Chapter 144 Management of Infections After Craniotomy
Incidence
Infection rates after craniotomy as reported during the 20th century are shown in Table 144–1.1–24 Such percentages, however, must be viewed with skepticism, given the great variation in obtaining and conducting data. The most reliable data are obtained in connection with research projects, in which an investigator, using strictly specified criteria, scrutinizes all patient records, complete with questionnaires after discharge. Inclusion criteria, detection methods, and cohort base may vary (e.g., number of operations, number of patients operated on, number of patients discharged, or number of patients admitted) from study to study. A continuous infection registry run by a hygienist, although not as reliable, is more useful for practical purposes because it allows detection of changes from year to year and even from month to month; for comparison between hospitals, it is of less value. The registry is important in keeping track of bacterial strains and resistance patterns, with special reference to highly resistant strains, such as methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus.
| Author | Year | Percentage |
|---|---|---|
| von Eiselsberg and Ranzi1 | 1913 | 12 |
| Tooth2 | 1913 | 12 |
| Olivecrona3 | 1934 | 15 |
| Cushing and Eisenhardt4 | 1938 | 2 |
| Cairns5 | 1939 | 2 |
| Woodhall et al6 | 1949 | 1 |
| Wright7 | 1966 | 6 |
| Skultety and Nishioka8 | 1966 | 2 |
| Balch9 | 1967 | 5 |
| Green et al10 | 1974 | 3 |
| Chou and Erickson11 | 1976 | 3 |
| Cruse12 | 1977 | 2 |
| Quadery et al13 | 1977 | 6 |
| Chan and Thompson14 | 1984 | 7 |
| Jomin et al15 | 1984 | 3 |
| Vlahov et al16 | 1984 | 7 |
| Blomstedt17 | 1985 | 8 |
| Tenney et al18 | 1985 | 4 |
| van Ek et al19 | 1986 | 8 |
| Rasmussen et al20 | 1990 | 6 |
| Gaillard and Gilsbach21 | 1991 | 9 |
| Mindermann et al22 | 1993 | 9 |
| NNIS 23 | 1996 | 1 |
| Korinek AM 24 | 1997 | 4 |
∗ When the article deals with prophylactic antibiotics, the infection rate for the group without antibiotics is given in this table. All figures are rounded to the nearest full number.
Factors Affecting Infection Rate
Infection rates fell sharply after the introduction of antisepsis and asepsis, but complete eradication of infections seems impossible to achieve. Human defense mechanisms usually have the edge over bacteria always present in the environment, and infection occurs as a result of a faltering balance among defenses, bacterial quantity, and bacterial virulence. Defenses are weakened with age, poor physical or mental condition, devitalized tissue, and foreign bodies. The quantity of bacterial contamination increases with large surgical fields, with reoperation, with long operating times, with inadequate aseptic technique, and with poor ventilation of the operating room.25 Efforts to pinpoint single factors predisposing to infection are rarely convincing: Isolation of one particular factor may be possible in laboratory conditions but is nearly always unreliable in a clinical situation; the impact of most factors is so small that extremely large groups are required to show statistical significance. Many recommendations are difficult to prove, such as that a shunt operation should be done as the first procedure in the morning, that only a minimum of traffic should be allowed in the operating room, and that infected patients should be operated on as the last procedure of the day. In fact most measures usually turn out to be nonsignificant when singled out. External CSF drainage is, however, a factor shown to predispose to infection.26–27 Most hospitals have their own routines, and although different, one may be just as good as the other. Nonetheless, it is probably advantageous to have a fairly rigorous scheme to be followed by all and taught to newcomers, the central issue being to minimize contamination.
Postoperative Cerebrospinal Fluid Leak
Cerebrospinal fluid (CSF) leak after surgery and operation in a patient with a concurrent non-CNS infection are, according to a retrospective study by Mollman and Haines25 on 9202 neurosurgical patients, factors increasing the risk of infection. In a retrospective study by Blomstedt17 of 1039 intracranial procedures (shunts excluded), postoperative CSF leak was also found to be the only highly significant risk factor: Of 15 patients with a CSF leak, 6 developed infection versus 54 patients among 1024 patients without a CSF leak (p = .0002, Fischer’s exact test). Every effort should be made to avoid CSF leak and to seal the leak if it occurs. If the dura cannot be completely closed, dural repair can usually be done with tissue obtained within the operating field, such as periosteum or fascia or, when big patches are required, fascia lata. Fat is quite efficient, especially for drilled-out cavities in the pneumatized bone. Fibrin glue resorbs too quickly to allow natural membranes to form, but it is helpful for fixation of the patch.28 The risk of accidental introduction of prions or other infectants through dura substitutes prepared from animals or cadavers has been reduced considerably through preparations of completely acellular material. Because dural substitutes are avascular, contamination during surgery can result in an infection that is difficult to eradicate. This haven for bacteria is, however, time limited when biodegradable material is used. 29,30,31,32
If CSF leak occurs after surgery, lumbar taps or draining may be helpful. Lumbar drainage causes reduction of CSF pressure, reducing the flow of CSF and allowing the fistula to scar. CSF drainage can, however, cause retrograde flow through the fistula, leading to pneumocephalus and an increased risk of infection. In such cases, early surgical repair is advisable. Persistent CSF leak can also result from low levels of factor XIII.33 In difficult cases, especially after repeated operations and radiotherapy, a microvascular tissue transplant may be necessary to seal the leak. If possible, a locally harvested piece of muscle, with the vasculature intact, is convenient, but a piece of anterior serratus muscle and its vessels can often be anastomosed to the superficial temporal artery and vein.34
Perioperative corticosteroids are widely used in connection with craniotomy to reduce postoperative brain swelling. These agents can have a negative effect on the healing of dura mater and arachnoid.35
The value of prophylactic antimicrobials with a CSF leak is debated.36 If one decides to use an antimicrobial in these circumstances, an agent is selected that is effective against the most likely organisms in the affected area (e.g., if leaking through the nose, the flora of the upper respiratory tract, such as Diplococcus pneumoniae and Haemophilus influenzae; if leaking through the skin, Staphylococcus epidermidis, S. aureus, and Propionibacterium acnes).17 Because CSF fistula increases the risk of intracranial infection substantially, antimicrobial prophylaxis seems justified, despite the lack of evidence of its effectiveness. In cases of CSF rhinorrhea, we use an oral penicillin, which covers most bacteria in the nose but causes few resistance problems. In CSF leak through the skin, we use a first-generation cephalosporin. The use of prophylactic antibiotics in this concept is controversial because the flora can change. When deciding to use prophylaxis, the regimen should be brief and the spectrum should be narrow.
Prophylactic Antimicrobials in Craniotomy
The value of antimicrobial prophylaxis in clean craniotomy is well documented in several controlled studies,21,22,37–44 although other studies do not support the prophylactic use of antimicrobials in this circumstance.24–45 A 2006 study indicates that prophylaxis indeed protects against wound infection but has no effect on postoperative meningitis.46 Negative effects of antimicrobial prophylaxis are possible induction of resistant organisms, inadvertent slackening of aseptic discipline, idiosyncratic reactions to the drug, and expense of the drug. Because the recommended regimen involves a single perioperative dose, only the issue of slackened aseptic discipline and the rare idiosyncratic reaction are relevant. Most investigators use a second-generation or third-generation cephalosporin, vancomycin, or fusidic acid. Because the prophylaxis is targeted at the most likely organisms, an antimicrobial with a fairly narrow spectrum can be selected. In our study,40 we chose vancomycin for three reasons: (1) At the time, bone flap infection was the commonest infection in the department (Table 144–2), and vancomycin covered well the likely agents staphylococci and P. acnes; (2) vancomycin was used sparingly; and (3) resistance against vancomycin was extremely rare, although resistance was reported before our study was finished.47 This single-dose vancomycin regimen brought about a substantial reduction of bone flap infections.40 Circumstances have changed since the mid 1980s in that the use of vancomycin has increased and resistance is reported more often. However, in our institution the prophylaxis in craniotomies is unchanged because no increase in the rate of bone flap infections have occurred. There is no antimicrobial agent that can be recommended universally. Each neurosurgical department needs to establish its types of infections, infection rate (prophylactic antimicrobial treatment is recommended if it is greater than 3%), and the likely bacteria, and the routines must be changed from time to time, if need be.
Postcraniotomy Infections
Central Nervous System Defenses Against Infections
Bacteria do not readily cross the blood-brain barrier and blood-CSF barrier, but when these defenses are breached, the CSF is ill equipped to mount an efficient defense against bacteria. Such a defense requires high-affinity or type-specific antibodies. The complement system reacts with the antigen-bacteria complex to induce lysis, and opsonins combine with bacteria and bacterial fragments and render them susceptible for phagocytosis by polymorphonuclear cells. In normal CSF, the levels of specific antibodies, complement, and opsonic proteins are low or absent, and polymorphonuclear cells are slow to reach the site of infection. If bacteria multiply rapidly, the opsonin supply can be quickly exhausted. Antimicrobial therapy, especially with a bacteriostatic antimicrobial, falls short in the absence of an efficient host response. The opsonic capacity seems to correlate with the leukocyte count of the infected CSF. A vigorous inflammatory response aids in the fight against infection by increasing penetration of opsonic protein and polymorphonuclear cells, but inflammation also causes cortical damage.
Most bacterial species (e.g., Pneumococcus, Haemophilus, Meningococcus, Klebsiella) that commonly cause meningitis possess a capsule that hampers opsonization, which in this situation is important for bacterial virulence. In contrast, the unencapsulated S. aureus rarely causes meningitis but commonly causes bacteremia, which the encapsulated bacteria rarely do.48–53
Antimicrobial Penetration into Cerebrospinal Fluid
Several factors limit antimicrobial penetration into the CSF. The capillary bed of the choroid plexus and the meninges are nonfenestrated, and all organic compounds must pass through a lipid membrane. The lipid solubility of the antimicrobial agent is a penetration factor. Ionization of the antimicrobial is another factor: The more ionized, the less lipid soluble. The pKa of penicillin G is 2.6, and at physiologic pH, the degree of ionization is high. The pH of plasma is normally 7.4 and of CSF is 7.3, and the pH gradient between plasma and CSF favors outflux of lipid-soluble antimicrobials from the CSF. In purulent meningitis, acid metabolites lower the pH of CSF, which increases this gradient and reduces antimicrobial penetration. Other mechanisms associated with meningeal inflammation increase penetration into the CSF several fold. A high degree of antimicrobial protein binding reduces CSF penetration, as do high molecular weight and complex molecular structure. Antibiotic transportation occurs passively, but there are also active bidirectional transport mechanisms.54–6061
Infections
Bacterial Meningitis
Postoperative bacterial meningitis is not common, but it can cause permanent brain damage or death. External CSF drainage devices are commonly used in the intensive care unit and are risk factors for meningitis (odds ratio [OR], 21.8; p = .001), as are extended antimicrobial treatment and infections elsewhere in the body.62,63,64 Staphylococci do not cause spontaneous meningitis but do produce postoperative meningitis, as do gram-negative bacilli.17,65–67,26,46 A rise in body temperature is common in the postoperative period. If the temperature is very high, persistently high, rising, or fluctuating, particularly with a rising C-reactive protein, stiff neck, deteriorating consciousness, new neurologic symptoms, or a CSF fistula, meningitis should be suspected. If the level of consciousness is altered, a computed tomography (CT) scan is done before lumbar puncture. This is to rule out an abscess, as well as other abnormal findings such as hydrocephalus or bleeding. The interpretation of CSF leukocyte criteria for postoperative meningitis is somewhat arbitrary, especially because CSF at this stage usually contains a lot of red blood cells. In our institution we have used the limit 100 × 106/L with a minimum of 50% polymorphonuclear cells or 400 × 106/L regardless of the polymorphonuclear percentage.68
