Review of clinical utility of pain classification part 1

Part I Mechanisms-based classification of pain: the algorithms

Understanding pain mechanisms is high on research agendas, but we’re still struggling translating this into practice. A number of frameworks have been published recently, which outline criteria for assessing dominant pain type in individual patients. These include:

  1. Guidelines on neuropathic pain assessment (Neuropathic Pain Special Interest Group of the IASP (NeuPSIG) [1])
  2. Classification of low back-related leg pain (Schafer et al. [2])
  3. Mechanisms based classification of low back pain (Smart et al. [3-5])

Algorithm 1, which we’ll refer to as the ‘NeuPSIG algorithm’ provides a framework for assessing the degree to which pain is neuropathic. It considers: clinical evidence of a possible nerve lesion concurrent with pain distribution that is neuroanatomically logical (Criterion 1) and signs of conduction loss and/or signs of sensory hypersensitivity within the pain distribution evaluated via clinical assessment (Criterion 2). Should these two criteria be met, the pain may be deemed to be ‘probable neuropathic pain’. A diagnosis of ‘definite neuropathic pain’ is reserved for when concurrent laboratory/radiological evidence supports clinical findings. If none of these criteria are met, the pain is deemed to be ‘unlikely neuropathic’ in nature.

Algorithm 2, the ‘Schafer algorithm,’ outlines a framework for assessing pain mechanisms underlying referred pain from the lumbar spine. This algorithm outlines a hierarchical system where four classifications are considered in order; namely, ‘neuropathic sensitization’, ‘denervation’, ‘peripheral nerve sensitization’ and ‘musculoskeletal pain’. Neuropathic sensitization is deemed present when a score of ≥12 is recorded on the Leeds Assessment of Neuropathic Symptoms and Signs. If this criterion is not met, then results from the assessment of neurological function are considered; if positive, then the pain may be classified as ‘denervation’. If negative, then results from neural tissue provocation testing are considered e.g. straight leg raise; if positive, the classification ‘peripheral nerve sensitization’ is allocated. Finally, if negative, by process of elimination, the pain can be deemed ‘musculoskeletal’ reflecting somatic referred pain.

Framework 3, which we’ll call the ‘Smart classification’ comprises clinical signs and symptoms indicative of ‘nociceptive’, ‘peripheral neuropathic’ and ‘central sensitization’ pain. These criteria were developed using a Delphi consensus approach and validated in a population of people with low back pain. Smart et al. have not indicated how these criteria should be used in clinical practice e.g. how to weight criteria.

There are pros and cons to each algorithm. The NeuPSIG algorithm is useful if you want to identify the presence or absence of neuropathic pain and the guidelines are easy to follow. This is important given the ongoing under-recognition of neuropathic pain in clinical practice. It also provides a method of identifying a discrete lesion that may warrant more specific investigation or treatment, e.g. anti-convulsant medication, and thus may be very useful in practice. One key issue around this algorithm is that, On the other hand, if the NeuPSIG algorithm is taken literally, it fails to recognize neuropathic pain beyond identifiable lesions of the somatosensory system. This may be problematic in situations like chronic neuropathic pain where pain often extends beyond conventional neuroanatomical boundaries. Further, there is debate about the definition of neuropathic pain given that processes of altered nociceptive processing involve the nervous system, and thus could be considered neuropathic. Yet, these pain conditions might be excluded from this classification.

With respect to the Schafer algorithm, the hierarchical framework makes it simple to use. A useful aspect is the distinction made between pain associated with conduction loss ‘denervation’ and pain arising from neural tissue sensitivity in the absence of conduction loss. This is important clinically for reasons outlined above. On the downside, one might challenge the classification of neuropathic sensitization on the basis of results from one questionnaire.

One issue with the Smart classification is that it includes neurological deficits together with features of neural tissue mechanosensitivity into one category. Although frequently co-existing, it is important to separate those with a neurological deficit from those without, consistent with spinal triage. It also doesn’t guide the clinician in terms of how to use the information, making it more difficult to use. However, this likely reflects clinical practice where clinicians weigh up information from the history and clinical examination and make judgments accordingly.

Semantics and diagnostic criteria related to neuropathic pain is a hot topic for debate. Within the three algorithms studied, approaches differed; the NeuPSIG criteria for classification of neuropathic pain are more stringent than those listed by Smart et al [7], while Schafer et al. [2] separate denervation from peripheral nerve sensitization, consistent with a basic triage approach. While this is potentially confusing, it’s important to acknowledge that different opinions exist. The development of algorithms is an important step in translating research into clinical practice; however, more work is needed to examine whether stratification according to these systems leads to better patient outcome.

About Niamh Moloney

Niamh MoloneyNiamh Moloney is a musculoskeletal physiotherapist who, following her PhD at University College Dublin, moved to Australia and is currently a lecturer in Physiotherapy at Macquarie University. Her research relates to sensory profiling and assessment of pain sensitization in clinical practice. Outside of work, she is committed to regular ‘de-sensitising of her nervous system’ through yoga, rowing and exploring Sydney’s great bush walks.

References

  1. Haanpää, M., et al., NeuPSIG guidelines on neuropathic pain assessment. Pain, 2011. 152(1): p. 14-27.
  2. Schafer, A., T. Hall, and K. Briffa, Classification of low back- related leg pain- A proposed patho-mechanism based approach. Manual Therapy, 2009. 14(2): p. 222-230.
  3. Smart, K.M., et al., Mechanisms-based classifications of musculoskeletal pain: part 1 of 3: symptoms and signs of central sensitisation in patients with low back (+/- leg) pain. Man Ther, 2012. 17(4): p. 336-44.
  4. Smart, K.M., et al., Mechanisms-based classifications of musculoskeletal pain: Part 2 of 3: Symptoms and signs of peripheral neuropathic pain in patients with low back (±leg) pain. Manual Therapy, 2012. 17(4): p. 345-351.
  5. Smart, K.M., et al., Mechanisms-based classifications of musculoskeletal pain: part 3 of 3: symptoms and signs of nociceptive pain in patients with low back (+/- leg) pain. Man Ther, 2012. 17(4): p. 352-7.

Comments

  1. Keith Smart says

    Hi Niamh, Hi Wax
    Niamh, thank you very much for discussing our mechanisms-based classification system. We took a very pragmatic approach and attempted to identify clinical criteria assumed to reflect a relative dominance of nociceptive pain (NP), peripheral neuropathic pain (PNP) and central sensitisation pain (CSP) that clinicians could, we think, relatively easily identify from a clinical history and examination. They are of course not a means with which to ‘diagnose’ NP, PNP and CSP but rather an indirect attempt to categorise patients pain in a way that might usefully inform treatment selection and prognostic judgements.

    Wax-re the issue of a gold standard, we acknowledged the difficulty here, and to help clarify, I’ve copied/pasted this from our original Clinical Journal of Pain paper (2011; 27(8): 655-63)

    Discriminative validity designs require the identification
    of the “extreme groups” (i.e. pain type) and in the absence of
    a diagnostic gold standard the best alternative “reference
    standard”—defined as, “the best available method for
    establishing the presence or absence of a condition of interest,” may be expert clinical judgement. “Under this
    assumption the development of classification criteria
    becomes an exercise in determining which history [and]
    physical examination findings match the impression of an
    experienced clinician.”

    Hope that helps.
    Best wishes.
    Keith

  2. Niamh Moloney says

    Hi Wax,
    So, for the generation of the features in each classification, Smart et al. used a Delphi study process where they asked a number of experts what features they would expect for each classification. briefly, the validation process asked musculoskeletal physiotherapists to assess patients with LBP and (a) allocate a classification to each patient, and (b) tick the features (from a list generated from the delphi study) the patients presented with.
    I totally agree that there is no gold standard for pain classification and I think we have to use such classifications in a sensible way, knowing their pros and cons and using them as part of our clinical reasoning. We need to recognise the overlap in mechanisms that contribute to various pain states. I also think we should build from these algorithms, refining them as our knowledge builds.

  3. Hi Niamh,

    I am full of respect to the Smart et al. classification an their clinical utylity. However, what was a gold standard in their validation study? I know that, delphi study included ‘consensus’ in terms of which items (paind fetures) should be used. But gold standard remains unclear.