- Email: [email protected]
Controversies regarding radiological changes and variables predicting amputation in a surgical series of diabetic foot osteomyelitis
Foot and Ankle Surgery 18 (2012) 233–236
Contents lists available at SciVerse ScienceDirect
Foot and Ankle Surgery journal homepage: www.elsevier.com/locate/fas
Controversies regarding radiological changes and variables predicting amputation in a surgical series of diabetic foot osteomyelitis Javier Arago´n-Sa´nchez MD, PhDa,*, Jose L. La´zaro-Martı´nez DPM, PhDb, Nalini Campillo-Vilorio MDc, Yurena Quintana-Marrero RNa, Maria J. Herna´ndez-Herrero MDd a
Surgery Department, Diabetic Foot Unit, La Paloma Hospital, Las Palmas de Gran Canaria, Spain Diabetic Foot Unit, Complutense University Podiatric Clinic, Madrid, Spain Diabetic Foot Unit, Diabetology Department, Plaza de la Salud General Hospital, Dominican Republic d Diabetic Foot Unit, Physical Medicine & Rehabilitation, La Paloma Hospital, Las Palmas de Gran Canaria, Spain b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 August 2011 Received in revised form 24 November 2011 Accepted 5 January 2012
Background: To investigate if radiological changes have any influence on the outcomes of surgical treatment of diabetic foot osteomyelitis. Methods: Data of patients included in a prospective cohort who underwent surgical treatment for definitive osteomyelitis were analyzed. Cases were classified according to radiological changes as ‘‘early osteomyelitis’’ when no radiological changes were found or in cases showing periosteal elevation and/or subcortical demineralization and/or cortical disruption. Cases showing sequestra and/or gross bone destruction were classified as ‘‘advanced osteomyelitis’’. Results: Early osteomyelitis was defined according to radiological findings in 37 cases (45.7%) and advanced in 44 (54.3%). Advanced osteomyelitis was not associated with the risk of undergoing amputation. Conclusions: The bone changes seen in simple X-rays in cases of osteomyelitis do not have any prognostic value when surgical treatment is undertaken. The outcomes are more related to soft tissue involvement than bone destruction seen in simple X-rays. ß 2012 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Diabetic foot Diabetic foot infections Diabetic foot osteomyelitis Bone infection Foot ulcer Radiology Diabetic foot surgery
1. Introduction The optimum management of patients with diabetes and foot osteomyelitis is not currently well defined [1]. There is agreement that plain radiology should be performed when bone infection is suspected [2,3]. Authors have reported that the presence of a positive X-ray is the single most important factor influencing surgical consultation and hospitalization [4]. However, radiological changes in cases of osteomyelitis may take several weeks to appear accounting in part for the low sensitivity of plain radiography [1,5]. In this way, early osteomyelitis may be present even when plain X-rays show no changes or just cortical involvement. The undetected infection might then spread via the Haversian system, leading to the involvement of the medullary bone and marrow where the infection may spread rapidly [2]. Thus, advanced cases will show severe bone destruction, which
* Corresponding author at: C/Eduardo 1,48D, 35002 Las Palmas de Gran Canaria, Canary Islands, Spain. Tel.: +34 928383161; fax: +34 928242840; mobile: +34 609569937. E-mail addresses: [email protected], [email protected] (J. Arago´n-Sa´nchez).
will only then be obvious on the plain X-ray. High performance of plain radiography has been reported in advanced cases and a diagnosis approach based only on bone exposed on testing with a probe and plain radiography of the foot in such cases have been suggested [6]. Despite the fact that plain radiography is extensively used for diagnosing osteomyelitis in the feet of patients with diabetes, we have not found previous reports analyzing the correlation between clinical presentation and the outcomes of surgical treatment according to radiological changes. The aims of this study were to analyze the clinical factors that determine the radiological findings in a prospective series of surgically treated diabetic foot osteomyelitis and to investigate if radiological changes have any influence on the outcomes of surgical treatment. 2. Methods From 1 November 2007 to 30 May 2010, we enrolled a prospective cohort of patients with diabetes who were admitted to our Diabetic Foot Unit. Inclusion criterion was the presence of foot infections [7] and no exclusion criteria were established. In the present study, data of patients included in the cohort who underwent surgical treatment for definitive osteomyelitis [2] were
1268-7731/$ – see front matter ß 2012 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.fas.2012.01.005
234
J. Arago´n-Sa´nchez et al. / Foot and Ankle Surgery 18 (2012) 233–236
analyzed. Our flow chart based on sequential combination of the bone exposed on testing with a probe and plain X-rays for diagnosing and managing diabetic foot osteomyelitis has been published elsewhere [6]. Laboratory values were determined at admission. Perioperative glycemic control was undertaken by premeal bedside glucose monitoring (three times a day) using capillary blood. Plain X-rays (two standard views) of the involved foot were obtained for each patient included in this series and the interpretation of the images was always performed by the same experienced surgeon (J A-S). Cases were classified according to radiological changes as ‘‘early osteomyelitis’’ when no radiological changes were found or in cases showing periosteal elevation and/ or subcortical demineralization and/or cortical disruption. Cases showing sequestra and/or gross bone destruction were classified as ‘‘advanced osteomyelitis’’. The location of the ulcer was classified as plantar or non-plantar (dorsal, lateral, on the tip of the toe or interdigital). Peripheral arterial disease was diagnosed if the patient met the following criteria: absence of both distal pulses and/or ankle brachial index below 0.9. Soft tissue infections, including necrotizing soft tissue infections (NSTIs), were diagnosed either preoperatively or during surgical procedures according to earlier published criteria [8,9]. During surgical intervention, bone samples were extracted for analysis by the microbiology and pathology laboratories. Only aerobic cultures were grown in this study. Cases with persistent postoperative infection either in bone or soft tissue that precluded wound healing and/or produced clinical infectious symptoms were re-operated. Results of the bone cultures were categorized into four types: no growth, exclusively gram-positive, exclusively gram-negative and mixture of gram-positive and negative in the same culture. Histopathology of bone samples was categorized into two types according to the presence of acute inflammatory changes. The first type included acute osteomyelitis and acute exacerbation of chronic osteomyelitis and the second type included chronic osteomyelitis and fibrosis [10]. 2.1. Statistical methods The median and interquartile ranges were used for nonnormally distributed variables. Mean values of premeal bedside glucose monitoring were determined for each patient and the distribution was divided into quartiles and patients in quartile 1 were compared with those in quartiles 2–4 as previously reported [11]. Qualitative variables were analyzed by x2 test and where appropriate, the Fisher’s exact test. Non-normally distributed quantitative variables were analyzed using the non-parametric Mann–Whitney U-test. The patients were divided into two groups: patients with early osteomyelitis according to radiological changes and patients with advanced osteomyelitis. Univariate analysis was undertaken with a level of statistical significance set at p < 0.05 between patients with early and advanced osteomyelitis. Univariate analysis was also undertaken between patients who required any amputation and who did not require amputation. Later, the variables that reached statistical significance on both univariate analyzes were introduced in separate stepwise logistic regression models in order to identify independent predictors of those presenting early osteomyelitis at admission and independent predictors of those undergoing amputation. A p value < 0.05 was set as the threshold of statistical significance. We calculated statistical values using G-Stat 2.0.1 (GlaxoSmithKline, S.A. Tres Cantos, Madrid, Spain). The Committee for Ethical Issues of Las Palmas Medical College reviewed the protocol of this study, gave their approval, and concluded that this observational, non-experimental and noninterventional study, which did not alter or exceed the scope of our
standard medical care, did not require subsequent approval by a central ethics committee for clinical research. 3. Results Eighty-one patients were included in this study. The median age of the patients was 65 years (IQR 16) with a median duration of diabetes of 20 years (IQR 15). Insulin treatment was given to 41 patients (50.6%). The median glycated hemoglobin was 0.082 (IQR 0.029) and 65 patients (80.2%) had glycated hemoglobin values 0.07. The median of glucose was 161.1 mg/dl with a range of 102–274 mg/dl (IQR 35). Twenty patients (24.7%) were included into quartile 1 and 61 (75.3%) into quartiles 2 to 4. Thirtyeight patients (46.9%) had had previous ulcerations and 25 (30.9%) had undergone previous amputations. The ulcer was located on the plantar aspect in 27 patients (33.3%) and the rest were located as follows: 27 lateral (33.3%), 13 dorsal (16%), nine interdigital (11.1%) and five on the tip of a toe (6.2%). Plantar lesions had a median duration of 65 days (IQR 189) and other located lesions a duration of 30 days (IQR 73) (p = 0.01). Charcot deformity was present in seven patients (8.9%) and peripheral arterial disease was diagnosed in 39 (48.1%). Suppuration was present in 56 patients (69.1%), fetid odor in 32 patients (39.5%), skin necrosis in 28 (34.6) and soft tissue infections in 30 (37%). Skin necrosis and/or soft tissue infection were found in 24 (61.5%) patients with and in 16 (38.1%) patients without peripheral arterial disease (p = 0.03). Bone exposed on testing with a probe was positive in 75 patients (94.9%). X-ray was taken in every case. Early osteomyelitis was defined according to radiological findings in 37 cases (45.7%) and advanced in 44 (54.3%). The duration of the wound was 30 days (IQR 55) in cases of early osteomyelitis and 47.5 (IQR 90) in cases of advanced osteomyelitis (p = 0.46). Antibiotics had previously been given before admission in 61 patients (75.3%) during a median period of 21 days (IQR 20). Twenty-two patients (59.2%) with early osteomyelitis and 39 (88.6%) with advanced had received antibiotics before admission (p = 0.002). In those 61 patients, antibiotics had been given during a median of 25 days (IQR 23) for patients with early osteomyelitis and 20.5 days (IQR 27) for patients with advanced osteomyelitis (p = 0.4). Histopathology showed acute inflammation in 23 cases (65.7%) of early osteomyelitis and in 33 (75%) of advanced osteomyelitis (p = 0.36). Bone cultures revealed no bacterial growth in seven cases of advanced osteomyelitis (53.8%), gram-positive bacteria in 15 cases (62.5%), gram-negative in 12 (52.2%) and a mixture of gram-positive and gram-negative bacteria in 10 (47.6%) (p = 0.78). Factors that showed significant differences (p < 0.05) between univariate analysis between early and advanced osteomyelitis included: plantar location of the wound, peripheral arterial disease, suppuration, antibiotic treatment before admission and skin necrosis. The logistic regression models are shown in Table 1. Conservative surgery was the initial procedure in 59 patients (72.8%) but 11 of them (18.6%) subsequently underwent amputation. Eight patients (21.6%) with early osteomyelitis and three (6.8%) with advanced osteomyelitis presented failure of conservative surgery and required subsequent amputation (p = 0.05). The definitive outcomes of the series were as follows: 48 patients (59.3%) underwent conservative surgery, 32 (39.5%) minor amputations and one (1.2%) major amputation. Nineteen patients (51.4%) with early and 14 patients (31.8%) with advanced osteomyelitis underwent amputation (p = 0.07). Univariate analysis of the patients regarding the need to perform any amputation is shown in Table 2 and the results of the logistic regression model are shown in Table 3. Postoperative antibiotics were given over a median period of 39 days (IQR 26) to patients with early osteomyelitis and 29 days (IQR
J. Arago´n-Sa´nchez et al. / Foot and Ankle Surgery 18 (2012) 233–236 Table 1 Results of the logistic regression model of variables associated with early osteomyelitis. Variables
p-Value
Odds ratio
95% Confidence interval
Necrosis Ischemia Plantar location Pus draining
0.004 0.001 0.001 0.003
9.5 13.8 19.6 0.97
2.0–44.2 2.9–65.5 3.4–113.8 0.021–0.45
29) to patients with advanced osteomyelitis (p = 0.139). Healing was achieved in 13 weeks (IQR 17) for patients with early osteomyelitis and 6 weeks (IQR 7) for patients with advanced osteomyelitis (p = 0.007). 4. Discussion Our results are certainly paradoxical. Patients with advanced osteomyelitis, according to radiological changes, were not associated with a failure of conservative surgery, a definitive need for amputation or a long duration of antibiotic treatment in this series. Moreover, patients with early osteomyelitis required a significantly longer period to heal. Our results are discordant with those from another group of authors, which reported that patients with incipient bone involvement, diagnosed by combined Tc 99bone and leukocyte scanning, presented a better response to conservative therapy than when radiological signs were found [12]. This has not been found in our experience since conservative surgery has been successfully performed regardless of early or advanced osteomyelitis diagnosis. We believe that our findings are related to the fact that ischemia and skin necrosis were significantly associated with early osteomyelitis. There is agreement that bacteria gain access to bone by contiguous spread, entering from overlying soft tissue and penetrating the cortex before involving the marrow [1]. We hypothesize that in many cases of early osteomyelitis, the infection was just about to reach the bone before surgical intervention was performed as a consequence of evident soft tissue destruction. We believe that this is one of the reasons we had such a high rate of osteomyelitis in our previous series of diabetic foot infections [6,10,13]. Patients with skin necrosis and/or soft tissue infection are not always
235
diagnosed with osteomyelitis in other settings because severe soft tissue destruction is more evident. However, we systematically look for bone involvement despite the appearance of soft tissue destruction. Consistent with our results, other authors have also reported no differences in the outcomes of medical treatment of diabetic foot osteomyelitis regarding the results of plain radiographs [14]. Even though these results agree with ours, this series cannot be compared because patients were exclusively treated with antibiotics and patients with necrosis and ischemia were excluded [14]. Plantar location of the lesion was associated with presenting early radiological changes. We have not been able to find similar findings in the literature and we do not have any explanation for it. This is also paradoxical because plantar lesions have a longer duration and should have undergone higher pressures than lesions located at other sites, although this has not directly been assessed in this study. Theoretically, bone destruction should have been more evident in such cases but the relationship between plantar pressures and radiological destruction in cases of bone infections has not previously been studied. Suppuration through the ulcer was inversely associated with early osteomyelitis. In this way, pus draining through the ulcer was associated with more extensive bone destruction. Bone destruction involves complex mechanisms not totally well understood, which involve as many bacterial bone-modulatory molecules [15] as a number of host cytokines that play a significant role in the pathogenesis of osteomyelitis [16]. Molecular and cellular pathogenesis of bone infection in the feet of patients with diabetes have not been studied. We did not find any relationship between histopathological changes and advanced bone destruction. In our experience, advanced osteomyelitis according to radiological findings has not been associated with the need for amputation. Variables related to the need for amputation in the logistic regression model were: previous ulceration, peripheral arterial disease, soft tissue infection, skin necrosis and perioperative glucose values into quartiles 2–4. Some of these variables have been previously reported by our group as associated with the failure of conservative surgery in cases of diabetic foot osteomyelitis [10]. The association between hyperglycemia and increased risk of complications and mortality in patients with diabetes
Table 2 Univariate analysis of patients undergoing conservative surgery with those who had any amputation.
Glucose at admission, median (IQR) Capillary glucose monitoring into quartiles 2–4, n (%) ESR, median (IQR) WBC count, median (IQR) White blood cells count > 11,000 cells/mm3, n (%) Platelets count, median (IQR) Insulin treatment, n (%) Glycated hemoglobin > 7%, n (%) Previous amputation, n (%) Previous ulceration, n (%) Ulcer duration in days, median (IQR) Plantar location of the wound, n (%) Charcot deformity, n (%) Peripheral arterial disease, n (%) Fetid smelling Suppuration Bone exposure at the bottom of the wound Previous surgery in another hospital, n (%) Antibiotic treatment before admission, n (%) MRSA Severe infection according to IDSA criteria, n (%) Advanced osteomyelitis according to X-ray, n (%) Histopathology showing acute inflammatory, n (%) findings Soft tissue infection accompanying osteomyelitis, n (%) Skin necrosis, n (%)
No amputation (n = 48)
Amputation (n = 33)
p-Value
170.5 (150) 32 (66.7%) 52 (55) 8815 (3483) 11 (22.9) 264,500 (133,000) 26 (54.2) 38 (79.2) 15 (31.3) 27 (56.3) 42.5 (143.9) 21 (43.8) 5 (10.4) 17 (35.4) 15 (31.3) 35 (72.9) 12 (25) 12 (25) 39 (81.3) 5 (10.4) 2 (4.2) 30 (62.5) 32 (68.1) 10 (20.8) 6 (12.5)
201 (141) 29 (87.9%) 75 (54) 9870 (3695) 12 (36.4) 313,000 (144,000) 15 (45.5) 27 (81.8%) 10 (30.3) 11 (33.3) 56.5 (131) 6 (18.2) 2 (6.1) 22 (66.7) 17 (51.5) 21 (63.6) 15 (45.5) 5 (15.2) 21 (63.6) 3 (9.1) 5 (15.2%) 14 (42.4) 24 (75) 20 (60.6) 22 (66.7)
0.04 0.03 0.03 0.01 0.18 0.01 0.44 0.76 0.92 0.04 0.9 0.01 0.49 0.006 0.06 0.3 0.05 0.28 0.07 0.8 0.08 0.07 0.5 <0.001 <0.001
J. Arago´n-Sa´nchez et al. / Foot and Ankle Surgery 18 (2012) 233–236
236
Table 3 Results of the logistic regression model of variables associated with undergoing amputation. Variables
p-Value
Odds ratio
95% Confidence interval
Previous ulceration Soft tissue infection accompanying osteomyelitis Capillary glucose monitoring into quartiles 2 to 4 Peripheral arterial disease Skin necrosis
0.03 0.02 0.04 0.009 0.001
0.23 4.9 5.9 6.2 12.2
0.06–0.9 1.2–20.0 1.0–33.5 1.6–24.7 2.9–50.9
undergoing surgery have previously been reported [17,18]. Authors using the mean of three HbA1C levels measured in the year preceding admission and in the year following discharge found that patients who underwent amputation had significant higher values [19]. However, HbA1C values in our series were not significantly different regarding the need for amputation. Other authors reported no differences on the outcomes of conservative treatment of diabetic foot infections according to HbA1C values [20]. We have now found that previous ulcerations are inversely related to the need for amputation. Other authors have found quite the opposite: a significantly higher percentage of patients who underwent amputation had previously had foot ulcerations (p = 0.001) [21]. A history of hospitalization for diabetic infection independently predicted failure of conservative treatment in another report [20]. The strengths of the study are: its prospective nature; surgery was performed by the same experienced surgeon that limited the possibility of bias; and histopathological and microbiological studies were routinely performed. The weaknesses of the study are: we were unable to know the exact duration of the infection since many patients had previously been treated; the division into two classes according to radiology was arbitrarily established because this has not previously been done; and these results could not be extrapolated when treating osteomyelitis without surgery. In conclusion, the bone changes seen in simple X-rays in cases of osteomyelitis do not have any prognostic value when surgical treatment is undertaken. The outcomes of surgical treatment of patients with diabetes and osteomyelitis are more related to soft tissue involvement than bone destruction seen in X-rays. For this reason, many patients with only early radiological changes must undergo amputations due to skin necrosis and/or soft tissue infections accompanying osteomyelitis. Conflict of interest statement Nothing to declare. References [1] Jeffcoate WJ, Lipsky BA. Controversies in diagnosing and managing osteomyelitis of the foot in diabetes. Clin Infect Dis 2004;39(Suppl. 2):S115–22. [2] Berendt AR, Peters EJ, Bakker K, Embil JM, Eneroth M, Hinchliffe RJ, et al. Diabetic foot osteomyelitis: a progress report on diagnosis and a systematic review of treatment. Diabetes Metab Res Rev 2008;24(Suppl. 1):S145–61. [3] Hartemann-Heurtier A, Senneville E. Diabetic foot osteomyelitis. Diabetes Metab 2008;34:87–95.
[4] Edelman D, Matchar DB, Oddone EZ. Clinical and radiographic findings that lead to intervention in diabetic patients with foot ulcers. A nationwide survey of primary care physicians. Diabetes Care 1996;19:755–7. [5] Becker W. Imaging osteomyelitis and the diabetic foot. Q J Nucl Med 1999;43:9–20. [6] Aragon-Sanchez J, Lipsky BA, Lazaro-Martinez JL. Diagnosing diabetic foot osteomyelitis: is the combination of probe-to-bone test and plain radiography sufficient for high-risk inpatients? Diabet Med 2011;28:191–4. [7] Lipsky BA. A report from the international consensus on diagnosing and treating the infected diabetic foot. Diabetes Metab Res Rev 2004;20(Suppl 1):S68–77. [8] Aragon-Sanchez J, Quintana-Marrero Y, Lazaro-Martinez JL, Hernandez-Herrero MJ, Garcia-Morales E, Beneit-Montesinos JV, et al. Necrotizing soft-tissue infections in the feet of patients with diabetes: outcome of surgical treatment and factors associated with limb loss and mortality. Int J Low Extrem Wounds 2009;8:141–6. [9] Eneroth M, Larsson J, Apelqvist J. Deep foot infections in patients with diabetes and foot ulcer: an entity with different characteristics, treatments, and prognosis. J Diabetes Complicat 1999;13:254–63. [10] Aragon-Sanchez FJ, Cabrera-Galvan JJ, Quintana-Marrero Y, Hernandez-Herrero MJ, Lazaro-Martinez JL, Garcia-Morales E, et al. Outcomes of surgical treatment of diabetic foot osteomyelitis: a series of 185 patients with histopathological confirmation of bone involvement. Diabetologia 2008;51:1962– 70. [11] Golden SH, Peart-Vigilance C, Kao WH, Brancati FL. Perioperative glycemic control and the risk of infectious complications in a cohort of adults with diabetes. Diabetes Care 1999;22:1408–14. [12] Balsells M, Viade J, Millan M, Garcia JR, Garcia-Pascual L, del Pozo C, et al. Prevalence of osteomyelitis in non-healing diabetic foot ulcers: usefulness of radiologic and scintigraphic findings. Diabetes Res Clin Pract 1997;38: 123–7. [13] Aragon-Sanchez J. Treatment of diabetic foot osteomyelitis: a surgical critique. Int J Low Extrem Wounds 2010;9:37–59. [14] Senneville E, Lombart A, Beltrand E, Valette M, Legout L, Cazaubiel M, et al. Outcome of diabetic foot osteomyelitis treated nonsurgically: a retrospective cohort study. Diabetes Care 2008;31:637–42. [15] Nair SP, Meghji S, Wilson M, Reddi K, White P, Henderson B. Bacterially induced bone destruction: mechanisms and misconceptions. Infect Immun 1996;64:2371–80. [16] Wright JA, Nair SP. Interaction of staphylococci with bone. Int J Med Microbiol 2010;300:193–204. [17] Clement S, Braithwaite SS, Magee MF, Ahmann A, Smith EP, Schafer RG, et al. Management of diabetes and hyperglycemia in hospitals. Diabetes Care 2004;27:553–91. [18] Pomposelli JJ, Baxter 3rd JK, Babineau TJ, Pomfret EA, Driscoll DF, Forse RA, et al. Early postoperative glucose control predicts nosocomial infection rate in diabetic patients. J Parenter Enteral Nutr 1998;22:77–81. [19] Lepore G, Maglio ML, Cuni C, Dodesini AR, Nosari I, Minetti B, et al. Poor glucose control in the year before admission as a powerful predictor of amputation in hospitalized patients with diabetic foot ulceration. Diabetes Care 2006;29:1985. [20] Pittet D, Wyssa B, Herter-Clavel C, Kursteiner K, Vaucher J, Lew PD. Outcome of diabetic foot infections treated conservatively: a retrospective cohort study with long-term follow-up. Arch Intern Med 1999;159:851–6. [21] Barberan J, Granizo JJ, Aguilar L, Alguacil R, Sainz F, Menendez MA, et al. Predictive model of short-term amputation during hospitalization of patients due to acute diabetic foot infections. Enferm Infecc Microbiol Clin 2010;28:680–4.
Contents lists available at SciVerse ScienceDirect
Foot and Ankle Surgery journal homepage: www.elsevier.com/locate/fas
Controversies regarding radiological changes and variables predicting amputation in a surgical series of diabetic foot osteomyelitis Javier Arago´n-Sa´nchez MD, PhDa,*, Jose L. La´zaro-Martı´nez DPM, PhDb, Nalini Campillo-Vilorio MDc, Yurena Quintana-Marrero RNa, Maria J. Herna´ndez-Herrero MDd a
Surgery Department, Diabetic Foot Unit, La Paloma Hospital, Las Palmas de Gran Canaria, Spain Diabetic Foot Unit, Complutense University Podiatric Clinic, Madrid, Spain Diabetic Foot Unit, Diabetology Department, Plaza de la Salud General Hospital, Dominican Republic d Diabetic Foot Unit, Physical Medicine & Rehabilitation, La Paloma Hospital, Las Palmas de Gran Canaria, Spain b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 August 2011 Received in revised form 24 November 2011 Accepted 5 January 2012
Background: To investigate if radiological changes have any influence on the outcomes of surgical treatment of diabetic foot osteomyelitis. Methods: Data of patients included in a prospective cohort who underwent surgical treatment for definitive osteomyelitis were analyzed. Cases were classified according to radiological changes as ‘‘early osteomyelitis’’ when no radiological changes were found or in cases showing periosteal elevation and/or subcortical demineralization and/or cortical disruption. Cases showing sequestra and/or gross bone destruction were classified as ‘‘advanced osteomyelitis’’. Results: Early osteomyelitis was defined according to radiological findings in 37 cases (45.7%) and advanced in 44 (54.3%). Advanced osteomyelitis was not associated with the risk of undergoing amputation. Conclusions: The bone changes seen in simple X-rays in cases of osteomyelitis do not have any prognostic value when surgical treatment is undertaken. The outcomes are more related to soft tissue involvement than bone destruction seen in simple X-rays. ß 2012 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Diabetic foot Diabetic foot infections Diabetic foot osteomyelitis Bone infection Foot ulcer Radiology Diabetic foot surgery
1. Introduction The optimum management of patients with diabetes and foot osteomyelitis is not currently well defined [1]. There is agreement that plain radiology should be performed when bone infection is suspected [2,3]. Authors have reported that the presence of a positive X-ray is the single most important factor influencing surgical consultation and hospitalization [4]. However, radiological changes in cases of osteomyelitis may take several weeks to appear accounting in part for the low sensitivity of plain radiography [1,5]. In this way, early osteomyelitis may be present even when plain X-rays show no changes or just cortical involvement. The undetected infection might then spread via the Haversian system, leading to the involvement of the medullary bone and marrow where the infection may spread rapidly [2]. Thus, advanced cases will show severe bone destruction, which
* Corresponding author at: C/Eduardo 1,48D, 35002 Las Palmas de Gran Canaria, Canary Islands, Spain. Tel.: +34 928383161; fax: +34 928242840; mobile: +34 609569937. E-mail addresses: [email protected], [email protected] (J. Arago´n-Sa´nchez).
will only then be obvious on the plain X-ray. High performance of plain radiography has been reported in advanced cases and a diagnosis approach based only on bone exposed on testing with a probe and plain radiography of the foot in such cases have been suggested [6]. Despite the fact that plain radiography is extensively used for diagnosing osteomyelitis in the feet of patients with diabetes, we have not found previous reports analyzing the correlation between clinical presentation and the outcomes of surgical treatment according to radiological changes. The aims of this study were to analyze the clinical factors that determine the radiological findings in a prospective series of surgically treated diabetic foot osteomyelitis and to investigate if radiological changes have any influence on the outcomes of surgical treatment. 2. Methods From 1 November 2007 to 30 May 2010, we enrolled a prospective cohort of patients with diabetes who were admitted to our Diabetic Foot Unit. Inclusion criterion was the presence of foot infections [7] and no exclusion criteria were established. In the present study, data of patients included in the cohort who underwent surgical treatment for definitive osteomyelitis [2] were
1268-7731/$ – see front matter ß 2012 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.fas.2012.01.005
234
J. Arago´n-Sa´nchez et al. / Foot and Ankle Surgery 18 (2012) 233–236
analyzed. Our flow chart based on sequential combination of the bone exposed on testing with a probe and plain X-rays for diagnosing and managing diabetic foot osteomyelitis has been published elsewhere [6]. Laboratory values were determined at admission. Perioperative glycemic control was undertaken by premeal bedside glucose monitoring (three times a day) using capillary blood. Plain X-rays (two standard views) of the involved foot were obtained for each patient included in this series and the interpretation of the images was always performed by the same experienced surgeon (J A-S). Cases were classified according to radiological changes as ‘‘early osteomyelitis’’ when no radiological changes were found or in cases showing periosteal elevation and/ or subcortical demineralization and/or cortical disruption. Cases showing sequestra and/or gross bone destruction were classified as ‘‘advanced osteomyelitis’’. The location of the ulcer was classified as plantar or non-plantar (dorsal, lateral, on the tip of the toe or interdigital). Peripheral arterial disease was diagnosed if the patient met the following criteria: absence of both distal pulses and/or ankle brachial index below 0.9. Soft tissue infections, including necrotizing soft tissue infections (NSTIs), were diagnosed either preoperatively or during surgical procedures according to earlier published criteria [8,9]. During surgical intervention, bone samples were extracted for analysis by the microbiology and pathology laboratories. Only aerobic cultures were grown in this study. Cases with persistent postoperative infection either in bone or soft tissue that precluded wound healing and/or produced clinical infectious symptoms were re-operated. Results of the bone cultures were categorized into four types: no growth, exclusively gram-positive, exclusively gram-negative and mixture of gram-positive and negative in the same culture. Histopathology of bone samples was categorized into two types according to the presence of acute inflammatory changes. The first type included acute osteomyelitis and acute exacerbation of chronic osteomyelitis and the second type included chronic osteomyelitis and fibrosis [10]. 2.1. Statistical methods The median and interquartile ranges were used for nonnormally distributed variables. Mean values of premeal bedside glucose monitoring were determined for each patient and the distribution was divided into quartiles and patients in quartile 1 were compared with those in quartiles 2–4 as previously reported [11]. Qualitative variables were analyzed by x2 test and where appropriate, the Fisher’s exact test. Non-normally distributed quantitative variables were analyzed using the non-parametric Mann–Whitney U-test. The patients were divided into two groups: patients with early osteomyelitis according to radiological changes and patients with advanced osteomyelitis. Univariate analysis was undertaken with a level of statistical significance set at p < 0.05 between patients with early and advanced osteomyelitis. Univariate analysis was also undertaken between patients who required any amputation and who did not require amputation. Later, the variables that reached statistical significance on both univariate analyzes were introduced in separate stepwise logistic regression models in order to identify independent predictors of those presenting early osteomyelitis at admission and independent predictors of those undergoing amputation. A p value < 0.05 was set as the threshold of statistical significance. We calculated statistical values using G-Stat 2.0.1 (GlaxoSmithKline, S.A. Tres Cantos, Madrid, Spain). The Committee for Ethical Issues of Las Palmas Medical College reviewed the protocol of this study, gave their approval, and concluded that this observational, non-experimental and noninterventional study, which did not alter or exceed the scope of our
standard medical care, did not require subsequent approval by a central ethics committee for clinical research. 3. Results Eighty-one patients were included in this study. The median age of the patients was 65 years (IQR 16) with a median duration of diabetes of 20 years (IQR 15). Insulin treatment was given to 41 patients (50.6%). The median glycated hemoglobin was 0.082 (IQR 0.029) and 65 patients (80.2%) had glycated hemoglobin values 0.07. The median of glucose was 161.1 mg/dl with a range of 102–274 mg/dl (IQR 35). Twenty patients (24.7%) were included into quartile 1 and 61 (75.3%) into quartiles 2 to 4. Thirtyeight patients (46.9%) had had previous ulcerations and 25 (30.9%) had undergone previous amputations. The ulcer was located on the plantar aspect in 27 patients (33.3%) and the rest were located as follows: 27 lateral (33.3%), 13 dorsal (16%), nine interdigital (11.1%) and five on the tip of a toe (6.2%). Plantar lesions had a median duration of 65 days (IQR 189) and other located lesions a duration of 30 days (IQR 73) (p = 0.01). Charcot deformity was present in seven patients (8.9%) and peripheral arterial disease was diagnosed in 39 (48.1%). Suppuration was present in 56 patients (69.1%), fetid odor in 32 patients (39.5%), skin necrosis in 28 (34.6) and soft tissue infections in 30 (37%). Skin necrosis and/or soft tissue infection were found in 24 (61.5%) patients with and in 16 (38.1%) patients without peripheral arterial disease (p = 0.03). Bone exposed on testing with a probe was positive in 75 patients (94.9%). X-ray was taken in every case. Early osteomyelitis was defined according to radiological findings in 37 cases (45.7%) and advanced in 44 (54.3%). The duration of the wound was 30 days (IQR 55) in cases of early osteomyelitis and 47.5 (IQR 90) in cases of advanced osteomyelitis (p = 0.46). Antibiotics had previously been given before admission in 61 patients (75.3%) during a median period of 21 days (IQR 20). Twenty-two patients (59.2%) with early osteomyelitis and 39 (88.6%) with advanced had received antibiotics before admission (p = 0.002). In those 61 patients, antibiotics had been given during a median of 25 days (IQR 23) for patients with early osteomyelitis and 20.5 days (IQR 27) for patients with advanced osteomyelitis (p = 0.4). Histopathology showed acute inflammation in 23 cases (65.7%) of early osteomyelitis and in 33 (75%) of advanced osteomyelitis (p = 0.36). Bone cultures revealed no bacterial growth in seven cases of advanced osteomyelitis (53.8%), gram-positive bacteria in 15 cases (62.5%), gram-negative in 12 (52.2%) and a mixture of gram-positive and gram-negative bacteria in 10 (47.6%) (p = 0.78). Factors that showed significant differences (p < 0.05) between univariate analysis between early and advanced osteomyelitis included: plantar location of the wound, peripheral arterial disease, suppuration, antibiotic treatment before admission and skin necrosis. The logistic regression models are shown in Table 1. Conservative surgery was the initial procedure in 59 patients (72.8%) but 11 of them (18.6%) subsequently underwent amputation. Eight patients (21.6%) with early osteomyelitis and three (6.8%) with advanced osteomyelitis presented failure of conservative surgery and required subsequent amputation (p = 0.05). The definitive outcomes of the series were as follows: 48 patients (59.3%) underwent conservative surgery, 32 (39.5%) minor amputations and one (1.2%) major amputation. Nineteen patients (51.4%) with early and 14 patients (31.8%) with advanced osteomyelitis underwent amputation (p = 0.07). Univariate analysis of the patients regarding the need to perform any amputation is shown in Table 2 and the results of the logistic regression model are shown in Table 3. Postoperative antibiotics were given over a median period of 39 days (IQR 26) to patients with early osteomyelitis and 29 days (IQR
J. Arago´n-Sa´nchez et al. / Foot and Ankle Surgery 18 (2012) 233–236 Table 1 Results of the logistic regression model of variables associated with early osteomyelitis. Variables
p-Value
Odds ratio
95% Confidence interval
Necrosis Ischemia Plantar location Pus draining
0.004 0.001 0.001 0.003
9.5 13.8 19.6 0.97
2.0–44.2 2.9–65.5 3.4–113.8 0.021–0.45
29) to patients with advanced osteomyelitis (p = 0.139). Healing was achieved in 13 weeks (IQR 17) for patients with early osteomyelitis and 6 weeks (IQR 7) for patients with advanced osteomyelitis (p = 0.007). 4. Discussion Our results are certainly paradoxical. Patients with advanced osteomyelitis, according to radiological changes, were not associated with a failure of conservative surgery, a definitive need for amputation or a long duration of antibiotic treatment in this series. Moreover, patients with early osteomyelitis required a significantly longer period to heal. Our results are discordant with those from another group of authors, which reported that patients with incipient bone involvement, diagnosed by combined Tc 99bone and leukocyte scanning, presented a better response to conservative therapy than when radiological signs were found [12]. This has not been found in our experience since conservative surgery has been successfully performed regardless of early or advanced osteomyelitis diagnosis. We believe that our findings are related to the fact that ischemia and skin necrosis were significantly associated with early osteomyelitis. There is agreement that bacteria gain access to bone by contiguous spread, entering from overlying soft tissue and penetrating the cortex before involving the marrow [1]. We hypothesize that in many cases of early osteomyelitis, the infection was just about to reach the bone before surgical intervention was performed as a consequence of evident soft tissue destruction. We believe that this is one of the reasons we had such a high rate of osteomyelitis in our previous series of diabetic foot infections [6,10,13]. Patients with skin necrosis and/or soft tissue infection are not always
235
diagnosed with osteomyelitis in other settings because severe soft tissue destruction is more evident. However, we systematically look for bone involvement despite the appearance of soft tissue destruction. Consistent with our results, other authors have also reported no differences in the outcomes of medical treatment of diabetic foot osteomyelitis regarding the results of plain radiographs [14]. Even though these results agree with ours, this series cannot be compared because patients were exclusively treated with antibiotics and patients with necrosis and ischemia were excluded [14]. Plantar location of the lesion was associated with presenting early radiological changes. We have not been able to find similar findings in the literature and we do not have any explanation for it. This is also paradoxical because plantar lesions have a longer duration and should have undergone higher pressures than lesions located at other sites, although this has not directly been assessed in this study. Theoretically, bone destruction should have been more evident in such cases but the relationship between plantar pressures and radiological destruction in cases of bone infections has not previously been studied. Suppuration through the ulcer was inversely associated with early osteomyelitis. In this way, pus draining through the ulcer was associated with more extensive bone destruction. Bone destruction involves complex mechanisms not totally well understood, which involve as many bacterial bone-modulatory molecules [15] as a number of host cytokines that play a significant role in the pathogenesis of osteomyelitis [16]. Molecular and cellular pathogenesis of bone infection in the feet of patients with diabetes have not been studied. We did not find any relationship between histopathological changes and advanced bone destruction. In our experience, advanced osteomyelitis according to radiological findings has not been associated with the need for amputation. Variables related to the need for amputation in the logistic regression model were: previous ulceration, peripheral arterial disease, soft tissue infection, skin necrosis and perioperative glucose values into quartiles 2–4. Some of these variables have been previously reported by our group as associated with the failure of conservative surgery in cases of diabetic foot osteomyelitis [10]. The association between hyperglycemia and increased risk of complications and mortality in patients with diabetes
Table 2 Univariate analysis of patients undergoing conservative surgery with those who had any amputation.
Glucose at admission, median (IQR) Capillary glucose monitoring into quartiles 2–4, n (%) ESR, median (IQR) WBC count, median (IQR) White blood cells count > 11,000 cells/mm3, n (%) Platelets count, median (IQR) Insulin treatment, n (%) Glycated hemoglobin > 7%, n (%) Previous amputation, n (%) Previous ulceration, n (%) Ulcer duration in days, median (IQR) Plantar location of the wound, n (%) Charcot deformity, n (%) Peripheral arterial disease, n (%) Fetid smelling Suppuration Bone exposure at the bottom of the wound Previous surgery in another hospital, n (%) Antibiotic treatment before admission, n (%) MRSA Severe infection according to IDSA criteria, n (%) Advanced osteomyelitis according to X-ray, n (%) Histopathology showing acute inflammatory, n (%) findings Soft tissue infection accompanying osteomyelitis, n (%) Skin necrosis, n (%)
No amputation (n = 48)
Amputation (n = 33)
p-Value
170.5 (150) 32 (66.7%) 52 (55) 8815 (3483) 11 (22.9) 264,500 (133,000) 26 (54.2) 38 (79.2) 15 (31.3) 27 (56.3) 42.5 (143.9) 21 (43.8) 5 (10.4) 17 (35.4) 15 (31.3) 35 (72.9) 12 (25) 12 (25) 39 (81.3) 5 (10.4) 2 (4.2) 30 (62.5) 32 (68.1) 10 (20.8) 6 (12.5)
201 (141) 29 (87.9%) 75 (54) 9870 (3695) 12 (36.4) 313,000 (144,000) 15 (45.5) 27 (81.8%) 10 (30.3) 11 (33.3) 56.5 (131) 6 (18.2) 2 (6.1) 22 (66.7) 17 (51.5) 21 (63.6) 15 (45.5) 5 (15.2) 21 (63.6) 3 (9.1) 5 (15.2%) 14 (42.4) 24 (75) 20 (60.6) 22 (66.7)
0.04 0.03 0.03 0.01 0.18 0.01 0.44 0.76 0.92 0.04 0.9 0.01 0.49 0.006 0.06 0.3 0.05 0.28 0.07 0.8 0.08 0.07 0.5 <0.001 <0.001
J. Arago´n-Sa´nchez et al. / Foot and Ankle Surgery 18 (2012) 233–236
236
Table 3 Results of the logistic regression model of variables associated with undergoing amputation. Variables
p-Value
Odds ratio
95% Confidence interval
Previous ulceration Soft tissue infection accompanying osteomyelitis Capillary glucose monitoring into quartiles 2 to 4 Peripheral arterial disease Skin necrosis
0.03 0.02 0.04 0.009 0.001
0.23 4.9 5.9 6.2 12.2
0.06–0.9 1.2–20.0 1.0–33.5 1.6–24.7 2.9–50.9
undergoing surgery have previously been reported [17,18]. Authors using the mean of three HbA1C levels measured in the year preceding admission and in the year following discharge found that patients who underwent amputation had significant higher values [19]. However, HbA1C values in our series were not significantly different regarding the need for amputation. Other authors reported no differences on the outcomes of conservative treatment of diabetic foot infections according to HbA1C values [20]. We have now found that previous ulcerations are inversely related to the need for amputation. Other authors have found quite the opposite: a significantly higher percentage of patients who underwent amputation had previously had foot ulcerations (p = 0.001) [21]. A history of hospitalization for diabetic infection independently predicted failure of conservative treatment in another report [20]. The strengths of the study are: its prospective nature; surgery was performed by the same experienced surgeon that limited the possibility of bias; and histopathological and microbiological studies were routinely performed. The weaknesses of the study are: we were unable to know the exact duration of the infection since many patients had previously been treated; the division into two classes according to radiology was arbitrarily established because this has not previously been done; and these results could not be extrapolated when treating osteomyelitis without surgery. In conclusion, the bone changes seen in simple X-rays in cases of osteomyelitis do not have any prognostic value when surgical treatment is undertaken. The outcomes of surgical treatment of patients with diabetes and osteomyelitis are more related to soft tissue involvement than bone destruction seen in X-rays. For this reason, many patients with only early radiological changes must undergo amputations due to skin necrosis and/or soft tissue infections accompanying osteomyelitis. Conflict of interest statement Nothing to declare. References [1] Jeffcoate WJ, Lipsky BA. Controversies in diagnosing and managing osteomyelitis of the foot in diabetes. Clin Infect Dis 2004;39(Suppl. 2):S115–22. [2] Berendt AR, Peters EJ, Bakker K, Embil JM, Eneroth M, Hinchliffe RJ, et al. Diabetic foot osteomyelitis: a progress report on diagnosis and a systematic review of treatment. Diabetes Metab Res Rev 2008;24(Suppl. 1):S145–61. [3] Hartemann-Heurtier A, Senneville E. Diabetic foot osteomyelitis. Diabetes Metab 2008;34:87–95.
[4] Edelman D, Matchar DB, Oddone EZ. Clinical and radiographic findings that lead to intervention in diabetic patients with foot ulcers. A nationwide survey of primary care physicians. Diabetes Care 1996;19:755–7. [5] Becker W. Imaging osteomyelitis and the diabetic foot. Q J Nucl Med 1999;43:9–20. [6] Aragon-Sanchez J, Lipsky BA, Lazaro-Martinez JL. Diagnosing diabetic foot osteomyelitis: is the combination of probe-to-bone test and plain radiography sufficient for high-risk inpatients? Diabet Med 2011;28:191–4. [7] Lipsky BA. A report from the international consensus on diagnosing and treating the infected diabetic foot. Diabetes Metab Res Rev 2004;20(Suppl 1):S68–77. [8] Aragon-Sanchez J, Quintana-Marrero Y, Lazaro-Martinez JL, Hernandez-Herrero MJ, Garcia-Morales E, Beneit-Montesinos JV, et al. Necrotizing soft-tissue infections in the feet of patients with diabetes: outcome of surgical treatment and factors associated with limb loss and mortality. Int J Low Extrem Wounds 2009;8:141–6. [9] Eneroth M, Larsson J, Apelqvist J. Deep foot infections in patients with diabetes and foot ulcer: an entity with different characteristics, treatments, and prognosis. J Diabetes Complicat 1999;13:254–63. [10] Aragon-Sanchez FJ, Cabrera-Galvan JJ, Quintana-Marrero Y, Hernandez-Herrero MJ, Lazaro-Martinez JL, Garcia-Morales E, et al. Outcomes of surgical treatment of diabetic foot osteomyelitis: a series of 185 patients with histopathological confirmation of bone involvement. Diabetologia 2008;51:1962– 70. [11] Golden SH, Peart-Vigilance C, Kao WH, Brancati FL. Perioperative glycemic control and the risk of infectious complications in a cohort of adults with diabetes. Diabetes Care 1999;22:1408–14. [12] Balsells M, Viade J, Millan M, Garcia JR, Garcia-Pascual L, del Pozo C, et al. Prevalence of osteomyelitis in non-healing diabetic foot ulcers: usefulness of radiologic and scintigraphic findings. Diabetes Res Clin Pract 1997;38: 123–7. [13] Aragon-Sanchez J. Treatment of diabetic foot osteomyelitis: a surgical critique. Int J Low Extrem Wounds 2010;9:37–59. [14] Senneville E, Lombart A, Beltrand E, Valette M, Legout L, Cazaubiel M, et al. Outcome of diabetic foot osteomyelitis treated nonsurgically: a retrospective cohort study. Diabetes Care 2008;31:637–42. [15] Nair SP, Meghji S, Wilson M, Reddi K, White P, Henderson B. Bacterially induced bone destruction: mechanisms and misconceptions. Infect Immun 1996;64:2371–80. [16] Wright JA, Nair SP. Interaction of staphylococci with bone. Int J Med Microbiol 2010;300:193–204. [17] Clement S, Braithwaite SS, Magee MF, Ahmann A, Smith EP, Schafer RG, et al. Management of diabetes and hyperglycemia in hospitals. Diabetes Care 2004;27:553–91. [18] Pomposelli JJ, Baxter 3rd JK, Babineau TJ, Pomfret EA, Driscoll DF, Forse RA, et al. Early postoperative glucose control predicts nosocomial infection rate in diabetic patients. J Parenter Enteral Nutr 1998;22:77–81. [19] Lepore G, Maglio ML, Cuni C, Dodesini AR, Nosari I, Minetti B, et al. Poor glucose control in the year before admission as a powerful predictor of amputation in hospitalized patients with diabetic foot ulceration. Diabetes Care 2006;29:1985. [20] Pittet D, Wyssa B, Herter-Clavel C, Kursteiner K, Vaucher J, Lew PD. Outcome of diabetic foot infections treated conservatively: a retrospective cohort study with long-term follow-up. Arch Intern Med 1999;159:851–6. [21] Barberan J, Granizo JJ, Aguilar L, Alguacil R, Sainz F, Menendez MA, et al. Predictive model of short-term amputation during hospitalization of patients due to acute diabetic foot infections. Enferm Infecc Microbiol Clin 2010;28:680–4.