Case 1. The patient was a male child who started walking at one year and two months of age. He stopped walking for no apparent cause at one year and four months of age and was brought to a physician. After two weeks of observation, the symptoms had not improved and the patient was referred to our hospital. At the time of the initial examination, radiographs of the lower limbs showed no apparent abnormalities. The peripheral white blood-cell count was 11,100/µL (11.1 × 109/L), the serum C-reactive protein level was 0.21 mg/dL (2.1 mg/L), and the erythrocyte sedimentation rates were 24 mm/hr and 59 mm/2 hr, indicating a mild inflammatory condition. During subsequent observation, the condition of the patient gradually improved and, in approximately two months, he was able to walk. At one year and ten months of age, the patient again had walking difficulties. Radiographs revealed a radiolucent lesion involving the left distal femoral physis (Fig. 1-A). Computed tomography showed a sequestrum-like opacity in the osteolytic area, with marginal sclerosis extending from the metaphysis to the epiphysis (Fig. 2). Magnetic resonance imaging demonstrated an abnormality at the same site, extending across the high-signal intensity area of the physis in the T2-weighted fat-suppressed spin-echo image. Serum tests done at that time showed a peripheral white blood-cell count of 9,400/µL (9.4 × 109/L), a C-reactive protein level of 0.55 mg/dL (5.5 mg/L), and erythrocyte sedimentation rates of 31 mm/hr and 69 mm/2 hr. At 2.0 years of age, the patient underwent endoscopic surgery for biopsy and curettage of the lesion.
Surgical Technique
The patient was placed under general anesthesia in the supine position. A 1-cm skin incision was made under fluoroscopic guidance at the lateral aspect of the distal femoral diaphysis. A bone tunnel was created percutaneously with use of a 4.8-mm drill. A 2.7-mm-diameter arthroscope was inserted, and the metaphyseal region was observed under saline-solution perfusion. Similarly, a 1-cm skin incision was made at the medial aspect of the distal femoral diaphysis. Another bone tunnel was created with use of a 4.8-mm drill, and a working portal was prepared. A punch was inserted into the working portal, and tissue was removed for pathological evaluation. A motorized shaver was then inserted. While the field of vision was gradually enlarged, the granulation tissue of the lesion was débrided (Fig. 3) until the physis was reached. The bone defect within the growth plate was identified by palpation with a probe. After a portion of the abnormal tissue seen in the center of the growth plate was removed for diagnostic pathology, the area was débrided. The arthroscope was inserted deeper through the bone defect of the physis, and the inside of the epiphysis was examined. The abnormal tissue inside the epiphysis was débrided carefully with a shaver, while caution was exercised to avoid injuring the subchondral bone. After thorough irrigation, subcutaneous fat near the lateral skin incision was removed with a punch and used to fill the physeal defect. After the placement of closed drainage in the metaphyseal region, the incision was closed.
Pathological Findings
Histologically, caseous necrosis, Langerhans giant cells, and granulation tissue composed of epithelioid cells, lymphocytes, and plasma cells were present. Although Ziehl-Neelsen staining failed to identify any acid-fast bacteria, tuberculous osteomyelitis was still suspected as a result of these histological findings.
Postoperative Course
Although the intraoperative diagnosis was chronic osteomyelitis, the bacterial cultures were negative. Mycobacterial cultures were not performed because clinically we did not suspect tuberculous osteomyelitis at the conclusion of the operation. The patient received intravenous cefazolin sodium (0.4 g/day) from the day of the surgery to six days after surgery. Postoperatively, the affected limb was not immobilized. The closed suction drain was removed two days postoperatively. Four days postoperatively, the patient began to walk, seemingly without pain, and he was discharged six days postoperatively (Fig. 1-B). Since we were uncertain as to the diagnosis of this unusual condition and therefore concerned about erroneously treating the patient with drugs that might cause serious side effects, we administered neither antibiotics nor antituberculosis drugs after he was discharged. At the time of an outpatient examination two weeks postoperatively, the patient had a normal gait. A radiograph made seven months postoperatively (Fig. 1-C) showed bone regeneration in the area of curettage. No subsequent problems have been reported. At four years and eight months postoperatively, radiographs demonstrated no growth disturbance and no apparent difference in limb length (Fig. 1-D).
Case 2. The patient was a male child with no particular medical history. He began to experience continuous pain in the right knee at four years and six months of age, and he was brought to our hospital at four years and nine months of age for an initial examination. Radiographic examination identified osteolysis in the proximal aspect of the tibia. Serum tests showed a peripheral white blood-cell count of 8,500/µL (8.5 × 109/L), a C-reactive protein level of 0.15 mg/dL (1.5 mg/L), and erythrocyte sedimentation rates of 28 mm/hr and 61 mm/2 hr. Open biopsy was performed, and a large amount of purulent discharge was observed. Débridement in the area of the lesion was performed under fluoroscopic guidance, and the area was irrigated. A smear of aspirated tibial bone-marrow fluid was positive for acid-fast bacilli, and the polymerase chain reaction test was positive for Mycobacterium tuberculosis. Antituberculosis drugs (isoniazid 7.5 mg/kg, pyrazinamide 20 mg/kg, and rifampicin 15 mg/kg) were administered, but the osteolysis, intermittent purulent discharge, and fistula did not improve. At the age of five years, the patient once again underwent débridement of the lesion under fluoroscopic guidance, and the area was irrigated. Despite continuation of antituberculosis drugs, fistula formation and purulent discharge persisted and enlargement of the lesion was observed on radiographs (Fig. 4-A). A computed tomography scan showed a sequestrum in the osteolytic area, with marginal sclerosis extending from the metaphysis to the epiphysis. Magnetic resonance imaging showed an area of high signal intensity extending across the physis on the T2-weighted image (Fig. 5) and a fistula extending to the anterolateral aspect of the skin. At five years and ten months of age, the patient underwent endoscopic surgery for curettage of the lesion in a manner similar to the technique described for the patient in Case 1, with use of a 4.0-mm arthroscope and two 4.8-mm bone tunnels. The sequestrum and all necrotic and infected tissue in the metaphysis, the physis, and the epiphysis were débrided. After thorough irrigation, the physeal defect was filled with subcutaneous fat. After placement of closed drainage in the metaphyseal region, the incision was closed. Histologically, the lesion contained granulation tissue composed of caseous necrosis, epithelioid cells, and plasma cells. Langerhans giant cells were also scattered. A diagnosis of tuberculous osteomyelitis was made.
Postoperative Course
Antituberculosis drugs (isoniazid 7.5 mg/kg, pyrazinamide 20 mg/kg, and rifampicin 15 mg/kg) were administered orally for one year. Postoperative purulent discharge was not observed, and the drain was removed at the end of the first week. A computed tomography scan six weeks postoperatively showed no sequestrum, and the blood tests had also normalized at this point in time. A radiograph made six months postoperatively showed bone regeneration (Fig. 4-C) in the previously curetted areas (Fig. 4-B). Radiographs made three years postoperatively showed no growth disturbance in the physis and no obvious difference in limb length (Fig. 4-D).
Previous reports contain descriptions of techniques for inserting an endoscope into bone to excise a femoral head chondroblastoma3, to enhance visualization during excision of a central physeal bar4, to prevent articular penetration in screw fixation for slipped capital femoral epiphysis5, to débride a pin-track infection after external fixation6, and to visualize the medullary canal after nail removal for the treatment of nonunion and osteomyelitis7. The techniques we describe in this study differ from those previously reported: by drilling dual opposing insertion paths, we were able to provide an enhanced field of vision and to manipulate instruments much as we do during arthroscopy, thus enabling us to reduce encroachment into the growth plates during resection of the lesion.
Although surgery for intramedullary lesions can be performed with use of hand instruments under fluoroscopic guidance8, removal of the lesion sometimes is incomplete, which points to the need for a good field of vision to better ensure complete removal of the lesion. One reported case of tuberculosis of the proximal part of the tibia involving the growth plate, similar to the case of our patient in Case 2, required a second surgical procedure involving aggressive curettage of the lesion through a posterior approach after a failed first curettage and débridement9. For the patient in Case 2, débridement under fluoroscopic guidance alone yielded unsuccessful results twice. In contrast, distinguishing between normal and abnormal tissues was facilitated with use of the endoscope because the sequestrum had a yellowish luster.
Reports have suggested that resection of a central physeal bar may be indicated when as much as 50% of the area of the growth plate has been lost1,8, which implies that there is substantial capacity for future bone growth when the physeal defect affects <50% of the physis. Therefore, we believe it prudent to preserve as much width of the normal physis as possible. Fortunately, no obvious growth disturbances have been observed in the one patient who was followed until twelve years of age9 or in our two patients who were followed for three or more years.
By facilitating improved three-dimensional visualization of the lesion, endoscopy offers the opportunity to minimize the area of excision of the physis and the epiphysis when the lesion involves these areas and when complete excision of the lesion is necessary. In cases of osteomyelitis, endoscopic surgery offers a second advantage by enabling thorough local irrigation intraoperatively.
Intramedullary bone endoscopy (medulloscopy) has been described in several basic research studies. Roberts et al. studied the area that could be visualized by inserting an endoscope into the medullary cavity of a cadaver tibia proximally. However, they found that an arthroscope was not suitable for visualizing the distal aspect of the tibia, and they were therefore required to use a flexible rhinolaryngoscope10. In our method, since we used a drill hole for endoscopic placement, we were able to choose the insertion path according to the site that needed to be visualized, and we therefore did not need a flexible scope.
Oberst et al.11 studied tibiae obtained from above-the-knee amputations. They measured the intramedullary pressure while pushing an endoscope into the medullary canal and demonstrated that endoscopy did not increase the risk of fat embolism or bone necrosis, nor did it reduce cortical blood flow. During our endoscopic procedures, we perfused saline solution by means of gravity. Our techniques with use of a gravity system resulted in no complications; however, an arthroscopy pump is considered to be more effective than a gravity system for obtaining a more satisfactory field of vision in the medullary cavity, especially in the presence of active bleeding. Therefore, basic research studies that examine whether perfusion pumps may increase the risk of fat embolism and bone necrosis due to excessive intramedullary pressures are advisable, as, to our knowledge, no basic research studies to date have investigated intramedullary pressures associated with perfusion fluid placed under pressure.
Although our two case reports focused on the benefits of the endoscopic procedure, we would note that the postsurgical management for the patient in Case 1 fell short of the recommended standard of care for tuberculous osteomyelitis. Since we would never condone treating chronic osteomyelitis with surgery alone, we administered intravenous cefazolin sodium for a week because we initially did not suspect tuberculosis in this patient and instead had made an intraoperative diagnosis of chronic osteomyelitis of unknown etiology. After receiving the final pathology report, which strongly suggested the diagnosis of tuberculous osteomyelitis despite a failure to identify acid-fast bacteria, we stopped the administration of intravenous cefazolin sodium. Since the patient had already received a course of antibiotics and because antituberculosis drugs are associated with the risk of serious side effects, we decided to carefully observe the postoperative course of the patient with the intention of starting antituberculosis therapy if we observed any sign of recurrence. In the case of this patient, fortunately, we did not need to initiate further antimicrobial treatment. However, for cases of tuberculous osteomyelitis, we definitely would recommend antimicrobial therapy for an appropriate duration of time after endoscopy, along the lines of our postsurgical management of the patient in Case 2.
Endoscopic surgery with drilled dual opposing insertion paths appears to be especially useful in the complete débridement of sequestra and the minimization of encroachment into the physis during surgery. Although surgery for intramedullary lesions can be performed with use of traditional open procedures under fluoroscopic guidance, our experience with these two unusual cases in which the lesion extended across the physis indicates that there may be a potential advantage to the use of endoscopic techniques in such patients. 