Dialysis Type May Shape What Heart Scans Predict in Kidney Patients

Patients with end-stage renal disease face a cardiovascular burden so severe that heart disease accounts for more than half of all deaths in dialysis populations. Because kidney transplantation itself carries perioperative cardiovascular risk, and invasive coronary angiography is avoided in this group due to contrast exposure and procedural hazards, non-invasive imaging has become central to pre-transplant risk assessment. A new prospective observational study now suggests that the type of dialysis a patient receives—hemodialysis or peritoneal dialysis—may influence what one of the most advanced cardiac imaging tools can actually predict about their future.

The study, published in EJNMMI Research, enrolled 41 patients with end-stage renal disease who were listed for kidney transplantation and had no known coronary artery disease. Thirty-one were on hemodialysis and ten were on peritoneal dialysis. All participants were asymptomatic at the time of imaging and underwent a multimodal workup: positron emission tomography with computed tomography using the nitrogen-13 ammonia tracer to quantify absolute myocardial blood flow and coronary flow reserve, coronary artery calcium scoring from low-dose CT, and three-dimensional carotid ultrasound to measure carotid plaque burden. The median follow-up period was 4.9 years.

The nitrogen-13 ammonia PET/CT technique allows researchers to measure myocardial blood flow in absolute terms, in milliliters per gram per minute, using dynamic scans fitted to a two-compartment kinetic model. Coronary flow reserve was calculated as the ratio of stress blood flow to resting blood flow after adenosine-induced vasodilation. In parallel, coronary artery calcium was quantified by the Agatston method, and carotid plaque burden was measured semiautomatically in cubic millimeters using dedicated three-dimensional ultrasound software. Two clinical reviewers adjudicated cardiovascular events—cardiac death, non-fatal myocardial infarction, unstable angina requiring hospitalization, coronary revascularization, or heart failure hospitalization—while blinded to the imaging findings.

The central comparison revealed a pattern that was numerically consistent but statistically unconfirmed. Hemodialysis patients exhibited lower stress myocardial blood flow than peritoneal dialysis patients, 2.67 versus 2.97 milliliters per gram per minute, a difference that did not reach significance with a p value of 0.24. Coronary flow reserve showed a similar trend, 2.5 versus 3.08, with a p value of 0.07. Carotid plaque burden and coronary artery calcium were also higher in the hemodialysis group, again without statistical significance. Stress blood flow was impaired, below 2.0 milliliters per gram per minute, in four patients—all of them on hemodialysis. Impaired coronary flow reserve, below 2.0, was found in eleven patients, ten of whom were hemodialysis patients.

Correlations between perfusion and atherosclerosis markers emerged only in the hemodialysis subgroup. There, stress myocardial blood flow correlated inversely with both coronary artery calcium and carotid plaque burden, with correlation coefficients of -0.32 and -0.44 respectively, both statistically significant. Coronary flow reserve showed no such relationship in this group. In the peritoneal dialysis group, no significant correlations were observed at all, and an unexpected positive correlation between coronary flow reserve and calcium score—r = 0.78, p = 0.08—was attributed by the authors to a single outlier in this very small subgroup.

During follow-up, five cardiovascular events occurred: four in the hemodialysis group, representing 13 percent of that subgroup, and one in the peritoneal dialysis group, or 10 percent. Within the hemodialysis cohort, patients with preserved stress blood flow above 2.0 milliliters per gram per minute or preserved coronary flow reserve above 2.0 had longer event-free survival by Kaplan-Meier analysis, with log-rank p values below 0.05 in each case. However, hazard ratios derived from Cox regression using the continuous variables were not statistically significant: 0.57 for stress blood flow and 0.43 for coronary flow reserve. In the peritoneal dialysis group, with only a single event, no prognostic association could be reliably assessed.

Overall survival did not differ between the dialysis modalities. Mean survival time for the entire cohort was 4.5 years, and four patients died during follow-up, two in each group. Neither perfusion parameter predicted death. Formal interaction analyses testing whether the prognostic association of stress blood flow or coronary flow reserve with survival differed by dialysis modality found no significant interaction, with hazard ratios of 0.16 and 0.44 respectively. The authors emphasize that these analyses were severely constrained by the very low number of outcome events.

The study builds on the investigators’ earlier work in the same cohort, which covered a shorter follow-up period without stratification by dialysis type and had shown stronger associations of stress blood flow than coronary flow reserve with coronary calcification and clinical outcomes. In the present analysis, the total duration of dialysis correlated inversely with stress blood flow across the whole cohort, with a correlation coefficient of -0.592, and positively with coronary calcium, but these correlations were confined to the hemodialysis patients. The authors note that elevated resting blood flow, common in chronic kidney disease and previously linked to anemia in this cohort, may complicate the interpretation of coronary flow reserve, since volume overload and chronic activation of the sympathetic nervous and renin-angiotensin-aldosterone systems can raise cardiac workload even at rest.

Why might hemodialysis and peritoneal dialysis produce different imaging signatures? The authors point to distinct pathophysiological mechanisms rather than inherent superiority of one modality. Hemodialysis exposes blood repeatedly to artificial membranes, driving chronic inflammation and oxidative stress that contribute to endothelial dysfunction and atherosclerosis. It also produces intradialytic hemodynamic instability, myocardial stunning, and repetitive ischemia, which may make absolute hyperemic flow particularly informative, whereas peritoneal dialysis offers more stable volume and hemodynamics. Long-term hemodialysis patients additionally accumulate more vascular calcification and arterial stiffness. The hemodialysis group had, on average, a longer median time on dialysis—1.97 versus 0.91 years—though this difference was not statistically significant, and cumulative dialysis time is itself a recognized promoter of atherosclerosis.

The authors are explicit that these findings are hypothesis-generating rather than definitive. The peritoneal dialysis subgroup contained only ten patients and a single event, which precludes meaningful statistical comparison, and the absence of detectable associations there should not be read as evidence of no effect. Residual confounding from nutrition, dialysate glucose, and membrane biocompatibility cannot be excluded, the single-center design limits generalizability, and event adjudication, while prespecified, was not centralized. The cohort itself was highly selected—asymptomatic patients without known coronary artery disease undergoing pre-transplant evaluation—so the results apply only to this specific population. Neither coronary artery calcium nor carotid plaque burden was significantly associated with outcomes, underscoring what the authors see as the distinctive prognostic potential of the quantitative PET parameters. Larger multicenter studies with more cardiovascular events will be needed to determine whether dialysis-specific risk stratification can genuinely improve pre-transplant assessment.

Subject of Research: Prognostic value of nitrogen-13 ammonia PET/CT myocardial perfusion imaging in end-stage renal disease patients on hemodialysis versus peritoneal dialysis

Article Title: Type of dialysis may influence the prognostic value of myocardial perfusion imaging in end stage renal disease: an observational [13N]NH3 PET/CT study

Article References: Wenning, C., Gosmann, T. M., Vrachimis, A., Schäfers, M. A., & Reuter, S. (2026). Type of dialysis may influence the prognostic value of myocardial perfusion imaging in end stage renal disease: an observational [13N]NH3 PET/CT study. EJNMMI Research, 16(1), Article 150. https://doi.org/10.1186/s13550-026-01522-9

Image Credits: AI Generated

DOI: 10.1186/s13550-026-01522-9

Keywords: end-stage renal disease, myocardial perfusion imaging, PET/CT, myocardial blood flow, coronary flow reserve, hemodialysis, peritoneal dialysis, coronary artery calcium, carotid plaque burden, kidney transplantation, cardiovascular risk, observational study