Epidemiology, microbiology and therapeutic consequences of chronic osteomyelitis in northern China: A retrospective analysis of 255 Patients

The study aimed to explore the epidemiology and clinical characteristics of chronic osteomyelitis observed in a northern China hospital. Clinical data of 255 patients with chronic osteomyelitis from January 2007 to January 2014 were collected and analyzed, including general information, disease data, treatment and follow-up data. Chronic osteomyelitis is more common in males and in the age group from 41–50 years of age. Common infection sites are the femur, tibiofibular, and hip joint. More g+ than g- bacterial infections were observed, with S. aureus the most commonly observed pathogenic organism. The positive detection rate from debridement bacterial culture is 75.6%. The detection rate when five samples are sent for bacterial culture is 90.6%, with pathogenic bacteria identified in 82.8% of cases. The two-stage debridement method (87.0%) has higher first curative rate than the one-stage debridement method (71.2%). To improve detection rate using bacterial culture, at least five samples are recommended. Treatment of chronic osteomyelitis with two-stage debridement, plus antibiotic-loaded polymethylmethacrylate (PMMA) beads provided good clinical results in this study and is therefore recommended.

Chronic osteomyelitis is a common clinical disease and a challenging disorder, characterized by its long disease course, difficult early diagnosis and high disability rate. The clinical characteristics of chronic osteomyelitis are varied, and may be affected by geography, time, and pathogenetic differences. Geographically, developing countries have a higher incidence of the disease than developed ones, likely caused by differences in economic foundation, lifestyle and healthcare level 1,2 . Over time, a shift has occurred from predominately hematogenous osteomyelitis several decades ago, to a predominance of chronic osteomyelitis that results from trauma, implant infection, and diabetes 3-6 . In recent years, the number of patients with open wounds and multiple fractures from road and industrial accidents has sharply increased in China, as the Chinese economy undergoes rapid development. Multiple injuries are difficult to treat and frequently lead to infection; thus, treatment exerts great pressures on patients, both economically and psychologically. It also poses great challenges for orthopedists 7 . Wang 8 and Jiang et al. 9 provide recent data on the epidemiology of chronic osteomyelitis in southwest and southern China, which can be used by local government policy-makers and by clinicians. However, China is a country of great diversity in its population, climate and culture. Currently, data and relevant research are lacking for northern China on the epidemiology of chronic osteomyelitis. To address this, we conduct an epidemiological statistical analysis on 255 patients at a prominent hospital and explore the clinical characteristics of chronic osteomyelitis in northern China.

Methods
A retrospective analysis was completed on 255 cases of patients with chronic osteomyelitis seen in the trauma orthopedic department at our hospital from January 2007 to January 2014. (1) osteomyelitis diagnosis; (2) local swelling and bone pain on examination; (3) draining fistula present; (4) imaging procedures completed; (5) microbiological and histopathological examinations completed; (6) biochemical examinations completed.
Exclusion criteria were: (1) acute osteomyelitis (less than 2 weeks); (2) no surgical treatment done; (3) osteomyelitis site in the spine, pelvis or skull; (4) complete data and examinations not available; (5) patient had diabetic foot infection and was not treated in our department.
All data were obtained from the case management system in the medical record room of the hospital prior to this research.
The following data were collected: 1. Demographic information: age and gender; 2. Disease data: infection site, Cierny-Mader classification, laboratory examination results (including erythrocyte sedimentation rate (ESR), C-Reactive protein (CRP) and white blood cell (WBC)), and bacterial culture results (respectively from debridement and sinus tract); 3. Treatment method: two-stage debridement and one-stage debridement; 4. Follow-up data were obtained by searching medical records and by telephone. All patients were successfully followed-up 9-93 months (46.2 month in average) after the operation.
The Ethics Committee of our hospital approved this study and any related procedures were performed in accordance with relevant guidelines and regulations.
Bacterial Culture and Identifying Pathogenic Bacteria. Different samples were collected during debridement and in sinus tract and sent for bacterial culture. The results are essential to identify pathogenic bacterial.
If the same bacteria type was detected in at least two different samples, it was identified as a pathogenic bacterium. Based on this principle, we examined the number of samples that tested positive with the number of samples analyzed, and the differences between them in identifying the pathogenic bacteria.
Bacterial cultures from the sinus tract samples can be an additional aid when examining chronic osteomyelitis and its result can serve as a reference for diagnosis and treatment. Samples from 58 patients with sinus tracts were sent for bacterial culture. We compared sinus tract positive result with the debridement bacterial culture result, to see their difference of ability in identifying pathogenic bacteria.
The coincidence rate between sinus tract positive result and identified pathogenic bacterial were examined, to assess its ability in identifying pathogenic bacteria. Treatment Methods. Two treatments were used in the study.
Step 1: Debridement and implantation of antibiotic-loaded PMMA. At the first stage of debridement, the antibiotics vancomycin and imipenem are used locally.
Step 2: During the one-week clinical observation, X-rays and CT scans are repeated, along with laboratory tests of CRP and ESR levels and reexamined while awaiting pathology results from the debridement bacterial culture. Patients received intravenous antibiotics during this time.
Step 3: Debridement was performed again after one week. The original antibiotic-impregnated cement beads were removed. Based on the pathogenic bacteria detected and its sensitivity test, appropriate antibiotic PMMA beads were implanted.
This treatment included only step 1 of the two-stage treatment above.
In the current study, we compared and analyzed their clinical effects and prognosis of each treatment. CRP; and, whether progressive periosteum thickening or osteolysis was no longer present on a recent X-ray image.
Ethics approval and informed consent. The current study was approved by the ethical medical committee of Qinghai Provincial People's Hospital. All participants gave informed written consent.

Statistics analysis.
Statistical analysis was performed with SPSS software (Version 21.0; SPSS Inc., Chicago, IL). Frequency data were compared by Pearson's chi-square test. For normally distributed data, independent groups were compared by a Student t-test or one-way analysis of variance (ANOVA). Where data were not normally distributed, the Mann-Whitney U test or Kruskal-Wallis H test was adopted. Results were considered significant at p < 0.05.

Results
Gender Ratio and Age at First Diagnosis. The present study included 202 males (79.2%) and 53 females (20.3%), an approximate gender ratio of 4:1.
The median age at first diagnosis was 45.5 years. Approximately 80% of affected patients were between 21 and 60 years of age (206 cases). The top three age groups represented were 41 to 50 years (29%), 21 to 30 years (20.8%), and 31 to 40 years (16.5%). Classifications. All 255 cases met the Cierny-Mader classification criteria for chronic osteomyelitis (Table 1).

Infection Site. All cases included in
Type IIA (23.1%), type IIIA (20.4%), and type IVA (18.0%), are the most frequent classifications in the study. Laboratory Test. The white blood cell count (WBC), ESR, and CRP of 255 patients before operation are presented in Table 2   In total, 41 types of bacteria were detected. The 13 most frequently detected bacteria are given in Table 4. Another 28 bacteria were detected three times or less, including Klebsiella oxytoca (3 times or 1.22%), Citrobacter braakii (3 times or 1.22%), Staphylococcus sciuri (2 times or 0.82%), S. kloosii (2 times or 0.82%), and Proteus mirabilis (2 times or 0.82%). The following bacteria were detected only once (0.    Different numbers of samples were sent for detection of bacteria using bacterial cultures. Correlations between the number of samples sent and the detection rate were examined through pairwise comparisons (Table 5). Five samples sent to bacterial culture achieved highest detection rate.
Identifying Pathogenic Bacteria. According to the results, it is statistically significant between 2 samples and 4 samples, 2 samples and 5 samples, 3 samples and 5 samples. Five samples sent for detection has the greatest likelihood of identifying pathogenic bacterial (Table 6).
Sinus Tract. Samples collected from 58 patients with sinus tracts were sent for bacterial culture (Table 7).
We compared the result of sinus tract bacterial culture with the pathogenic bacteria. Among 57 positive samples, only 24 samples (42.1%) from sinus tract were in consistent with the pathogenic bacteria and more than half samples (33 samples, 57.9%) were different with pathogenic bacteria.
In addition, 82.8% pathogenic bacterial were identified when 5 samples in debridement sent to bacterial culture (Table 6), while sinus tract bacterial culture only have 42.1% coincidence rate with pathogenic bacteria. Obviously the former one has stronger ability in identifying pathogenic bacteria.   Two-Stage Debridement and One-Stage Debridement. Two-stage debridement treatment has higher union rate (87.0%) than the one-stage debridement (71.4%) and the result was statistically significant (p < 0.05).

Times of detection
In addition, one-stage debridement has more recurrence than two-stage debridement (p < 0.05) ( Table 8).

Discussion
Among patients with chronic osteomyelitis, the male:female ratio was 4:1. This is in accordance with Kremers's report 10 , which argued that increasing road and industrial accidents contributed to the greater number of male patients, since males are more likely to engage in heavy physical labor or high risk activities.
The current study showed chronic osteomyelitis was highest in the age group of 41 to 50 year-olds (29.0%), probably because road accidents frequently occur among those aged 20-50 years 11 . Traffic trauma is a high energy trauma that often leads to an open wound, a potential pathogenic factor for traumatic osteomyelitis. In addition, this age group has a high incidence of diabetes, which increases the frequency of diabetes-related osteomyelitis. Kremers 10 reports that American patients with diabetes-related osteomyelitis have increased from 2.3/100,000 in the late 1970s, to 10.5/100,000 in the 1990s, although currently its incidence appears to be stable at approximately 7.6/100,000. According to Cierny et al. 12 , immunity of an organism is a significant factor in the occurrence and transformation of osteomyelitis. Different age groups have different resistance to infection, and older individuals may be more likely to have osteomyelitis because they have weaker immune systems.
The top three infection sites of chronic osteomyelitis -the femur (30.2%), tibiofibular (25.9%) and hip joint (15.3%) -account for 71.4% of all infections. This is consistent with other clinical reports 13,14 , but differs from the clinical reports of south China. Wang et al. 8 reports that in southwest China, the top two infection sites are tibia (57.5%) and femur (26.8%). Jiang et al. 9 reported a similar result for southern China, finding the most frequent single infection site was the tibia (39.00%), followed by the femur (24.46%), and calcaneus (11.46%). One possible reason for this difference between north and south China could be the dense population found in southern China, leading to more frequent road and industrial accidents. As noted above, such accidents are more likely to create trauma with an open wound, and the thin soft tissue surrounding the tibia and lack of blood supply can increase the likelihood of wound infection, followed by chronic osteomyelitis. Thus, tibia osteomyelitis has the highest incidence in the north China.
The top three classifications in the current study are type IIA (23.1%), type IIIA (20.4%), and type IVA (18.0%), accounting for 61.5% of cases in total. In the domestic literature, the Cierny-Mader classification information given is often insufficient, with only anatomical type staging provided and no classification of the physiological type of the host 15 . In foreign literature using the Cierny-Mader classification, different literature focuses on different case groups, and the results also differ from each other. Hence, there is no uniform consensus on the statistical distribution of Cierny-Mader types [16][17][18] .
Cierny et al. 12 discussed osteomyelitis classification in detail. They held the view that osteomyelitis classifications are not fixed for a particular case, but rather that osteomyelitis may transform from one classification to another in the same patient, as a process of natural progression or treatment. Thus, different types of osteomyelitis can be thought of as different pathological/physiological stages of disease 12 . The current study suggests that most chronic osteomyelitis falls into types II, III, and IV. For the physiological class, host A is more common than host B, and host C type is seen only rarely.
Identifying the pathogenic bacteria involved is an important step in treating chronic osteomyelitis. Based on characteristics of the pathogenic bacteria, effective antibiotics specific to the bacteria found should be used in clinical practice; this is a precondition to guaranteeing a curative effect. Sending more samples at once was an  Table 8. Clinical effects and prognosis of the two different treatments.
effective method to improve the detection rate of pathogenic bacteria. Lew et al. 22 suggested that five samples or more should be sent when treating chronic osteomyelitis, and that this could significantly improve the microbial detection rate. Our study similarly showed that the more sample sent for bacterial culture, the higher the detection rate. Sending only one sample had the lowest detection rate (66.7%), while five samples gave the highest rate (90.6%). Therefore, five or more samples should be sent from debridement to improve detection rates. Typically, pathogenic bacteria and contamination bacteria are both observed in the results from bacterial cultures. If the same bacteria is detected twice or more in the culture results, it is presumed to be the pathogenic bacteria 23 . Using this principle, contaminant bacteria can be distinguished from pathogenic bacteria in the results. If all cultures give negative results, or if only one or differing results are observed among samples, then pathogenic bacteria cannot be determined. The current study indicates that the more samples that are sent for examination, the higher the probability of identifying the pathogenic bacteria. The probability of identifying the pathogenic bacteria was significantly improved (82.5%) if five or more samples were sent for bacterial culture; this is higher than the rates reported from Wang (71.5%) and Kremers et al. (75%).
In the current study, results from sinus tract bacterial culture were consistent with pathogenic bacteria in only 42.1% of cases. This might be caused by contamination with nonpathogenic bacteria during the sampling process, or other naturally-occurring random factors 8,24 . Clinicians may also be collecting these samples in a non-standard manner. Regardless, the results from sinus tract bacterial culture can only be used as a reference for finding antibiotics to which bacteria are susceptible, not as the basis of the diagnosis and treatment decision.
In treating chronic osteomyelitis, the patient's overall condition and local lesion must be considered to accurately assess the patient's state before making careful decisions regarding treatment. The published literature includes a number of clinical treatments for chronic osteomyelitis [25][26][27][28] , such as radical debridement, bone transport, acute limb shortening, two-staged reconstruction, local antibiotic therapy (antibiotic-impregnated beads, spacers, cements, intramedullary nails) and soft tissue grafting (free flap, myocutaneous flap, skin graft). Nonetheless, a common standard of diagnosis and treatment has not been established. Two basic consensuses are present regarding the treatment of chronic osteomyelitis. The first is that debridement is key to all treatments. The second is that guaranteeing a curative effect requires identifying the pathogenic bacteria involved and the use of appropriate antibiotics. The site of bone defect should be treated with local antibiotic-loaded PMMA spacers. External fixation or internal fixation can stabilize fracture ends.
Cierny 29 and Forsberg et al. 30 repeatedly emphasize radical debridement as the foundation for treating chronic osteomyelitis. A thorough scrape is needed at the dead space, necrotic tissue and sinus tract, until fresh blood is exuded, indicating healthy tissue, and plenty of irrigation must be done. Results of the current study suggest a good curative effect (87%) is achieved by repeated radical debridement, and implantation of appropriate antibiotic bone cement beads in the local lesion, with external or internal fixation. The reasons are as follow. 1) In the two-stage debridement method, debridement is completed in a more thorough manner. Within one week, debridement is performed twice. The second debridement removes all possible contaminated tissues based on the first debridement, thus reducing the possibility of recurrence 22,31,32 . 2) Sending five samples for bacterial culture in the first debridement improves the detection rate of pathogenic bacterial. Further, an appropriate (i.e., one the bacteria was susceptible to) antibiotic release-carrier can be identified; it is more successful to use antibiotics based on the results of the bacterial culture and susceptibility testing. This also ensures rare bacteria are quickly recognized, and reduces the side effects of antibiotics since targeted antibiotics are used 33 . 3) An appropriate antibiotic release-carrier implanted after one week maintains the effective blood concentration of the antibiotic in local tissue longer. Previous research has shown that, after the blood antibiotic concentration reached its peak, it would gradually decline, so that after a few weeks, local antibiotic blood concentration was less than the effective blood concentration 34,35 . Therefore, implanting a second, targeted antibiotic release carrier after one week could extend the duration of the local blood concentration, thus better inhibiting the growth of pathogenic bacteria. The method of repeating radical debridement and local antibiotic-loaded PMMA spacers ± external fixation gave good clinical results and can be recommended. One drawback of this approach is that it increases the number of operations required and prolongs hospitalization, thus increasing expenses.
China is a country of great diversity. The previous studies by Wang 8 and Jiang et al. 9 have summarized the recent data on the epidemiology of chronic osteomyelitis in southwest and southern China. Hence the current study is of great significance to report the data in northern part of China. Apart from the geographic significance, the novelty of current study is shown in a few other aspects. Firstly, we concluded the bacterium-spectrum of chronic osteomyelitis and identified the differences of their ability to identify pathogenic bacteria for different numbers of sample. Secondly, we compared the differences of ability to identify pathogenic bacteria between sinus tract bacterial culture and debridement bacterial culture. Thirdly in terms of treatment method, we recommended repeated radical debridement and antibiotic-loaded PMMA spacers ± external fixation is more superior in treating chronic osteomyelitis.
The present study had several limitations. Since it was conducted in a single medical center in north China, it may not characterize chronic osteomyelitis in the whole area. Therefore, multicenter studies should be performed to obtain more detailed data. Second, this study lacked detailed statistical data on oral antibiotic therapy because most patients received oral antibiotics out of hospital.

Conclusions
In summary, our study showed that chronic osteomyelitis is more common in males and in the age group from 41-50 years of age. Common infection sites are the femur, tibiofibular, and hip joint. Using Cierny-Mader classification, type IIA, IIIA and IVA infections were most common. More g+ than g-bacterial infections were observed, with S. aureus the most commonly observed pathogenic organism.
To improve pathogen detection rate, five or more samples should be sent for bacterial culture. Bacterial culture of the sinus tract performed poorly in identifying pathogenic bacteria, so choosing the appropriate antibiotic based on the result of sinus tract bacteria is not advised. Repeated radical debridement and identification of pathogenic bacteria are keys to successful treatment of chronic osteomyelitis. The current study suggests a treatment of two stage debridement + antibiotic-loaded PMMA spacers ± external fixation is the most effective.