The clinical starting point for the study is well established. Moderate-to-severe chronic kidney disease-associated pruritus, abbreviated CKD-aP, affects up to 40 per cent of people with kidney failure receiving haemodialysis. Beyond the obvious suffering it causes, the condition has been linked to increased mortality, higher rates of depression and sleep disorders, and markedly diminished health-related quality of life. Crucially for the new analysis, patients with CKD-aP also experience higher rates of infection and infection-related hospitalisation, with risks rising as itch severity worsens. Because hospitalisations for sepsis, catheter infections, endocarditis and skin infections are among the most expensive events in the dialysis population, the researchers reasoned that any therapy capable of resolving pruritus might pay back part of its cost through avoided admissions.

Difelikefalin, a peripherally restricted kappa-opioid receptor agonist administered intravenously at the end of each dialysis session, is currently the only therapy approved in both the United States and Europe specifically for moderate-to-severe CKD-aP in patients on haemodialysis. Its approval rested on the double-blind, randomised, placebo-controlled phase 3 trials KALM-1 and KALM-2, which demonstrated significant reductions in pruritus severity as measured by the Worst Itch Intensity Numerical Rating Scale, with clinically meaningful improvement defined as a score of 4 or higher at baseline. While alternatives such as gabapentin, pregabalin and antihistamines are sometimes used off-label, they are not specifically indicated for CKD-aP and carry variable effectiveness and tolerability in this population, including concerns about adverse outcomes with gabapentinoids in dialysis patients.

To quantify the infection burden attributable to pruritus, the team turned to the Stockholm Creatinine Measurements project, or SCREAM, a healthcare cohort study encompassing the entire population of the Stockholm region and linked through Sweden’s unique personal identity numbers to the Swedish Renal Registry, the National Patient Registry, the Prescribed Drug Registry and the Cause of Death Registry. From 2329 patients on maintenance haemodialysis who were free from CKD-aP at baseline, the researchers identified 411 patients, or 17.6 per cent, who developed clinically recognised pruritus during follow-up between 2006 and 2021. Because Sweden lacks a dedicated itch registry, the team defined clinically recognised pruritus using a combination of surrogate markers: ICD-10 diagnostic codes for pruritus, at least two consecutive dispensations of antipruritic medications such as hydroxyzine, clemastine, gabapentin or pregabalin in the absence of diabetes, peripheral neuropathy or epilepsy, use of an extemporaneously compounded topical itch ointment known in Swedish as klådsalva, or ultraviolet therapy without a psoriasis diagnosis. These criteria were chosen because they align with the moderate-to-severe threshold used in the KALM trials and are likely to capture patients experiencing clinically meaningful itch.

The epidemiological finding at the heart of the study is stark. Infection-related hospitalisation rates were 1.5-fold higher among patients who developed incident CKD-aP than among those who remained itch-free, with an unadjusted incidence rate ratio of 1.48 and a 95 per cent confidence interval of 1.19 to 1.83, a difference that was highly statistically significant. Infection-related hospitalisations occurred at a rate of 16.98 per 100 person-years among patients with incident CKD-aP, compared with 11.60 per 100 person-years in those without. The burden was driven chiefly by sepsis due to Staphylococcus species, which showed a significantly elevated rate in the pruritus group with an incidence rate ratio of 1.55, followed by skin infections including erysipelas, cellulitis, abscess and impetigo, which were also more common in the itch group although the difference did not reach statistical significance. The biological plausibility of these associations rests on immune dysregulation, skin breakdown from chronic scratching and excoriation, and the frequent presence of central venous catheters in this population.

The economic modelling then married these real-world event rates to the treatment effect observed in the clinical trials. Using pooled data from KALM-1 and KALM-2, supplemented by longer-term extrapolation from the SHAREHD stepped-wedge cluster randomised trial, the model projected that over a two-year horizon patients treated with difelikefalin would spend 61.7 weeks free of moderate-to-severe CKD-aP, compared with just 38.1 weeks for those receiving standard care, a difference of roughly 23.6 weeks. The model, adapted from a previously published cost-effectiveness framework, used four health states defined by itch severity, with moderate and severe states combined to represent the presence of CKD-aP and none and mild combined to represent its resolution. It operated on 28-day cycles over two years, applied a 3.5 per cent discount rate in the second year, and adopted the perspective of each country’s healthcare payer. Importantly, drug acquisition costs were deliberately excluded because the study aimed to isolate the infection-related contribution rather than perform a full economic evaluation, meaning the figures represent cost offsets rather than net savings.

Scaled to national dialysis populations using prevalence data from the European Renal Association Registry and country-specific unit costs, the projected two-year infection-related cost offsets were substantial everywhere the model was applied. In the base case using SCREAM-derived assumptions, offsets ranged from approximately €231,000 in Sweden to just over €2 million in France. Across all scenario analyses, the estimated national two-year offsets spanned €2 million to €6.4 million in France, €1.2 million to €3.5 million in Spain, €1.1 million to €2.6 million in Italy, €1 million to €3 million in the United Kingdom, €500,000 to €1.4 million in Portugal, €250,000 to €730,000 in Switzerland, and €230,000 to €660,000 in Sweden. Probabilistic sensitivity analyses, running the model 1000 times with parameters sampled from their confidence intervals and Dirichlet distributions, produced narrow confidence bands, indicating that the conclusions were stable despite uncertainty in individual inputs.

The scenario analyses are particularly revealing about how conservative the base case may be. Catheter-associated infection rates in the SCREAM data were only 1.89 per 100 person-years, a figure the authors attribute to coding practices in hospital records, where catheter infections may be displaced by codes reflecting kidney failure or secondary infections, and to the requirement that events result in hospital admission to be captured. Alternative sources paint a different picture: a systematic review and meta-analysis by van Meurs and colleagues reported pooled catheter infection rates of 20.3 per 100 person-years, while analyses from the Dialysis Outcomes and Practice Patterns Study, or DOPPS, found rates between 3.5 and 10.2 per 100 person-years depending on country. Similarly, while SCREAM suggested a CKD-aP prevalence of 17.6 per cent, the European CENSUS survey and DOPPS reported figures between 26.8 and 40.1 per cent. Substituting these higher prevalence and infection rates increased the estimated cost offsets by 48 per cent with higher prevalence and by 103 to 179 per cent with higher catheter infection rates, and a scenario adding pruritus-associated excess mortality slightly reduced offsets because more surviving patients continue to incur costs.

The authors are careful to acknowledge the limitations inherent in real-world evidence of this kind. Misclassification of pruritus is possible where cases go unrecognised by clinicians or unreported by patients, and gabapentin and pregabalin dispensations may partly reflect treatment of restless legs syndrome rather than itch, although use of these drugs was low overall and similar between groups. Patients who developed incident CKD-aP also differed systematically at baseline, with more heart failure, prior skin infection, sleep and mood disorders, and longer time on dialysis, factors that could confound the association with infection; however, a previously published adjusted analysis showed a similar elevation in infection-related hospitalisation risk, with an adjusted hazard ratio of 1.36, suggesting the association is robust. The model also assumes that the relationship between incident pruritus and outcomes is exchangeable with symptom resolution in a prevalent population, whereas in clinical practice most difelikefalin is prescribed to patients with existing pruritus, which is why the higher-prevalence scenarios may better reflect reality.

Beyond the headline figures, the study’s significance lies in how it reframes the economics of a symptom that has historically been underdiagnosed and undertreated in dialysis units. Because the analysis counted only infection-related hospitalisations, it almost certainly understates the full economic case: previous work by the same Scandinavian collaboration showed that CKD-aP is associated with increased use of antidepressants, anxiolytics and sleep aids, and the documented improvements in health-related quality of life from trials represent additional value not captured in a cost-offset framework. The authors suggest that extending the model over a longer horizon would likely yield greater offsets, since early reductions in itch severity with difelikefalin would be sustained over time, and that further treatment-related gains through reduced mortality are plausible given the established association between pruritus and death in dialysis populations. For health system decision-makers weighing whether to fund CKD-aP treatment within bundled dialysis payments or as a separate per-patient mechanism, the message of this analysis is that treating the itch may not merely make patients more comfortable; it may also keep them out of hospital, and that difference carries a price tag measured in millions of euros across Europe every two years.