The incidence of cardiac surgery in adults with treated kidney failure in Australia: a retrospective cohort study

Introduction

In Australia, chronic kidney disease is a risk for one in three adults and if untreated may lead to kidney failure (KF).¹ People with KF live with an increased burden of cardiovascular disease.² A high prevalence of traditional (e.g. diabetes and hypertension) and non-traditional factors (e.g. inflammation and oxidative stress) contribute to their excess risk.^3–5 Cardiovascular disease remains their leading cause of mortality.⁶ For coronary artery disease (CAD), coronary artery bypass grafting (CABG) is an important strategy for revascularisation in this cohort.⁷ Mitral annular and aortic valve calcifications are common and progress to stenosis or regurgitation more rapidly for adults with KF than in the general population.⁸ Surgical interventions carry a substantial risk, especially for mitral valve (MV) replacement or repair.⁸

The prevalence of adults with KF treated with either dialysis or kidney transplantation (collectively, kidney replacement therapy [KRT]) in Australia is projected to increase to approximately 37,000 by 2030, most rapidly among people over 60 years of age.⁹ Greater comorbidity is also anticipated in the KRT population, associated with the rise of diabetes and other diseases in the general population.^10,11 However, the frequency and nature of cardiac surgeries for these cohorts and their preoperative risk profiles are not well documented. Postoperative outcomes for adults with KF in Australia suggest they are a high risk for cardiac surgery^12–14 and this is reflected in risk prediction models.^15,16 Estimating the recent burden and risk of cardiac surgery is therefore critical to inform patients and clinicians, and facilitate accurate risk assessment and peri- and postoperative planning.

Limited data are available on cardiac surgery trends for KRT cohorts^17,18 and the specific consideration of kidney transplant recipients is rare. This study aimed to bridge this gap by estimating the incidence of cardiac surgeries among adults receiving KRT in a population-based retrospective cohort based on data linkage between two binational clinical quality registries.

Methods

Results

A total of 114,667 surgeries could be assigned to either of the three study cohorts for the period 2010–2019, with 1243 (1.08%) for adults receiving dialysis and 298 (0.26%) for those with a functioning kidney transplant. Excluded from analyses were surgeries for which the patient was coded in the ANZDATA Registry as in renal recovery (n = 11), lost to follow-up (n = 1) or deceased (n = 3).

Most surgical procedures were in males (n = 74%), with no difference in sex ratio among the three cohorts (Table 1). Median age at the time of surgery varied significantly, with the dialysis and transplant cohorts younger than the non-KRT cohort. Almost all other risk factors varied significantly (P < 0.01) among the sample; risk for the dialysis cohort tended to be highest, followed by transplant and non-KRT cohorts. Isolated CABG procedures exceeded 60% of the procedures for adults on dialysis, but were only 36.2% for transplant recipients, as compared with 51.7% for the general population.

The unstandardised cardiac surgery rate was highest for adults receiving dialysis, 10.1 per 1000 persons per year (95% CI 9.6–10.7) followed by transplant recipients (2.9 [2.6–3.3]) and those not receiving KRT (0.62 [0.61–0.62]). Incidence rate ratios (IRRs), relative to the non-KRT cohort, exceeded one for all combinations of modality, sex and age except for 18–39-year-old transplant recipients (Fig. 1); IRRs tended to be higher for females than for males, for both dialysis and transplant cohorts. For both sexes, IRRs were higher for adults receiving dialysis than for transplant recipients, and for all ages except for those over 70 years (Supplementary material Table S1).

Fig. 1.

Incidence rate ratios for dialysis and transplant cohorts relative to the non-KRT population for 2010–2019, by modality at the time of surgery, sex and age group.

The age–sex-standardised incidence rates for dialysis and transplant cohorts relative to the non-KRT cohort significantly exceeded one for all procedure types except for isolated MV surgeries in transplant recipients (n = 8) (Table 2). The procedure type with the highest relative rate for the dialysis and transplant cohorts was isolated CABG and isolated AVR respectively. For all the cohorts defined by diabetes status, BMI and clinical status, the dialysis rate tended to exceed the respective transplant rate and both were significantly higher (P < 0.05) than the non-KRT rate (Tables 3–5). The highest relative rates for any measured group were for adults with insulin-treated diabetes: for adults receiving dialysis (65.2 per 1000 persons per year [95% CI 56.6–74.7]) and those with a transplant (11.4 [8.3–15.5]). Other elevated relative rates included those for adults receiving dialysis with BMI < 25 kg/m² and undergoing non-elective surgeries.

For isolated CABG, the incidence rate for adults commencing dialysis less than 30 days prior to surgery was higher relative to those 30 days to <1 year receiving dialysis, and relative rates for those receiving dialysis for ≥2 years were lower (Table 6). Adults on home haemodialysis had a lower cardiac surgery incidence than those on facility haemodialysis (Table 7).

Discussion

This is the first population-based Australian study to estimate the incidence of cardiac surgeries for adults receiving KRT relative to the general population. Although people receiving KRT are approximately 0.1% of the population, they comprised over 1.3% of cardiac surgeries in 2010–2019. The relative incidence of each major cardiac surgical procedure type was elevated for the KRT cohorts: at least 9- and 2-fold higher for adults receiving dialysis and transplant recipients, respectively. Given the observed clustering of comorbidities and risk, especially for adults receiving dialysis, that finding was expected. Unmeasured, non-traditional risk factors are also likely to have accelerated arterial and valve disease.⁴ Those undergoing cardiac surgery were predominantly males aged 50–69 years. Decreased relative incidence of surgery for those over 70 years of age receiving dialysis likely reflects a reduction in referral for cardiac work-up pre-kidney transplantation in those older adults.

Relative surgery risk for adults with KF who were comorbid with insulin-treated diabetes was substantial: over 65-fold for those receiving dialysis. Postoperative outcomes are also known to be worse for this cohort.^7,23 The projected increase in the diabetic kidney disease population (even accounting for widespread treatment limiting its progression) raises concern at the consequent increase in surgical burden for the individual and the health system.²⁴ In contrast to the relationship with diabetes, adults with a BMI <25 kg/m² receiving KRT had a higher relative rate than those with a higher BMI. Those findings, while contradictory at face value, may illustrate the competing factors of surgical disease burden and comorbidity versus risk stratification and patient selection in determining who ultimately undergoes a procedure. Notably, high BMI is associated with lower postoperative risk.²⁵

Despite significant elevation over the general population risk, the relative incidence of cardiac surgery was lower for transplant recipients compared with adults receiving dialysis. They had less comorbidity and a lowered risk profile, and incidence rates were not as strongly associated with diabetes status. Those findings were anticipated given the routine cardiovascular evaluation and treatment that occurs prior to kidney transplantation. Adults receiving home haemodialysis had a lower incidence of cardiac surgery than those receiving facility haemodialysis. While peritoneal dialysis and haemodialysis both accelerate vascular disease, findings remain mixed on any differences in their cardiovascular outcomes.²⁶

Over a similar period to this study, the CABG case volume for adults receiving dialysis (for those with stable CAD) decreased in the USA, reflecting a broader downward trend reported from the USA and internationally.^17,27,28 A combination of improved results of non-invasive treatment, a shift in patient selection, and the advent of percutaneous coronary intervention (PCI) have contributed, despite a trend toward an older and more morbid underlying population. Trends owing to advances in CAD revascularisation strategies with PCI, and the use of transcatheter aortic valve replacement for older and/or more comorbid adults are likely to continue to influence cardiac surgery rates in Australia, for the general and KRT populations.

Strengths of this study include the 10-year study period and the generalisability of data extracted from binational clinical quality registries. One limitation is that the population coverage of the ANZSCTS Cardiac Surgery Database changed measurably over the study period owing to an increase in the number of contributing units. To address this, the subset of surgeries from 2015 to 2019 were analysed to elucidate the difference in incidence rates between that period and 2010–2019. No sizeable differences were identified (Table S2). Finally, the actual burden of surgical disease is likely to be greater than we have estimated, owing to selection for surgery and biases in care provider and consumer decision-making in the perioperative period.

Conclusion

In conclusion, data linkage between two binational clinical quality registries allowed an estimation of cardiac surgery incidence rates for adults with treated KF at the time of surgery in 2010–2019 that were significantly higher than in the general population. Adults receiving dialysis, especially those comorbid with diabetes, were at the highest risk. These data will aid in determining how best to prevent the need for surgery and optimise outcomes for the growing population of adults receiving KRT at the time of their surgery.

Supplementary material

Supplementary material is available online.