Acquired hemolytic diseases are characterized by premature red blood cell destruction and can, regardless of etiology, lead to transient or permanent elevated bilirubin levels, which may promote the formation of pigment gallstones 1-3. Previous studies have demonstrated an increased risk of gallstone disease in congenital hemolytic disorders, which are often characterized by chronic hemolysis 3-7. By contrast, the effect of the often transient hemolysis in acquired hemolytic diseases remains unknown 8. Observations in patients undergoing heart–lung-assisted cardiac surgery suggest that even this transient hemolysis may predispose to pigment gallstones 3, 9. This raises the possibility that increased gallstone formation is not restricted to congenital hemolytic disorders but may also occur in acquired hemolysis. To address this question, we conducted a nationwide, population-based cohort study using more than four decades of data from the Danish health registries to estimate the risk of clinically significant gallstone disease in patients with acquired hemolysis. The Danish healthcare system is tax-funded and provides universal free of charge access, capturing primary diagnosis, reflection the condition prompting admission, and relevant secondary diagnoses from all hospital contacts 10. Hematological diagnostics as well as acute surgical treatments are exclusively performed in public hospitals, while planned surgeries may occur in both public and private hospitals 10. We included incident patients with autoimmune hemolytic anemia (AIHA), cold agglutinin disease (CAD), paroxysmal nocturnal hemoglobinuria (PNH), drug-induced hemolytic anemia, and the miscellaneous group of other acquired hemolytic anemias (e.g., March hemoglobinuria) on their first day of diagnosis between January 1980 and December 2021 11, 12. At inclusion, each patient was age- and sex-matched with up to 25 general population comparators free of hemolytic diseases. Follow-up continued until the first occurrence of a gallstone event (primary outcome), death, permanent emigration, or end of observation time (31 December 2021). Patients and comparators with a history of gallstones were excluded. Comorbidities were identified from the hospital diagnoses and prescription data (Table S1). The primary outcome was the first of either a discharge diagnosis of gallstone disease or gallstone-related surgical procedure (Table S2). Analysis was performed overall and by subgroup comparing patients and comparators. We estimated cumulative incidences of gallstone events until 20 years after diagnosis, with death and emigration as competing events. Cumulative incidences and absolute risk differences were tabulated at 5 and 10 years. We estimated Cox proportional cause-specific hazard ratios (csHR), adjusting for age, sex, diabetes, dyslipidemia, and obesity. Results were reported with 95% confidence intervals (CI). We performed multiple sensitivity analyses including alternative outcome definitions, sex-specific analyses, subtypes of gallstone, and the gallstone risk associated with ITP 13. Further details on data sources, patients and comparators, analyses, and literature review are provided in the Supporting Information: Methods. We identified 6147 patients with acquired hemolytic diseases and 143 265 matched comparators. Patients were slightly older at inclusion (median 67.6 vs. 66.9 years, p 0.05). Patients contributed 44 554 person-years (PY; median 4.2), and comparators contributed 1 652 986 PY (median 9.1). During follow-up, 370 gallstone events occurred in patients and 5348 in comparators. Median age at gallstone event was lower amongst patients (67.9 years, IQR 50.9–78.9) than comparators (74.9 years, IQR 62.3–82.2). The adjusted csHR for gallstone events amongst patients versus comparators was 2.9 (95% CI 2.6–3.2), Table 1 and Figure 1. Generally hemolysis showed a stronger or comparable association with gallstone than demographic and metabolic covariates in both univariate and adjusted analyses (Table S3). The exceptions were CAD, where diabetes had the strongest univariate association, and drug-induced hemolysis, where obesity remained the dominant predictor in both models. Gallstone events occurred earlier in patients than in comparators, and for the majority of the hemolytic diseases the excess risk accumulated within the first 5 years following diagnosis, Figure 2, Table S4. At 5 years, the cumulative incidence of gallstone events was 3.5% (95% CI 3.1–4.0) amongst patients and 1.5% (95% CI 1.4–1.5) amongst comparators. At 10 years, the corresponding estimates were 5.3% (95% CI 4.7–5.9) in patients and 2.9% (95% CI 2.5–2.7) in comparators. Disease-specific results are described below, and in Table 1, Figures 1 and 2, and Tables S4–S9. The AIHA cohort included 2989 patients with 182 gallstone events and 69 305 comparators with 2530 events. Median time from inclusion to gallstone event was shorter in patients (5.6 years, 95% CI 4.7–6.5) than in comparators (9.0 years, 95% CI 8.7–9.3), Table 1 and Table S5. The adjusted csHR was 2.8 (95% CI 2.4–3.3), Figure 1. The cumulative incidence curve indicated a high incidence during the first few months after diagnosis, Figure 2. There were 159 patients with CAD and 8 gallstone events, and 3406 comparators with 83 events (Table 1 and Table S6). Despite the adjusted csHR of 2.5 (95% CI 1.2–5.1), CAD was the only group where the patients were older than comparators at the time of gallstone event. There was a slow but continuous increase in cumulative risk: Figure 2. The PNH cohort included 129 patients with 12 gallstone events, and 3127 comparators with 118 events (Table 1 and Table S7). The adjusted csHR was 3.4 (95% CI 1.9–6.2) and the risk difference continued to increase even beyond 10 years after diagnosis, Figure 2. We identified 372 patients with drug-induced hemolytic anemia, who contributed 16 gallstone events. Their 8640 comparators contributed 390 events (Table 1 and Table S8). The adjusted csHR was 1.9 (95% CI 1.1–3.1). The difference in cumulative incidence was most prominent shortly after diagnosis and then decreased to near-equal risk after 10 years, Figure 2. For sensitivity we included 8599 patients with ITP and 206 139 comparators, having 375 and 6269 gallstones, respectively (Table 1 and Supporting Information: Results). The adjusted csHR was 1.7 (95% CI 1.5–1.9). Unlike the hemolytic diseases, ITP showed no steep early rise in cumulative incidence (Figure S1). The sensitivity analyses were consistent with the main results across all outcome definitions and inclusion criteria. For further descriptions of the results, including extensive results for other acquired hemolytic disorder, ITP, and sensitivity analyses see Supporting Information: Results and Tables S4–S9. The magnitude of the risk of gallstone events was comparable to that observed in congenital hemolytic disorders, where gallstone formation and complications are documented as summarized in Table S10 4, 5. The age at first gallstone event was lower in all patient groups, except CAD, and the excess risk accumulated most rapidly within the first 5 years after diagnosis, particularly in AIHA, PNH, and drug-induced hemolysis, consistent with increased risk before hemolysis is effectively controlled. In PNH, the risk difference continued to widen beyond 10 years, likely reflecting persistent hemolysis despite treatment 14. In contrast, the difference in cumulative incidence between patients with drug-induced hemolysis and comparators narrowed after 5 years, likely reflecting drug cessation and resolution of hemolysis. CAD had the highest cumulative incidence during the first years, a pattern likely explained by the increase in baseline risk of gallstones with advancing age 1, 5, 15. Adjustment for sex, age, and metabolic comorbidities increased the csHRs across subgroups. In both univariate and adjusted models, hemolysis generally showed a stronger effect than the other covariates. Collectively, these findings make metabolic or demographic confounding unlikely and instead support hemolysis as the primary driver of the increased gallstone risk. Our results indicate that even brief hemolytic episodes (e.g., warm AIHA or drug-induced hemolysis) and compensated chronic hemolysis (e.g., PNH and CAD) are sufficient to increase gallstone risk. This aligns with prior observations of gallstone formation in mechanically induced hemolysis, for example after artificial heart valve replacement 3, 9, 16. The increased gallstone risk could reflect underlying inflammation or immunosuppressive treatment, and we therefore assessed gallstone risk in patients with ITP 17. The csHR in ITP closely mirrored those reported by Park et al. in rheumatoid arthritis, and the cumulative incidences were higher in rheumatoid arthritis than in ITP 18. The lower estimates in ITP compared with AIHA and PNH support that hemolysis is the primary driver of the increased gallstone risk. Key strengths of this study include its nationwide design, minimal loss of follow-up, and the validated algorithms to identify patients. The universal healthcare coverage in Denmark minimizes selection bias, and the sensitivity analyses support that the increased risk is not attributable to coding practices, outcome definitions, or calendar period. Key limitations include the absence of clinical detail, such as gallstone type, biochemical measures, and treatment of hemolysis. Further considerations are described in the Supporting Information: Discussion. While asymptomatic gallstones are common and of limited clinical relevance, symptomatic disease with pain or infections requiring medical or surgical intervention represents a meaningful burden for patients 5, 19. Patients with acquired hemolytic diseases have a substantially increased risk of symptomatic gallstone events. This risk emerges early after diagnosis and persists over time, particularly in conditions marked by ongoing or recurrent hemolysis. Our findings highlight the systemic consequences of acquired hemolysis and underscore gallstone disease as an under-recognized complication in these disorders. Whether early and sustained suppression of hemolysis can mitigate gallstone risk remains uncertain and warrants further investigation. D.L.H. conceptualized the study. A.B.N. and D.L.H. planned analysis. D.L.H. performed analysis and wrote first draft. All authors participated in subsequent improvements. This work was supported by the Novo Nordisk Foundation, Denmark, grant number NNF22OC0078228. The project was registered by the Southern Danish Region record no. 17/10885. According to Danish law, register-based research without patient contact does not require ethical approval. The authors declare no conflicts of interest. According to Danish law the authors cannot grant direct access to the data from the national health registers. Researchers may apply for use of research data to Statistics Denmark. Data analysis files (“Do-files”) is available upon request. Data S1: Supplementary text description of methods and results, and supplementary elaborated points concerning strengths and limitations. Figure S1: Depicting cumulative incidence of gallstone events amongst patients with immune thrombocytopenia. Tables S1: and S2. Diagnosis, surgery, and drug codes used to identify comorbidity and outcome. Tables S3–S9: Results from sensitivity analyses. Table S10: Comparing risk estimates for gallstone disease amongst patients with hemolysis from the literature. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Naamansen et al. (2026) studied this question.