Dear Sir,
We read with interest the manuscript by Bustos et al.,1 which provided data on the validity of their own diagnostic criteria for ventricular neurocysticercosis (NCC).2 We agree that improving the diagnosis and management of people with NCC is a priority, and any tools that would help to improve this are always welcomed.
In their discussion, the authors state that when our previously published validated diagnostic criteria3 were applied to the same cases, they found poor specificity (39%). We were surprised by this, as this step was not mentioned in the Methods section—nor were any results provided. We can, therefore, surmise that this seems to have been done ad hoc. We were also bewildered by the specificity value provided and attempted to reproduce specificity for both criteria (as data were provided).
To estimate specificity, the cases used by Bustos et al.1 were independently reviewed by two of our team (A. F., D. S.-J.), and they obtained different results than those reported (Table 1). Of the 40 cases that we retrieved, the specificity for each of the evaluations of our criteria was 82.5% and 78.1% (confirmed by an external statistician). We were also surprised that we could not reproduce the specificity reported by the authors for their own criteria as we again found differing values (see Table 1). We did not attempt to estimate sensitivity as it seems only reasonable to assume that they are also not reproducible.
Specificities of Carpio et al. and Del Brutto et al. criteria independently evaluated
| Specificity (95% CI) | ||
|---|---|---|
| Evaluator 1 | Evaluator 2 | |
| Carpio et al.3 criteria | 78.1% (62.4–89.4%) | 82.5% (67.2–92.7%) |
| Del Brutto et al.2 criteria | 31.7% (18.1–48.1%) | 55.0% (38.5–70.7%) |
CI = confidence interval.
Our finding generated concerns as to why the results of Bustos et al.1 are not reproducible. One possibility is that their review of cases and controls was not blind as this is not mentioned in their methods. Of course, the authors may well have an alternative explanation for this lack of reproducibility, and we look forward to receiving such an explanation if this is the case.
Acknowledgment:
We would like to thank Dr. Willem M. Otte (University Medical Center Utrecht) for his expert advice on the statistical analysis.
REFERENCES
- 1.↑
Bustos JA, García HH, Del Brutto OH, 2017. Reliability of diagnostic criteria for neurocysticercosis for patients with ventricular cystic lesions or granulomas: a systematic review. Am J Trop Med Hyg 97: 653–657.
- 2.↑
Del Brutto OH, Nash TE, White AC Jr, Rajshekhar V, Wilkins PP, Singh G, Vasquez CM, Salgado P, Gilman RH, Garcia HH, 2017. Revised diagnostic criteria for neurocysticercosis. J Neurol Sci 372: 202–210.
- 3.↑
Carpio A et al. 2016. New diagnostic criteria for neurocysticercosis: reliability and validity. Ann Neurol 80: 434–442.