Authors: S. Munirama; G. McLeod
Br J Anaesth. 2015;114(4):543-545.
Anaesthetists provide personalized care. Preoperative assessment guides decision-making, and awareness of risk based on clinical experience allows anaesthesia to be conducted in a way that aims to achieve the best possible clinical outcome for each patient while minimizing side-effects. This traditional approach has led to the use of a wide variety of good, safe anaesthetic techniques.
Making sense of clinical outcomes becomes much more complex when one considers that the nature of anaesthesia may change simply according to an intelligent clinical ‘hunch’ rather than written evidence; what the anaesthetist feels and fears based on training and experience is important in delivering good care.
Conversely, a number of well-intended protocol-based approaches to delivery of anaesthesia and postoperative pain relief have recently been undertaken. The Prospect group[1,2] took a surgery-specific approach; for specific surgical procedures, a consensus was reached between experts and recommendations for anaesthetic management published. Accelerated recovery programmes have gone further and distributed guidelines stipulating drugs and dosages for all patients undergoing anaesthesia as part of a perioperative rehabilitation package. A recent national audit of the anaesthetic management of hip fracture has recommended standardization of spinal anaesthesia with regard to patient position, mode of sedation, dose of intrathecal bupivacaine, and choice of intrathecal opioid.
There is no doubt that when using ‘one size fits all’ therapeutic regimes, some patients experience good outcomes; pain relief is good, rehabilitation is rapid, and hospital discharge is achieved within 2–3 days after surgery. On the other hand, other patients suffer moderate to severe pain and are difficult to manage; pain limits mobility and discharge is delayed. Related pain outcomes, which should be considered but are rarely measured, include the poorly understood profile or trajectories of postoperative pain during the first month after hospital discharge, and the now well-recognized association between surgery and emergence of chronic pain.
Physicians have long recognized the highly variable response to drug therapy; only 30–70% of patients respond positively to any drug, and patient response to analgesics is also bimodal. Recent classification of patients into analgesic ‘responders’ or ‘non-responders’ describes the distribution of pain in a much more meaningful way. Put simply, the bounds of patients experience are much more meaningful than the average and standard deviation of a clinical outcome. Nevertheless, the basic central nervous system mechanisms that account for such variability remain unknown, and randomized controlled trials (RCTs), in the main, are restricted to comparisons of anaesthetic drugs or technique, rather than exploring the relationships between fundamental cellular mechanisms and outcomes.
Stratified medicine has been described as the targeting of treatments according to the biological or risk characteristics shared by subgroups of patients[10–12] and forms an important component of the Medical Research Council’s research strategy. It provides a holistic framework for investigating and understanding why some patients fail to respond to standard therapy by consideration of not just the physical characteristics of disease but also the overarching effect of psychological well-being, environmental exposure, and genetics.
Personalized medicine uses genetic biomarkers, imaging techniques, or validated questionnaires in order to discriminate between low- or high-risk groups and predict which patients are more likely to respond to particular therapies, providing ‘the right therapy, for the right patient, in the right dose, at the right time’.
The potential clinical advantages of a stratified approach are: more targeted therapy; improved drug efficacy; less side-effects; and greater certainty of effect in all patient groups. Ultimately, knowing which patient characteristics accurately predict outcomes, a shift in medical culture should occur from treatment to prevention. On a broader level, and in the longer term, clinical trials will focus on investigating the impact of new drugs on biological mechanisms rather than symptoms, reducing drug development times, and lowering costs.
How might the speciality of anaesthesia adopt stratified medicine? Taking postoperative pain as an example, we already know that psychological factors such as depression, anxiety, and catastrophizing; physical tests such as endogenous modulation and suprathreshold temperature; and clinical symptoms such as severe postoperative pain play some part in predicting postoperative pain, and we are starting to understand more about the underlying genetics of pain. Polymorphisms of genes associated with the μ-opioid receptor, ATP-binding cassette subfamily B, interleukin-1 receptor antagonist, catechol-O-methyltransferase, cytochrome 2D6 enzyme, and melanocortin-1 receptor play a role in patient response to analgesics. The UK biobank of tissue and genetic data offers a ready resource for all clinicians and scientists to develop future preventive health strategies (http://ukbiobank.ac.uk).
Unsurprisingly, our own in-house mathematical models only account for 60% of the variation in postoperative pain (unpublished data) from routinely collected postoperative epidural data. Thus, the wide variation in patient pain experience that still exists, suggests a stratified mechanistic approach to postoperative pain is justified.
Clearly, many independent covariates have a profound influence on pain outcomes after surgery, and it would seem rationale to investigate the relative influence or weight of each predictor on outcome. Only when postoperative pain is viewed broadly, taking all independent predictors and their interactions into account, will we be able to construct a validated model of outcome and, from this, understand the relative influence of anaesthesia and analgesia using visual tools such as nomograms.
Practically, how should we investigate, say, pain after surgery? We would suggest pinpointing genetic biomarkers from tissue samples, identifying psychological predictors using validated questionnaires, and characterizing pain pathways using quantitative sensory testing and fMRI. Models or biomarkers identifying patients at greatest risk would be externally validated in another patient data set then selected for stratification within a randomized stratified clinical trial. The ideal clinical trial design comparing, for example, a new analgesic or intervention with a standard control group, would incorporate four groups by randomly stratifying patients within each treatment group according to their risk status. Typical group descriptors would be: (i) active group, low risk; (ii) active group, high risk; (iii) control group, low risk; and (iv) control group, high risk. The advantage of this model is that it provides two comparisons of relative risk between the new intervention and the control group to be calculated, one for low-risk patients and the other for high-risk patients. While little difference would be expected for the former, the greatest benefit would be expected in the latter.
In order to achieve a stratified approach, we need to form multidisciplinary groups incorporating clinicians, neuroscientists, psychologists, imagers, economists, and informatics specialists, and break down the barriers between universities and neighbouring teaching hospitals. Key to success will be the formation of large informatic databases, or clinical quality registries, linking preoperative and intraoperative clinical data, imaging, genomics, and cancer and drug prescription databases.
From a clinical perspective, we forsee patients undergoing the same operation receiving a different anaesthetic depending on their outcome risk. Taking knee arthroplasty as an example, patients deemed at low risk of experiencing outcomes such as postoperative pain or chronic pain may receive intra-articular local anaesthesia infiltration; patients at medium risk may obtain additional analgesia via a saphenous nerve catheter; and patients at high risk may also receive nerve blocks and anti-hyperalgesics. Best possible treatment in low- and high-risk groups, based on the clinical balance of efficacy and side-effects, would be determined by stratified RCTs.
We believe that, by adopting this philosophy, a stratified, holistic approach to research will bring a more personalized approach to anaesthesia.
- Fischer HB, Simanski CJ, Sharp C,et al.Aprocedure-specific systematic review and consensus recommendations for postoperative analgesia following total knee arthroplasty. Anaesthesia 2008; 63: 1105–23
- Joshi GP, Group PW, Schug SA,et al. Postoperative pain management: number-needed-to-treat approach versus procedure specific pain management approach. Pain 2013; 154: 178–9
- Scott NB. Wound infiltration for surgery.Anaesthesia 2010; 65(Suppl. 1): 67–75
- The National Hip Fracture Database: Anaesthetic Sprint Audit of Practice. London, 2014
- Andersen LO, Gaarn-Larsen L, Kristensen BB, Husted H, Otte KS, Kehlet H. Subacute pain and function after fast-track hip and knee arthroplasty.Anaesthesia 2009; 64: 508–13
- Lavand’home PM, Grosu I, France MN, Thienpont E. Pain trajectories identify patients at risk of persistent pain after knee arthroplasty: an observational study.Clin Orthop Relat Res 2014; 472: 1409–15
- Macrae WA. Chronic post-surgical pain: 10 years on.Br J Anaesth 2008; 101: 77–86
- Hingorani AD, Windt DA, Riley RD,et al. Prognosis research strategy (PROGRESS) 4: stratified medicine research. Br Med J 2013; 346: e5793
- Moore A, Derry S, Eccleston C, Kalso E. Expect analgesic failure; pursue analgesic success.Br Med J 2013; 346: f2690
- Stratified Medicine in the UK. Vision and Roadmap. Technology Strategy Board, 2013
- MRC Stratified Medicine Strategy Workshop. London: Medical Research Council, 2013
- Realising the Potential of Stratified Medicine. London: Academy of Medical Sciences, 2013
- Landau R, Kraft JC, Flint LY,et al. An experimental paradigm for the prediction of Post-Operative Pain (PPOP). J Vis Exp 2010; 35: http://www.jove.com/index/Details.stp?ID=1671 (accessed 18 September 2014)
- Strulov L, Zimmer EZ, Granot M, Tamir A, Jakobi P, Lowenstein L. Pain catastrophizing, response to experimental heat stimuli, and post-cesarean ection pain.J Pain 2007; 8: 273–9
- Puolakka PA, Rorarius MG, Roviola M, Puolakka TJ, Nordhausen K, Lindgren L. Persistent pain following knee arthroplasty.Eur J Anaesthesiol 2010; 27: 455–60
- Fernandez Robles CR, Degnan M, Candiotti KA.Curr Opin Anesthesiol 2012; 25: 444–9
- Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. New York: Springer, 2006