Background. Needle puncture has been suggested as a method for identifying bacteria in the bones in patients with diabetes with osteomyelitis of the foot. However, no studies have compared needle puncture with concomitant transcutaneous bone biopsy, which is the current standard recommended in international guidelines.
Methods. We conducted a prospective study in 2 French diabetes foot clinics. Transcutaneous bone biopsy specimens, needle puncture specimens, and swab samples were collected on the same day for each patient.
Results. Overall, 31 patients were included in the study from July 2006 through February 2008. Twenty-one bone biopsy specimens (67.7%), 18 needle puncture specimens (58%), and 30 swab samples (96.7%) had positive culture results. Staphylococcus aureus was the most common type of bacteria that grew from bone samples, followed by Proteus mirabilis and Morganella morganii. The mean number of bacteria types per positive sample were 1.35, 1.32, and 2.51 for bone biopsy specimens, needle puncture specimens, and swab samples, respectively. Among the 20 patients with positive bone biopsy specimens (69%), 13 had positive needle puncture samples. Overall, the correlation between microbiological results was 23.9%, with S. aureus showing the strongest correlation (46.7%). Results of cultures of bone biopsy and needle puncture specimens were identical for 10 (32.3%) of 31 patients. Bone bacteria were isolated from the needle punctures in 7 (33.3%) of the 21 patients who had positive bone biopsy specimen culture results. If the results of cultures of needle puncture specimens alone had been considered, 5 patients (16.1%) would have received unnecessary treatment, and 8 patients (38.1%) who had positive bone culture results would not have been treated at all.
Conclusions. Our results suggest that needle punctures, compared with transcutaneous bone biopsies, do not identify bone bacteria reliably in patients with diabetes who have low-grade infection of the foot and suspected osteomyelitis.
There is still no consensus on the most appropriate strategy for treating osteomyelitis of the foot in patients with diabetes . Some authors suggest that the infected bone should be removed to arrest the infection, but others have obtained satisfactory clinical success rates with use of a strictly medical approach [2–6]. In all but one of these studies, the decision to administer antimicrobial agents, such as fluoroquinolones, clindamycin, and rifampin, was not based on bone culture results . Conversely, some authors have stated that osteomyelitis of the diabetic foot could not be treated without resecting the infected bone [7, 8]. A recent study reported by our group showed that, after a mean of 12 months of follow-up, a significantly greater percentage of patients with diabetes recovered from osteomyelitis of the foot without surgery when the choice of antibiotics was based on bone culture results, rather than based on another method (77% vs. 46%) . The standard procedure for diagnosing osteomyelitis of the foot is generally considered to be culture of a bone biopsy specimen obtained without traversal of an open wound [10, 11]. Bone biopsies are rarely performed in routine care, however, especially in diabetes centers, where antibiotics are generally chosen on the basis of the results of culture of superficial samples or of deep samples obtained during an amputation or surgical debridement [2, 3, 5, 6, 12]. Kessler et al.  have recently suggested that culture of needle puncture specimens could be used to identify bone bacteria in patients with diabetes with osteomyelitis of the foot in cases in which surgical debridement is not recommended. Although the needle puncture is an appealing method, because any physician can perform it without requiring an orthopedic surgeon or a radiologist, Kessler et al.  did not compare the results of needle puncture and bone biopsy specimen cultures. The aim of the present prospective study was to assess the correlation between needle puncture specimen, bone biopsy specimen, and swab sample culture results when these procedures were performed concomitantly for patients with diabetes who had suspected osteomyelitis of the foot.
We conducted a prospective study in the General Hospital in Béthune and the Gustave Dron Hospital in Tourcoing, which are 2 referral diabetes foot clinics in northern France.
Study population. To be included in the study, patients had to be >18 years of age, have type 2 diabetes mellitus, have a foot ulcer overlying a bony prominence, and have radiological abnormalities, such as demineralization, periostal reaction, and bone destruction, that suggested active osteomyelitis. Patients were excluded from the study if they had been receiving antibiotic therapy up to 2 weeks before hospital admission, if they presented with gangrene, if they were found to need an amputation at physical examination, or if they lacked dorsal-pedal and posterior-tibial pulses.
Sample collection. To obtain a swab sample, the ulcer first had to be cleaned with a sterile saline solution, to reduce the number of contaminating bacteria. Swab samples were collected by applying a sterile compress onto the surface of the ulcer. A sterile compress imbibed with a polyvidone-iodine solution was used to clean the periphery of the ulcer. The area was then washed with a sterile saline solution (0.9%). To obtain the needle puncture sample, a syringe with 1 mL of sterile saline solution (0.9%) and an 18-gauge needle was inserted transcutaneously at least 20 mm from the periphery of the ulcer, to limit the risk of contamination by colonizing flora. Radiographs of the foot guided the needle extremity toward the bony structure underlying the wound without penetrating the cortical surface of the bone. Once a drop of fluid was obtained through aspiration, it was placed in a sterile receptacle. If no fluid was obtained, the needle was rinsed with a saline solution and placed in a sterile receptacle. Needle puncture procedures were in accordance with the protocol described by Kessler et al. .
In Tourcoing, an orthopedist performed transcutaneous bone biopsies in the surgical room, and in Béthune, a radiologist performed the biopsies in the radiology room. After disinfecting the skin that surrounded the ulcer, the surgeon used fluoroscopy, as did the radiologist in Béthune hospital, to guide the insertion of an 11-gauge biopsy needle (Becton Dickinson) into a skin incision 5 to 10 mm in length and at least 20 mm from the periphery of the ulcer, in accordance with previously described procedures . In the case of severe peripheral neuropathy resulting in superficial and deep anesthesia, no additional anesthesia was administered. Swab, needle puncture, and transcutaneous bone biopsy samples were transported in sterile receptacles and sent to the microbiology laboratory. The laboratory maintained aerobic and anaerobic cultures of all samples for 5 days and 2 weeks, respectively. We used API (Biomérieux) to identify bacterial isolates at the species level. Culture and identification procedures were identical at both centers. Histopathological examinations were performed at both centers, although not for all patients. The results of these examinations included aggregates of inflammatory cells (neutrophils, lymphocytes, histiocytes, and plasma cells), erosion of trabecular bone, and marrow changes (from loss of normal marrow fat to fibrosis and reactive bone formation), and they were assessed for consistency with culture results.
Overall, 451 patients with diabetes presented with foot ulcers from July 2006 through February 2008. Of these patients, 79 (17.5%) had suspected osteomyelitis, and 31 met the inclusion criteria. The mean age of the patients included in the study was 63.4 years, and the mean time since the onset of diabetes was 12.5 years. The main reasons for study exclusion were absence of a foot ulcer (in 10 patients), urgent need for surgical drainage of the foot (17 patients), and antibiotic therapy within 2 weeks of hospital admission (21 patients). The locations of the suspected osteomyelitis were the metatarsal head and the phalanx in 21 (67.7%) and 10 (32.3%) patients, respectively. Mean time (±SD) at inclusion since receipt of the most recent antibiotic was 4.2±2.7 weeks (range, 2–12 weeks). Twelve patients underwent surgical bone biopsies in Tourcoing, and 19 patients underwent radiological biopsies in Béthune. No adverse events related to bone biopsies or needle punctures were reported. Culture results were positive for 21 (67.7%) of the bone biopsy specimens, 18 (58%) of the needle puncture specimens, and 30 (96.8%) of the swab samples. All needle puncture and bone biopsy specimens had negative Gram stain results. Staphylococcus aureus was the most common type of bacteria that grew from bone biopsy specimens, followed by Proteus mirabilis and Morganella morganii (table 1). The distribution of microorganisms was comparable for cultures of bone biopsy specimens, needle puncture specimens, and swab samples (table 1). The mean number of bacteria types per positive sample was 1.35, 1.32, and 2.51 for bone biopsy specimens, needle puncture specimens, and swab samples, respectively. Thirteen (69%) of the 20 patients with culture-positive bone biopsy specimens also had culture-positive needle puncture specimens. Overall, the results of these 2 techniques had a 23.9% correlation, with S. aureus being associated with the highest correlation (46.7%) (table 2). Ten patients (32.3%) had identical bone biopsy specimen and needle puncture specimen culture results. Bone bacteria were isolated from needle puncture specimens in 7 (33.3%) of the 21 patients who had bone biopsy specimens with positive results. Overall, the correlation between bone biopsy specimen culture results and swab sample culture results was 41.7%, with S. aureus having the highest correlation (82.3%) (table 3). If needle puncture alone had been used in this study, 5 patients (16.1%) with negative bone biopsy specimens culture results would have received unnecessary treatment, and 8 patients (38.1%) with positive bone biopsy specimen culture results would not have been treated. We found that the microorganisms associated with these discrepancies were mostly gram-negative rods, especially Pseudomonas species and Proteus species; gram-negative rods accounted for 76% of the inconsistencies. Results of histopathological examinations were consistent with the diagnosis of osteomyelitis in 5 (35.7%) of 14 patients. Results of histological examinations and bone biopsy specimen cultures matched in 6 (42.9%) of 14 cases (table 4).
The results of our prospective study suggest that needle punctures, compared with transcutaneous bone biopsies, do not detect bone bacteria reliably in patients with diabetes who have suspected osteomyelitis of the foot. Our data also confirm other study findings that have shown that S. aureus is the most common form of bacteria found in matching bone biopsy specimen, needle puncture specimen, and swab sample cultures [14–16]. Our finding that needle punctures isolate a lower number of microorganism types than do superficial swabs (1.32 vs. 2.51) is consistent with the findings of Kessler et al. . The number of bacterial isolates was almost identical for bone biopsy specimens and needle puncture specimens (1.35 and 1.32, respectively). The correlation that we found between bone biopsy specimen and swab sample culture results was higher than the correlation that our group recently reported . This difference may be related to the prospective design of the present study, in which all microbiological samples were obtained concomitantly. In our previous study , samples were obtained within 3 days of each other. Because patients with recent antibiotic use, patients with gangrene, and patients who required amputation were excluded from the present study, the results presented here only apply to patients with low-grade infections of the foot. In patients with more fulminant infection caused by virulent pathogens, the correlation rate between needle puncture specimen, swab sample, and bone biopsy specimen culture results might be higher than that established in the present study. In addition, some of our patients who had negative bone biopsy specimen cultures probably had no bone infection but did have neuro-osteoarthropathy of the foot. When we used bone biopsy specimen cultures to diagnose osteomyelitis, we found that one-third of patients who had both a foot ulcer overlying a bony prominence and radiograph abnormalities that suggested osteomyelitis did not, in fact, have a bone infection. It is generally accepted that neuropathic bone disorders, which are frequent among patients with diabetes, may cause radiograph abnormalities [10, 11, 17–19]. Antibiotic treatment most likely played no role in false-negative bone culture results, because the mean time since previous receipt of antibiotics was 4.2 weeks before inclusion in the present study. Results of histological examinations and bone culture results were not strongly correlated (42.8%), perhaps because there are no well-designed criteria for the histological diagnosis of osteomyelitis of the foot in patients with diabetes . Microbiological testing may, therefore, be more useful than histological examination, as has been previously suggested by Lavery et al. .
Our results confirm the high rate of negative needle puncture specimen culture results reported by Kessler et al. (13 [41.9%] of 31 vs. 5 [23.8%] of 21) . This may explain why we found that bone biopsy specimen results were more closely correlated with results from swab samples than with results from needle puncture specimens. If we had administered antibiotic therapy on the basis of needle puncture specimen culture results, 38.1% of our patients with positive bone biopsy specimen culture results would not have been treated. Conversely, 16.1% of patients with positive needle puncture specimen culture results had negative bone culture results and would have received unnecessary treatment in this scenario. Although the latter situation is less harmful than the former, it is associated with an increased risk of drug-related adverse effects, including Clostridium difficile diarrhea and the emergence of antibiotic-resistant organisms [20–23]. Some authors have stated that the use of non-bone specimens as a guide to antibiotic therapy in chronic osteomyelitis could cause an unacceptably high rate of therapeutic error, because bone and non-bone microbiologies are very different [14–16, 24, 25]. However, this suspicion has not been confirmed in prospective, randomized studies. Although transcutaneous bone biopsy is not as precise as open-surgery biopsy, it is the most reliable technique within international guidelines for detecting osteomyelitis of the foot microbiologically in patients with diabetes who do not require surgery [26–28].
Although needle puncture is easy to perform and, therefore, constitutes an attractive method for detecting osteomyelitis of the foot, our results show that it is less reliable than transcutaneous bone biopsy. An important limitation of the bone biopsy is its lack of availability for most physicians. However, our study confirms that the procedure is simple and safe and can be performed by either orthopedic surgeons or radiologists. In addition, patients often do not require anesthesia because of the severe peripheral neuropathy associated with many cases of osteomyelitis of the foot .
Our results suggest that needle punctures identify bone bacteria less reliably than do transcutaneous bone biopsies in patients with diabetes who have low-grade infection of the foot and suspected osteomyelitis. Microbiological examination of the bone biopsy specimens appears to be more useful than histological examinations in this setting. For patients who do not require surgery, transcutaneous bone biopsy remains, at present, the best method for identifying microorganisms in bone.
We thank Catherine Sloan for proofreading the manuscript.
Potential conflicts of interest. All authors: no conflicts.