Postoperative analgesia reduces mortality and morbidity after esophagectomy

Postoperative Analgesia Reduces Mortality Morbidity after Esophagectomy

and

Siu Lun Tsui, MB, BS, FANZCA, Simon Law, MB, BChir, FRCS(Ed), Manson Fok, MB, BS, FRCS(Ed), J. Ronald Lo, MB, BS. MMed, FANZCA, FFICANZCA, Edward Ho, MB, BS, MRCP, FRCA, Joseph Yang, MD, FFARCSI, John Wong, MD (Hon), PhD, FRACS, FACS (Hon), Hong Kong

BACKGROUND: To study the influence of postoperative analgesia on morbidity and mortality after esophagectomy. METHODS: The outcomes of 578 patients who underwent one-stage resection between 1988 and 1995 were analyzed. Patients who received either epidural morphine, patient-controlled analgesia, or continuous intravenous morphine infusion supervised by an anesthesiology-based acute pain service (group APS, n = 299) were compared with those for whom conventional intramuscular meperidine injections were used (group CON, n = 279). RESULTS: For patients who underwent transthoracic esophagectomy, group APS (n = 228) had a lower incidence of pulmonary complications (13% versus 25%, P = 0.002), cardiovascular complications (21% versus 43%, P < O.OOl), and hospital mortality (8% versus 14%, P = 0.038) when compared with group CON (n = 189). No similar difference was demonstrated in patients who underwent esophagectomy without thoracotomy. The hospital stay (days) was shorter in group APS than in group CON for both transthoracic esophagectomy (22 5 20 versus 30 -+ 37, P = 0.005) and nontransthoracic esophagectomy patients (19 k 13 versus 25 k 21, P = 0.029). CONCLUSION: Adequate postoperative analgesia is associated with lower cardiopulmonary complications, lower mortality and reduced cost in patients undergoing transthoracic esophagectomy. Am J Surg. 1997;173:472-478. 0 1997 by Excerpta Medica, Inc.

E

sophageal cancer surgery is a major procedure performed in a group of high-risk patients and is associated with significant morbidity and mortality.’ Cardiopulmonary complications are the most common causes of perioperative morbidity and mortality.2-5 Adequate postoperative analgesia after esophagectomy, like after other From the Departments of Anesthesiology (SLT, JRL, EH, JY), and the department of Surgery (SL, MF, JW), University of Hong Kong, Queen Mary Hospital, Hong Kong. Requests for reprints should be addressed to S. L. Tsui, MB, Consultant Anaesthesiologist, Head, Pain Management Team, Department of Anaesthesiology, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong. Manuscript submitted February 2, 1996 and accepted in revised form July 1, 1996.

472

0 1997 by Excerpta All rights reserved.

Medica,

Inc.

major operations involving laparotomy or thoracotomy, may help reduce these complications and mortality.6-s Good pain relief allows for better preservation of pulmonary function, chest expansion, effective cough effort, and cooperation with chest physiotherapists.’ Also, adequate analgesia minimizes increase in sympathetic activity, which raises myocardial oxygen consumption and reduces perfusion to the left ventricle (through tachycardia and hypertension). Conventional intermittent intramuscular (IM) opioid injections often fail to provide satisfactory postoperative analgesia after major operations.i”~” Modern analgesic technics are more effective in relieving postoperative pain. These analgesic technics include epidural opioid administration,“-I5 patient-controlled analgesia ( l’CA),16-‘8 and continuous intravenous (IV) infusion of opioid. Delivery of these analgesic techniques should be supervised by an anesthesiology-based acute pain service (APS).‘9-23 The aim of this report is to evaluate the effect of AI’S supervised modern analgesic techniques on surgical outcome after esophagectomy.

METHODS Patient Selection Between 1 January 1986 and 31 July 1995, 1,007 patients with esophageal carcinoma were managed in the Department of Surgery, The University of Hong Kong at Queen Mary Hospital. A one-stage resection was performed in 578 patients. Prospectively collected data from these 578 patients were analyzed. The patients were allocated into two groups. group AI’S (n = 299) were patients under the care of the AI’S for their postoperative pain management. This service covered the period from 1989 to 1995. Group CON (n = 279) were patients given conventional intermittent IM meperidine injection for postoperative analgesia. These patients were managed in this manner from 1986 to 1990, before AI’S was fully established. Surgical Management Preoperative investigations included barium swallow, flexible esophagoscopy and biopsy, bronchoscopy, abdominal ultrasonography, cardiopulmonary assessment, and general biochemical and hematological evaluation. The choice of operation depended on the location and extent of tumor, the presence or absence of metastases, and the cardiopulmonary status of the patient. Cardiopulmonary risk assessment was based on history, symptoms and signs of chronic lung or heart diseases, exercise tolerance, chest radiograph, arterial blood gas assessment, and pulmonary function tests. A patient was considered to have moderate pulmonary risk if the forced expiratory volume at 1 minute (FEVi) was less 0002-961 O/97/$1 7.00 PII 50002-961 0(97)00014-7

POSTOPERATIVE TABLE

ANALGESIA

AFTER

ESOPHAGECTOMYITSUI

ET AL

I Standard

Acute for

Pain Service Analgesic Esophagectomy

Regimen

A.

Epidural morphine infusion First epidural bolus: 0.1 mg/kg, given after induction of general anesthesia. Infusion: morphine 0.05 mg/mL, initially at 0.1 ml/kg/h commencing postoperatively in intensive care unit, subsequent titration nursing staff, within range for adjustment: 0.05 to 0.15 ml/kg/h. Monitoring: hourly RR, SpOp; 4-hourly BP, HR, NRS and sedation score, side effects (nausea, vomiting, pruritus and dizziness). B. Patient-controlled analgesia (PCA) intravenous morphine Initial loading in recovery room, 1 to 1.5 mg every 5 minutes until the patient is comfortable. PCA setting: patient controlled bolus 1 to 1.5 mg, lock-out interval: 5 minutes, background infusion: 0 to 0.5 mg/h, 1 -hour maximum dose: 0.1 mg/kg. Monitoring: hourly RR, SpOz; 4-hourly BP, HR, NRS and sedation score, side effects. C. intravenous morphine infusion (for patients following postoperative mechanical ventilation) Initial intraoperative loading. Infusion: initially at 2 to 3 mg/h, subsequent titration by nursing staff within range for adjustment at 1 to 6 mg/h. Monitoring: hourly RR, SpO,; 4-hourly BP, HR, NRS and sedation score, side effects. RR = respretory rate/minute; BP = systemic numerical rating scale for pain.

b/o&

pressure;

HR = heart rate/minute;

than 70% with normal arterial blood gases, and high pulmonary risk if the FEVr was less than 50% or arterial blood gas showed hypoxemia or carbon dioxide retention. Moderate cardiac risk included a history of ischemic heart disease or cardiac failure, or evidence of arrhythmia on electrocardiogram. Patients were considered to be at high cardiac risk if there was a history of myocardial infarction or if medical treatment for cardiac failure was ongoing. For the majority of patients who had an intrathoracic squamous cancer, resection was performed via an abdomino-right thoracic approach (Lewis-Tanner operation). A three-phase esophagectomy, including a neck incision, was performed for lesions that extended above the level of the aortic arch. For tumors that involved the cardia, an esophagogastrectomy was carried out with or without a thoracotomy. In patients with increased pulmonary risk, a transhiatal esophagectomy was performed. For esophageal substitution, the stomach was most frequently used. Resections were considered curative when there was no tumor left after surgery. Resections were palliative when there was tumor infiltration beyond the esophagus into mediastinal organs, with gross lymph node metastases, and when there was actual or a high chance of residual tumor. All operations were performed by a small team of three surgeons with minor modification of techniques and surgical management protocol throughout the study period. Anesthetic Management During the preanesthetic visit, in addition to a thorough assessment of major organ functions, the attending anesthesiologist explained the details of postoperative pain control to the patient. For patients receiving epidural analgesia, an epidural catheter was inserted under aseptic technique before induction of general anesthesia, at the high lumbar or low thoracic spine (T8-L3) level. General anesthesia was then induced with IV thiopentone followed by muscle relaxation with neuromuscular blockers, endotracheal or endobronchial intubation, and mechanical ventilation. Anesthesia was maintained with nitrous oxide, volatile anesthetics, and IV opioid. One-lung ventilation was employed during the thoracotomy phase. Most patients were THE

AMERICAN

SpOz = oxygen

saturation

% on puke

ox/meter

monitoring;

by

NRS = verbral

routinely extubated immediately after the operation. Elective ventilation was employed in patients with poor cardiopulmonary status or when the surgical procedure was complicated. All patients, following esophagectomy, were managed in the intensive care unit (ICU) for monitoring and usually stayed there for 3 to 4 days. Vigorous chest physiotherapy with breathing and coughing exercises supported by incentive spirometry were employed. Clearance of sputum and secretions was aided by frequent flexible fiberoptic bronchoscopic suction. In patients with poor chest condition, tracheostomy was established either at the time of operation or when the need arose, for ventilatory support and sputum evacuation. Postoperative Pain Management All patients in group APS received either epidural or systemic analgesia according to a standardized regimen (Table I). The choice of analgesic technique depended on the feasibility of epidural catheter insertion, and the preference and expertise of the attending anesthesiologist. For epidural analgesia, the patient received the first epidural bolus of preservative-free morphine 0.1 mg/kg before skin incision. Subsequent epidural morphine infusion was commenced in the ICU. For patients receiving systemic analgesia, morphine was given either as continuous IV infusion through a syringe pump or as IV bolus injections through a PCA pump. This infusion was started intraoperatively and continued into the postoperative period. Pain at rest and during coughing was assessed every 4 hours using the O-to-10 verbal numerical rating scale (NRS).‘4 Zero was defined as no pain and 10 as the worst pain the patient could imagine. Patients under the care of AI’S were asked to rank their overall satisfaction on pain management as “good,” “fair,” or “unsatisfactory” upon termination of APS. All patients from 1 January 1986 onward, before the commencement of AI’S, belonged to group CON. They received conventional IM meperidine 1 to 1.5 mg/kg or morphine 0.1 to 0.2 mg/ kg at 4 to 6-hour intervals, or more frequently as requested. The NRS was not assessed in this group of patients as this was not routine before the implementation of APS. JOURNAL

OF

SURGERYa

VOLUME

173

JUNE

1997

473

TABLE II Demographic

Age mean Gender Male Female Preoperative Normal

Data,

Preoperative

Cardiopulmonary Operative Procedures

+ SD

pulmonary

CON = conventiona/

P

62.8

63.4

k 9.9

NS

256 43

230 49

NS

80% 18% 2%

75% 20% 5%

NS

93% 5% 2%

94% 6% 0%

NS

1% 6% 55% 21% 17%

1% 4% 48% 25% 22%

5% 13% 3% 75% 4%

4% 13% 6% 69% 8%

pain treatment;

2 9.3

(76%) (24%) + 49.1 -+ 997.4

Group APS (n = 299) and group CON (n = 279) were comparable with respect to age, gender, preoperative cardiopulmonary risk, level and staging of tumor, operation duration, and blood loss (Table II). More patients in group

MS = not sfafisficaiiy

OF SURGERY@

VOLUME

173

1 a9 90 203.8 971.0

(68%) (32%) k 47.3 t- 1051.8

NS

0.037 NS NS

significant.

APS underwent transthoracic resection (76% versus 68%, P = 0.037). The overall morbidity and mortality data are summarized in Table III. Group APS was associated with a significantly lower incidence of pulmonary complications, cardiac complications, necessity for tracheostomy, and hospital mortality. Further analysis showed that these differences apply only to the subgroup of patients who underwent transthoracic esophagectomy (Table IV). No difference was demonstrated in the group for patients without thoracotomy. However, hospital stay was significantly shorter in group APS than in group CON regardless of whether or not a thoracotomy was performed for esophageal resection (Table IV). The efficacy of analgesic techniques in 206 patients under the care of APS between 1992 and 1995 is summarized in Table V. Satisfactory analgesia was achieved in most patients, with median NRS around 0 to 1 at rest and 0 to 4 during cough. Comparison between the 165 patients who received epidural analgesia and 41 who received systemic analgesia (PCA in 23, IV morphine infusion in 18) showed that the former technique achieved better analgesia during cough on the day of operation and the first postoperative day. Respiratory depression, which was defined as respiratory rate less than 10 per minute or SpOz less than 90% for longer than 1 minute, occurred in 3 patients. Administra-

RESULTS

JOURNAL

and

Group CON (n = 279)

226 73 218.3 832.9

Statistical Techniques Statistical technics used were Student’s t test for interval scale data, and the chi-square test or Fisher’s exact test for categorical data. The Mann-Whitney U test was employed for comparison of NRS. A P value of less than 0.05 was regarded as statistically significant.

AMERICAN

Staging,

NS

Data Analysis The two treatment groups were compared for demographic data, preoperative risk factors, anatomic location of esophageal tumor, types of operation and incidence of complications, which included the incidence of anastomotic leakage and respiratory and cardiovascular complications. Respiratory complications included sputum retention requiring frequent bronchoscopic suction, radiologic and microbiologic evidence of lung consolidation, pneumonia, and respiratory failure. Cardiov&cular complications included cardiac arrhythmias, cardiac ischemia, and acute myocardial infarction or congestive heart failure. Postoperative mortality and length of hospital stay were also studied.

THE

Site,

risk

High Preoperative cardiac risk Normal Moderate High Level of tumor Cervical Superior mediastinal Mid third Lower third Cardia Stage I Ila Ilb Ill IV Operation Transthoracic Nontransthoracic Operation duration: mean 2 SD (minutes) Blood loss: mean 2 SD (ml)

474

Tumor

Group APS (n = 299)

Moderate

APS = acute pain sewice;

Risk,

JUNE

1997

/ POSTOPERATIVE TABLE

Ill Overall

Postoperative

Morbidity,

Pulmonary complications Cardiovascular complications Tracheostomy Postoperative mechanical ventilation Anastomotic leakage* Hospital mortality Mean hospital stay (days 2 SD)

ANALGESIA

Mortality,

and

AFTER

Hospital

ET AL 1

Stay

Group APS (n = 299)

Group CON (n = 279)

39 65 48 77 11 25 21.3

61 106 70 84 11 38 28.6

(13%) (22%) (16%) (26%) (4%) (8%) ? 18.9

ESOPHAGECTOMY/-WJI

P

(22%) (38%) (25%) (30%) (4%) (14%) k 33.0

0.005
figures represent number of patients (percent). unless otherwise stated * /r&ding subciuvcal leakages. APS = acute pawn service; CON = convenbonaipain treatment; NS = not signkant.

TABLE

IV Postoperative

Morbidity,

Mortality,

and

Transthoracic Group APS (n = 226) Mean age (yrs 2 SD) Level of tumor Cervical Superior mediastinal Mid third Lower third Cardia Stage I Ila Ilb Ill IV Pulmonary complications Cardiovascular complications Postoperative mechanical ventilation Anastomotic leakage* Hospital mortality Mean hospital stay: (days 2 SD) figures

61.9

Hospital Stay Thoracotomv

in Group APS and for Resection

Transthoracic Group CON (n = 169)

2 9.1

62.1

P

2 9.0

Group

CON

Stratified

Nontransthoracic Group APS (n = 73)

NS

65.5

According

to Use

of

Nontransthoracic Group CON (n = 90)

? 9.5

66.3

P

+ 11.0

NS

1% 2% 69% 17% 11%

0 2% 62% 24% 12%

NS

0 18% 12% 33% 37%

1% 10% 18% 28% 43%

NS

4% 12% 4% 77% 3% 30 (13%)

2% 12% 9% 74% 3% 48 (25%)

NS

10% 13% 0 67% 10% 9 (12%)

9% 14% 0 59% 18% 13 (14%)

NS

48 (21%)

81 (43%)

17 (23%)

25 (28%)

NS

59 (26%) 9 (4%) 17 (8%)

62 (33%) 6 (3%) 26 (14%)

18 (25%) 2 (3%) 8 (11%)

22 (24%) 5 (6%) 12 (13%)

NS NS NS

22.2

k 20.3

represent number of patients (percent), subclinicai leakages. APS = acute pain service; CON = conventional

30.4

unless otherwise

0.002
2 37.3

18.5

0.005

2 13.3

24.7

NS

t 21 .l

0.03

stated.

* hciuded

pain treatment;

NS = not s/gnificant.

tion of analgesia was temporarily stopped in these patients, and only 1 patient required naloxone reversal and none suffered any sequalae. Pruritus occurred in 13.9% of patients who received epidural morphine, which was significantly higher than with systemic morphine. Most patients who had epidural morphine rated their overall pain management highly. No major complications (such as infection, spinal cord injury or hematoma formation) related to epidural catheter insertion occurred in our series. Minor technical problems occurred in 10 patients, and these include slippage of epidural catheter in the ICU in 7 cases and blockage in 3 cases. THE

AMERICAN

COMMENTS Esophagectomy is a major operation performed in a group of patients with high risk factors including advanced age, chronic smoking, frequent association with abnormal chest conditions, and poor nutritional status.5.“.13,15,23X252i ln spite of this, the hospital mortality has declined from 29%’ to approximately 10% or less in recent years4,’ This is due to an overall improvement in preoperative care, better surgical technique, anesthetic management, and postoperative care.15 Surgical complications (such as anastomotic leakage) are no longer a significant cause of morbidity and mortality. 4~‘8~2’)On the other hand, medical complications are JOURNAL

OF

SURGERY@

VOLUME

173

JUNE

1997

475

OSTOPERATIVE TABLE

ANALGESIA

AFTE

V Analysis

on Postoperative Comparison

Analgesia Between

in 206 Patients Epidural and

Under Systemic

score

(25%-75%

206 -

165 7.3 k 7.0

41 30.9 2 1.8

1 .o (O-3.0) 4.0 (2.0-6.0)

1 .o (O-2.5) 4.0 (2.0-5.5)

1.5 (O-3.0) 5.0 (2.0-6.0)

NS 0.039

0 (O-2.0) 4.0 (2.0-5.0)

0 (O-2.0) 4.0 (2.0-5.0)

0 (O-2.0) 4.5 (3.0-5.5)

NS 0.045

0 (O-l .O) 3.0 (2.0-5.0)

0 (O-l .O) 3.0 (2.0-5.0)

0 (O-l .O) 3.0 (2.0-4.0)

NS NS

0 (O-O) 0 (O-2.0)

0 (O-l 0 (O-l

0 (O-0.5) 0 (O-3.0)

NS NS

AMERICAN

JOURNAL

OF SURGERY”

P
quartile)]

.O) .5)

3 (2%) 18 (9%) 24 (12%)

2 (1%) 12 (7%) 23 (14%)

1 (5%) 6 (15%) 1 (2%)

NS NS 0.04

181 (88%) 25 (12%)

150 (91%) 15 (9%)

31 (76%) 10 (24%)

0.007

not markedly reduced and have replaced surgical problems as the major cause of postoperative morbidity and mortality.“28 After esophagectomy and general anesthesia, the limited preoperative respiratory reserve of these patients is further compromised by deterioration in pulmonary mechanics, sputum retention, and immobilization in the immediate postoperative period. Pain aggravates pulmonary complications because it discourages the patient from taking deep inspirations, coughing effectively and cooperating with the chest physiotherapist. Pain also causes sympathetic discharge, which increases myocardial oxygen consumption (through hypertension and tachycardia) and predisposes to arrhythmia and deep venous thrombosis. Hence, effective pain relief is most important. Intermittent IM opioid is the conventional postoperative analgesic technique because it is simple and safe. However, it has been shown that analgesia is not adequate in most patients.’ ‘-’ ’ Whether increasing the frequency of IM opioids can provide adequate analgesia is uncertain. Effective analgesia can now be achieved with epidural analgesia” or PCA.‘“>” The efficacy of these techniques was demonstrated by the low median NRS at rest in our patients that decreased from 1.0 on the day of operation to 0 on subsequent days. However, mild to moderate pain (NRS 3.0 to 4.0) still occurred during cough in the early postoperative period. As in other reported series,7 a comparison of pain scores between group AT’S and group CON was not possible as pain assessment was not routinely charted before the imTHE

Service: Systemic Morphine*

’ Systemic morphine includes 23 patients wth PCA morphine and 18 p&rents with intravenous morphine. + Respiratory depression = respiratory rate < 1O/mnute or pulse oximetry reading iess than 90% for more than 1 mrwte, (ep,dural group 17 6, systemic group 79). MS = not signkant.

476

Pain

Epidural Morphine

All Cases Number of patients Mean dose (mcg/kg/h ? SD) Pain assessment [median NRS On day of operation At rest During cough First posoperative day At rest During cough Second postoperative day At rest During cough Third posoperative day At rest During cough Side effects Respiratory depression+ Nausea Pruritus Overall satisfaction Good Fair or unsatisfactory

Care of Acute Morphine

VOLUME

173

in 135 spontaneously

breathing

patients

plementation of APS. However, the superior analgesic efficacy of epidural morphine over conventional IM morphine after esophageal surgery has been demonstrated in a randomized, prospectively controlled study conducted in our institution.” The incidence of pulmonary as well as cardiovascular complications was significantly lower in group APS compared with group CON in patients who underwent transthoracic esophagectomy. This significance was lost when compared with resections not involving thoracotomy. The transthoracic approach was associated with more postoperative pain as well as derangement in pulmonary functions than procedures without thoracotomy. Effective analgesia was therefore more important in the thoracotomy group to minimize postoperative morbidity and mortality. The overall hospital mortality was also lower in group APS compared with group CON (8% versus 14X, P = 0.043). Our hospital mortality included all patients who were not discharged after operation and comprised patients who died even up to 6 months after operation; many of these patients remained in the hospital because of social reasons or advancing malignancy. Our study supports a previous report on a favorable outcome after esophageal cancer resection with the introduction of better analgesic methods,7 and, likewise, we note that the main limitation of our study was that the two groups of patients belonged to two different time periods. A prospective randomized controlled study to compare APS with the conventional approach is not considJUNE

1997

1 POSTOPERATIVE

ered clinically appropriate since the former’s effectiveness is now well established.19~?“~“~‘3 In our study, the comparability of the two study groups was substantiated by the similarity in preoperative risk analysis, tumor location and staging, and treatment by the same small surgical team. Apart from the provision of APS, there has been only minor change in surgical technique and patient management policy. In further support of the causal relationship with APS, there was a sudden decrease in pulmonary complications (27% to lZ%), cardiac complications (43% to 19%), and hospital mortality (19% to 10%) from 1988 to 1989 that coincided with the implementation of APS. Improvement in postoperative pain relief should therefore be considered a significant contributor of good results, although the contribution of a combination of small changes, such as antibiotics, policy of bronchoscopic suction, selective early removal of chest drains and nasogastric tubes, should not be underestimated. The best quality of pain relief is achieved with epidural morphine. This is the most popular analgesic technique employed in our institution for esophagectomy. Since morphine is administered in close proximity to its site of action in the spinal cord, good quality analgesia can be achieved regionally at one quarter dose of IV morphine. This is particularly suitable for postesophagectomy patients because better analgesia can be achieved with less sedation and other systemic side effects.‘,” We insert the epidural catheter at T8 to L3 level because it is effective and technically simple. Since epidural morphine diffuses extensively in cerebrospinal fluid to block higher dermatome levels, the level of epidural catheter insertion at T8 to L3 is adequate to cover both the thoracotomy and abdominal incisions. One further advantage of an epidural morphine over epidural local anaesthetic agent is circulatory stability as the latter causes sympathetic blockade, which may require fluid loading to prevent hypotension. However, special precautions are necessary for all epidural opioid administrations in view the risk of delayed respiratory depression. When there are technical difficulties in inserting epidural catheter, PCA or continuous IV morphine are good alternatives. Patient-controlled analgesia is simple, safe, and effective in postoperative analgesia.‘6*” It must be acknowledged that intermittent IM opioid cannot satisfactorily cope with the variability in postoperative pain intensity and fluctuating analgesic requirement.‘” With PCA, the patient can titrate to individual analgesic needs and minimize the time interval between making a request and receiving an analgesic dose. Adequate analgesia also reduces hospital costs because of the saving of expenses on treatment of cardiopulmonary complications, more complex nursing care, and a shorter hospital stay.6,7 With the average daily cost of about US$400 per bed in the Queen Mary Hospital, this translates into an annual saving, on shortened hospitalization alone, of about US$ZOO,OOO for the 60 patients who underwent esophagectomy in our institution. In Hong Kong, public health service is provided by government and is available to the public at low cost. Many patients stay in hospital longer because the home environment is not satisfactory, and home nursing service is not widely available. The THE

AMERICAN

ANALGESIA

AFTER

ESOPHAGECTOMY/TSUI

ET AL 1

length of hospital stay is therefore longer than that in the United States. We conclude that the provision of effective postoperative analgesic techniques by APS contributes to the reduction of cardiopulmonary complications, hospital stay, and mortality in patients undergoing transthoracic esophagectomy for cancer. The authors demonstrate that the type of postvperatiue analgesic technique really does influence the frequency of cardiovascular com@ications, morbidity, and mortality, as well as length of hospital stay fey patients undergoing transthoracic esophagectomy

REFERENCES 1. Earlam R, Cunha-Melo cinoma: I. A crltical review 390.

JR. Esophageal c>f surgery. Br./

squamous cell carSurg. 1980;67:381-

2. Gurkan N, Terzioglu T, Tezelman S. Sasmaz 0. Transhiatal oesophagectomy for oesophageal carcinoma. Br.l Surg. 1991;78:13481351.

3. Yamanaka H, Hiramatsu Y, Kawaguchi Y, et al. Surgical treatment for poor risk patients with carcmuma ofoesophagus. JpnJ Surp. 1991;21:178-181.

4. Law SYK, Fok M, Wang J. Risk an,llysis in resection of squamous cell carcinoma of the ebc,phagus. World J Surg. 1994;18:339-346. 5. Fok M, Law SYK, Wang J. Operable clesophageal carcinoma: current results from Hong Kong. World J Surg. 1994;18:355360. 6. Smedstad KG, Beattie WS, Blair WS, Buckle): DN. tive pain relief and hospital stay afwr total esophagectomy. Pain. 1992;8:149%153.

PostoperaClin I

7. Watson A, ,411en PR. Influence c>fth<)racic epidural analgesia on nutcOme after resection for esc&geal cancer. Surgery. 1994;115:429-432. 8. Yeager MI’, Glass DD, Neff RK, Brmck-Johnsen T. Epidural anesthesia and analgesia m high-risk surgical patients. Anesthesiology. 1987;66:729-736. 9. Shulman MS, Brehner J. Sandier A. The effect of epldural mor-

phine on postoperative pain relief and pulmonary function on rhoracotomy patients. Anesthesiology 1983;59:A192. 10. Austin KL, Srapletcm JV, Mathcr LE. Multiple intramuscular injections: a major source of variability *II ,m&esic response t0 meperidme. Pain. 1980;8:47762. 11. Tsui SL, Chan CS, Chan ASH, vr al. Postoperative analgesia for oesophageal surgery: a cumparwn of three analgesic regimens. Anaesth Intens Care. 1991;19:329-337. 12. Bsnica JJ. Postoperati\;e pain. In: Bonica JJ, ed. Tb Management of Pain. 2nd ed. 1990:461-480. 13. Smith G. Postoperative paIn. In: Nlmmo WS, Smith G, eds. Anaesthesia. 1st ed. Vol 2. London: Blackwell; 1989:1175% 1197. 14. Cousins MJ, Mather LE. Intrathec‘~l ,md cpidural administration of opioids. Anesthcs~ology 1984;h 1:276-J 10. 15. TSUI SL. Anaesthesia for oesophage~~l cancer surgery. South African] Surg. 1993;31:17-18. 16. Bennett RL, Batenhorst RL. B~\ms B,4, et al. Patient-controlled analgesia: a new concept of p<,.stoperatlve pain relief. Ann Surg. 1982;195:700-705. 17. Camp JF. Patient-cuntroll~~l ,rnnlgesia. Am Fam Phys. 1991;44:2145%2150. 18. Ready LB, Edwards WT. Introductlr~n. 1n: Ready, Edwards WT, eds. Management of Acute Pam: A Praztrcnl Guide. International ASsociatmn for the Study of Pain; 1992: I ~ IO. JOURNAL

OF SURGERYE

VOLUME

173

JUNE

1997

477

19. Read\7 LB, Oden R, Chadwick HS, et al. Development of an anesthesiology-hased acute postoperative pain management service. Anesthesioloa. 1988;68:1@0~106. 20. Wheatley RG, Madej TH, Jackson IJB, Hunter D. The first year’s experience of an acute pain service. Br J Anaesth. 1991;67:353-359. 21. Zimmetmann DL, Stewart J. Postoperative pain management and acute pain service actwiq in Canada. Can J Awsth. 1993;40:568-575. 22. Cartwright PD, Helfinger RG, Howell JJ, Siepmann KK. Introducing an acute pain service. Anaesthesia. 1991;46:188-191. 23. Tsui SL, Lo JR, Tong WX, et al. A clinical audit for postoperative pain control on 1443 surgical patients. Acta Anaesthesiologicu Sinica. 1995;33:137-148. 24. Katz J, Melzack R. Measurement of pain. Anesthes Clin North Am. 1992;10:229-245.

478

THE

AMERICAN

JOURNAL

OF

SURGERYa

VOLUME

173

25. Wong J. Surgery in esophageal cancer: how radical should it be. Dig Surg. 1993;10:164-166. 26. Fok M, Law S, Stipa F, et al. A comparison of transhiatal and transthoracic resection for oesophageal carcinoma. Endoscopg. 1993;25:660-663. 27. Fok M, Wang J. Oesophageal cancer treatment: curative modalities. Eur J Gastroenteml Hepd. 1994;6:676-681. 28. Law SYK, Fok M, Cheng SWK, Wang J. A comparison of outcome after resection for squamous cell carcinomas and adenocarciSurg Gynecol Obstet. nomas of the esophagus and canha. 1992;175:107-112. 29. Paterson IM, Wong J. Anastomotic leakage: an avoidable complication of Lewis-Tanner oesophagect
JUNE

1997