A population-based study of reoperations after back surgery.

STUDY DESIGN

Longitudinal follow-up study of back surgery reoperations using an administrative database.

OBJECTIVES

To identify population-based rates and factors that determine the need for reoperation after back surgery.

SUMMARY OF BACKGROUND DATA

Reoperation after lumbar surgery has poorer results than the initial surgery, yet the population-based incidence and determinants of reoperation are not known. Reported rates of reoperation are derived from retrospective case series and range from 4% to 15%. There are conflicting data on the rate of reoperation after different types of initial surgery.

METHODS

All patients who had back surgery in the Province of Ontario (population 10,000,000) between April 1990 and March 1991 were identified using hospital discharge abstracts and an ICD-9 code algorithm. Patients who had undergone prior surgery were excluded. Patients were observed from the index operation to subsequent readmission and reoperation with a maximal time to follow-up examination of 4 years. Basic demographic information and information regarding diagnoses, surgery performed, complications, comorbid factors, reoperation diagnosis, and surgery type were obtained. Patients were divided into surgical treatment groups, and their subsequent reoperations were identified. Multivariate analysis using proportional hazards modeling was conducted.

RESULTS

The index surgery group consisted of 4,722 patients, of whom 449 (9.5%) underwent reoperations in the follow-up period. Complications from surgery were significantly higher in the fusion and fusion with decompression groups. The reoperation rate was not significantly different among individual surgery groups. Diagnosis, operation performed, complications after the index surgery, comorbid conditions, and sex did not predict the need for spine reoperation. Younger age was predictive of the likelihood of reoperation (P = 0.04)

CONCLUSION

The incidence of reoperation after back surgery is independent of diagnosis and type of surgery performed. Despite different anatomic reasons for surgical intervention, the success of different types of surgery are not influenced by the factors identified in this study. More extensive surgery does not prevent nor predispose a patient to the need for further surgery.