van de Beek and de Gans discuss important information related to the treatment of post–dural puncture meningitis (PDPM). They recommend that ceftazidime or cefepime be substituted for the empiric third-generation cephalosporin recommendation in my review.1The former drugs have been found to have greater in vitro activity against Pseudomonas aeruginosa . (However, their reference 54does not mention either P. aeruginosa or cefepime.)
Their admonition against adjunctive steroids in PDPM needs further clarification. Korinek et al. 5(van de Beek and de Gans' reference 3) studied 6,243 consecutive craniotomies to evaluate the effect of preoperative antibiotic prophylaxis on the incidence of postoperative meningitis. They found 50% reduction in incision infections (skin, bone) but no reduction in meningitis for the group that received prophylaxis. van de Beek and de Gans give as the reason for withholding adjunctive dexamethasone in PDPM, that PDPM resembles postcraniotomy meningitis more than it resembles community-acquired meningitis. Postcraniotomy meningitis is clearly different from community-acquired meningitis. Adjunctive dexamethasone would be inappropriate in the presence of a wound infection. But PDPM may have more in common with the community-acquired type than it has with the neurosurgical type. Should not PDPM be considered as a distinguished subset? Each of the three types of meningitis has a different spectrum of causative organisms, but the high incidence of viridans streptococci in the PDPM group (a primarily upper respiratory, mouth, and skin commensal) makes it more like the community-acquired (Streptococcus pneumoniae , N. meningitidis ) than the neurosurgical type (staphylococci [skin commensals] and enterococci). In addition, the complication of the concomitant surgical wound is absent in PDPM. Therefore, would not the desirable antiinflammatory effect of dexamethasone be the same in postpuncture meningitis as in community-acquired meningitis (other things being equal, e.g. , no other infected foci)?
van de Beek and de Gans' reference 7,6covering meningitis due to viridans streptococci, reports seven cases of meningitis due to viridans streptococci—three of which are termed iatrogenic. Two patients had undergone thermocoagulation of the gasserian ganglion, and one had endoscopic sclerotherapy for esophageal varices. These three cases suggest meningitis of the post–dural puncture type (albeit inadvertent). It is not clear whether van de Beek's unpublished cases of meningitis due to viridans streptococci are community-acquired, neurosurgical, or post–dural puncture in type. Two other cases described in van de Beek and de Gans' reference 76died of S. milleri meningitis. The underlying cause for these cases of meningitis is not given, but the mechanism of death in each case6is given as brain herniation accompanied by cerebral edema. Whether adjunctive dexamethasone is appropriate for treatment in puncture-type meningitis needs further study.
I thank Dr. Edsall for drawing attention to the practice, by some, of injection of air into the epidural space as a possible source of infection.
Dr. Lambert's letter is puzzling. Science is the wedding of logic and observation. Dr. Lambert's readiness to jettison logic is perverse. He is encouraged that statistics show “… that the odds that I will have an infection are exceedingly low.” However, the data reflect incidence expected when the anesthesiologist washes his or her hands. Lambert admits that he has “… not routinely done so …” but he wants credit for sterile gloves and cap. No mention of a mask. May we conclude that he does not “routinely” wear one? He states, “If the neuraxial anesthesia infection rates were the same as for [central venous catheter] insertions, no one would argue against the use of neuraxial full-barrier precautions.” (Isn't it logical to surmise that if all anesthesiologists were as cavalier about sterile technique, the infection incidence might approach that of inserting a central venous catheter?) He then implies that because the incidence rate of central venous catheter infection is much higher than that for dural puncture, that makes applying the central venous catheter precautions “illogical” because a 1:10,000 infection rate is a “nonproblem.” (It is a serious problem for that unlucky 1 in 10,000.) He asks, “… what risk is acceptable?” The answer is that (in the case of prevention of PDPM) the acceptable risk is the lowest possible under conditions of meticulous attention to sterile technique. The idea that there must be a cost–benefit analysis for every intervention, no matter how trivial, is impractical. Why does Dr. Lambert wear a cap and gloves while invading a patient's nervous system? Where are the data, and where is the logic?
Drs. Moen, Dahlgren, and Irestedt point out an error in the article text,1stating that in their series, 28 cases (instead of 11 cases) of PDPM were culture positive for α-hemolytic streptococci. (Table 1 and Table 4 have the correct value.) They ask, “Why do physicians have such an aversion to facemasks?” (Perhaps the situation is analogous to the reaction to automobile seat belts and bike helmets when they were first introduced.) They opine that the unnecessary deaths could reflect, among other things, “… lower diagnostic preparedness.” Not only does this serious complication often go unrecognized, but in several of the case reports, the existence of PDPM is denied and attributed to other causes.
The absence of inclusion of PDPM in the US Practice Guidelines7for the management of bacterial meningitis is inexplicable, as is the Centers for Disease Control and Prevention's exclusion of PDPM as a nosocomial disease. Perhaps forums such as this one will reduce the incidence, as well as heighten awareness, of this preventable disease.
Estelle Traurig Baer M.D. retired. Previous affiliation: Department of Medicine, Kaiser/Permanente Medical Center, Richmond, California. firstname.lastname@example.org