Shunt infections: a review and analysis of a personal series
Introduction and purpose
CSF diversion shunts are notoriously prone to complications. The most difficult to manage among them is shunt infection, which warrants a prolonged hospital stay. The aim of this paper is to review the pattern of infections, the pathology, and management of shunt infections with special reference to a tertiary pediatric center in a developing country.
Materials and methods
This is a review of shunt infections in general and a retrospective study of all cases operated in the hospital from 2000 to 2015.
The authors analyze the data and try to discern patterns, which may enable newer interventions to treat as well as decrease the burden of shunt infections in the future.
It is difficult to determine the true incidence of shunt infections as there is no definition of what constitutes a shunt infection. There are no standardized international guidelines as to how to deal with an infected shunt. Though the ability to treat shunt infection has improved and the incidence of shunt infection has decreased over time, there is still no consensus on the best way to manage it. The prevention is predominantly based on common sense and has helped but a more scientific algorithm is the need of the hour.
KeywordsShunt infection Hydrocephalus Cerebrospinal fluid CNS infections CSF infections Antibiotic therapy Ventriculoperitoneal shunt
CSF diversion shunts are notoriously prone to complications. Among these, the most troublesome and difficult to manage is perhaps a shunt infection. The mechanical complications can be treated by surgical procedures and may not require prolonged hospitalization or antibiotic therapy. Infections, however, require both. Several trials and protocols have been conducted in an effort to minimize infections. Choux et al. have shown that infections can be reduced by simple measures and strict adherence to protocols .
An infected shunt increases morbidity and can even cause mortality. Mortality of up to 30% has been noted. It often entails prolonged hospital stay and increased expenses. Although the incidence has decreased in the recent past, infections have not been eliminated altogether. Each surgeon follows a different line of management. Though this makes the analysis of results difficult, each unit has to evolve a protocol tailored to suit the conditions they are working under. A uniform standardized system may be difficult to formulate.
As there is no uniform definition of what constitutes a shunt infection, the estimation of the incidence is often impossible. The rates have been estimated to range from almost 0% to as high as 60%. Realistically, the incidence lies between 2 and 5% in most centers. A wide variation in the incidence between centers and even between surgeons is apparent. Also, it has to be noted that the incidence of infections is waning in all kinds of shunts and the incidence is similar for all types of shunts. A better understanding of the pathophysiology, stricter regimens in preventing and combating infections, and reduction in operating time may be the important factors for this. Furthermore, the operating team’s experience, better understanding of the presurgical and post-surgical processes involved, and better biocompatible materials can be cited as the reasons for this reduction in the infection rate.
Pathology of infections
The shunt is alien to the body and naturally the body mounts a response to it. But the rejection rate is low these days because of better and more biocompatible material and other advances.
The immediate response of the host is to form a protein coat for the shunt, which confers protection against infections by limiting bacterial adherence to the walls. The bacteria, however, form micro colonies at the site of their adherence. The next step by the bacteria is the secretion of a mucopolysaccharide, which forms a slime capsule (glycocalyx) around the micro colony. Staphylococcus epidermidis and Staphylococcus aureus are well known to do this [2, 3, 4]. This slime capsule prevents the ingress of the antibiotic to the interior of the capsule; thus, the bacteria are “safe” and continue to flourish. In addition, the bacteria escape from the colony in “batches” to further colonize the shunt distally and perpetuate the infection. The leucocytes, however, do not adhere as well to the shunt walls as the bacteria do. The walls of the shunt have tiny irregularities, pits, and cracks, which serve as a harbor for the bacteria. Certain bacteria like S. aureus dig into the catheter and find a safe haven there. Besides this, the problem of incompletely treated infections exists [2, 5, 6]. It is therefore logical to conclude that the shunt hardware has to be removed to eradicate the infection completely. Antibiotic and silver impregnated shunts aim to lessen this bacterial burden and studies are in progress to assess their value in prevention of pediatric shunt infections .
Perhaps the most common organism infecting a shunt is S. epidermidis—being encountered in up to half of the cases. This stands to reason since it is a resident flora of the skin. Next in line is S. aureus. Escherichia coli, Pseudomonas sp., and Klebsiella sp. are also not infrequently encountered. Exotic flora and fungal sepsis indicate an immune compromised state. The fact that the commonest organism is S. epidermidis indicates that it is implanted at the time of surgery. Indeed, the wound is “contaminated” by the end of the surgical procedure.
Almost 90% of the infections are seen within the first 6 months and half of them being encountered in the first 2 months after surgery. There is no difference, however, in the incidence of infections or the causative flora between the various types of shunts.
As discussed earlier, the formation of the slime capsule results in downstream colonization of the shunt necessitating the removal of the hardware to eradicate the infection. MRSA infection is a very difficult problem to manage, needing prolonged therapy and isolation. E. coli infection is common in children with neural tube defects and often follows urosepsis since these children are chronically colonized in the bladder with this organism. Group B streptococcus, meningococcus, pneumococcus, and H. influenza infections are different as these can be managed without removal of the hardware.
Two features stand out in an infected shunt—fever and features of shunt malfunction. Nonspecific features like failure to thrive, poor feeding, lethargy, malaise, and irritability may also be present. Only a high index of suspicion will lead to the diagnosis being made in these. History of frequent shunt blocks should alert the neurosurgeon to the possibility of shunt sepsis.
Increased risk and incidence are seen in extremes of age, immunocompromised state, incompletely treated sepsis anywhere in the body, CSF leaks, and in those with neural tube defects.
Certain iatrogenic causes are worthy of mention: shunt done at the end of a long operating list, inexperienced team, several shunt surgeries in 1 day, increased operation time, and OT traffic.
Colonization causes mild fever while serious sepsis like meningitis or peritonitis leads to high grade fever.
A high index of suspicion is mandatory to diagnose shunt sepsis.
A good history and thorough examination are the initial steps. Points worthy of note are age, past history of sepsis or sepsis anywhere in the body, time since last shunt, number of revisions, etiology of hydrocephalus, presence of neural tube defects, immune status, and if the child is/has been treated for malignancy, especially with chemotherapy.
Imaging will reveal features of shunt malfunction like ventriculomegaly, broken or disconnected shunts, and abdominal sepsis with previously described features.
CSF analysis is mandatory in all cases. In some instances, the infection may be obvious like redness along the tube.
There is no ideal place to sample CSF but a shunt tap is the best since it samples CSF from the shunt and the ventricles and the yield is often good. It can diagnose hardware colonization/sepsis and ventriculitis.
Analysis is often variable, the cell count being raised, and the sugar and protein levels too variable. A positive culture clinches the diagnosis. A marginally elevated cell count has to be interpreted with caution since this can happen even in non-infected states because the shunt is a foreign body and it may just be a response to it. A differential count of the white cells in the CSF may resolve the issue—neutrophils will be high in infections and eosinophils in allergic response.
It has been the senior author’s experience that only a positive culture is indicative of a shunt infection. The results of the rest of the analysis can be highly variable and have to be interpreted with caution taking the overall picture into account. If there are overt clinical signs of sepsis, that obviously takes precedence over the lab data and the child has to be treated for sepsis.
It has to be emphasized that a single negative culture does not mean clearance of the infection or full treatment of the sepsis. This is because the administration of the antibiotics may reduce the virulence of the organism and not clear it completely. The organism may still be present. Hence, it is the senior author’s policy to get at least five consecutive negative cultures before proceeding with reinsertion of the shunt.
These result from low grade or anaerobic/microaerophilic sepsis. In the presence of infection, the shunt is exteriorized and revised later. ETV is a viable option in suitable cases. If there is no evidence of active sepsis, laparoscopic intervention may be attempted. Shunt revision can only be performed once the abdominal fluid has shown no evidence of infection and is preferably done in another quadrant of the abdomen.
This is the same as the foregoing but is also seen in children with chasmal/hypothalamic gliomas in the absence of sepsis. Revision with placement of shunt in another quadrant of abdomen is not feasible in these cases. Here, an atrial shunt or where possible ETV is done since a ventriculopleural shunt may also result in hydrothorax.
Abdominal sepsis/acute abdomen
Here, the shunt has to be immediately exteriorized. After the abdominal problem is resolved, it can be replaced in the peritoneum or if not possible ETV or pleural shunt can be done.
Antibiotics alone cannot eradicate the infection—except in the situations referred to earlier in the section on “Microbiology.”
The shunt hardware has to be removed in most other situations.
It has to be emphasized that there is no uniform protocol for the use of antibiotics. The choice is governed by the individual surgeon’s experience and preference and policy of the microbiology team of the hospital.
A simple or basic first generation antibiotic is started pending the culture report. If, however, the causative organism is known, then the appropriate antibiotic may be started. A broad spectrum drug is started if there is serious sepsis like peritonitis.
The antibiotic is usually administered at least for 2 to 3 weeks in all but the superficial infections. After the cultures of the CSF are obtained and if sterile (vide supra), the shunt is reinserted. The antibiotics have to be continued—irrespective of the duration—till the CSF is sterile.
The role of intraventricular antibiotics is limited. Access to the ventricle through an EVD or exteriorized shunt is mandatory. Giving the drug through a lumbar puncture is of not much use since adequate concentrations are not seen above the cervical level. In addition, the choice of antibiotics is limited, the only advantage being no consideration of CSF penetration.
The drugs commonly used are gentamycin 0.05 to 0.1 mg/kg od or q 12 h; vancomycin 0.5 mg/kg od or q12hr; amphotericin 0.1 mg to 0.3 mg/kg od; amikacin 5–50 mg/day; colistin 10 mg/day, and tobramycin 5–20 mg/day. Adjustments have to be made for age, renal, and hepatic status. Penicillin and cephalosporin are not used for fear of precipitating seizures.
The role of surgery
Removal of the shunt is mandatory in almost all the cases.
The surgical management has to proceed in a simple, logical, and stepwise manner.
Baseline imaging is done. If it shows ventriculomegaly, the shunt is removed straightaway and an EVD is placed. If, on the other hand, the ventricles are not dilated, the shunt is exteriorized and after a few days is closed overnight to allow the ventricles to dilate. The shunt is then removed and an EVD is placed.
Intraventricular antibiotics are given if indicated.
The shunt is placed—the choice of the site or ETV depends on the clinical and imaging scenario.
Senior author’s (CB) series
Materials and methods
The data presented are a retrospective study of cases from 2000 to 2015 from the senior author’s personal series. The total 477 children were shunted. The ages ranged from 32 weeks to 17 years. The total number of procedures was 1026. Primary shunts were 477 in number and revision procedures were 549.
Shunt infections were observed in 31 cases (3%). These were differentiated into primary shunt infections: 25 (2.4%) and residual shunt infections which were 6 (0.58%). Residual infections were those wherein a child is treated for an infection and then subsequently the same microbial organism was isolated from the shunt. This has been reported with difficult to treat pathogens . These are due to incompletely treated infections and were seen between 1 and 6 months post shunt surgery while primary infections were seen between 1 week and 1 month after the shunt surgery (except for two cases which occurred after a year).
The senior author’s shunt protocol
The shunt protocol followed in our institute is akin to the Choux protocol. Other intra-operative practices followed by us are as follows:
The child is given a shampoo the night before and the morning of surgery. Most shunts are preferably done only in the theater designated for neurosurgery. Not more than two shunts are done in a day. If there are more cases, one of the following is done, ETV, EVD, or postponement of the procedure to the next day. If only a single shunt is posted in the list, then the shunt is the first case of the day. If there are two cases in the list, the shunt is done first. A shunt can follow only another shunt. Neonates and infants are operated before older children. If for any unavoidable reason a shunt has to be done later in the day or has to follow another “non-shunt” procedure, the OT is thoroughly cleaned and disinfected before the shunt. Bilateral shunts are treated as two independent procedures, the whole team de-scrubs and re-scrubs, patient is re-prepped and re-draped, and a fresh set of instruments are used.
The doors of the operation theater are locked before commencement of surgery. The intra-theater personnel are kept to a bare minimum. No talking or moving is permitted once the surgery is commenced. The table is positioned such that the body is closest to the ventilator and care is also taken to see that the anesthetic circuit does not course along the patient’s skin.
The hair is not shaved but clipped to lessen the incidence of skin injury and commensal infection .The skin is thoroughly prepared with povidone iodine and is allowed to dry for at least 3 min before draping is commenced. Intravenous cefazolin 100 mg/kg/day at commencement of anesthesia is administered . This is continued only for 24 h.
The skin contact is kept to a minimum. The use of double gloves for all involved in the surgery (powderless gloves) . This practice has been well described . Two layers of sterile disposable drapes are used. The shunt tube is not opened till the tunneling is completed. The shunt is not soaked or rinsed in antibiotic solution. Intraventricular gentamycin (1–2 mg) is administered while tapping the ventricle. The use of cautery is kept to a minimum.
The distal catheter is cut to an adequate length and the complete length is never placed in children. The incidence of pseudocysts in our series is very low (7 out of 1026, i.e., 0.68%). None of the pseudocysts were of significant size and none required intervention. One pseudocyst was aspirated under ultrasound guidance but the culture was sterile.
The dressing is removed on post op day 1 and a shampoo bath is given on day 2. The child is discharged on day 2 or 3.
The most important part of the management is to define shunt infection. Unfortunately, there is no standard definition of what constitutes a shunt infection. But generally the diagnosis is based on a set of clinical and laboratory features. “Shunt infection is generally defined as the identification of a bacterial pathogen from the CSF both by gram stain and culture, in conjunction with CSF pleocytosis, fever and neurologic symptoms and signs of shunt malfunction” [2, 3]. Most centers follow this definition though the emphasis on various aspects of the definition may be different. The time interval post shunt is irrelevant, though the longer the interval between the surgery and the presentation, the lesser the degree of suspicion for an infection.
Clinical signs such as fever and inflammation along the shunt tract with/without swelling (in the immediate post-operative period) are often noted. The presence of shunt dysfunction is an important but not an absolute feature of shunt infection. However, recurrent events of shunt blocks in quick succession are viewed with suspicion as the hardware might be harboring a pathogen. Coupled with this the complaint of malaise are the clinical features wherein a probable shunt infection is suspected.
The initial CSF sample is usually from the shunt chamber/reservoir and is done at the bedside prior to administration of any antibiotic. The laboratory features include elevated total counts, elevated neutrophils or lymphocytes in the differential counts, raised CRP and a CSF picture of low sugar, elevated counts, and a positive culture. The single important feature given credence to in our institute is the CSF culture. A positive CSF culture is taken as definitive infection.
When a shunt is infected, the shunt is removed in toto and EVD is placed. Intravenous antibiotics are commenced as per culture reports. The average duration of antibiotics is 14–21 days. Once the antibiotic course is completed, it is our practice to repeat five consecutive CSF cultures and only if they are all negative do we proceed with a new shunt preferably on the opposite side or at least through a new burr hole. The CRP is only used as an adjunct to monitor treatment response and is not used as a diagnostic tool. This protocol has led to a low shunt infection rate (3%) in the senior author’s experience.
Some studies have more specific definitions for shunt infections. For example, in a study by Kestle et al. , CSF shunt infection and reinfection was defined by 1 or more of the following criteria: (i) presence of bacteria in a Gram stain or culture of CSF, wound swab, or pseudocyst fluid; (ii) documentation of visible hardware; (iii) abdominal pseudocyst; and (iv) in children with a ventriculoatrial shunt in place, presence of bacteria in a blood culture was also considered a CSF shunt infection.
Organisms that grew in broth only were considered to be infections.
In some special scenarios where the organism cultured is one such as Pneumococcus, the shunt is not revised unless dysfunctional. If a new shunt procedure is done in the presence of Pneumococcal infection, it is placed under cover of antibiotics, which have already been given for at least 72 h. This protocol has worked well for us in Pneumococcal infections, either primary or secondary. This is probably because the Streptococcus pneumoniae has a low incidence of adherence to the shunt lumen unlike staphylococcus . A few well-documented cases wherein shunt infections with pneumococcus have been treated with antibiotics with the shunt in situ have been reported [15, 16].
A feature quite unique to our series is the presence of more than one organism in the culture which was seen in 12 cases who were all children of less than 1 year. This is probably because these infants had incompletely or inadequately treated sepsis resulting in residual infection.
All infected shunts were tackled by removing the existing hardware and placement of EVDs. EVDs were changed every 7–10 days. The use of ventriculo subgaleal shunts for infections in this situated is reported [17, 18] but has not found to be reliable by us.
Shunt revisions in our series were not associated with an increased risk of infection though multiple revisions are well known as a cause for shunt infection . In our cases, multiple cases of shunt revisions were usually associated with mechanical failure such as block or migration.
The use of endoscopic third ventriculostomy routinely in the presence of infections or for post-infectious hydrocephalus in children less than 2 years is not favored by us though several large case series have been reported especially from Uganda [19, 20] with success rates of up to 52%.
Number of cases (n)
Shunt infection incidence (%)
Simon DT et al. (2009) 
Simon DT et al. (2014) 
Yakut N et al. (2018) (Turkey) 
Kestle JR et al. (2006) 
Kulkarni AV et al. (2001) 
Gathura E et al. (2010)  (Africa)
65 (also 7.1% mortality)
Braga MHV et al. (2009) (Brazil) 
Choux M et al. (1992) (France) 
Kanangi SMR et al. (2018)*
Most shunt infections are reported to occur within 1 month of procedure . A high percentage of pathogens originate from skin and include Staphylococcus epidermidis and Staphylococcus aureus. Gram negative infections in VP shunts originate from the enteric system [3, 35]. Thus, most CSF shunt infections are introduced during surgical manipulation of the site. Prematurity predisposes to shunt infection probably due to the need for multiple revisions and this was seen in our series as well . Investigators have noted considerable variation in treatment duration and methods of medical and surgical management of shunt infections that might explain reinfections [37, 38].
Given the large and diverse nature of the medical practice, it is difficult to formulate a uniform set of guidelines. However, regional standardization would help. The first step in this direction would be to set up a regional shunt registry as the Hydrocephalus Clinical Research Network (HCRN) . The data procured would help in quantifying the burden of disease and also in formulating guidelines.
The management of shunt infections has been significantly influenced by the work of Choux et al. . The results of newer trials like the NHS-England BASICS trial  are awaited and these will hopefully shed more light on the best modalities of managing and preventing shunt infections.
Antibiotic impregnated shunts
Antibiotic and antibacterial material shunts have been introduced. But these have not stood the test of time. Given the low incidence rates by following simple protocols and measures, the infection rate can be greatly reduced. Whether these are really useful and the justification for using these more costly hardware can only be answered by multicenter studies [39, 40, 41]. Currently in the UK, the BASICS (British antibiotic and silicone impregnated catheters for VP shunts trial) is recruiting patients who will be randomized to assess efficacy of these special catheters. There is some reports that although this is a novel idea for prevention of infection, in reality, it does not bestow much extra protection . However, these shunts have proved useful in infants and against Staph infections. However, the infection rate is higher if they come up for revisions.
Sonication and its use in shunt infections
Sonication is a useful technique which dislodges biofilm bacteria from the surface of implanted materials before culture. In one study , removed devices were sonicated in saline (40 kHz, 1 min, 0.25 W/cm(2)), the resulting fluid was cultured aerobically and anaerobically at 37 °C, and bacterial growth was counted. Ventricular cerebrospinal fluid (CSF) was cultured separately. In the EVD group, sonication cultures grew significantly more bacteria (64%, 9/14) than cultures of aspirated ventricular CSF (14%, 2/14). In the VPS group, the difference was not significant. Positive sonication cultures of EVD catheters yielded a median of > 100 colony forming units (CFU) (range, 60–800). For positive sonication cultures of VPS, the median was 1000 CFU (range, 20–100,000). All patients with bacteria in their CSF also had positive sonication cultures from the removed device.
The diagnostic yield of conventional microbiological techniques is much lower than after sonication.
Shunt infection is a serious clinical problem. It is time consuming and expensive to manage and can be quite frustrating to clear the infection too. Modern microbiological techniques like using sonication to break up the biofilm yields much better microbiological results making the choice of antimicrobial medication easier. A meticulous protocol during shunt surgery is sine qua non for the prevention of shunt infections. Simple measures can reduce the infection rate substantially. Care should be taken to treat any pre-existing systemic infection adequately. Shunt surgeries are not to be treated casually and should be given proper importance.
Compliance with ethical standards
Conflict of interest
The authors have no disclosures to report.
- 1.Choux M, Genitori L, Lang D, Lena G (1992) Shunt implantation: reducing the incidence of shunt infection. JNS 77(6):875–880Google Scholar
- 5.Gathura E, Poenaru D, Bransford R, Albright AL (2010a) Outcomes of VP shunt insertion in Subsaharan Africa. JNS Paeds 6(4):329–335Google Scholar
- 7.Jenkinson MD, Gamble C, Griffiths M, Mallucci C et al (2014) BASICS: British antibiotic and silver impregnated catheters for VP shunts multi centre randomized controlled trial (BASICS trial). Trials 15:4. Published online 2014 Jan 3. https://doi.org/10.1186/1745-6215-15-4 CrossRefPubMedPubMedCentralGoogle Scholar
- 11.Tulipan N, Cleves MA (2006) Effect of an intra-operative double gloving strategy on the incidence of CSF infection. JNS 104(1 Suppl):5–8Google Scholar
- 12.Drake JM (2006) Does double gloving prevent CSF shunt infection? JNS 104(1 Suppl):3–4 discussion 4Google Scholar
- 16.O’Keefe PT, Bayston R (1991) Pneumococcal meningitis in a child with a ventriculo-peritoneal shunt. J Inf Secur 22:77–79Google Scholar
- 20.Warf B, East AFRICAN Neurosurgical Research Collaboration Paediatric hydrocephalus in East Africa: prevalence, causes, treatments and strategies for the future. World Neurosurg. https://doi.org/10.1016/J.WNEU.2010.02.009
- 31.Yakut N, Soysal A, Kadayifci EK et al. (2018) VP shunt infections and re-infections in children: a multicentre retrospective study. Br J Neurosurg 1-5 DOI: https://doi.org/10.1080/02688697.2018.1467373
- 39.Govender ST, Nathoo N, van Dellen (2003, JNS) Evaluation of antibiotic impregnated shunt system for the treatment of hydrocephalus. 99(5):831–839Google Scholar