TABLE 2. Demographic Characteristics and Comorbidities
for Sigmoid and Cecal Volvulus Cases That Underwent
Resection
N
Sigmoid Volvulus
19,220
Cecal Volvulus
23,392
Age, yr
71 (62–82)
63 (52–78)
Sex, %
Male
65.5
27.0
Female
34.5
73.0
Ethnicity (%)
White
53.3
66.8
Black
14.9
4.9
Hispanic
4.6
2.2
Asian or Pacific Islander
0.9
0.2
Native American
0.4
0.4
Other
1.8
1.1
Missing
24.1
24.4
Primary payer type (%)
Medicare
49.6
38.1
Medicaid
4.3
3.5
Private including HMO
10.3
26.5
Self-pay
1.6
2.4
No charge
0.1
0.2
Other
1.0
1.6
Missing
33.1
27.7
Comorbidities (%)
Anemia
16.6
11.7
Congestive heart failure
16.8
10.4
Chronic pulmonary disease
16.6
20.6
Diabetes
17.2
9.5
Hypertension
43.9
36.1
Liver disease
0.8
1.4
Fluid and electrolyte disorders
48.1
36.4
Obesity
2.7
2.4
Renal failure
5.8
4.6
Weight loss
12.9
9.9
Neurological disorder or paralysis
31.3
8.0
Comorbidity scores
7.4 (3.00–12.00)
4.7 (0.00–8.00)
Continuous variables such as age and comorbidity scores are reported as mean
and interquartile range; categorical variables (sex, ethnicity, payer type, and comor-
bidities) are reported as percent proportions.
HMO indicates health maintenance organization.
Laparoscopy was most commonly used in cases involving fixation of
the colon (Table 1).
Among patients with sigmoid or cecal volvulus who underwent
resection, mean age was higher in the sigmoid volvulus group and
male patients accounted for the majority of cases, whereas female sex
was more predominant in the cecal volvulus (Table 2). Looking at the
age histogram, the incidence of sigmoid volvulus peaked in the mid-
70s for both women and men. This in contrast to the age histogram
of cecal volvulus in which women demonstrated 2 incidence peaks:
1 in the mid-50s and another in the late 70s. Moreover, the frequency
of cecal volvulus in women seems to increase rapidly starting in their
early 30s (Fig. 1).
African Americans were present at a higher frequency in the
sigmoid volvulus group. Patients with sigmoid volvulus group tended
to have higher incidence of comorbidities, such as anemia, conges-
tive heart failure, hypertension, and fluid and electrolyte disorders.
Perhaps, the major difference was in the frequency of diabetes and
neuropsychiatric disorders including dementia and paralysis related
to cerebrovascular events. Comorbidity scores were higher in the
sigmoid group, reflecting a sicker patient population (Table 2).
FIGURE 1. Age histogram per sex for cecal volvulus and sig-
moid volvulus.
Looking at hospital characteristics in the sigmoid and cecal
volvulus groups, we observe that the majority of cases were performed
in nonteaching, urban, and large hospitals. The use of laparoscopy
was similar in the 2 groups; however, conversion rates were higher
in cases of cecal volvulus. A stoma was required in almost half of
sigmoid volvulus cases that underwent resection, whereas this was
much lower in cecal volvulus cases (Table 3).
The management of sigmoid volvulus and the associated out-
comes were similar in different hospital settings. The small differ-
ences observed were in the higher use of laparoscopy and the lower
associated conversion rates when comparing urban hospitals with ru-
ral hospitals. Stoma use and mortality rates were similar in different
hospital settings (Table 3). In cecal volvulus, the use of laparoscopy
was more common in urban than in rural hospitals. Stoma use was
higher in teaching than in nonteaching hospitals and higher in large
hospitals than in medium and small hospitals. Mortality rates were
similar in different hospital settings (Table 3).
Looking at surgical outcomes in sigmoid and cecal volvulus
cases that underwent resection, we note a longer length of stay in
the sigmoid volvulus group and a higher total charge. Mortality was
9.4% for sigmoid volvulus and 6.7% for cecal volvulus. Of note is that
anastomotic complications were high in both groups. The incidence
of respiratory failure, pneumonia, acute renal failure, urinary tract
infection, urinary retention, and deep vein thrombosis was higher in
the sigmoid volvulus group (Table 4).
The LASSO algorithm was applied to cases that underwent
resection for sigmoid volvulus and identified the presence of peri-
tonitis and bowel gangrene as the strongest predictors of mortality.
This was followed by the use of an ostomy and coagulopathy. Other
factors, such as the presence of chronic kidney disease, age more
than 70 years, chronic pulmonary or cardiac disease, and fluid and
electrolyte disorders, were also found to predict mortality (Table 5).
Performing surgery on an emergent or semielective basis and the use
of laparoscopy were not found to affect mortality. Hospital factors
were not found to impact mortality. The area under the curve (AUC)
for the predictive model was 0.74 (Fig. 2).
Looking at cecal volvulus, the LASSO algorithm identified
coagulopathy as the strongest predictor of mortality, followed by age
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
4
| www.annalsofsurgery.com
C
2013 Lippincott Williams & Wilkins
Annals of Surgery
r
Volume 00, Number 00, 2013
Colonic Volvulus in the United States
TABLE 3. Use of Laparoscopy, Conversion Rates, Use of Ostomy and Mortality Per Hospital Type, Location, and Bed Size
Sigmoid Volvulus (N
= 19,220)
Cecal Volvulus (N
= 23,392)
All (%)
Use of
Laparoscopy,
%
Conversion
Rate, %
Stoma
use, %
Mortality,
%
All, %
Use of
Laparoscopy,
%
Conversion
rate, %
Stoma
use, %
Mortality,
%
Hospital type
Teaching
41.5
7.2
18.8
49.8
8.8
41.8
7.0
33.1
14.2
6.1
Nonteaching
57.9
5.9
17.2
49.3
10.0
57.6
7.2
31.5
8.6
7.1
Missing
0.6
0.6
Location
Urban
84.8
6.8
16.8
50.0
9.5
84.5
7.7
33.0
11.1
6.5
Rural
14.6
4.4
28.0
46.5
9.6
14.9
3.8
22.2
10.1
7.4
Missing
0.6
0.6
Bed size
Small
8.7
7.7
23.1
45.9
5.9
9.7
9.3
32.6
8.9
7.0
Medium
17.7
8.2
21.0
50.3
9.4
18.2
9.8
33.3
8.9
7.8
Large
39.9
8.6
15.7
46.4
8.0
44.0
8.6
33.5
12.0
6.2
Missing
33.7
28.1
TABLE 4. Surgical Outcomes for Sigmoid Volvulus and Cecal Volvulus of Patients Who
Underwent Resection
N
Sigmoid Volvulus (N
= 19,220)
Cecal Volvulus (N
= 23,392)
Total charge ($)
80,352 (33,685–90,800)
68,935 (26,712–73,525)
Length of stay, d
15 (8–18)
11 (6–13)
Mortality
Died
9.4
6.7
Missing
0.28
0.04
Postoperative complications
Cerebrovascular accident
0.1
0.2
Cardiac complications
2.7
3.1
Respiratory failure
13.6
11.9
Pneumonia
10.0
7.5
Ileus/bowel obstruction
20.5
19.4
Anastomotic complications
∗
15.8
15.2
Acute renal failure
14.5
11.8
UTI
18.1
8.9
Urinary retention
3.2
2.3
Postoperative bleeding
2.6
3.0
Wound complications
6.3
6.6
DVT
1.0
0.6
Continuous variables such as total charge and length of stay are reported as mean and interquartile range; categorical variables
are reported as percent proportions.
∗
Including anastomotic leak, fistula, and intra-abdominal abscess.
DVT indicates deep vein thrombosis; UTI, urinary tract infection.
more than 60 years. The presence of peritonitis and bowel gangrene,
the use of an ostomy, and several comorbidities listed in Table 5 were
also found to be associated with worse outcomes. Again, the use of
laparoscopy and hospital factors were not found to affect mortality.
Interestingly, female sex, private insurance, and hypertension were
found to be protective against mortality relative to the variables identi-
fied in the model. The AUC for the predictive model was 0.82 (Fig. 3).
DISCUSSION
Colonic volvulus is a rare cause of bowel obstruction in
the United States, accounting only for 1.9% of admissions which
is in the range of previously published studies in the United
States that attributed 1.2% to 20% of all intestinal obstructions to
colonic volvulus.
20–23
The wide variation of incidence observed in
previous studies is due to demographic differences among the study
populations.
24
This was observed in our results with a higher inci-
dence of sigmoid volvulus in African Americans. The latter group
seems to be prone to the development of sigmoid volvulus because
of a long sigmoid mesentery with a narrow stalk.
8,9
This suggests
that anatomical factors are more important than dietary factors in the
pathogenesis of sigmoid volvulus. The incidence of colonic volvulus
in the United States is also lower than in other parts of the world, such
as Africa, Middle East, Russia, India, and Brazil, where colonic volvu-
lus accounts for 13% to 42% of all intestinal obstruction.
5,10,25,26
Most of these data are, however, outdated and more recent data sug-
gest that colonic volvulus seems to be decreasing in parts of the
African continent
27,28
and Middle East
29,30
possibly due to western-
ization of the diet,
31
population migration,
29
or change in etiological
patterns of bowel obstruction.
32
The epidemiology of the different types of colonic volvulus is
also changing. With an aging population in the United States,
33
one
would expect to see increasing numbers of sigmoid volvulus. This,
however, was not the case as yearly admissions for sigmoid volvu-
lus remained stable. The prevalence of diverticular disease in the
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
C
2013 Lippincott Williams & Wilkins
www.annalsofsurgery.com
| 5
Halabi et al
Annals of Surgery
r
Volume 00, Number 00, 2013
TABLE 5. Predictors of Mortality of Patients With
Sigmoid Volvulus and Cecal Volvulus Who Underwent
Resection Based on the LASSO Algorithm
Coefficient
LASSO OR
Sigmoid volvulus
Intercept
− 2.95
0.05
Peritonitis/bowel gangrene/necrosis
0.61
1.84
Stoma use
0.46
1.58
Coagulopathy
0.46
1.58
Chronic kidney disease
0.24
1.27
Age
>70 yr
0.17
1.18
Chronic pulmonary disease
0.13
1.14
Congestive heart failure
0.04
1.04
Fluid and electrolytes disorders
0.01
1.01
Cecal volvulus
Intercept
− 3.45
0.03
Coagulopathy
0.91
2.49
Age
>60 yr
0.71
2.04
Metastatic cancer
0.54
1.71
Chronic kidney disease
0.43
1.53
Congestive heart failure
0.39
1.47
Stoma use
0.35
1.42
Fluid and electrolyte disorders
0.32
1.38
Weight loss
0.20
1.22
Pulmonary circulation disorder
0.17
1.18
Peritonitis/bowel gangrene/necrosis
0.12
1.13
Chronic pulmonary disease
0.07
1.07
Hypertension
− 0.04
0.96
Private insurance including HMO
− 0.08
0.92
Female
− 0.15
0.86
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
).
HMO indicates health maintenance organization; LASSO, least absolute
shrinkage and selection operator; OR, odds ratios.
FIGURE 2. ROC curve and C-statistic (AUC) describing the dis-
criminative power of the mortality predictive model for sig-
moid volvulus cases.
FIGURE 3. ROC curve and C-statistic (AUC) describing the dis-
criminative power of the mortality predictive model for cecal
volvulus cases.
United States and the associated high number of sigmoidectomies
34
may explain this finding. In contrast, we observed an increase in ce-
cal volvulus cases. Although this is neither supported nor refuted by
our data, one may try to explain this observed trend by the parallel
increase of screening colonoscopies as a result of large colorectal
cancer screening campaigns
35,36
and an overall increase in the use of
laparoscopic techniques. Air insufflation during colonoscopy leads
to cecal dilatation and may play a role in the development of cecal
volvulus.
37,38
During laparoscopy, pneumoperitoneum, patient posi-
tioning, lateral tilting of the operating table, and mobilization of parts
of the right colon have been implicated as causative factors in patients
with a mobile cecum.
39,40
Reports from the United States identified nearly equal pro-
portions of cecal volvulus and sigmoid volvulus.
5,13,41
Because of
limitations of the NIS database, these proportions are difficult to
obtain. However, if we consider all cases of endoscopic decompres-
sion for sigmoid volvulus to represent definitive treatment with no
recurrence, and moreover, if we consider cases of detorsion, fixation,
or enterostomy to be all performed for sigmoid volvulus, we find
cecal volvulus to account for at least 38% of all volvulus cases.
Cecal volvulus and sigmoid volvulus demonstrate differ-
ent demographics. While sigmoid volvulus was more common in
elderly men, the majority of patients with cecal volvulus were
younger women. Our findings are in line with previously reported
series.
5,6,13,25,42
Sigmoid volvulus in the United States affects an
older population compared with other countries where it is more
endemic.
43,44
The age histogram shows rising frequencies of cecal
volvulus in women in child-bearing age. This confirms previous re-
ports that described a link between cecal volvulus and pregnancy
when the gravid uterus comes out of the uterine cavity displacing the
cecum, thereby elongating its mesentery and making it more prone to
torsion during or immediately after delivery.
45,46
The incidence peak
of cecal volvulus observed in women in their mid-50s may also be due
to previous pelvic surgery. This corresponds to the age group where
many women in the United States have already had a hysterectomy.
47
Previous pelvic surgical procedures may create a mobile cecum or
lead to postoperative adhesions that may create an axis around which
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
6
| www.annalsofsurgery.com
C
2013 Lippincott Williams & Wilkins
Annals of Surgery
r
Volume 00, Number 00, 2013
Colonic Volvulus in the United States
the cecum can rotate.
42,48,49
The lower incidence of sigmoid volvulus
in women can be explained anatomically by the fact that women have a
capacious pelvis with lax abdominal musculature that can accommo-
date and allow the untwisting of a floppy sigmoid.
50
The previously
observed high incidence of sigmoid volvulus in patients with neu-
ropsychiatric diseases
12,13,24,51–53
and diabetes
54
is also confirmed
by our results and points toward an acquired pathology.
Examining the different management methods in colonic
volvulus, we observed that nonsurgical methods of colonic decom-
pression are not commonly used as the sole treatment measure and
are usually followed by more definitive surgery. Contrast enema in
the management of colonic volvulus was historically used for diag-
nostic and therapeutic reasons.
4,5,13,53,55,56
The rare identified cases
were older patients with multiple comorbidities, who probably were
not good surgical candidates. Colonoscopic decompression which
has a success rate of 70% to 90%
12,29,53,57
is mainly used for sigmoid
volvulus, as it is mostly unsuccessful for cecal volvulus.
5,13,58
It is,
however, considered a temporizing measure allowing surgery to be
performed on either an elective or a semielective basis after correction
of underlying fluid and electrolyte imbalance.
5,30
It is rarely advocated
as a definitive treatment because of the associated high-recurrence
rates of 20% to 70%.
1,4,30,53,57,59,60
Moreover, because the mortality
rate for recurrent sigmoid volvulus presenting emergently can be
as high as 33%,
53
several authors advocate resection as a form of
definitive management after initial colonoscopic decompression.
53,60
Surgical management of colonic volvulus can be broadly di-
vided into resective versus nonresective procedures. The use of non-
resective procedures such as detorsion with or without fixation or
enterostomy seems to be infrequent in the acute setting according
to our results. The reason for these findings are the high recurrence
rates after these procedures which are reported to be 22% to 25% for
detorsion of cecal volvulus,
61,62
20% to 30% for cecopexy,
6,63
9%
to 44% for simple detorsion of sigmoid volvulus,
6,64,65
and 28.5%
for sigmoidopexy.
25,65
In contrast to previously published series, we
found that operative detorsion with or without fixation was used in
younger populations. This may be explained by younger patients’ re-
fusal to have a resective procedure or in certain cases because the over-
all condition of the patient would not permit a colonic resection.
53,66
Sigmoidostomy and cecostomy with the use of a decompression tube
were uncommonly performed. These procedures that were more com-
mon decades ago
67
seem to have fallen out of favor as they are as-
sociated with high morbidity and recurrence rates.
68
Furthermore,
it seems from our analysis that their use is limited to patients not
suitable for resection.
Resective surgery with or without stoma was found to be the
most commonly performed surgical procedure in the acute setting.
Although a right hemicolectomy is usually sufficient to treat cecal
volvulus, the extent of resection for sigmoid volvulus has been the
subject of debate.
69
Recurrence has been reported after resection for
sigmoid volvulus depending on the extent of colonic involvement;
patients whose disease is limited to the sigmoid colon recur less than
those with associated megacolon and colonic atony.
70,71
The other
reason for performing an extended resection is that gangrene can
extend beyond the area of constriction in a patchy and ill-defined
pattern.
43
This likely explains the significant proportion of cases that
required a total or subtotal colectomy in our results.
The high use of stoma during resective surgery for sigmoid
volvulus as observed in our results is in line with previous reports
13,59
and can be explained by the high fecal and bacterial content of the
left colon
72
and the advanced age and comorbidities of patients with
sigmoid volvulus.
59
Although performing a primary anastomosis in
the setting of gangrene may lead to a high rate of anastomotic leak,
high fecal content load should not be the only reason to perform
a stoma, as a primary anastomosis may be safe in this setting.
73
The use of stoma was also observed in right hemicolectomy for cecal
volvulus. Cecal volvulus often presents late and sometimes represents
a diagnostic challenge delaying surgical treatment.
49,68
The thin cecal
wall is especially sensitive to dilatation and perforation, and a delay
in diagnosis and management can be detrimental.
Mortality rates of the different surgical procedures are in line
with previously reported data.
5,6,10,13,57,59,74,75
Compared with older
data, mortality rates did not change significantly. Cases managed la-
paroscopically seemed to have lower mortality rates. However, the
use of laparoscopy was observed in patients with lower comorbidity
scores, and these procedures are usually performed on a semielec-
tive basis after initial endoscopic decompression.
46,76,77
In contrast to
previously published data,
43,78
we did not identify the performance of
semielective surgery after initial endoscopic decompression to pro-
tect against mortality. It may be the case in which this effect may be
masked by other factors that have a higher impact on mortality. Delay-
ing surgery after successful colonoscopic decompression to correct
underlying fluid and electrolyte imbalance has been found to de-
crease mortality rates.
60
The effect of fluid and electrolyte imbalance
on mortality rates was not very pronounced.
The presence of bowel gangrene and peritonitis was a strong
predictor of mortality from cecal volvulus and sigmoid volvulus. In
sigmoid volvulus, this finding alone doubled mortality rates. This was
also seen for cecal volvulus cases to a lower extent. These finding
echo findings from smaller observational studies
6,24,25,43,52,60,62,79,80
;
however, most of these studies were small and did not achieve sta-
tistical significance. The use of stoma was also found as a predictor
of mortality; however, this could be a surrogate to bowel gangrene
and peritonitis. Coagulopathy was another strong predictor of mor-
tality. The definition of coagulopathy in NIS includes the presence
of disseminated intravascular coagulopathy that occurs in the setting
of sepsis and septic shock. This suggests that prompt management
is essential in the management of colonic volvulus.
65
The effect of
age on mortality was less pronounced in sigmoid volvulus than in
cecal volvulus. Again, this is in line with previous findings that noted
higher mortality rates for sigmoid volvulus in patients older than 70
years.
59,79,81
The AUC of the ROC for both sigmoid and cecal volvu-
lus groups reflected good predictive power, and the 10-fold cross-
validation performed ensures that these models can be generalized
beyond the sample analyzed.
The main limitation of this study lies in its retrospective nature
and its inherent biases. The potential for coding errors exist when
using an administrative database.
82
The incidence of volvulus could
be slightly underestimated as we excluded patients who died before
surgery and the rare cases that may have detorsed spontaneously as
previously reported.
13,64,69,79,83
One ICD-9 diagnosis code for colonic
volvulus exists and hence differentiation between cecal volvulus and
sigmoid volvulus in cases of operative detorsion or fixation was not
possible. Mortality rates could be higher than observed, as the NIS
only provides information related to the index hospital stay and hence
30-day mortality and mortality rates are not available. Because each
record in NIS is for a single hospitalization, there could be multi-
ple records for an individual if that individual had several hospital-
izations. This would affect the number of admissions for sigmoid
volvulus that may recur after nondefinitive treatment. The inability
to track individual cases did not allow us to calculate recurrence
rates for sigmoid volvulus cases that underwent initial endoscopic
decompression.
Dostları ilə paylaş: |