| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University Hospital of Geneva, 24, rue
Micheli-du-Crest, Geneva 14 1211, Switzerland.
2 Department of Internal Medicine, University Hospital of Geneva, Geneva 14
1211, Switzerland.
3 Present address: Spitalul Clinic Sf. Spiridon, University Gr. T. Popa, Iasi,
Rumania.
4 Clinic/Policlinic of Visceral Surgery, University Hospital of Geneva, Geneva
14 1211, Switzerland.
Received July 31, 2003;
accepted after revision November 10, 2003.
Address correspondence to P.-A. Poletti
(pierre-alexandre.poletti{at}hcuge.ch).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We retrospectively analyzed the medical files
and CT scans of 312 consecutive patients who were diagnosed as having
diverticulitis on an admission CT report or who had a final diagnosis of left
colonic diverticulitis. Patients who did not undergo nonoperative treatment or
were lost to follow-up (n = 144) were excluded from the study.
Admission CT scans of 168 consecutive patients with a diagnosis of left
colonic diverticulitis who underwent nonoperative treatment and had an
18-month follow-up were reassessed by three radiologists unaware of the
clinical findings. Nonoperative treatment was defined as an attempt to treat
the patient with only antibiotics without scheduling them for elective
(delayed) surgery. Unfavorable outcome was defined as a failure of
nonoperative treatment 18 months after admission that required either surgery
or rehospitalization for antibiotic treatment. The risk of unfavorable outcome
was modeled using logistic regression as a function of sex, age, and CT
criteria including the maximum number of diverticula per 10 cm of colon; the
presence of intraabdominal abscess or extraintestinal gas bubbles (< 5 mm
diameter) or gas pockets (
5 mm); the length and location of the abnormal
colonic segment; the maximum thickness of the colonic wall; the presence of
associated free intraperitoneal fluid; and the extent of fatty
infiltration.
RESULTS. Among these 168 patients, 115 (68%) had an uneventful outcome, but nonoperative treatment failed in 53 (32%). The presence of an abscess (n = 19) or extraintestinal gas pocket (n = 14) were the only CT findings significantly associated with failure of nonoperative treatment. Adjusted odds ratios (95% confidence interval) for failure were 6.18 (1.76–21.68) when an abscess was diagnosed and 4.26 (1.04–17.57) when pockets of free air were observed. Sex and age were not significantly associated with unfavorable outcome of nonoperative treatment.
CONCLUSION. Abscess and pockets of extraintestinal gas 5 mm in diameter or larger correlated with unfavorable outcome of nonoperative treatment. None of the other criteria evaluated were predictive of failure of nonoperative treatment, including bubbles of extraintestinal gas smaller than 5 mm in diameter.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The purpose of this study was to determine whether CT and demographic criteria can be used to predict the outcome of first events of acute left colonic diverticulitis in cases in which nonoperative treatment is chosen after the initial evaluation.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Nonoperative treatment was defined as an attempt to treat the patients with antibiotics only, without scheduling them for elective surgery. A percutaneous drainage procedure was considered nonoperative treatment if it was performed in an attempt to avoid surgery. Of the 312 consecutive patients with a medical discharge diagnosis of diverticulitis, 144 (47%) were excluded from further analysis for the following reasons: No CT was performed at admission (n = 4) because a massive pneumoperitoneum was observed on the abdominal radiograph and the patient's clinical condition required immediate surgery. A surgical procedure was performed within 24 hr after admission (n = 29); 24 of these patients had pneumoperitoneum at admission, three had intraperitoneal abscess (one of them had also colovesical and coloenteral fistulas), and two had CT signs of diverticulitis and severe sepsis. Patients with a second episode of diverticulitis and men younger than 50 years with CT criteria of severe diverticulitis including abscess, extraluminal air, or extraluminal contrast (n = 62) were scheduled for elective surgery on the basis of the recommendation of the American Society of Colon and Rectal Surgery, the Committee of the American College of Gastroenterology, the European Association for Endoscopic Surgery, and other authors [1, 2, 9, 10, 12, 19]. Nine patients had no clinical follow-up available or were lost to follow-up. CT scans were not found on hard copies or on PACS (picture archiving and communication system), and only a radiologic report was available in nine patients. Thirteen patients died of non–diverticulitis-related causes before their 18 months' follow-up, in spite of a successful nonoperative treatment. Eighteen patients showed evidence of diverticulitis on initial CT that was not confirmed by further examination. Their final diagnoses included pseudomembranous or ischemic colitis (n = 4), adenocarcinoma of the large bowel (n = 4), mechanical ileus (n = 2), pelvic appendicitis (n = 1), villous adenoma (n = 1), ovarian carcinoma with extension to the sigmoid colon and peritoneal carcinomatosis at surgery (n = 1), inflammatory bowel disease (n = 1), and abdominal pain of unknown origin (n = 4).
CT
The CT examinations were performed using a single-slice helical CT scanner
(PQ 5000, Marconi Medical System or CT/i HiSpeed, General Electric Medical
Systems). Scanning was routinely performed with IV contrast enhancement using
a power-injected bolus of 140 mL of 240 mg I/mL injected at 3 mL/sec. A
uniphasic IV contrast injection with a scanning delay of 60 sec was used. Oral
contrast material was given (4% ioxithalamate meglumine) at a dose of 300 mL
15 min before scanning and an additional 100 mL immediately before scanning.
Colonic contrast material (4% ioxitalamate meglumine) was systematically
administrated via gravity drip using a soft rubber rectal catheter without
taping the catheter or inflating the balloon. For each patient, the contrast
material volume was determined by the subjective sensation of fullness; this
procedure usually allows an optimal opacification of the left colon. All
initial abdominal CT scans were obtained within 24 hr after admission. CT was
performed from the lung base to the pelvis with 5-mm contiguous sections and a
table speed of 5 mm/sec (pitch, 1).
Image Interpretation
CT images were reviewed and assessed by consensus by three board-certified
radiologists on the basis of a predetermined protocol, without knowledge of
the clinical findings and without consulting the original interpretation. The
maximum number of diverticula seen in the total circumference of the
10-cm-long colonic segment with the highest concentration of diverticula was
reported on a 3-level scale (0–5, 6–10, and > 10 diverticula
per 10 cm). The amount of mesenteric fat was reported as minor, moderate, or
extensive. Specific predictors of diverticulitis were assessed as the maximum
thickness of the bowel wall (in millimeters) of the affected segment (Figs.
1 and
2) reported on a 3-level scale
(< 6 mm, 6–10 mm, and > 10 mm), the maximum length of the inflamed
bowel wall segment (wall thickening or presence of a spiky diverticulum)
reported on a 3-level scale (< 5 cm, 5–8 cm, and > 8 cm), and the
maximum extent of the soft-tissue density in the fat (length, width, and
height in centimeters). If the fatty infiltration was crossed by a bowel
segment, the bowel width was deducted from the measure (Figs.
1 and
3A,
3B); the three variables were
multiplied to obtain a volume of infiltration. A fluid confluence of the
inflammation was also reported. Extraintestinal gas was defined as bubbles
(< 5 mm) (Fig. 2) or pockets
( 5 mm) (Fig. 3A,
3B) of extraintestinal air,
measured in their largest dimension, without wall (not located inside an
abscess). The term "pocket" included all collections of
extraintestinal air larger than 5 mm; if a patient had both bubbles and
pockets of extraintestinal gas, only the presence of the pockets was
considered. Abscess (including number, size, and location) was defined as a
fluid-containing collection, surrounded by a contrast-enhancing rim, that
might or might not contain contrast material or gas
(Fig. 4). Other predictors
included extraintestinal contrast material leak, defined as a spot of
extradigestive contrast medium; free intraperitoneal fluid or confluence of
the fatty infiltration (Fig.
5); or an associated small-bowel ileus.
|
|
|
|
|
|
Assessment of Outcome
Eighteen months after discharge, patients undergoing nonoperative treatment
were interviewed over the phone by one physician. If they were unavailable or
deceased, information was obtained from their referring physician. We asked
the following questions: Had they experienced symptoms of diverticulitis (left
lower quadrant abdominal pain and fever) after discharge from hospital? if so,
did they need to be hospitalized for antibiotic therapy or collecting because
of recurrent diverticulitis? In the latter event, the histologic report was
obtained to exclude other causes. Hospitalization or a diverticulitis-related
surgical procedure was considered to constitute failure of nonoperative
treatment. If the patient had been readmitted for reasons other than relapse
of diverticulitis, nonoperative treatment was considered successful.
Statistical Analysis
A univariate statistical analysis was performed for the CT and demographic
variables to determine their correlation with patient outcomes over the
18-month follow-up period.
We modeled the failure of nonoperative treatment as a function of demographic parameters and CT findings. Predictors used in the model included patients' sex and age (patients were divided as older or younger than the median of 68 years), patients' characteristics observed on CT (amount of abdominal fat and number of diverticula), and diverticulitis-related CT findings (length and thickness of inflamed segment; volume of fat infiltration; and presence of abscess, free air, or liquid). The median value was arbitrarily chosen as the cutoff for the analysis of the volume of fat infiltration. The variables were used to calculate unadjusted odds ratios and introduced in the model to obtain adjusted odds ratios using logistic regression. Odds ratios and p values were computed using SPSS version 11.0 (Statistical Package for the Social Sciences) for Windows (Microsoft). An odds ratio of 1.0 indicates no difference between groups. A predictor was considered statistically significant for p values below 0.05.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Age Distribution Among Excluded Patients with Immediate or Elective Surgery
Of the 294 patients with a discharge diagnosis of diverticulitis, 55 (19%)
were 50 years old or younger and 239 (81%) were older than 50 years.
Ninety-five patients underwent surgery: 33 patients within 24 hr of admission
and 62 assigned for elective surgery without attempting nonoperative
treatment. Of the 95 surgery patients, 26 were 50 years old or younger (47% of
young patients) and 69 were older than 50 years old (29% of older patients).
Five (9%) of the 55 patients 50 years old or younger (four men, one woman) and
28 (12%) of the 239 patients older than 50 (14 men, 14 women) underwent
surgery within 24 hr after admission.
Twenty-one (38%) of the 55 patients 50 years old or younger (16 men, five women) and 41 (17%) of the 239 patients over 50 years old were assigned for elective surgery without attempting nonoperative treatment.
CT Findings
An abscess was found in 19 patients (11%), only one of whom was under 50
years old. Fourteen patients (8%) had a pocket of extraintestinal gas; all of
them were over 50 years. One of the 14 patients had both an abscess and a
pocket of extraintestinal gas. All pockets of extraintestinal gas consisted of
small collections of air entrapped in the mesenteric fat; no free
pneumoperitoneum was found. Twelve patients (7%) had bubbles of
extraintestinal gas. Two of them (17%) were less than 50 years old.
A leak of bowel contrast material was suspected in four patients (2%). In all four patients, this finding was a subject of controversy among observers because of the difficulty in differentiating leaked material from the normal presence of contrast material in a diverticulum of a small sinus tract. The absence of reliable criteria to distinguish the two situations led to a consensus decision to exclude this factor from analysis.
A small bowel ileus was found in two (1%) of the 168 patients; both underwent a successful nonoperative treatment.
Outcome
Nonoperative treatment was successful in 115 (68%) of the 168 patients and
failed in 53 patients (32%). Six (11%) of these 53 patients developed
peritonitis and sepsis during the first week after admission and underwent
immediate surgery. Two had pockets of extraintestinal gas, one had an abscess,
and one had both a pocket of extraintestinal gas and an abscess. Five patients
(9%) did not improve under nonoperative treatment while admitted to the
hospital and were assigned for elective surgery. Two had small abscesses (<
2.5 cm in diameter, too small to undergo a percutaneous drainage); another had
a pocket of free extraintestinal gas. Forty-two (79%) of the 53 patients for
whom nonoperative treatment failed were discharged free of symptoms after
their first episode of diverticulitis but were readmitted for a second episode
later.
Of the 53 patients in whom nonoperative treatment failed, 14 (26%) relapsed during the first month after admission, and 43 (81%) relapsed within 12 months. The median time to relapse was 4.0 months after admission among the 53 with a failed nonoperative treatment and 6.4 months in the group of 42 patients in whom nonoperative treatment failed after they were discharged home. The subsequent treatment for the 53 patients with failed nonoperative treatments consisted of emergent surgery for 12 (23%) in IV antibiotic treatment followed by elective colectomy for 24 (45%), and IV antibiotic treatment alone for 17 patients (32%). The latter 17 patients either refused surgery or were considered to pose unacceptable risks from anesthesia.
Failure of nonoperative treatment was observed in 12 (63%) of 19 patients with an abscess, in eight (57%) of 14 patients with pockets of extraintestinal gas, and in none of the patients with bubbles of extraintestinal gas. The complete healing of the abscess in the seven patients who underwent successful nonoperative treatment was confirmed on follow-up CT. No statistically significant difference appeared in the diameter of the abscesses in the 12 patients with failed nonoperative treatment (mean, 4.03 cm; median, 3 cm) when compared with the diameter of the abscesses in the seven patients with successful treatment (mean, 2.9 cm; median, 3 cm). None of the parameters analyzed helped predict which patient with an abscess would have successful nonoperative treatment. Five (26%) of the 19 patients with an abscess underwent a percutaneous drainage; the remainder did not because the abscess was too small (n = 13) or inaccessible (n = 1). In four (80%) of these five patients, nonoperative treatment failed within 3 months after admission.
All three patients less than 50 years old who had an abscess (n = 1) or extraintestinal gas bubbles (n = 2) refused surgery. The patient with the abscess failed nonoperative treatment 6 months after admission. The two patients with extraintestinal gas bubbles did not relapse during the follow-up period.
CT and Demographic Predictors of Failure of Nonoperative Treatment
Table 1 summarizes the value
of CT parameters and demographic data to predict the patient's outcome using
univariate and multivariate analysis. Of all CT parameters analyzed, only two
were statistically associated with failure of nonoperative treatment: the
presence of an abscess and a pocket of extraintestinal gas (odds ratio of 6.18
[95% confidence interval (CI), 1.76–21.68, p = 0.004] and 4.26
[95% CI, 1.04–17.57, p = 0.045], respectively). Failure of
nonoperative treatment was observed in 63% (12/19) of patients with an abscess
and 57% (8/14) of patients with a pocket of extraintestinal gas.
|
No other CT parameter was predictive of failure of nonoperative treatment at univariate analysis. The finding of extraintestinal gas bubbles did not correlate with the outcome.
With regard to demographic findings, patients 68 years old or older had a higher rate of failed nonoperative treatment (37/88, 42%) than younger patients (16/80, 20%). However, this difference was not statistically significant when other factors were controlled.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Only one of our five patients who underwent percutaneous drainage of an abscess had successful nonoperative treatment. Despite the small number of patients who underwent this procedure, our data suggest that percutaneous drainage does not improve the success of nonoperative treatment beyond the level obtained with antibiotic treatment alone. This finding fits with the fact that percutaneous drainage is generally advocated to obviate immediate surgical resection and to allow a delayed single-stage colectomy rather than to replace surgery [21–23].
The presence of one or more pockets of extraintestinal gas ( 5 mm) was
the second CT parameter to be significantly associated with failure of
nonoperative treatment. The presence of free air seen on CT has already been
reported to be a useful predictor of failure of medical treatment and to
assess the severity of diverticulitis
[5,
6,
12]. However, these series did
not make a distinction between large pneumoperitoneum and smaller amounts of
extraintestinal air. In our series, we could not include patients with free
pneumoperitoneum on CT; they all underwent surgery within 24 hr after
admission. Our results show that failure of nonoperative treatment is
associated with not only pneumoperitoneum but also collections of
extraintestinal air entrapped in the mesenteric fat, if they are larger than 5
mm.
The finding that bubbles smaller than 5 mm were not associated with failure of nonoperative treatment is instructive. Earlier series that reported extracolonic gas to be predictive of failure of nonoperative treatment did not distinguish between bubbles and larger collections of air [5, 6]. To our knowledge, no other series has related the volume of free extraintestinal gas collections to patient follow-up. Our data suggest that the presence of extraintestinal air seen on CT can be categorized in two types: bubbles as large as 5 mm, which seem not to be associated with a poor outcome, and larger air collections. This distinction could help therapeutic decision-making in difficult clinical cases, when surgery versus nonoperative treatment for a patient with extraintestinal bubbles visible on CT is being considered. The distinction is probably more relevant in the treatment of patients less than 50 years old because it has been suggested that these patients should undergo elective surgery when criteria of severe diverticulitis are found on CT [6, 12].
In our study, no other CT parameters correlated with failure of nonoperative treatment. Intraabdominal fluid and the maximum colonic wall thickness are considered among the "disease-dependent criteria" for the treatment decision of diverticulitis [1, 8]. Our data suggest that none of these CT criteria predict the outcome of the nonoperative treatment after a first episode of diverticulitis. Similarly, no association has been established between the severity of the diverticulosis, the length of the abnormal segment, or the extent of fatty infiltration for predicting the outcome of nonoperative treatment. Therefore, our data suggest that the presence of diverticula, bowel wall thickening, and fatty infiltration—which are considered the most frequent CT findings associated with diverticulitis [3, 4]—cannot be used quantitatively to predict patient outcome.
The relationship between patient age and the outcome of diverticulitis is controversial [6, 11, 14, 15, 24]. Elderly patients treated conservatively may be more subject than younger patients to recurrent abdominal pain but not to recurrent confirmed diverticulitis [15]. Other studies have reported that men less than 50 years old have a greater risk than older men for poor outcome as determined by persistent or recurrent diverticulitis, symptomatic colonic stenosis, fistula, or abscess [12]. Surgery has been recommended as treatment for the first episode of diverticulitis in young patients [6, 11].
In our series, age and sex were not correlated with failure of nonoperative treatment although we noted an insignificant trend for more failure in older patients. However, we have to consider that 38% (21/55) of the young patients (< 50 years) were scheduled for elective colectomy whereas 17% (41/239) of older patients underwent surgery. Besides, 76% (16/21) of the young patients who were elected for surgery were men, and 41% (17/41) were men in the group of older patients. Indeed, in our institution, surgery is usually proposed for young men with CT findings of severe diverticulitis. This triage procedure caused a major selection bias between younger and older patients and between men and women because most young men with CT findings suggestive of severe diverticulitis (according to the criteria of Ambrosetti et al. [24]) have been excluded from comparison. We can therefore not assert from our results that age and sex do not play a role in the outcome of diverticulitis. Despite these limitations, these parameters were included in our model to limit confounding factors with regard to other criteria.
Our study design was mainly focused on CT criteria, so we did not evaluate the influence of clinical and laboratory data on patient outcome. Therefore, we cannot say whether they represent confounding factors with regard to CT findings.
Some other limitations of our series should be discussed. First, the
follow-up period for relapse was arbitrarily fixed at 18 months, which
underestimates the true amount of failure of nonoperative treatment. The true
percentage of recurrence of diverticulitis is difficult to assess from the
literature because of the differences in methodologies between series. The
rate of recurrent diverticulitis has been reported from 7–35%, with
follow-up time intervals ranging from 1 to 11 years
[7]. In the current series, the
rate of failure of nonoperative treatment was 31.5% (53/168), which is close
to the higher rate of relapse found in the literature. Most (
82%)
recurrent attacks of diverticulitis have been reported to occur in the first
year after initial treatment
[15]. The remaining failures
of nonoperative treatment were scattered in time, because 90% of patients with
relapse were readmitted in 5 years
[7].
Some biases linked to the retrospective nature of the study must be highlighted. First, some patients (n = 18) were excluded from the study because important information was not available for them. The actual influence of these exclusions on the final data results cannot be assessed. The telephone survey information could also have brought some limitation to the follow-up data collection, especially when patients were not interviewed directly; the physicians who were contacted may have been unaware of some relevant clinical information. Fortunately, because of the short period of follow-up, this problem concerned only a few patients and probably contributed little or no bias.
The large percentage of patients who were excluded from our series probably constitutes the major limitation on our analysis. Indeed, of the 294 patients with a discharge diagnosis of diverticulitis, 126 (43%) were excluded, mainly because they underwent surgery at admission or were scheduled for surgery later. However, the decision to elect patients for immediate or delayed surgery is based on multiple disease-dependent and patient-dependent criteria, including radiologic findings [1]. Therefore, we could not determine the value of CT parameters for predicting failure of nonoperative treatment on all admission CT scans unless the surgeon was unaware of the CT findings, which of course was not possible. CT criteria for predicting severity of diverticulitis have already been established [5, 12, 20] and have probably been taken into consideration in the decision for initial or elected colectomy, so our sample for analysis was already selected using some CT criteria. This selection explains the fact that we did not find any pneumoperitoneum on our CT scans. This bias does therefore not allow us to determine CT findings associated with the clinical severity of the diverticulitis. This has been done by others and requires a different study design.
In conclusion, our data suggest that the presence of an intraabdominal abscess and pockets of extraintestinal gas 5 mm or larger in diameter are the only CT parameters to be statistically predictive of failure of nonoperative treatment in patients admitted for a first episode of diverticulitis. CT findings like the presence of extraintestinal gas bubbles (< 5 mm), peritoneal fluid, and the severity of the bowel wall involvement should not be considered predictive of relapse of diverticulitis any longer. Further prospective studies are required to substantiate these observations and determine how they should be integrated with clinical and biologic criteria in the treatment decision process of patients admitted after a first onset of diverticulitis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
