Statistics from Altmetric.com
By contrast with the weaknesses of anecdotal observations from case series, the power of randomised clinical trials for deciding the benefit of therapy has become increasingly evident and indisputable world wide. Nowadays, to argue against the validity of this assertion may seem superfluous; however, a recent paper reported by Del Brutto1 regarding treatment in neurocysticercosis ignores basic procedures for well performed clinical trials by using inappropriate and misleading methodology to evaluate medical therapy.
By definition, a clinical trial is a prospective study comparing the effect and value of treatment against a control in human subjects. The main drawback of Del Brutto’s report is that it does not include a control group against which the intervention group is compared; therefore, its results are definitely flawed. Additionally, a basic experimental study design requires at least minimal information regarding inclusion and exclusion criteria, randomisation, and definitions of response or outcome variables. This information is not provided by Del Brutto’s report; its design fails to protect against potential bias in patient selection or evaluation of outcome. The definition of subarachnoid cysterci used by Del Brutto was based on “appearance on CT of hypodense cystic lesions located over the convexity of the cerebral hemispheres, the sylvian fissure, or the CSF cisterns at the base of the brain”. It is well known that there are many other diagnostic possibilities to be considered in the differential diagnosis of subarachnoid hypodense lesions.2 3 Besides, CT is not a reliable procedure for diagnosing subarachnoid cysterci, as is MRI. In fact, we cannot be completely sure, for example, that the CT images shown in the report of Del Brutto correspond to subarachnoid cysterci. If we were to use MRI on this patient, they might correspond to a parenchymal cyst which resolved as a reflection of the natural history of the condition. There is no evidence that objectively confirms or rejects this assertion.
Del Brutto’s report1 maintains that evaluation of the therapeutic response to albendazole included comparison of the size of the cysts, as well as clinical evaluation of patients before and after treatment. To consider the size of cysts as a response variable is certainly useless because of the obvious difficulties in measuring cyst size in the subsequent follow up CT. It is also widely accepted that the clinical manifestations of neurocysticercosis are polymorphic, and their clinical course is unpredictable; therefore, the clinical manifestations as an outcome variable is entirely biased. Another personal appreciation of Del Brutto1 is that albendazole reaches high concentrations in CSF, and has been used with success in some patients with subarachnoid cysts; nevertheless, studies used as support1 of this presumption are similarly flawed in that they are not randomised or blinded, having historical control groups or patients who served as their own control, and regarding clinical evaluation as an outcome variable.
Whereas it is generally assumed that albendazole is effective treatment for neurocysticercosis, a critical review of the literature3 4 suggests that the studies on which these assumptions are based are defective in terms of patient selection, assignment to treatment, and selection and measurement of outcome variables. Many authors have warned that this therapy in some patients might sometimes be harmful, particularly in the subarachnoideal localisation, because some patients have developed arteritis and hydrocephalus after the administration of antihelminthic drugs.3 According to these authors a parasite may be easily removed surgically at a cystic stage before an inflammatory reaction develops.3 A randomised clinical trial of treatment of neurocysticercosis4 considers the question of to what extent and in which patients treatment with either praziquantel or albendazole is effective. The improvement attributed to these drugs in several studies may be related to the lack of appropriate controls and is likely to be a reflection of the natural history of the condition. The authors point out the need to conduct a long term, placebo controlled trial with precise end points, proper randomisation, sample size calculations, and predetermined statistical calculations, to evaluate properly the effectiveness and determine the indications of aetiological treatment for neurocysticercosis.
In the era of evidence-based medicine, we neurologists and general practitioners should be demanding regarding use of sound scientific information with methodological rigour for improving our clinical decision making.5 Medical information from reports that do not conform to the minimal requirements of a clinical trial should be avoided.
The author’s reply:
I celebrate the rigid academic standards of Carpio’s medical practice, but wish they were matched with knowledge on the available literature on albendazole therapy for neurocysticercosis. That albendazole actually has a cysticidal effect is beyond all doubt.1-1-1-6 The drug has been used to treat patients with neurocysticercosis since 1987, and physicians who are familiar with the disease know that it is effective. Moreover, the single study in which albendazole has not been useful for adults with viable parenchymal brain cysticerci—published by Carpio et al 1-7—has been questioned due to inaccurate recollection of data.1-8 In our study, we did not attempt to verify the cysticidal effect of the drug (it has already been demonstrated) but to document if albendazole could also be useful in a severe form of neurocysticercosis that has been associated with a grim prognosis.1-9 Under these circumstances, it is not ethical to deprive a group of patients of a safe and inexpensive treatment just for the sake of science. In addition, Carpio’s concerns about the criteria we used for the diagnosis of subarachnoid cysticerci are typical of those who are not familiar with the disease. The problem with CT is that this imaging method may misdiagnose some subarachnoid cysts as parenchymal cysts, but the opposite is not true,1-10
As a physician interested in the advancement of science, I applaud Carpio’s interest in evidence-based medicine but I completely disagree with him in that information from reports other than clinical trials should be avoided. He must remember that outstanding contributions to medical knowledge have been made through single case reports, small clinical series, and open trials. On the contrary, thecliché “randomised” has been the shield of major medical frauds.1-11 Medicine is art and science, and wise physicians know that information from clinical findings actually have a “significant” impact on everyday clinical practice.