Chapter 72. Infection Control in Regional Anesthesia

Infectious complications related to regional anesthesia are rare. Since the only information is available in case reports and retrospective surveys, it is likely that these complications are underreported. The objective of this chapter is to summarize information from the literature on infections associated with regional anesthesia, as well as to discuss the mechanism and to suggest strategies to prevent these complications.

Microorganisms from exogenous or endogenous sources may gain access to the subarachnoid, epidural, or tissue space surrounding peripheral nerves in several ways. Microorganisms from the patient's or anesthesia practitioner's flora can be inoculated directly when a catheter or needle is inserted into those spaces. Several reports in the literature suggest that infections are on occasion caused by the anesthesia practitioner's flora.1–3 For example, Trautmann and colleagues reported a case of meningitis caused by a Staphylococcus aureus strain that was identical by pulsed-field gel electrophoresis to the S. aureus isolate from the anesthesiologist's nose.2 Microorganisms can also enter the epidural space by hematogenous spread from other body sites, such as infected skin,2,4 or by migrating along the catheter tract.5,6 Several case reports suggested that infection was caused by spread of bacteria from infected sites through the bloodstream to the epidural space.7–9 Others maintain that infections at distal sites are not contraindications to epidural anesthesia. For example, Newman concluded that distal infections did not increase the risk of epidural infection because traumatic injuries are often infected and no epidural catheter-related infections were identified among over 3000 patients who had epidural neural blockades for postoperative or posttraumatic analgesia.10 The anesthetic agents injected into the patient's subarachnoid or epidural space are another possible source of infection. Infections from contaminated multidose vials are likely to be rare because most anesthetic drugs are weak bases dissolved in acidic solutions that inhibit growth of bacteria and fungi11–13; besides most multidose local anesthetic solutions contain a bacteriostatic agent. Nevertheless, the case report by North and Brophy suggests that contaminated multidose vials still can be a source of infection. These authors reported an infection in which S. aureus isolates with matching phage types were recovered from an abscess and a multidose lidocaine vial.1

To assess whether contamination of the anesthetic agent or the equipment (needles, syringes, tubing) is related to subsequent infections, investigators have cultured these items after they have been used with patients or during simulations. In four studies, 0–29% of used catheters were contaminated,14–17 and James and coworkers found that 5 of 101 syringes used to inject anesthetic agents were contaminated.14 Ross and coworkers drew up 0.25% bupivacaine into control syringes and into syringes used to induce continuous lumbar epidural neural blockade (test syringe) in 18 obstetric patients.18 After each dose from the test syringe, the investigators cultured the contents of both the test and control syringes. Six of 18 test syringes were contaminated with bacteria, compared with only 1 of 18 control syringes. Raedler and associates cultured 114 spinal and 20 epidural needles after use for single lumbar injections.19 Twenty-four cultures (17.9%) grew microorganisms: 15.7% coagulase-negative staphylococci, 1.5% yeasts, 0.8% each enterococci, pneumococci, and micrococci. These authors concluded that it is easy to contaminate the needle and that anesthesiologists need to improve their hygienic measures. Despite finding contaminated equipment or anesthetic solutions, these investigators did not identify any infected patients,14–18 and thus, none of the authors were able to correlate contamination with infection.

There are numerous case reports in the literature of infections occurring after epidural neuraxial blockade, attesting to the fact that such complications can be severe (Table 72–1).1,7,9,19–55 Thirty-five of 48 patients in these case reports acquired epidural or intraspinal abscesses. Three patients had injections only, 1 patient had injections and several catheters, and 47 patients had catheters. Among the 38 patients who had catheters and for whom the duration of catheterization was specified, the median duration of catheterization was 3 days (range 50 min to 6 weeks). The median time to onset of the first signs or symptoms of infection was 4 days (range 1–4.8 months) after catheter placement. S. aureus caused 27 of 42 infections from which bacterial pathogens were isolated. P. aeruginosa caused five infections and Streptococcus spp. caused five. Three patients died and 27 nearly or fully recovered.

It should be kept in mind that the number of reported cases does not allow us to assess the true frequency of infections after epidural neural blockade. However, several investigators have done studies to assess this risk. When reviewing 350 reports in the literature, Dawkins found no reports of infection after thoracic or lumbar epidural block, but he identified 8 (0.2%) reports of infection after 3767 sacral epidural blocks used for operative procedures and for obstetrics 1969.56 More recently, Dawson reviewed the literature and found rates of deep infection ranging from 0 to 0.7% and rates of superficial infection ranging from 1.8 to 12%.57

Scott and Hibbard surveyed all obstetrics units in the United Kingdom and identified 1 epidural abscess in approximately 506,000 epidural neural blocks.58 In contrast, Palot and colleagues identified 3 cases of meningitis in 300,000 patients who had undergone epidural blocks.59 Three smaller series of obstetric epidural neural blockades (some 12,000 patients) did not identify any infections.60–62 Similarly, in a recent study by the French SOS group on complications of regional anesthesia Auroy and coworkers did not identify any infections in 29,732 epidural neural blocks given for obstetrical procedures.63 Together, the results of these five studies suggest that 4–5 serious infectious complications (ie, epidural abscesses or meningitis) occur per 1 million obstetric epidural neural blocks.

A number of studies have assessed infections associated with epidural neural blockades done for operative procedures or for short-term pain relief. However, these studies report fewer patients than the studies of epidural neural blockade for obstetric procedures. We summarized findings from nine studies in Table 72–2.47,51,63–69 Brooks and associates found 4 infections among 4832 (0.08%) patients undergoing epidural neuraxial blockade for surgical procedures or for labor and delivery.70 All four infections occurred in healthy young women who underwent cesarean sections; two infections were superficial (0.04%), and two involved the epidural space (0.04%). In contrast, Holt and colleagues reported 53 (1.8%) local infections and 11 (0.4%) central nervous system infections related to approximately 3000 epidural catheters.71,72 The median duration of catheterization was 8 days for patients with local infections and 15 days for those with generalized symptoms (p = 0.01). Catheters removed from patients with clinical symptoms were more heavily colonized than those removed from asymptomatic patients. However, 59 of 78 catheters with positive cultures were removed because patients were symptomatic, suggesting that this observation may have been affected by ascertainment bias.

Given that the number of infections identified in any study has been low, the results reported by investigators who calculated the upper boundaries of the infection risk associated with epidural neural blockade are particularly important because they provide a better estimate of the true risk than do studies that report only the number of infections and the number of procedures. For example, Strafford and coworkers did not identify skin infections or epidural abscesses among 1458 pediatric patients who had epidural analgesia to control perioperative pain.73 These investigators calculated the incidence of clinical infection to be 0 with a 95% confidence interval from 0% to 0.03%, or 3 infections per 10,000 procedures. Auroy and colleagues, as noted previously, did not identify any infections among 29,732 procedures done for deliveries.63 They calculated 95% confidence intervals of 0/10,000 to 1/10,000 procedures. Darchy and associates evaluated 75 patients, 9 (12%; incidence density rate of 2.7/100 catheter days) of whom acquired local infections. None of the patients acquired deep infections.65 Based on these data, Darchy and associates estimated the upper risk of spinal space infections to be 4.8% for catheters that remained in place for 4 days. Of note, these estimates are considerably higher than those of Strafford and coworkers73 and higher even than the rates found by Du Pen and associates among patients with epidural catheters for long-term pain control74 (see next paragraph).

In general, epidural catheters inserted for long-term pain control become infected more frequently than those used for short periods of time. Du Pen and associates identified 30 superficial (9.3/10,000 catheter-days), 8 deep catheter track (2.5/10,000 catheter-days), and 15 epidural space (4.6/10,000 catheter-days) infections among 350 patients who had long-term epidural catheters.74 Similarly, Zenz and colleagues identified two cases of meningitis among 139 patients (1.4%, or 2.1/10,000 catheter-days) treated for pain due to malignancy.75 Coombs reported that 10 of 92 (10.9%) cancer patients acquired local infections, and 2 (2.2%) acquired meningitis.76 Malignancy and reduced immunocompetence might be additional risk factors in the long-term catheter population.

Case reports in the literature indicate that serious infections can occur as complications of subarachnoid neural blockade (Table 72–3).2,3,5,6,77–91 Of the 26 infections reported in these case reports, 8 were meningitis, 4 were epidural abscesses, 2 were soft tissue abscesses, and 2 were infections of a disk or of a disk space. The median time to onset of signs or symptoms of infection was 1 day (range 1 h to 2 months) for all infections and 18 h (range 1 h to 10 days for meningitis). Streptococcal species caused 11 of the 23 infections from which bacterial pathogens were identified, and S. aureus caused 2 infections and Pseudomonas spp. caused 4. Twenty-one patients recovered fully. Compared with infections after epidural neural blockade, infections associated with subarachnoid neural blockade were more likely to be caused by streptococci, and patients were more likely to recover fully. Table 72–4 reviews data from nine studies or reviews, which, if taken together, suggest that the rate of infection is approximately 3.7 per 100,000 subarachnoid neural blockades.63,91–98

At present there are few reports in the literature about infectious complications of using combined epidural and subarachnoid neural blockade (CSE). We identified eight case reports of infections (nine infections) after combined procedures51,99–105 (Table 72–5). The median time to onset of signs or symptoms or infection was 21 h (range 8 h to 9 days) for all infections and 18 h (range 8 h to 3 days) for meningitis.

Signs or symptoms of epidural abscesses were first noted 1–9 days after the procedures. Streptococcal species caused three of six cases of meningitis and S. aureus caused all three epidural abscesses. Eight of nine patients recovered fully. We identified only one study that assessed rates of infection associated with combined procedures. Cascio and Heath identified 1 case of meningitis after about 700 (≈0.1%) combined epidural and subarachnoid neural blockades.99

Continuous regional anesthetic techniques utilizing peripheral nerve blocks have become more popular in recent years for postoperative pain management especially for orthopedic procedures.106,107 Only a few studies have addressed infectious complications related to these procedures. The study by Auroy and coworkers of French anesthesiologists did not identify any infections after 43,946 peripheral blocks.63 Bergman and colleagues identified 1 patient among 368 patients (405 axillary catheters) who had a local S. aureus skin infection in the axilla after 48 h of axillary analgesia.108 The patient recovered fully with antibiotic treatment. Meier and colleagues reported 8 superficial skin infections among 91 patients who had continuous interscalene catheters for an average of 5 days.109 Nseir describes a case of fatal streptococcal necrotizing fasciitis following axillary brachial plexus block.110 Adam reported a psoas abscess complicating femoral nerve block catheter.111

Cuvillion and coworkers obtained cultures of 208 femoral catheters when they were removed after 48 h.112 Fifty-four percent of the catheters were colonized with potentially pathogenic bacteria (71% Staphylococcus epidermidis, 10% Enterococcus spp., and 4% Klebsiella spp.). These investigators also reported three episodes of transient bacteremia, but they did not identify any abscesses or episodes of clinical sepsis.110 None of the groups provided information about the aseptic techniques used for catheter insertion.

Other reports include cases of osteomyelitis following digital blocks113 and hematoma block for fracture repair,114 as well as orbital cellulites from sub-Tenon's anesthesia.115,116 All these reports emphasize the importance of maintaining strict asepsis when performing continuing peripheral nerve blocks.

To date there is no consensus regarding infection control measures that are necessary when administering central neuraxial blockades. Most anesthesiologists agree that they should prepare the patient's skin with an antimicrobial agent, wash their hands with an antimicrobial soap, and wear sterile gloves. However, anesthesiologists disagree about the necessity of other precautions.117–123 For example, several surveys indicate that only 50–66% of anesthesia staff wear masks when performing epidural and subarachnoid neural blockades.124–126

The review of studies on infections associated with epidural anesthesia indicates that there is no consensus regarding patient risk factors for infectious complications of epidural neural blockade.107 Few studies have been designed to assess risk factors for infection associated with epidural or subarachnoid neural blockades, possibly in part because these infections are rare. In fact, we identified only one case-control study that was done to evaluate risk factors for infections associated with epidural neural blockade.127 Dawson and colleagues evaluated epidural neural blockades done for postoperative pain relief and found that procedures done between April and August had a sixfold higher risk than those done during other months (95% CI 1.28–28.12, p = 0.009). The risk of infection was lower if a bag rather than a syringe was used to administer the anesthetic agent (odds ratio 0.17, 95% CI 0.02–1.34, p = 0.05). Of the two risk factors identified by this study, only the latter, use of syringes, could be addressed by practice changes.

Assuming that the respiratory tract of anesthesia personnel could be a source of infection, Philips and associates conducted a simulation to assess the efficacy of masks.128 They seated anesthesia staff in a room with controlled ventilation and asked them to speak directly at blood agar plates placed 30 cm away. The number of bacteria on the plates was significantly lower when masks were worn. However, the clinical significance of this finding is unknown.

Chlorhexidine has been shown to reduce the risk of catheter-associated bloodstream infections significantly compared with povidone–iodine.129 Several investigators have tried to determine whether a particular disinfectant provides more effective skin antisepsis before epidural neural blocks than do other agents.130–133 However, none of the studies were large enough to assess rates of infection; instead the outcomes evaluated were catheter or skin colonization. Kasuda and colleagues randomly assigned 70 patients to have their skin prepared with either 0.5% alcoholic solution of chlorhexidine or 10% povidone iodine.131 After a median of 49 ± 7 h, the investigators removed the catheters and obtained cultures of the insertion sites and catheter tips. There was no difference in rates of positive cultures. Kinirons and associates (the only investigators who reported a power calculation) obtained cultures from catheters removed from 96 children who had epidural catheters longer than 24 h.132 The colonization rate was lower for catheters removed from children whose skin was prepared with 0.5% alcoholic solution of chlorhexidine (1/52 catheters, 0.9/100 catheter days) than for those removed from children whose skin was prepared with povidone–iodine (5/44 catheters, 5.6/100 catheter days) (relative risk 0.2, 95% CI 0.1–1.0). Sato and coworkers enrolled 60 patients who were undergoing back operations under general anesthesia.133 After preparing the site with either 0.5% alcoholic chlorhexidine or 10% povidone–iodine, the investigators obtained skin biopsies. Cultures from skin prepared with the alcoholic chlorhexidine were less likely to be positive (5.7%) than were cultures from skin prepared with povidone–iodine (32.4%; p < 0.01). However, microscopy was as likely to identify bacteria in the hair follicles of skin prepared with the alcoholic solution of chlorhexidine (14.3%) as skin prepared with povidone–iodine (11.8%).

The fact that infections rarely complicate neuraxial blockades suggests that the infection control practices used for these procedures are usually adequate. Given the very low rates of infection associated with epidural and subarachnoid neural blockade, it will be very difficult to prove that additional infection control practices such as wearing masks and using full barrier precautions (ie, the anesthesiologist wears a cap, mask, sterile gloves, and sterile gown and uses a large drape to cover the patient) reduce the risk of infection. However, bacteria that colonize the skin, respiratory tract, or water caused most reported infections after epidural and subarachnoid neural blockades. Masks have been shown to decrease spread of organisms when anesthesiologists are talking.128 Thus, a mask would allow the anesthesiologist to talk to the patient while doing the procedure and could decrease the risk of contaminating the insertion site with oral or respiratory flora. Moreover, in the United States, anesthesiologists usually wear masks during operative procedures, even though they are at distance from the operative site. Therefore, it probably makes sense that anesthesia personnel wear masks while performing those procedures.117 Furthermore, epidural and subarachnoid neural blockades are at least as invasive as placing central venous catheters and the consequences of subsequent infections are at least as bad as those for catheter-associated bloodstream infections. Because the use of full barrier precautions reduces the incidence of catheter-related bloodstream infections,134 it is possible that hygienic measures similar to those used for placing central venous catheters should be used during the placement of catheters that will remain in place for several days or longer.57,135

In addition, we believe that the infusion system should be opened as infrequently as possible and that chlorhexidine should be considered as the agent for preparing the skin before the procedure. Anesthesia personnel should observe their patients closely for signs and symptoms of infection so that infections can be diagnosed and treated immediately. Pegues and coworkers reviewed medical records from 1980 to 1992 of patients who had short-term epidural catheters to identify those who acquired infections. They followed patients prospectively from January 1993 until June 1993.136 In 1990, they introduced a standardized procedure for inspecting temporary epidural catheters. During the entire 12.5-year period, the investigators identified seven infections, all of which occurred after catheters were inspected routinely. The increased incidence of infection could have resulted from ascertainment or misclassification bias associated with the retrospective review or from increased use of epidural catheters for pain management during the later time period. On the other hand, it could indicate that infections were not diagnosed when catheters were not inspected routinely for signs of infection.

Because it may be difficult to draw up opioids sterilely from ampules, some have suggested that these drugs be drawn through a filter into a syringe that is then double-wrapped and sterilized in ethylene oxide.137,138 However, the benefit of such extreme precautions is highly hypothetical. Brooks and colleagues implemented the following precautions for continuous central neuraxial blocks done in their hospital:

1. The anesthesiologist should evaluate whether the infusion can be continued if the system becomes disconnected.

2. The exit site should be dressed with a transparent dressing and should be viewed every 8 h or as necessary.

3. A 0.22-μm, 96-h filter should be included in the infusion system.

4. The tubing, filter, and solution should be changed or the system should be discontinued every 96 h.

5. The primary anesthesia and adjuvant bolus dose should be given with a larger syringe that remains attached to the catheter until a continuous analgesia infusion is begun.70

To date, there are no published guidelines for placing catheters for peripheral nerve blocks. Recently, new catheter systems have been introduced that require anesthesiologists to perform several manipulations while inserting the catheter.139 Given the complexity of these manipulations, it is reasonable to assume that aseptic technique similar to that recommended for placement of central venous catheters is warranted.135

Other investigators have demonstrated that surveillance for surgical site infections and reporting rates to surgeons is one of the most effective methods for reducing the incidence of and consequence of such infections.140,141

Although rare, infectious complications from regional anesthesia and analgesia do occur and can be serious. Additional studies are necessary to define the rate of infection after different types of regional anesthesia and to identify risk factors for infections. Studies of these are necessary for development of appropriate guidelines for placement and care of catheters used for regional anesthesia. Table 72–6 contains our recommendations for means of decreasing the risk of infections related to regional anesthesia procedures. Future studies and a multicenter surveillance system should be beneficial in addressing some of the unanswered questions about infections after regional anesthesia procedures.