Effectiveness of a written clinical pathway for enhanced recovery after transthoracic (Ivor Lewis) oesophagectomy




This study assessed the feasibility of a protocol-driven written clinical pathway for multidisciplinary postoperative management after oesophagectomy for oesophageal neoplasia, and examined whether the application of such a protocol could shorten hospital stay and reduce postoperative morbidity and mortality.


Consecutive patients undergoing transthoracic oesophagectomy for oesophageal neoplasia were divided into those treated between 2003 and 2008 to whom a clinical pathway was applied for postoperative management (group 1), and a control group treated between 1998 and 2002 when no clinical pathway was applied (group 2).


There were 74 patients in each group. Morbidity rates were similar in the two groups: 31 per cent in group 1 and 38 per cent in group 2. There were more pulmonary complications in group 2 (23 versus 14 per cent; P = 0·025). One patient (1 per cent) in group 1 and four (5 per cent) in group 2 died after surgery (P = 0·010). The median (range) length of hospital stay was 9 (5–98) days for group 1 and 13 (8–106) days in group 2 (P = 0·012).


Use of a written clinical pathway in patients undergoing oesophageal resection significantly reduced pulmonary complications, postoperative mortality and hospital stay. Copyright © 2010 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.


Oesophageal resection is a complicated procedure, historically associated with high rates of morbidity and mortality1. In recent series mortality rates have progressively dropped to between zero and 10 per cent in specialized centres2, 3. Most centres advocate thorough perioperative management by a multidisciplinary team, including surgeons, anaesthetists, intensive care unit staff and nurses, and the use of agreed written protocols throughout the patient's hospital stay seem beneficial4, 5. These protocols are essentially designed to enhance recovery and reduce the length of hospital stay6.

The average length of hospital stay in most oesophagectomy series ranges from 11 to 26 days, with a mean of 15 days5, 7–9. Even groups performing oesophageal resection using a laparoscopic or thoracoscopic approach report mean hospital stays of 7–13 days10–13. Cerfolio and colleagues5 reported a median hospital stay of 7 days for a series of 90 patients undergoing oesophageal resection with postoperative follow-up using a clinical pathway.

The aims of the present study were to test the feasibility of applying a protocol-driven clinical pathway for multidisciplinary postoperative management of patients undergoing oesophagectomy for oesophageal neoplasia, and to assess whether the application of such a protocol influenced postoperative morbidity, mortality and hospital stay.


Consecutive patients undergoing transthoracic oesophagectomy (Ivor Lewis procedure) for oesophageal neoplasia were divided into two groups. Patients in group 1 underwent surgery between 2003 and 2008, and their postoperative management in the surgical intensive care unit (SICU) and on the surgical ward was carried out according to a clinical pathway. Patients in group 2 had surgery between 1998 and 2002, and no protocol or agreed written clinical pathway for postoperative care was applied. Patients with junctional neoplasia of the cardia (Siewert type II and III) were excluded. Demographics and clinical features of the two groups are shown in Table1.

Table 1. Epidemiological, clinical and tumour characteristics in patients managed according to a fast-track clinical pathway (group 1) and controls (group 2)
 Group 1 (n = 74)Group 2 (n = 74)
  • Values in parentheses are percentages unless indicated otherwise;

  • *

    values are median (range).

  • ASA, American Society of Anesthesiologists; RCT, radiochemotherapy; R0, no macroscopic or microscopic residual tumour; R2, macroscopic residual tumour; R1, invasion of circumferential or longitudinal microscopic borders.

  • P = 0·015 versus group 1 (χ2 test).

Age (years)*59 (23–79)60·5 (17–78)
Patient > 65 years24 (32)20 (27)
Sex ratio (M : F) 64 : 10 65 : 9
ASA grade  
 I10 (14)10 (14)
 II32 (43)31 (42)
 III32 (43)33 (45)
History of smoking40 (54)41 (55)
Alcohol use24 (32)27 (36)
Duration of symptoms (months)*2 (0–10)2 (0–12)
 Squamous cell carcinoma18 (24)35 (47)
 Adenocarcinoma56 (76)39 (53)
Location14 (19)16 (22)
 Proximal60 (81)58 (78)
Neoadjuvant RCT6 (8)5 (7)
Tumour stage  
 I16 (21·6)8 (11)
 II16 (21·6)24 (32)
 III37 (50)39 (53)
 IV5 (6·8)3 (4)
Type of surgery  
 R0 (radical)67 (90)61 (82)
 R1–R2 (non-radical)7 (10)13 (18)

All patients underwent a transthoracic oesophageal resection via the abdomen and right chest. Restoration of continuity of the digestive tract was performed with a circular stapler, creating an end-to-end anastomosis above the azygos vein. A pyloroplasty was performed in all patients. Feeding jejunostomy was never performed.

The management of both patient groups was by the same clinicians in the same environment using intensive care then ward-based facilities. The same principles of care were applied to both groups, the difference being that specific actions were detailed in a written protocol for the first 7 days after operation in group 1. The day-to-day postoperative protocol is detailed in Table2. The clinical pathway was applied in both the SICU (days 1–3 after operation) and the surgical ward (days 4 to 7–9). All patients were extubated in the operating theatre or immediately on arrival in the SICU14 and received an early mobilization strategy, negative fluid balance15, intense respiratory physiotherapy and pain control with epidural analgesia over the first few days. Fluids were given according to urinary volume, central venous pressure and previous history, the aim being to achieve a negative balance 4 days after surgery. Total fluid per day was 1500–2000 ml total parenteral nutrition plus 500 ml normal saline. Epidural analgesia was provided through a 16-G epidural catheter inserted under local analgesia at T6–T9 using 1 µg/ml fentanyl and 0·1 per cent bupivacaine at 6–8 ml/h. Pain control also required infusion of 6 mg metamizole.

Table 2. Clinical pathway for fast-track daily management
 Day after operation
  • *

    Fentanyl 1 µg/ml and bupivacaine 0·1 per cent at 6–8 ml/h.

  • Metamizole is approved for use in Spain.

  • SICU, surgical intensive care unit; FiO2, fraction of inspired oxygen; TPN, total parenteral nutrition; ABG, arterial blood gases; i.v., intravenous; VTE, venous thromboembolism; s.c., subcutaneous; ARDS, acute respiratory distress syndrome.

Treatment areaSICUSICUSICUDischarge from SICUSurgical wardSurgical wardSurgical ward
Oxygen deliveryFace maskNasal cannulaNasal cannulaNasal cannulaNasal cannulaNoNo
Oxygen flow8 l/min4 l/min4 l/min4 l/min4 l/min  
 FiO2 40%      
Liquids (balance)NegativeNegativeNegativeNegativePositivePositivePositive
NutritionTPNTPNTPNTPNTPNDextrose 10%No
Epidural analgesia*YesYesYesRemoveNoNoNo
Tubes and drains 
 Central lineYesYesYesYesYesYesRemove
Physical activityChairChairToiletToiletCorridorWalk
Blood test and ABGYesYesYesYesYesNoYes
RadiologyChestChestChestChestContrast swallowNoChest
 AnalgesiaMetamizoleMetamizoleMetamizoleMetamizoleMetamizoleMetamizoleMetamizole on request
  6 mg 6 mg 6 mg 6 mg 2 mg i.v. 2 mg orally 
  infusion infusion infusion infusion every 8 h every 8 h 
 AntibioticAmoxicillin + clavulanic acid 2 g i.v. every 8 hAmoxicillin + clavulanic acid 2 g i.v. every 8 hAmoxicillin + clavulanic acid 2 g i.v. every 8 hAmoxicillin + clavulanic acid 2 g i.v. every 8 h
 VTE prophylaxisEnoxaparinEnoxaparinEnoxaparinEnoxaparinEnoxaparinEnoxaparinEnoxaparin
  40 mg s.c. 40 mg s.c. 40 mg s.c. 40 mg s.c. 40 mg s.c. 40 mg s.c. 40 mg s.c.
 Proton-pump inhibitorEsomeprazole 40 mg i.v.Esomeprazole 40 mg i.v.Esomeprazole 40 mg i.v.Esomeprazole 40 mg i.v.Esomeprazole 40 mg i.v.Esomeprazole 40 mg i.v.Esomeprazole 40 mg i.v.
Alarm signals
equation image

Patients remained in the SICU for 3 days. Pleural drains, epidural and urinary catheters were removed on day 4 after surgery, and the patient was moved to the surgical ward. On day 5 after operation oesophagography was performed using Gastrografin® (Bayer Schering Pharma, Barcelona, Spain) to check for leaks and adequate passage of contrast to the duodenum. If this was normal, the nasogastric tube was removed and oral diet started with liquids. Parenteral postoperative nutrition was used for the first 4 days. Total parenteral nutrition was discontinued and the central venous catheter removed when the patient had progressed to an oral diet. If there were no complications, with satisfactory haematology, biochemistry and a normal chest radiograph, the patient was discharged between days 7 and 9 after surgery. The clinical pathway allowed for a margin of 2 days (days 7–9) for patient discharge, particularly to accommodate older patients or those needing longer to recover.

An important aspect of postoperative follow-up was careful attention to ‘alarm signals’ on the part of surgeons, anaesthetists and nurses to establish an early diagnosis of complications (usually pulmonary and/or anastomotic leak with mediastinitis). In the presence of fever, hypoxaemia, tachyarrhythmia, leucocytosis or abnormal pleural drainage, thoracoabdominal computed tomography with contrast was performed to diagnose and treat any complications as quickly as possible. Administration of methylene blue by nasogastric tube (while pleural drainage was in place) or contrast oesophagography was also used for suspected anastomotic leakage. This philosophy was applied to both groups.

Statistical analysis

Continuous variables were expressed as median (range) and compared using Student's t test. Categorical data were compared with the χ2 test. In both cases P < 0·050 was considered significant. Statistical analysis was performed with SPSS® version 8.0 for Windows® (SPSS, Chicago, Illinois, USA).


There were 74 patients in each group. Results are shown in Table3. The morbidity rate was 31 per cent in group 1 and 38 per cent in group 2, with no significant difference (P = 0·120). The rate of anastomotic leak was similar in the two groups (7 and 8 per cent respectively). There were more pulmonary complications (pneumonia, pleural effusion, atelectasis and respiratory distress) in the control group than in the protocol-driven group (23 versus 14 per cent; P = 0·025). Postoperative mortality rates were 1 per cent in group 1 and 5 per cent in group 2 (P = 0·010). One patient in group 1 died 5 days after surgery from fulminant acute respiratory distress. Deaths in group 2 were due to pulmonary complications in three patients, and anastomotic leak and mediastinitis in one patient.

Table 3. Morbidity, mortality and length of hospital stay in patients managed according to a fast-track clinical pathway (group 1) and controls (group 2)
 Group 1 (n = 74)Group 2 (n = 74)P
  • Values in parentheses are percentages unless indicated otherwise;

  • *

    values are median (range).

  • χ2 test,

  • Student's t test.

Morbidity23 (31)28 (38)0·120
 Pulmonary without leak10 (14)17 (23)0·025
  Pleural effusion33 
  Respiratory distress53 
  Acute oedema01 
 Anastomotic leak5 (7)6 (8)0·251
 Other8 (11)5 (7)0·154
  Wound complications43 
  Atrial fibrillation22 
  Pericardial effusion10 
  Upper gastrointestinal bleeding10 
Death1 (1)4 (5)0·010
Length of hospital stay (days)*   
 All patients9 (5–98)13 (8–106)0·012
 Patients without complications8 (7–11)12 (8–23)0·009

The length of hospital stay was significantly shorter in group 1 (median (range) 9 (5–98) versus 13 (8–106) days; P = 0·012). For patients who had no complications, the median duration of hospital stay was 8 (7–11) days in group 1 and 12 (8–23) days in group 2 (P = 0·009).

Twenty-three patients (31 per cent) in group 1 were unable to complete the fast-track protocol because they developed complications. The fast-track daily follow-up protocol was therefore applied successfully to 51 patients. Seven patients without complications failed to adhere to the protocol for social reasons (five did not feel well enough to return home and were reluctant to leave hospital; in two the clinical pathway was not applied properly by staff and discharge was delayed unnecessarily). Of these seven patients, five were aged over 65 years and two lived more than 200 miles from the hospital. Thus, 44 patients in group 1 (59 per cent of the total and 86 per cent of those without complications) completed the fast-track protocol and were in hospital for 7–9 days. Only two of these 44 patients re-presented within 15 days of discharge: one for a small wound seroma and one with mild regurgitation. Neither required readmission.

Only 14 patients in group 2 (19 per cent) had a hospital stay of less than 10 days (P = 0·001 versus group 1), despite the fact that 46 patients had no complications in the postoperative period. Three patients returned to the emergency department, two because of wound infection and one owing to deep venous thrombosis, the latter requiring readmission.


Application of protocols for postoperative management using clinical pathways has reduced mortality16 as it enables a more objective, homogeneous and intensive multidisciplinary follow-up. A benefit from using standardized perioperative clinical pathways has also been described after oesophagectomy4, 5. Low and colleagues4 reported excellent results in 340 patients undergoing oesophageal resection with the use of clinical pathways (complication rate 45 per cent but postoperative mortality rate 0·3 per cent).

In the present series two groups of patients underwent oesophagectomy for oesophageal cancer, either with or without a written clinical pathway. The groups were homogeneous, as all patients were operated on by the same surgical team using the same transthoracic (Ivor Lewis) oesophagectomy technique. From 2003 onwards a written agreed postoperative protocol was used by the entire multidisciplinary team involved in patient management. Patients undergoing surgery before this time were looked after by the same professionals and in the same way, although it was felt that by not applying a written protocol a combination of subjectivity and clinical inexperience might influence decision making with adverse effects on outcome. The application of a written clinical pathway seemed useful for professionals with little experience when joining the oesophageal surgery unit and also for experienced specialists, to ensure that diagnostic tests and therapeutic decisions were performed in an objective manner. The application of a clinical pathway led to a significant reduction in postoperative mortality and in the rate of pulmonary complications.

Cerfolio and co-workers5 published the only study with specific reference to fast tracking in oesophagectomy in 2004. The authors applied a daily clinical pathway after Ivor Lewis oesophagectomy in 90 patients. The protocol proposed discharge on day 7 after operation, which was achieved in 78 per cent of the patients. Twenty-one patients were unable to complete the protocol, 16 owing to complications and five for other reasons. Median hospital stay was 7 days but four patients had to be readmitted for minor complications. Morbidity and mortality rates were 27 and 4·5 per cent respectively. The percentage of patients who had to be excluded from the clinical pathway, and were therefore not discharged on day 7, was higher among patients aged over 70 years, those who received preoperative radiation and chemotherapy, and those with major complications. In these respects there are clear similarities with the present findings. Application of a postoperative clinical pathway to patients undergoing oesophageal resection reduced the length of hospital stay. If patients had no complications, they were nearly all discharged between 7 and 9 days after surgery.

The present retrospective study compared two groups of patients who had surgery in different periods, the control group between 1998 and 2002 and the clinical pathway group between 2003 and 2008. It is true that this bias could weaken the study, but the bias was minimized because, before these 148 patients, the surgeons involved had experience of oesophageal resection in more than 200 patients.

Written multidisciplinary management protocols for patients with oesophageal resection significantly reduce pulmonary complications and postoperative mortality. Their use should be encouraged.


This work was supported by grants from Centro de Investigacion Biomedica en Red de Enfermedades Hepaticas y Digestivas (CIBEREHD), Instituto de Salud Carlos III, Ministerio de Sanidad, Spain.

The authors declare no conflict of interest.