ObjectiveTo analyse patients with recurrent atlantoaxial dislocation and give a criterion of an ideal patient who can benefit from redo surgery. MethodsThis retrospective study was conducted in a tertiary care centre, which included 20 patients who failed atlantoaxial surgery from January 2013 to June December 2021. They were evaluated using X-ray, CT, and MRI examinations, and their clinical data were accessed from the hospital's medical records department and the picture archiving and communication system. They were given a trial of traction to look for clinical and/or radiological improvement. Those showing clinical and/or radiological improvement underwent redo fixation. The operative steps involved removing joint capsules, denuding articular cartilage and joint preparation followed by reduction of basilar invagination by the combination of spacer and/or bone graft and putting screws in C1/Occiput and C2. A strut graft was placed between Occiput/ C1 and C2. ResultsThe mean change in mJOAS and Nurick grade following the first surgery was statistically significant (1.00 ± 0.73, p-value 0.002 and -0.15 ± 0.27, p-value 0.046, respectively). On similar paths, the mean change in mJOAS and Nurick grade following the second surgery was also statistically significant (4.25 ± 0.32, p-value <0.001 and -1.2 ± 0.11, p-value <0.001, respectively). Improper usage of constructs (31.57%), inadequate/no joint preparation (42.10%/57.90%) and poor choice of graft (100%) were the leading causes of failure of index surgery. ConclusionsThe best candidates who can benefit after re-do surgery are the ones who exhibit either clinical and/or radiological improvement on the trial of traction, as the pathological C1-C2 joints are either not fused or have undergone pseudoarthrosis. Those patients who do not exhibit significant clinical or radiological improvement post-trial of traction should not be offered subsequent surgical intervention.