1. ORIGINAL ARTICLES
Postoperative Complications After Esophagectomy for
Adenocarcinoma of the Esophagus Are Related to Timing
of Death Due to Recurrence
Sjoerd M. Lagarde, MD,* Johannes D. de Boer, MSc,* Fiebo J. W. ten Kate, MD,†
Olivier R. C. Busch, MD,* Huug Obertop, MD,* and Jan J. B. van Lanschot, MD*
Background: Esophagectomy is frequently accompanied by sub-
stantial complications with secondary disturbance of the immune
T he incidence of adenocarcinoma of the esophagus is
rapidly rising.1–3 Esophageal adenocarcinoma is an
aggressive disease with early lymphatic and hematogenous
system. After esophagectomy for adenocarcinoma of the distal
dissemination. Surgery is the best curative treatment op-
esophagus and/or gastroesophageal junction, the majority of patients
tion. Centralization of surgical resections, advances in
develops an early recurrence and dies within 2 years. The aim of this
surgical techniques and improvements in perioperative
study was to determine the relevance of perioperative complications
care (eg, epidural analgesia, early extubation) have re-
on the timing of death due to recurrence.
duced the risk of esophagectomy to an acceptable level.4 – 8
Methods: A consecutive series of 351 patients who underwent
esophagectomy for adenocarcinoma of the esophagus and gastro-
However, these advances have failed to translate into a
esophageal junction was reviewed.
substantial improvement of long-term survival after surgi-
Results: Of the 351 included patients, 191 patients (54%) died due
cal resection with 5-year survival rates rarely exceeding
to recurrence of esophageal adenocarcinoma. Of these 191 patients,
30%.6,8 –11
77 (40%), 138 (72%), and 186 patients (97%) died before 12, 24,
Despite comprehensive preoperative staging to select
and 60 months, respectively. Multivariate Cox regression analysis
patients for potentially curative surgery, many have unrecog-
demonstrated that T-stage, lymph node ratio 0.20, the presence of
nized metastatic disease at the time of first presentation and
extracapsular lymph node involvement, but not complications were
later present with locoregional, hematogenous, and/or trans-
significant factors for the prediction of death due to cancer recur-
celomic recurrences. Disappointingly, even in patients who
rence. However, in the patients who died, multivariate Cox regres- were operated on with curative intent, the majority develops
sion analysis demonstrated that not only the presence of extracap- recurrence and dies within 2 years.7,12–16 Only a minority of
sular lymph node involvement but also the occurrence of patients developing recurrence suffers from relatively late
complications were significantly related with a shorter time interval recurrence. This marked difference in biologic behavior prob-
until death due to recurrence. ably depends both on tumor characteristics and on host
Conclusion: The relation between perioperative complications and factors. It has been suggested that potent immunosurveillance
cancer recurrence per se is not causal. However, postoperative is of pivotal importance to eradicate microscopic residual
complications are independently associated with the early timing of disease after surgical resection.17–19 Major surgical resection
death due to cancer recurrence. A possible explanation for this is frequently accompanied by substantial complications with
phenomenon is that immunologic host factors enhance microscopic secondary disturbance of the immune system.17–19 It could be
residual disease to develop more rapidly into clinically manifest hypothesized that these complications might have a negative
recurrence. impact on immunosurveillance and thus on tumor recurrence
and long-term survival.
(Ann Surg 2008;247: 71–76) To our knowledge, no attention has been paid to the
predictors of timing of death due to recurrence after
esophagectomy. It remains unknown whether patients who
From the *Departments of *Surgery and †Pathology, Academic Medical die due to an early recurrence had more advanced disease
Center at the University of Amsterdam, Amsterdam, The Netherlands. at the time of operation or had a secondary disturbance of
Huug Obertop and Jan J. B. van Lanschot are currently at Department of the immune system due to the presence of perioperative
Surgery, Erasmus Medical Center Rotterdam.
Supported by a grant (04-77) from the Maag Lever Darm Stichting (Dutch
complications. Therefore, the aim of the present study was
Digestive Foundation) (to S. M. L.). to determine the factors associated with survival due to
Reprints: S. M. Lagarde, MD, Department of Surgery, St. Lucas Andreas cancer recurrence after intentionally curative esophagec-
hospital, Jan tooropstraat 164, 1061 AE Amsterdam, The Netherlands. tomy for adenocarcinoma of the distal esophagus and
E-mail: sjoerdlagarde@hotmail.com, s.lagarde@slaz.nl.
Copyright © 2008 by Lippincott Williams & Wilkins
gastroesophageal junction (GEJ) and especially to deter-
ISSN: 0003-4932/08/24701-0071 mine the relevance of complications to the time interval
DOI: 10.1097/SLA.0b013e31815b695e until death due to cancer recurrence.
Annals of Surgery • Volume 247, Number 1, January 2008 71

2. Lagarde et al Annals of Surgery • Volume 247, Number 1, January 2008
PATIENTS AND METHODS Follow-up
Between January 1993 and January 2003, a consecutive All patients were seen at the outpatient clinic at 3 to 4
series of 351 patients underwent esophagectomy for adeno- months intervals during the first 2 years and every 6 months
carcinoma of the esophagus and/or GEJ. Patients did not thereafter for 3 years. After 5 years, follow-up was obtained by
receive pre- or postoperative chemo- and/or radiation ther- telephone from the patient or the patient’s family practitioner.
apy. Clinicopathologic data, including complications, from Follow-up extended to August 2005 ensuring a minimal poten-
all operated patients were permanently collected in a prospec- tial follow-up of 32 months. Follow-up was complete in all
tive database. patients. Recurrence of disease was only diagnosed on clinical
Preoperative staging was done with endoscopy with grounds. Radiologic modalities were not routinely used. When
histologic biopsy, endosonography, external sonography of recurrence was suspected, additional investigations were per-
the neck, computed tomography scan and, on indication, formed on indication.
positron emission tomography. Surgery was performed/su-
pervised by 1 of 3 experienced surgeons (ORCB, HO, JJBvL) Definition of Recurrence Pattern
at the Academic Medical Center at the University of Amster- Recurrences were classified as hematogenous recur-
dam, a tertiary referral center with a wide experience in esoph- rence, locoregional recurrence, and transcelomic (pleural/
ageal surgery. Operations were performed with curative intent, peritoneal) recurrence. Recurrence at cervical-, celiac-, or
ie, in the absence of local irresectability and/or distant metasta- paraaortic lymph nodes was classified as locoregional recur-
ses (including tumor positive cervical lymph nodes or irresect- rence. Simultaneous locoregional and hematogenous disease
able celiac nodes). Operations were performed via the transhiatal was classified as hematogenous recurrence.
or transthoracic approach. During transhiatal esophagectomy, Statistics
the mid- to distal esophagus was dissected under direct vision
Statistical calculations were performed using SPSS
through the widened hiatus of the diaphragm. The tumor and its
version 12.0 (Statistical Package for the Social Sciences,
adjacent lymph nodes were dissected en-bloc. During transtho-
Chicago, IL). The primary end point of the analysis was
racic esophagectomy, a right-sided postero-lateral thoracotomy
disease specific survival (defined as time from date of surgery
was performed. The esophagus was resected en-bloc with all
to date of death as a result of disease recurrence). To perform
(peri-) esophageal tissue in the mediastinum including the tho-
statistical analysis in a homogeneous group, patients who
racic duct, azygos vein, ipsilateral pleura, and lymph nodes.
died due to postoperative complications and patients who
The subcarinal nodes and the origin of left gastric artery
died due to unrelated causes were excluded. Cox regression
were marked in the resection specimen. Pathologic findings
models were used to examine the association between poten-
were described on a standardized form. The pTNM-stage,
tial predictors and the time until death as a result of esoph-
differentiation grade, radicality of resection, total number of
ageal cancer. The following potential predictors were ana-
resected lymph nodes and total number of positive lymph
lyzed; age, sex, operative approach, American Society of
nodes, including their location were recorded. The lymph
Anesthesiologists classification, pT-stage, pN-stage, presence
node ratio was defined as the ratio between the total amount
of positive truncal nodes, lymph node ratio (categorized into
of positive lymph nodes divided by the total amount of
0.20 or 20%20), the presence of extracapsular LNI, pres-
resected nodes. Lymph nodes were cut in 2 and routine
ence of Barrett mucosa, differentiation grade, microscopi-
hematoxylin and eosin staining was performed using a stan-
cally irradical resection, and the presence of medical or
dardized protocol. Extracapsular lymph node involvement
surgical complications. The Cox proportional-hazards regres-
(LNI) was defined as metastatic adenocarcinoma extending
sion model was used for both univariate and multivariate
through the nodal capsule into the perinodal fatty tissue.
analyses. The models in the present analysis were constructed
(1) to identify factors associated with death due to recurrence
Definition of Complications and (2) to identify prognostic factors in patients who died due
Complications were categorized as medical (including to recurrence.
infectious complications) or complications directly attribut- Variables achieving a probability value of less than
able to surgical technique. Medical complications included 0.1 in the univariate analysis were subsequently introduced
cardiac complications (atrial fibrillation, myocardial isch- in a multivariate stepwise proportional-hazard analysis
emia, heart failure, cerebral infarction), respiratory failure (Cox model) to identify those variables significantly asso-
(diagnosed with blood gas criteria with or without mechanical ciated with death due to recurrence and timing of death due
ventilatory support), atelectasis, pulmonary embolism, and to recurrent disease. Results are given as hazard ratios with
renal failure. Infectious complications included wound infec- their 95% confidence interval. P-values 0.05 (2-sided)
tion, pleural empyema, pneumonia (defined as a new infiltrate were considered statistically significant.
on chest x-ray, accompanied by purulent sputum or fever, for
which treatment with antibiotics was started), and sepsis. RESULTS
Technical complications were recorded as chylothorax, Between January 1993 and January 2003 a consecutive
chyloperitoneum, anastomotic leakage (including both radio- series of 351 patients underwent esophagectomy for malig-
logic and clinical leakages), ischemia of the gastric tube, nant disease of the esophagus and/or GEJ. The following
hemorrhage, recurrent laryngeal nerve palsy, diaphragmatic patients were excluded from this study; 8 patients in whom
herniation, and dehiscence of the fascia. tumor was left behind macroscopically and 10 patients who
72 © 2008 Lippincott Williams & Wilkins

3. Annals of Surgery • Volume 247, Number 1, January 2008 Postoperative Complications After Esophagectomy
died of postoperative complications (3%). Furthermore, 22 Risk Factors for Death Due to Cancer
patients who died during follow-up due to unrelated disease Recurrence
were excluded from this analysis; 8 patients due to myocar- The clinicopathologic characteristics of 191 patients
dial infarction, 6 patients due to cerebral infarction, 5 patients (62%) who died of recurrence and 120 patients (38%) who
due to pulmonary disease, 1 patient due to pathologically were alive are given in Table 1. In univariate analysis, the
proven primary pancreatic cancer, 1 patient due to patholog-
radicality of resection, the depth of invasion (T-stage), the
ically proven primary lung cancer, and 1 patient with Parkin-
presence of lymphatic dissemination, the presence of positive
son disease. Five (1%) patients who experienced recurrence
at last follow-up, but were still alive, were censored in celiac nodes, a lymph node ratio 0.20, the presence of
survival analyses. extracapsular LNI and the differentiation grade of the tumor
and the presence of medical or surgical complications were
all significantly associated with death from recurrence of
TABLE 1. Clinicopathologic Characteristics of Patients With adenocarcinoma of the distal esophagus and/or GEJ (Table
Adenocarcinoma of the Distal Esophagus or 2). There were no significant differences in survival time
Gastroesophageal Junction. Patients Are Divided Into Groups between the 3 operating or supervising surgeons (P 0.258).
of Those Who Died of Cancer Recurrence and Those Who Multivariate Cox regression analysis demonstrated that
Were Alive At Least 32 Months After Date of Operation
T-stage, a lymph node ratio 0.20 and the presence of
Died of Recurrence, Alive, extracapsular LNI were significant factors for the develop-
Factor N 191 N 120
ment of cancer recurrence (Table 3). The presence of medical
Age* 64 (35-85) 64 (37–83) or surgical complications was not related with cancer recur-
Sex rence in multivariate analysis.
Male 162 (85%) 100 (83%)
Female 29 (15%) 20 (17%)
Operative approach
THE 137 (72%) 88 (73%)
TABLE 2. Univariate Cox Regression Analysis for Disease
TTE 52 (28%) 32 (27%)
Specific Survival After Esophagectomy for Adenocarcinoma
ASA classification of the Esophagus or Gastroesophageal Junction. Results Are
1 43 (23%) 25 (21%) Given as Hazard Ratios (HR) and Their 95% Confidence
2 113 (56%) 78 (65%) Interval (CI)
3 33 (17%) 17 (14%)
Factor Hazard Ratio 95% CI P
4 2 (1%) 0
pT-stage Age, 1 yr increment 0.99 0.978–1.01 0.327
pT1 12 (6%) 48 (40%) Male gender 1.02 0.69–1.52 0.918
pT2 19 (10%) 17 (14%) Transthoracic approach 0.91 0.66–1.24 0.538
pT3 148 (77%) 54 (45%) ASA classification (compared with ASA 1)
pT4 12 (7%) 1 (1%) 2 0.87 0.61–1.23 0.419
pN-stage 3 1.04 0.66–1.64 0.878
N0 28 (15%) 69 (58%) 4 1.88 0.45–7.75 0.385
N1 163 (85%) 51 (43%) pT-stage (compared with pT1)
Presence of positive celiac 52 (27%) 16 (13%) pT2 3.27 1.58–6.74 0.001
node (M1a) pT3 6.52 3.61–11.76 <0.001
Lymph node ratio* 0.30 (0.0-1.0) 0.0 (0.0-0.6) pT4 12.89 5.76–28.88 <0.001
Presence of extracapsular 122 (64%) 19 (16%) Presence of lymph node 4.69 3.13–7.03 <0.001
LNI metastasis (pN1)
Presence of Barrett 101 (53%) 75 (63%) Presence of positive celiac 2.01 1.46–2.77 <0.001
mucosa node (M1a)
Differentiation grade Lymph node ratio 0.20 4.57 3.38–6.20 <0.001
Good 4 (2%) 13 (11%) Presence of extracapsular 4.42 3.28–6.00 <0.001
Moderate 61 (32%) 63 (53%) LNI
Poor 126 (66%) 44 (37%) Presence of Barrett mucosa 1.27 0.96–1.69 0.110
Microscopically irradical 47 (25%) 9 (8%) Differentiation grade (compared with good)
resection (R1) Moderate 2.19 0.80–6.04 0.128
Medical or surgical 121 (63%) 55 (46%) Poor 4.53 1.67–12.28 0.003
complications Microscopically irradical 2.34 1.68–3.26 <0.001
ICU stay* 2 (1–70) 2 (1–73) resection (R1)
Hospital stay* 17 (10–104) 15 (10–129) Medical or surgical 1.32 0.99–1.77 0.062
complications
*Values depicted are median (range).
ASA indicates The American Society of Anesthesiologists; THE, transhiatal esoph- Statistically significant results shown in bold.
agectomy; TTE, transthoracic esophagectomy; LNI, lymph node involvement; ICU, ASA indicates The American Society of Anesthesiologists; LNI, lymph node
Intensive Care Unit. involvement.
© 2008 Lippincott Williams & Wilkins 73

4. Lagarde et al Annals of Surgery • Volume 247, Number 1, January 2008
TABLE 3. Multivariate Cox Regression Analysis for Disease TABLE 4. Univariate Cox Regression Analysis for Timing
Specific Survival After Esophagectomy for Adenocarcinoma Until Death in Patients Who Died Due to Recurrence From
of the Esophagus or Gastroesophageal Junction. Results Are Adenocarcinoma of the Esophagus or Gastroesophageal
Given as Hazard Ratios (HR) and Their 95% Confidence Junction. Results Are Given as Hazard Ratios (HR) and Their
Interval (CI) 95% Confidence Interval (CI)
Hazard 95% Confidence Hazard 95% Confidence
Factor Ratio Interval P Factor Ratio Interval P
pT-stage (compared with pT1) Age, 1 yr increment 0.99 0.97–1.00 0.071
pT2 2.28 1.07–4.88 0.034 Male gender 1.02 0.69–1.52 0.920
pT3 3.05 1.58–5.88 0.001 Transthoracic approach 1.13 0.82–1.54 0.467
pT4 5.43 2.28–12.87 <0.001 ASA classification (compared with ASA 1)
Presence of lymph node 1.24 0.73–2.09 0.430 2 0.85 0.60–1.21 0.370
metastasis (pN1) 3 0.98 0.62–1.56 0.941
Presence of positive 1.05 0.74–1.49 0.780 4 1.83 0.44–7.56 0.404
celiac node (M1a) pT-stage (compared with pT1)
Lymph node ratio 0.20 6.91 3.27–14.56 <0.001 pT2 0.94 0.45–1.98 0.871
Presence of extracapsular LNI 1.91 1.28–2.85 0.002 pT3 1.60 0.86–2.96 0.135
Differentiation grade (compared with good) pT4 1.87 0.81–4.30 0.141
Moderate 0.80 0.279–2.30 0.679 Presence of lymph node 2.07 1.36–3.15 0.001
Poor 1.24 0.44–3.52 0.688 metastasis (pN1)
Microscopically irradical 0.98 0.66–1.43 0.897 Presence of positive 1.67 1.20–2.31 0.002
resection (R1) celiac node (M1a)
Medical or surgical 1.21 0.90–1.63 0.214 Lymph node ratio 0.20 1.95 1.44–2.64 <0.001
complications Presence of extracapsular LNI 2.17 1.58–2.98 <0.001
Statistically significant results shown in bold. Presence of Barrett mucosa 1.04 0.78–1.39 0.780
Differentiation grade (compared with good)
Time Interval Until Death Due to Cancer Moderate 1.96 0.62–6.19 0.253
Poor 0.81 0.59–1.10 0.167
Recurrence
Microscopically irradical 1.38 0.99–1.93 0.058
Of the 191 patients who died due to recurrence of esoph- resection (R1)
ageal adenocarcinoma, 77 patients (40%), 138 patients (72%), Medical or surgical 1.47 1.08–2.00 0.013
and 186 patients (97%) died before 12, 24, and 60 months, complications
respectively. Almost 3 quarters died within 2 years and only 5 Statistically significant results shown in bold.
patients (3%) died due to recurrence after 5 years. To identify
which factors are responsible for the time interval until death due
to recurrence, univariate Cox regression analysis was performed. TABLE 5. Multivariate Cox Regression Analysis for the Time
This analysis revealed that besides conventional pathologic fac- Interval Until Death Due to Cancer Recurrence. Results Are
tors, (such as depth of invasion, the presence of lymphatic Given as Hazard Ratios and Their 95% Confidence Interval (CI)
dissemination, the lymph node ratio, the presence of extracap- 95% Confidence
sular LNI), also the occurrence of medical and/or surgical Factor Hazard Ratio Interval P
complications was associated with a shorter time interval until Age, 1 yr increment 0.99 0.98–1.00 0.214
death due to cancer recurrence (Table 4). Multivariate Cox Presence of lymph node 1.07 0.62–1.85 0.807
regression analysis demonstrated that the presence of extracap- metastasis (pN1)
sular LNI and the occurrence of medical and/or surgical com- Presence of positive celiac 1.36 0.96–1.94 0.087
plications were significantly related with a shorter time interval node (M1a)
until death due to cancer recurrence (Table 5). Lymph node ratio 0.20 1.42 0.98–2.04 0.061
The presence of medical or surgical complications (P Presence of extracapsular LNI 1.72 1.18–2.52 0.005
0.030), especially more wound infections (P 0.033) and Microscopically irradical 1.08 0.76–1.53 0.681
more sepsis (P 0.013) was related to a shorter time to death resection (R1)
due to recurrence. Also the presence of surgical complica- Medical surgical 1.46 1.06–2.00 0.020
complications
tions, especially chylothorax (P 0.001) was related with a
shorter time to death due to recurrence (Table 6). The oper- Statistically significant results shown in bold.
ating surgeon was not related with timing of death due to
recurrence (P 0.439).
most 3 quarters of cancer related deaths occurred within 2
DISCUSSION years after surgery, which is in line with other studies.7,12–16
Despite extensive preoperative staging, 60% of the Only a small proportion developed a recurrence after 5 years.
patients treated in our hospital developed cancer recurrence In various malignancies, a relation has been described
within 60 months after (intentionally) curative surgery. Al- between postoperative complications and recurrence of can-
74 © 2008 Lippincott Williams & Wilkins

5. Annals of Surgery • Volume 247, Number 1, January 2008 Postoperative Complications After Esophagectomy
might indicate that the primary tumor biology determines
TABLE 6. The Influence of Specific Complications for the
Time Interval Until Death Due to Cancer Recurrence. Results who will die of recurrence and not a hypothetical negative
From Univariate Analysis Are Given as Hazard Ratios and impact of postoperative complications on oncological immu-
Their 95% Confidence Interval (CI). Hazard Ratios and P- nosurveillance. Possibly, the (indirect) relation between post-
Values If 5 or More Events Were Present operative complications and death due to recurrence can be
Medical and Infectious Hazard 95% Confidence
explained by the fact that the operative procedure is techni-
Complications N Ratio Interval P cally more demanding in patients with more advanced disease
and therefore is accompanied with more postoperative com-
Cardiovascular 27 1.37 0.91–2.06 0.134
complications* plications. Another explanation might be that patients with
Atrial fibrillation 20 1.55 0.98–2.48 0.064 more advanced disease have a higher load of unrecognized
Myocardial ischemia 3 (micro) metastatic tumor that is left behind and makes the
Heart failure 5 patient more susceptible for complications.
Cerebral infarction 1 Interestingly, the presence of postoperative complica-
ARDS 6 1.37 0.91–2.06 0.134 tions did not have a significant effect on recurrence per se, but
Atelectasis 51 1.27 0.92–1.75 0.154 was related to the timing of death due to recurrence in
Pulmonary embolism 4 multivariate analysis. The occurrence of postoperative com-
Renal failure 1 plications was an independent predisposing factor for a short
Wound infection 6 2.44 1.07–5.53 0.033 time interval until death due to cancer recurrence. It is
Pleural empyema 3 unlikely that this effect can be explained by a difference in
Sepsis 8 2.48 1.21–5.09 0.013 preoperative comorbidity because American Society of An-
pneumonia 49 1.24 0.89–1.72 0.202 esthesiologists classification was not a significant factor in
Total medical complications 96 1.38 1.03–1.83 0.030 univariate analysis. It could be hypothesized that complica-
Chylothorax 9 3.47 1.75–6.9 <0.001 tions are accompanied by a secondary disturbance of the
Chyloperitoneum 1 immune system. After major surgery there is a profound but
Anastomotic leakage 28 0.88 0.59–1.33 0.556 self-limiting systemic inflammatory response in patients with
Radiological 16 1.36 0.81–2.28 0.243 an uncomplicated postoperative course. However, patients
Clinical 12 0.61 0.347–1.09 0.096 who develop postoperative complications suffer a “double
Hemorrhage 4 hit,” the first as a result of surgery and the second from their
Recurrent laryngeal 15 1.48 0.87–2.52 0.152 complications. Possibly, these circumstances enhance that
nerve injury residual microscopic disease can develop more rapidly to
Diaphragmatic herniation 3 become clinically manifest and fatal.17–19,22 This is further
Dehiscence of the fascia 5 1.33 0.54–3.25 0.531 underlined by the fact that especially patients who developed
Total surgical 57 1.33 0.97–1.82 0.077 sepsis and patients who had chyle leakage died earlier. The
complications first group of patients probably suffers from a major double
Medical or surgical 121 1.47 1.08–2.00 0.013 hit and the second group of patients suffers from an ongoing
complications loss of fluids and proteins (consisting of eg, lymphocytes and
Statistically significant results shown in bold. immunoglobulines),30,31 which further deteriorates the al-
*There were 2 patients with atrial fibrillation and myocardial ischemia. ready malnourished and immunocompromised patient with
ARDS indicates Acute Respiratory Distress Syndrome.
adenocarcionoma of the esophagus and GEJ.30 –33
In conclusion, the results of the present study indicate
that the relation between postoperative complications and
cer.16,21,22 In esophageal cancer, a prognostic relation be- cancer recurrence per se is not causal. However, postopera-
tween the presence of complications after esophagectomy and tive complications are significantly associated with a short
overall survival has been reported before.6,23 However, in time interval until death due to cancer recurrence in multi-
those earlier studies, no effect of complications was identified variate analysis. A possible explanation for this phenomenon
on tumor specific survival and thus the impact of complica- is that immunologic host factors enhance microscopic resid-
tions on cancer recurrence remained unclear. A study focus- ual disease to develop more rapidly into fatal recurrence.
ing only on anastomotic leakage did not find a long-term
survival effect due to this complication, but the number of ACKNOWLEDGMENTS
patients with anastomotic leakage (n 14) was too small to The authors are very grateful to J. B. Reitsma, PhD,
draw definite conclusions.24 from the Department of Clinical Epidemiology and Biosta-
In the present study, classic pathologic factors, which tistics of the Academic Medical Center at the University of
indicate advanced disease, such as advanced T-stage, lymph Amsterdam for his statistical suggestions.
node ratio 0.20, and the presence of extracapsular LNI,
were all independent prognostic factors associated with death
due to cancer recurrence.11,25–29 Although the presence of REFERENCES
1. Botterweck AA, Schouten LJ, Volovics A, et al. Trends in incidence of
complications was related with survival in univariate analysis adenocarcinoma of the oesophagus and gastric cardia in ten European
in multivariate analysis, the occurrence of complications was countries. Int J Epidemiol. 2000;29:645– 654.
not an independent predictor of death due to recurrence. This 2. Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in he incidence
© 2008 Lippincott Williams & Wilkins 75

6. Lagarde et al Annals of Surgery • Volume 247, Number 1, January 2008
of esophageal and gastric carcinoma in the United States. Cancer. prognosis after surgery for colorectal cancer. Danish RANX05 Colorec-
1998;83:2049 –2053. tal Cancer Study Group. Br J Surg. 2000;87:1553–1562.
3. Pohl H, Welch HG. The role of overdiagnosis and reclassification in the 19. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow?
marked increase of esophageal adenocarcinoma incidence. J Natl Can- Lancet. 2001;357:539 –545.
cer Inst. 2005;97:142–146. 20. van Sandick JW, van Lanschot JJ, ten Kate FJ, et al. Indicators of
4. Wu PC, Posner MC. The role of surgery in the management of oesoph- prognosis after transhiatal esophageal resection without thoracotomy for
ageal cancer. Lancet Oncol. 2003;4:481– 488. cancer. J Am Coll Surg. 2002;194:28 –36.
5. Stein HJ, Brucher BL, Sendler A, et al. Esophageal cancer: patient 21. de Melo GM, Ribeiro KC, Kowalski LP, et al. Risk factors for postop-
evaluation and pre-treatment staging. Surg Oncol. 2001;10:103–111. erative complications in oral cancer and their prognostic implications.
6. Stein HJ, Siewert JR. Improved prognosis of resected esophageal cancer. Arch Otolaryngol Head Neck Surg. 2001;127:828 – 833.
World J Surg. 2004;28:520 –525. 22. McArdle CS, McKee RF, Finlay IG, et al. Improvement in survival
7. Mariette C, Balon JM, Piessen G, et al. Pattern of recurrence following following surgery for colorectal cancer. Br J Surg. 2005;92:1008 –1013.
complete resection of esophageal carcinoma and factors predictive of 23. McArdle CS, McMillan DC, Hole DJ. Impact of anastomotic leakage on
recurrent disease. Cancer. 2003;97:1616 –1623. long-term survival of patients undergoing curative resection for colo-
8. Hulscher JB, van Sandick JW, de Boer AG, et al. Extended transthoracic rectal cancer. Br J Surg. 2005;92:1150 –1154.
resection compared with limited transhiatal resection for adenocarci- 24. Rizk NP, Bach PB, Schrag D, et al. The impact of complications on
noma of the esophagus. N Engl J Med. 2002;347:1662–1669. outcomes after resection for esophageal and gastroesophageal junction
9. Ellis FH Jr, Heatley GJ, Krasna MJ, et al. Esophagogastrectomy for carcinoma. J Am Coll Surg. 2004;198:42–50.
carcinoma of the esophagus and cardia: a comparison of findings and 25. Junemann-Ramirez M, Awan MY, Khan ZM, et al. Anastomotic
results after standard resection in three consecutive eight-year intervals leakage post-esophagogastrectomy for esophageal carcinoma: retro-
with improved staging criteria. J Thorac Cardiovasc Surg. 1997;113: spective analysis of predictive factors, management and influence on
long-term survival in a high volume centre. Eur J Cardiothorac Surg.
836 – 846.
2005;27:3–7.
10. Orringer MB, Marshall B, Iannettoni MD. Transhiatal esophagectomy:
26. Feith M, Stein HJ, Siewert JR. Pattern of lymphatic spread of Barrett’s
clinical experience and refinements. Ann Surg. 1999;230:392– 400.
cancer. World J Surg. 2003;27:1052–1057.
11. Rudiger SJ, Feith M, Werner M, et al. Adenocarcinoma of the esopha-
27. Lerut T, Coosemans W, Decker G, et al. Extracapsular lymph node
gogastric junction: results of surgical therapy based on anatomical/
involvement is a negative prognostic factor in T3 adenocarcinoma of the
topographic classification in 1,002 consecutive patients. Ann Surg. distal esophagus and gastroesophageal junction. J Thorac Cardiovasc
2000;232:353–361. Surg. 2003;126:1121–1128.
12. Wayman J, Bennett MK, Raimes SA, et al. The pattern of recurrence of 28. Rice TW, Zuccaro G Jr, Adelstein DJ, et al. Esophageal carcinoma:
adenocarcinoma of the oesophago-gastric junction. Br J Cancer. 2002; depth of tumor invasion is predictive of regional lymph node status. Ann
86:1223–1229. Thorac Surg. 1998;65:787–792.
13. Hulscher JB, van Sandick JW, Tijssen JG, et al. The recurrence pattern 29. Rice TW, Blackstone EH, Adelstein DJ, et al. Role of clinically deter-
of esophageal carcinoma after transhiatal resection. J Am Coll Surg. mined depth of tumor invasion in the treatment of esophageal carcinoma.
2000;191:143–148. J Thorac Cardiovasc Surg. 2003;125:1091–1102.
14. Dresner SM, Griffin SM. Pattern of recurrence following radical oesoph- 30. Steup WH, De Leyn P, Deneffe G, et al. Tumors of the esophagogastric
agectomy with two-field lymphadenectomy. Br J Surg. 2000;87:1426 – junction. Long-term survival in relation to the pattern of lymph node
1433. metastasis and a critical analysis of the accuracy or inaccuracy of pTNM
15. Dresner SM, Wayman J, Shenfine J, et al. Pattern of recurrence follow- classification. J Thorac Cardiovasc Surg. 1996;111:85–94.
ing subtotal oesophagectomy with two field lymphadenectomy. 31. Lagarde SM, ten Kate FJ, de Boer DJ, et al. Extracapsular lymph node
Br J Surg. 2000;87:362–373. involvement in node-positive patients with adenocarcinoma of the distal
16. de MG, Pedrazzani C, Pasini F, et al. Pattern of recurrence after surgery esophagus or gastroesophageal junction. Am J Surg Pathol. 2006;30:
in adenocarcinoma of the gastro-oesophageal junction. Eur J Surg 171–176.
Oncol. 2003;29:506 –510. 32. Omloo JM, Lagarde SM, Vrouenraets BC, et al. Compartimentalization
17. Abramovitch R, Marikovsky M, Meir G, et al. Stimulation of tumour for chylothorax originating from the abdomen after extended esopha-
growth by wound-derived growth factors. Br J Cancer. 1999;79:1392– gectomy. Report of two cases and review of the literature. Dig Surg.
1398. 2006;23:86 –92.
18. Mynster T, Christensen IJ, Moesgaard F, et al. Effects of the combina- 33. Lagarde SM, Omloo JM, de JK, et al. Incidence and management of chyle
tion of blood transfusion and postoperative infectious complications on leakage after esophagectomy. Ann Thorac Surg. 2005;80:449 – 454.
76 © 2008 Lippincott Williams & Wilkins