O
RIGINAL
A
RTICLE
Colonic Volvulus in the United States
Trends, Outcomes, and Predictors of Mortality
Wissam J. Halabi, MD,
∗
Mehraneh D. Jafari, MD,
∗
Celeste Y. Kang, MD,
∗
Vinh Q. Nguyen, PhD,
†
Joseph C. Carmichael, MD,
∗
Steven Mills, MD,
∗
Alessio Pigazzi, MD, PhD,
∗
and Michael J. Stamos, MD
∗
Introduction: Colonic volvulus is a rare entity associated with high mortality
rates. Most studies come from areas of high endemicity and are limited by
small numbers. No studies have investigated trends, outcomes, and predictors
of mortality at the national level.
Methods: The Nationwide Inpatient Sample 2002–2010 was retrospectively
reviewed for colonic volvulus cases admitted emergently. Patients’ demo-
graphics, hospital factors, and outcomes of the different procedures were
analyzed. The LASSO algorithm for logistic regression was used to build a
predictive model for mortality in cases of sigmoid (SV) and cecal volvulus
(CV) taking into account preoperative and operative variables.
Results: An estimated 3,351,152 cases of bowel obstruction were admitted
in the United States over the study period. Colonic volvulus was found to be
the cause in 63,749 cases (1.90%). The incidence of CV increased by 5.53%
per year whereas the incidence of SV remained stable. SV was more com-
mon in elderly males (aged 70 years), African Americans, and patients with
diabetes and neuropsychiatric disorders. In contrast, CV was more common
in younger females. Nonsurgical decompression alone was used in 17% of
cases. Among cases managed surgically, resective procedures were performed
in 89% of cases, whereas operative detorsion with or without fixation pro-
cedures remained uncommon. Mortality rates were 9.44% for SV, 6.64% for
CV, 17% for synchronous CV and SV, and 18% for transverse colon volvulus.
The LASSO algorithm identified bowel gangrene and peritonitis, coagulopa-
thy, age, the use of stoma, and chronic kidney disease as strong predictors of
mortality.
Conclusions: Colonic volvulus is a rare cause of bowel obstruction in the
United States and is associated with high mortality rates. CV and SV affect
different populations and the incidence of CV is on the rise. The presence of
bowel gangrene and coagulopathy strongly predicts mortality, suggesting that
prompt diagnosis and management are essential.
Keywords: sigmoid volvulus, cecal volvulus, transverse colon volvulus,
trends, mortality
(Ann Surg 2013;00: 1–9)
C
olonic volvulus refers to torsion of the bowel around its own
mesentery. This condition occurs in a long redundant colonic
segment that has an elongated mesentery with a narrow base.
1,2
It
is usually seen in the sigmoid colon, cecum, and less commonly, the
transverse colon and splenic flexure.
3–5
Colonic volvulus is thought
to account for 3.4% of all cases of bowel obstructions in the United
States
5
and 10% to 50% in areas of higher endemicity, such as Africa,
the Middle East, and South America
6,7
This geographic variation is
From the
∗
Department of Surgery, University of California Irvine Medical Center,
Orange, CA; and
†
Department of Statistics, University of California Irvine,
Irvine, CA.
Disclosure: The authors declare no conflicts of interest.
Reprints: Michael J. Stamos, MD, Department of Surgery, University of California
Irvine Medical Center 333 City Boulevard, West Suite 700, Orange, CA 92868.
E-mail: mstamos@uci.edu.
Copyright
C
2013 by Lippincott Williams & Wilkins
ISSN: 0003-4932/13/00000-0001
DOI: 10.1097/SLA.0b013e31828c88ac
thought to be due to anatomical differences
8,9
; differences in diet,
altitude, cultural factors; and endemic infections.
10,11
If left unattended, colonic volvulus can compromise the blood
supply of the involved segment, leading to ischemia, gangrene, perfo-
ration, and death.
1,12
The mainstay of sigmoid volvulus management
has been through proctoscopic or colonoscopic decompression when
feasible, followed by surgery either during the same admission or
electively. Cecal volvulus is mostly managed surgically as the vast ma-
jority of these cases are not amenable to endoscopic decompression.
13
Most published literature on colonic volvulus is limited by
small numbers accumulated over decades, with significant disparities
in management techniques and outcomes. Moreover, these studies
mainly come from areas of high endemicity where volvulus usually
presents in a younger population and is thus associated with lower
mortality rates
11
compared with countries of low endemicity such as
the United States.
In the United States, the few available reports that investigated
the incidence of colonic volvulus relative to other causes of bowel
obstruction are either outdated or limited to large centers in specific
regions. National-level data investigating incidences, practices trends,
and outcomes in different hospital settings are thus lacking. Moreover,
because prior data were limited by small sample sizes, a meaningful
analysis of the predictors of mortality in sigmoid and cecal volvulus
was never undertaken. This is a large retrospective analysis of colonic
volvulus in the United States over a 9-year period, investigating trends,
outcomes, and predictors of mortality of the different procedures
performed for this disease entity.
METHODS
Patient Population and Data Source
Data were extracted from the Healthcare Cost and Utilization
Project Nationwide Inpatient Sample (NIS) database from January 1,
2002 to December 31, 2010. We performed a retrospective analysis
of colonic volvulus cases that underwent operative and nonoperative
management. The NIS is the largest all-payer inpatient care database
in the United States and contains information from nearly 8 million
hospital stays each year across the country. The data set approxi-
mates a 20% stratified sample of American community, nonmilitary,
nonfederal hospitals, resulting in a sampling frame that comprises ap-
proximately 95% of all hospital discharges in the United States. Data
elements within the NIS are drawn from hospital discharge abstracts
that allow determination of all procedures performed during a given
hospitalization.
14
Approval for the use of this database was obtained
from the institutional review board of the University of California-
Irvine Medical Center and the NIS.
Study Aims
The aim of our analysis was to investigate the management
trends of colonic volvulus in the United States over a 9-year period.
Mean patient age, comorbidity scores based on the Elixhauser-Van
Walraven model,
15
and mortality rates for the different procedures
were listed. The yearly number of admissions for all causes of bowel
obstruction in the US was provided as a reference. These include
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery
r
Volume 00, Number 00, 2013
www.annalsofsurgery.com
| 1
Halabi et al
Annals of Surgery
r
Volume 00, Number 00, 2013
small and large bowel obstruction due to adhesions, strictures, tu-
mors, impaction, and hernia with obstruction. In the remainder of the
analysis, we selected what we expected to be the largest groups of
cases, namely, cecal volvulus and sigmoid volvulus that underwent
resection. Patient characteristics, surgical management in different
hospital settings, and the associated mortality rates were examined.
Surgical management includes the use of laparoscopy and the as-
sociated conversion rates and the use of an ostomy. The outcomes
of surgical resection were listed for the 2 groups. Finally, we built
predictive models for in-hospital mortality in patients undergoing re-
section for cecal volvulus and sigmoid volvulus. The model takes into
account comorbidities present on admission, hospital and operative
factors, excluding postoperative complications.
Inclusion Criteria
All patients with an International Classification of Diseases,
Ninth Revision, Clinical Modification (
ICD-9 CM) diagnosis code
for colonic volvulus (560.2) admitted urgently or emergently who
underwent any of the procedures identified by the associated ICD-9
CM procedure codes were included in our analysis: enema decom-
pression, endoscopic decompression, operative reduction, operative
reduction and fixation, cecostomy or sigmoidostomy, resection with
primary anastomosis, resection with the use of an ostomy, and total or
subtotal colectomy. Laparoscopic modifier codes (54.21, 54.51) were
used to identify laparoscopic cases for the years before 2009.
Because colonic volvulus has 1 single ICD-9 diagnosis code,
we used the associated ICD-9 procedure codes to differentiate among
cecal, sigmoid, and transverse colon volvulus that underwent resec-
tion. A binary model was used to ensure that no overlap occurred
among identified cases and that each case is counted once.
Exclusion Criteria
Patients with a diagnosis of colonic volvulus who did not un-
dergo any of the aforementioned procedures or who died before
surgery were excluded from our analysis to minimize the risk of
coding errors. Elective cases were excluded as our aim was to inves-
tigate management trends and outcomes in the acute setting. Missing
variables listed in Tables 2 to 4 were also excluded from our analysis.
These include data on ethnicity, hospital factors, and mortality.
Study Variables
Patient factors, such as age, sex, ethnicity, primary payer type,
and comorbidities, provided by NIS were considered in our predic-
tive models for mortality. The list of comorbidities is based on the
Elixhauser model for comorbidity measures.
16
Presence of peritonitis
or bowel gangrene identified by their respective
ICD-9 codes were
also used in our predictive model. Hospital factors, such as teaching
status, location, and size, were considered in our model, as previ-
ous reports have underlined differences in mortality rates in sigmoid
volvulus cases managed in different hospital settings.
17
Operative
factors, such as the use of stoma, laparoscopic versus open surgery,
and emergent versus semielective surgery, were considered in our
model-building routine as well. Cases that underwent endoscopic de-
compression on admission followed by surgery during the same hos-
pitalization were considered in the semielective surgery group. Those
who underwent surgery on admission without prior endoscopic de-
compression were considered in the emergent surgery group. In the
cecal volvulus model, we excluded the semielective versus emergent
surgery variable, as these cases are usually not amenable to endo-
scopic decompression. This list of variables we hypothesized would
predict that mortality was chosen a priori.
Statistical Analysis
All statistical analyses were conducted using SAS, version 9.3,
and the R Statistical Environment. The LASSO algorithm for logistic
regression
18
was used to identify variables predictive of mortality in
patients who underwent resection for sigmoid or cecal volvulus in a
complete case analysis. Ten-fold cross-validation together with the
1-SE rule was used to determine the model size (number of variables)
to control for overfitting.
19
In contrast to the classic multivariate
logistic regression in which odds ratios are independent of each other
and cannot be added together to predict mortality, LASSO assigns a
coefficient to each predictor. These coefficients can be added together
to calculate the predicted inhospital mortality risk for each individual.
For a coefficient total of x, the inhospital mortality risk is e
x
/(1
+
e
x
). Separate models for sigmoid and cecal volvulus were built. The
receiver operating characteristic curve (ROC) and C-statistic were
used to describe how well our model predicts mortality.
RESULTS
From 2002 to 2010, an estimated 3,351,152 patients were ad-
mitted with a bowel obstruction in the United States. Colonic volvulus
accounted for 63,749 cases (1.90%). Admissions for bowel obstruc-
tion increased steadily from 2002, reaching a peak in 2008 and then
decreasing over the last 2 years of the study period, whereas the num-
ber of admissions for colonic volvulus increased over the last 3 years
of the study period. This increase was mainly driven by admissions
for cecal volvulus that showed an upward trend, increasing by an
average of 5.53% per year over the study period. Sigmoid volvulus
cases remained relatively stable over the study period (Table 1).
Nonsurgical methods, such as enema or endoscopic decom-
pression for sigmoid volvulus, not followed by surgery during the
same admission were used in 16.6% of cases. Enema was used in-
frequently, primarily in relatively elderly patients with high comor-
bidity scores. The use of endoscopic decompression alone showed a
relatively stable incidence and had an associated mortality of 6.4%
(Table 1).
Surgical methods not involving resection remained relatively
infrequent in the acute setting. Surgical detorsion without fixation
(4.2% of surgical cases) was associated with a relatively high mortal-
ity of 7.8%, considering the younger age of this population. Surgical
fixation such as cecopexy or sigmoidopexy (3.3% of surgical cases)
was performed in patients with relatively low comorbidity scores and
was associated with low mortality rates. Enterostomy procedures such
as cecostomy or sigmoidostomy (3.3%) were used in relatively sick
patients and had a 13.0% mortality rate (Table 1).
Among cases managed surgically, resection was the most com-
monly performed procedure in 89.3% of cases. Resective procedures
were performed for cecal, sigmoid, synchronous cecal and sigmoid,
and transverse colon volvulus. A subtotal or total colectomy was re-
quired in 16.0% of sigmoid volvulus cases. In these cases, patients
had relatively high associated comorbidity scores and mortality rates
(14.6%). It is interesting to note that we identified 576 cases of syn-
chronous cecal volvulus and sigmoid volvulus. Patients in this group
had the highest comorbidity scores and a mortality rate of almost
18%. The other interesting finding is the 597 identified cases of trans-
verse colon volvulus: these had high comorbidity scores and mortal-
ity rates considering again the younger age group of this population
(Table 1).
Laparoscopic techniques were rarely applied in the manage-
ment of colonic volvulus, accounting for 3.7% of surgical cases.
However, the use of laparoscopy has increased over the study period,
especially in the last 3 years. Laparoscopy was used in relatively
younger patients with lower comorbidity scores and the associated
mortality rates were lower than their corresponding open counterpart.
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
2
| www.annalsofsurgery.com
C
2013 Lippincott Williams & Wilkins