Fractures and Osteoporosis in Patients With Diabetes With Charcot Foot

The Charcot foot, also known as neuropathic arthropathy, is a rare yet serious disease described as a progressive degeneration of the bone, joints, and soft tissue of the foot and ankle (1). The disease presents as a hot, red, swollen foot, and in the chronic phase, the foot loses its original architecture, causing deformities such as the “rocker-bottom” deformity (1). The pathogenesis behind Charcot foot is yet to be fully understood, but previous studies have shown that patients with diabetes with a Charcot foot may have increased bone metabolism, increased osteoclast activity, increased bone resorption, and increased inflammatory markers (2–4). It is well known that patients with diabetes in general have an increased risk of fracture compared with individuals without diabetes (5–8), but whether patients with diabetes with a Charcot foot have an increased risk of fracture, beyond that of patients with diabetes without a Charcot foot, is unknown.

Some studies of bone mineral density (BMD) measured by DXA in smaller populations of patients with diabetes with a Charcot foot indicate low BMD in the affected foot compared with the unaffected foot (2,9–15). However, the decrease in BMD only seems to be applicable for the lower extremities, as no decrease in BMD of the hip, spine, or total body of patients with diabetes with a Charcot foot has been reported (9–11,13). This indicates that the bone pathology of Charcot foot may be a local condition, but no studies have investigated whether patients with diabetes with Charcot foot have an increased risk of osteoporosis. Thus, the aim of this study was to investigate whether patients with diabetes with a Charcot foot have an increased risk of fracture and/or an increased risk of osteoporosis compared with patients with diabetes without a Charcot foot.

This study is a register-based, nationwide, population-based matched cohort study, where exposed patients are matched with unexposed patients at the onset of exposure.

The study population was based on data from the Danish National Patient Register (NPR). The NPR contains information on all hospitalizations from 1977 to present and additionally, information on outpatient visits and emergency department contacts from 1995 to present (16). Since 1995, the ICD-10 has been used in Denmark. Data about age, sex, immigrations, and emigrations were found in the Danish Civil Registration System (CPR) (17), in which all Danish individuals born and residing in Denmark are registered with an unique identification (ID) number. This number is used in all contacts with the health care system as well as public authorities, which makes it possible to link data from civil registers to health care registers. Dates of death were found in the Death Register (DR), Statistics Denmark (18).

The study population consisted of all individuals in the NPR with an ICD-10 code for diabetes (DE10, DE11, DE13, or DE14) from 1 January 1995 to 31 December 2018. We linked the patients in NPR, CPR, and DR using anonymized versions of the unique Danish ID numbers.

Only information on hospitalizations, and not information on outpatient visits and emergency room contacts, are available from before 1995. All patients with a closed hospital contact related to diabetes before 1995 were excluded. Furthermore, patients with a diabetes code registered after the date of death or after the date of emigration (temporary or permanent) were excluded.

Patients with Charcot foot were identified as having the ICD-10 codes DM146 (neuropathic arthropathy) or DM142 (diabetic arthropathy) and were matched on sex and date of diabetes ±365 days with 10 random patients with diabetes without Charcot foot. This ensured that the patients with and without Charcot foot were diagnosed with diabetes within 365 days of each other. Patients with Charcot foot with fewer than three matches were excluded. Patients with unspecified type of diabetes (ICD-10 code DE13 and DE14) were excluded.

Exposure, outcome, and covariates were based on ICD-10 codes from NPR. First outcome “any fracture” was identified as having any of the following ICD-10 codes: DS02, DS12, DS22, DS32, DS42, DS52, DS62, DS72, DS82, or DS92. Outcome “fracture lower leg” was identified as having the ICD-10 code DS82 (includes fractures in patella, tibia, or fibula but not in the tarsal bones or distal femur), and outcome “fracture foot” was identified as having the ICD-10 code DS92. The outcome “osteoporotic fractures” included vertebral fractures in the thoracic or lumbar spine, hip fractures, proximal humerus fractures, and wrist fractures (ICD-10 codes DS22.0, DS22.1, DS32.0, DM48.5, DS42.2, DS42.3, DS42.4, DS42.9, DS52.5, DS52.6, DS52.8, DS62.8A, and DS72.X).

The outcome “osteoporosis” was defined as having any of the following ICD-10 codes: DM80, DM81, or DM82. Osteoporosis was divided into the subgroups “prior to date of diagnosis of Charcot foot” and “after date of diagnosis of Charcot foot.”

Diabetic complications were defined as diabetic neuropathy (DE1X0.4), diabetic nephropathy (DE1X0.2), diabetic retinopathy (DH360), diabetes with eye complications (DE1X0.3), diabetic foot ulcer (DE1X0.5B), diabetes with complications in the peripheral vascular system (DE1X0.5), and diabetes with multiple complications (DE1X0.7). Cardiovascular comorbidities were defined as hypertension (DI10), apoplexia (DI61, DI63, or DI64), ischemic heart disease (DI259), atherosclerotic cardiovascular disease (DI251), angina pectoris (DI20), and acute myocardial infarction (DI21).

We summarized categorical values using frequencies and percentages and used the χ² test to test for differences between patients with diabetes with and without Charcot foot. Continuous variables were summarized using mean and SD, and differences between diabetes patients with and without Charcot foot were tested using the t test. Odds ratios (ORs) with 95% CIs of fractures and osteoporosis were calculated using conditional logistic regression. The Wald χ² test was used to test for statistical significance. All analyses and data management were completed using SAS 9.4 software (SAS Institute, Cary, NC). P values <0.05 were considered statistically significant.

The study was approved by the Knowledge Center for Data Reviews, the Capital Region, Hillerød, Denmark (number P-2019-23). As data were accessed through a secure connection provided by Statistics Denmark and ID numbers were anonymized, person-sensitive data are well protected. Ethics approval is not required for register-based studies in Denmark.

As shown in Fig. 1, we identified 350,736 patients with diabetes from 1 January 1995 until 31 December 2018. A total of 1,084 patients with a “completed hospital contact” related to diabetes ending before 1995 were excluded. Furthermore, 53 patients with a diabetes code registered after the date of death were excluded, and 40,042 patients were excluded due to emigration (temporary or permanent). After these initial exclusion criteria, the number of patients with diabetes was 309,557, and of these, 1,655 were diagnosed with Charcot foot. After matching, the study cohort therefore consisted of 16,550 patients with diabetes with Charcot foot and 16,550 without Charcot foot. Finally, 307 patients with unspecified diabetes were excluded. Thus, the final study cohort consisted of 17,898 patients, namely 1,602 patients with diabetes with Charcot foot and 16,296 patients with diabetes without Charcot foot. We found 64% of the patients with Charcot foot and 37% of those without Charcot foot were registered with both a DE10.X and a DE11.X code, which is why separation into patients with type 1 diabetes or type 2 diabetes was considered invalid.

In the patients with diabetes with Charcot foot, the mean age at the diagnosis of Charcot foot was 60.1 ± 11.9 years, and 64% were men. These patients were slightly younger when diagnosed with diabetes (52.8 ± 13.5 years) than those without a Charcot foot (55.2 ± 17.6 years, P < 0.001), but their duration of diabetes was similar (i.e., 13.7 ± 6.3 years from diagnosis of diabetes until death or follow-up).

The patients with diabetes with a Charcot foot had more microvascular complications, hypertension, foot ulcers, and complications in the peripheral vascular system (P < 0.001) than those without Charcot foot, but not in cardiovascular disease (Table 1).

The patients with diabetes with a Charcot foot had a higher proportion of any fracture, osteoporotic fractures, fractures of the lower legs, and fractures of the foot compared with the patient with diabetes without a Charcot foot (Table 2). Accordingly, the patients with diabetes with Charcot foot had a higher OR of any fracture, osteoporotic fractures, fractures of the lower legs, and fractures of the foot compared with the patients with diabetes without Charcot foot, as shown in Fig. 2, with ORs of 1.8, 1.3, 2.4, and 2.9, respectively.

The time interval between diagnosis of Charcot foot and any subsequent fracture was on average 4.3 ± 3.9 years. A fracture in the foot occurred in 15% within 3 months and in 26% within 12 months after the diagnosis of Charcot foot.

Patients age <60 years when diagnosed with Charcot foot had a higher frequency of subsequent fractures than those age >60 years when diagnosed with Charcot foot (OR 1.4 [95% CI 1.1–1.7], P = 0.002). This may be explained by confounding by longer observation periods for the younger patients, as the frequency of subsequent fractures was significantly higher 12 years after the diagnosis of Charcot foot compared with the first years after the diagnosis of Charcot foot (OR 2.6 [95% CI 1.7–3.9], P < 0.001).

More women than men had fractures, both in the patients with diabetes with Charcot foot (all fractures: 54.7% vs. 47.7%, P < 0.001), as well as in those without Charcot foot (all fractures: 44.0% vs. 31.2%, P < 0.001).

A total of 11.5% of the patients with diabetes with Charcot foot had osteoporosis compared with 9.1% of the patients without Charcot foot (Table 2). The OR of osteoporosis was 1.3 (95% CI 1.1–1.5, P < 0.01) for patients with diabetes with Charcot foot compared with those without Charcot foot. This OR of osteoporosis appearing after the diagnosis of Charcot foot was 1.4 (95% CI 1.2–1.8, P < 0.001), whereas the OR of osteoporosis occurring prior to the diagnosis of Charcot foot was 1.1 (95% CI 0.9–1.4, P = 0.12). More women than men had osteoporosis, both in the patients with diabetes with Charcot foot (18.9% vs. 7.4%, P < 0.001) and in those without Charcot foot (15.8% vs. 5.4%, P < 0.001).

The current study of 1,602 patients with diabetes with Charcot foot and 16,296 matched patients with diabetes without Charcot foot is so far the largest study of the disease Charcot foot in patients with diabetes, and it is the first nationwide study. In our study, the patients with diabetes with Charcot foot had a significant increased risk of fracture compared with patients with diabetes without Charcot foot. The increased fracture risk was most pronounced in the lower extremities. Furthermore, we found that patients with diabetes with Charcot foot had an increased risk of osteoporosis, especially after the diagnosis of Charcot foot.

An increased risk of fracture in patients with diabetes with Charcot foot is in line with previous studies, where patients with diabetes with a Charcot foot seem to have bone weakening beyond that of diabetes in general (9–11,13), with increased bone turnover, increased osteoclast activity, and increased local as well as systemic inflammatory markers (2–4). The increased risk especially of fracture of the lower leg and foot observed in our study is also in accordance with the current evidence that Charcot patients may have decreased BMD in the lower extremities and in the Charcot foot. Furthermore, the patients with diabetes with Charcot foot in our study had an increased risk of osteoporotic fractures as well as having the diagnosis of osteoporosis compared with those without a Charcot foot. However, it cannot be ruled out that fractures in the foot may be an indicator of a new onset of—or existing—Charcot foot, but a long duration between the diagnosis of the Charcot foot and fracture makes it more unlikely.

In previous smaller studies of patients with diabetes with Charcot foot, the BMD in the skeleton in general, as in the lumbar spine and hip, was not decreased (15–17). In this study with thousands of patients, the statistical power to detect an increased risk of fracture and osteoporosis is much better. Furthermore, the association of low BMD and fracture is not good in patients with diabetes, where the increased risk of fracture may be due to other components of the bones, such as the quality, geometry, or collagen (19). Nonbone-specific risk factors of fractures, such as likelihood of falling, balance problems, disturbed vision, etc., also increase the risk of fractures. In this study, as in previous studies (20), the patients with diabetes with Charcot foot had a higher proportion of almost all types of diabetes complications compared with those without Charcot foot, which emphasizes that patients with diabetes with Charcot foot have more serious disease compared with those without Charcot foot. Thus, the increased frequency of neuropathy, retinopathy, and nephropathy may all contribute to the increased risk of fracture and osteoporosis seen in the diabetes patients with Charcot foot.

In acknowledgment of diabetes patients having an increased risk of fractures, a working group on bone and diabetes of the International Osteoporosis Foundation (IOF) has recently suggested, that the T-score, based on bone mineral measurement by DXA, in the lumbar spine or hip below −2.0, instead of the usual −2.5, should be used as the threshold for diagnosis of osteoporosis and treatment for osteoporosis in patients with diabetes (19). Thus, given the even more increased risk of fracture and osteoporosis in patients with diabetes with Charcot foot, these patients should be evaluated aiming at decreasing their risk of fractures. Several drugs used for treatment of osteoporosis have been investigated as treatment of the acute Charcot foot, but so far all have failed that purpose. But there have been no studies of pharmacologic treatment of osteoporosis and fractures in the patients with diabetes with Charcot foot (21,22). Such randomized clinical trials are recommended. Markers of inflammation and bone resorption (IL-6 and the RANK-L–to–OPG ratio) are increased in the acute Charcot foot. Therefore, an optimal medical treatment of patients with diabetes with an acute Charcot foot, as well as of osteoporosis, may be through an inhibition of both bone resorption and inflammation by targeting the RANK-L/OPG system. This can be done with denosumab, a monoclonal antibody against RANK-L that inhibits bone resorption. Denosumab is approved for treatment of osteoporosis as well as for prevention of skeletal-related events in patients with bone metastases (22).

On the other hand, as suggested by IOF, patients with diabetes in general should be evaluated and treated for osteoporosis, as it seems that patients with diabetes may benefit from this (19) and probably patients with diabetes with Charcot foot may as well. Thus, the patients with diabetes with Charcot foot did not have increased frequency of osteoporosis diagnosis prior to the diagnosis of Charcot foot, which may be because patients with diabetes are not screened for osteoporosis until they have had a Charcot foot with long-term immobilization.

This is the largest study ever to investigate patients with diabetes with Charcot foot, and it includes all patients with diabetes with Charcot foot in Denmark during a 23-year period. The study population is a large, unselected population of patients with diabetes. Patients with diabetes with Charcot foot are often found in highly specialized diabetes clinics and are generally compared with other patients with diabetes in the same clinic. Our study population of patients with diabetes is more comparable with the population with diabetes in Denmark. This study matches patients with diabetes with and without Charcot foot on the date of diagnosis of diabetes. Therefore, all patients with Charcot foot and their matches have had diabetes during the same time period, thus ensuring that the same treatment possibilities were available.

Our study does, however, have some selection bias, as we included only individuals with an ICD-10 code of diabetes in the Danish NPR. Patients with diabetes who have never been hospitalized will not be registered with an ICD code for diabetes. This may underestimate the number of patients with diabetes in Denmark. However, the total number of patients with diabetes in Denmark in 2019 has been estimated to be ∼370,000 (23), which is not that different from our population with diabetes of ∼350,000. Patients with diabetes who have never been hospitalized and are only being treated in the primary sector tend to be healthier, with fewer complications, compared with patients with diabetes treated in the hospital, which may underestimate our findings, because the healthiest patients with diabetes are not included in the comparator group of patients with diabetes without Charcot foot.

On the other hand, it is likely that all patients with diabetes with a Charcot foot, a serious health-threatening condition in patients with multiple complications, will have received treatment at a hospital for the Charcot foot and will therefore be registered with the ICD-10 code for Charcot foot.

In conclusion, in the present register-based, nationwide, population-based matched cohort study of 1,602 patients with diabetes with Charcot foot and 16,296 patients with diabetes without Charcot foot, the patients with diabetes with a Charcot foot had an increased risk of fractures and osteoporosis compared with patients with diabetes without a Charcot foot.

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

Author Contributions. O.C.R. wrote the protocol, researched and analyzed the data, and wrote the manuscript. M.W.-J. researched and analyzed the data, contributed to the discussion, and reviewed and edited the manuscript. K.H.A. contributed to the discussion and reviewed and edited the manuscript. O.L.S. contributed to the protocol and to the discussion and reviewed and edited the manuscript. O.C.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.