The primary complications associated with ICP monitor placement are infection, hemorrhage, and malfunction. Ventriculostomy-related infection has been reported most commonly around 10% of drains placed (2%-25%), or 7.5 to 32 infections per 1000 drain-days, though reported rates depend on the definition used (ie, surveillance data, including the Centers for Disease Control and Prevention [CDC] definition of health care–associated meningitis, are frequently blinded to clinical diagnosis, bacterial culture growth, and treatment ramifications). Associated clinical ventriculomeningitis (a diagnosis made on the basis of CSF cellular and metabolic profile, culture data and clinical parameters including fever, peripheral leukocytosis, and findings of meningismus) is a significant cause of morbidity.
Risk factors for ventriculostomy-related infection include duration of catheterization, catheter irrigation/sampling or replacement, neurosurgical operation, systemic infection, CSF leak, and intraventricular or subarachnoid hemorrhage. Responsible microorganisms most commonly include skin flora (eg, Staphylococcus aureus and epidermidis, streptococci, and Propionibacterium acnes), and gram-negative bacteria including Pseudomonas and Enterobacteriaceae. Evidence suggests reduced infection risk with minocycline or clindamycin plus rifampin-impregnated catheters (1.3% vs 9.4% infection and 17.9% vs 36.7% colonization rates), potentially with silver-impregnated ones, and with institution of ICU-based placement bundles and hygiene interventions.
Parenchymal pressure monitors may be associated with a seemingly lower infection rate, though this may be influenced by sampling bias given the ease of CSF access for ventriculostomy surveillance and with few studies culturing device tips in all patients or CSF in all patients. However, retrospective series generally describe clinical infection rates of less than 5% (and mostly < 2%, with higher colonization rates) and either noninferiority or reduced infection risk versus EVD.
While the majority of studies evaluating hemorrhage following invasive ICP device placement do not report radiographic hematoma volume or clinical significance, surgical evacuation of ICP monitor-related hematomas is reported on the order of 0.5% to 3.8%, with one pediatric report describing a 2.7-fold hemorrhage risk reduction with parenchymal monitors versus EVD.
ICP monitors are also subject to technical malfunction. All ICP monitors may provide inaccurate data; in the case of parenchymal sensors this is most often due to insidious zero-drift, which may require replacement (see the section Device Use and Monitoring), or to filament kinking. Zero-drift, occurring on the order of up to 2 mm Hg/8 h in parenchymal monitors (the Integra Camino system is specified to drift up to 2 mm Hg in the first 24 hours and 1 mm Hg/d over the following 5 days), is less of an issue in EVDs as ventriculostomy pressure transducers can be recalibrated to atmospheric pressure. In one large series, some 38% of parenchymal monitors required replacement for drift (Shapiro, n = 244). Unlike fiberoptic transducers, ventricular ones can be recalibrated, and rezeroing should be performed each time the reading is in question and when the patient or drain height is altered.
EVDs are additionally susceptible to compromise of CSF drainage. Poor drainage with clinically consistent ICP data (including with a waveform varying appropriately with respiration and head position) may be due to ICP normalization (which can be ruled in via drain-lowering). Alternatively, this may be due to occlusion by blood or brain/choroidal debris, which can be assessed with visual inspection of EVD tubing and instillation of saline in the distal direction via access ports using sterile technique. Drains also commonly fail due to collapsed lateral ventricle (or ventricles, depending on the degree of compartmentalization) or intraventricular occlusion; CT imaging with particular emphasis on ventricular size/configuration and catheter placement aids in prompt diagnosis and serves as a new baseline prior to replacement should this be required; a nonfunctioning EVD should be promptly removed—regardless of intent to replace—to minimize the risk of infection.