Infection after anterior spinal fusion for idiopathic scoliosis using the Cotrel-Dubousset-Hopf system: A clinical case series of three patients

Discussion

Complications in surgical treatment of idiopathic scoliosis are encountered in about 5.7% to 15.4% of the patients. The use of an anterior or a posterior approach does not appear to be of influence on the overall complication rate.^{6, 7} Infections after PSF for idiopathic scoliosis are reported in 1.4% to 5% of cases, whereas after ASF these percentages are reported to be considerably lower; 0% to 0.2%.^{7, 10, 11} The low postoperative infection rates after ASF may be due to less intraoperative soft tissue damage and devascularisation as compared to posterior procedures.¹²

In contrast to the previously reported infection rates after ASF, we found a relatively high infection rate in our CDH-series, with 3 out of 16 patients affected. This is an exceptionally high infection rate for anterior instrumented spinal fusion. A possible explanation is a break in sterile technique within the operation theatre. Review of the local complication registration database of the department of Orthopaedic surgery did not show an increased infection rate for instrumented spinal surgeries or total hip or knee arthroplasty in the years 1999-2002 when the CD-Hopf system was implanted in our clinic. There were no breaks in sterile technique described either. The most likely sources of infection are intra-operative inoculation or hematogenous spreading of the pathogens.¹³ The P. Acnes bacterium, as cultured in the first case, is known to cause delayed infection with a long interval between inoculation of the pathogen and onset of symptoms.¹⁴ The insidious onset in the second case suggests a low virulent microorganism as the causative agent as well. The cultures probably were unreliable due to start of antibiotic treatment, before representative cultures were obtained. In the third case, S. Aureus was the causative bacterium, suggesting a hematogenous origin. These findings confirm the need for obtaining multiple representative intra-operative cultures and prolonged culturing if initially no micro-organism is found.

Implant factors influence the infection risk as well. In comparison to other ASF-systems the CDH-system is relatively bulky. This may increase the risk of damaging vital structures due to friction with the implant.¹⁵ In PFS increasing volume of the implant has been proposed as a risk factor for higher infection rates.¹⁶ In anterior procedures this relation seems less obvious. The spinal implant alloy is considered to influence the development of late onset infection as well. Stainless steel implants, as the CDH system, seem more subject to late post-operative infection as compared to titanium systems. The latter having greater bone adhesion on the hardware, resulting in the production of thinner biofilm.^{13, 17}

To our knowledge, only two reports have addressed the clinical presentation of infection after ASF and instrumentation.^{8, 9} Robertson and Taylor described three cases of late onset infection after anterior scoliosis correction using Dwyer instrumentation.⁸ Two of these patients presented with malaise and an elevated ESR level but none of the patients experienced back pain. One patient developed a draining sinus in the groin and one patient an abscess in the groin, both originating from the spinal fixation.⁸ Van Luijk and de Nies reported one case of infection after anterior scoliosis surgery, with scoliosis curve progression and a hip flexion contracture one year after surgery.⁹ The patient did not experience back or abdominal pain and the infection parameters were within normal range.⁹ We found similar clinical presentations as Robertson and Taylor, with absence of specific symptoms (Table 2).⁸ Infection was suspected, but only discovered after formation of a fistula or an abscess. Late onset infections after PSF often present with diffuse back-pain after an interval of normal postoperative recovery, malaise, slightly elevated infection parameters and absence of fever.^{12, 16, 18–20} In contrast with late PSF infections, after ASF the pain, if present, can be poorly localized and may even mimic an abdominal problem.

The role of diagnostic imaging in infection after spinal fusion and instrumentation has not been clearly established.^{8, 9, 12}, ^{19, 21} Plain radiographs may show a pseudarthrosis or screw loosening associated with infection. Technetium bonescan, Gallium-scan and Indium-111-labeled leukocyte scintigraphy may be useful diagnostic tools, but have been reported to have high false negative rates.^{12, 15, 21} MRI with gadolinium enhancement and CT have 91% sensitivity when a fluid collection or abscess is present.²¹ Robertson and Taylor reported negative findings with plain radiographs and Technetium bone scan.⁸ In their patient with a draining sinus, the infection was confirmed with CT scan and Gallium-scan.⁸ Van Luijk and de Nies reported negative or inconclusive findings with initial plain radiographs, CT and MRI imaging. After one year, MRI showed the presence of an abscess, confirming the suspected infection.⁹ In our series initial imaging was inconclusive in two of the three cases (Table 2). Only in case 2 plain radiographs showed clear loosening of the screws, indicating infection. In two out of three cases a bone scintigraphy failed to show signs of infection. Only if patients developed an abscess or fistula, infection could be confirmed by CT or MRI.

The recommended treatment for delayed infection after PSF for scoliosis, consists of surgical debridement, lavage and implant removal, followed by culture guided antibiotic treatment.^{16, 18, 20–23} For infections after ASF comparable treatment has been suggested.^{8, 9} We treated our patients with removal of the instrumentation, irrigation and debridement. The same approach as the index surgery was used in case 1 and 2, in case 3 only the lumbotomy part of the incision was used. Implant removal followed by two to three weeks parenteral antibiotics, and six to ten weeks oral antibiotics. After treatment the infection was cleared in all three cases. An alternative treatment option might be suppressive antibiotic treatment, with drainage of abscesses if present. Thus evading the risks of revision surgery. Because of biofilm formation in late onset infection, eradication without implant removal cannot be expected with this strategy. The indefinite administration of antibiotics might induce bacterial resistance and in delayed infection after posterior spinal fusion, this treatment strategy showed to fail in 67% of cases.¹⁸

Complication rates for revision anterior lumbar surgery are reported to be three to five times higher as compared to primary interventions.²⁴ There is an additional risk for injury to vascular and visceral structures by the formation of scar tissue from the original exposure.²⁵ Ureter injury after revision anterior spinal surgery, as occurred in patient 3, has been reported in 0,3 to 0,5% of cases.²⁴ Removal of the instrumentation may increase the scoliotic curve. When PSF instrumentation is removed, even when solid fusion is present, more than 50 percent of cases have curve progression exceeding ten degrees.^{23, 24} In our series all patients had complete fusion at the time of implant removal, and there was no visible motion between segments upon manual intraoperative testing. Therefore, solid fusion was assumed to be present. We observed no, or only mild curve progression after CDH-system removal of 0 to 7° before a stable situation was reached.