Protocol of a short post-surgical antibiotic therapy in spine infections - prospective, randomized, unblinded, non-inferiority trials(SASI trials)

Background There are several open scientific questions regarding the optimal antibiotic treatment of spine infections (SI) with or without an implant. The duration of post-surgical antibiotic therapy is debated.Methods We will perform and perform two unblinded randomized -controlled RCTs. We hypothesize that shorter durations of systemic antibiotic therapy after surgery for SI are non-inferior (10% margin, 80% power, ɑlpha 5%) to existing (long) treatment durations. The RCTs allocate the participants in two arms of 2 x 59 episodes each: 3 vs. 6 weeks of targeted post-surgical systemic antibiotic therapy for implant-free spine infections (two positive microbiological samples); or 6 vs. 12 weeks for implant-related spine infections. This equals a total of 236 adult SI episodes (randomization schemes 1:1) with a minimal follow-up of 12 months. All participants have a concomitant multidisciplinary surgical, re-educational, internist and infectious diseases care. We perform three interim analyses that are evaluated, in a blinded analysis, by an independent Study Data Monitoring Committee. Besides the primary outcome remission, we also assess adverse events of antibiotic therapy, changes of the patient’s nutritional status, the influence of immune suppression, total costs, functional scores, and the timely evolution of the (surgical) wounds. We define infection as the presence of local signs of inflammation (pus, wound discharge, calor, rubor) together with microbiological evidence of the same pathogen(s) in at least two intraoperative samples; and remission as absence of clinical, laboratory and/or radiological evidence of (former or new) infection.Discussion Provided that there is adequate surgical debridement, both RCTs enable to potentially prescribe less antibiotics during the therapy of SI; with potentially less adverse events and reduced overall costs.


Background and rationale
Surgical site infections are feared complications of spine surgery of which the volume is expected to increase every year; worldwide [1]. Likewise, community-acquired spine infections (SI) are associated with enhanced morbidity; costs and prolonged hospital stay for the patients [1]. Most scientific papers are rather interested in the epidemiology of SI and risk factors for surgical site infections after spine women and their children. Additionally, the study population is likely not to reveal women at procreating age. Thus, pregnant and breast-feeding women are not excluded. The investigators will avoid agents that are not liberated for pregnant or breast-feeding women, according to the Swiss Compendium (www.compendium.ch).

Outcomes of interest
For the RCTs and BioBanking, we will collect data and biological material. Concerning the randomized trails, Table 1 (bottom) summarizes the outcome parameters. Regarding the investigation of the dynamic changes of the nutritional status during SI care, specialist nutrition nurses will assess the status at baseline and the end of treatment. In case of severe malnutrition, they are allowed to propose corrective measures already during the SI therapy, because it would be unethical not to intervene only because of study purposes. Finally, the database will be sufficient large to estimate the influence of an underlying chronic immune-suppression (i.e. diabetes mellitus, chronic steroid therapy, dialysis, untreated HIV disease, active cancer in therapeutic or palliative treatment, cirrhosis CHILD C) on SI outcomes and related nutritional status. We also reminder that patients with very severe iatrogenic immune suppression, such as recent solid organ or bone transplantations in the last five years, are exempted from the SASI trials ( Figure 1).

Allocation and timetable
After written informed consent will been given to participants (until Day 5 of debridement), the unblinded allocation occurs electronically with a randomization scheme of 1:1 (randomization without blocked or matched variables). The study nurse of the Unit for Clinical and Applied research and/or the co-investigators will implement the allocation sequence into the trial. For both RCTs, we need 36 months of study time; starting in August 2019. Table 3 highlights some key time events.

Statistical analyses and sample size
Both RCT are non-inferiority trials. Remission incidence (at the first attempt of therapy) is set at 5% (5% recurrence in both arms). The clinically maximum acceptable difference (unidirectional noninferiority margin with binary-outcome categorical variables) is arbitrarily fixed at 10% regarding the primary outcome remission [1]. Assuming a risk of alpha at 0.05 and a power of 80%, it will be necessary to recruit 59 patients in each antibiotic duration arm (short or long). Together with the distinction of the RCT into implant-related and implant-free SI, we would finally need 2 x 2 x 59 episodes, equalling a total of 236 SI episodes within three years. For assessment the formal noninferiority requirement (regarding the primary outcome "remission"), we will compute with a unidirectional p-value limit of 0.025. We do not predefine a non-inferiority margin for secondary outcomes such as costs, adverse events, functional outcomes, underlying immune suppression, dynamic changes in the nutrition status and BioBanking.

Interim analyses
When the first 20 episodes of any randomization branch will have a complete follow-up, and again 60 and 120 SI episodes, we perform three interim analyses. On this occasion, we equally check if the expected statistical power for the final analysis will be acceptable. If it is lower than 30%, we will consider the trial will not be able to demonstrate the result, and the recruitment is no more ethical.
The most frequent conditional power evaluated under the current trend (i.e. using the information from the collected data) will be assessed [7,8]. The Study Data Monitoring Committee will consist of independent surgeons or physicians, with clinical and statistical experience, not participating in the study. They will decide about the future of the trial, entirely or partially, after each of the three interim analyses. The PI and the Sponsor will present the data in a blinded form to the Data Monitoring Committee. Their members will only know if there is an implant, but ignore allocations to the antibiotic arms.
The intent-to-treat (ITT) population will consist of all randomized patients who signed for the participation. Patients will be analysed according to treatment group assignment regardless of whether the patient receives any treatment or the wrong treatment or is lost to follow-up. The perprotocol (PP) population will consist of all patients who complete the study and who have not deviated significantly from the protocol. The statistical analyses will mostly base on descriptive analyses, group comparisons and a multivariate, unmatched, eventually cluster-controlled, Cox regression analysis adjusting for the large case-mix that we expect. Equally, a Generalized Estimation Equations (GEE) model might adjust for clustering in case of multicentre origin of the patients. The Biostatistician will analyze the datasets in a blinded form (as group A or B), but the PI, the Study Nurses, and the Sponsor will ultimately unblind the allocations for data verification and definition of the ITT and PP populations.

Study registration, ethical conduct and categorization
The study is approved by the Ethical Committee of Zurich (no. 2019-00646) and registered in the

Patient Information and Informed Consent
Participants will be recruited by any of the investigators of the study. Our institution has a standardized procedure for recruiting participants as participant studies are common. Each participant will be informed that the participation in the study is completely voluntary and that he/she may withdraw from the study at any time and that withdrawal of consent will not affect his/her medical assistance and treatment in the future. All participants of the study will be provided a Participant Information Sheet and Informed Consent Form entailing sufficient information. For the BioBank, the participants will sign the General Consent for the further use of personal data and biologic material. The investigators affirm and uphold the principle of the participant's right to privacy and that they shall comply with applicable privacy laws and/or the corresponding section of the study specific consent.

Monitoring
The Unit for Clinical and Applied Research of Balgrist University Hospital will assign an independent monitor. Regular monitoring visits at the investigator's site prior to the start and twice during the course of the study will help to follow up the progress, to assure utmost accuracy of the data and to detect possible errors at an early time point. The monitor will review all or a part of the Case Report Forms (CRF) and written informed consents. The accuracy of the data will be verified by reviewing the above referenced documents. There will be a close-out visit at the study end. During the monitoring, all documents including source data/documents will be accessible for the monitor.

Audits and Inspections
An audit/inspection of this study may be conducted by the competent authority. The quality assurance auditor/inspector could have access to all medical records, the investigator's study related files and correspondence, and the informed consent documentation that is relevant to this clinical study. The investigator will allow the persons being responsible for the audit or the inspection to have access to the source data/ documents and to answer any questions arising. All involved parties will keep the patient data strictly confidential.

Early termination of the study (participation)
The investigators may terminate the study prematurely according to certain circumstances, for example: ethical concerns, insufficient participant recruitment, when the safety of the participants is doubtful or at risk, respectively, alterations in accepted clinical practice that make the continuation of a clinical trial unwise, early evidence of benefit or harm of the experimental intervention. If a patient is withdrawn, the reason will be noted. When possible, evaluations required at the next scheduled visit will be performed at early termination.

Treatment by specialists
All surgeries will be performed in the supervision and participation of an experienced spine surgeon.
The antibiotic therapy is ordered and supervised by internists and infectious diseases physicians with therapeutic and academic experience in SI treatments. The current medications of the study patients, as well as possible interactions, will be controlled by the internists several times a week during hospitalization.

Definition and assessment of (serious) adverse events and other safety related events
An Adverse Event (AE) is any untoward medical occurrence in a patient, and which does not necessarily have a causal relationship with the study procedure. A Serious Adverse Event (SAE) is classified as any untoward medical occurrence that: results in death, is life-threatening, requires inpatient hospitalization or prolongation of existing hospitalization, results in persistent or significant disability/incapacity. In addition, important medical events that may not be immediately lifethreatening or result in death, but may jeopardize the patient or may require intervention to prevent one of the other outcomes listed above should also usually be considered serious. Participants with ongoing SAEs at study termination will be further followed up until recovery or until stabilization of the disease after termination. The investigators will make a causality assessment of the event to the study. All SAEs must be reported immediately and within a maximum of 24 hours to the Sponsor-Investigator of the study. SAEs resulting in death are reported to the local Ethics Committee (via local Investigator) within 7 days. Patients with adverse events, and leaving the study, will be treated offstudy, without restriction, at the study site.

Case Report Forms, procedure of data analysis and BioBank archiving
An electronic CRF will be generated for every patient. All relevant study data are recorded by authorized persons in the REDCap ® electronic data capture tool [9] and archived for a minimum of 10 years. Participating patients will be registered in an enrolment log assigning the participant to his/her study ID. Corrections can only be made by the authorized persons. For data analysis, subject-related data from REDCap will be exported and analyzed in statistics software (IBM -SPSS and/or STATA).
Before data export, all patient identifiers will be removed. Patient-source and BioBank data will be registered using subject identifiers. Collection, disclosure, storage of patient-related data is carried out in accordance with Swiss data protection regulations and the Human Research Act. The BioBank will store the intraoperative tissue samples in accordance with their guidelines. Likewise, radiological data are stored in the PACS system according to the standard at the Balgrist University Hospital.

Theoretical risk of the study
Besides the retrospective identification of patients, we do not see any particular risk for the patients regarding the cohort. For BioBanking specifically; a theoretical additional risk could be the detection of unknown pathologies, if there would be a further work-up of the intraoperative samples.
Concerning the RCTs, a theoretical risk could be a higher incidence of recurrences in the corresponding short antibiotic arms.

Discussion
Our cohort with two embedded RCTs seeks to demonstrate a clinically relevant non-inferiority of a shorter systemic antibiotic treatment in adult SI patients; with and without implants [1]; and independently of the surgical drainage technique, the number of debridement, underlying individual chronic immune suppression, the infection localization or the pathogens. Importantly, all study participants have an accompanying multidisciplinary surgical, re-educational, internist and infectious diseases treatment and follow-up. We equally collect intraoperative soft tissues and bone for future (laboratory) studies and assess adverse events, overall costs, functional outcomes and the dynamic changes in the nutritional status of the infected patients; with relation to their therapy and outcome.
The studies start in Zurich, but are expendable to other study centers with experience in treating SI.
The primary outcome is "remission at the last follow-up", but the RCT enable to adjust for different important variables such as the number of surgical debridement, the use of a negative pressure therapy, administraton of a parenteral antibiotic regimen or the total duration of antibiotic therapy. As in many field of septic orthopedic surgery, the number of surgical debridement does not formally influence remission rates, which has been shown for chronic osteomyelitis [10], septic native joint arthritis [11], fracture device infections [12], infected open fractures [11], or prosthetic joint infections [14]. There is very little evidence to guide surgical treatment of patients who require a  [15].
Certainly, the most important variables retrieved form our trials will be antibiotic-related. Most author groups advocate a minimum length of parenteral antibiotic courses of 2-4 weeks and a total duration up to three months [16,17] for SI, although some groups only recommend two weeks of parental therapy [18,19], or even only 2-3 days [20], without further compromising the success. To cite examples, Clark and Shufflebarger treated delayed infections with surgery and 48-72 hours of parenteral antibiotics followed by ten days of targeted oral antibiotics. All infections were eradicated [21]. Likewise, Richards and Emara prescribed systemic antimicrobials only for three weeks, of which 2-5 days parenterally, followed by a 7 to 14 day-course of oral treatment [22].
In the entire field of "orthopedic infections", there are no formal scientific data proving the benefit of a systemic antibiotic therapy beyond six weeks; compared to four to six weeks or even less.
Exceptions are by nature expert opinions in previous book chapters or past publications without own database analyses; or the therapy of special microorganisms requiring long-lasting antibiotic therapies such as mycobacteria [23], Nocardia spp. [24], actinomyces or fungi [25]. To cite recent and own examples of investigations regarding the overall antibiotic duration, sacral osteomyelitis [26], long bone osteomyelitis [10], fracture-device-related infections [12], spondylodiscitis [26], prosthetic joint infections [14], diabetic foot osteomyelitis [27,28] and many more failed to enhance remission rates, if antibiotics were prolonged beyond four to six weeks; even in presence of an infected implant.
Farhad et al. resumed that six weeks of antibiotic therapy was sufficient for all bone-related infections [14]; together with an early switch of oral medication [6]. These emerging and relatively short durations are equally acknowledged by international consensus meetings [30] of surgeons and infectious diseases physicians, who treat these infections and who perform research on them.
There are also studies with less than six weeks of total antimicrobial therapy, especially in the pediatric literature for hematogenous osteomyelitis. In this particular setting, a three-week antibiotic course appears to be sufficient as highlighted by many authors [31][32][33][34]. For adults, 38 case series with antibiotic treatment durations of 3 to 4 weeks, including 5 to 36 patients each, revealed cure rates of approximately 80% according to a review published in 2005 [35].
A second issue is the distinction between intravenous and oral antibiotic administration; at least initially. Current textbooks recommend the parenteral route for at least the first two weeks for all osteoarticular infections [1, [36][37][38], but this recommendation is not evidence-based either. There are no predictive clinical markers that would justify prolonged initial intravenous administration. In addition, up to one-third of patients with chronic bone and implant infections may experience antibiotic-related or catheter-related problems during parenteral treatment [38]. For economic reasons, as well as patient and nurse comfort, parenteral administration should be kept to a minimum [39]. Good bone penetration during parenteral and oral administration has been proven in several reports [40][41][42] and data suggest that an early switch to oral antibiotics is as effective as prolonged parenteral regimens [43].
A Cochrane review investigated five trials comparing oral vs. parenteral antibiotics in osteomyelitis.
There was no statistically significant difference between the two groups in the remission rate twelve months or more after treatment [44]. Glassman et al. successfully treated two patients with SI with oral ciprofloxacin from the start, an antibiotic with excellent oral bioavailability and bone penetration [45]. Even in cases of diabetic foot osteomyelitis, a frequent disease with its hallmark of vascular insufficiency and tissue ischemia, there are no data indicating the superiority of any particular route of delivery of systemic antibiotics [46]. Byren et al. demonstrated that an intravenous course of antibiotics for over four weeks did not enhance cure for the treatment of arthroplasty infections [47]. Zimmerli et al. summarized observational studies that showed the same failure rates of arthroplasty infection treatment despite a prolonged (four to six weeks) period of intravenous treatment [48]. For the treatment of bone infections, there are some antibiotics that have already proven to be effective in oral form. Quinolones, rifampicin, co-trimoxazol, tetracycline or clindamycin have such a good and sufficient oral bio-availability [49].
Our future patient population will comprise all co-morbidities and chronic immune suppressions. For example, we expect 20-25% diabetic patients [50] in our center, along with other immune suppressions such as cancer, advanced cirrhosis and steroid medication. While immune suppression (especially diabetes mellitus) is an acknowledged independent risk associated with healthcareassociated surgical site infections [51], its influence on the remission during a therapy for SI is unknown. Indeed, all current therapeutic concepts for osteoarticular infections in general, do not relay on the presence or absence of immune suppression [1,10, 12,26]; suggesting that the surgical debridement and the long antibiotic administration overcome eventual shortcomings of the patients' immunity. Although our SASI trials do not target the association of immune suppression with SI outcomes, we will see if immune suppression tends to decrease remission when we shorten the antibiotic duration.
Finally, our RCT will also give insight in the nutritional status of the infected spine patients. Current literature is departed between experts advocating a causal relationship between malnutrition and occurrence of surgical site infections in orthopedic surgery, while others have retrospectively investigated this relationship and mostly found no associations [52]. Both fractions know even less about the associations and the dynamics of the nutritional status in already infected orthopedic spine patients, and the association of these alterations with remission, functional outcomes and underlying immune suppression; let alone the question of the benefits of nutritional interventions during the combined surgical, physiotherapeutic and antibiotic treatment [53]. This will be terra nova that we embed into our trials.
We do not expect major difficulties performing our studies. Despite two prospective-randomized designs (for SI with and without implants) and only 236 different episodes anticipated, the patients' voluntary participation might be low. Likewise, patients who are continued to be treated outside of our center may have been lost to our follow-up or have their treatment changed, because the following physicians do not agree. However, our center is the largest public hospital for surgical SIs in the region, and it is the University Spine Center; so this is unlikely to be a major bias. Lastly and formally, our study participants will benefit from an initializing surgical debridement of infections.
Hence, our results will not be valid for the conservative treatment of SI, which must not be confounded with.

Ethics approval and consent to participate
The study protocol was submitted for approval to the Cantonal Ethical Commission of Zurich,   Study Flowchart