Pain disaggregation theory—statistical nonsense or a pointer to a paradigm for quantum nociception?

I. D. Conacher*,1, G. Sudarshan2 and A. K. Soni3

1 Department of Thoracic Anaesthesia, Freeman Hospitals Trust, Freeman Road, Newcastle upon Tyne NE7 7DN, UK. 2 University Cincinnati College of Medicine, Ohio, USA. 3 Westwood, Massachusetts, USA

Corresponding author. E-mail: i.d.conacher@btinternet.com

Accepted for publication: April 2, 2003


    Abstract
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
Background. The various patterns of patients’ experience of treated acute post-thoracotomy pain exemplify the phenomenon of disaggregation. The intent in this study was to define a theory of disaggregation with a hard-wired neuroanatomical model of thoracotomy pain.

Methods. In order to distinguish the disaggregated nociception conducted along one of three possible pathways, the vagus, the phrenic and, in this study, the intercostal nerves, data from 143 patients undergoing thoracic surgery, and that from two previously conducted studies of multimodal analgesic regimens, were reviewed. The values of one subjective outcome measure (verbal rating score) at different stress levels—at rest, on raising the arm, and on coughing (dynamic pain scores)—were used to construct individuals’ charts (pain profiles) of the progress of pain relief over time. These were batched, and analysed using statistics of summary measures.

Results. This was a crude exercise in the handling of redundant data, but there is a suggestion that it is possible to distinguish a disaggregated route by an effect of a treatment on a mass of nociception.

Conclusions. This information could underpin a paradigm of quantum nociception, and has potential to quantify aspects of analgesia practice and current and future neurophysiological theories of pain. Prospective studies are warranted.

Br J Anaesth 2003; 91: 279–81

Keywords: pain, postoperative; statistics


    Introduction
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
Recently, the term disaggregation has been used to describe situations in which analgesic treatment alters the pain profile but not necessarily the degree of discomfort experienced by patients.1 By interfering with the routes of nociception, therapy often does not influence the magnitude of the pain but merely its perception, in effect revealing and changing one discomfort for another.2

In thoracic surgery, there is an example of significance because of the distress caused. The different pathways involved in nociception, and which contribute to the total post-thoracotomy pain experience, typically become broken up into constituents, or disaggregated, by analgesia regimens that interfere with nociception conducted by intercostal nerves. As a result, in the absence of intercostal nerve stimulation, patients complain of, and are discomforted by, shoulder-tip pain resulting from predominance of nociception via the phrenic nerve. This classic of referred pain affects 80% of patients after posterolateral thoracotomy treated with clinically effective analgesic regimens.3 4


    Methods and results
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
Some old data were re-examined in order to test a hypothesis that if a hard-wired neuroanatomical model for thoracotomy pain was used, disaggregation can be defined and measured. The model prescribes that most of the incisional and non-incisional pain from a thoracotomy is conducted by the intercostal, vagus and phrenic nerves. A premise was made – that movement of the chest wall following a posterolateral thoracotomy triggers constituents of nociception, mainly incisional, that are transmitted by ipsilateral intercostal nerves. The stimulus, previously used as a stress, was a deliberate arm movement (as though using a comb) ipsilateral to the thoracotomy.5 6 It was presumed that the action of coughing, in contrast, would stimulate nociception, both incisional and non-incisional, along all ipsilateral hard-wired routes (i.e. intercostal, phrenic and vagus nerves). By inference, in those subjects who had some form of intercostal nerve blockade, subtraction of the nociception measure (e.g. pain analogue score) of arm movement from that on coughing would signal the proportion of nociception being conducted along the phrenic and vagus nerves, and be a comparator for those whose analgesic regimens did not include intercostal nerve deafferentation.

The pain analogue scores (five-point verbal rating score (VRS)) were taken from two previously conducted ethically approved, double-blind placebo-controlled studies involving 143 patients, all of whom had had a posterolateral thoracotomy.5 6 All patients, as part of the test regimen or as controls, had access to a patient-controlled-analgesia system containing morphine. For each patient, a 10-h pain profile was constructed from on-the-hour VRS data at each of the three levels of stress (rest, arm movement, coughing). The area under the curve (AUC) of the pain profile was calculated and tested by analysis of variance using the Minitab statistical package.7

Statistics were prepared for each of the different types of analgesic regimen from the batched data of those from each study who, in effect, received the same analgesia regimen. These were: paravertebral block regimen (intercostal nerve blockade, labelled PVB); intrathecal fentanyl regimen (no intercostal nerve block, neuraxial opioid, labelled ITF); and patient-controlled analgesia only (non-intercostal nerve block, systemic opioid, labelled PCA).

Table 1 shows the information derived from the AUC summary measures. The mean baseline (at rest) value for each analgesic regimen has been subtracted from that when the arm movement or cough stimulus is activated. In addition, a measure of the different nociception magnitude is achieved by subtracting the value for arm movement from that on coughing for the treatment regimens ITF, PVB and PCA. There are no statistical differences in the values that represent change from rest to movement, but a highly significant change (P<0.001) in those from rest to coughing, and between movement and coughing, as a result of intercostal nerve blockade (Table 1). Quality of pain control, in terms of reduced magnitude of pain in the ITF group, was superior overall to that experienced by the other groups. However, when analysed in this fashion, it was the ITF group in which the negative value represents pain appearing to have been worse on movement rather than on coughing (–3.5 vs 2.13 or 1.14)—the reverse of the situation in the other groups, and of that anticipated.


View this table:
[in this window]
[in a new window]
 
Table 1 Subtraction of summary measures (area under the curve) at three levels of nociception. The values represent changes in nociception from background/baseline (at rest) to when the two stresses (arm movement and cough) are activated and compared. ITF, intrathecal fentanyl; PVB, paravertebral blockade; PCA, patient-controlled analgesia (morphine); CI, confidence interval. Data are mean (SEM)
 

    Comment
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
Primarily it must be stated this was an exercise in data manipulation. The numbers in one group are very small (n=10). Proof of any hypothesis will always be questionable for these reasons. Therefore, the prosaic conclusion is that this is a data-dredging exercise. Analysing a patient’s pain experience as though it were a kind of waveform that can be segmented and revealed by an abstract prism is simplistic. Although it is intuitive that analgesia reduces the mass, nociception generally is not regarded either as a quantum phenomenon or as a dynamic. And, logically, according to current concepts of postoperative pain, removal of nociception from one hard-wired route will not necessarily mean reduction in the amount of pain perceived, complained of, or measured. Other confounders are present, for example there are separate afferents from nociception initiated by the contrived stimuli of arm movement and coughing. In addition, there are all the problems of subjectivity and pain score measurement against a background of changing pain magnitudes with time and clinical circumstance. Therefore, in recognising these as fundamentals it was presumed that the null hypothesis would have been proved beyond doubt.

The null hypothesis was that it would not be possible to distinguish intercostal nerve nociception from the mass of nociception (i.e. no detectable difference between those who had intercostal nerve deafferentation and those who did not). Implicit was that the responses of patients to the requests to grade pain would be the same irrespective of the regimen or the pain-provoking stimulus; that the sentient would be unable to detect a difference between two contrived painful stimuli; and that, even if they could, the methodology of pain analogue scoring is so suspect that it would not be quantifiable. In casting doubt on the proof of the null hypothesis, it can be argued that intercostal-conducted nociception can be abstracted from that conducted by the vagus and phrenic nerves, and may be quantifiable. Besides, other intriguing issues have been raised.

Quantifying nociception could clarify the sites of action of analgesics. For example, in the field of thoracic analgesia, there has been debate about interpleural local anaesthetics. Whether action is by diffusion to intercostal nerves or on pleural sites that also feed nociception along the phrenic and vagus nerves could be demonstrated with a disaggregation methodology.

And it is not only local anaesthetic pharmacodynamics to which pain disaggregation theory could be applied. The observation that the fentanyl group were discomforted more by arm movement than coughing, in comparison to those who had intercostal nerve block or systemic opioid, is probably a reflection of the small numbers in the former and not repeatable. But it may have other significance. That fentanyl suppresses the cough reflex can be discounted. The cough manoeuvre was contrived for the studies and is not a natural observation. A distant possibility is that the dynamic pain scoring technique has shown a marker for a segmental action of opioids such as fentanyl. Also, a new problem is emerging in clinical practice – that of acute tolerance to opioids following the intraoperative use of remifentanil. The descriptions of the pain experiences of some of these patients are new and there is anecdotal evidence of disaggregation phenomena to such an extent that old operations are requiring the application of novel multimodal analgesic regimens.

Applying the logic that follows from a theory based on quantum nociception extends to clarification of the activity of a third component of multimodal regimens: the non-steroidal anti-inflammatory drugs. Current belief is that these drugs reduce a ‘nocigenic soup’ of inflammatory mediators. This may be a quantifiable reduction measurable with double-blind, placebo-controlled methodology. The effect and influence of pre-emptive techniques on postoperative pain could similarly be measured. Therefore, it is to be hoped that, besides enlightening the science of clinical analgesia, a paradigm of quantum nociception ultimately will lead to improvements in patient care.


    References
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
1 Conacher ID. Anaesthesia for thoracoscopic surgery. In: Badner N, ed. Anaesthesia and Minimally Invasive Surgery. Clin Anaesthesiol 2002; 16: 63–80

2 Conacher ID. Pain relief following thoracic surgery. In: Gothard JWW, ed. Thoracic Anaesthesia. Baillieres Clin Anaesthesiol 1987; 1: 233–57[ISI]

3 Burgess FW, Anderson DM, Colonna D, Sborov MJ, Cavanaugh DG. Ipsilateral shoulder pain following thoracic surgery. Anesthesiology 1993; 78: 365–8[ISI][Medline]

4 Scawn N, Pennefather SH, Soorae A, Wang JY, Russell GN. Ipsilateral pain after thoracotomy with epidural analgesia: the influence of phrenic nerve infiltration with lidocaine. Anesth Analg 2001; 93: 260–4[Abstract/Free Full Text]

5 Sudarshan G, Browne BL, Matthews JNS, Conacher ID. Intrathecal fentanyl for post-thoracotomy pain. Br J Anaesth 1995; 75: 19–22[Abstract/Free Full Text]

6 Soni AK, Sudarshan G, Conacher ID. Evaluation of a clinical methodology to detect pre-emptive analgesia. International Monitor (Regional Anesthesia). Abstract Issue. 11th Annual ESRA Congress 1993; 1 [Suppl.]: 47

7 Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. Br Med J 1990; 300: 230–5[ISI][Medline]