Relationship Between ABO/Rh Blood Groups and Knee Osteoarthritis: A Retrospective Cohort Study

Background/Objectives: The association between ABO/Rh blood groups and knee osteoarthritis (OA) remains controversial, with inconsistent findings reported across different populations. This study aimed to evaluate the distribution of ABO and Rh(D) blood groups in patients undergoing primary total knee arthroplasty (TKA) for primary knee OA and to compare these distributions with a regional external reference population. Methods: This retrospective, single-center, observational study reviewed hospital records of patients who underwent primary TKA between January 2011 and October 2024. After applying predefined exclusion criteria (different ethnic background, age < 50 years, secondary knee OA, and missing blood group data), 4969 patients with primary knee OA were included. ABO/Rh(D) data were obtained from the institutional electronic hospital information system and transfusion/laboratory records. The external reference population consisted of a previously published dataset of regional blood donors (10,867 unique donors). Observed blood group frequencies in the study cohort were compared with expected frequencies derived from the reference distribution using chi-square goodness-of-fit tests for (1) 8-category ABO/Rh(D) distribution, (2) ABO-only distribution, and (3) Rh(D)-only distribution. Results: Among 4969 patients, 4096 (82.4%) were female and 873 (17.6%) were male. Mean age was 66.8 ± 7.0 years (range, 50-94) in females and 68.8 ± 7.3 years (range, 50-88) in males. The most frequent blood groups were A Rh (+) (39.3%), O Rh (+) (30.0%), and B Rh (+) (14.5%). The sex-specific ABO/Rh distribution did not differ significantly (p = 0.052). Compared with the regional reference distribution, the overall 8-category ABO/Rh(D) distribution showed a borderline difference (χ² (7) = 14.04, p = 0.050; Cramér's V = 0.020). However, neither the ABO-only distribution (χ² (3) = 5.26, p = 0.153; Cramér's V = 0.019) nor the Rh(D)-only distribution (χ² (1) = 0.11, p = 0.737; Cramér's V = 0.005) differed significantly from the regional reference. The observed deviations were numerically small and not suggestive of a clinically meaningful difference. Conclusions: In this large single-center cohort of patients undergoing primary TKA for primary knee OA, the ABO and Rh(D) blood group distributions were largely comparable to those of the regional population. Although the overall 8-category ABO/Rh(D) comparison showed a borderline difference, separate ABO-only and Rh(D)-only analyses were not significant. These findings do not support a strong association between blood group status and surgically treated primary knee OA in this population.