Conservative Versus Surgical Treatment of Native Vertebral Osteomyelitis

INTRODUCTION AND EVIDENCE. Native vertebral osteomyelitis (NVO) is mainly caused by hematogenous spread from a distant infection focus (e.g. respiratory, urogenital, intestinal tract, oral cavity, skin or intravascular device), typically involving the intervertebral disc space and adjacent vertebral bodies ("spondylodiscitis"). Rarely NVO develops after percutaneous injection or intervention in the spine region without hardware implantation. Recalcitrant back pain unresponsive to conservative measures along with increased systemic inflammatory biomarkers (C-reactive protein) and/or fever or chills is the initial manifestation prompting additional investigations such as collecting blood cultures and spine imaging. Nevertheless, NVO is often diagnosed late or the diagnose is missed, which has a high impact on patient morbidity and mortality, life quality and healthcare costs (prolonged hospital stay, surgical intervention, antimicrobial therapy). Unrecognized primary infection focus causing NVO may lead to further hematogenous spread, potentially affecting other organs, joints and heart valves. The annual incidence of NVO is reported at 5.78 per 100.000 per year and is increasing in the past decades due to aging population, rise of immunocompromised patients and more invasive treatments. The incidence also increased, in part, due to improvements in diagnostic possibilities, including imaging and non-culture-based microbiological techniques. The mortality of NVO ranges between 5 and 20 %, depending on the primary infectious focus, causing pathogen and patient comorbidities.

The treatment goal of NVO is pain relief, prevention or reversion of neurologic deficits, eradication of infection, and establish spinal stability. The indication for conservative versus surgical treatment of uncomplicated NVO (no relevant abscesses, no neurologic impairment, no spine instability) are not uniform and largely depend on institutional/ regional /national preferences, traditions or random choice. A recent systematic review identified only three studies that comparing conservative to surgical treatment. All three studies are retrospective in design. The indication for surgical treatment in the first two studies was neurologic impairment, extensive bone destruction, epidural abscess formation, and failure of non-operative treatment or intractable back pain. These studies reported the complications and reoperation rates, but no formal statistical analysis was performed, in particular the long-term outcome of both treatment modalities is lacking. Other authors (Nastro et al.) treated with orthesis or spinal instrumentation and found no significant differences in both groups after 9 months. No study so far used uniform definition and stratification criteria for infection, neither used a standardized antimicrobial therapy, allowing to evaluate the outcome on the pathogen, pathogenesis, or primary focus of infection. Therefore, evidence-based data is missing to develop rational treatment guidelines for optimal treatment of NVO.

We hypothesize that the conservative treatment of uncomplicated NVO is non-inferior to surgical therapy regarding infection outcome.

AIM. The aim of this multicenter, randomized, controlled study is to compare the long-term outcome including the proportion of infection-free subjects, functional outcome and socioeconomic outcome between conservative and surgical treatment approaches providing novel high-class evidence to resolve this unknown.

MATERIAL AND METHODS. A total of 350 subjects from 8 different centers (Berlin, Munich, Dresden, Leipzig, Jena, Bonn, Magdeburg, Aachen, Cologne, Hamburg) will be screened. With 8 participating study sites, 200 patients are expected to be randomized per year. With a recruitment period of two years (i.e. 400 patients) and a dropout rate of 15%, 140 patients could be analyzed regarding the primary endpoint at 12 months.

Included patients will be randomized 1:1 into 2 arms: surgical arm and conservative arm. Randomization should avoid selection bias for each group. Allocation of treatments will be performed by a computer-generated list. Randomization will be stratified by center. To ensure balanced group sample sizes in the course of patient recruitment, blockwise randomization with varying block sizes will be conducted within the centers. An independent review committee reviews possible failure cases.

1. SURGICAL ARM: Surgical treatment of native vertebral osteomyelitis consisting of debridement, spinal decompression and instrumentation using hardware. Additional standard antimicrobial therapy, according to the pathogen and its antimicrobial susceptibility (intravenous route during first 2 weeks, followed by oral route for additional 4 weeks).
2. CONSERVATIVE ARM: No surgical intervention, antimicrobial therapy only, bed rest or temporary external immobilization of affected spinal segment with orthopedic orthesis. Antimicrobial therapy will be administrated according to the pathogen type and antimicrobial susceptibility, with initial intravenous administration for 2 weeks, followed by oral antibiotics for 4 weeks to complete 6 weeks of total therapy.

Follow-up will be performed 3, 6 and 12 months after surgery (experimental group) or start of antimicrobial therapy (control group).

STATISTICAL ANALYSIS. For analysis of the primary endpoint (proportion of infection-free subjects in both intervention groups one year after treatment initiation), a continuity corrected chi-squared test will be performed at significance of α=0.05. The Log-rank test (two-sided) will be performed and visualized by Kaplan-Meier method. Potential confounders will be considered by using a Cox proportional-hazards regression model. The primary analysis will be conducted following the intention-to-treat (ITT) principle. The number of patients lost to follow up is assumed to be about 15% and will be replaced. Secondary analyses based on the per-protocol and the as-treated participants will be performed. A logistic regression model will be fit to identify prognostic factors for treatment failure by application of resampling techniques for variable selection. We expect a compliance of taking oral antibiotics as prescribed of about 80%. The incompliance is expected to be equally distributed between both groups and should therefore not represent a bias.

ETHICAL CONSIDERATIONS. As both treatment options are considered part of daily practice, a comparative study between both treatment options should be considered as ethically justifiable. However, ethical concerns in both directions are possible. The surgical treatment may be associated with additional surgical-related complications, whereas the conservative treatment group (antimicrobial treatment alone) may be associated with higher treatment failure rate. The Principle Investigator and the Data Monitoring and Safety Board (DMSB) will analyze the intermediate reports and take appropriate actions, if necessary.

STRATEGIES FOR DATA HANDLING. An electronic Case Report Form (eCRF) will be used for data collection. Regular back-up of the database will be performed. The study database is protected by an authentication method. The database will be accessible anytime to the monitor organization and Data Management and Safety Board (DMSB). Data protection is guaranteed by study software: this generates automatically for each patient a pseudonym, consisting of a combination of alphanumeric characters. All study data will be saved in a pseudonymized way. The patient identification list will be stored only locally by each study center.

TRIAL EXPERTISE. The interdisciplinary spine center at Charité has developed into a leading orthopedic and neurosurgical spine centers, certified by the German Spine Society (DWG). The high-volume clinical workload (>10.000 spinal surgeries in the past 5 years) is utilized for extensive training and research, in particular in the field of diagnosis and treatment of spinal infection. Several multicenter IIT´s were initiated by the interdisciplinary septic team.

TRIAL-SUPPORTING FACILITIES. The support will be provided by the coordinating Center for clinical trials, a member of the German network of coordinating centers for clinical trials (KKS). In addition, the trial is supported by the spine section of the German association of Neurosurgery.