for liver transplantation and cardiovascular risk factors were estimated as crude and adjusted hazard ratios (HRs) using a Cox regression model [7]. The analysis was conditioned on the matching factors (age, sex, municipality, year of liver transplantation) and in the adjusted model further adjusted for cardiovascular risk factors (defined above) at baseline. The hazard ratio of incident CAD was compared between LT recipients by subtypes of liver disease leading to the need for transplant and their respective matched controls. Follow-up time was measured from baseline. Subgroup analyses were performed in prespecified groups, including age (≤50 vs >50 years), sex, hypertension, hyperlipidaemia, diabetes, prior CAD, stroke, obesity, chronic kidney disease, and chronic lung disease. These analyses were conducted separately in LT recipients and their matched controls to explore potential differences in the associations. The proportional hazards assumption was assessed using Schoenfeld residuals [8].
Sensitivity analyses assessed the incidence of post-transplant CAD in LT recipients and matched controls without a history of CAD prior to baseline. The Aalen–Johansen estimator (STATA: stcompet) estimated cumulative incidence of CAD, CAD related death, angina, and myocardial infarction, while accounting for the competing risk of death from other causes other than the outcome (non-CAD death). STATA 16.1 (StataCorp, College Station, TX) and R version 4.2.2 (2022–10–31 ucrt) ("Innocent and Trusting"), ggplot2, and forestplot were used for statistical analysis and figures.
2.6Ethical considerationsThe DELIVER study was approved by the Regional Ethics Committee in Stockholm, Sweden (reference number 2017/1019–31/1). The research was conducted in accordance with the Declaration of Helsinki and relevant national ethical and legal requirements for registry-based studies. All data were obtained from Swedish national health and population registers and were pseudonymized before access by researchers; therefore, informed consent from individual participants was not required. Data linkages were authorized and performed by the National Board of Health and Welfare and Statistics Sweden in compliance with Swedish data protection legislation.
3Results3.1Prevalence of coronary artery disease and baseline characteristics before liver transplantationA total of 3587 LT recipients were identified. After excluding 662 patients (18.4%) based on predefined criteria, 2925 remained in the final analysis (Fig. 1 depicts a flowchart of exclusions). The matched control cohort included 27,589 individuals. The median age was 54 years (IQR: 44–61), with 37% women (Table 1 presents baseline characteristics). Autoimmune liver disease (28.2%) and hepatocellular carcinoma (22.9%) were the leading indications for LT. Compared with controls, LT recipients had higher prevalence of cardiovascular risk factors. Diabetes was present in 21.8% vs. 3.3% and hypertension in 21.4% vs. 8.2% (both p < 0.001). The baseline prevalence of CAD was 4.5% in LT recipients versus 3.7% in controls (p = 0.02).
Baseline characteristics of matched controls and liver transplant recipients.
MASLD = Metabolic dysfunction-associated steatotic liver disease, HCC = Hepatocellular carcinoma, CAD = Coronary artery disease, Diabetes = included both type 1 and 2. Multiple diseases counted at least two liver diseases apart from hepatocellular carcinoma. Other = liver transplant indications not among predefined categories. * Age was reported as a continuous variable in years and as categorical variables for the intervals of 18–50 and >50 years. **“Note that the control population did not have the liver disease of interest. Here, the absolute numbers (percentages) reflect matched individuals based on age, sex, and municipality.”.
A total of 395 (14%) LT recipients experienced a new-onset CAD event, compared with 2680 (9.7%) controls over a median follow-up of 8.0 years (Fig. 2). The incidence rate was 16.9 per 1,000 person-years in LT recipients and 8.5 per 1,000 person-years in controls (Table 2). After multivariable adjustment, LT recipients had more than twice the rate of CAD (aHR=2.02; 95%CI = 1.80–2.30).
Risk of Coronary Artery Disease by Indication for Liver Transplantation. Forest plot displaying hazard ratios (HRs) for coronary artery disease (CAD) among liver transplant (LT) recipients, stratified by underlying liver disease. Each HR compares LT recipients to matched controls from the general population without transplantation (No LT). Estimates of HRs are derived from multivariable models adjusting for demographic and cardiovascular risk factors. Error bars represent 95% confidence intervals. ALD – Alcohol-related liver disease, MASLD – Metabolic dysfunction-associated steatotic liver disease, RareLD – Rare liver disease, AILD – Autoimmune liver diseases, HCC – Hepatocellular carcinoma, Multiple – Two or more concomitant liver diseases (excluding HCC), LT– Liver transplantation.
Incidence rates and hazard ratios (HRs) of coronary artery disease in liver transplant recipients (cases) and the background population (controls), categorized by transplantation indications and demographic factors.
MASLD = Metabolic-dysfunction associated steatotic liver disease, ALD = alcoholic-related liver disease, Multiple diseases counted at least two liver diseases apart from hepatocellular carcinoma, Other = liver transplant indications not among predefined categories, HCC = hepatocellular carcinoma, Age was measured in years, aHR = adjusted multivariable analysis including: previous coronary artery disease (CAD), diabetes, hypertension, hypercholesterolemia, obesity, chronic obstructive pulmonary disease, and stroke. Other = liver transplant indications not among predefined categories.
The cumulative incidence of CAD was consistently higher in LT recipients compared with matched controls across all follow-up periods (Fig. 3 and Table 3). At 90 days, the cumulative incidence was 1.61% (95%CI=1.20–2.12) in LT recipients and 0.47% (95%CI=0.39–0.56) in controls. At 1 year, the cumulative incidence increased to 2.73% (95%CI=2.18–3.73) and 1.35% (95%CI=1.21–1.49), respectively. At 5 years, cumulative incidence reached 6.87% (95%CI=5.94–7.88) in LT recipients and 4.14% (95%CI=3.90–4.39) in controls, and at 10 years, 12.63% (95%CI=11.27–14.07) and 7.63% (95%CI=7.28–7.99), respectively. In both groups, the cumulative incidence increased of CAD increised at a relatively steady rate over time, with no clear evidence of acceleration or deceleration.
Cumulative incidence of coronary artery disease in liver transplant recipients. Cumulative incidence curves of coronary artery disease (CAD) in liver transplant (LT) recipients versus matched general population controls (No LT), accounting for the competing risk of non-CAD death. Follow-up time is shown in years.
Cumulative incidences of coronary artery disease, considering death as a competing risk, stratified by the indications for liver transplantation and their matched controls.
95% CI = 95% confidence interval, HCC = hepatocellular carcinoma, MASLD = Metabolic-dysfunction associated steatotic liver disease, NA = Not applicable (due to few patients), ALD = Alcohol-related liver disease, LD = Liver disease, Multiple (except HCC) = counted at least two liver diseases apart from hepatocellular carcinoma. Other = liver transplant indications not among predefined categories.
Among individuals with baseline cardiovascular risk factors, incidence rates of CAD were generally higher in controls than in LT recipients (Table 2). For instance, CAD incidence in LT recipients with diabetes was 29 per 1,000 person-years, compared with 39 in controls with diabetes (Supplementary Table 6A–B). In contrast, LT recipients without such risk factors consistently showed higher CAD rates than controls.
In within-group analyses, most traditional risk factors were statistically significant predictors of CAD in controls but not in LT recipients, although point estimates were largely overlapping (Fig. 4). For example, diabetes was significantly associated with CAD in controls (aHR = 1.65; 95%CI = 1.41–1.93) but was not significant in the LT group. Among LT recipients, only previous CAD (aHR = 7.47; 95%CI = 5.71–9.77) and chronic kidney disease (aHR = 1.60; 95%CI = 1.15–2.23) remained significant predictors, whereas nearly all traditional risk factors were associated with CAD in controls.
Cardiovascular Risk Factors and coronary artery disease (CAD) Risk in Liver Transplant Recipients vs. Controls. Forest plot showing adjusted hazard ratios (HRs) for coronary artery disease (CAD) by subgroups of study participants with established cardiovascular risk factors at baseline, analyzed separately for liver transplant (LT) recipients and matched controls. Models were adjusted for age, sex, municipality, and all listed comorbidities. Error bars indicate 95% confidence intervals.
Across liver disease aetiologies, CAD risk was elevated in all subgroups except those with acute liver failure and MASLD (Table 2), where the analysis had low precision due to a low sample size. The highest rates of CAD were observed in LT recipients with rare liver diseases (aHR=4.82; 95%CI,1.90–12.0) and ALD (aHR=2.73; 95%CI=1.80–4.10).
3.4Sensitivity analyses3.4.1Myocardial infarctionWhen examined separately, the rate of myocardial infarction was higher in LT recipients compared with controls (Fig. 5).
Cumulative incidence of myocardial infarction in liver transplant recipients. Cumulative incidence curves of myocardial infarction (MI) in liver transplant (LT) recipients versus matched general population controls (No LT), accounting for the competing risk of non-MI death. Follow-up time is shown in years.
Among patients without a prior history of CAD at the time of LT, the incidence rate of CAD remained higher in LT recipients (13.7 per 1,000 patient-years) than in controls (6.4 per 1,000 patient-years). Over a median follow-up of 9.9 years, new-onset CAD was observed more frequently in LT recipients (n = 2,793) than in controls (n = 25,427). The cumulative incidence of the composite CAD outcome was significantly higher in LT recipients (11%) compared with controls (8%), with an aHR of 2.07 (95%CI=1.80–2.38).
Although none of the patients had a history of CAD at baseline, the cumulative incidence of each component of the composite CAD outcome was significantly higher among LT recipients compared to the control group (Supplementary Figures 1–4).
4DiscussionIn this large register-based cohort study of all LT recipients in Sweden between 1987 and 2020, we found that LT recipients had a twofold higher rate of CAD events compared with matched controls from the general population. The rate of incident CAD was higher across most liver diseases leading to transplant.
Baseline CAD risk factors were potentially stronger predictors of new CAD in the matched controls compared to the LT recipients. Chronic kidney disease and previous CAD were the only cardiovascular risk factors that reached statistical significance in the time-to-event analysis of CAD among LT recipients. Although not statistically significant, most other risk factors showed point estimates in the expected direction. In contrast, cardiovascular risk factors were more commonly statistically significant predictors of incident CAD in the matched controls. This was further supported by the higher incidence of CAD among controls with cardiovascular risk factors, compared to LT recipients with the same risk factors. Together, this could suggest that traditional cardiovascular risk factors may not be as important in predicting CAD risk in LT recipients as in the general population, although the lack of a statistically significant effect may be due to a limited sample size. This finding should be confirmed or refuted in other studies, preferably with more detailed data on CAD risk factors. A potential alternative explanation may be that generally sicker patients with more advanced disease are represented in the control population than in the LT population. For instance, controls may have more severe type 2 diabetes than LT recipients. Given no data on disease severity, we cannot further examine this, but it should be a topic for future studies with more granular data.
In general, our findings should be interpreted with caution, particularly in relation to the differential predictive value of traditional cardiovascular risk factors. The lack of matching for these factors between liver transplant recipients and controls may have limited direct comparability and could have contributed to an underestimation of their prognostic relevance within the transplant cohort. This methodological constraint, inherent to the use of registry data, introduces potential residual confounding that future studies should address through more granular matching strategies or advanced statistical approaches to more fully account for baseline cardiovascular risk profiles.
Beyond traditional cardiovascular risk factors, the cardiovascular effects of immunosuppressive therapy represent a crucial consideration in LT recipients. Calcineurin inhibitors (e.g., tacrolimus, cyclosporine), corticosteroids, and mTOR inhibitors (e.g., sirolimus, everolimus) are known to induce or exacerbate dyslipidemia, arterial hypertension, de novo post-transplant diabetes mellitus, and endothelial dysfunction—each contributing to a proatherogenic and prothrombotic milieu [9–12]. These therapy-related metabolic and vascular alterations may modify the presentation and progression of cardiovascular disease after transplantation, potentially attenuating or amplifying the predictive value of classical cardiovascular risk factors [13]. Understanding the interplay between immunosuppressive exposure and established cardiovascular risks warrants dedicated longitudinal research with more granular data on drug regimens, dosing, and cumulative burden. The absence of detailed immunosuppressive therapy information in our dataset limits deeper exploration of these complex mechanisms.
Previously published studies have usually followed smaller cohorts of LT recipients or for shorter periods of time in relation to incident CAD. To the best of our knowledge, no study has reported on CAD rates in LT recipients compared to a background population, which is important to put risk estimates in context and less sensitive to bias, given that CAD events are very likely to be accurately captured and not missed by healthcare.
During a one-month follow-up in 2118 LT recipients, previous CAD was not associated with new myocardial infarction in a study by Moon et al. [14]. A total of 2.8% experienced myocardial infarction during follow-up, and pre-transplant computed tomographic coronary angiography with signs of CAD, increasing calcium score, hypertension, and diabetes were all significant predictors of new incident myocardial infarction. Our 90-day incidence of CAD was lower at 1.6%, which may reflect that cardiovascular risk factors were more common in the US population than in Sweden. In a single-centre study, Weick et al. followed 803 LT recipients for an average of five years and found that 5.4% developed incident CAD [15]. In multivariable analysis, only increasing age and diabetes were significantly associated with CAD risk. In contrast, the 5-year cumulative incidence of CAD in our multicentre cohort was somewhat higher at 6.9%. However, diabetes was not associated with higher CAD rates among LT recipients in our study, which may be due to the lack of granular data, even though the baseline prevalence of diabetes was similar in both our study and that of Weick et al.
4.1ImplicationsOur study indirectly suggests that screening for CAD can be intensified both before and after liver transplantation for several reasons. First, risk stratification for chronic CAD and the type of screening test are based on symptoms and cardiovascular risk factors. Here, CAD rates were significantly higher in LT recipients compared with matched controls when both groups were free of cardiovascular risk factors at baseline. Therefore, many patients may not fit the standard risk stratification for CAD before liver transplantation, even though leading cardiology guidelines include solid organ transplantation in their assessment [16]. Thus, both LT recipients with and without cardiovascular risk factors can be at moderate to high risk of future CAD event at the time of liver transplantation.
Second, if LT candidates have three or more cardiovascular risk factors, non-invasive CAD testing is indicated according to guidelines [2]. Further studies are needed to examine if CAD risk prediction can be refined using more novel invasive tests such as advanced coronary angiography with intravascular ultrasound or optical coherence tomography, which are routinely used in heart transplant recipients to diagnose cardiac allograft vasculopathy and can differentiate between subtypes of CAD both before and after liver transplantation [17–18].
Third, rejection in both heart and kidney transplant recipients has been associated with incident CAD [19–20]. In contrast, the impact of liver graft rejection on cardiovascular disease is unclear [21]. Therefore, research into advanced testing for CAD after liver transplantation and its association with acute and chronic liver rejection would provide essential knowledge in the field of solid organ transplantation. We could not reliably examine graft rejection due to uncertainty in ICD-coding for this condition.
4.2Strengths and limitationsStrengths include the use of a national cohort comprising all LT recipients in Sweden, which minimizes selection bias; structured follow-up of up to 33 years with minimal loss to follow-up; and the use of Swedish administrative data for CAD diagnoses, which have a positive predictive value of over 95% [22]. Our risk estimates may therefore be more precise and externally valid to countries similar to Sweden, than estimates from studies without such data. Another strength is the use of a matched cohort design, where each LT recipient was matched to up to 10 control individuals. This approach enabled robust estimation of the relative risk of incident CAD, with a background population serving as an appropriate comparator group. This is important for communicating post-transplant risks to patients who are candidates for LT, and for clinicians to guide follow-up decisions.
A key limitation of our study is the absence of matching on major cardiovascular risk factors (e.g., diabetes, hypertension, and hyperlipidemia) between liver transplant recipients and controls. This may have introduced residual confounding, potentially affecting the estimated associations between these risk factors and cardiovascular events in the transplant cohort and reducing the comparability of their predictive value across the two populations.
Other limitations include that LT recipients may be better diagnosed and tested than the matched controls. For example, primary care does not report to the Swedish Patient Register. Matched controls from the general population with e.g. hypertension, who are only treated by primary care, may not be captured in our data set. The true prevalence of the underlying cardiovascular risk factors is therefore likely to be underestimated, with a risk of differential misclassification bias. Another weakness was that we did not have detailed data on laboratory tests, electrocardiograms, and echocardiograms, or non-invasive and invasive tests for CAD.
5ConclusionsCAD rates were twofold higher in LT recipients compared with matched controls from the general population. In LT recipients, previous CAD and chronic kidney disease predicted CAD events after liver transplantation.
Declaration of generative AI in scientific writingDuring the preparation of this work the author(s) used Copilot and ChatGPT to enhance the grammar and clarity of the English language. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication
Author contributionsConception of the study design: HH, CL; Data acquisition: HH; Statistical analysis: LW; Interpretation of data: CL, LW, MM, AW, YS, HH; Drafting and revision of the work: CL, LW, MM, AW, YS, HH; Final approval of the submitted version: CL, LW, MM, AW, YS, HH.
HH’s institutions have received research funding from Astra Zeneca, EchoSens, Gilead, Intercept, MSD, Novo Nordisk, Takeda and Pfizer. He has served as consultant, speaker or on advisory boards for Astra Zeneca, Boehringer Ingelheim, Bristol Myers-Squibb, GSK, Echosens, Ipsen, MSD and Novo Nordisk and has been part of hepatic events adjudication committees for Arrowhead, Boehringer Ingelheim, KOWA and GW Pharma.
AW was supported by Mag-tarmfonden, the Bengt Ihre foundation (SLS-999079), Professor Nanna Svartz foundation (2022–00448), the Stockholm County Council (FoUI-985859), Gastroenterologisk forskningsfond (SLS-999141), and Ruth and Richard Julin foundation (2025–00282).
HH was supported by grants from The Swedish Research Council, The Swedish Cancer Society, Region Stockholm (CIMED and clinical researcher award), The Swedish Heart and Lung Foundation and others.
