JSTOR

Case Reports and Case Series Investigation

The first patient was a 27‐year‐old man who presented to hospital A on November 26, 2004, with a temperature of 39°C, headache, vomiting, and nuchal rigidity. The patient (patient 1) had undergone arthroscopy under spinal anesthesia 24 hours previously in another institution (hospital B). Lumbar puncture was performed on admission, and the cerebrospinal fluid (CSF) was found to have 6,000 white blood cells (WBCs) per mL, of which 95% were polymorphonuclear (PMN) cells; the glucose level was 16 mg/dL, and the protein level was 118 mg/dL. The presumptive diagnosis was bacterial meningitis, for which the patient was treated with intravenous vancomycin and meropenem. He showed rapid clinical improvement within 24 hours. Cultures of CSF specimens grew no pathogens.

The infectious diseases consultant recalled that 4 years earlier he had treated a similar case of bacterial meningitis that presented 24 hours after arthroscopy performed under spinal anesthesia. This patient (patient 2) had had surgery at the same institution as the index patient. The causative organism in that case had been Streptococcus salivarius. The infectious diseases consultant thus notified the Haifa District Health Office of the index patient's case. In view of the long interval between the 2 cases, and in the absence of an identified pathogen in the index case, it was decided that there was no indication to proceed with an investigation.

Patient 3 was a 70‐year‐old man who had undergone arthroscopy in hospital B on February 24, 2005. On February 26, 2005, the Haifa District Health Office received a report of meningitis in this individual. The patient had been admitted to hospital A 24 hours after the surgical procedure with headache, nausea, and a confusional state. His peripheral WBC count was 24,800 WBCs/mL. Lumbar puncture revealed turbid CSF with 15,750 WBCs/mL, of which 84% were PMN cells; a glucose level of 5 mg/dL (peripheral glucose level, 147 mg/dL); and a protein level of 1,065 mg/dL. The clinical picture was consistent with bacterial meningitis, and the patient was treated with intravenous vancomycin and meropenem. Within 24 hours, his clinical condition improved. After 2 weeks, he was discharged in good condition. Bacterial cultures of CSF specimens grew no pathogens. In a further attempt to identify the pathogen, DNA was extracted from the CSF samples using a commercial kit (QIAamp DNA Mini Kit; Qiagen). Amplification was then performed for a variable segment from the center of the gene encoding for 16S rRNA, using primers that identify universally conserved sequences. The resulting nucleotide sequences were compared with sequences in public databases by means of BLAST analysis software (Basic Local Alignment Search Tool; National Center for Biotechnology Information) and found to be homologous with Streptococcus salivarius.¹⁰

Epidemiological investigation was initiated at this point. The medical record of the patient who was diagnosed in February 2000 (patient 2) was obtained and reviewed. He, too, had presented with a clear clinical picture of bacterial meningitis less than 24 hours after undergoing arthroscopy in hospital B. Culture of CSF specimens grew S. salivarius, and he was discharged in good condition after treatment. The 3 patients had all undergone arthroscopy at the same medical center but with different surgeons. Spinal anesthesia had been performed for all 3 by the same anesthesiologist. An initial review of standard operating room procedures used at the surgical center was unremarkable. Samples of povidone‐iodine solutions used for disinfection and the spinal anesthetic solutions were sent for culture and grew no pathogens. In the absence of any significant findings, no further interventions were made.

A 59‐year‐old male (patient 4) was admitted to hospital C in early May 2005, who had undergone lithotropy under spinal anesthesia at hospital B 24 hours previously. On admission he was febrile (temperature, 38.7°C), obtunded, and had nuchal rigidity. Lumbar puncture revealed turbid CSF with 15,350 WBCs/mL, of which 99% were PMN cells; a glucose level of 75 mg/dL (peripheral glucose level, 228 mg/dL); and a protein level of 350 mg/dL. He received parenteral antibiotics and demonstrated clinical improvement within 24 hours. Cultures of CSF and blood grew no pathogens. Further testing showed the CSF to be positive for bacterial DNA, which was identified as belonging to S. salivarius by sequencing of the 16S rRNA gene.

Patient 5 was a 25‐year‐old man admitted to hospital C on the day after patient 4 was admitted; he had undergone arthroscopy performed under spinal anesthesia at hospital B the day before. Ceftriaxone had been administered intraoperatively. The patient developed nausea and vomiting, and he was febrile and confused on admission. CSF examination revealed 5,200 WBCs/mL, of which 96% were PMN cells. The CSF glucose level was 36 mg/dL. He, too, demonstrated clinical improvement within 24 hours after initiation of antibiotic treatment. CSF and blood cultures was negative for pathogens.

Because CSF specimens obtained from all of these recent cases yielded no pathogens on culture, we suspected that there might have been other, unreported cases. Meningitis is a reportable disease in Israel, but there is probably significant underreporting of cases that are culture negative, as well as underreporting of cases not caused by Neisseria meningitidis or Haemophilus influenzae b. S. salivarius infections are not reportable by law. Therefore, we actively searched for other possible unreported cases.

Patient 6 was a 58‐year‐old woman admitted to hospital D on April 4, 2004, with fever, obtundation, and nuchal rigidity. CSF findings were consistent with acute bacterial meningitis. The patient had undergone a procedure for varicose vein stripping 24 hours earlier, which was performed under spinal anesthesia at hospital B, the same surgical center where all the other case patients underwent surgical procedures. She was treated with parenteral antibiotics and recovered. Blood and CSF cultures grew no pathogens.

One additional patient with culture‐positive S. salivarius meningitis was identified. The patient had been diagnosed with meningitis at a fourth hospital (hospital D) in 1999 after having undergone surgery with spinal anesthesia at hospital D. No connection with the present series was found. No other cases of S. salivarius meningitis had been reported in Israel in the previous 3 years.

All 6 patients with meningitis described in the present study had presented within 24 hours after a surgical procedure performed under spinal anesthesia. Symptoms and CSF findings were clinically consistent with bacterial meningitis, yet only 1 of the patients had a CSF culture positive for S. salivarius. The Table summarizes the key features of the cases.

Table.  Summary of a Series of 6 Patients With Meningitis After Spinal Anesthesia

Open New Window

A second investigation was performed at hospital B that included prolonged observation of surgical and anesthetic technique. We found that a mask was not used for every lumbar puncture. When masks were used, they were put on in the morning and not changed for every procedure. They were also worn without closing the chin strap. Demonstration of the puncture technique revealed inadequate sterilization of the puncture site. The puncture site was rapidly washed for less than the recommended 30 seconds with povidone‐iodine solution and swabbed dry, rather than being allowed to air dry. All cases of meningitis had followed procedures that were performed at midday in a private institution where the surgical schedule was intense with little time between procedures. Medical record review revealed that 5 of the 6 patients had received intravenous antibiotics intraoperatively, which could account for the lack of pathogen growth on culture. Medical record review as well as patient interviews revealed that nonsteroidal anti‐inflammatory drugs that could have been responsible for aseptic meningitis were not used.

Surgical procedures were halted at hospital B until all staff were provided with in‐service training to improve aseptic technique and infection control in the operating theatre. The hospital's administration was instructed to institute a formal infection control team to monitor procedures and perform follow‐up of cases. No further cases have been reported.

Discussion

The cases of meningitis following spinal anesthesia described above appear to be one of the largest series reported. Despite the absence of positive culture results for patients except patient 1, whose case occurred in 2000, or 4 years before the others, the association with a single anesthesiologist during a relatively short interval and the finding of CSF positive for bacterial DNA that was consistent with S. salivarius lead us to believe that these cases are related in etiology. The use of intraoperative antibiotic therapy (which was unnecessary and not indicated) probably masked the growth of the pathogens responsible and impeded the identification of the etiologic connection between the cases.

Clusters of meningitis associated with individual anesthesiologists have been previously described.¹¹ Schneeberger et al.¹² reported 4 cases that occurred over more than 2 years that were related to an anesthesiologist who did not use facial masks during procedures. It was thought that the cluster was probably related to certain characteristics of the anesthesiologist that favored the organism’s dispersal.

While contamination of the needle has been mentioned as a possible source of infection, skin contaminants are rarely reported as causative pathogens in cases of infection after lumbar puncture. Raedler et al.¹³ found a 16.7% rate of needle colonization after the administration of spinal anesthesia and a 25% rate after the administration of epidural anesthesia. The organisms identified were Staphylococcus aureus, coagulase‐negative staphylococci, micrococci and enterococci. They documented no infections and no Streptococcus contamination.

The overwhelming prevalence of viridans streptococci as the pathogens responsible for meningitis after spinal procedures suggests that respiratory droplet contamination involving commensals of the oral cavity and upper respiratory tract is the mechanism of this infectious complication. Specifically, S. salivarius is the most predominant species detected on the tongues of healthy subjects.¹⁴ Droplet or smear contamination of the needle used for lumbar puncture was postulated as the most probable source of contamination in 3 cases reported by Trautmann.¹⁵ It was hypothesized that the surgical masks worn by the anesthesiologists were open at the neck or were not covering the nose. The culture of identical organisms from the nasopharynx of an anesthesiologist and a patient treated by that anesthesiologist who developed meningitis following myelography also point to droplet contamination as the source of infection.¹⁶

Phillips et al.¹⁷ found that facial masks were effective at reducing forward dispersal of bacteria. Masks that had been worn for 15 minutes were less effective than fresh masks at protecting against contamination, although the difference was not found to be statistically significant. Masks have also been shown to reduce the amount of vertical shedding of bacteria, even when the person was talking and turning his head.^18,19 While good masks may be effective as bacterial filters for up to 8 hours, it is recommended that they be changed for each new procedure.³ Even those experts arguing against the routine use of masks for all lumbar punctures suggest that they be used during high‐risk procedures or when circumstances favor dispersal of bacteria, such as procedures that involve frequent talking and procedures of long duration.²⁰ The Centers for Disease Control and Prevention recommends that surgical masks be worn for all procedures in surgical theaters.²¹

In addition to the importance of preventing iatrogenic infections by the use of proper aseptic technique, it is critical that hospitals, infection control teams, and health departments recognize outbreaks and investigate them appropriately in real time. Because the infectious complications of a surgical procedure may be diagnosed at an institution other than the one where the procedure was performed, as in the series we describe here, it is imperative to report such cases promptly so that they can be properly investigated. We suggest that the identification of even a single case of suspected iatrogenic meningitis should be followed up with a thorough examination of the techniques that might have contributed to its cause, via direct observation of practice. Had this been done when the earliest case we describe was identified (2000), the violations of aseptic technique might have been identified and corrected much earlier, possibly preventing the subsequent cases.

Finally, we wish to point out the importance of using DNA fingerprinting techniques in the investigation of outbreaks.²² In the absence of positive culture results, the findings of bacterial DNA consistent with S. salivarius in 2 different CSF specimens were important in identifying the connection between the cases described above, particularly in light of the first positive culture result 4 years earlier.