Postoperative pain management: are we ready to move beyond the ‘kitchen‐sink’ approach?

Surgery is a well-known cause of acute pain, and this pain experience can vary between individuals [1] and by type of surgery [2]. How should modern-day anaesthetists approach acute postoperative pain treatment for surgical procedures anticipated to cause severe pain after surgery? One commonly used strategy is to employ every type of applicable analgesic intervention (non-pharmacological, oral and intravenous opioid and non-opioid medications, multiple nerve blocks, continuous catheters) pre-emptively in all patients, with the goal of limiting or even eliminating the possible occurrence of poorly managed pain and long-term opioid use. We have some evidence that the use of more non-opioid analgesics will decrease the short-term use of opioids after surgery [3], and there are multi-society recommendations to use a multimodal analgesic approach [4]. While imprecise, this ‘kitchen-sink’ approach may appeal to our altruistic sense of obligation to prevent pain or treat it as effectively as possible in a standardised way for as many patients as possible. However, this approach is not likely to be the ideal target for the future of acute postoperative pain management, and there are important caveats to consider when employing it. Universal concurrent application of multiple analgesic modalities in anticipation of pain is an example of ‘practice-based evidence’, where practitioners use evidence supporting individual interventions to reinforce their combined use, despite the lack of evidence backing up the benefits of the combination. Pain management protocols are typically implemented in bundles, without clear measurement of the incremental benefits of each added modality or evidence of comparison of one combination of modalities vs. another [5]. The inherent evidence leap does not consider that combining various individual interventions also means stacking their potential complications, prolonging procedural time when procedures are applicable, increasing costs and requiring additional skills. To demonstrate the limitations of this strategy, we will examine its hypothetical application in a surgical population known to experience moderate-to-severe acute pain, namely total knee arthroplasty (TKA). Current recommendations for TKA clinical pathways [6] advocate for a kitchen-sink approach, whereby multimodal analgesia, motor-sparing regional anaesthetic techniques (including nerve block catheters) and peri-articular local anaesthetic infiltration are all administered concurrently to all patients. The practical application of this strategy in most centres that use it is a complex, resource-intensive clinical pathway that features some combination of the following: peri-operative oral paracetamol; non-steroidal anti-inflammatory drugs; opioids for severe breakthrough pain; adductor canal block with or without catheter infusion; and peri-articular local anaesthetic infiltration by surgeons or ultrasound-guided infiltration of the posterior capsule of the knee by anaesthetists. If we look more closely at the local and regional anaesthetic interventions specifically, it is noteworthy that the evidence underlying adductor canal block for TKA originates from trials that excluded local anaesthetic infiltration [7, 8]; the efficacy of anaesthetist-administered infiltration of the posterior capsule has been proven in studies that also excluded local anaesthetic infiltration; and the addition of postoperative continuous adductor canal infusion to single-injection adductor canal block did not yield clinically important benefits in a recent meta-analysis [9]. Therefore, one can argue that this particular regional analgesic kitchen-sink approach for patients having TKA includes interventions whose incremental benefits may not be applicable to every patient. Indeed, newer evidence suggests that, when performed well, local anaesthetic infiltration alone may be sufficient in the context of multimodal analgesia for many patients without additional nerve blocks [10]. This is important information given that many patients having TKA may not have access to a regional anaesthetist. What is the alternative? We believe that the future of acute pain management will employ a procedure-specific and personalised approach that matches the selection and duration of analgesic interventions to the projected person-specific experience of acute postoperative pain following surgery. Examination of pain trajectories following surgical procedures lends credence to this strategy, as available data indicate that not all patients will require the most aggressive acute pain management [1]. Indeed, only a portion of patients develop moderate-to-severe acute pain [1], which may vary by surgery [2], and certain patients may develop persistent post-surgical pain [11]. Therefore, in an ideal world, a procedure-specific personalised pain management approach would maximise clinical pathway efficiency and selectivity, thereby avoiding unnecessary interventions and associated costs, improving patient satisfaction and enhancing patient outcomes. While this personalised approach sounds amazing, it currently has some impediments. Although procedure-specific pain management strategies have been developed and published for a variety of surgical procedures [12] using standardised methodologies [13], they do not account for inherent differences between individuals [1]. Truly personalised pain management requires perfect prediction: the ability to predict who, among a group of patients undergoing the same surgical intervention, will proceed to develop moderate-to-severe pain, who will specifically benefit from a more inclusive approach to selecting analgesic modalities and who will benefit from longer vs. shorter duration of initial postoperative pain treatment. In this issue of Anaesthesia, Armstrong et al. report on a secondary analysis of the UK Peri-operative Quality Improvement Programme (PQIP) dataset with the goal of advancing our ability to predict severe postoperative pain [14]. Drawing on a dataset of 17,079 patients undergoing various surgical procedures at UK hospitals from 2016 to 2020, the authors developed a regression-based prediction model that uses 25 pre-operative variables to predict severe pain on the first day after surgery. Only 18% of patients in the sample experienced severe pain, and important factors that they found to be associated with severe postoperative pain included: smoking status; patient psychological well-being; and certain patient and surgical characteristics. The paper has numerous strengths from utilising a rich, national multicentre data source and including a range of potential factors, such as measures of patient psychology, function and pre-operative pain status. Typical assessments of prediction model performance consider both calibration (how well the model fits the data at hand) and discrimination (how well the model distinguishes between those who do and those who do not experience the outcome of interest). Here, the authors report a model with good calibration but modest discrimination [15]. The area under the receiver-operator characteristic curve or ‘c-statistic’ is the most commonly used approach to characterising model discrimination; in this case, the c-statistic of the main model using only pre-operative factors was 0.66. In general terms, this means that a randomly selected patient who truly had severe pain would have a higher predicted probability only 66% of the time using this model when compared with another patient who did not have severe postoperative pain. When intra- and postoperative factors (e.g. presence of epidural anaesthesia, pain and mobility data) were added to the model, the authors achieved better discrimination (c-statistic 0.7); however, this value still falls on the border of ‘poor discrimination’ and ‘acceptable’ according to their own interpretation guide [14]. As a result, despite identifying potential peri-operative factors that may be associated with the outcome, this model cannot reliably predict which patients will have severe postoperative pain before they have surgery. What does all this mean? Most importantly, as the authors acknowledge, the prediction of postoperative pain based on pre-operative factors is a fraught task and a problem not yet solved. Perfect prediction of postoperative pain before surgery may never be possible if an individual's unique physiologic or pathophysiologic response to injury is a critical factor and cannot be known pre-operatively. Indeed, the modest discrimination afforded by the model by Armstrong et al., despite access to a richly detailed clinical dataset, offers a humbling take on the complexity and challenge of consistently identifying those patients at highest risk of experiencing severe pain after surgery [14]. At the same time, the work still offers important insights for future research and practice. The authors' observation that nearly 1 in 5 patients in this large national cohort experience severe pain after surgery emphasises the importance of ongoing efforts to improve pain preventation and postoperative pain care. Despite its limitations, the model developed by Armstrong et al. serves as a step forward towards personalised peri-operative pain management by identifying 25 factors that serve as partial predictors to identify at-risk patients before surgery who may develop severe postoperative pain [14]. These predictors may be combined with others that have been associated with greater than expected postoperative pain, such as pain catastrophising [16] or serum biomarkers [17], to aid modern-day anaesthetists and acute pain services in the assignment of patients to specific pain management protocols, prioritisation for nerve block techniques and allocation of resources. Although we are not ready for individualised pain medicine yet, a stratified approach such as this represents an advance over the kitchen sink and may be the best way to maximise the value of specialised acute pain services at the present time. This material is in part the result of work supported with resources and the use of facilities at the Veterans Affairs Palo Alto Health Care System. The contents do not represent the views of the Department of Veterans Affairs or the United States Government. EM is an Editor of Anaesthesia. No other competing interests declared.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.