Related Experiment Video
Updated: May 17, 2025
Assessment of Bone Fracture Healing Using Micro-Computed Tomography
Published on: December 9, 2022
An External Validation of the Pathologic Fracture Mortality Index for Predicting 30-day Postoperative Morbidity Using
Joseph O Werenski1, Marie W Su, Ryan K Krueger
1From the Orthopaedic Oncology Service, Department of Orthopaedic Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, MA (Werenski, Su, Krueger, Groot, Clunk, Sodhi, Patil, Bell, and Lozano-Calderon), and the Division of Oncology, Department of Orthopaedic Surgery, Johns Hopkins Medicine, Baltimore, MD (Levin).
Introduction:
Skeletal metastases increase the risk of pathologic fractures, causing functional impairment and pain. Predicting morbidity in patients undergoing surgical fixation for these fractures is challenging due to the complexity of metastatic disease. The Pathologic Fracture Mortality Index (PFMI) was developed to predict 30-day postoperative morbidity in long bone fractures caused by metastases. External validation is necessary for clinical use. This study aims to evaluate the following: (1) How well does the PFMI predict 30-day medical, surgical, utilization, and all-cause morbidity after pathologic fracture fixation in an external cohort of patients with long bone metastases? (2) How does the performance of the PFMI compare to established predictive indices including the American Society of Anesthesiologists (ASA) classification score, the modified 5-Item Frailty Index (mF-I5), and the modified Charlson Comorbidity Index (mCCI)?
Methods:
We analyzed 978 patients who underwent internal fixation for pathologic fractures at two urban tertiary centers. The area under the receiver operating characteristic curve (AUC) was calculated for each predictive index to assess their accuracy in predicting 30-day morbidity across medical, surgical, utilization, and all-cause categories.
Results:
All four predictive indices demonstrated suboptimal performance, with AUC values ranging from 0.51-0.62, 0.45-0.51, 0.51-0.62, and 0.50-0.57 for medical, surgical, utilization, and all-cause morbidity, respectively. The PFMI outperformed the ASA ( P < 0.001), mF-I5 ( P = 0.018), and mCCI ( P = 0.034) in predicting utilization morbidity. It also better predicted medical ( P = 0.021) and all-cause ( P = 0.009) morbidity than ASA but did not outperform mF-I5 or mCCI in these areas. The PFMI did not surpass any indices in surgical morbidity.
Conclusion:
None of the indices reached the ideal AUC of 0.80 for any morbidity type, emphasizing the need for refinement. Updating these tools with contemporary data and exploring new prognostic factors is critical to improve morbidity risk stratification in metastatic bone disease.
Insights
The Pathologic Fracture Mortality Index (PFMI) showed limited accuracy in predicting 30-day morbidity for patients with bone metastases. While it outperformed other indices for utilization morbidity, further refinement is needed for better risk stratification in metastatic bone disease.
Area of Science:
- Orthopedic Surgery
- Oncology
- Medical Statistics
Background:
- Skeletal metastases frequently cause pathologic fractures, leading to pain and functional decline.
- Predicting postoperative morbidity in these complex cases is challenging.
- The Pathologic Fracture Mortality Index (PFMI) was developed to assess 30-day morbidity after surgical fixation of long bone fractures due to metastases.
Purpose of the Study:
- To externally validate the PFMI's predictive accuracy for 30-day medical, surgical, utilization, and all-cause morbidity in patients with long bone metastases.
- To compare the PFMI's performance against established indices: American Society of Anesthesiologists (ASA) classification, modified 5-Item Frailty Index (mF-I5), and modified Charlson Comorbidity Index (mCCI).
Main Methods:
- Analysis of 978 patients undergoing internal fixation for pathologic fractures at two tertiary centers.
- Calculation of the area under the receiver operating characteristic curve (AUC) for each index to evaluate predictive accuracy.
- Comparison of PFMI against ASA, mF-I5, and mCCI for various morbidity outcomes.
Main Results:
- All tested indices, including PFMI, demonstrated suboptimal predictive performance (AUCs 0.45-0.62).
- PFMI showed superior prediction for utilization morbidity compared to ASA, mF-I5, and mCCI.
- PFMI also outperformed ASA for medical and all-cause morbidity, but not mF-I5 or mCCI.
Conclusions:
- Current predictive indices, including PFMI, require significant refinement for accurate morbidity risk stratification in metastatic bone disease.
- Updating these tools with current data and identifying novel prognostic factors are crucial for improving patient care.
