One of the most feared complications of cerebrospinal fluid (CSF) diversion surgeries, using for instance, a ventriculoperitoneal shunt (VPS), is a surgical site infection (SSI) leading to a iatrogenic shunt infection. Presumably, this infection occurs as a result of the contamination of the shunt with bacteria during surgery [10, 15]. Efforts to reduce SSIs have been around since the early 1980s [13]. Broadly, these efforts involve the use of prophylactic systemic antibiotics, the use of standardized protocols [8], the use of antibiotic-impregnated shunts [9], and the use of topical antibiotics [3].

The use of topical antibiotics is prevalent among different surgical specialties [1, 4, 6, 13]. The rationale for using topical antibiotics is their supposed minimal systemic effect, together with high direct concentrations of antibiotics in the surgical field. This might, in theory, reduce the need for systemic antibiotic use but might incur other risks, such as topical dermatitis.

Interestingly, most of the reviews published that attempt to summarize the effect of topical antibiotics in different surgical fields show mixed results, with effects only present in subgroups [11]. Moreover, despite their widespread use, evidence from conclusive randomized controlled trials is still missing [7]. The issue is further confounded by the large variation in clinical practice regarding the antibiotic used and the manner of administration (e.g., intraoperative washes, injections, lotions, solutions, powders, gels, impregnated beads, collagen implants, or even stitches). An illustration of the difficulty in defining the term ‘topical antibiotics’ is the Cochrane review, which included only studies on topical antibiotics administered in the wound dressing applied after wound closure that subsequently showed an overall positive effect [5]. As a result, the presence of an effect, as well as an indication of which form of topical antibiotics should be used, remains ambiguous.

Antibiotic-impregnated catheters, which may also be considered a form of topical antibiotic administration, have been around for 20 years. They have shown various effects in reducing VPS-related infections up until a couple of years ago [9, 16]. Underpowered, often single-centre observational studies failed to provide compelling evidence of their effectiveness. In 2019, however, a large, phase III randomized controlled trial (RCT) run in the UK managed to demonstrate that their use led to the reduction of VPS catheter-associated infections [9].

Frequently, the addition of topical antibiotics to routine surgical care is part of the implementation or update of a protocol. Increased attention to the surgical procedure is a consequence. Generally, it is not possible to claim a reduction of the postoperative infection rate, or a part of it, to one of the items of this protocol, including the addition of antibiotics.

It has been shown that the introduction of a protocol coupled with subsequent high-level adherence successfully reduces shunt infections [8]. A protocol, however, is a bundle of care which makes it difficult to pinpoint exactly which elements of the care bundle contribute most effectively to the desired effect and which do not [2]. Furthermore, the so-called ‘Hawthorne effect’ may also come into play, as simply ‘paying attention’ to a specific procedure or to the way the steps of a protocol are followed may, by itself, lead to the desired effect [14].

Within this backdrop of uncertainty, in the latest issue of Acta Neurochirurgica, Ganesh et al. report their systematic review and meta-analysis of the use of topical antibiotics to reduce VP shunt infections [3]. The authors included nine studies, of which one RCT encompassed 2,764 patients. Despite the reduction in the risk of infection, the authors point out that the overall risk of bias of the included reports was high, which hampers drawing any definitive conclusions from this data.

As with other neurosurgical topics, what this topic lacks most is standardization. In addition, with the original set up of some of these studies, the population sampling is nevertheless expected to differ. What kind of protocols were used and were they adhered to? Did they all use antibiotic-impregnated catheters? Was the length of the surgical wounds, their closure, and the length of surgery comparable? Were the postoperative instructions and postoperative care the same?

Furthermore, different patient populations with different risk profiles were included. For example, the included RCT showed a benefit of the use of vancomycin but only included patients up to 1-year-old [12]. The largest study with a historical cohort also showed a reduction in the rate of VPS infection, but the intervention cohort was sampled after their protocol was updated, so whether the effect is based on the use of vancomycin or the updated protocol (and increased attention due to this change) is debatable [17].

The investigational and control group of a future study should be carefully defined. A multiple centre design is warranted with similar protocol, uniform administration, and dosage of antibiotics. Neurosurgeons with extensive experience in shunt surgery preferentially participate in order to maximize the efficiency of the shunting surgery and, thereby, facilitate the reduction in time of wound exposure. If these conditions are minimally met, it is possible to estimate the true effect size of the use of intraoperative antibiotics in shunt surgery. Due to the incidence of shunt infection, and the estimated, relative low effect size, the sample size would include several thousands of patients, making its realization unlikely.

Since VPS infections are a major iatrogenic healthcare burden, involving high costs and potentially devastating neurological consequences, especially in children, alternative methods, and designs should be investigated to collect scientific data considering the effect of intraoperative antibiotics.

Meanwhile, efforts must be concerted in reducing the rate of shunt infections. If the introduction of a protocol is effective, it might even be worthwhile to introduce it without knowing what factor contributes to the effect. One final challenge remains: the term ‘topical antibiotic administration’ is so heterogeneous and encompasses so many treatment modalities that it is uncertain if it will ever truly qualify, without further specification, as one single intervention.