• Summary
  Click here to print a Quick Reference Guide
• Background
• Clinical Diagnosis
• Investigation
• Treatment
• Non-Antimicrobial Treatment
• Empirical Antimicrobial Treatment
• Directed Antimicrobial Treatment (when microbiology results are known)
• Duration of Treatment
• Treatment Failure

Infectious complications following the surgical treatment of hydrocephalus with shunt implantation remains one of the most serious problems in neurosurgical practice.

The clinical manifestations of infections related to CSF shunts tend to be quite variable and often non-specific. Any patient with fever or signs of shunt malfunction must be evaluated for shunt infection and definitive diagnosis requires culture of the CSF obtained via shunt tap.

Treatment of shunt infections is difficult. Medical management with antimicrobial therapy often fails because of the recalcitrant and indolent nature of prosthetic device infections and consequently the shunt must be surgically replaced (See clinical algorithm).

Back to top

Ventriculo-peritoneal (VP) shunt is one of the commonest procedures in neurosurgical practice. Ventriculo-atrial shunts (VA) are now rarely used in UK as they require more revisions and are more prone to infections which may lead to bloodstream infection and nephritis. These shunts are called CSF shunts in this guideline. CSF Shunt infection is associated with an increased risk of seizure disorder, decreased intellectual performance, and a two-fold increase in the long-term mortality rate. The published incidence of shunt infection has varied widely from 1.5 to 39%, although more recently rates of 10 to 15% have been reported. Shunt infections are more likely to occur within first four weeks after surgery; 90% are reported within first 6 months. Infection risk may be especially high in those undergoing three or more shunt revisions.

Risk factors for CSF shunt infection:

• Aetiology of hydrocephalus: increased risk of infection with open head trauma, severe intraventricular haemorrhage and previous neurosurgery, including shunt revisions
• Age of the patient:  much more in premature infants compared to older age groups
• Type of shunt : VA shunts more prone to infections
• Surgeon’s experience and duration of shunt surgery: Shunt infections can be reduced by strict adherence to peri-operative principles of vigorous asepsis, limited personnel in theatre, peri-operative antimicrobial prophylaxis, double-gloving and antiseptic surgical technique.

Microbiology:

• Coagulase-negative staphylococci (CoNS) (50–90%)
• Staphylococcus aureus (13–27%)
• Streptococci (8–10%)
• Aerobic Gram-negative bacilli (GNB) (10–20%)
• Other Gram-positive organisms, including Propionibacterium spp and Corynebacterium spp are occasionally encountered.
• Fungal shunt infections are quite rare though frequency of Candida shunt infections has increased in recent years

Classification of Shunt infections, based on aetiology and site of initial infection
Internal shunt infections
These constitute the majority of cases. The organism colonises the lumen of the shunt. Most internal infections are caused by CoNS which adhere to device materials by producing biofilms.

External shunt infections
Usually reflect surgical wound infections around shunt tubing. May be caused by Staphylococcus aureus and aerobic GNBs.

Back to top

Clinical:

• Fever (variable) in the absence of other recognised causes
• Signs of shunt malfunction (raised intracranial pressure): headache, vomiting, altered mental status.
• Traditional meningeal symptoms less common as mostly low grade pathogens
• Erythema/ tenderness of skin over shunt tubing
• Abdominal pain, focal or generalised peritonitis if distal end infection in VP shunts. Rarely, the distal portion of the shunt may perforate the bowel, leading to peritonitis and abscess formation. Occasionally large abdominal pseudo-cysts (best detected by ultrasound) are the first sign of a low grade infection. U/S guided aspiration may assist in treating the symptoms associated with the cyst and also provide CSF for culture. VP shunt infections may also result from other intra-abdominal pathologies or surgery.
• VA shunts - Signs of sepsis, right sided endocarditis, hepatosplenomegaly

Early detection and diagnosis of shunt infections may be difficult because of non-specific clinical features, however, any deterioration of consciousness should warrant immediate neuro-imaging and CSF examination. It may be difficult at times to differentiate between a shunt infection and a shunt malfunction because shunt malfunctions can mimic the signs and symptoms of a shunt infection (Barnes et al., 2002) and shunt infection itself may result in shunt blockage/malfunction.

Neurosurgical review for shunt tap is critical for assessing shunt viability, checking pressures, and diagnosing infection.

Back to top

Laboratory:
Shunt tap (CSF from accessible reservoir):

Recommendation: Neurosurgical assessment for shunt tap should be undertaken in all patients with suspected CSF shunt infection [Evidence Level C].

Recommendation: Sampling of CSF should be performed as a sterile procedure as shunt tapping may introduce infection in the system [Evidence Level C].

Positive CSF culture from the shunt is the most important test to establish the diagnosis of shunt infection.
           
CSF may show leucocytosis. However cell counts and biochemistry may be normal.

Gram stain. A negative Gram stain does not exclude infection.

Staphylococcus aureus and GNB shunt infections are likely to reveal more striking ventricular fluid abnormalities compared to CoNS.

CSF may be completely normal in distal shunt infections without shunt malfunction. Note lumbar puncture may not reflect the state of ventricular fluid because lumbar fluid is not in direct connection with the proximal shunt.

Blood Cultures
These are usually positive in patients with VA shunts (90%). However they have low yield in those with VP shunts (<20%).

Recommendation: As CSF shunt infections present in a non-specific way, blood cultures must be done to investigate for other foci of infection [Evidence Level D].

Further samples
Recommendation: Samples such as wound swabs shunt tubing or fluid from abdominal collections may be sent on a case-by-case basis [Evidence Level C].

Back to top

The decision to commence empirical treatment (surgical and antimicrobial) can be difficult to make. The clinical signs of shunt infection are very similar to those of shunt malfunction. CSF white cell counts, protein and glucose may be normal or only very mildly deranged and CSF Gram stains are often negative. In the absence of clear microbiological findings (e.g. a positive CSF Gram stain) the decision to intervene surgically and commence antimicrobial therapy is ultimately a clinical one.

There are no published, well-designed studies comparing different methods of therapy for shunt infections.

OPTIONS

1. Two stage procedure whereby the entire colonized shunt is removed, CSF is drained externally and antibiotics administered by the systemic or intra-ventricular route or both and the shunt replaced when the CSF is sterile.
2. One stage procedure whereby the colonized shunt is removed and immediately replaced with a new shunt, followed by a course of antibiotics.
3. Antimicrobial therapy alone

There have been very few reviews of comparison of efficacies of all three major categories of intervention. Yogev et al reported cure rates of 96%, 65% and 36% for the two stage, one stage and conservative management, respectively. A more recent analysis also reported similar cure rates. Poorest results have been observed with antibiotic treatment alone (34-36%).

Recommendation: Complete removal of the shunt, external drainage of CSF by external ventricular drain (EVD) and antibiotic therapy is recommended as the treatment of choice [Evidence Level B].

The only exception to this is shunted patients who contract purulent primary bacterial meningitis (e.g. Neisseria meningitidis, Streptococcus pneumoniae) who could be treated with antibiotic therapy alone without shunt removal, provided the shunt continues to function normally.

Re-shunting
A clinical and microbiological response is usually seen by day 4 or 5 and the patient can usually be re-shunted, if needed, between 7-10 days of treatment. For most CoNS infections, 5-7 days of treatment should suffice.

Back to top

The decision to commence empirical treatment (surgical and antimicrobial) can be difficult to make. The clinical signs of shunt infection are very similar to those of shunt malfunction. CSF white cell counts, protein and glucose may be normal or only very mildly deranged and CSF Gram stains are often negative. In the absence of clear microbiological findings (e.g. a positive CSF Gram stain) the decision to intervene surgically and commence antimicrobial therapy is ultimately a clinical one.

Intrathecal (IT)¹ Vancomycin 20mg daily (note the dose may be split as 10mg x 2 in non-communicating ventricles) plus IV Cefotaxime 2g four times daily or IV Meropenem 2g three times daily.  The final choice will depend on risk factors for resistance and other factors – please discuss with Microbiology.

Recommendation: If CSF Gram stain is positive, initial therapy can be given according to Gram result [Evidence level C]:

Gram positive bacteria seen in CSF:
Intrathecal (IT)¹ Vancomycin 20mg daily (note the dose may be split as 10mg x 2 in non-communicating ventricles)

Gram negative bacteria seen in CSF:
IV Cefotaxime 2g four times daily or IV Meropenem 2g three times daily. The final choice will depend on risk factors for resistance and other factors (such as recent antimicrobial therapy) – please discuss with Microbiology

Back to top

Coagulase negative staphylococci (CoNS)
The addition of Rifampicin 600mg twice daily (oral or IV) should be considered in CoNS infections which fail to respond to IT Vancomycin monotherapy.

S.aureus (meticillin-susceptible) – add IV Flucloxacillin ² 2g four times daily.
S. aureus (meticillin-resistant) – add IV Vancomycin (dose determined using dosing guidance) or Linezolid* as advised by microbiology.

¹ Intrathecal Vancomycin is used to achieve therapeutic CSF concentrations. It appears to be safe, despite high CSF levels, and monitoring is not required. Penicillins and cephalosporins should not be given by the intrathecal route because they have been associated with significant neurotoxicity, especially seizures.

²In cases of penicillin allergy – use IV Vancomycin  or Linezolid * as advised by microbiology.  

* Linezolid has a number of drug interactions/contraindications. Please see full guidance to check suitability for the patient. This is an off-label use of Linezolid and the patient should be informed of this.

Gram negative shunt infections (e.g. "coliforms", Escherichia coli, Enterobacter spp.)

IV Cefotaxime 2g four times daily or IV Meropenem 2g three times daily. The final choice will depend on risk factors for resistance and other factors (such as recent antimicrobial therapy) – please discuss with Microbiology

The addition of intrathecal (IT) Gentamicin should be considered if the isolate is confirmed to be susceptible to Gentamicin. It should be started at a dose of 1mg once daily on day 1, increased to 3mg once daily on day 2 and then 5mg once daily thereafter.

Monitoring of CSF level of Gentamicin is NOT required.

Fungal shunt infections

IV ambisome 3mg/kg per day.

Alternative therapy may be recommended on the basis of the causative organism and antifungal susceptibilities. Please discuss on a case-by-case basis with Microbiology.

Back to top

Depending on the clinical and microbiological response, treatment for Gram negative shunt infections should continue for a minimum of 14 days.

Back to top

Clinical suspicion of VP shunt infection – empirical therapy
(Therapy commenced to be agreed in discussion with Microbiology as CSF microscopy and then culture results are obtained)

Back to top

Back to top