251

Distinguishing Cerebrospinal Fluid Abnormalities in Children with Bacterial Meningitis and Traumatic Lumbar Puncture William A. Bonadio, Douglas S. Smith, Susan Goddard, Jean Burroughs, and Gaziuddin Khaja

From the Department of Pediatrics, Medical College of Wisconsin and Children's Hospital of Wisconsin, Milwaukee

When lumbar puncture is traumatic, contamination with peripheral blood elements may obscure the accurate analysis of cerebrospinal fluid (CSF) parameters, especially in determining the presence of pre-tap pleocytosis. Prior studies examining the effect of traumatic LP have compared the ratios of white blood cells (WBC) to red blood cells (RBC) in CSF and peripheral blood to determine whether the pre-tap CSF WBC count was normal or abnormally elevated [1-3]. For obscure reasons, this method has proven unreliable in most cases of culture-negative CSF. No study has examined the effect of traumatic lumbar puncture on other traditionally analyzed CSF parameters. Here we report CSF characteristics of children with. and without bacterial meningitis who experienced traumatic lumbar punctures.

Methods To obtain adequate numbers for analysis, two separate retrospective reviews were performed of patients at the Children's Hospital of Wisconsin, Milwaukee. Group 1 was derived from consecutive previously healthy children >1 month of age with CSF culture positive for bacterial pathogens and CSF RBC count >1000/mm3 during 1980-1989. Group 2 was similar to group 1 except their CSF

Received 11 October 1989; revised 28 December 1989. Reprints and correspondence: Dr. William A. Bonadio, 1240 Pioneer Trail, Waukesha, WI 53186. The Journal of Infectious Diseases 1990;162:251-254 © 1990 by The University of Chicago. All rights reserved. 0022-1899/90/6201-0040$01.00

was culture-negative for bacterial pathogens and they were hospitalized during January-June 1988. Infants I month of age; 30 had bacterial meningitis and 62 had negative CSF cultures. The purpose was to distinguish CSF profiles of the two groups despite contamination with peripheral blood elements. In each case, white blood cell (WBC) counts were observed (0) and compared with those predicted (P), calculated as P = CSF RBC x (blood WBClblood RBC). Comparison ofO:P ratios revealed that all 30 patients with bacterial meningitis had ratios ~1, 28 (93%) had ratios >10, and 24 (80%) had ratios >100; by contrast, only 2 patients (3%) with culture-negative CSF had ratios >10,21 (34%) had ratios of I-tO, and 39 (63%) had ratios 100,000/mm3 • The pathogens isolated were Haemophilus infiuenzae type b (22), Streptococcus pneumoniae (4), and Neisseria meningitidis (4). CSF analysis revealed that all 30 patients had pleocytosis, 29 had a differential cell count predominance (>50 %) of polymorphonuclear leukocytes (PMNL), 22 had hypoglycorrhachia, 30 had an abnormally elevated protein concentration, and 24 had a positive Gram's-stained smear for pathogenic organisms. The range of calculated O:P ratios was 1-55,000. Both patients with an O:P ratio of 1-10 had a CSF differential cell count with :?>90% PMNL; one had hypoglycorrhachia and the other had a positive Gram's-stained smear. During January-June 1988, 522 patients received lumbar puncture; of these, 101 were traumatic. Eliminated from study were 32 neonates, 5 patients with a ventriculoperitoneal shunt, and 2 with computed tomography scan evidence of central nervous system hemorrhage; 62 patients met the criteria for study and formed group 2. The distribution of CSF RBC counts was: 36, 100,000/mm3 • Analysis of CSF revealed 48 patients had pleocytosis, 7 had a differential cell count predominance of PMNL, 2 had hypoglycorrhachia, 57 had an abnormally elevated protein concentration, and none had positive Gram's-stained smear. The predicted CSF WBC count exceeded that actually observed in 39 patients. The range of O:P ratios was 0-25. Both patients with an O:P ratio >10 had a differential cell count predominance of mononuclear cells, a normal CSF glucose concentration, and negative Gram's-stained smear for pathogenic organisms. All 21 patients with an O:P ratio of 1-10 had a normal CSF glucose concentration and negative Gram's-stained smear for patho-

genic organisms, and 17 had CSF differential cell count predominance of mononuclear cells. The distribution of 0:P ratios for groups 1 and 2 is depicted in figure 1. There was a significant difference in the rate of CSF differential cell count predominance of PMNL, hypoglycorrhachia, positive Grams-stained smear, O:P ratios ~1, and O:P ratio >10 between groups (P < .0001). The results of sensitivity, specificity, and predictive values of various CSF parameters for bacterial meningitis are given in table 1.

Discussion Traumatic lumbar puncture is common in pediatric patients, as evidenced by the 20 %'rate of this complication documented in our series. Previous studies [1-3, 5] spanning the past 50 years have sought a method for determining the effect of trauTable 1. Predictive value of abnormal cerebrospinal fluid parameters for bacterial meningitis. Predictive value

O:P ratio ~1 O:P ratio >10 Pleocytosis PMNL predominance Hypoglycorrhachia Abnormally elevated protein Organisms on gram'sstained smear NOTE.

Sensitivity

Specificity

+

100 93 100 97 73

62 95 23 89 97

57 93 38 81 92

100 97 100 98 88

100

8

34

100

80

100

100

91

Data are percentages. PMNL = polymorphonuclear leukocyte.

Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on November 20, 2014

5

JIO 1990;162 (July)

Concise Communications

all patients with bacterial meningitis had a CSF differential cell count predominance of PMNL and either hypoglycorrhachia or positive Grams-stained smear. There was a significant difference in rates of hypoglycorrhachia between those with (73 %) and without (3 %) bacterial meningitis and traumatic lumbar puncture. The 73 % rate is identical to that reported previously in children with bacterial meningitis and nontraumatic lumbar puncture [7], suggesting that contaminationwith peripheral blood elements does not affect the CSF glucose concentration. Prior studies that performed in vitro artificial contamination of CSF of known composition with varying amounts of blood found no detectable change in CSF glucose concentration [8, 9]. Hypoglycorrhachia associated with traumatic lumbar puncture is an important finding that usually represents a pre-tap CSF abnormality and may indicate bacterial meningitis. By contrast, an abnormally elevated CSF protein concentration was consistently present in both those with (100%) and without (92 %) bacterial meningitis associated with traumatic lumbar puncture due to the presence of RBC. This finding, in conjunction with traumatic lumbar puncture, is of little diagnostic aid in distinguishing underlying cause. As expected, the CSF Gram's-stained smear was specific in distinguishing bacterial meningitis from culture-negative CSF when lumbar puncture was traumatic. A positive gramstained smear was present in 80 % with bacterial meningitis, a rate identical to that previously documented in children with bacterial meningitis and nontraumatic lumbar puncture [7]. A negative Gram's-stained smear was universally present in those with culture-negative CSF. When lumbar puncture is traumatic, all elements of the CSF profile should be examined for possible aberrations. In children >1 month of age, cerebrospinal fluid abnormalities associated with bacterial meningitis are rarely obscured by blood contamination from traumatic IUJTIbar puncture. Because bacterial meningitis may rarely be associated with normal CSF parameters at the time of initial evaluation, physicians should scrutinize the entire clinical picture when planning management strategies for children who warrant lumbar puncture.

References 1. Novak RW. Lack of validity of standard corrections for white blood cell counts of blood-contaminated cerebrospinal fluid in infants. Am J Clin Pathol 1984;82:95-97 2. Rubenstein JS, YogevR. What represents pleocytosis in blood-contaminated ("traumatic tap") cerebrospinal fluid in children? J Pediatr 1985;107:249-251 3. Osborne JP, Pizer B. Effect on the white cell count of contaminating cerebrospinal fluid with blood. Arch Dis Child 1981;56:400-401 4. Krieg AF. Cerebrospinal fluid and other body fluids. In: Henry JB, ed. Clinical diagnosis and management by laboratory methods. Vol 1. Philadelphia: W. B. Saunders, 1979:635-637 5. Merritt HH, Fremont-Smith F. Technique of lumbar puncture and cistern puncture. In: Merritt HH, Fremont-Smith F, eds. The cerebrospinal fluid. Philadelphia: W. B. Saunders, 1937:89-93

Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on November 20, 2014

matic lumbar puncture-induced blood contamination of CSF analysis. Most compared the WBC: RBC ratios of peripheral blood and CSF, arguing that when the pre-tap CSF WBC count is normal these values should essentially be identical; on the other hand, when CSF pleocytosis precedes traumatic lumbar puncture, the observed CSF ratio should exceed that predicted by the complete blood count profile. Although intuitively correct, this correlation has been shown to be inaccurate - in most studies examining culture-negative CSF associated with traumatic lumbar puncture, the predicted CSF WBC count consistently overcorrected that actually observed. Our results were consistent with this, as in 63 % of culturenegative cases the predicted CSF WBC count exceeded that observed. A prior study [6] of mostly adult patients with traumatic lumbar puncture (arbitrarily defined as only >200 RBC/ mm') compared the CSF WBC O:P ratios of those with and without bacterial meningitis to determine if this index had utility in distinguishing these two conditions. They noted that t"V90 % of cases of bacterial meningitis were characterized by an O:P ratio >10, whereas 10% with culture-negative CSF had a ratio >10. Yet there was considerable overlap of ratio ranges noted between the two groups, as 7 % with bacterial meningitis had a ratio 1. Similarly, Rubenstein and Yogev [2] noted several children (their ages were not given) with traumatic lumbar puncture and bacterial meningitis whose observed CSF WBC count was less than that predicted. The correlative frequency of other CSF abnormalities with culture-positive and culture-negative CSF was not delineated in either study. We found a distinct difference between O:P ratios characterizing those with and without bacterial meningitis. In contrast to most culture-negative patients, no patient with bacterial meningitis had a predicted CSF WBC count that overcorrected that observed. As a reflection of underlying CSF pleocytosis, all but one patient with bacterial meningitis had an O:P ratio well above unity, with more than 90% ofthem >10 and 80% >100. By contrast: only 3% of culture-negative patients had an O:P ratio >10 (none were >25), and 63% were 10 was used, the sensitivity remained high (93 %) and the specificity increased to 95 %. We also noted that the more commonly described aberrations in CSF parameters, such as differential cell count predominance of PMNL, hypoglycorrhachia, and positive Gram's-stained smear, were present in most cases of bacterial meningitis associated with traumatic lumbar puncture despite blood contamination. These traditionally analyzed features were accurate in distinguishing the CSF profiles of those with and without bacterial meningitis in the critical "overlap"group with O:P ratios of 1-10; all culture-negative patients had a normal CSF glucose concentration and 80 % had a CSF differential tell count predominance oflymphocytes, whereas

253

254

Concise Communications

6. Mayefsky JH, Roghmann KJ. Determination ofleukocytosis in traumatic spinal tap specimens. Am J Med 1987;82:1175-1181 7. Bonadio WA. Acute bacterial meningitis. Cerebrospinal fluid differential count. Clin Pediatr (Phi1a) 1988;27:445-447

JID 1990;162 (July)

8. Mehl AL. Interpretation of traumatic lumbar puncture. A prospective experimental model. Clin Pediatr (Phila) 1986;25:523-526 9. Reske A, Haferkamp G, Hopf HC. Influence of artificial blood contamination on the analysis of cerebrospinal fluid. J NeuroI1981;226:187-193

A Human Monoclonal Antibody that Protects Mice against Pseudomonas-Induced Pneumonia From the Department of Immunology Research, Merck Sharp & Dohme Research Laboratories, Rahway, New Jersey

X-linked immunodeficient (xid) (CBA/N x DBA/2 0-) F1 male mice, when treated with cyclophosphamide, were much more susceptible to challenge with aerosolized Pseudomonas aeruginosa serotype 11 than were control F1 female littermates. Mortality of F1 males was decreased significantly after intravenous administration of human P. aeruginosa serotype 11o-specific monoclonal antibody. Antibody treatment reduced bacterial titers in the lungs as well as the severity of Pseudomonas-induced lung histopathology.

Pseudomonas aeruginosa is an important cause of nosocomial infections in immunocompromised patients [1], with pneumonia being a common complication that is associated with high mortality [2]. Antibiotic therapy may be an effective treatment, but alternatives such as immunization or passive antibody administration are being considered, especially in cases of antibiotic resistance and refractory infections. We have generated lymphoblastoid cell lines that secrete human monoclonal antibodies against O-specific carbohydrates on the lipopolysaccharide (LPS) of P. aeruginosa serotypes most commonly associated with nosocomial infections. These monoclonal antibodies protect mice against systemic bacterial challenge effectively when used prophylactically [3]. To circumvent the need for high challenge doses because of naturally occurring protective antibodies, we used X-linked immunodeficient (xid) mice (CBAlN mice or the male F1 progeny from a cross between CBA/N females and DBA/2 males) as an animal model to test the efficacy of Pseudomonas-specific monoclonal antibodies for protection against systemic bacterial challenge. These xid mice are defective in generating an antibody response to high-molecular-weight carbohydrates and other antigens [4]. We observed that neutropenic (i.e.,

Received 14 August 1989; revised 9 January 1990. Reprints and correspondence: Dr. Hans 1. Zweerink, Department of Immunology Research, Merck Sharp & Dohme Research Laboratories, P.o. Box 2000, Rahway, NJ 07065. * Present address: Department of Comparative Medicine, University of Maryland School of Medicine, Baltimore. The Journal of Infectious Diseases 1990;162:254-257 © 1990 by The University of Chicago. All rights reserved. 0022-1899/90/6201-0041$01.00

cyclophosphamide-treated) F1 males were significantly more sensitive to systemic challenge with P. aeruginosa than were control female littermates. This increased sensitivity was due to reduced levels of (presumably naturally occurring) LPSspecific antibodies in the males [5]. Intravenous administration of low doses of P. aeruginosa-specific monoclonal antibodies to neutropenic males increased resistance to levels similar to those of neutropenic females [3, 5]. These studies were extended to develop a mouse model of pseudomonas-induced pneumonia using neutropenic (CBA/N 9 X DBA/2 a)Fl males (referred to as Fl xid male mice) and control female littermates.

Materials and Methods Bacteria. Mouse challenge experiments were conducted using a clinical isolate of P. aeruginosa serotype 11 obtained from Dr. Daniel Shungu (Merck Sharp & Dohme Research Laboratories). The serotyping system of P. aeruginosa has been described [6]. Bacteria were grown in trypticase soy broth (Difco Laboratories, Detroit) to a concentration of rv109 cells/ml, harvested by centrifugation (6500 g for 20 min), washed once in cold phosphate-buffered saline (PBS; 0.15 MNaCI and 0.01 Mphosphate buffer, pH 7.2), and carefully suspended (to avoid clumping) in cold PBS at an optical density of 0.8 measured at 610 nm. This suspension was used immediately. Viable bacteria were quantitated by plating on trypticase soy agar. Monoclonal antibody. Properties of the human IgM monoclonal antibody RM5, secreted by lymphoblastoid cells and specific for the O-specific carbohydrate moiety of P. aeruginosa serotype 11, have been described [3]. The antibody preparation used in this study was secreted by a heterohybridoma cell line (RSIGll) maintained in serum-free medium (HB104; NEN Research Products,

Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on November 20, 2014

Hans J. Zweerink, Louis J. DetoUa,* Maureen C. Gammon, Cameron F. Hutchison, Jane M. Puckett, and Nolan H. Sigal

Distinguishing cerebrospinal fluid abnormalities in children with bacterial meningitis and traumatic lumbar puncture.

The characteristics of cerebrospinal fluid (CSF) associated with traumatic lumbar puncture, defined as CSF red blood cell (RBC) count greater than 100...
433KB Sizes 0 Downloads 0 Views