Pain characterization in cancer patients and the analgetic response to epidural morphine

Pain, 46 (1991) 3-X 0 1991 Elsevier Science Publishers ADONIS 030439599100145X

3 B.V. 0304-3959/91/$03.50

PAIN 01809

Clinical Section Pain characterization

in cancer patients and the analgetic response to epidural morphine Hikan Samuelsson and Thomas Hedner

Departments of Anaesthesiology, Central Hospital, Bora”s (Sweden), and Clinical Pharmacology, Sahlgrenska University Hospital, Giiteborg (Sweden) (Received

26 April 1990, revision

received

16 October

1990, accepted

14 December

1990)

In 48 patients with pain related to malignancy, a pain characterization was performed during oral Summary opioid therapy. After an optimal epidural morphine regimen had been established, the alteration in pain relief was evaluated by means of a visual analogue scale. The CSF and plasma morphine concentrations at minimum steady state were then analysed in 28 patients and related to the degree of pain relief. The efficacy of the spinal treatment ranked in the following order: somatic > visceral > radiating = 0, but the difference was only significant between the somatic and radiating pain groups. There was a tendency for continuous pain to be better relieved than intermittent pain. No correlations were found between the CSF or plasma morphine concentrations and the degree of pain relief, suggesting that not all pain impulses are modulated in a dose-dependent manner by morphine at the spinal level. Pain characterization may be instrumental in providing an optimal spinal opioid analgesia in malignancy. Moreover, there is a need for better defined diagnostic criteria for clinical pain characterization. Key words: Cancer pain; Analgesia; Epidural; Morphine;

Introduction

Opioids administered through the spinal route are claimed to produce clear and selective changes in virtually all of the experimental and clinical pain measures employed in man and in a variety of nociceptive tests in animals [32]. However, clinical practice in the treatment of pain related to malignancy reveals that the degree of analgesia obtained with spinal opioids is a matter of great variability 120,301. As yet there is no clear distinction between the extent to which pharmacokinetic or pharmacodynamic causes may form the main source of variability in analgetic responses in man. On the one hand, there is ample experimental evidence that, dependent on the dose, spinal opioids inhibit the nociceptive response in various animal models [31,32]. The degree of spinal modulation of thermal nociceptive stimuli in rats has also been shown to correlate with the concentration of morphine within the spinal cord 1131. On the other hand, based on

Correspondence to: Dr. HIkan Samuelsson, Dept. of Anaesthesiology, Central Hospital, Box 850, S-501 15 Boris, Sweden.

Differential

effects; Cerebrospinal

fluid

differential analysis of pain in patients with malignant disease, ArnCr and ArnCr [41 suggest that the origin of pain impulses and its temporal pattern could be of major importance for the analgetic response to spinal opioids. In particular, continuous pain, originating from deep somatic structures, seemed to respond favourably, while the outcome in visceral or intermittent somatic pain was more variable. Cutaneous pain and pain classified as neurogenic were reported to be relieved only occasionally. In order to test this hypothesis, we performed a study in which the analgetic response to epidural morphine was related to the actual morphine concentrations in plasma and CSF as well as to the nature of the pain in patients with pain due to cancer.

Patients

and methods

Forty-eight Caucasian patients, referred to the Pain Section, Dept. of Anaesthesiology, Central Hospital, Barbs, Sweden, between Aug. 1985 and May 1988, were included in the study. The study was approved by the Ethics Committee, Medical Faculty, University of Goteborg, Sweden, and all patients gave their oral and

I

PRIMARY TUMOUR LOCATIONS IN 4X (‘ANC’ER RECEIVING EPIDURAL. MORPHINI~ Tumour

location

Gastrointestinal Pulmonat) Female urogenital Malt uropcnital 13rea\t Unspecified

MAIN PAIN REGIONS R,ZI MORPHINE

PATIENTS

RF<‘L:.IVINCi FIPlDf~

No. of patient5

Pain region

No. of patient\

II 2 1 2.5 2

I Jpper extremities

.! 1 15

I’horax i\hdomen Lumbo-sacral and/or lotier extremities

37

I

pharmacokinetic steady-state conditions, the patients had to be on a certain dose regimen for at least 3 days before CSF and plasma sampling were performed. The mean daily epidural dose was then 27 + 34 mg. Sampling was carried out immediately before the next dose was to be given, thus representing the minimum steady-state concentration CC,, min). Lumbar puncture was carried out at the L2-L4 interspace in 28 of the 48 patients, and blood was collected from an antecubital vein for plasma morphine measurements. The plasma was separated by centrifugation; the samples were frozen at - 20 o C and later analysed by gas chromatography with electron capture detection as described previously [9]. Prior to epidural therapy, differential pain analysis was performed based on diagnosis of primary tumour, knowledge of localization and spread of metastasis, patient history and pain assessment from clinical examination. The pain was categorized as deep somatic or visceral as well as radiating with or without concomitant signs of neuropathic lesion. The temporal pattern was estimated as continuous or intermittent. Pain relief was evaluated using a 10 cm visual analogue scale (VAS) before the treatment was instituted (mean VAS 7 f 2) and when the first sample was obtained, i.e., 3-l 1 days after initiation of epidural therapy. The

written informed consent prior to inclusion in the study. Primary tumour locations and main pain regions of the 48 patients are presented in Tables I and II. Mean age was 65 + 13 years, and male to female ratio was 35/13. All patients suffered from advanced cancer and were receiving oral or systemic opioid therapy, with insufficient pain relief or unacceptable side effects, when the epidural treatment began. The mean daily opioid dose, equivalent to oral morphine, at inclusion in the study are given in Table III. An intramuscular to oral relative potency of 3 : 1 was used in the calculations [18]. Standard 16-gauge polyamide epidural catheters were used (Portex epidural minipack). The catheter was introduced at a level that approximately corresponded to the spinal cord segments involved in pain transmission. The treatment was routinely initiated by 2 or 4 mg of preservative-free morphine hydrochloride and then adjusted in increments until optimal pain relief was obtained or a level of intolerable side effects was reached. The patients were admitted to the intensive care unit for at least 12 h during this procedure. When an appropriate dose was established, the patients were prescribed intermittent doses on a fixed time schedule. In order to ensure

TABLE

IN -1XPATlE:NTS

111

PREVIOUS OPIOID CONCENTRATIONS

EXPOSURE, DAILY EPIDURAL MORPHINE DOSE IN SUBGROUPS OF CANCER PATIENTS RECEIVING

Pain subgroups

Deep somatic

Opioid dose at inclusion in study

Epidural morphine dose at steady state

(mg/day)

(mg/day)

*

174 f 15s

(n = 17)

1s+

7

AT STEADY EPIDURAL

STATE, PLASMA AND MORPHINE THERAPY

Morphine

concentrations

Plasma

CSF MORPHINE

(ng/ml) CSF

780 f 1 146

4.3 + 4.2 (n = IO)

Visceral

I66 * 244

(n = 0)

36 & 40

5.2+

1537 + 262’)

I.6 (n = 4)

Radiating

227 t_ IS7

(n = 4)

46 f 33

25 900

34.2 (n=

Somatic

plus visceral (n = 5)

97 + 65

15t

6

8.8 i

I) 279 _t

7.0

134

(n = 4) Somatic

plus radiating

(n = 12)

107+

64

36 + 47

17sY’)t7Y3

17.1 f 19.5 (rl = 9)

* Equianalgetic observations.

to oral morphine,

Intramuscular:

oral relative

potency

of 3 : I is used in the calculations.

Shown are means Z!ZS.D. n = number

of

reduction in the VAS score after introduction of epidural morphine was calculated and expressed as percent (% VAS) improvement compared to the initial VAS score. Results are presented as means + S.D. Statistical calculations were performed using the Microsoft Statview. Linear regression was calculated by the method of least squares as well as by Spearman’s rank-order correlation. Statistics were evaluated by the Mann-Whitney U test. A P value of 0.05 or less was considered to be significant. CSF

morphine

Css

min

ng/ml

Fig. 2. Pain relief in relation to CSF minimum steady-state concentrations of morphine after administration of epidural morphine. No correlation was found between the degree of pain relief and CSF morphine concentrations.

Results The CSF and plasma morphine concentrations related to the degree of pain relief at C,, min are shown in Figs. 1 and 2. It is evident that there is large interindividual variation in pain relief caused by morphine. For example, while a CSF morphine concentration of 40 ng/ml could provide almost complete pain relief in one patient, a lOO-fold or higher concentration in another case failed to produce acceptable analgesia. For the study population from which samples were obtained (n = 281, no correlation could be detected between CSF or plasma concentrations and pain relief. The plasma morphine levels were too low in most cases to contribute to any considerable analgetic effect at minimum steady-state concentrations. Following pain analyses, 30 patients were found to have 1 type of pain when included in the study, while 17 patients had 2 and another patient had 3 different types of pain. Significant differences in analgetic responses were found between radiating pain and somatic pain alone or in combination with the other pain types (Fig. 3). Seventeen patients who suffered from

.

.$$

. .

Plasma

morphine

C,,

min

ng/ml

Fig. 1. Pain relief in relation to plasma minimum steady-state concentrations of morphine in cancer patients receiving epidural morphine. No correlation was found between the degree of pain relief and plasma morphine.

deep somatic pain obtained adequate and uniform reduction in pain scores with an improvement of 76 + 15%. In contrast, visceral pain (n = 9) was affected in a most varying way, ranging from -5 to 100% (49 * 43%). Four patients with only radiating pain reported worsened or poor pain relief (2 + 37%) ranging from -40 to 50%. In combined pain types, a significant degree of alleviation was obtained when somatic pain was one of the components. The mean daily doses of epidural morphine in the subgroups tended to be higher in the groups of patients who obtained the lowest degree of pain relief (Table III and Fig. 3). Moreover, there was a large interindividual variation in plasma and CSF morphine concentrations within the different subgroups (Table III) and, although measurements were not obtained from all patients, there were no significant differences in mean plasma and CSF concentrations between subgroups. The distribution of continuous and intermittent pain are shown in Fig. 4 together with the number of patients with signs of neuropathic lesions. In somatic and visceral pain, the relative distribution of intermittent pain tended to shift towards poor relief, while continuous pain tended to deviate towards higher degrees of relief. No case of pain due to pathological fracture was present during the course of the study, but 2 cases occurred after the study had been completed. Both cases were found to be unresponsive to spinal morphine therapy. Pain caused by intestinal obstruction (n = 3) was also unaffected by epidural morphine. In pain originating from other visceral structures, the outcome varied and no uniform conclusion could be drawn. Four patients with radiating pain had concurrent somatosensory and motor disturbances, and their neuropathic pain component was poorly influenced or was even caused to deteriorate by epidural morphine

Pain 4c

~20

I

36'40

46243

l5f6

36!47

Fig. 3. Pain relief in relation

to character

Visceral alone

0

fi

relief 20

11

Rddlating alone

visceral

8om.t pair

radiating

n=5

n-12

of pain (means + S.D.).

Shown are 30 patients with a single pain type and 17 patients with combined

= P < 0.05; n.s. = non-significant.

and was discontinued in favour of other methods of pain control. No plasma or CSF samples were obtained from these patients.

llnn 61-80

81-100

% pain

SOMATIC

neg-20

21-40

61-80

81-100

% pain m Continuous =

IntermIttent

- Number

Fig. 4. Distribution (open

columns)

of

of pain relief

and the relative

41-60

61-80 !4 pain

81-100 relief

RADIATING

10

41-60

21-40

relief

VISCERAL

neg-20

100

I1

n=4

I

0

80

pain

8om.t pain

41-60

I

n=9

as compared to conventional oral or systemic opioid administration. In 7 of the patients reported above, epidural morphine treatment was judged to be a failure

21-40

60

I

pain

pain. *

neg-20

I;

40

ncg-20

21-40

41-60

relief

61-80 %

pain

81-100 relief

pain pain patients

with

neuropathic

in cancer patients distribution

lesion

after epidural

of continuous

morphine

at steady state, tneg = negative).

and intermittent pain within different types of pain.

the subgroups.

Shown is the whole population

One specific patient

may exhibit

l-1

Discussion The contribution of plasma morphine to the analgetic response following epidural administration is debatable. Although shown to be low at minimum steady state in this study, the peak morphine concentrations seen immediately after an epidural injection may well influence the supraspinal analgetic mechanisms. In the present study, the CSF morphine concentrations following spinal administration did not show any correlation with the analgesia obtained. The high CSF concentrations tolerated by some patients without serious side effects can to some extent be attributed to cross tolerance due to high oral or systemic doses prior to spinal therapy. Such cross tolerance has recently been described regarding morphine’s respiratory depressant action, but it was not found to be valid for its analgetic effects [Zl]. However, these high concentrations in the present study did not result in improved analgesia. This suggests that the lack of improved analgetic activity by spinal morphine, compared to other routes of administration, which was seen in some pain states may be related to pharmacodynamic factors, presumably at the level of the dorsal horn in the spinal cord. hypothetically, however, the morphine metabolites morphine-3-glucuronide and morphine-6-glucuronide, which are formed in the liver and in the brain [291, may also influence the analgetic response to spinal morphine. It is interesting that morphine-3-glucuronide has recently been shown to antagonize the analgetic effect of morphine [25]. It should also be considered that other factors like age and cognition may influence the effect of opioid analgesics. Pain characterization, based on the criteria used in the present study, can in some areas only be tentative. For exampIe, it is difficult to establish the degree of nerve involvement in radiating cancer pain in clinical practice unless there is evidence of neuropathic lesions [1,15,22]. Such signs will develop gradually and in their absence the pain might also be considered as nociceptive of somatic or neurogenic origin [2,5,28]. Thus, in this study the radiating pain was neuropathic in 4 patients due to compression of the spinal cord (n = If, nerve roots (n = 11, brachial plexus (n = 1) and lumbar plexus (n = 1). Another common feature was pain on movement, radiating along the dorsal part of the legs, found in patients with bone metastasis in the lumbar spine and pelvis, mainly in prostatic cancer and in most cases combined with a continuous lumbo-sacral pain component. Some cases later developed neurological symptoms which made the neurogenic pain component obvious; however, other patients improved with adequate pain relief and no neurological signs developed. Distinction between visceral pain and the activation of somatic nerves close to the diseased organ may also he difficult [6]. However, in the group with visceral

pain, 2 patients with urgency of the urinary bladder due to advanced spread of tumour from the pelvic organs did not respond to epidural morphine. One patient was then treated with intrathecal neurolysis and became pain free. Pain from pancreatic carcinoma with local spread was successfully treated with epidural morphine in 1 case but not in another; the latter showed partial improvement on coeliac neurolytic blockade. The same difference in results was seen in 2 patients with carcinoma of the stomach. Finally, 2 patients with continuous pain from local peritoneal growth due to large bowel carcinoma did respond favourably. In the clinical evaluation of pain related to malignancy, it should also be noted that the intensity often follows an alternating course, and an initial pain evaluation may, therefore, have to be reconsidered at a later stage. Despite the need for diagnostic refinement, pain characterization in relation to analgetic response may be instrumental for 2 main purposes. First, it might provide better possibilities of predicting the outcome of spinal opiate therapy and, second, it might offer a conceptual bridge between experimental and clinical research. In the latter case, pain analyses should focus on what types of peripheral receptors are likely to be excited by painful stimuli. Tissues like periosteum, muscle, cutis, peritoneum and hollow visceral structures may convey their noxious messages through different afferent fibres [6,11,26] possibly using differential chemical codings [12,16,17] terminating at different laminae in the dorsal horn [14]. They are probably also subjected to a varying degree of modulation of the transmission of nerve impulses [8,32]. In man, it should be possible to ascribe a pain sensation to specific receptor areas related to more or less defined afferent fibers or to nerve structures in neuropathic pain and then to observe what effect certain spinal receptorspecific agents exert on such a defined pain transmission [lo]. Furthermore, human CSF determination of putative neurotransmitters, believed to be involved in the transmission or modulation of sensory impulses, has been related to various pain states [3,27]. In an ongoing study investigating CSF neuropeptides, we have found indications of low concentrations of substance P and Met-enkephalin in visceral pain as compared to somatic pain [23]. Summing up observations reported by others [4,24] as well as our own experiences, it appears that somatic pain from mucocutaneous ulcers and from fractures, visceral pain from distention of the gut as well as neuropathic pain respond poorly to modulation by spinal morphine. Other types of pain exhibit various responsiveness to morphine analgesia, and this may be a matter of noxious stimulation intensity or of pharmacokinetic differences 171.Finally, continuous pain originating from deep somatic tissues can be extensively

s

inhibited or abolished by opioid administration at the spinal level. It remains to be shown how these findings can be fitted into a physiological or pathophysiological model like the A& and C-fibre concept and how the transmitter coding is arranged and modulated at the central primary afferent terminals. Despite the extensive knowledge of pain transmission derived from experimental studies, many clinical findings are as yet unexplained. This ought to stimulate further research in order to improve pain treatment in malignancy as well as to correlate clinical observations to cxperimental results.

myclinated tivr

prima0

Jaeckle.

K.. Young,

lumhosacral primary

affcrent

S.E.. Pain mechanisms 1135-l

in lumbar

radiculopathy.

Anesth.

7 Almay. B., Diagnosis of chronic pain syndromes, Nord. Psykiatr. Tidskr..

43, Suppl. 20 (1989)

3 Almay.

B., Johansson,

Terenius,

L., Substance

syndromes,

4 Am&, in

F.. Von

treatment

of

Stand..

29 (1985)

32-36.

5 Asbury.

AK.

In: P. Wall

(‘hurchill

Livingstone.

M.R.

Bond (Eds.),

Amsterdam,

7 Cousins.

pain

effects of epidural pain.

Acta

morphine

Anaesthesiol.

H.L..

Pain

Neurology, pain.

due to peripheral

34 (1984)

1:h

World

G.F.

Gebhart

and

Congress on Pain. Elsevier.

1988, pp. 216-226.

M.,

Cherry,

D.

and Gourlay.

pain: use of spinal opioids. (Eds.1, Neural

Blockade.

G..

Acute

and chronic

In: M. Cousins and P. Bridenbaugh

Lippincott.

Philadelphia.

PA.

1988, pp.

x

Duhner.

R.. Specialization

crimination, Fields.

R.

Research

in nociceptive

sensory modulation Dubner

and

and Therapy,

pathways: sensory dis-

and neural

F. Cervero

connectivity.

(Eds.),

In: 11.

Advances

Vol. 9. Raven Press. New York,

in Pain 1985. pp.

P.. Determination

glass capillary 10 Eisenach,

of opiates

gas chromatography

tion. J. Chromatogr.,

206 (1981)

J.. Rauck,

ology. 7 I

( 1989)

Fitzgerald, fihres.

17 Fumes\.

R., Buzzanell.

York.

samples by

capture

detec-

C. and Lysak,

S., Epidural

cancer pain: phase I. Anesthesi-

and R. Melzack Edinburgh,

J.. Costa, M..

Morris.

and the chemical

of the lnt. Union

afferent

Textbook

of Pain.

(Eds.).

J. and Gibbins,

I.. Novel neuro-

coding of neurons.

In: Proc. 30th

of Physiol. Sci.. Plenum

Press. New

concentration

rats. Acta Pharmacol. I4 Hunt.

Toxicol.,

S. and Rossi. J., Peptide

after intrathecal 58 (lY86)

administration

to in

243-248.

and non-peptide

containing

un-

transmitter\

at

(Eds.). Advance5 in Pain

chemist0

Edinburgh,

lYS5. pp.

equianalgesic

I :1.

ot primal-y afferent

(Eds.).

Textbook

dose ot intramuscular/oral

In: K. Foley and C‘. lnturriai and Therapy.

Vol.

Hirota.

N.,

Sate,

Y.,

Hino.

Y..

Satoh.


related

intormation

M.,

Inturrisi.

and plasma

Thcr..

in patient9

R., Galer,

s21 I.

Sj(ibery,

Influence

Smith.

M.

antagonist 26 Srinivasa.

J.. Ekman,

R. and Hedner,

S., Hultman. pain.

Acta Anaesthesiol. T.,

statin-like

R. and Campbell,

2X V&t.

Stand.,

from

the

in press. a potent

J., Peripheral

mccha-

68 (19Xx) S71-5YO.

G., Spielman.

F. and Nemeroff,

fluid concentrations

in chronic

pain

of somato-

patients.

Pain.

33

IhY- 172. (1.. Nociceptive

Pain. 3Y (19x9) hrain

R., Bisette.

in cerehrospinal

immunoreactivity

morphine

I. Results

Pain, Suppl. 5 (1990) SIX&

nisms of somatic pain. Anesthesiology, B.. France.

5

E.. Llndel,

Morphine-3-glucuronide

analgesia.

with

Pain. Suppl.

I.. Long term intrathecal cancer

T.. Suh-

in patients

therapy,

L.. Einarsson.

and Crnmond.

C‘.. Alterations

in cancel

M. and I-lnkcl. J..

spinal fluid content

in ‘refractor)’

Meyer.

morphine

O., Freilich.

of spinal opioid

of morphine R..

fluid

pain. (‘Iin.

U. and Porges. S.. (‘ro\s

and epidural

P. and Curelaru.

and hupivacain

cancer

neoplasm, Cancer, 64 (1989) 2207-221.;.

M., Appelgrcn.

FI.. Nite\cu.

with chronic

B., Salamon.

of epidural

H.. Lindqvist.

(IY90)

Li. (‘.I I. and

ccrebrospinal

I81 - 1X7.

and {omatostatin pain.

I’..

opiates:

11.. ‘Ter fiorst,

systemic

Pain, 39 (lYX9)

cancer

and

3X (1985) 631-041.

hetween

,tancr-P

R., Grabinsky.

and intrathecal

B., Scrnaker.

Identification

M.

transmit

to pain in the spinal dorsal horn.

c‘.. Kaiko.

profiles

Pharmacol.

patients.

Pres\.

(1985)7~14-20~

K., Epidural

tolerance

(Eds.).

7, Raven

that substance P and somatostatin

Max.

of Pain.

198X. pp. X2--YY.

nerve pain and neuropathic

pain (Letter).

243-243.

A..

metabolism

Winblad,

B.. Bixo.

of morphine

and

M.

and

Kane.

naloxone.

A..

Pain,

lluman

3.5 (19X8)

111~177. morphine

for intractable

pain secondary to

cancer of pelvic organs, Pain. ?I (19x5) 99-102. 31 Yaksh,

T., Spinal opiate

of action. Pain, 11 (1981)

analgesia: characteristic5

nociceptive

and principles

293-346.

37 Yaksh. T. and Stevens. c‘.. Properties

IYX7, pp. 143-165.

I3 Gustavshon, L. and Post, C., The degree of analgesia correlates spinal morphine

and clew exitston

and R. Melzack

versus

30 Wang. J.. Intrathecal

19X8, pp. 46-62.

history 01

19X6. pp. X7-YJ. Y.,

7Y Wahl\triim. of primary

natural

35 (1985) X- 15.

I‘akagi. H.. Evidence

(10X8)

647-h52.

Livingstone,

transmitters (‘ongrc\\

with electron

M., The course and termination

In: P. Wall

(‘hurchill

in biological

10%1 16.

clonidin analgesia for intractable II

York.

27 Urban.

111-137. Y Edlund.

K., The

and F. (‘ervero

first series of 52 patients,

YS5m 1ll30.

( IVX~~

B. .?OX

in the dorsal horn ot the spinal cord

in Pain Research

Samuelsson.

15X7-1590.

In: R. Dubner.

Proc. Vth

nerve

Lc~ntl

Vol. Y. Raven Press. New York.

Kaiko. R., Commentary:

Portenoy,

and Fields,

F., Visceral

with chronic

J.. Fa\t

\ynapsec

neurons.

Pteifer.

cancer-related

damage: a hypothesis, h Cervero,

P. and

3-9.

B., Differential

the

L., Le Greves,

P in CSF of patients

Pain, 33 (1988)

S. and Arm?.

Knorring,

and Foley.

Jessel. T. and Dodd. J.. Functional

Foley.

IS-23.

prcjcch\mg 01 noclc~p~

Sot

3 I --30.

Brain Res., 325

137.

Roy.

in cancer. Neurology.

and Therapy.

Kuraihhi.

h7 (1988)

D.

In: }I. Fields. R. Dubner Research

Ncu

Analg.,

the parallel

Trans.

plexopathy

Jcsael. T. and Craig,

,\dvances

I Abram.

Phil.

2X3-289.

morphine,

References

afterent\:

information.

of the modulation

transmission by receptor-selective

01 spinal

agents. In: R. Dub-

ner. G.F. Gebhart

and M. Bond (Eds.). Proc. Vth World C’ongreah

,m Pain. Elsevier,

Amsterdam,

IYXX, pp. 417&t35.