No clear consensus exists regarding the optimal approach in treatment of post-operative pain. We have demonstrated a considerable methodological variability by reviewing the current literature regarding postoperative pain treatment after total hip and total knee arthroplasty (THA and TKA).
Materials and Methods
In two systematic reviews of all analgesic interventions in the immediate postoperative period after THA or TKA we included a total of 171 randomized controlled trials. The clinical hetero-geneityof the included trials were described for the following issues; risk of bias (Cochrane); type of anesthesia, rescue analgesics, and basic analgesic regimen (underlying non-opioid analgesic regimen that was administered to all patients in both intervention and control groups). Basic regimens may influence the level of pain from which the intervention should provide the analgesic effect and thereby the propensity to detect a clinical relevant effect (assay sensitivity). Furthermore, we evaluated the assessment of pain scores, adverse effects and length of stay along with assay sensitivity of morphine consumption and pain scores.
Results and Discussion
Substantial clinical heterogeneity in study methodology and setup amongst trials regarding assessment tools and underlying perioperative analgesic regimens was demonstrated. Incomplete reporting of relevant outcomes regarding pain and adverse effects was present in the majority of the studies. Large variation was demonstrated for mean cumulated 0-24 hours morphine consumption in the control groups of the included trials; for THA 4-63 mg and for TKA 5-116 mg. This also applied for pain scores as we demonstrated a wide range in VAS (0-100) at rest; for THA 3-74 mm and for TKA 0-82 mm. High risk of bias.
Reaching a global consensus regarding optimal trial methodology of pain studies, including evaluation of pain and adverse effects, use of basic analgesic regimen and rescue medication may reduce the variability and heterogeneity in trials investigating postoperative pain. The registration of adverse effects was appallingly low. We recommend registration nausea and vomiting, sedation, dizziness, pruritus, intervention-specific adverse events (e.g. motor blockade after nerve block), and serious adverse events. The majority of unclear bias domains identified in the included trials, could have been avoided by use of proper methodology and study design according to CONSORT.
We believe that the heterogeneity demonstrated in administration of perioperative analgesics, registration of outcome assessments, and the risk of bias could be reduced prospectively by achieving international consensus regarding well-designed protocols for postoperative pain research. Furthermore, the demonstrated shortcomings should be taken into account when drawing conclusions based on the current literature.