The authors reviewed a consecutive series of 93 patients between March 2009 and February 2016 in our hospital. There were 60 patients (48 males and 12 females), and 61 wounds (35 tibia, 15 femur, 5 fibula, 2 radius, 2 ulna and 2 humerus) with an average age of 43.4 years (range, 18–82 years) in modified group. Cultures that performed by direct biopsy from the involved bone at time of debridement, and local symptoms, clinical examination (X-ray, CT, fever, white blood cell count, erythrocyte sedimentation rate, and C-reactive protein concentration) were used to diagnose osteomyelitis. 53 wounds were found to have positive cultures, and 8 wounds of culture-negative. The bacterial species cultured were 17 MRSA, 12 Staphylococcus Aureus, 6 Pseudomonas Aeruginosa, 4 Escherichia coli, 4 MRSE, 3 Enterobacter cloacae, 2 Acinetobacter baumannii and 5 of Enterococcus Faecium, Klebsiella, Hemolytic Streptococcus, Staphylococcus Epidermidis and Proteus species. There were 12 wounds that was associated with Cierny-Mader type 1, 10 wounds belong to type 2, 39 wounds belong to type 3 and type 4. 56 patients belong to Class A, 4 patients belong to Class B with systemic factors of 1 patient and diabetes mellitus, 1 patient with steroid therapy and 2 patients with tobacco abuse.
There were 33 patients (22 males and 11 females) in conventional group, and 34 wounds (26 tibia, 3 femur, 5 of radius, ulna, elbow, humerus and fibula) with an average patient age of 44.7 years (range, 22–80 years). The bacterial species Culture and clinical examination are same as modified group.30 wounds were found to have positive cultures, and 4 cases with negative cultures. 30 wounds were found to have positive cultures, and 4 wounds of culture-negative. The bacterial species cultured were 16 Staphylococcus aureus, 3 Pseudomonas aeruginosa, 3 Escherichia coli, 2 Klebsiella, 2 MRSE, and 4 with Hemolytic streptococcus, Staphylococcus epidermidis, Acinetobacter and MRSA. There were 5 wounds that were associated with Cierny-Mader type 1, 6 wounds belong to Cierny–Mader type 2, and 23 wounds for both type 3 and type 4. 30 patients belong to Class A, 3 patients for Class B with systemic factors of 1 patient who had diabetes mellitus, and 2 patients with tobacco abuse.
The average course of infection in the modified group was 3.6 months (rang, 3 week-40 months), the average course of infection in conventional group was 3.3 months (range, 3 week–22 months).The NPWT sponges (KCI, TX, USA) were cut to fill and cover the wound after the procedure of surgical debridement, the distal end of the drainage tube connected a vacuum bottle with 20-60KPa, or a cyclical negative pressure container, and the sponges were changed every 3–4 days.
Antibiotics were started empirically in patients after cultures have been obtained, at the time of debridement. Antibiotic treatment guideline and the antibiotics used were tailored to the recovered bacteria and their susceptibility pattern [7, 8]. In all cases, appropriate antibiotic coverage was gained and maintained for the duration of the treatment protocols .
Conventional treatment consisted of re-exploration, removal and debridement of all necrotic nonviable tissues including free sequestrum. The wound was washed with hydrogen peroxide and rinsed with saline solution. The extent of infection and debridement determined the subsequent treatment. The treatment algorithm for the implant was modified with reference to Ziran  (Table 1).
Treatment algorithm for the implant
I) Stable hardware + Bone Not Healed = Retain hardware, antibiotics until healed, then hardware removal.
II) Unstable hardware + Bone Not Healed = Remove hardware, antibiotics, temporary stabilization, spacer, and reconstruction when clean.
III) Stable hardware + Bone Healed = Remove hardware, debride with effort not to destabilize, control dead space, and antibiotics.
IV) Stable hardware + Bone Not Healed + Systemic Effects = Remove hardware, temporary stabilize, spacer, antibiotics, and reconstruction when able, consider amputation if bad host.
1. Stable hardware + bone not healed Cierny–Mader 1 type = remove hardware, temporary stabilize + Antibiotic cement-coated (ACC) rods/ Ilizarov technique, debridement, soft-tissue coverage, and reconstruction when clean.
2. Stable hardware + bone not healed Cierny–Mader 2 type = retain hardware, debridement, soft-tissue coverage, bone healed then hardware removal.
3. Stable hardware + bone not healed Cierny–Mader types 3 and 4 = remove hardware, temporary stabilize/ Ilizarov technique, debridement, soft-tissue coverage and reconstruction when clean.
4. Unstable hardware + bone not healed = remove hardware, temporary stabilize/ Ilizarov technique, debridement, soft-tissue coverage and reconstruction when clean.
5. Stable hardware + bone healed = remove hardware, debridement, soft-tissue coverage.
Autodermoplasty, flap transfer, myocutaneous flap and the other methods including antibiotic irrigation and drug delivery system (DDS) were used in wound repair. The local symptoms, fever, white cell count, erythrocyte sedimentation rate, and C-reactive protein concentration were monitored to evaluate the infection.
Statistical evaluation of the data was carried out using the SPSS Statistics for Windows, version 21.0 (IBM Corp., Armonk, NY, USA). The significant differences between variables were tested using x2-test. P value<0.05 was considered to be statistically significant.