1
An Evaluation of International Contemporary Operative Outcomes and
Management Trends Associated with Esophagectomy: A Four-Year Study of Over
6000 Patients Using ECCG Definitions and the Online Esodata Database
Madhan K. Kuppusamy, FRCS (C-Th) and Donald E. Low, FACS, FRCS On Behalf of the
International Esodata Study Group (IESG)
List of Institutions: Agaplesion Markus Krankenhaus, Frankfurt, Germany; Amsterdam UMC, University of Amsterdam, Amsterdam,
Netherlands; Cambridge Oesophago-Gastric Centre, Addenbrookes Hospital, Cambridge, UK; Claude Huriez University Hospital, Lille, France;
Erasmus Medical Center, Rotterdam, Netherlands; Esophageal and Lung Institute, Allegheny Health Network, Pittsburgh, PA, USA; Guy’s & St
Thomas’ NHS Foundation Trust, London, UK; Hirslanden Medical Center, Zürich, Switzerland; Hôpital Nord, Aix-Marseille Université,
Marseille, France; Hospital Universitario del Mar, Barcelona, Spain; Karolinska Institutet and Karolinska University Hospital, Stockholm,
Sweden; Katholieke Universiteit Leuven, Leuven, Belgium; Keio University, Tokyo, Japan; Massachusetts General Hospital, Boston, MA, USA;
MD Anderson Cancer Center, Houston, TX, USA; Memorial Sloan Kettering Cancer Center, New York City, NY, USA; National University
Hospital, Singapore, Singapore; Newcastle Upon Tyne Hospitals, Newcastle upon Tyne, UK; Nottingham University Hospitals NHS Trust,
Nottingham, UK; Odense University Hospital, Odense, Denmark; Oregon Health and Science University, Portland, OR, USA; Oxford University
Hospitals NHS Foundation Trust, Oxford, UK; Princess Alexandra Hospital, University of Queensland, Brisbane, Australia; Queen Elizabeth
Hospital University of Birmingham, Birmingham, UK; Queen Mary Hospital, The University of Hong Kong, Hong Kong SAR, China; Royal
Victoria Hospital, Belfast, Northern Ireland; Sichuan Cancer Hospital & Institute, Chengdu, China; St. James’s Hospital Trinity College, Dublin,
Ireland; Tata Memorial Centre, Mumbai, India; The University of Chicago Medicine, Chicago, IL, USA; Toronto General Hospital, Toronto,
Canada; University Hospital of Cologne, Cologne, Germany; University Hospital Southampton NHS Foundation Trust, Southampton, UK;
University Medical Center, Utrecht, Netherlands; University of Michigan Health System, Ann Arbor, MI, USA; University of São Paulo School
of Medicine, São Paulo, Brazil; University of Verona, Verona, Italy; Virginia Mason Medical Center, Seattle, WA, USA; Vita-Salute San
Raffaele University, Milan, Italy
Word Count
Title: 28
Abstract: 269
Paper w/o Tables: 4694
Running Head: Esophagectomy outcomes Esodata database
CORRESPONDENCE TO:
Donald E. Low, FACS, FRCS(C), FRCSI(Hon), FRCSE(Hon), FRCSEdin(Hon)
Virginia Mason Medical Center
1100 Ninth Ave
Seattle, WA 98111
Phone: 206-223-6164
Fax: 206-625-7245
donald.low@virginiamason.org
2
INTRODUCTION
Major complications following esophageal cancer surgery are associated with increased
operative mortality,
1-3
cancer recurrence, diminished long-term survival,
4-8
longer hospital stay,
8-
10
more readmissions,
11-14
greater hospital costs,
15,16
and worse long-term health-related quality
of life (HR-QL).
17,18
Mortality rates are best in high volume centers, but 90-day mortality
remains between 2.5% and 7%, and, even in a minimally invasive era, morbidity rates are
high.
3,19,20
Until standardized and generally accepted definitions were developed by the Esophageal
Complications Consensus Group (ECCG) in 2011, it was impossible to make comparisons
between studies. The first internet-based international oncologic dataset (Esodata.org), has
previously provided a contemporary benchmark for the incidence of perioperative esophageal
outcomes including complications.
19
This has provided a validated methodology for
institutions,
20
a tool for systematic reviews,
9
a secondary outcome data set used within
international randomized clinical trials (RCTs),
21
and national datasets
22
as well as an
infrastructure for comparative national and international audits, thereby supporting the
development of quality improvement programs.
An ECCG publication on benchmarks comprehensively documented all perioperative
complications as well as 30- and 90-day mortality in 2,704 resections from 24 high volume
centers representing 24 countries collected over a two-year period between January 2015 and
December 2016.
19
The report herein expands this dataset to over 6000 patients, with the primary
aim of the present study to compare outcomes, including rates of four specific complications
(anastomotic leak, conduit necrosis, recurrent laryngeal nerve injury, chyle leak) between the
original ECCG dataset (January 2015-December 2016) to a more recent two-year period
3
(January 2017- December 2018). Secondary goals included reporting on recommended quality
improvement measures developed by the ECCG
23
(90-day mortality rate, 30-day readmission
rates, requirement for documenting change/escalation in level of care, blood product utilization
(quantity and timing) and documentation of discharge location). This study will also assess the
utility of an internet-based high volume international standardized dataset to document short-
term evolution in patient and tumor demographics, operative technique, mandatory quality
measures as well as individual complication incidence and overall complication severity.
METHODS
The Esodata dataset, based on the standardized platform developed by the ECCG,
23
was
initiated as a secure, web-based database in March 2015. The original 24 centers of the ECCG
committed to entering all esophageal resections done at their institution starting in January 2015
(see Membership Agreement Supp1). Data was collected from these 24 centers until December
2016 when the preliminary benchmark of complications was published.
19
Over the subsequent year, 15 additional high volume international esophagectomy
centers applied and were admitted to the International Esodata Study Group (IESG), so currently
there are 39 centers representing 19 countries (Table 1). All centers were responsible for
complying with institutional and national ethics and IRB requirements and the dataset remained
in a format where all patient information within the database was anonymized to comply with
international data privacy agreements.
The original Esodata dataset was designed using the consensus-based data fields and
definitions agreed by the ECCG, and only registered contributors from participating institutions
could enter patients into the database. The data collection process via a secure web interface was
4
constructed using a standardized “user-friendly” online database platform to encourage not only
compliance with entering all esophageal resections but also encouraging data completeness and
integrity. In the context of the wide variation in computer systems between the 39 participating
centers, the integrity and security of the dataset was guaranteed by utilizing a modern web
browser based interface using encrypted network communications. In addition, pre-existing
computer systems did not require institutional IT support or the downloading of complex
software systems onto individual institutional hard drives. The dataset is available to surgeons,
data managers and cancer coordinators through a secure sign-on at any location that they had
internet connections allowing encrypted network communications. The dataset was constructed
with a database interface using multiple level data validation algorithms that allowed only
targeted data to be submitted.
The database and web portal were hosted, as previously described,
19
in a private
dedicated web server and database interface that was accessible only through an authenticated
and encrypted secure network connection (SSL Client and Server Certificate with Extended
Validation issued by Symantec Corporation, Mountain View, CA). Open-source database server
package (MariaDB V10.1.21 by MariaDB Foundation, Redwood City, CA) with regular
encrypted backup system arrangements in combination with Drupal Content Management
System (Distributed under the terms of GNU General Public License) was used for user and data
access management. The system as originally designed provided portability, dynamic live data
analytics, modularity and flexibility in content access management. The registered data
contributors from each institution were authenticated individually to access the database
interface in the members-only area of the Esodata.org web portal. Importantly, registered
5
contributors had instant access to their own institutional results on the Esodata website that was
available to them anywhere in the world where secure internet access was available.
Statistical Methods
The cohort was stratified according to time period of inclusion. Comparisons between
the groups were done using Chi-square test. Quantile regression model was used to analyze the
association between grade of anastomotic leak, conduit necrosis, recurrent laryngeal nerve
injury, and chyle leak to Clavien-Dindo score and length of hospital stay. Significance level was
set at 0.05. Analyses were performed using STATA® version 13 software (StataCorp LP,
College Station, Texas, USA).
RESULTS
Outcome Analysis of Entire Study Population (2015-18)
The current study population includes 6,022 patients entered into the Esodata dataset
between January 2015 and December 2018. Patient demographics are shown in Table 2. The
majority were male (78.3%), with a mean age of 63.2 years. Patients presented with a BMI >30
in 20.5% of cases, and 5.4% had a BMI <18.5. ASA II was most common, reported in 44.3%,
with ASA III at 40.0% and ECOG 0/1 in 93.9%. Most tumors were located in the distal
esophagus (54.4%) or at the esophagogastric junction (33.8%). Minimally invasive
esophagectomy (MIE) surpassed open cases, at 52.8% vs. 47.2% cases, respectively, and 53.1%
of MIE cases were accomplished with MI approaches in both abdomen and thorax (Table 3). A
