Ann R Coll Surg Engl 1995; 77: 202-209

Has postoperative pain been eradicated?

Sabaratnam Sabanathan FRCSConsultant Thoracic Surgeon

Department of Thoracic Surgery, Bradford Royal Infirmary, Bradford

Key words: Thoracotomy; Postoperative pain; Analgesia; Pre-emptive analgesia; Continuous extrapleural intercostal nerveblock; NSAID; Balanced analgesia

Recent evidence suggests that surgical trauma inducesa process of central nervous system sensitisation thatcontributes to and enhances postoperative pain. Thesechanges are also thought to be the underlying cause ofmuch chronic pain. Central sensitisation is generatednot only during surgery, but also postoperatively as aresult of the inflammatory response to the damagedtissue. This knowledge provides a rational basis forpro-active, pre-emptive and postoperative analgesicstrategies to reduce the neuronal barrage associatedwith tissue damage. Reduction or elimination ofpostoperative pain is therefore possible.We advocate the use of continuous extrapleural

intercostal nerve block for postoperative analgesia inpatients undergoing thoracotomy. When this is begunpre-emptively (by precutaneous, pre-incisional para-vertebral block) it is combined with an opiate and anon-steroidal anti-inflammatory drug premedication.In a randomised study of 56 patients, pain scores of

less than 0.5 cm on a 10 cm scale were produced,postoperative lung function was preserved andglucose and cortisol responses were significantlyunchanged from preoperative values. Evidence thateffective perioperative analgesia reduces the incidenceof chronic post-thoracotomy chest wall pain wasfound in a retrospective study of 1000 consecutivethoracotomies. The endpoints of a zero pain score,complete preservation of preoperative lung functionand prevention of the stress response to trauma arecurrently achievable and should be provided forvirtually all patients undergoing chest surgery.Pre-empting pain must be the goal for all those

involved in the postoperative care of patients.

Based on a Hunterian Lecture delivered at Bradford RoyalInfirmary on 6 October 1994

Correspondence to: Mr S Sabanathan FRCS, ConsultantThoracic Surgeon, Bradford Royal Infirmary, DuckworthLane, Bradford BD9 6RJ

The cruel and callous disgrace of inadequate attention bythe medical profession towards acute postoperative painhas been highlighted by lay people and doctors alike whohave found themselves at the receiving end of a surgeon'sscalpel. Despite lurid descriptions, editorials and surveys,progress in acute pain management has been painfullyslow, only recently attention being paid by governmentagencies (1) and professional organisations (2). Againstthis apparent stagnation must be set great advances in theunderstanding of underlying pathophysiological mechan-isms of pain.Major alterations in ventilatory mechanics and pul-

monary gas exchange occur in all patients after anaesthesiaand thoracotomy because of a decrease in the functionalresidual capacity (FRC) with minimal changes in theclosing volume. Shallow monotonous breathing withoutperiodic maximal inflation occurs as a consequence of thepostoperative pain and these changes are most severe inthe elderly, the obese, smokers and in those with pre-existing cardiopulmonary disease (3). These factors leadto airway closure during tidal breathing producinghypoxaemia as a result of venous admixture fromregional ventilation perfusion inequality and shunting ofpulmonary capillary blood through closed alveoli (3).Mucous plugs and infection develop secondarily toalveolar collapse and airway closure.

Postoperative hypoxaemia is associated with complica-tions such as myocardial insufficiency and infarction,pulmonary complications, cerebrovascular accidents,thromboembolism, delirium, delayed wound healing andprolonged convalescence with fatigue and inability towork (4-6). My work and that of others has showneffective postoperative analgesia will improve and evenreverse the adverse effects of surgery on the pulmonarymechanics (3,5-8). Such analgesia may reduce the nutri-tional and immunological consequences of the endocrine-metabolic response to surgical trauma (7,8).

Recent emphasis on a multimodal approach (9) to thetreatment of postoperative pain has come about by betterunderstanding of its pathophysiological mechanisms.

Has postoperative pain been eradicated? 203

Peripheral tissue injury can lead to hyperexcitability andneuronal plasticity in the spinal cord dorsal horn, that isresponsible for the maintenance of postoperative pain (9-12). Wall (12) has pointed out that these findings providea rational basis for pre-emptive measures which mightreduce the peripheral neuronal barrage associated withtissue damage and thus reduce or eliminate postoperativepain.Although recent clinical papers suggest the effective-

ness of pre-emptive analgesia (13-16), this has not so farproved to be of major clinical importance (15,17-19), asthe inflammatory reaction to tissue damage duringoperation provides a source of sensory signals that couldinduce central sensitisation, even it it had been preventedduring operation (19). It follows that the postoperativeanalgesic regimen should be continued until these sensorysignals have subsided as a result of normal wound healing.Furthermore, a truly effective analgesic regimen, bypreventing the prolonged sequelae of central sensitisa-tion, ought to be able to reduce the incidence and severityof chronic post-thoracotomy neuralgia, responsible forprolonged pain in up to 60% of patients in one survey(20), by preventing or reducing the development of any'memory' of the pain stimulus in the central nervoussystem (CNS). This concept has never been testedclinically.We undertook a review of the effects of a continuous

extrapleural intercostal nerve block balanced analgesicregimen on 1000 consecutive patients who underwent aposterolateral thoracotomy at the Bradford Royal In-firmary.We have previously demonstrated the effectiveness of a

continuous extrapleural intercostal nerve block for post-thoracotomy analgesia (5,21). As the paravertebral spaceis catheterised just before chest wall closure this techniqueis unable to provide intraoperative analgesia. In order toextend this method of pain relief into the intraoperativeperiod we commenced the block with pre-incisionalpercutaneous paravertebral injection of local anaesthetic.In a randomised study we assessed the efficacy onpostoperative pain and pulmonary function of pre-emptive analgesia in conjunction with postoperativeextrapleural continuous intercostal nerve block and rectaldiclofenac compared with on-demand opiates.

Postoperative acute pain studyStudy design

All patients undergoing elective thoracotomy who wereable to give informed consent were admitted to the trial.The study was approved by the hospital ethics committee.Before operation, patients were instructed in the use of ahand-held spirometer and a linear visual analogue scalefor pain assessment. Those who were treated with non-steroidal anti-inflammatory drugs (NSAID) and opiatesbefore operation, and those who had an additional incisionor procedure were excluded. Also excluded were patientswith a known allergic reaction to anti-inflammatory drugs,a history of peptic ulceration, impaired renal function and

Table I. Pre-emptive analgesia used according to groups

Premedication

Groups Pre-incisional block Opiate NSAID

1 Yes Yes Yes2 Yes Yes3 Yes Yes4 Yes5 Yes Yes6 Yes7 Yes8

those with spontaneous bleeding tendencies. All patientshad standard general anaesthesia using a double lumenendobronchial tube and single lung ventilation.The patients were allocated by random permutated

blocks to one of eight groups (Table I). Premedicationwas administered 1 h before the estimated time ofsurgery. The non-steroidal anti-inflammatory premedica-tion was diclofenac sodium (100 mg) given in the form ofa suppository (Groups 1, 2, 5, 6). The opiate premedica-tion used was morphine sulphate 10 mg intramuscularly(Groups 1, 2, 3, 4). Pre-incisional paravertebral blockade(Groups 1, 3, 5, 7) using 10 ml of 0.5% bupivacaine wasperformed at the proposed level of incision with thepatient in the lateral position fully anaesthetised, 10 minbefore incision. An identical postoperative analgesicregimen was used in all patients throughout the studyperiod of 48 h: continuous extrapleural intercostal nerveblock, 12 hourly diclofenac suppositories (100 mg) andon-demand opiates. No attempt was made to withholdrequested analgesia. The technique of insertion of theextrapleural cannulas has been described previously(21,22). A bolus injection of 10-20 ml (10 ml to thosewho had pre-incisional block and 20 ml to those who didnot) of 0.5% bupivacaine was first given. This wasfollowed by a continuous infusion of the same solutionat a rate of 0.1 ml/kg body weight per hour for 5 days (5).

Postoperative assessment

Subjective pain relief

Postoperative pain was assessed by the patients on a10 cm linear visual analogue scale. Pain scores wererecorded at 4 h intervals for the first 48 h after operation.The requirement for additional analgesia was recorded foreach patient.

Objective respiratory function

Before operation and 12 hourly during the first 48 h afteroperation, peak expiratory flow rate (PEFR), forcedexpiratory volume in 1 s (FEV1) and forced vital capacity(FVC) were measured with the patient in a sitting positionusing the spirometer. The best of three measurements wasrecorded. PEFR, FEVI, and FVC were expressed as apercentage of the preoperative values.

204 S Sabanathan

Hormonal response (group 1)

Blood samples for analysis of cortisol and glucose weretaken before induction of anaesthesia, at surgical incision,and at 1, 4, 8, and 24 h after surgical incision. Glucose wasanalysed using a Yellow Springs Glucose Analyser(Clandon Scientific Ltd, Sheffield, UK) and cortisol wasestimated by radioimmunoassay (ADL, Amersham,Bucks, UK).

Statistics

The data for each criterion were analysed statistically bythe methodology of analysis of variance. Pain data weredivided into two 24 h periods and then analysed as above.Pulmonary function data fitted into a general linear modelby multiple regression for statistical analysis. Changes inplasma glucose and cortisol were analysed by two-wayanalysis of variance. A P value <0.05 was taken asstatistically significant.

Results

There was no significant difference between the groupsregarding age distribution, surgical procedure, sex ratio,smoking habits, body weight, height, and duration of theoperation. There were no complications related either tothe infusion or the use of diclofenac sodium.

Pain relief (Fig. 1)

The balanced analgesia group (Group 1) had consistentlysignificantly better pain relief (P<0.05) compared withthe other groups. Only one patient required a dose ofopiate postoperatively (P< 0.05). Those patients who had

pre-incisional block had better pain relief (P<0.001) andsignificantly less postoperative opiate requirement in theimmediate postoperative period (P< 0.02). Premedicationwith opiates and NSAIDs provided better pain relief onlyas part of the balanced analgesia regimen. None of thepatients in this study required additional opiate analgesiaafter the first 24 h.

Postoperative pulmonary function (Figs 2-4)

Throughout the study the balanced analgesia group(Group 1) had significantly better pulmonary functionthan any other group (FEV1, P=0.02; FVC, P=0.04).These patients uniquely maintained their preoperativepulmonary functions postoperatively and often exceededthe baseline values. Those groups of patients who hadpre-incisional block had significantly better pulmonaryfunction at 12 h than those without pre-incisional block(FVC, FEV1, P<0.001; PEFR, P<0.02). Opiate pre-medication and NSAID premedication alone or incombination had no advantage unless as part of abalanced analgesic regimen. Maximal reduction ineffort-dependent lung volumes occurred at 12 h in allpatients. No pulmonary complications occurred duringthe 48 h study period.

Plasma levels of cortisol and glucose

There was no significant change in the plasma levels ofglucose or cortisol throughout the study period. Themean (SEM) preoperative levels of plasma cortisol andglucose were 452 (41) mmol/l, 5.2 (0.6) mmol/l comparedwith postoperative maximum levels of 474 (45) mmol/l,5.91 (0.6), respectively.

30 -

25 -

C:n*=cB

20-

15-

10-

5-

u

\ x #I "k #

Groups1

***** 2a a ... 3oooo 4.... 5

----- 7------7------ 8

'I

4 8 12 16 20 24 28 32 36 40 44 48

Hours Post-Op

Figure 1. Mean pain scores during the first 48 h after operation. The accumulatedpain scores were significantly lower in the balanced analgesia group (Group 1).

Has postoperative pain been eradicated?

Twelve Hourly F,VEC. Recrding9

At 1::. . E

Figure 2.'Box and whisker' plot of forced vital capacity (FVC) values as a percentageof preoperative values. Balanced analgesia group (Group 1) maintained theirpreoperative pulmnonary function postoperatively, uniquely exceeding the baselinevalues.

tr;Xi^;.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~R :AJf1I-:

140%~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1

tR:#01,pFt0m1f::N~~~~~~~~~~~~~~~~~51i,i, 20% 0 06 0

[80%!

40%-

NTW

Figure 3. 'Box and whisker' plot of forced expiratory volume in 1 s (FEV1) values as

a percentage of preoperative values.

205

.:.Ii

-.i

i..f

A:.lAi

206 S Sabanathan

Twev Hourly P.EPF.R. Recordings

12~~~ 24 36 46

4~ IkPu~fts00.raonFigure 4. 'Box and whisker' plot of peak expiratory flow rate (PEFIpercentage of preoperative values.

Post-thoracotomy neuralgia study (PTN)

Patients and methods

A series of 1000 consecutive thoracotomies were reviewedretrospectively to assess the incidence of post-thoraco-tomy neuralgia and to identify and evaluate possible riskfactors. Post-thoracotomy neuralgia was defined as chestwall pain unrelated to recurrent or persistent tumour orinfection, but persisting beyond or recurring 2 months ormore after thoracotomy. Insufficient information elimi-nated 117 patients from the study. For the remaining 883we analysed age, gender, ethnicity, diagnosis, type andside of surgical procedure, rib removal, reoperation,insertion of chest drain, adjuvant therapy and periopera-tive analgesia.

Statistics

All data were entered into a Smartwares statisticalsoftware database. Most analyses of the significance ofthe difference in various characteristics were analysed bymeans of the x2 test and contingency tables. Unpaired ttests were used for comparison of means. A P value < 0.05was considered significant.

Results (Fig. 5)

PTN was much more common in Asians and Afro-Caribbeans (P= 0.05). Age and sex were irrelevant. Therewas a strong association with benign oesophageal disease,

Paliant GroupiUi

1_23

R~) values as a

less so with malignant oesophageal disease or benign lungdisease. The least association was after operations formalignant lung disease. Postoperative adjuvant radio-therapy increased the likelihood of PTN (P=0.01). Ribremoval at operation reduced it (P=0.02). The relation-ship of acute pain management to the development ofPTN was also examined. Cryoprobe neurolysis ofintercostal nerves increased the incidence of pain(P= 0.01), whereas continuous extrapleural intercostalnerve block with bupivacaine, especially when combinedwith NSAIDs, reduced the incidence (P= 0.01). NSAIDsreduced chronic pain when used in simple combinationwith opiates (P= 0.02).

400

12panat I =momI

0 300 .X feal

~0

laden 0.. Mang. 0Q. DnM Lung Mal. Long%with.pin 41.2 11.1 14.8 20.4 -13.4% r.bnud 30.4 4.4 10.2 13.2 7.5

tology and Site

Figure 5. Chronic pain related to histology and site.

..I

Has postoperative pain been eradicated? 207

Discussion

Postoperative pain control is finally receiving the attentionthat it deserves. It is particularly important in thoracicsurgery in view of the damaging effects of chest wall painand the high-risk population coming to thoracotomy. Ascomplete pain control cannot be achieved with a singleagent or technique without significant serious side-effectsor the need for surveillance, a balanced analgesic regimenis more appropriate (9). 'Pain prophylaxis' is by opiatepremedication together with a NSAID followed by pre-incisional afferent blockade, the most effective and logicalpostoperative technique being continuous paravertebralanalgesia. Regular NSAIDs postoperatively are anessential part of pain management.

After thoracotomy, patients are unable or unwilling toinspire deeply. The reduced inspiratory capacity limitsthe ability to cough effectively. Impaired lung volumes,particularly FRC, ventilatory mechanics and oxygenationare the consequences of this shallow monotonous breath-ing without periodic maximal inflation (3). Sighingreverses these changes (3). The accepted view of theliterature is that these changes in the pulmonarymechanics can be minimised but not prevented (3). Ourfinding of preserved preoperative lung function post-operatively in Group 1 patients in this study disputes thiscontention. These patients had a balanced pre-emptiveand postoperative analgesic regimen. The effectiveness oftotal afferent blockade was confirmed by the absence ofhormonal and metabolic response to surgery as measuredby plasma cortisol and glucose levels. The data presentedhere suggests that postoperative mechanical abnormalitiesof pulmonary function are not an inevitable consequenceof thoracotomy. Effective postoperative analgesia willimprove and even reverse the adverse effects of surgery onthe pulmonary mechanics.The peripheral neuronal barrage resulting from tissue

damage can be blocked with pre-incisional regionalanaesthetic techniques (13,14,23,24). Afferent blockademust be continued during and for several days afteroperation to ensure that neuronal plasticity is preventedand not simply delayed (14). Evidence suggests that theperipheral afferent block is more effective than the centralblock in preventing nociceptive impulses from enteringthe central nervous system (14,25,26). Paravertebralblock would also reduce neurogenic inflammation oftraumatised tissues that is dependent on efferentfunctions of peripheral nerves (24).The mechanism of action of the continuous intercostal

nerve block is via the paravertebral spread of a localanaesthetic agent (27). The results presented here show thata pre-incisional afferent block significantly decreases theintensity of the postoperative pain and more importantlysignificantly reduces the early loss of postoperativepulmonary function. Postoperative opiate consumptionwas also less in patients who had pre-incisional paraverte-bral block. These findings provide further support to thepathophysiological mechanism of pain proposed by Woolf(10) and Woolf and Thompson (11).

Small doses of opiates given before incision prevent

central sensitisation by diminishing the sustained hyper-excitation of the central nervous system caused byintraoperative painful stimuli (12,25,28), whereas sup-pression of established neuronal hyperexcitability re-quires very large doses (12). In our study, opiatepremedication was found to minimise the loss ofpostoperative lung function and was effective in reducingthe immediate postoperative pain only when it wascombined with a NSAID premedication.

Suppression of prostaglandin synthesis is probably thebasis of the analgesic mechanism of action of NSAIDs(29). Prostaglandins alone are weak stimulators of afferentpain-transmitting nerves, but they facilitate the response toother stimuli (29). Total afferent blockade is not likely tobe achieved by prostaglandin synthesis inhibition, andthus NSAIDs have a synergistic effect with other analge-sics (29). To prevent prostaglandin-mediated sensitisationof nociceptors during surgery it seems logical to adminis-ter NSAIDs prophylactically and then to maintain thempostoperatively (30). In the literature, pretreatmentwith NSAIDs has been found to provide variably effectivepostoperative analgesia. We were unable to demonstrateany advantage of NSAID premedication in respect ofpostoperative pain and pulmonary function unless it waspart of the balanced analgesic regimen.The immediate postoperative period is a potentially

high-risk period for the occurrence of hypoxaemia whichmay be persistent or episodic. Persistent hypoxaemia isdue to ongoing abnormalities in pulmonary mechanicsinduced by operation, while the episodic hypoxaemiaoccurs exclusively during sleep and is often opiateinduced in susceptible individuals (31). Because of thedeleterious effect of hypoxaemia on cardiac, neurologicaland psychiatric status of the patients, systemic opiatesshould no longer be used as the mainstay of treatment inthe postoperative period (31). Furthermore, a significantproportion of postoperative pain is caused by the painarising from chest drains transmitted via the sympatheticsfrom the pleura and A6 and C fibres in the intercostalnerve. As opiates are primarily just C fibre inhibitors, paindue to the chest drains would not be expected to beprevented by reliance upon this method of analgesia (32).Regional analgesia is logical and more effective.

Epidural techniques using either narcotic or localanaesthetic agents provide effective analgesia and im-prove respiratory mechanics (33), but are difficult toapply to large numbers of surgical patients because of thedifficulty in providing sufficient staff for the necessarymonitoring. These limitations would be acceptable if aprofound afferent block was being effected. Unfortu-nately, this is not the case for lumbar epidural blockadewith 26.9 ml of 0.5% bupivacaine, despite dermatomalanalgesia, results in suppression of somatosensory evokedpotentials in only 50% of patients (26). This is probablythe explanation for the failure of epidural techniques toprevent stress responses in anything more rostral thanlower abdominal surgery (9).Cryoprobe neurolysis of intercostal nerves does not

include the posterior primary rami and sympathetic fibresand so does not provide complete postoperative pain relief

208 S Sabanathan

and may cause prolonged anaesthesia or dysaesthesia(2,3,34). As repeated multiple intercostal blocks per-formed postoperatively are painful and time-consuming,continuous infusion techniques are more appropriate.

Paravertebral infusion of a local anaesthetic agenteffectively blocks not only the intercostal nerves but alsothe thoracic sympathetic chain and the posterior primaryrami which mediate pain from straining of the ligamentsof the costovertebral and costotransverse joints and theposterior spinal muscles (3). Unilateral blockade of thesympathetic nervous system is important as it is involvedin each stage of the nociceptive response to acute injury(35). Paravertebral block was thought to be comparable toepidural block in respect of pain relief and restoration ofpulmonary mechanics (36). We have demonstrated in thisstudy that as part of balanced analgesia beginning pre-emptively it is superior. The side-effects of hypotensionafter thoracic epidural instillation of local anaesthetic, ofurinary retention and the fear of respiratory depressionwith neuraxial opiates are not a problem (36).Although there have been some surveys of the problem

of post-thoracotomy neuralgia (37), data on this conditionremain scant and anecdotal owing to lack of defineddiagnostic and therapeutic criteria. Our finding thatalmost one-quarter of patients (22.3%) were stillcomplaining of pain 2 months after thoracotomy andthat 13.9% were still in pain at 12 months, comparesfavourably with other surveys of this condition.An increase in incidence of post-thoracotomy neuralgia

with cryotherapy has been noted in this study, which issimilar to several other studies reported in the literature(2,3,34), and hence only a few enthusiasts continue toadvocate its use (2).

In this survey, a large group of patients receivedcontinuous extrapleural intercostal nerve blocks withbupivacaine. They were found to have significantlyreduced chronic pain (P<0.01); moreover the additionof a NSAID produced a further reduction in the incidenceof post-thoracotomy neuralgia (P< 0.02). The implicationof this is that the afferent barrage on the central nervoussystem has been prevented and therefore stimulus fordorsal horn hyperexcitability and hence chronic painavoided (20,38).The first clear clinical evidence that management of

acute pain has a major effect on the development ofchronic pain is provided in our findings. The highlyeffective analgesic regimen of continuous extrapleuralintercostal nerve block and NSAIDs prevented not justthe acute post-thoracotomy neuralgia but to a significantextent led to the avoidance of chronic pain.

I believe that continuous extrapleural intercostal nerveblock used pre-emptively (by percutaneous paravertebralblock) and used as part of a balanced analgesic regimenprofoundly affects the outcome after thoracotomy. Acutepain is prevented, pulmonary function is preserved andchronic pain is prevented. We have the knowledge and theanalgesic agents necessary to prevent postoperative pain.What we need now is logical, rational, and universalapplication of this information. The principle of aneffective afferent blockade as part of a balanced analgesic

regimen beginning preoperatively and maintained untilwound healing is established deserves to be considered forall surgical procedures.

References

1 Agency for Health Care Policy and Research. Acute PainManagement and Trauma: Operative or Medical Procedures.Publication 92-0032. US Department of Health and HumanServices Rockville (MD), 1992.

2 Commission on the Provision of Surgical Services. Report ofthe Working Party on Pain after Surgery. London: TheRoyal College of Surgeons of England and The College ofAnaesthetists, 1990.

3 Sabanathan S, Eng J, Mearns AJ. Alterations in respiratorymechanics following thoracotomy. J R Coll Surg Edinb 1990;35: 144-50.

4 Knill RL, Moote CA, Skinner MI, Rose EA. Anaesthesiawith abdominal surgery leads to intense REM sleep duringthe first postoperative week. Anesthesiology 1990; 73: 52-61.

5 Sabanathan S, Mearns AJ, Bickford Smith PJ et al. Efficacyof continuous extrapleural intercostal nerve block on post-thoracotomy pain and pulmonary mechanics. BrJ Surg 1990;77: 221-5.

6 Berrisford RG, Sabanathan S, Mearns AJ, Bickford SmithPJ. Pulmonary complications after lung resection: the effectof continuous extrapleural intercostal nerve block. Eur JCardiothorac Surg 1990; 4: 407-11.

7 Kehlet H. The importance of postoperative pain relief. ActaAnaesthesiol Scand 1993; 37(S100): 122-3.

8 Brown DL, Mackey DC. Management of postoperative pain:influence of anaesthetic and analgesic choice. Mayo Clin Proc1993; 68: 768-77.

9 Kehlet H. Surgical stress: the role of pain and analgesia. Br JAnaesth 1989; 63: 189-95.

10 Woolf CJ. Evidence for a central component of post-injurypain hypersensitivity. Nature 1987; 306: 686-8.

11 Woolf CJ, ThompsonWN. The induction and maintenance ofcentral sensitization is dependent on N-methyl-D-asparticacid receptor activation; implications for the treatment ofpost-injury pain hypersensitivity states. Pain 1991; 44: 293-9.

12 Wall PD. The prevention of postoperative pain. Pain 1988;33: 289-90.

13 Tverskoy M, Cozacov C, Ayache M, Bradley EL, Kissin I.Postoperative pain after inguinal herniorrhaphy withdifferent types of anaesthesia. Anesth Analg 1990; 70: 29-35.

14 Ejlersen E, Anderson HB, Eliasen K, Mogensen T. Acomparison between preincisional and postincisional lido-caine infiltration and postoperative pain. Anesth Analg 1992;74: 495-8.

15 Richmond CE, Bromley LM, Woolf CJ. Preoperativemorphine pre-empts postoperative pain. Lancet 1993; 342:73-5.

16 Kiss IE, Kilian M. Does opiate premedication influencepostoperative analgesia?: a prospective study. Pain 1992; 48:157-8.

17 Kehlet H, Dahl JB. The value of 'Multimodal' or 'Balancedanalgesia' in postoperative pain treatment. Anesth Analg1993; 77: 104I-56.

18 McQuay HJ. Pre-emptive analgesia. Br J Anaesth 1992; 69:1-3.

19 Woolf CJ, Chong M. Preemptive analgesia-treating post-operative pain by preventing the establishment of centralsensitization. Anesth Analg 1993; 77: 362-79.

Has postoperative pain been eradicated? 209

20 Dubner R. Pain and hyperalgesia following tissue injury: newmechanisms and new treatments. Pain 1991; 44: 213-14.

21 Sabanathan S, Bickford Smith PJ, Pradhan GN, Hashimi H,Eng J, Mearns AJ. Continuous intercostal nerve block for painrelief after thoracotomy. Ann Thorac Surg 1988; 46: 425-6.

22 Berrisford RG, Sabanathan S. Direct access to theparavertebral space at thoracotomy. (Letter.) Ann ThoracSurg 1990; 49: 854.

23 Dahl JB, Rosenberg J, Dirkes WE, Mogensen T, Kehlet H.Prevention of postoperative pain by balanced analgesia. Br JAnaesth 1990; 64: 518-20.

24 Coderre TJ, Malzack R. Cutaneous hyperalgesia: contribu-tions of the peripheral and central nervous system to theincrease in pain sensitivity after injury. Brain Res 1987; 404:95-106.

25 McQuay HJ, Carroll D, Moore RA. Postoperative orthopae-dic pain-the effect of opiate premedication and localanaesthetic blocks. Pain 1988; 33: 291-5.

26 Lund C, Hansen OB, Kehlet H. Effect of surgery on sensorythreshold and somatosensory evoked potential after skinstimulation. Br J Anaesth 1990; 65: 173-6.

27 Eng J, Sabanathan S. Site of action of continuousextrapleural intercostal nerve block. Ann Thorac Surg 1991;51: 387-9.

28 Magora F, Ready LB. The problem of acute pain. It hurts.Pain Clin 1990; 3: 65-7.

29 Dahl JB, Kehlet H. Non-steroidal anti-inflammatory drugs:rationale for use in severe postoperative pain. Br J Anaesth1991; 66: 703-12.

30 Matila MAK, Ahlstrom-Bengs E, Pekkola P. Intravenousindomethacin or oxycodone in prevention of post-operativepain. Br Med J 1983; 297: 1026.

31 Jones JG, Sapsford DJ, Wheatley RG. Postoperativehypoxaemia: mechanisms and time course. Anaesthesia1990; 45: 566-73.

32 Sabanathan S, Richardson J, Mearns AJ. Management ofpain in thoracic surgery. Br J Hosp Med. 1993; 50: 114-20.

33 Shulman M, Sandler AN, Bradley JW, Young PS, Brebner J.Postthoracotomy pain and pulmonary function followingepidural and systemic morphine. Anesthsiology 1984; 61: 569-75.

34 Orr IA, Feenan DJM, Dundee JW. Improved pain reliefafter thoracotomy: use of cryoprobe and morphine infusion.Br Med J 1981; 283: 945-8.

35 Sabanathan S, Richardson J, Mearns AJ. Efficacy ofcontinuous intercostal bupivacaine for pain relief afterthoracotomy. Br J Anaesth 1993; 71: 463-4.

36 Richardson J, Sabanathan S, Eng J et al. Comparisonbetween continuous epidural morphine and extrapleuralbupivacaine for relief of post-thoracotomy pain. Ann ThoracSurg 1993; 55: 377-80.

37 Kalso E, Perttunen K, Kaasinen S. Pain after thoracicsurgery. Acta Anaesthesiol Scand 1992; 36: 96-100.

38 Coderre TJ, Katz J, Vaccarino AL, Melzack R. Contributionof central neural plasticity to pathological pain: review ofclinical and experimental evidence. Pain 1993; 53: 259-85.

Received 31 Azgust 1994