Indian Pediatrics 2001; 38: 1281-1286
An Indigenous Leucocyte Esterase Test Along with Pandy’s Test for the Diagnosis of Bacterial Meningitis
Polymorphs contain several esterases that are not present in serum. The leukocyte esterase (LE) act on esters breaking them down to corresponding alcohols. The alcohol thus produced may be detected by color change brought on an indicator. The LE activity in the urine is used to detect pyuria(1). Many such processes have been patented in dipstick test for pyuria. The esterase test is not dependent on the integrity of the leukocytes and would be present even after cell lysis(2). In this respect it scores over routine microscopy. The dipstick used for detection of pyuria has been used previously to diagnose pyogenic meningitis(3-7). The advantage of such a test in developing countries, which obviates the need for 24-hour laboratory services for CSF examination, is obvious. With the help of the test, antibiotic treatment can be initiated in bacterial meningitis, with benefit to the patient. Combur 9 (Boehringer Mannheim), a 9 patch test strip that is used to test urine for protein, glucose, leukocyte, nitrites, pH, blood, ketone, bilirubin and urobilinogen has been used(3) to measure CSF protein, glucose and leukocytes. Molyneux and Walsh(4) used another urine reagent strip (Multistix 8SG Bayer Diagnostics Ames) which has LE test but whose sensitivity range for protein and sugar does not lend itself to test the CSF range of these compounds. They therefore depended entirely on the detection of LE to diagnose meningitis. We have previously published our study of the same imported urine reagent stix (Multistix 8SG) to diagnose bacterial meningitis(5). We had concluded that if the test were used in combination with a test of protein more than 40 mg/dl, it would enhance the pickup rate.
In a project funded by the Department of Science and Technology New Delhi, to develop appropriate technology for India, one of the authors (PU) had developed a test for leukocyte esterase, which involves the use of Naphthylacetate as substrate. The resulting naphthol is detected by a color change of azo salt (turn red in presence of alcohol). This study evaluated the indigenously prepared leukocyte esterase test used in combinations with Pandy’s test (saturated solution of phenol to detect protein >40 mg/dl)(8) in the diagnosis of bacterial meningitis.
Subjects and Methods
Eighty seven consecutive CSF samples obtained from children (aged 1 month - 12 years) with suspected meningitis were analyzed in the study. Each sample was divided in two. One was sent to laboratory for routine diagnostic evaluation and the other was tested with the indigenously developed reagent strip and Pandy’s test to detect leukocyte esterase activity and protein. The study protocol was approved by the institutional research protocol review board.
Preparation of the LE Strips
Pads of size 2 mm × 2 mm were cut from thick (0.5 mm) water absorbing paper. The pads were washed for 1 hour with distilled water and dried at 110º C for 2 hours. They were cooled in dry air and dipped in a solution of alpha napthyl ester (80 mg in 10 ml alcohol). After 10 minutes the excess solution of alpha napthyl ester was decanted and pads were dried at room temp and dry air and stored in a sealed dry container in a refrigerator.
Procedure for Use
The procedure comprised of the following steps:
One drop of CSF was dropped on to 1 ml of Pandy’s reagent in a test tube and turbidity developed was noted, by holding the test tube against a black background. The mildest milkiness was noted as positive(8).
One researcher (RS) did the all the LE and Pandy’s tests and results were noted down before the laboratory report was known. The diagnosis of meningitis was made on the basis of laboratory results using the criteria outlined earlier(9):
(i) Bacterial meningitis – Glucose <50 mg/dl, Protein >40 mg./dl and Polymorph >1 mm3; and (ii) Viral meningitis – Glucose >50 mg/dl or 75% of the blood glucose, protein 40 mg - 200 mg/dl and cells >5 mm3 predominantly lymphocytes.
Table I shows the Pandy’s test, LE test and the laboratory results in all 31 cases where any of the above tests was abnormal. The remaining CSF samples were completely normal. The lowest count of polymorphs in our study of abnormal CSF was 4/mm3. All cases with polymorphs were detected also by LE test. Two cases of viral meningitis (one with 33 lymphocytes with no polymorphs and other with 6 lymphocytes and no polymorphs) were leukocyte esterase negative. Two traumatic CSF with polymorphs were LE positive. One CSF without polymorphs on microscopic examination was reported as LE positive. Twenty four CSF had protein greater than 40 mg/dl and all cases were Pandy’s positive. Two false positive Pandy’s test were noted where laboratory value of protein was less than 40 mg/dl (The actual values of protein were 35 mg/dl and 38 mg/dl). One sample with 40 mg/dl protein was Pandy’s test negative.
Table II shows LE test results against laboratory diagnosis of bacterial meningitis. LE test was positive in all 20 cases. In 2 traumatic CSF samples and 1 normal CSF it was falsely positive. Table III shows Pandy’s test against laboratory diagnosis of bacterial meningitis. It was positive in 19 of the 20 cases and false positive in 6 others. Table IV shows the combination of the two tests against the diagnosis of bacterial meningitis. There was no case, which had both tests positive but did not have bacteiral meningitis.Table I__Abnormal CSF Results
Moosa et al.(3) using Combur 9 reagent strip (urine stix) in CSF for detection of meningitis found sensitivity of 97% and specificity of 100%. Molyneux and Walsh used Bayers Multistix 8SG for detection of bacterial meningitis and found a sensitivity of 33% and a specificity of 83%(4). An earlier study from this hospital using the same Bayer Multistix 8SG for detection of meningitis found a higher sensitivity of 85.26% and specificity of 77.27%(5). In the present study, using an indigenously prepared leukocyte esterase test, sensitivity for bacterial meningitis was 100% and specificity is as 95.52%. Thus the indigenous LE seems to be far more sensitive and specific than the commercially available Multistix 8SG. The commercial test kits are made to detect pathological urine (upto 5 cells are normal in urine). Any polymorph in CSF is pathological and it is therefore appropriate that the CSF test be more sensitive. Pandy’s test was positive in all CSF with protein more than 40 mg/dl. While reasonably good correlation of Pandy’s and laboratory protein levels was found, CSF protein is increased in a large number of conditions other than bacterial meningitis and this test cannot be used by itself, for diagnosis of pyogenic meningitis. The addition of Pandy’s test to the LE test improved the specificity of diagnosis to 100%. Thus presence of both tests together confirmed bacterial meningitis in our study. The LE test was positive in this study even when the polymorphonuclear count was as low as 4/cu mm. This is better than the Combur 9, which is sensitive when the cell count is more than 10/cu mm. The presence of even one polymorph denotes meningitis and the ability to pick up smaller numbers of cells makes LE a better test.
Sensitivity – 100; Specificity – 95.2%; Predictive value of Positive test – 86.96%; Predictictive value of Negative test – 100%; Kappa – 0.91%; SE – 0.1068; P – <0.001.
Table III__ Comparison between Pande’s test and Laboratory Diagnosis of Bacterial Meningitis
Sensitivity – 95%; Specificity – 89.55%; Predictive value of Positive test – 73.08%; Predictictive value of Negative test – 98.36%; Kappa – 0.77%; SE – 0.1055; P – <0.001.
Sensitivity – 95%; Specificity – 100%; Predictive value of Positive test – 100%; Predictictive value of Negative test – 98.53%; Kappa – 0.97%; SE – 0.1072; P – <0.001.
The ability of LE test to show up positive even after cell lysis suggests that this test may be used to detect meningitis even if there has been considerable delay in testing the CSF.
The LE test can discriminate between different types of white cells. The leukocyte esterase is present in polymorphonuclear cells but not in lymphocytes. Thus, in the 2 samples of viral meningitis with 33 lymphocytes and 100 lymphocytes, respectively the LE test was negative. However, polymorphs are some-times present in early viral meningitis and the LE test will pick these up as pyogenic meningitis.
The LE test was positive in both the partially traumatic CSF. This must have to do with the polymorphs present in the con-taminating blood. Of the 3 false positive tests 2 were positive because of blood contamina-tion. One had a RBC count of 200 cells/cu mm and the other 280 cells/cu mm suggesting that the blood contamination was minimal. The resultant rise in CSF protein was not significant and the protein was not significant and the protein was 17 mg/dl in both cases. The protein level was far below the threshold for Pandy’s test and this test helped to discriminate the samples as probably normal.
If LE test alone was used to diagnose pyogenic meningitis, all those needing antibiotics would have been treated. Three extra children would have been administered antibiotics unnecessarily. This study has clubbed pyogenic meningitis and tubercular meningitis under the umbrella of bacterial meningitis. Distinction cannot be made between these types of meningitis using the LE test, just as the distinction is sometimes difficult with available techniques.
A limitation of this study is that we did not use the gold standard of CSF culture and Antigen detection for the diagnosis of meningitis. Any misclassification based on CSF criteria will be reflected in this study.
The test seems to have potential in the diagnosis of bacterial meningitis. However, larger studies are required to confirm these findings. A problem we encountered while conducting this evaluation was the short shelf life of the reagent strip of 3 months, when kept in a refrigerator. Facilities for refrigeration are not widely available in the rural setting.
A limited suply of the LE strips used in the study can be made available for non-commercial clinical trials from the Center for Science Education and Communication (PU).
Contributors: RKS, SG, MB and NK were involved in data collection and drafting the manuscript. PU developed the indigenous test and conceived the project. JMP conceived the project, co-ordinated and supervised the study and helped in final drafting; he will act as the guarantor for the paper.
Competing interests: None stated.