using all the evidence

Anesthesiologists and intensivists face a variety of different patient problems. Often the best treatment approach to a given problem is unclear. Evidence-based medicine has been advo­cated as the most logical approach for clinicians to use to apply research to clinical practice (1, 2). Unfortunately, the database of suitable clinical studies upon which the clinician bases important therapeutic decisions is often incomplete. It is in this setting that the anesthesiologist and col­leagues must nonetheless endeavor to make logical decisions while avoiding an uncertainty-induced pa­ralysis of action—a paralysis that may ensure that an adverse situation will result in an adverse outcome. Another source of information that can be logically applied to some clinical situa­tions is animal studies. Unfortunately, conclusions from animal studies may not always translate to hu­mans, and clinical studies may be too specific or too generalized to be relevant to some cases. There are advantages and disadvantages of basing clinical decisions on animal research and of applying the results of clinical studies to specific patient populations. The full text has been previously published (

Kofke WA: Making clinical decisions based on animal research data: Pro. J Neurosurg Anesth 8:68-72, 1996).

In a hypothetical multi-institutional clinical study, the ability of drug A to protect the brain from focal cerebral ischemia was tested. Preclinical animal studies had been carefully designed and performed, and all showed a protective effect. Thousands of patients undergoing procedures with anticipated in­traoperative ischemia were entered into the study. Unfortunately no overall group differences in mortality or other outcome variables were ob­served. Subsets of patients in whom the drug may have been effective were lost in the noise. As a result, the drug was abandoned and patients contin­ued to have intraoperative strokes.

 In another hypothetical, multiinstitutional study, the brain-protective effects of drug B were assessed in patients undergoing carotid endarterectomy. This time a statistically significant effect was noted, and the drug achieved widespread use. The drug com­pany’s investment paid off handsomely. In this case, due to the narrow clinical situation studied, clinical judgment is required in deciding whether the data are applicable to a given patient who may not be undergoing carotid endarterectomy.   A therapeutic decision to use a drag released for clinical use based on a narrow group of patients (e.g., nimodipine for subarachnoid hemorrhage) is highly  judgmental, although possibly justified.

The scenario, wherein a patient’s problems are not exactly modeled by any clinical or laboratory study, is quite common. In situations with this degree of uncertainty, one can reasonably argue that the well-read clinician would be remiss in not considering all available data including noisy clinical studies and focused clinical and animal studies, despite the fact that a patient may not fit precisely into the entry criteria of any of these studies.
 

There are several compelling reasons supporting the use of data from animal models in clinical deci­sion-making. Compared with clinical studies, important variables that influence   outcome can be better controlled and maintained for several days to allow   lesions to mature in a homo­genous group of animals under strictly controlled conditions (15,16). With such studies, the power to determine the potential for efficacy is increased due to fewer biologic and therapeutic variables; thus, determinations of whether it is possible for a given therapy to confer brain protection are more likely.

The principles for interpreting clinical trials  may also be relevant in this context. For example, if a patient is in a similar situation as that modeled in an animal study and the potential morbidity of the therapy is acceptable, it’s logical to apply the therapy to the patient. An im­portant aspect of using such animal data in humans is the implicit expectation that potential morbidity of therapy is low (i.e., “can’t hurt, might help”). In some situations, the morbidity of the therapy is low enough that it will do no harm or the morbidity of the disease is so severe that using such therapy can not make things worse. Some examples of these therapies relevant to neuroanesthesia include anti­hypertensive therapy, mild hypothermia, and barbiturates.

 

Ideally, clinical decisions should be evidence-based from clinical trials. However, it is not reasonable to expect that all clinical questions will be answerable in clinical studies, and animal studies  may be the best option available.  In the absence of good data or appropriate clinical studies, the clinician is justified in using his or her best judgment to synthesize clinical and animal studies to make decisions relevant to a particular clinical situation.

 

Comments

[mkeamy]

I've read this post about six times. It keeps getting better; why is that? send me a pdf of the original article, if you can AK.

Mitch