Primary pyogenic infection of the spine in intravenous drug users: a prospective observational study

A prospective comparative analysis of surgically-treated nontuberculous primary pyogenic infection of the spine (PPIS).

To evaluate and compare the demographics, presentation, treatment and outcomes of surgically-treated PPIS between intravenous drug users (IVDU) and non-IVDU patients.

This is the first prospective cohort comparative analysis of the clinical outcomes of surgically-treated PPIS between IVDU and non-IVDU populations. This is also the largest cohort of consecutive surgically-treated PPIS in a population of IVDU.

Data on all patients with PPIS presenting to a quaternary referral center during a 4-year period from 2004 were collected in a prospectively maintained customized database.

During the study period there were 102 patients treated for PPIS of which 51 were IVDU. Of this IVDU group, the mean age was 43 years (range: 25-57). Twenty-three had human immunodeficiency virus, 43 had hepatitis C, and 13 had hepatitis B. All were using cocaine, 26 were also using heroin and 44 were using at least 3 recreational drugs. Thirty patients presented with axial pain of a mean duration of 51 days (range: 3-120 days). Of the IVDU patients with neurological deficit on presentation, the mean American Spinal Injury Association (ASIA) motor score was 58.6. The most common ASIA motor levels involved were C4 and C5. Mean duration of neurological symptoms was 7 days (range: 1-60 days). Twenty-six were already receiving i.v. antibiotics for known spinal infection and 33 patients had an identifiable organism on blood cultures (19 methicillin-sensitive Staphylococcus aureus, 9 methicillin-resistant S. aureus). Forty-four of the 51 IVDU patients were treated surgically. Thirty-four of the 44 surgically treated cases involved the cervical spine. Twenty-two had a posterior approach alone, 13 had anterior only while 9 required combined anterior and posterior approaches at the index surgery. Seven required early revision for hardware failure (none of whom has combined approach) and 2 developed a postoperative surgical site infection (SSI). Thirty-seven of the 51 IVDU patients were apprehended at least once using illicit drugs while in hospital. Mean duration of antibiotic treatment after surgery was 62 days. At discharge, 28 of 44 patients had neurological improvement (mean = 20 ASIA points, range: 1-55), 11 had neurological deterioration during treatment (mean = 13, range = 1-50), and 5 were unchanged. Among the IVDUs there were no in-hospital deaths. At 2 years after index admission 13 IVDU patients were dead, and none were attending for follow-up despite all efforts to locate the patients. In the non-IVDU group, the mean age of the 51 patients was 56 years (range 25-83). Thirty-four patients presented with axial pain with a mean duration of 105 days (range 2-365). Mean ASIA motor score of patients with neurological deficit on admission was 74. Most common ASIA level was T12. Mean duration of neurological symptoms was 12 days (range 1-84). Thirteen patients were receiving i.v. antibiotics for known spinal infection and 20 patients had an identifiable organism on blood culture (30% methicillin-sensitive S. aureus, 50% methicillin-resistant S. aureus). Forty-four of these 51 non-IVDU patients were treated surgically. Thirty-five of the 44 surgically-treated cases involved the thoracic or lumbar spines. Twenty-nine had a posterior approach alone, 3 had anterior alone while 12 required combined approaches. No early hardware failures were seen in the non-IVDU group while 4 developed SSI. Mean duration of antibiotic treatment after surgery was 45 days. At discharge 21 patients had neurological improvement (mean 9 ASIA points, range: 1-17). Five had neurologic deterioration with a mean motor loss of 16 points (10-23). There were 4 in-hospital deaths among the non-IVDU group. At 2 years after index admission, 19 patients were dead and the remainder were all available for follow-up.

There are significant differences in demographics, presentation, treatment and outcomes of primary spinal pyogenic infection between a population of IVDU and a comparable cohort of non-IVDU. The IVDU group presents with cervical quadriplegia while it is the thoracolumbar spine that is almost exclusively involved in the non-IVDU group. Among the IVDUs, surgical management is complex with a high incidence of early hardware failure. SSI is significantly more common among non-IVDU. Significant neurological improvement can be expected in the majority of IVDU patients with a high mortality rate among the non-IVDU. IVDU are unreliable patients and in-hospital, in-halo incarceration is recommended where possible.