To the Editor:—
First, I would like to congratulate Van Zundert et al. for their efforts to elucidate one of the putative mechanisms associated with pulsed radiofrequency (PRF), which may help us to understand its analgesic effect in clinical settings. Unfortunately, the explicit and implicit critique in the editorial by Richebéet al. about PRF in general may leave readers not familiar with this technique with a false impression that this modality is all but a speculative, experimental treatment.
The use of PRF is not taken lightly. Last year, more than 350 pain specialists from all over the world met on April 24–25, 2004, in Amsterdam for the First European Scientific meeting of the International Spinal Injection Society. During this 2-day meeting, we launched numerous multicenter clinical and basic science research protocols and created the European Collaborative Group for PRF research, while exchanging among us vast accumulated clinical experience. It is therefore that I read with some surprise the unsubstantiated remark that “there has been a mass migration to the use of pulsed radiofrequency with few data to support efficacy of this new technique.” I wish to clarify this statement.
In a simple, straightforward, systematic search in MEDLINE®, EMBASE, and Cochrane on PRF, one can generate 269 relevant reports in many fields, including pain medicine. Even by excluding all reports on electrical field research not directly relevant to the nervous system (such as biology, biochemistry, and physics), 38 reports remain available for critical reading. Of these, 1 is a prospective, randomized controlled trial (RCT); 5 are prospective uncontrolled trials; 7 are case series and clinical audits; 18 are letters, comments, and editorials; 7 are neurobiologic reports; and more than 30 are abstracts from important scientific meetings, including my own presentation at the American Society of Anesthesiologists annual meeting on October 11–15, 2003 (A-1090). The accumulation of these data is impressive and shows unequivocally that PRF is a genuine neurobiologic and clinical phenomenon and is different when compared with continuous radiofrequency (also known as thermocoagulation). Although the clinical advantages of this modality are not yet clear, what is clear is that PRF is not merely a whim of “wishful thinking” for those who practice it. Furthermore, exciting data on the effect of electrical fields on neural substrates suggests that PRF may have positive effects on synaptic strength and long-term potentiation, and if indeed central sensitization and long-term potentiation share similar mechanisms, these findings are of great interest.
My second comment regards the implicit critique “Neurobiology in Need of Clinical Trials,” suggesting that clinical trials are indispensable to determine the utility of PRF and that neurobiology is only of intellectual and theoretical interest. Implying lack of knowledge and thus lack of value to any treatment in the presence of marked variation in response is not a trivial epistemological matter. Causality in conditions of uncertainty require a careful approach to data analysis, and even perfect data synthesis may create “statistical alchemy” devoid of logical thought, resulting in “mega-silliness.” Asking the question “Does PRF work, or does PRF cause pain relief in patients?” is like asking what caused the fire in a house that was burned down as a result of a lit candle, a piece of paper, an open window, a strong wind, and the fact that the house was wooden. This situation, known in philosophy as INUS (insufficient nonredundant component of unnecessary sufficient complex), means in lay terms that by analyzing each component, none of them is a single sufficient cause, but their conjunction gives origin to an overall sufficient result. That is, assuming a hierarchy of causes that can be proved by a hierarchy of evidences, as evidence-based medicine (EBM) pretends, in conditions of uncertainty is at best misleading and at worst simply incorrect.
Current knowledge is fallible, EBM is limited by its probabilistic nature, and knowledge and evidence are not interchangeable terms. RCTs may sometimes be difficult or unattainable for methodologic or ethical reasons, and one must not forget that the decision to “downgrade” other forms of knowledge, as suggested by the editors in the spirit of EBM (such as clinical audits and basic science experiments), is a decision, not a truism. Systematic reviews may as well be victim to publication bias sullied by commercial influence, sponsorship pressures, and forces beyond control of reviewers. Reviews may suffer from poor delineation of references, simplistic concepts of pathophysiology, and the thought that stand-alone therapies (such as PRF) can be the cause of pain relief. Therefore, I argue that just as “basic scientific studies in the neurobiology of pain models and analgesic techniques are not a substitute for randomized controlled clinical trials,” RCTs are not a substitute for knowledge. EBM is only a methodologic solution to clinical epistemology, and it is blind to and disinterested in mechanisms of explanation and causation. EBM does not test causal hypotheses but merely establishes correlations or contributes to probability kinematics of various degrees of belief.
Finally, science is about discovering, recognizing, and changing paradigms, and that is the beauty of the PRF story that I think should be told. For years, heat has been thought to be the cause of pain relief in radiofrequency lesioning, until the ingenuity of Professor Sluijter and others suggested that perhaps it is only an epi-phoneme, and it is the electrical fields that are responsible for the analgesic effect. Saying (again) that we need to perform RCTs is too easy, and perhaps methodologic modesty is the order of the day.
Geneva University Hospital, Geneva, Switzerland. alex.cahana@hcuge.ch