Failure of Standards in Reporting the Composition of Artificial Cerebral Spinal Fluid in Studies of the Pial Blood Vessels
To the Editor:
Reproducible experimental results require the replication of the experimental conditions under which the experiment was first performed. Relevance to humans may depend on the degree to which those conditions resembled those to which humans are exposed. With respect to studies of vascular reactivity, a most important parameter is the composition of the fluid bathing the vessel. For vessels on the brains surface, so-called pial vessels, this fluid is the cerebrospinal fluid (CSF). Experiments studying these vessel, in vivo or in vitro, generally bathe the vessels in an artificial CSF, sometimes called a “mock” CSF, or some other fluid, sometimes called a physiological salt solution and sometimes simply a buffered solution. A question that should immediately arise, if relevance to humans is an issue, is whether the fluid used has the composition of cerebrospinal fluid, and if so, is it the composition of the experimental species’ CSF or of human CSF. At present there is no data to tell us whether or not it would be better to always replicate the composition of human CSF, even if this differs from that of the animal in use, or whether, instead, the mock CSF should resemble that of the animal itself. CSF contains urea, but not everyone adds urea to the artificial CSF, and there are no studies showing the effect of the urea on the vascular responses. Indeed there is virtually no information on the effect of altering one or more aspects of CSF composition, other than pH and changing the concentration of one ion or another so that it is clearly out of mammalian physiological range. Moreover, recent studies show that under some circumstances, when a mock CSF flows over the pial surface in a continual suffusion, some vascular responses may be dramatically altered because essential substances, not thought of as part of CSF but in the tissue and its extracellular space, may be washed away. In the absence of the requisite studies, this letter will not address the question of “which formulation of mock CSF is better—human or animal?” Nor will it address the importance, or lack thereof, of the addition of urea or of the presence or absence of continuous rapid suffusion. It will, however, direct itself to the most basic requisite for a scientific presentation—the need to describe the fluid one is using so others may replicate your experimental conditions.
Unfortunately, with respect to studies concerning cerebrovascular reactivity in vivo, our leading journals and, presumably, others have demonstrated a dramatic failure to maintain the standards displayed for studies of other vascular beds. Why this is so cannot be stated. But that it is so is evident from the following analysis.
For studies of pial vessels, 3 journals were consulted over the past 4 years. These journals were Stroke, The Journal of Cerebral Blood Flow and Metabolism, and The American Journal of Physiology (AJP). As a “control,” papers concerning extracerebral vascular beds were surveyed over the same time period in AJP and in Microcirculation. A total of 34 in vivo studies of extracerebral vascular reactivity were located. Of these, only 2 (6%) failed to specify the composition of the bathing fluid. In striking contrast, of the 49 studies of pial vessels, 21 (43%) failed to specify the composition of the bathing solution. In only 3 of these 21 articles was there even a reference to an earlier paper that described the bathing solution. In some of the remaining 18 cases a reference to the preparation as a whole was given, but perusal of that reference failed to give the composition of the bathing fluid.
In the 18 papers without even a helpful reference (37% of the total of 49 studies), an effort was made to determine what the composition of the fluid might be, by locating past papers of the authors. Sometimes, the first description of the fluid was found in a paper published many years earlier in conjunction with some other author. In the absence of a reference to the earlier paper, it seemed unwise to assume that the composition of the fluid used in the recent year was identical to that described in the much earlier paper.
The absence of a description of the composition of the fluid was even more glaring than suggested by the simple fact that in 18 of 49 papers there was neither description nor a useful reference to the composition of the fluid. That is because of the 31 papers which did give a description, 8 came from a single laboratory. In other words some authors publish a lot and consistently provided essential experimental details and thus bias the statistics toward the side of compliance with a necessary standard.
Finally, when the composition of the irrigating solution is provided, it is necessary to consider the units in which it is presented. The preferred units would be in molar or millimolar terms or in terms of milliequivalents. If, instead, the grams or milligrams per liter of each substance are given, it falls to the reader to make the required calculation.
It should be clear from this analysis that journal editors, reviewers, and authors have been delinquent in meeting a reasonable standard for the publication of in vivo studies of pial blood vessels, a standard that is almost always met when other vessels and/or in vitro studies are the subject of the reports, and a standard that is frequently violated for pial vessels even when it is met for other vascular beds in the very same journal. It is hoped that these remarks will lead to correction of that situation.
- Copyright © 2000 by American Heart Association