Evidence-based medicine (EBM) may be defined as “the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients” (1). The Cochrane Collaboration is the largest organization in the world engaged in the production and maintenance of evidence-based reviews. The first issue of the Cochrane Database of Systemic Reviews was published in 1995 and included 36 Cochrane reviews. Currently, the database contains the full text of 2000 Cochrane reviews plus an additional 1400 published protocols for reviews in process. As demonstrated by the exponential growth of the Cochrane Collaboration, EBM is highly fashionable.
EBM has attracted much debate and has been adopted by different specialties of medicine to various degrees. However, acceptance of EBM within the field of anesthesiology has been delayed compared with other medical specialties. For example, the Cochrane Anesthesia Review Group (CARG) was not established until February 2000; to date, only 17 completed reviews exist within the Cochrane database (Table 1). Pronovost et al. (2) speculated that difficulty in applying the principles of EBM in our daily practice is in part due to 1) the relative paucity of properly performed randomized clinical trials, 2) the dispersal of anesthesia-related literature across medical and surgical subspecialties and basic science, and 3) the need to make rapid clinical decisions, which precludes an “on-the-spot” systematic review. To overcome these barriers, and distribute EBM in anesthesiology to a wider readership, the CARG and Anesthesia & Analgesia formed a collaborative link in 2002 (3). As part of the collaboration, selected Cochrane reviews are highlighted within the journal. It was anticipated that feedback from the readers of Anesthesia & Analgesia would help focus CARG’s future projects (3).
Ideally, the clinical question evaluated during a systematic review should be answered using the highest level of evidence available. The evidence for clinical questions varies in quality. Level 1 evidence, as proposed by the Oxford Centre for Evidence-based Medicine (www.cebm.net/levels_of_evidence), would be considered extremely strong evidence to guide practice. This requires multiple randomized controlled trials (RCTs) with homogeneity, an individual RCT with a narrow confidence interval, or an “all or none” situation (Table 2). However, there has been little standardization for grading evidence (4) and recommendations for a single question often vary greatly depending on the methodology of the review.
The first “Cochrane Corner” report assessed the effects of ventilation with smaller tidal volumes on the morbidity and mortality of patients affected by acute lung injury and acute respiratory distress syndrome (5). Five RCTs involving 1202 patients were included in the review. The authors concluded that mortality at day 28 was significantly reduced by smaller lung volumes (7 mL/kg or less) and lower airway driving pressure (plateau pressure 30 cm H2O or less). However, the effect on long-term mortality was less certain. Although there were concerns regarding the heterogeneity of the clinical trials and the potential adverse effects of using smaller tidal volumes, this systematic review determined the intervention (smaller tidal volumes) to be effective and without increased risk.
It is estimated that approximately 10,000 Cochrane reviews are needed to cover all health care interventions (www.cochrane.org/index2.htm). This may be an underestimate, based on the narrow focus of each systematic review. For example, in this issue of Anesthesia & Analgesia, Zaric et al. (6) compared the frequency of transient neurologic symptoms (TNS) after spinal anesthesia with lidocaine compared with other local anesthetics. A total of 1349 patients in 14 RCTs or quasi-RCTs, regardless of blinding, were included. The authors concluded that the relative risk of developing TNS after spinal anesthesia with lidocaine was 7 times that of bupivacaine, prilocaine, and procaine, and similar to that of mepivacaine. These results are consistent with those of previous (individual) RCTs and observational series published during the past decade (7,8). However, since no subgroup analysis of patient positioning, ambulation status, or lidocaine concentration/dose/baricity could be performed, the conclusions fall well short of the accepted medical knowledge of the subject; additional systematic reviews would be required to evaluate the effect of these interventions on TNS.
Clinical questions pertinent to the practice of anesthesiology frequently do not meet criteria for high-level evidence as judged by EBM advocates. This may lead to a sense of frustration when reading EBM appraisals of clinical questions related to anesthesiology. Anesthesiologists and evaluators of the specialty must appreciate that the practice of anesthesiology is supported by varying levels of evidence. It is critical to remember that although a clinical question may not be supported by high-level evidence, evidence exists along a spectrum; there is always evidence to guide patient care. In the absence of higher level evidence, observational data or even pathophysiological reasoning may be the best evidence available to support a clinical decision. Pathophysiological principles used to determine clinical practice are referred to as “first principles.” For example, the use of perioperative β2-adrenergic receptor antagonists in patients with cardiac disease undergoing major surgery is supported by Level 1 evidence. Conversely, the use of pulse oximetry to decrease intraoperative mortality is not supported by a similar level of evidence (9). The authors of this Cochrane review concluded, “we have found no evidence that pulse oximetry affects outcome of anesthesia.” Furthermore, “the value of perioperative monitoring with pulse oximetry is questionable in relation to improved reliable outcomes, effectiveness and efficiency.” These statements should not be confused with the suggestion that pulse oximetry be abandoned. Rather, support for pulse oximetry is provided by a lesser level of evidence—“first principles” and observational data—which show that pulse oximetry can detect hypoxemia and related events.
While clinicians need to appreciate that the level of evidence supporting a given practice is important, the hierarchy is not absolute. For example, if treatment effects are sufficiently large and consistent, observational studies may provide compelling evidence in the absence of a RCT. Another Cochrane review investigated the role of epidural blood patching for the prevention and management of postdural puncture headache (PDPH) (10). The authors included only randomized trials. As a result, only 3 trials with a total of 77 patients were involved. On this basis the authors concluded, “until further randomized evidence becomes available, epidural blood patching should not be used routinely for the prophylaxis or treatment of PDPH. At present, we believe that epidural blood patching should be reserved for exceptional cases only.” They also noted that randomized trials including at least a few hundred patients “must be carried out before the balance of risks and benefits of this potentially helpful intervention can be properly assessed.”
This meticulously performed systematic review demonstrates that clinical guidelines are only as good as the evidence and judgments upon which they are based. Most anesthesiologists would believe it was a critical flaw in the methodology to evaluate the effectiveness of epidural blood patching on both the treatment and prophylaxis of PDPH. Likewise, it is unfortunate that the authors did not follow the basic approach to EBM: when adequate higher level evidence is unavailable, then studies of a weaker strength are reviewed and/or pathophysiologic reasoning is applied. These conclusions, as stated, are not helpful to the clinician, and may indeed be detrimental. For example, it is very likely that the recommendation to perform a large RCT on the effectiveness of epidural blood patching will go unheeded- possible randomization into the placebo group would make recruitment of subjects problematic. In addition, it is doubtful that there would be equanimity among clinicians should such a study be attempted. Thus, we will likely never have results from an RCT, much less multiple RCTs, addressing this topic. Furthermore, as a result of the Cochrane review, neurologists and neurosurgeons may be hesitant to refer patients with PDPH for epidural blood patching, and patients will have to wait for spontaneous resolution. Another Cochrane review by the same authors (11) concluded, “there is no good evidence from randomized trials to suggest that routine bed rest after dural puncture is beneficial and the role of fluid supplementation in the prevention of postdural puncture headache remains uncertain.” While the evidence for these practices may not be convincing, it is debatable that treatment associated with limited potential harm (but possible benefit) be withheld due to lack of high-level evidence.
Finally, EBM appraisal of a single clinical question may arrive at different conclusions depending on the methods used. In addition, a systematic review of outcomes after institution of evidence-based recommendations would be helpful to demonstrate that EBM truly improved patient care. For example, the concept for EBM, as described by Sackett (12), has been applied as the basis for recommendations regarding antithrombotic therapy since 1986 (13). The recommendations, based on the effectiveness of a treatment modality in reducing the frequency of asymptomatic deep venous thrombosis (DVT) as detected by imaging, have progressively increased the level and duration of thromboprophylaxis over the last two decades. However, the efficacy of these measures in reducing the frequency of nonfatal/fatal pulmonary embolism (PE) or post-phlebitic syndrome—the clinical outcomes of importance—has not been documented. A recent Cochrane review examining the effectiveness of antithrombotic therapy for prevention of thromboembolism after surgery for hip fracture concluded that while heparin and low molecular weight heparin (LMWH) protected against DVT, there was insufficient evidence to confirm protection against PE or overall benefit (14). In contrast, the American College of Chest Physicians (ACCP) strongly recommends, with certainty that the benefits outweigh the risks, that these patients receive thromboprophylaxis with LMWH or standard heparin perioperatively (15).
The conflicting conclusions, despite “systematic review” methodology and inclusion of the same randomized clinical trials result from the ACCP reliance on a surrogate end-point–asymptomatic DVT. Without longer follow-up of patients and sample sizes large enough to evaluate the incidence of PE and post-phlebitic syndrome, it is impossible to know if the evidence has been applied correctly. Thus, even the most respected evidence-based guidelines may require “alterations.” This has significant practice implications for anesthesiologists who tailor their analgesic plan to reduce the risk of spinal hematoma.
The examples described above illustrate some of the strengths and weaknesses of EBM as it pertains to the practice of anesthesiology. The underlying principles of EBM will serve the clinician well. However, it is important to remember that EBM will never tell the clinician what to do. Rather, EBM provides the tools by which clinicians determine the best available evidence to combine with caregiver experience and patient preferences when practicing medicine.
Performing a systematic review is much like weaving cloth. Although the design may be the same, depending on the age, strength, and quantity of the material, the final product may be either silk brocade or polyester. While both cover the subject, they vary greatly in intrinsic worth. As clinicians, it is our responsibility to achieve the best “fit.”
1. Sackett DL, Haynes RB, Tugwell P. Clinical epidemiology: a basic science for clinical practice. Boston: Little Brown & Co, 1985.
2. Pronovost PJ, Berenholtz SM, Dorman T, et al. Evidence-based medicine in anesthesiology. Anesth Analg 2001;92:787–94.
3. Pedersen T, Moller AM, Cracknell J. The mission of the Cochrane Anesthesia Review Group: preparing and disseminating systematic reviews of the effect of health care in anesthesiology. Anesth Analg 2002;95:1012–8.
4. Ebell MH, Siwek J, Weiss BD, et al. Strength of recommendation taxonomy (SORT): a patient-centered approach to grading evidence in the medical literature. Am Fam Physician 2004;69:548–56.
5. Petrucci N, Iacovelli W. Ventilation with smaller tidal volumes: a quantitative systematic review of randomized controlled trials. Anesth Analg 2004;99:193–200.
6. Zaric D, Christiansen C, Pace NL, Punjasawadwong P. Transient neurologic symptoms after spinal anesthesia with lidocaine versus other local anesthetics: a systematic review of randomized controlled trials. Anesth Analg 2005;100:1811–6.
7. Freedman JM, Li D, Drasner K, et al. Transient neurologic symptoms after spinal anesthesia: an epidemiologic study of 1863 patients. Anesthesiology 1998;89:633–41.
8. Pollock JE, Neal JM, Stephenson CA, Wiley CE. Prospective study of the incidence of transient radicular irritation in patients undergoing spinal anesthesia. Anesthesiology 1996;84:1361–7.
9. Pedersen T, Dyrlund Pedersen B, Møller AM. Pulse oximetry for perioperative monitoring (Cochrane Review). In: The Cochrane Library
, Issue 4, 2004. Chichester, UK: John Wiley & Sons, Ltd.
10. Sudlow C, Warlow C. Epidural blood patching for preventing and treating post-dural puncture headache (Cochrane Review). In: The Cochrane Library, Issue 4, 2004. Chichester, UK: John Wiley and Sons, Ltd.
11. Sudlow C, Warlow C. Posture and fluids for preventing post-dural puncture headache. (Cochrane Review). In: The Cochrane Library
, Issue 2, 2001. Chichester, UK: John Wiley & Sons, Ltd.
12. Sackett DL. Rules of evidence and clinical recommendations on the use of antithrombotic agents. Chest 1986;89:2S–3S.
13. American College of Chest Physicians and the National Heart, Lung and Blood Institute National Conference on Antithrombotic Therapy. Chest 1986;89:1S–106S.
14. Handoll HH, Farrar MJ, McBirnie J, et al. Heparin, low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures (Cochrane Review). In: The Cochrane Library
, Issue 4, 2004. Chichester, UK: John Wiley and Sons, Ltd.
15. Geerts WH, Pineo GF, Heit JA, et al. Prevention of venous thromboembolism: the Seventh ACCP conference on antithrombotic and thrombolytic therapy. Chest 2004;126(3 Suppl):338S–400S.