There is insufficient evidence to support a strong or moderate strength recommendation regarding antibiotic prophylaxis in pTBI.

Prophylactic antibiotics in patients with pTBI can be considered.

No specific antibiotic regimen or duration of treatment has demonstrated superiority for infection prophylaxis in pTBI patients as compared with others.

In the absence of direct scientific evidence, EXPERT CONSENSUS concluded that:

• It is reasonable to treat pTBI patients who have violation of the dura or who are undergoing invasive procedures with prophylactic antibiotics (87.5% consensus).
• It is reasonable to treat patients with pTBI related to blasts, air sinus penetration, retained bone fragments, grossly contaminated objects with prophylactic antibiotics as these are at higher risk of infections (95.8% consensus).
• in pTBI patients at higher risk of intracranial infections it is reasonable to administer broad-spectrum antibiotic prophylaxis (95.8% consensus) including anaerobic coverage (81.8% consensus) for duration of up to 5 days (91% consensus).
• In patients who are NOT at higher risk of intracranial infections it is reasonable to administer antibiotic prophylaxis (87.5% consensus) for duration of up to 3 days (100% consensus).
• When prophylactic antibiotics are administered, they should be administered as soon as possible after injury (100% consensus).

No evidence or expert opinion supported distinct recommendations based on patient gender, age, wounding mechanism, or military vs. civilian context.

Intracranial infection following pTBI is not uncommon and carry significant morbidity.1 There is still a scarcity of data regarding the utility of prophylactic antibiotics for preventing subsequent intracranial infections and informing the optimal timing for the initial administration of these antibiotics. Historically, infection rates were high, and therefore led to routine prophylactic use of antibiotics in such patients, with the belief that antibiotics were 'essential' to prevent intracranial infections. However, with improving surgical debridement techniques, rates of infection have been decreasing overall since earlier studies from the world wars. Moreover, bacteria isolated from pTBI-related infections are different from those obtained when surveillance wound swabs are performed at the time of an initial surgery. In line with this, an expanding body of evidence and experience has begun to question this belief, offering insights into which patients may face a higher risk and challenging the universal application of prophylactic antibiotics.

The risk of post-pTBI intracranial infection can range from 7% in civilian populations to as high as 25% in military populations.2,3 Additionally, complicating the overall infection rate in these patients is a large number who go on to have invasive surgical procedures which can also lead to increased infection rates, making it unclear whether or not prophylactic antibiotics help prevent infections.4 There have been no randomized controlled trials evaluating the use of prophylactic antibiotics and the best time for initiation.

Physicians treating pTBI have routinely given prophylactic antibiotics aiming to prevent the increased morbidity and mortality associated with intracranial infections. Historically, this represented broad spectrum of coverage against gram positive, gram negative, and anaerobic bacteria. This has often involved giving cefazolin 2gm IV every six to eight hours with the addition of metronidazole if there is clear contamination. (CPG JTS). Additionally, the time course of the antibiotics and when to starts has been incredibly variable throughout the literature ranging from one to five days. Additionally, multiple studies have evaluated the risks factors for intracranial infection after pTBI and found that continued CSF leak and operative procedures were placing patients at the highest risk of infection.2,4 With rising rates of antimicrobial resistance within communities globally, it is also importance to be good stewards in the use of antibiotic prophylaxis.

In the prior edition of the guidelines (2001), an 'optional' recommendation (equivalent to a current Level III/IV) was made supporting the prophylactic use of antibiotics in patients with pTBI.5 Recommendations created in this first edition were based on 14 articles defined as Class III. Level III/IV recommendations stated that the use of prophylactic broad-spectrum antibiotics is recommended for patients with penetrating brain injury.

Data supporting this recommendation noted increased rates of infection in pTBI patients with associated CSF leak, air sinus penetration, or retained bone fragments. Use of prophylactic antibiotics for patients with pTBI was an option based on the decrease in infection rate after the onset of routine antibiotic use for these injuries and extrapolation from data on clean, elective neurosurgical cases. In the literature examined for these recommendations (1966-2000), routine use of prophylactic antibiotics and absence of direct comparative studies led to comparison of these data with data from World War I wherein routine infection rate in pTBI patients was 58.5%.1 However, in comparison since, the risk of cranial infections has been found to be significantly lower, thought to be due to less contaminated wounds, decreased time to medical care and improved intensity of care. In addition, since all the studies included use of antibiotics and thus did not provide direct evidence in the prior guidelines, they were not included in this edition.

There is scant evidence of benefit from routine use of prophylactic antibiotics in patients with pTBI. We identified no clinical trials that evaluate the use of prophylactic antibiotics, antibiotic regimen, or duration of antibiotics in this setting. The studies that were included are in civilian populations and are summarized below in the table. Seven studies are described for use of prophylactic antibiotics, while one study describes duration of use.

The largest study included (Harmon et al) was a retrospective review of 11 years (2006 to 2016) of records from 17 trauma centers in the United States that identified 763 adult patients with penetrating brain injuries, including 62% (N=475) who were treated with antibiotics within 24 hours of admission.2 The overall rate of central nervous system infections was 7% and this was not statistically significantly different for patients who did or did not receive antibiotics. Another large study of 192 non missile penetrating head injuries treated in one medical center in South Africa published by Harrington et al focused on categorizing management and complications of these injuries and how they differed from gunshot wounds.6 The authors reported that 33% of the patients received no antibiotics, 52% received antibiotics in less than 24 hours from the time of injury, and 15% received antibiotics more than 24 hours after injury. The rate of infections was significantly lower among patients who received antibiotics at any time compared with those who did not (8% vs. 26%, p=0.002) in both univariate and adjusted models. However, the characteristics of the patients who did and did not receive antibiotics and the variables used in the adjusted model were not reported, as this was not the focus of the study.

Three studies were limited to patients with gunshot wounds: Johnson et al included 75 patients over 16 years old treated in a single U.S. civilian hospital for gunshot wounds to the head, surviving over 48 hours are compared them based on antibiotic regimens: no antibiotics, a single agent, or multiple antibiotics.4 In cases from 6 years (2012 to 2018), there were only a total of 8 (10.7%) intracranial infections documented over 1 week to 3 years from injury. There were more infections in patients who received the multiple antibiotics in unadjusted analyses, but in the adjusted analysis, neither the antibiotic group (single antibiotic or multiple antibiotics) nor duration of antibiotic prophylaxis were statistically significant. There were also 13 nosocomial infections reported in this study, but these were also not different across the antibiotic groups and were not associated with later intracranial infections. A study in Colombia followed 160 adult patients with gunshot wounds to the head who did or did not receive prophylactic antibiotics at the discretion of the treating neurosurgeon.3 A higher proportion of the patients who received antibiotics (33.9%) developed infections than those who did not (19.8%), but this difference was not statistically significant, and receipt of antibiotics was not included in multivariate analysis of risk factors for infection. Another study by Doherty et al examined patients treated for gunshot wounds to the head in a single trauma center in the United States. It identified 59 patients who survived more than 48 hours.7 Thirty-eight (64%) were given systemic antibiotics within 36 hours of admission; 26 as preparation for a surgical procedure and 12 without surgery. Twenty-nine percent of these patients developed infections (N=11), while none of the patients who did not receive antibiotics developed infections. Two further studies examined routine use of prophylactic antibiotics: Meister et al8 included 106 military personnel or dependents with pTBI or an open skull fracture while deployed in Iraq or Afghanistan between 2009-2014 and who were treated in US hospitals. Of these only 2 did not receive early (<48 hours from injury) prophylactic antibiotics. Infection rates were 11.3% with a total of 12 infections, all of which occurred amongst those receiving antibiotics. Harrington et al6 provide data from cranial stab wounds found predominantly in young males with a variety of weapons ranging from knives (81%), screwdrivers (9%) to garden forks etc. with 18% having retained weapon. No antibiotics were administered to 33% (64), in less than 24 hours in 52% (99) and later than 24 hours in 15% (29); intracranial infections were seen in 8% of those who received antibiotics at any time and 26% in those who did not receive any antibiotics, with an adjusted odds ratio of 0.3 (95% CI 0.1-0.8) in favor of antibiotics. This study included a majority of penetrating injuries from knives which tend to occur infrequently in clinical practice and may differ from other penetrating injuries due to contamination of knives.

In regard to extracranial infections, a non-randomized study by Marut et al reviewed charts of patients with penetrating brain injuries treated at a single Level 1 trauma center in Northeast Ohio (U.S.) from December 1, 2015 through July 31, 2018.9 Patients were included if they were over 16 years old and were not transferred to a pediatric institution or transferred from another facility 24 hours after presentation. The primary objective was to determine what proportion of patients received prophylactic antibiotics to inform quality improvement. Thirty-three patients were included and 24 (73%) received antibiotics with a median time to administration of 52.8 minutes from admission and a median duration of 24 hours. There were no nervous system, skin, or soft tissue infections during the hospital admission, but four patients (12%) were treated for pneumonia. Thirty-three percent of the patients died. Neither the pneumonia cases nor deaths are reported by whether the patient received an antibiotic or not.

Etheridge et al provided the one study on the duration of antibiotics that was identified, used a registry from one trauma center to identify a small sample (N=22) of adult patients with penetrating gunshot wounds to the head who received short-term prophylactic antibiotics, for 48 hours or less (N=9) or extended treatment for over 48 hours (N=11).10 In follow-up that lasted for a mean of 15 months, no infections occurred in the patients with short-term treatment and four patients (30.8%) developed infections in the extended treatment group. One patient in each group died within 30-days of any cause and neither infection nor mortality were statistically significantly different, however, the study was not sufficiently powered to detect a difference in infection rates.

Studies not included in evidence synthesis: Below are several case series that did not meet criteria for inclusion in evidentiary tables but are described:

Jankovic (1997) and Hecimovic (2000) and Tudor (2005) report case series from the Croatian War (1991 to 1994).11-13 In all three, the authors state all patients were given prophylactic antibiotics as part of initial emergency treatment, according to policy, but few additional details are provided and no information is given on patients who did not receive antibiotics. One study (N=138) reviewed the cases of Croatian soldiers treated for penetrating intracranial injury that focused on the role of computer tomography in diagnosis and treatment.11 In the 138 soldiers given prophylactic antibiotics, nine (6.5%) developed infections, including eight with meningitis cerebritis and one abscess; four of whom died. In another study (N=88) of patients treated for missile brain wounds at a university hospital, 15 (17%) developed infections: meningitis alone (4 patients), abscess alone (3 patients), cerebritis alone (1 patient), meningitis plus abscess (4 patients), and one patient each developed meningitis plus subdural empyema, meningitis plus abscess plus osteomyelitis, and meningitis plus abscess plus epidural empyema.11 Eight of the 15 patients died (53.3% of those who developed infections) and two additional patients had poor outcomes (Glasgow Outcomes Scale score of 3). In 176 patients treated for penetrating craniocerebral injuries at a different university hospital where all were given routine antibiotic therapy on admission, 28 patients (15.9%) developed infections; 13 were superficial and 15 (8.52%) were central nervous system infections (meningitis, meningoencephalitis or ventriculitis).12 Reoperations were required for 60% of the patients with central nervous system infections.

The two additional case series were from the Iran-Iraq war. One study (N=125) compared patients treated from April 1983 to November 1984 based on whether there was operative exploration, then if primary debridement was definitive or if a secondary exploration was needed.14 All patients were started on antibiotics at the casualty clearing station, seven developed infections; six meningitis, one a brain abscess, and one a brain abscess and meningitis. Mortality for patients with infections was 23% (2 of 7). None of the patients with positive early cultures of the wound, bone, or brain developed infections. In a study of patients at the same hospital from February 1981 to August 1987, 397 patients with dural penetrating missile head wounds were followed after primary debridement closure who had antibiotics started at the frontline and continued for 10 days after surgery.15 Eighteen patients developed infections; 16 patients developed meningitis, one patient an abscess, and one patient developed both meningitis and an abscess. Twelve of these 18 patients had cerebrospinal fluid fistulas. Seven (38.8%) of the patients with infections died, one was lost to follow-up, and two had minimal, six moderate, and two with severe disability at follow-up (23.7 ±11 months).

One case series reported 32 missile injuries to the brain were treated between 1981 and 1988 during the Lebanese conflict.16 Patients were selected who were neurologically stable with a Glasgow Comas Scale score greater than 10, examined within 6 hours of the time of injury, with an entry wound less than 2 cm, no exit wound, missile tract that did not cross the proximal sylvian fissure, no epidural or subdural hematoma, and no intracerebral hematoma that enlarged or measured more than 4 cm. Patients were given 1 g of methicillin on arrival and continued for 14 days. Their wound was debrided and closed in the emergency department without drainage. One patient (3.1%) developed a brain abscess indicated by a seizure and left hemiparesis 20 days after injury. The abscess was excised, treatment included three weeks of cefotaxime and the patient recovered.

In another case series (N=82) of patients with penetrating gunshot wounds treated at an urban U.S. hospital from 1973 to 1975, all patients received antibiotics (oxacillin 2 g, ampicillin 2 gm) with resuscitative measures on admission.17 Thirty-seven patients did not undergo an operation, of whom 34 died, and three patients with more minor injuries recovered. The 45 patients treated with surgery were the subject of infection surveillance: two developed meningitis and none had brain abscesses. Two patients had scalp infections and 13 patients with retained fragments had no infections during the 3.5-year follow-up period.

Together these studies provide low strength of evidence that routine prophylactic antibiotic administration following pTBI make no significant difference in infection rates compared with no antibiotics.

The use of prophylactic antibiotics has long been thought to potentially be beneficial after pTBI in order to prevent intracranial infections due to dural penetration, even though there has been lack of evidence to support this. Furthermore studies have shown that early bacteriological swab isolates do not correlate with bacteria causing subsequent intracranial infections. Despite the significant morbidity and mortality associated with pTBI and the consequent infections, current evidence is insufficient to provide a strong recommendation for the routine use of prophylactic antibiotics.

In the above studies there is little evidence to support the routine use of antibiotics following pTBI. However, the landscape of pTBI has also changed significantly, from highly contaminated wound in WW1 to shrapnel injury in warfare. Meanwhile, civilian epidemiology of pTBI varies by regions; in US the commonest cause of pTBI being gunshot wounds which have a lower incidence of infections, compared to farm or other handheld metal objects in other countries. While intracranial infections can be fatal, on the other hand antimicrobial resistance (AMR) also poses an important problem. This was witnessed in the COVID pandemic which saw an associated jump in incidence of AMR due to routine use of antibiotics to treat viral pneumonias.

The available evidence limited the recommendations that could be generated for penetrating brain injuries. DELPHI consensus ws was thus used to bridge available evidence with the need to make clinical decisions.

Limitations

All studies that have attempted to look at rates of infection and use of prophylactic antibiotics in pTBI were retrospective. The heterogeneity of injuries, the process of decision-making for the administration of antibiotics, and which antibiotics were selected for use are all significant sources of bias in these studies. Whether the patients who received prophylactic antibiotics were actually at high risk of infections remains undetermined. For example, pTBI from blast injuries as opposed to gunshot injuries are at high risk of contamination, retained fragments, and resulting infection. This additional selection bias must also be taken into account when considering the results of these studies, as they limit the external validity of some findings. Of note, in the main supportive study, there was a significant delay in presentation for many patients. Furthermore, no study described complication rates specifically associated with antibiotic use. While some patients with pTBI may have other systemic injuries that could predispose to sensitivity to toxicity of antibiotics, there is no reason to assume that the head injury itself would do so.

Prophylactic antibiotics are a complex issue in neurotrauma. While pre-procedural antibiotics are now a standard of care for surgical procedures, there is concern that they do not reduce the rate of infection in patients with CSF leaks nor do they reduce the risk of ventriculitis in patients with external ventricular drains in situ. There is concern with these latter indications that they may lead to infections with antibiotic resistant bacteria and that they could provoke clostridium difficile diarrhea/pseudomembranous colitis. Post-insult antibiotic prophylaxis was recommended for patients with pTBI in the first pTBI guidelines making such an application of prophylactic antibiotics after an event unique within neurotrauma. Here we have revisited this recommendation.

On balance with one study suggesting some benefit8 of prophylactic antibiotics for pTBI, the expert panel decided that a consensus would be helpful in addressing risk-stratification of patients for the use of prophylactic antibiotics. High risk patients were considered to be those with violation of dura, patients who are undergoing invasive procedures, air sinus penetration or retained contaminated object fragments. The consensus concluded that it is reasonable to utilize prophylactic antibiotics in high-risk patients with the goal of reducing the risk of infection. Along the same lines, consensus on prophylactic antibiotic use was obtained on TBI-related blast injuries where air sinus penetration occurs, or there are retained bone fragments of penetration from grossly contaminated objects.

To make a comprehensive consensus recommendation, we also reviewed type of antibiotics to be used, when they were used. The consensus was that if prophylactic antibiotics were to be used, the choice should be in favor of broad-spectrum antibiotics including anerobic coverage and for a duration of up to 5 days in high-risk patients and up to 3 days in low-risk patients. Finally, consensus was clear on when prophylactic antibiotics were utilized, they should be administered as soon as possible.

In the face of critically rising antimicrobial resistance, and the balance of immediate benefit to patient but delayed societal harm, further research into the routine prophylactic antibiotic use in pTBI patients is required. It would be easy to conduct relevant studies in centers treating a high volume of pTBI; this guideline suggests that it would not be unethical to withhold prophylactic antibiotics for such investigations. Key elements are classification of contamination in civilian and combat settings, which constitute different phenotypical injuries, and subsequent risk-stratification for the development of intracranial infections following pTBI. This can also be performed by large registries and observational studies, for which data may already exist in national databases.

Randomized controlled trials provide very valuable information. Given that the incidence of civilian pTBI from firearms is rising it may, however, be more efficient to perform these studies by cohort or comparative effectiveness methods. Specific areas of research required remain the utility of prophylaxis versus early treatment, the type of antibiotics i.e broad-spectrum versus narrow spectrum, the timing and importantly duration of prophyaxis.