Lower-Extremity Amputation Risk After Charcot Arthropathy and Diabetic Foot Ulcer

From a national diabetic population treated in the VA in 2003 (6,7), we identified patients who were newly diagnosed with Charcot arthropathy (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] 713.5) and with a diabetic foot ulcer (707.1x or 707.9) but without Charcot arthropathy in the years 2002–2007. In both groups, a condition was determined as newly diagnosed in 2003 if it was not found in any utilization records in 2002. We used 2002 as a full-year washout period (8).

Major and minor amputations were identified from the VA inpatient and outpatient records using ICD-9-CM and Current Procedural Terminology (CPT)-4 codes (foot, 84.10–84.12, 28800–28825; ankle or leg, 84.13–84.15, 27880–27889; knee or above, 84.16–84.17, 27590–27598). Data for known risk factors of amputation including age, sex, race, marital status, diabetes duration and control, and all coexisting conditions in the Elixhauser comorbidity method (9) were obtained from patient records for 2003.

We computed the time to event by following patients in either group from the date of the first diagnosis for up to 5 years to the first date of amputation and analyzed it using Cox proportional hazards regression.

Our initial analysis suggested that amputation risks were not significantly different between the two groups but that Charcot patients with foot ulceration had remarkably higher risks than individuals without. We thus stratified the study cohort into three groups (patients with Charcot alone, patients with Charcot and foot ulcers, and patients with foot ulcers alone) and compared amputation risks for patients in the three groups. For Medicare beneficiaries, some amputations might have been performed in non-VA hospitals and were not observable in the VA data. We therefore analyzed the data separately for patients who were <65 years old at the end of follow-up and patients who were ≥65 years old. This study was approved by our institutional review board.

From the VA cohort of diabetic patients in 2003, we identified 911 patients with incident Charcot arthropathy and 15,117 patients with incident diabetic foot ulcers after eliminating patients with previous history of lower-extremity amputations. Crude amputation rates were 14.7% for Charcot patients and 14.5% for foot ulcer patients. After a mean follow-up of 37 ± 20 and 43 ± 18 months for Charcot and ulcer groups, respectively, patients with Charcot arthropathy experienced 4.1 amputations per 100 person-years compared with 4.7 for patients with diabetic foot ulcers (Mantel-Haenszel rate ratio = 0.88; P = 0.15).

Among Charcot patients, 538 (59%) were treated for foot ulceration between 2002 and 2007; 66% (354 patients) were treated for foot ulceration immediately before or concurrently with Charcot arthropathy, and 34% (184 patients) experienced it as a complication. Compared with patients with Charcot alone, those with ulcer alone had 7 times higher risk and those with both an ulcer and Charcot had 12 times higher risk of amputation among patients <65 years (Table 1). Similarly, the two groups had 9 and 13 times higher risks than the reference for patients ≥65 years old.

Our results show that amputation risk for Charcot arthropathy overall is not significantly different from that for diabetic foot ulcers. When Charcot patients were stratified by foot ulceration, Charcot alone was associated with low risk (<2%) and ulcerations were responsible for most amputations experienced by Charcot patients.

These results are consistent with the current practice guideline suggesting that prevention of ulceration is critical for Charcot limb salvage (10). They further call into question whether surgery is advisable early in the disease process. Feet affected by Charcot arthropathy are unlikely to ulcerate when they remain clinically plantigrade and the radiographic weight-bearing relationship between the hind foot and forefoot is collinear (11,12). These results suggest that amputation risk for Charcot arthropathy may be reduced by reserving corrective surgeries for patients with a high risk of Charcot-related ulceration.

Our results also suggest that patients with Charcot arthropathy in the community may have considerably higher risk of amputation than previously believed. The rate of 6.6% in a meta-analysis mentioned above represent half the rate observed for the VA Charcot patients (14.7%). This may be attributable to the secondary ulceration rate in this VA cohort (34% during a 5-year follow-up), which is higher than others with corrective surgical intervention (12,13) and comparable to some series reported from specialty clinics (14,15).

Our results are consistent with a previous study of amputation risks of diabetic foot ulcer. Moulik et al. (4) followed diabetic foot ulcer patients for a median of 28 months and reported a 5-year rate of 19%. The VA cohort consisting of foot ulcer patients, who were overall older than the Moulik et al. sample, had a 5-year rate of 20%.

One limitation of this study is that it relied on patient data recorded for administrative purposes. Regarding the accuracy of data, our chart reviews suggest that diabetic foot ulcers can be identified from administrative data with 93% sensitivity and 91% specificity. We also confirmed the presence of Charcot arthropathy for 12 of 13 patients (92%) through medical records. Importantly, however, we could not link the affected limb at study entry with the limb amputated during follow-up because of the lack of information in ICD-9-CM codes. Also, we did not have access to data for Medicare use by elderly VA users, and some amputations performed in the non-VA hospitals could not be observed. For this reason, the findings for patients aged ≥65 years need to be interpreted carefully.

In summary, diabetic foot ulcers pose a significant risk for amputation, while Charcot arthropathy does not unless it is complicated by an ulcer. Further research is needed to evaluate comparative effectiveness of corrective surgery versus accommodative treatments in preventing amputations for Charcot patients (10,13).

The authors acknowledge the financial support from the Center for Management of Complex Chronic Care, Hines Veterans Affairs Hospital, Hines, IL (LIP 42-512; Elly Budiman-Mak, MD, principal investigator).

No potential conflicts of interest relevant to this article were reported.

We thank Elissa Oh, Huiyuan Zhang, and Lishan Cao, who helped in manuscript preparation and the analytic data construction. The corresponding author had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.