AbstractPurposeWhether the longevity of total knee arthroplasty (TKA) differs between postoperative phenotypes has not been investigated. This study aims to examine which phenotype has a worse long‐term survival rate than the reference phenotype (neutral alignment‐parallel joint line), and whether joint‐line obliquity (JLO) affects the survivorship of TKA.MethodsA total of 945 knees that underwent primary TKAs for primary osteoarthritis from January 2000 to January 2009 were included. These were classified into nine postoperative phenotypes based on the combined assessment of the hip–knee–ankle (HKA) angle and JLO, measured on standing radiographs. The 5‐, 10‐ and 15‐year survival rates were analysed using Kaplan–Meier methods and log‐rank tests. The long‐term survival rates of each phenotype were compared with the reference phenotype.ResultsThere were 55 aseptic mechanical failures within a period of 10.4 ± 5.0 years. The most frequently observed phenotypes were the reference phenotype (n = 527), neutral alignment‐lateral joint‐line inclination (n = 162), varus alignment‐lateral joint‐line inclination (n = 104) and varus alignment‐parallel joint line (n = 101). The overall failure rate for each phenotype was 3.6%, 3.7%, 18.3% and 7.9%, respectively. Only the 10‐ and 15‐year survival rates of the varus alignment‐lateral joint‐line inclination phenotype were significantly different from those of the reference phenotype (97%–93% vs. 90%–69%; p = 0.017, <0.001).ConclusionThe lateral joint‐line inclination phenotype had an inferior long‐term survival rate after varus‐aligned TKA. This suggested that alignment and JLO affected the long‐term survival rate of patients who underwent TKA.Level of EvidenceLevel III, retrospective cohort study.