Prediction of complications in a high-risk cohort of patients undergoing corrective arthrodesis of late stage Charcot deformity based on the PEDIS score

The aim of any treatment concept involving neuroosteoarthropathy (Charcot arthropathy) for diabetic or other neuropathic diseases is a long-term infection-free, ulcer-free, and sufficiently plantigrade stable foot in order to maintain the ability to walk independently and avoid amputation [1, 2]. The search for the ideal treatment strategy is a challenge for foot and ankle specialists; furthermore, general evidence-based treatment algorithms are lacking and the literature is inconsistent regard to both the ideal treatment option and treatment timing because of the disease’s unrelenting progression. The disease progression may lead to a loss of osteo-ligamentous architecture and consequently loss of the plantigrade foot alignment; thus, inducing subsequent soft tissue complications such as skin breakdown, recurrent ulcerations, and infections [1, 3]. Neuroosteoarthropathic patients with chronic ulceration have a 12 times higher risk of amputation, compared to those with ulcer-free feet [4]. Ulceration characteristics are described by the widely-used PEDIS classification system developed by the International Working Group on Diabetic Foot (IWGDF) [5, 6]. It describes five specific ulcer criteria; perfusion, extent, depth/tissue loss, infection, and sensation; each criterion is graded by severity (Fig. 1).

With non-operative treatment with total contact casts or walkers, a long-term ulcer-free foot can be achieved in approximately 60 % of patients if plantigrade foot alignment can be established (early stage I/II applying the Eichenholtz classification system) [1]. Patients with an advanced degree of deformity and mid- and/or hind-foot instability experience less favourable results after non-operative treatment. Multiple recastings and prolonged disability are the consequences, and 40–51 % of Charcot patients may need surgical treatment at some point in the disease’s progression [1, 7, 8]. At this time point, when signs of instability and progressive malalignment are present, reconstruction arthrodesis techniques are indicated [1, 9, 10]. Reconstruction arthrodesis techniques vary from external fixation techniques such as ring fixators to internal fixation techniques such as plates, screws, or bolts (or combinations of these techniques) [1, 2, 9, 11, 12]. Due to the absence of general algorithms, the treatment option is usually determined by individual patient factors and the surgeon’s expertise [1].

For surgery in Charcot disease patients, prolonged healing periods, high rates of infectious complications, non-union and malunion, stress fractures, fixation failure, metal-induced soft-tissue irritation, implant breakage or loosening, and concomitant high re-operation rates are frequently described [1, 7, 11‐18]. Due to these multiple disadvantages, in 2007 Pinzur et al. [1, 15] proposed that patients with high risk criteria such as a large bone deformity, an longstanding ulcer overlying infected bone, regional osteopenia, obesity, or immunocompromising illnesses do not qualify for open reduction and internal fixation. Instead, in these cases, percutaneous correction and fixation with an external ring fixator is recommended. Only in cases of available low-risk criteria (no open wounds, no history of deep infection, good bone quality, minimal diabetes-associated comorbidity, no morbid obesity), internal corrective arthrodesis is recommended [1, 15]. However this recommendation was based on the experience of one single surgeon and no detailed algorithm was provided [1]. Thus, it remains unclear how many high- or low-risk criteria are applicable to achieve the desired outcome.

In view of the foregoing, we retrospectively evaluated a consecutive series of neuroosteoarthropathic patients who underwent corrective arthrodesis of late stage Charcot mid- and hind-foot neuroosteoarthropathy with at least two out of the five positive Pinzur high-risk criteria. A 4.5 years follow-up was conducted to determine whether internal corrective arthrodesis could provide the desired outcome of a foot that was infection-free and ulcer-free for a long period, despite the presence of positive high risk criteria. The results were then quantified based on the PEDIS criteria in order to evaluate those patients that may qualify for internal corrective arthrodesis despite Pinzur high-risk criteria.

The study comprised 43 feet in 37 patients with severe Charcot neuroosteoarthropathy who underwent internal corrective arthrodesis from November 2005 to March 2012 and were retrospectively reviewed. The indication criteria for corrective arthrodesis were: (1) a clinically and radiographic non-plantigrade foot alignment; (2) a high degree of instability of the medial or/and lateral midfoot and/or hindfoot region proven by clinical examination and visualized by the changes of radiologic axes as described by Sammarco et al. [19]: positive talar-first metatarsal angle in anterior-posterior (AP) views, negative talar-first metatarsal angle in lateral views, negative calcaneal-fifth metatarsal angle in lateral views, and deviations from the neutral axis in Saltzman views; (3) the foregoing, plus clinically manifest or impending ulceration of soft tissues overlying bony deformity; and (4) failed primary conservative treatment.

The mean patient age (24 males; 13 females) was 56.7 ± 8.5 years (range 29–76). All but two patients suffered from type 2 diabetes (n = 31) or type 1 diabetes (n = 4) with insulin dependency in 94 % of the cases (n = 33). Twenty-eight patients (76 %) suffered from three or more than three secondary diagnoses. Thirty-three feet (77 %) were operated during the consolidation phase (Eichenholtz stage III). Ten feet (23 %) presented with stage II arthropathy (coalescence). Applying the topographic classification, according to Sanders and Frykberg, a mid-foot affection type II and/or III, corresponding to the Lisfranc and Chopart joint region, was visible in all feet; however, 15 radiographs showed an additional involvement of the subtalar and talocalcaneal joints, according to Sanders and Frykberg type IV.

The preoperative assessment included anterior-posterior and lateral weight-bearing radiographs of the foot and ankle joint, including Saltzman views based on the following described angles. In case of doubt regarding the peripheral vascular status, color-coded duplex sonogram was performed before surgery to reveal patients with relevant macroangiopathy. Subsequently, angiographic dilatation and stenting was performed in two patients preoperatively.

The location of the corrective osteotomy and arthrodesis was determined according to radiologic and clinical assessment, and stabilization was performed using specific implants in order to achieve maximum stability. The procedure was as follows. A medial utility incision exposed the medial column as well as the talonavicular, naviculocuneiform, and tarsometatarsal joints for either osteotomy, excision of the bony deformity, or denuding the joint surfaces from cartilage was performed as well as resection of necrotic bone; this was done in preparation for fusion and reconstruction of a plantigrade foot position. Stabilization of the medial column was performed in 21 feet (49 %) using extramedullary devices, of which 9 feet (21 %) received angular stable plates and 9 feet (21 %) received intramedullary implants only. Seven feet (16 %) were stabilized using a combination of intra- and extra-medullary implants (Table 1). With isolated collapse of the medial column, the hindfoot and lateral foot region remained untouched. Additional stabilization of the lateral column was performed in 22 feet (51 %); 6 of these feet (14 %) were stabilized with angular stable plates. Hindfoot arthrodesis was performed in 16 cases (37 %) using compression screws (n = 9; 21 %), angular stable plates (n = 6; 14 %) or locking nails (n = 1; 2 %). In order to reconstruct the osseous foot geometry, resection of the necrotic midfoot/hindfoot bones was necessary in 7 feet (16 %). In 30 cases (70 %), osseous defects were filled with autologous iliac crest grafts. One case (2 %) required lengthening of the gastrocnemius-soleus complex. In 3 cases (7 %), Achilles tendon lengthening via the Dockery technique was performed as indicated by intraoperative evaluation of tightness of the Achilles and gastrocnemius tendon complex.

Postoperatively, a lower leg splint was applied and replaced by a total contact cast with soft tissue consolidation. Mobilization was performed with partial weight-bearing of 20 kg if possible. Routine follow-up e.g., for cast replacement and radiographs, was conducted two weeks after hospital discharge and then at monthly intervals until radiographic bony consolidation was proven. Subsequently annual follow-up visits were established. Follow-up ended if amputation occurred. Mean follow-up averaged 4.5 ± 6.6 years (range 1.8–11.2).

The 4.5 years follow-up included a detailed failure analysis for the peri- and post-operative time period, focusing on complication and re-operation rates. Therefore, early (30 days postoperatively), intermediate (30 days to 5 months postoperatively) and late complications (from postoperative month 6 onward) were recorded and assessed according to their degree of severity and the need for either further surgery or conservative regime. All patients were then graded according to the PEDIS classification system [5, 6, 20] and Pinzur criteria [1, 15]. The former, developed by the International Working Group on the Diabetic Foot (IWGDF), describes the diabetic foot depending on severity of the following characteristics: lower leg/foot perfusion, ulcer extent, ulcer depth/tissue loss, infection, and sensation [5, 6, 20]. The latter, published by Pinzur in 2007 [1, 15], distinguishes high-risk criteria as a large bone deformity, a longstanding ulcer overlying infected bone, regional osteopenia, obesity, or immuno-compromising side illnesses, as well as low-risk criteria such as absence of open wounds, no history of deep infection, good bone quality, minimal diabetes-associated comorbidity, and absence of morbid obesity. On the basis of this Pinzur grouping [1, 15] and PEDIS criteria [5, 6, 20] a complication analysis was performed with the goal of evaluating those patients at risk who do not qualify for internal corrective arthrodesis. Furthermore, patient mobilization and casting length were registered.

A detailed radiologic follow-up included measurement of the AP talar-first metatarsal angle, lateral talar-first metatarsal angle, calcaneal-fifth metatarsal angle, dorsal midfoot displacement in pre-, post- and follow-up radiographs. The mean values are given in absolute numbers representing the change of angulation for the AP talar-first metatarsal angle; however, the range is given with algebraic signs; thus, negative values correspond to abduction deformity and positive values correspond to adduction deformity of the forefoot. Twenty-nine feet were available with complete radiologic follow-up.

For diabetic neuroosteoarthropathy treatment general evidence-based treatment algorithms are lacking and the literature is inconsistent regarding both the ideal treatment type and timing of treatment due to its unrelenting progression. With disease progression and unstable non-plantigrade alignment of the mid- and hind-foot, a high degree of skin breakdown and ulceration at the site of bony deformities occurs [1, 3]. Pinzur [1, 8] and other authors [21, 22] proposed the main goals: an infection-free and ulcer-free foot for a long duration together with the ability to use commercially-available depth-inlay shoes and custom-accommodative foot orthoses for maintaining long-term walking independence. Non-surgical measures such as total contact casting represent the treatment option of choice in cases of plantigrade foot positioning [19, 23, 24]. However, the progressive character of instability is associated with a serious impairment of the quality of life and an estimated risk of up to 49 % to develop recurrent ulceration together with a high risk for further complications such as infections or eventually amputation [23, 25]. Therefore, early reconstructive surgery is suggested by several authors as a valuable treatment option in the presence of severe deformity [25‐28]. Evidence-based literature does not exist, but clinical reports stress that stability can be restored with precise surgical technique, appropriate perioperative education, and postoperative therapy (assuming adequate patient compliance) [2, 27]. Nevertheless, according to Pinzur [1], certain high-risk criteria such as a large bone deformity, a longstanding ulcer overlying infected bone, regional osteopenia, obesity, or immuno-compromising illness showed increased complication rates and therefore interfere with open reduction and internal fixation. Instead, percutaneous correction and fixation with an external ring fixator is recommended for high-risk constellations

In patients with low-risk criteria, internal corrective arthrodesis is recommended [1]. Low-risk criteria include absence of open wounds, no history of deep infection, good bone quality, minimal diabetes-associated comorbidity, and absence of morbid obesity. Actually, the diagnosis of Charcot deformity is frequently delayed and only made in the deformed stages [29]. Conversely, despite prompt immobilization and protected weight-bearing, some patients develop severe deformities [29].

We retrospectively reviewed 43 feet with a deformation stage consistent with Charcot disease and determined as a high-risk group, which had at least two positive high-risk criteria according to Pinzur [1]. During a 4.5 years follow-up, a detailed complication analysis was performed which revealed high complication rates. The early postoperative period was characterized by soft tissue complications in which 75 % of patients were affected by immediate postoperative wound infections (47 %); this finding may be attributable to the endangered diabetic patient population [3] and concurs with rates of 7–26 % reported by other studies with an mean of 2.5 years of follow-up [13, 23]. The osteomyelitis rate was 9 % in our study population. It is reported to range from 12 % to 16 % after reconstructive surgery in other studies [13, 30]. However, our osteomyelitis rate is markedly lower than the 33 % osteomyelitis rate after conservatively-treated diabetic foot infections [31].

About one third (31 %) of the patients suffered from hardware breakage and 18 % developed hardware loosening with consecutive non-union. Sammarco et al. [19] observed similar rates with 32 % implant breakage in their 4.3 years of follow-up after midtarsal arthrodesis. A recent review [2], including 95 Level IV and V studies, revealed similar results and ulcer-free feet in most cases; however, it described a 22.4 % non-union rate. Implant or technique-associated problems, as well as the prolonged healing period in the altered diabetic metabolism, are likely to be causal [32]. Nevertheless, currently there is consensus that even with incomplete union, an ulcer-free plantigrade foot may be achieved [2]. We also observed 9 cases (20 %) with amputation, resulting in an annual amputation rate of 4.4 %, which concurs with Saltzman et al. [25] who reported on a cohort of 115 patients with a mean of 3.8 years of follow-up; they found an annual amputation risk of 2.7 % of cases without ulceration and an amputation risk of 28 % in cases with ulceration. These cases incurred a 23 % risk of requiring more than 18 months of casting and a high risk of about 49 % for recurrent ulceration. In our study population, 36 % of the patients suffered from recurrent ulcerations. Illgner et al. [11] reported on 205 patients with an external fixator and a follow-up period of 21 months; they found a 25 % re-ulceration rate. A study of 115 conservatively treated patients with 3.8 years of follow-up by Saltzman et al. [25] showed 49% re-ulceration rates. Altogether, we observed 2.9 complications/foot leading to the question: “Can those patients at risk for a complicated postoperative course be detected in preoperative planning?”.

In view of the foregoing we developed a potential predictor that would reveal those high-risk patients qualifying for internal corrective arthrodesis despite this risk. It was to take the PEDIS single group of criteria into account: feet without signs of PAD, an ulcer extent <0.9 cm², an ulcer depth up to erosion, and inflammation including only skin and subcutaneous tissue. This group of criteria alone showed significantly lower overall complication rates (P1, E1, D1, I2, (S)). For this degree of ulceration, the data also significantly predict lower amputation rates and for this degree of ulceration depth, they significantly predict lower amputation and reoperation rates. Since all factors but sensation showed significant influences on complication rates, we think that the cumulative PEDIS count rather than the PEDIS single criteria is practicable for clinical use because the former takes the entire patient into consideration. This study significantly revealed a favourable outcome in terms of overall complication, re-ulceration, and amputation rates for patients/feet with a cumulative PEDIS count below 7. The cutoff value of 7 may aid clinical decision-making in preoperative planning for a Charcot deformity and an internal corrective arthrodesis. By implication, these results endorse the importance of early diagnosis for optimal treatment planning.

In regard to radiologic evaluation, the goal of reconstruction is to correct the non-plantigrade foot position and to establish a neutral AP tarso-first metatarsal angle, a positive lateral tarso-first metatarsal angle, a calcanear-fifth-metatarsal angle, and a dorsal midfoot displacement correction to normal values. Correction was completely achieved in 83 % of the AP talar-first metatarsal angle, in 97 % of patients for the lateral talar-first metatarsal angle, 90 % of patients for the calcaneal-fifth metatarsal angle and 83 % of patients with dorsal midfoot displacement; these findings concur with those of other studies [19, 30]. In addition, at final follow-up a significant recollapse of the longitudinal foot arch for both the medial column and, to a greater extent, the lateral column had to be observed in order to minimize the number of patients with persistent correction to 62 % medially and 35 % laterally. Correction of AP talar-first metatarsal angle recollapsed in 17 % of feet. Recollapse of the midfoot, measured by the dorsal midfoot displacement, lowered radiographic success rates to 59 % of the patients. Wiewiorski et al. [30] similarly experienced the worst recollapse rate for the lateral tarso-first metatarsal and calcanear-fifth-metatarsal angles. Those radiographic measurements reveal a good overall reliability for reconstructive techniques in our patient population. The significance of those radiographic measurements in regard to patient satisfaction is unknown [30]; this adds to the difficulty of radiographic evaluation, due to its putative measurement error, to determine weaknesses of this study. In addition, we noted another limitation: patient satisfaction was not evaluated in this study. This limitation is of significant importance. By forming groups based on clinically defined boundaries, heterogeneous cohorts may result, thus, leading to a statistical error. The study design was retrospective with specific inclusion criteria. Therefore, the limited patient sample negatively impacted generalizations.

Internal corrective arthrodesis in the more advanced stages of Charcot deformity can provide adequate reconstruction with intraoperative radiographic measurements. Superior stability was achieved for the medial column in this 4.5 years follow-up period with the lateral column showing less stability and a tendency to recollapse. Despite the favourable radiologic results, overall complication rates were high. A good predictor implementing Pinzur and PEDIS criteria was developed to identify those high-risk patients qualifying for uncomplicated internal corrective arthrodesis. Despite two-to-five positive Pinzur high-risk criteria, a favourable outcome in terms of overall complication, re-ulceration, and amputation rates for patients/feet with a cumulative PEDIS count below 7 was found. The cutoff value of 7 may aid clinical decision-making during preoperative planning for Charcot deformity and internal corrective arthrodesis.