Risk Factors for Surgical Site Infections

A multitude of factors have been identified that result in SSI. Although information remains contradictory on their respective validity, risk factors can be categorized as patient- or procedure-related.17

Wound Categories

The National Academy of Sciences has proposed the widely accepted classification of the contamination status of sites of surgery as clean, clean–contaminated, contaminated, and dirty.15 With the correct administration of antibiotic prophylaxis and appropriate aseptic technique, the infection rates are lowered in all categories.1,15,20,21 In neurosurgery, most cases are considered clean; however, reported infection rates for shunt implantations, which are considered clean surgeries, are higher than those for other cranial procedures. To account for this discrepancy, the cases in the clean subdivision have been separated into those with and those without foreign body implantation.34 Although the question remains of how much material constitutes the difference between clean and clean with foreign body, this seems like a reasonable subdivision, in particular with regard to shunt procedures, cranioplasty, and surgeries involving dural substitutes.34

The term clean–contaminated describes procedures in which the respiratory or alimentary membranes are entered under controlled conditions, as in trans-sphenoidal or endoscopic approaches30,33 to the skull base. With proper attention to antiseptic techniques, these cases can have similar infection rates (1.8%),33 as can clean elective craniotomies. Contaminated and dirty procedures are rarely performed in neurosurgery and are mostly associated with emergency situations, which in themselves represent a risk factor.

Shaving of the Surgical Site

Hair has previously been considered a risk factor for wound infections. However, evidence exists that shaving itself poses a risk. A clear association between the manner of hair removal and SSI has been found in that the use of a razor results in a higher risk for SSI than do clippers or a depilation cream.26 A razor produces microscopic abrasions of the skin, which are subsequently colonized by bacteria. Shaving the surgical site the evening before surgery poses the greatest risk for SSI because of the extended period in which bacterial colonization can occur, and this practice has largely been abandoned.

A comprehensive article by Winston26 offers an enjoyable analysis of this issue of hair removal. Hair can be rendered void of infecting organisms by using proper local preparation and surgical scrubbing. The infecting flora is found in the deeper layers of the skin, which are not affected by shaving. There is no increase in infection rates if the hair is not removed. Subsequently, various groups have published their data on surgery without shaving.26–28 A prerequisite to prevent infection is thorough preparation of the site before surgery with shampooing and antiseptic cleansing. Furthermore, in closing the wound, the surgeon must ensure that no hair remains in the wound, thus making the final part of the procedure more involved. Staples and interrupted sutures26–28 are used to close the skin. The studies were carefully designed and should be read attentively before a “no-shave” policy is adopted. Bekar et al27 administered antibiotics for 3 days after surgery. This practice may have resulted in a skewed representation of the true infection risk posed by not shaving. Tokimura et al28 administered antibiotics for 24 hours and had the patients’ hair washed on the second, fourth, and sixth postoperative days. Accordingly, all studies that pertain to “no-shaving” routines instituted additional safety measures. Thus, “not shaving” poses no additional risk for infection if proper precautions are taken.

Estimating the Risk for Surgical Site Infection: NNIS Risk Index

In an effort to identify patients with a susceptibility to postoperative infection, the CDC (National Nosocomial Infections Surveillance [NNIS] system) developed a risk index, which has been validated and found to correlate well with infection rates.15 The risk index is based on an ASA (American Society of Anesthesiologists) preoperative assessment, the wound category, and the duration of surgery (75th percentile of the NNIS data for craniotomies rounded to the nearest whole number of hours: 4 hours). The index is divided into four categories (0 to 3), which correlate with increasing risk for postoperative wound infections. The categories are calculated by adding one point for each of the following when present: an ASA score of 3, 4, or 5; contaminated or dirty wounds; and a duration of surgery exceeding 4 hours. This NNIS risk score allows an estimate of the risk for developing an SSI.1,10,20

Surgical Site Infections in Postoperative Cranial Surgery

To distinguish general postoperative from wound infections, the latter are appropriately termed SSIs.19 The neurosurgical site is divided into extra- and intradural parts. Extradural infections involve every level from the skin to the dura mater, including the bone flap. Intradural infections can be localized to the subdural space (empyema) or to an intracerebral location (abscess), or they can be generalized (meningitis and ventriculitis). To allow a more thorough analysis of SSIs, the CDC introduced a classification19 that differentiates between incisional (superficial and deep) and organ-related SSIs. In neurosurgery, incisional SSIs represent extradural infections. Organ-related SSIs include every intradural infection (subdural empyema, intracerebral abscess, meningitis, and ventriculitis). In the CDC definition, organ SSIs are specified as “all infections… [which]… involve any part of the anatomy other than the incision opened or manipulated during the operative procedure,”19 which includes the cranial opening. Thus, epidural abscess and bone flap infection are considered deep wound infections, whereas osteomyelitis as a CDC organ/space SSI refers to orthopedic surgery.

Although this clear categorization of location can facilitate appropriate reporting, a gray zone exists for potential under-reporting. In the neurosurgical literature, a particularly pragmatic approach is taken whereby only those SSIs that required reoperation are considered. This practice covers only those complications with immediate serious consequences, whereas “near-complications” such as superficial SSIs treated on the ward (e.g., additional stitches, lumbar drainage for CSF collections without leakage) may be incompletely reported. However, these near-complications or incidents allow us to understand and address the issue of SSI more comprehensively. In risk management and prevention, special emphasis is given to near-incidents, which are more frequent than actual events and therefore represent a larger database. This more extensive pool of information facilitates the analysis of potential risk factors and the assessment of surgical and preventive techniques.

Additional concern exists that SSIs may be underreported because of cases that are not detected by the surgeon,14 as well as infections that occur after discharge7,10,16,35 and are handled either on an outpatient basis or at other institutions. This is particularly relevant for infection of the bone flap, which may manifest months after surgery.6 With shorter hospital stays and increased patient transfer to specialized centers,12 referral and admission patterns may indeed result in problems with accurately reporting SSIs. Patient transfer, which is less developed in Europe, has to be considered when infection rates9 are compared by countries or continents.

The CDC criteria for wound infection include “purulent discharge” and the “surgeon′s impression” of the wound as key features. Infections present mostly with local signs of inflammation. Fever is an explicit symptom, but in milder courses of infection, temperature elevation can be either absent or slight. Neurologic symptoms in intracranial infections range from headaches to altered mental status and focal deficits.

Currently, the most useful postoperative laboratory parameter is the C-reactive protein (CRP) level.36–38 Although elevated levels are not surprising after surgery (correlating with tissue damage), sustained elevation beyond the fourth day or renewed elevation after normalization should raise suspicion for infection. In the presence of neurologic symptoms, imaging is indicated to assess for the presence of intracranial suppuration. Computed tomography (CT) is the most readily available imaging modality. Contrast-enhanced CT provides a firm basis for the diagnosis of intracranial and epidural infectious complications. Although nonspecific contrast enhancement may appear a few days postoperatively, when considered together with clinical suspicion and laboratory findings, the results rarely remain ambiguous. Generally, infections present as fluid collections with pronounced ring enhancement. To plan the surgical procedure, the extent of the infection must be defined. The increased availability of magnetic resonance (MR) imaging provides further structural and pathophysiologic (perfusion) information.

Most often, the offending organisms are gram-positive skin flora (i.e., coagulase-negative staphylococci, S. aureus, and Propionibacterium acnes). Various reports on rare and resistant bacterial strains emphasize the importance of adhering to the general principles of antibiotic prophylaxis and of coordinating treatment closely with microbiologists. Fungal SSIs are rare but should not be disregarded.

Although incisional and organ SSIs manifest at different time intervals after surgery, the range is too wide to provide a solid basis for differential diagnosis. In a population of 2,944 patients in a multicenter study,10 scalp infections (5 ± 9 days; median, 13 days) and meningitis/ventriculitis (10 ± 8 days; median, 7 days) occurred earlier than intradural infections (empyema and abscess, 25 ± 31 days; median, 15 days), with the longest latency seen for bone flap infection (42 ± 56 days; median, 27 days). In a single-institution2 follow-up study with a larger population of 4,578 patients, the authors corroborated the order with time to manifestation—with, however, an even larger range (mean time between surgery and onset of infection in days: meningitis, 12 ± 14; scalp infection, 23 ± 51; brain abscess or empyema, 25 ± 27; bone flap, 118 ± 157), reemphasizing the long latency period for bone flap infection.