Dural puncture has long been considered a risk factor in the pathogenesis of meningitis. Exactly how bacteria cross from the blood stream into the spinal fluid is unknown. The presumed mechanisms include introduction of blood into the intrathecal space during needle placement and disruption of the protection provided by the blood–brain barrier. However, lumbar puncture is often performed in patients with fever or infection of unknown origin. If dural puncture during bacteremia results in meningitis, definite clinical data should exist. In fact, clinical studies are few and are often antiquated.

Initial laboratory and clinical investigations were performed over 80 years ago (Table 9.3) [33–39]. In 1919, Weed et al. demonstrated that lumbar or cisternal puncture performed during septicemia (produced by lethal doses of an intravenously administered gram-negative bacillus) invariably resulted in a fatal meningitis [39]. In the same year, Wegeforth and Latham [38] reported their clinical observations on 93 patients suspected of having meningitis who received a diagnostic lumbar puncture. Blood cultures were taken simultaneously. The diagnosis was confirmed in 38 patients. The remaining 55 patients had normal cerebrospinal fluid (CSF). However, 6 of these 55 patients were bacteremic at the time of lumbar puncture . Five of the six bacteremic patients subsequently developed meningitis. It was implied, but not stated, that patients with both sterile blood and CSF cultures did not develop meningitis. Unfortunately, these lumbar punctures were performed during two epidemics of meningitis occurring at a military instillation, and it is possible that some (or all) of these patients may have developed meningitis without lumbar puncture. These two historical studies provided support for the claim that lumbar puncture during bacteremia was a possible risk factor for meningitis.

Subsequent clinical studies reported conflicting results. Pray [35] studied the incidence of pneumococcal meningitis in children who underwent a diagnostic lumbar puncture during pneumococcal sepsis. The incidence of meningitis was no greater among patients who were subjected to lumbar puncture, which produced normal CSF (8 of 30 patients, or 27 %), than among those who did not undergo diagnostic spinal tap (86 of 386 patients, or 22 %). Eng and Seligman retrospectively reviewed the records of 1089 bacteremic patients, including 200 patients who underwent lumbar puncture [34]. The authors reported that the incidence of meningitis after lumbar puncture did not significantly differ from the incidence of spontaneous meningitis and concluded: “If lumbar puncture induced meningitis does occur, it is rare enough to be clinically insignificant.”

However, not all studies have been as reassuring as those described earlier. In a review of meningitis associated with serial lumbar punctures to treat posthemorrhagic hydrocephalus in premature infants, Smith et al. attempted to identify risk factors [36]. Six of 22 (27 %) infants undergoing multiple (2–33) therapeutic dural punctures during a period of 2–63 days developed meningitis. Bacteremia, a risk factor for meningitis in this report, was associated with central venous or umbilical artery catheters. However, 11 septic infants who underwent dural puncture did not develop meningitis. The number of dural punctures, incidence of “difficult or traumatic” procedures and use of antibiotics did not differ between infants who developed meningitis and those who did not. A causal relationship between the dural puncture and onset of meningitis was not clear. Teele et al. retrospectively reviewed the records of 277 bacteremic children during a 10-year interval from 1971 to 1980 [37]. Meningitis occurred in 7 of 46 (15 %) children with normal CSF obtained during a bacteremia. However, only 2 of 231 (1 %) children who did not undergo lumbar puncture developed meningitis. These results were significantly different. In addition, children treated with antibiotics at the time of lumbar puncture were less likely to develop meningitis than children who were not treated until after lumbar puncture. The authors admitted that clinical judgment may have allowed the pediatricians to select the child in whom meningitis is developing before the CSF is diagnostic; these patients may appear more ill and thus suggest the performance of a lumbar puncture.

Prevention of lumbar puncture-induced meningitis with antibiotic therapy is supported by a more recent animal study. Carp and Bailey investigated the association between meningitis and dural puncture in bacteremic rats [40]. Twelve of 40 rats subjected to cisternal puncture with a 26-gauge needle during an E. coli bacteremia subsequently developed meningitis. Meningitis occurred only in animals with a blood culture result of ≥50 colony forming units/mL at the time of dural puncture, a circulating bacterial count observed in patients with infective endocarditis. In addition, bacteremic animals not undergoing dural puncture, as well as animals undergoing dural puncture in the absence of bacteremia did not develop meningitis. Treatment of a group of bacteremic rats with a single dose of gentamycin immediately prior to cisternal puncture eliminated the risk of meningitis; none of these animals developed infection.

This study demonstrates that dural puncture in the presence of bacteremia is associated with the development of meningitis in rats, and that antibiotic treatment before dural puncture reduces this risk. Unfortunately, this study did not include a group of animals that were treated with antibiotics after dural puncture. Since many surgeons defer antibiotic therapy until after cultures are obtained, the actual clinical scenario remains unstudied. There are several other limitations to this study. While E. coli is a common cause of bacteremia, it is an uncommon cause of meningitis. In addition, the authors knew the sensitivity to the bacteria injected, allowing for appropriate antibiotic coverage. The authors also performed a cisternal puncture (rather than lumbar puncture) and utilized a 26-gauge needle, producing a relatively large dural defect in the rat compared to humans and no local anesthetic was injected. Local anesthetic solutions are bacteriostatic, which may theoretically reduce the risk of meningitis in normal clinical settings. While these results may apply to the performance of spinal anesthesia in the bacteremic patient, they do not apply to administration of epidural anesthesia in the febrile patient, which is associated with a higher incidence of vascular injury and typically involves placement of an indwelling foreign body.

Even when meningitis occurs temporally after spinal anesthesia, it is often difficult to establish a cause-and-effect relationship. The following case report describes a probable case of lumbar puncture-induced meningitis [41]. A 60-year-old man underwent kidney stone removal under general anesthesia. On postoperative day six, the patient remained afebrile, but was taken to the operating suite for transurethral clot evacuation. Spinal anesthesia was performed under aseptic technique. Cerebrospinal fluid was clear. Forty minutes later, shaking chills developed. Initial blood and urine cultures were negative. The following day, the patient became febrile and complained of headache and back pain and appeared confused. CSF examination revealed cloudy CSF with a leukocytosis (80 % polymorphonucleocytes), decreased glucose concentration consistent with bacterial infection, but no growth on culture. Three days later, a repeat lumbar puncture was performed with similar results. A third lumbar puncture was performed 2 days later; culture yielded group D streptococcus (enterococci). Group D enterococci are unusual sources of meningitis. In this case it is possible, though unlikely, that the patient was bacteremic prior to administration of the spinal anesthetic. It is more likely that the bacteria entered the blood stream during bladder irrigation (since bacteremia occurs in perhaps 60 % of urologic procedures), and traversed the dura at the puncture site, similar to the animals in the study by Carp and Bailey [40].

Bacterial meningitis can also present after epidural blockade with or without a localized epidural abscess [3, 42]. Ready and Helfer described two cases of meningitis following the use of epidural catheters in parturients [3]. In the first case, a healthy 28-year-old parturient underwent lumbar epidural catheter placement for elective cesarean section. The epidural analgesia was provided for 48 h postoperatively with an opioid. At the time of removal, a 4 cm erythematous indurated area, which was tender to palpation, was noted at the catheter entry site. Three days later, the patient complained of severe headache, nuchal rigidity, and photophobia. An area of cellulitis was present at the epidural insertion site. CSF examinations revealed an elevated protein (308 mg/dL), decreased glucose (27 mg/dL), and 3000 leukocytes/μL (73 % polymorphonucleocytes). Culture of the CSF was positive for S. faecalis. Urine and blood cultures were negative. There was no evidence of epidural abscess on MRI scan. Antibiotic therapy was initiated and the patient recovered completely.

In the second case, a lumbar epidural was placed in a healthy 25-year-old parturient. Delivery occurred uneventfully 50 min later, and the catheter was removed. No local inflammation was noted at the catheter insertion site. The patient reported a nonpositional headache and neck stiffness 24 h later. Lumbar puncture revealed elevated protein (356 mg/dL), decreased glucose (5 mg/dL), and 4721 leukocytes/μL (90 % polymorphonucleocytes). CSF cultured positive for S. uberis (a strain of α-hemolytic streptococcus). However, urine, blood, and vaginal cultures also grew the same organism. Antibiotic therapy was initiated, and recovery was complete. The short duration of the indwelling catheter; the lack of physical findings suggestive of infection at the catheter insertion site; and the presence of the organism in vaginal secretions, blood, and urine suggest that the source of the meningitis was most likely hematogenous spread of the infecting organism from the vagina. The case reported by Berman and Eisele [41] and the two cases by Ready and Helfer [3] demonstrate how a cause-and-effect relationship should not be assumed between the regional anesthetic and the CNS infection, but rather other possible sources should be investigated.

Several relevant studies have specifically examined the risk of epidural abscess in bacteremic patients receiving epidural anesthesia and/or analgesia. Few data exist regarding the placement and maintenance of epidural catheters in patients with an infection at a site distant from the neuraxis. Darchy et al. studied 75 patients in the intensive care unit receiving epidural analgesia (median 4 days), including 21 patients with a known localized concomitant infection [43]. Although five patients had catheter insertion site inflammation/erythema (with or without positive epidural catheter culture) the frequency was not increased by the presence of an infectious source distant to the epidural catheter site. However, the authors recommended a meticulous daily inspection of the catheter insertion site and immediate removal of the catheter if both erythema and local discharge are present, as these two signs of local inflammation are predictors of positive epidural catheter colonization/infection.

Jakobsen et al. examined the records of 69 patients with localized infections who had a total of 120 epidural catheters placed, undergoing on average 4 epidural anesthetics with catheters left in place for a mean of 9 days [44]. On 12 occasions the catheter was removed due to local infection, no specific therapy was instituted, and the infection resolved. There was one case of spondylitis, which was not apparently related to epidural catheterization. The retrospective nature of this study and the small number of patients limit the conclusions but suggest that placing an epidural catheter in a chronically infected patient may not be associated with a high risk of epidural infection.

The obstetric patient presents a unique challenge, since the decision to not perform a neuraxial block may result in less than satisfactory analgesia and patient dissatisfaction. Despite these advantages, the anesthesiologist is frequently faced with the management of the parturient with suspected chorioamnionitis, approximately 8 % of whom are bacteremic. Bader et al. investigated the use of regional anesthesia in women with chorioamnionitis [12]. Three hundred nineteen women were identified from a total of 10,047 deliveries. Of the 319 women, 100 had blood cultures taken on the day of delivery. Eight of these had blood cultures consistent with bacteremia. Two hundred ninety-three of the 319 patients received a regional anesthetic, in 43 patients antibiotics were administered prior to needle or catheter placement. No patient in the study, including those with documented bacteremias, had infectious complications. In addition, mean temperatures and leukocyte counts in patients who received blood cultures showed no significant differences between bacteremic and nonbacteremic groups. Goodman et al. also retrospectively reviewed the hospital records of 531 parturients who received epidural or spinal anesthesia and were subsequently diagnosed with chorioamnionitis [17]. Blood cultures were drawn in 146 patients; 13 were positive. Antibiotics were administered before the regional block was placed in only 123 patients, while nearly one-third of patients did not receive antibiotic therapy in the entire peripartum period. As with the study by Bader et al., leukocytosis, fever, abdominal tenderness, or foul smelling discharge was not predictors of positive blood cultures [12]. There were no infectious complications. These authors continue to administer spinal and epidural anesthesia in patients with suspected chorioamnionitis because the potential benefits of regional anesthesia outweigh the theoretical risk of infectious complications. However, the small number of patients with documented bacteremias in both studies defies a definitive statement regarding the risk of CNS infections in patients suspected of chorioamnionitis undergoing regional anesthetic techniques.

Herpes simplex virus type-2 (HSV-2) is an incurable, recurrent disease characterized by asymptomatic periods alternating at variable periods with recrudescence of the genital lesions. The primary infection is associated with viremia and can be accompanied by a variety of symptoms including fever, headache, lymphadenopathy, and, in rare cases, aseptic meningitis. In contrast, recurrent or secondary infections present as genital lesions without evidence of viremia. When obstetric patients present for delivery with evidence of active HSV-2 infection, cesarean section is usually recommended to avoid exposing the neonate to the virus during vaginal delivery. The use of central neuronal block has been considered controversial by some because of the theoretical concern of introducing the virus into the CNS. Although this issue is usually discussed in the context of obstetrical anesthesia, the incidence and prevalence of genital herpes has increased dramatically in the past two decades. Therefore, the theoretical risk of CNS contamination is present in the general surgical population as well.

Bader et al. reviewed management of 169 HSV-2 infected patients undergoing cesarean delivery. Five were classified as having primary infections with the remaining 164 being secondary [45]. General (59), spinal (75), and epidural (35) anesthetic techniques were used. One patient with primary HSV-2 developed transient unilateral leg weakness following bupivacaine spinal anesthesia. The problem resolved within 1 week. While this patient was classified by the obstetrician as having a primary infection, genital lesions had appeared 3 weeks prior to delivery and there was an active lesion at the time of delivery. The number of patients with primary HSV-2 infections was very small.

These recommendations are consistent with previous studies. Crosby et al. reviewed a 6-year experience with active HSV-2 infections in obstetrical patients in two institutions [46]. Cesarean section was performed on 89 affected parturients, all with recurrent herpes disease. There were no neurologic or infectious complications. In a similar retrospective review, Ramanathan et al. reported 43 epidural anesthetics in parturients with HSV-2 infection who had either active lesions (71 %) or had at least one recurrence during the pregnancy [47]. Again, no complications were noted in the parturient or neonate. One patient who was treated prenatally with steroids to promote fetal lung maturity developed a lesion in the postnatal period which resolved within 10 days. Neither of these studies included patients with primary infections.

Herpes simplex virus type-1 (HSV-1) the infectious agent for oral herpes rarely causes genital lesions. However, recurrent HSV-1 has been described in parturients receiving intrathecal and epidural morphine for pain management [48]. The postnatal association is controversial since several other factors such as emotional or physical stress, other infections, and parturition have been cited as causes of recurrent HSV infection. Valley et al. reported a case of thoracic and labial HSV-1 infection in a patient receiving epidural fentanyl [49]. While surgical stress may have been a factor, this patient had no other known risk factors, and lesions developed near the site of the epidural catheter.

The risk of performing regional anesthesia procedures in HIV-infected patients is largely unknown. Hughes et al. reported the safe administration of central neuronal block in 18 HIV-infected parturients [50]. The patients studied showed no postpartum change in immune, infectious, or neurologic status. Avidan et al. and Bremerich et al. also reported a low complication rate for parturients with HIV infection on antiretroviral therapy that underwent spinal anesthesia. However, in all three series (with a combined total of 117 patients), the patients were relatively healthy and in the early stage of their disease [51, 52]. The effects of anesthesia on patients with more advanced disease are unreported.

In a report on the use of epidural blood patch for postdural puncture headache in HIV-positive males, Tom et al. followed nine patients longitudinally for periods ranging from 6 to 24 months [53]. No complications were attributable to the epidural blood patch, although the authors noted the high incidence of neurologic manifestations in this population. Approximately 40 % of patients with the diagnosis of acquired immune deficiency syndrome (AIDS) have clinical signs of neurologic disease and at autopsy, patients with AIDS have a 70–80 % incidence of neuropathologic changes. While many of the neurologic symptoms are unrelated to complications associated with spinal or epidural anesthesia, some such as aseptic meningitis, chronic headaches, and polyneuropathy may be mistaken for problems related to needle placement. A clear understanding of the association of CNS symptoms with HIV infection is important in order to interpret postblock neurologic pathology.