Aim: The aim of this study is to evaluate the functional outcomes and complications after non-fusion knee arthrodesis with a modular segmental intramedullary implant used for infected total knee arthroplasty revisions. Methods: A retrospective review of the patients who had been surgically treated with a modular intramedullary arthrodesis implant for recurrent infection after revision TKA between January 2016 and February 2020 were included. The indications for arthrodesis were failed infected TKA with massive bone loss, deficient extensor mechanism and poor soft tissue coverage that precluded joint reconstruction with revision TKA implants. Clinical outcomes were assesed with visual analogue scale for pain (pVAS), Oxford knee score (OKS) and 12-item short form survey (SF-12). Full-length radiographs were used to verify limb length discrepancies (LLD). Results: Fourteen patients (4 male and 10 female) patients with a mean age of 69.3 (range, 59 to 81) years at time of surgery were available for final follow-up at a mean of 28.8 months (range, 24–35 months). All clinical outcome scores improved at the final follow-up (pVAS, 8.5 to 2.6, p =.01; OKS, 12.6 to 33.8, p =.02; SF-12 physical, 22.9 to 32.1, p =.01 and SF-12 mental, 27.7 to 40.2, p =.01). The mean LLD was 1.0 cm (range, + 15 – 2.3 cm). Re-infection was detected in three patients (21.4%). Two patients were managed with suppressive antibiotic treatment and a third patient required repeat 2-stage revision procedure. In one patient, a periprosthetic femur fracture was observed and treated with plate osteosynthesis. Conclusion: Uncontrolled infection after total knee arthroplasty can be effectively treated with arthrodesis using a modular intramedullary nail and satisfactory functional results can be obtained. Level of evidence: Level 4, Retrospective cohort study.