Spinal deformities in Atlantic salmon, Salmo salar L., have been described as a disease of multifactorial origin for which vaccines and time of vaccination have been suggested as risk factors. A vaccine efficacy trial where spinal deformity became evident was continued by the observational study reported here. In the preharvest part of the study 17 months post-sea transfer, there was a prevalence of 11.3% spinal deformity, with deformities present only in one vaccine group indicating a strong vaccine involvement. At slaughter, the prevalence of spinal deformities was 11.7%, and deformed fish had only 62% of normal slaughter weight. Visual analogue scales (VAS) were used for continuous recordings of vaccine-induced abdominal lesions and deformity. A logistic regression model associating presence of spinal deformity with markers of abdominal lesions was developed. The odds ratio for spinal deformity was 5.7 (95% CI: 3.4-9.4) for each unit increase in adhesion score (0-6) and 4.9 (2.9-3.4) for each unit increase in melanin on abdominal organs (0-3). Lesions in the dorsal caudal part of the abdomen gave an odds ratio for spinal deformity of 2.2.