Does a Competing Risk Analysis Show Differences in the Cumulative Incidence of Revision Surgery Between Patients with Oncologic and Non-oncologic Conditions After Distal Femur Replacement?

BACKGROUND

Distal femur replacement is frequently used for limb salvage after bone tumor resections. It is also used in patients with severe bone loss because of traumatic conditions or revision TKA. Some studies on distal femur replacement reported on revision-free survival without distinguishing between patients with oncologic diagnoses and those without, although these patients might be incomparable because of their differences in important patient- and disease-specific characteristics. This may lead to an inaccurate and undifferentiated interpretation of the results of survival analyses.

QUESTIONS/PURPOSES

(1) What is the overall cumulative incidence of revision surgery after cemented and cementless distal femoral replacement, as determined with a competing risk analysis? (2) Does the cumulative incidence of revision surgery change over time? (3) Are there differences in the cumulative incidence of revision surgery between patients with oncologic conditions and those without who are treated with cemented or cementless distal femoral replacement?

METHODS

A total of 403 patients were possible candidates for distal femoral replacement. Of these, 56 patients elected to undergo different procedures, 83 were excluded because an expendable growing prosthesis was implanted, and 28 were lost to follow-up. Therefore, 229 patients who underwent distal femoral replacement for oncologic or non-oncologic reasons between 1983 and 2016 were retrospectively included in this study. The type of fixation method (cemented or cementless) was obtained from the patients' medical records, operation reports, and radiographic analyses from plain radiographs. All radiographs were standardized and obtained at standard time intervals in our institution. No algorithm regarding the fixation approach was followed. According to our data, patients receiving cementless fixation were younger and therefore likely to be more active than those receiving cemented fixation. The median follow-up duration of the overall cohort was 85 months (range 0.1-391 months). Patients who died or had revision surgery before the 2-year minimum follow-up interval were adequately considered using competing risk calculation. The reasons for revision surgery were classified using the classification system proposed by the International Society for Limb Salvage. A competing risk analysis was performed to estimate the cumulative incidence function of revision, accounting for death as a competing event. To evaluate the influence of potential prognostic factors, including diagnosis (oncologic versus non-oncologic), fixation (cemented versus cementless), year of distal femoral replacement, age, and sex on the occurrence of revision surgery, univariate and multivariable Fine and Gray models were applied.

RESULTS

The competing risks analysis revealed cumulative incidences of revision surgery for any cause (Types 1 to 5) of 26% (95% CI, 20.3%-31.9%) at 12 months, 37.9% (95% CI, 31.3%-44.4%) at 24 months, 52.6% (95% CI, 45.1%-59.5%) at 5 years, and 58.2% (95% CI, 50.1%-65.4%) at 10 years for all patients. Rotating hinge-type prostheses showed a lower cumulative incidence of revision surgery (41.6%; 95% CI, 31.8%-51%) than fixed-hinge prostheses did (64%; 95% CI, 50.5%-74.5% ) at 5 years (Gray's test: p = 0.01). According to the multivariate Fine and Gray model, the year of surgery did not have any effect on the risk of revision surgery (1994 to 2003: hazard ratio 0.70; 95% CI, 0.46-1.07); 2004 to 2016: HR 0.83; 95% CI, 0.52-1.34; p = 0.26). The multivariate analysis, adjusted for disease, sex, age, cementation, and year of surgery, revealed a difference in the risk of revision surgery between patients with oncologic disease and those with non-oncologic disease (HR 0.44 for oncologic versus non-oncologic; 95% CI, 0.22-0.87; p = 0.02) and a reduction in the risk of overall revision with cemented fixation in patients with oncologic disease (HR 0.53; 95% CI, 0.29-0.98; p = 0.03).

CONCLUSION

This study indicates that even with newer implants, there was a high incidence of revision surgery after distal femoral replacement. According to our analysis, patients with oncologic diagnoses have a lower likelihood of revision when the stem is cemented whereas the type of fixation did not impact patients with non-oncologic diagnoses. Because of differences in patient demographics (age, etiology of disease, and use of chemotherapy) and outcomes of fixation, oncologic and non-oncologic patients should be analyzed separately in survival studies about distal femoral replacement.

LEVEL OF EVIDENCE

Level III, therapeutic study.