Letters to the Editor

Indian Pediatrics 2002; 39:603-604

Reply

We appreciate the interest in our paper(1) and are happy to clarify the queries. In the Lancet, Moosa et al.(2) suggested that 3 components of a 9 strip urine dipstick test (Combur 9 Urine stix available in the West) could be used to test CSF and to initiate antibiotics for bacterial meningitis in circumstances where 24 hour CSF cytology and biochemistry are not available. This study was replicated by others using different reagent strips, with lesser levels of success. Our own pilot study with an imported dip stick (Bayers Multistix 8SG) was published previously and this suggested the need to combine leukocyte esterase (LE) with an additional test for protein of more than 40mg/dl(3). Prompted by the findings of the pilot study, the present study uses Pandy’s test along with an indegenous LE test, to pick up cases of bacterial meningitis (BM).

Experienced clinicians know that even when CSF cytology and biochemistry are available, the diagnosis of BM cannot be made with cent percent certainty in all cases. Dr. Locham e t al. say that CSF can be normal in early meningitis (paragraph 1 of their letter). The Table used by them suggests that CSF of BM have more than 75% polymorphs while in viral meningitis there is less than 25% polymorphs. Where would one classify CSF with 50% polymorphs? It is unrealistic to expect bed-side/side-laboratory tests to resolve these issues. Safe practice mandates that such cases be classified as BM and treated with antibiotics till proven otherwise. With such safety considerations in mind, the criteria we have used to classify as BM are listed in our paper (last paragraph of Subject and Methods). Presence of even 2 of the 3 factors suggestive of BM was sufficient to classify the case as bacterial meningitis. It can be seen that in adapting the table mentioned by Dr. Locham, where there was an overlap of findings between viral meningitis and BM, we have put it categorically under BM. That an extremely rare case of fungal meningitis (Point 2) will be treated unnecessarily with antibiotics does not worry us excessively.

A study of the half-life of LE in CSF was not one of the objectives of this investigation. We have quoted a reference that suggests that LE activity persist even after cell lysis. We have ourselves retested CSF kept at room temperature for 24 hours and found it stayed positive for LE activity. More detailed investigation into this was not part of our study protocol. Locham et al suggest that CSF of babies between 1 and 2 months may be Pandy’s false positive (Point 4). We disagree. Babies less than 3 months old, may normally have CSF protein more than 40 mg/dl. Under these circumstances, if Pandy’s test is positive it could be ‘true positive’ and not ‘false positive’.

Regarding point 5, we have reported in the text (Paragraph 1 of Results), "two cases of viral meningitis (one with 33 lymphocytes and no polymorphs and other with 6 lymphocytes and no polymorphs) were leukocyte esterase negative". In this context the mention of 100 lymphocytes is an error. Locham and colleagues also note that CSF of cases with traumatic CSF showed 0% polymorphs on microscopy but their LE test was positive. Logically traumatic CSF will have polymorphs and these were detected by the LE test. Lysis of the cells can explain their absence on microscopy.

Points 6 and 7 have been addressed in the second paragraph of this response.

Dr. Locham et al. suggests that Cases 9, 29 and 30, with protein more than 40 mg/dl should be diagnosed as meningitis (Point 8), although they had normal cytology and normal CSF sugar. We cannot agree with them on this.

They also suggest that Cases 3, 4, 14, 16, 18-21 and 25 be classified as viral meningitis because they had less than 75% polymorphs cells. They will note that all these cases had more than 25% polymorphs and so they do not fit into their criteria for viral meningitis either. As explained in our introductory paragraph, safe practice prompts them to be classified and treated as bacterial meningitis.

We thank them for pointing out the typographical error in the JAMA reference (Point 9). In reference 2 of the original paper the page numbers should read 1221-1224 and not 1121-1124.

In summary, 56 CSF samples which were normal on cytology and biochemistry were also LE test negative and Pandy’s test negative. The indigenous LE test and Pandy’s test were both positive in all cases of bacterial meningitis. In the sample of cases we have studied, this combination of bed-side tests was adequate to indicate the initiation of treatment of bacterial meningitis. More studies are required to confirm our observation.