Medium-term outcomes of multi-disciplinary surgical management of non-ischemic diabetic heel ulcers

Abstract

Background

Heel ulcers in patients with peripheral neuropathy and diabetes pose a significant challenge to treating physicians. Infection spreading to the os calcis is associated with a poor prognosis. There is no consensus on which method of surgical treatments results in better outcomes. The aim of this study was to assess patients’ survival, rate of ulcer healing following surgical treatment, along with limb salvage rate, time taken for healing, ulcer recurrence and patients’ functional outcome after healing.

Methods

We studied 29 patients (20 men, 9 women) presenting with diabetic neuropathic heel ulcers (30 feet) and no critical limb ischemia, were managed surgically in our unit and followed-up for a minimum of 12 months. We assessed their clinical and functional outcomes within a mean follow-up period of 28 months (12–83).

Results

11 out of 29 patients died (38%) with mean duration of post op survival being 31months (range 4–70). 14 ulcers (50%) healed fully within a mean of 360 days (131–1676). Limb salvage was achieved in 29 feet (97%). Only 5 out of 17 patients with diabetic nephropathy (p value 0.016) and 9 out of 24 ulcers with calcaneal osteomyelitis (p value 0.044) achieved full ulcer healing. Ulcer recurrence rate was 36% (5/14) within 12 months of achieving ulcer healing. Six patients were able to return to independent walking in surgical shoes while 11 patients were mobilising using either a crutch or frame.

Conclusion

While excellent limb salvage can be anticipated from the outcome of surgically managed infected heel ulcers in patients with diabetes, complete healing can still be slow and unpredictable. Significant medical co-morbidities in these patients make them vulnerable to medium-term post-operative complication and survival.

1. Introduction

In 2019, the International Diabetes Federation reported nearly 463 million adults were living with diabetes and this figure has the potential to rise to a staggering 700 million by 2045.¹ With the rising trend in the number of people affected with diabetes globally, owing to associated peripheral neuropathy,^2,3 the incidence of diabetic foot ulcers (DFU) is expected to rise as well. An earlier study suggested that the life time risk of a person with diabetes developing a foot ulcer stretches to 25%⁴ but a recent report estimated it to be as high as 34%.⁵ Close to 50% of these ulcers become infected⁶ thereby increasing the risk of major amputation.

The existing body of evidence reveals that a major amputation in patients with diabetes is associated with a high rate of mortality.⁷,⁸,⁹,¹⁰ A 2018 study by Gurney et al. reported a substantial 11% death rate within 30 days following a major amputation, and 18% within 90 days.⁷ Thorud et al. conducted a systematic review in 2016 on mortality following a major amputation among patients with diabetes and peripheral vascular disease. They found a significant 40%–82% mortality after below-knee and 40%–90% after above-knee amputation within 5 years.⁸ Putting aside 5-year mortality, their likely established state of lower limb neuropathy raises their propensity for ulceration in the remaining contra-lateral foot, following a major amputation. A new ulcer developing in the surviving foot can potentially also affect patients’ optimism for recovery and return to function.

Whilst the exact incidence of heel ulcers in patients with diabetes is still unclear¹¹,¹², the heel is the second most common anatomical site for developing a pressure ulcer.¹³ Once the infection seeds to the underlying bone, the prognosis is believed to be poor. Paisley et al. have reported a 33% healing rate at 30 weeks in those with diabetic calcaneal osteomyelitis in comparison to 94% healing in those with heel ulceration without osteomyelitis.¹⁴

Treatment usually demands resection of a key weight bearing area of the tuberosity within the proximity of Achilles tendon insertion. Earlier studies have reported varying healing rates from 35% to 74%.¹⁵,¹⁶ Nevertheless, with evolving surgical techniques and treating teams becoming more proficient in the management of these calcaneal heel ulcers, superior outcomes, with 100% limb salvage, have been reported.¹⁷,¹⁸

In this single unit case series, we examined the clinical outcomes of patients presenting to our unit with infected heel ulcers, without critical limb ischemia, with or without calcaneal osteomyelitis (COM), and who underwent surgical debridement performed by the orthopaedic team. Our aim was to assess patients’ survival rate, limb salvage, rate of ulcer healing following surgical treatment, time taken for healing to occur, ulcer recurrence and patients’ functional outcome after healing.

2. Patients and methods

2.1. Patient population

A retrospective study of consecutive cases was conducted using prospectively collected data from our hospital records. We carried this out as an internal service evaluation, hence no ethics committee approval was required. All patients, who underwent a surgical debridement of a heel ulcer carried out by the orthopaedic team, between June 2010 and April 2019 were included in this study and those that had a minimum of 12-month follow-up were subjected to a detailed analysis. Patients were identified from our departmental database and cross checked with the theatre activity information register. We identified 29 (20 men and 9 women) patients with a heel ulcer (n = 30) who underwent surgical debridement under the orthopaedic team. All debridements were carried out as in-patients. Baseline characteristics of our patient group are outlines in Table 1. Whilst we work closely with our vascular colleagues within in the Multidisciplinary Diabetic Foot Team (MDFT), patients with heel ulcers accompanied by features of critical limb ischemia were treated by the vascular team independently and not included in this study.

Table 1.

Patient characteristics.

N = 29 patients (30 ulcers), mean age 59 years (range 28–91)
Male: Female	20:9
Type 1: Type 2 diabetes	11:18
Mean HbA1c (in patients with diabetes)	9.6% (6.2–15.2)
History of previous revascularization	6
Number of patients with nephropathy	17
Number of patients on renal dialysis	4

The mean age of our cohort was 59 years (28–91). All patients suffered from peripheral neuropathy related to diabetes, with a vibration perception threshold (VPT) measured using a neurothesiometer (Horwell, S.L.S Ltd, Nottingham, U.K), of greater than 25 V¹⁹. Lower limb blood flow was assessed prior to surgical debridement in our vascular laboratory, in the form of arterial duplex scan and considered optimal if flow beyond level of trifurcation, especially in posterior tibial artery was either triphasic or biphasic. Probe-to-bone test was routinely performed, to aid the diagnosis of osteomyelitis. Despite the obvious presence of either active or low-grade soft tissue infection, all patients had undergone MRI imaging studies of the foot prior to surgery to assess the presence and extent of COM. Routine blood tests including full blood count, C-reactive protein, and collection of deep tissue specimens from the base of ulcer for microbiology culture and sensitivity, were performed in all patients, as a part of our initial assessment. The presence of diabetic nephropathy, defined as a creatinine level of higher than 120 μmol/L, was assessed in all patients.

The indication for surgery included a combination of the following features: (1) The presence of active infection of the ulcer not effectively controlled by a combination of off-loading and culture-specific antibiotic therapy. (2) Failure to show signs of ulcer healing progression after a minimum of 4 weeks of non-operative treatment. (3) The presence of osteomyelitis in the calcaneum (Fig. 1). All patients were managed by the MDFT, with the core team consisting of Endocrinologists, Orthopaedic surgeons, Vascular surgeons, Podiatrists, Diabetic Foot Practitioners, Orthotists, and Microbiologists.

Fig. 1.

Severely infected heel ulcer on presentation (A), Radiological appearance of COM (B,C), VAC assisted wound management following debridement (D), granulating wound base at 2 weeks (E), wound healing progress at 3 months (F), complete healing at 5 months (G) and XR appearance of healing (H).

2.2. Surgical treatment

All patients were treated with surgical debridement of the heel ulcer with or without partial calcanectomy based on the presence and degree of osteomyelitis. The procedures were performed under general anaesthetic with patients positioned either on their side or supine, dependent on the operating surgeon’s preference. All infected soft tissue was excised until a clear margin of healthy tissue was reached.²⁰ Multiple tissue specimens were collected intraoperatively and sent for microbiological analysis and antibiotic sensitivities.

When COM was present, partial calcaneal excision was performed using an oscillating saw. The amount of bone resection was based on the extent of existing OM in the calcaneum as noted on MR imaging and substantiated by intraoperative visual confirmation of a healthy cancellous bone resection margin. From 2016, we started using Antibiotic Loaded Bone Substitute (ALBS) in appropriate cases to facilitate local delivery of high concentrations of antibiotic into the calcaneum and surrounding tissues. ALBS was placed either as beads in the bone and soft tissue voids (Stimulan® Biocomposites Ltd, Keele, United Kingdom) or, injectable biocomposite (CERAMENT®, Bonesupport, Lund, Sweden), was injected into multiple drill holes made in the calcaneum using a silo technique as described by Dramploas et al.¹⁸ Choice of antibiotic in ALBS was influenced by the type of bacteria grown from the initial the wound swab.

The debridements were carried out as single stage procedures. Thorough wound lavage using normal saline was undertaken after aggressive debridement of all infected tissue. Once cleaned, it was packed with diluted Povidone-iodine (Betadine® solution) soaked gauze and a well-padded back slab splint was applied. In order to allow us to monitor the control of infection, we do not routinely attempt to close infected heel wounds.

2.3. Post-operative management

All patients received culture specific systemic antibiotic therapy and were closely monitored by the MDFT clinically and serologically. The duration of antimicrobial therapy was determined by the MDFT based on culture results, degree of bone involvement and healing response. Stringent glycemic control was maintained peri-operatively, under the guidance of the Diabetic Physicians. The Diabetic Foot Practitioners provided advanced wound care by applying negative pressure wound therapy (NPWT) and performing regular bedside sharp wound debridement to promote healing.

Patients were discharged from the hospital in a bivalved off-loading total contact cast when satisfactory wound healing progress and decline of inflammatory markers were evident. Regular out-patient reviews were arranged, initially at weekly intervals, gradually spaced out to a 4–6 week intervals, as determined by clinical progress. NPWT was continued in the community until there was a healthy granulation tissue covering the entire ulcer to skin level. The off-loading cast was continued until skin epithelialisation was achieved. Those feet that did not achieve skin epithelialisation underwent skin graft or fasciocutaneous flap procedures. All patients with healed ulcers were referred to the orthotist for bespoke footwear.

2.4. Statistics

Chi squared test was used to examine the mean differences for statistical significance. A p-value < 0.05 was considered statistically significant.

3. Results

Out of 29 patients included in the study, two died within 12 months of medical causes (one at 4 months and another at 8 months) post-operatively and these have been excluded from analysis of functional outcomes. An additional 9 patients died during the follow-up period, giving an overall mortality rate during the study period of 38% (11/29). Their deaths were due to various medical causes and not related to heel ulcers. Mean duration of post op survival in this group was 31months (range 4–70).

The mean duration of heel ulcers prior to presentation to our multi-disciplinary clinic was 1.7 years (range 4 months–9 years) and that of post-operative follow-up was 27 months (range 12–83). Signs of acute infection such as discharging pus and significant erythema were noted on two patients while the rest of the group presented with heel ulcer associated chronic infection. Of the 30 heel ulcers (29 patients), 24 were associated with COM confirmed on imaging studies. There was no significant difference in the size of ulcers between those with and without COM. 3 patients in the COM group presented with an associated pathological fracture of the calcaneum (Fig. 2). Partial calcanectomy was performed for all patients with COM (24 feet). The ulcer characteristics and treatment outcomes are shown in Table 2.

Fig. 2.

Acute on chronic heel ulcer in a diabetic patient with pathological fracture in osteomyelitic calcaneum (A,B), after debridement with partial calcanectomy where fracture stabilization was achieved using K-wires (C,D), after complete healing at 5 months with weight-bearing lateral X-rays (E,F).

Table 2.

Ulcer Characteristics and post-operative outcomes.

	Heel ulcer with OM (N = 24)	Heel ulcer without OM (N = 6)
Mean ulcer size in cm2 (Range)	15.4 (0.9–52.4)	14 (0.6–41.5)
Presence of associated calcaneal fracture	3	0
Ulcer healing achieved	9	5
Recurrence of ulcer	4	1
Major Amputation	1	0
Death within 1 year	2	0
Total death in follow up period	9	2

Limb salvage was achieved in 29 feet (97%). One patient with bilateral heel ulcers opted for below knee amputation on one side when healing become stagnant, despite a series of surgical debridements. Equally, this patient’s heel ulcer on the opposite leg went on without healing. Two patients with protracted non-healing ulcers, despite surgical debridement followed by close medical and podiatric management, were discussed for potential below knee amputation, but the patients opted for limb retention at that stage.

Only 14 ulcers (50%) achieved complete healing following surgical debridement and wound care during the follow-up period. Seven of these patients received either skin graft or dermal substitute and a further two needed fasciocutaneous flaps to achieve skin coverage. The mean time taken to reach full healing was 360 days (range 131–1676 days). Seventeen patients (17/29, 59%) in our group suffered from diabetic nephropathy and four of these were on renal dialysis. Full ulcer healing was achieved in only five feet (5/17, 30%) in the nephropathy group. This contrasts with the ulcer healing noted in 9 out of 12 (75%) feet without nephropathy (p value 0.016). Ulcer healing rate was also significantly low in the COM group, with healing noted in only 9 in this group (9/24), in contrast to 5 out 6 in those without COM (p value P. 0.044).

Re-ulceration occurred in the same location in four patients and one patient developed new ulceration in another part of the same foot within 12 months of healing. Therefore, foot ulcer recurrence rate at one year among those that achieved full healing, was 36% (5/14).

We have analysed the functional outcomes of 27 patients who had a minimum of 12 months follow-up. Six patients (22%) were able to return to independent walking in either normal or custom-made surgical footwear. Eleven patients (41%) returned to their life with assisted walking using either crutches or a frame, in custom-shoes or off-loading cast. Ten patients (37%) remained wheelchair bound for various reasons.

Nine infected heel ulcers with osteomyelitis of the calcaneal tuberosity had ALBS supplementation (8 cases of Cerament and 1 case of Stimulan) at the time of ulcer debridement. All patients had an appropriate duration of deep tissue culture specific parenteral antibiotic therapy post-operatively. However, despite such an aggressive infection eradication approach, no statistically significant difference in ulcer healing was noted with ALBS; 3 out of 9 in this group achieved full healing, whereas 11 out of 21 healed in those without ALBS (p value 0.34).

4. Discussion

The prospect of achieving a successful outcome in the management of heel ulcers in patients with diabetes is usually met with scepticism from care providers. From initial off-loading, to the point of the patient going into custom footwear following ulcer healing, is usually a protracted course.²¹,²² Multifaceted challenges ranging from ulcer location, tissue loss, presence of infection, diabetic control, and patient compliance demand thoughtful and considered steering from MDFT, which not every hospital may have access to. Once the infection has seeded to the deep tissue and the calcaneum, the need for surgical debridement becomes inevitable. Often, multiple debridements are necessary to achieve adequate infection control.

A number of studies have described various surgical techniques to achieve limb salvage in patients with calcaneal osteomyelitis¹⁵,¹⁶,¹⁷,¹⁸,²³,²⁴,.²⁵ However, there is no consensus on which technique is superior. Historically, Crandall et al. performed 20 partial and 11 total calcanectomies in 29 patients including 18 with diabetes.¹⁵ They reported a success rate of only 35% in their series and concluded that primary wound healing occurred in only four of eighteen patients with diabetes. In contrast, Cook et al. reported a 71.4% healing rate during the first year, with an average time to full closure of 201 days in their study of 50 partial calcanectomies in 46 patients. They noted a wide variation in the number of days required for complete wound healing despite primary wound closure being attempted in most of their cases.¹⁶ They also found that there was no statistically significant difference in healing time between patients who required vascular intervention before surgery and those who did not. This information lends support to us limiting our study to only those patients operated on by the orthopaedic surgeons.

Despite promising results emerging from centres in recent years¹⁷,¹⁸,²⁵ the long held view of poor outcome of diabetic heel ulcers remains. The synergy of poor wound healing potential in patients with diabetes, loss of protective sensation due to peripheral neuropathy, diminished local circulation and impaired renal and cardiac function favours a raised peri-operative morbidity and mortality.²⁶ Our results showed that a heel ulcer in patients with diabetic neuropathy is still a complex challenge.

In our study, we observed an ulcer healing rate of only 50% taking almost a year to achieve. This contrasts with Paola et al. reporting 100% healing rate in a mean time of 69 days (±64) in their cohort of 18 patients treated with open partial calcanectomy and application of circular external fixator for hindfoot stabilization and off-loading.¹⁷ All of their patients underwent skin grafting. Their follow up duration was less than one year. Three out of 18 patients (17%) re-ulcerated during their follow-up period. The author advocated the use of a ring external fixator to keep the hindfoot in equinus and ensure off-loading. In our diabetic foot unit, previous experience with a ring external fixator was of poor patient tolerance and a higher incidence of pin-site infections and wire loosening. Accordingly, we prefer to achieve off-loading, in this group of patients, with total contact casting and stringent non-weightbearing. Although seven patients in our cohort underwent split skin grafting, only one achieved full healing. Furthermore, we observed that the sustained use of post-operative NPWT helped achieve good granulation tissue coverage of exposed calcaneal bone and gradual skin epithelialisation, thereby rendering skin grafting not essential in some patients. However, this healing process reached a static state, short of full healing, in some patients.

Similarly, our findings differ to the results published by Dramploas et al.¹⁸ They reported 100% infection eradication and healing rate in their series of 12 patients, with a mean duration of 16 weeks follow-up after partial calcanectomy and application of injectable ALBS using the silo technique. 50% of their patients had primary wound closure while the rest had a vacuum assisted closure dressing. This study focused on presenting their surgical technique - the Silo technique - and there was less information on the medium-term outcomes, as their follow up duration was only 16 weeks. We feel that such short follow-up may not have captured any data on ulcer recurrence or infection flare up. In contrast, only three out of nine patients who had ALBS in our cohort went on to achieve complete heeling. While we appreciate that the number of feet (9 out of 30) that were managed with ALBS following osteotomy was not large, our findings did not support that faster wound healing with this method. However, we acknowledge that this cohort of patients presented with the most severe form of COM, which is why we opted to use ALBS.

Recently, Qin et al. reported 81% infection remission with no recurrence in their cohort of 26 patient with COM treated with extensive debridement and ALBS.²⁵ Their 18% recurrence rate at a mean follow-up of 36 months indicated that achieving radical cure is not an easy task in such patients. The weakness of their study was not including patients with significant comorbidities, especially with diabetes. It is possible that their healing rates would be inferior had they included patients with complex medical issues.

We have given thought as to why only 48% of our cohort achieved full ulcer healing. Perhaps blood sugar control and the level of HbA1c plays a significant role in reduced wound healing rates in our cohort. Mean HbA1c level in our diabetic patient population was greater than 9% and this is consistent with the findings of Christman et al.,²⁷ where HbA1c of more than 8% was associated with longest duration for wound healing. By the same token, renal impairment has long been believed to affect wound healing. Progressive loss of kidney function leads to an elevated concentration of urea in the circulation, uraemia, which has a negative effect on would healing. Earlier research has reported on the adverse effects of uraemia on important wound repair elements such as fibroblast proliferation, hydroxyproline level and collagen production²⁸,²⁹,.³⁰ Our data supported the fact that wound healing rate is poor in patients with significant diabetic nephropathy (p value 0.016). Additionally, patients’ compliance, dietary patterns and nutrient intake fundamentally contribute to tissue growth and wound healing.

COM is known to negatively affect the healing of infected diabetic heel ulcers and this was observed in our study too. Only 9 ulcers with COM healed (9/24), in contrast to 5 out of 6 without (p value 0.044). Concomitant calcaneal fractures on a background of osteomyelitis and heel ulcer complicates the situation greatly, and the chances of the ulcer healing become remote. Three patients in our group had such fractures and only one reached full ulcer healing, which took 152 days.

Re-ulceration in our patients most commonly occurred due to mechanical reasons such as the presence of a residual bone spike or cortical ridge from the remaining calcaneal tuberosity (4/5) requiring revision debridement (Fig. 3). In one patient it was due to friction occurring between footwear and a slightly mobile flap over the heel area. But ulcer recurrence responded well to a period of offloading. We noted that in this patient no bone-tissue adherence occurred under the fasciocutaneous flap, leaving the area to form a bursa which facilitated gliding of the overlying flap. This movement of the tissue against footwear can result in re-ulceration over a period. However, no further surgery was required on this patient during the follow-up period.

Fig. 3.

Ulcer recurrence due to residual bone spike in calcaneum (A,B), MRI showing high signal activity (C). Radiograph of calcaneum after revision debridement and application of antibiotic loaded bone cement (D), complete wound healing (E,F) and leucocytes labelled scan showing no evidence of residual infection (G).

An important observation is that, with an overall 38% mortality (11/29), and with 2 dying within one-year, our patient population had significant co-existing diabetic related medical illnesses, which could have potentially hindered their ulcer healing and negatively impacted on their long-term survival.

Our study does have some limitations in that it is a retrospective study with a relatively small number of subjects. We did not include heel ulcer patients with critical limb ischemia requiring emergency vascular interventions. Perhaps, incorporating such patients would have increased our sample size and provided generalisation. However, we feel that exclusion of those with critical limb ischemia allowed us to design this study with a focus on one specific aspect of patient presentation; i.e. with an infected diabetic neuropathic heel ulcer.

5. Conclusion

Surgical management of infected diabetic neuropathic heel ulcers, without critical limb ischemia, in a structured MDFT setup, allows an excellent limb salvage rate but is associated with unpredictable and slow ulcer healing. Our results reveal a greater chance of wound healing in the absence of calcaneal osteomyelitis and nephropathy. Significant medical comorbidities in these patients contribute to a high mortality rate in the medium term and we recommend the treating physicians consider this aspect of survivorship when counselling patients for a complex treatment regimen. Larger multicentre studies are warranted in order to identify optimal management of this group of patients with complex presentations.

Author’s contributions

1: Conception and design- Phyo, Kavarthapu.

2: Administrative support- Phyo, Tang.

3: Provision of study materials or patients- Phyo, Kavarthpau

4: Collection and assembly of data- Phyo, Tang.

5: Data analysis and interpretation: All authors.

6: Manuscripts writing: All authors.

7: Final approval of manuscript: All authors.

Funding

This research did not receive any specific grant from funding agencies, the public, commercial, or not-for-profit sectors.

Originality of figures and tables

None of the Tables or Figures have previously been published.

Informed consent

Institutional ethics committee approval was not required as this study was part of our ongoing audit work. Informed consent was obtained for the prospective data collection used in this study.

Ethical statement

The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Declaration of competing interest

None.