AJR 2005; 184:915-919
© American Roentgen Ray Society
Gas Distribution in Intraabdominal and Pelvic Abscesses on CT Is Associated with Drainability
Gladwin C. Hui1,
Joao Amaral1,
Derek Stephens2,
Eshetu Atenafu2,
Philip John1,
Michael Temple1,
Peter Chait1 and
Bairbre Connolly1
1 Division of Image Guided Therapy, Department of Diagnostic Imaging, University
of Toronto, Hospital of Sick Children, 555 University Ave., Toronto, ON M5G
1X8, Canada.
2 Population Health Sciences, University of Toronto, Hospital of Sick Children,
Toronto, ON, Canada M5G 1X8.
Received May 8, 2004;
accepted after revision August 10, 2004.
Address correspondence to B. Connolly
(bairbre.connolly{at}sickkids.ca).
Abstract
OBJECTIVE. Intraabdominal and pelvic abscesses are treated by
percutaneous image-guided drainage, under sedation or general anesthesia. This
study attempts to determine if the CT features of gas distribution are
associated with "drainability." Our premise was that gas may be
trapped deep in a collection as bubbles, if the material is thick. Gas may
rise to the surface if the material is thin, forming either an air–fluid
level or superficial bubbles.
MATERIALS AND METHODS. Patients with intraabdominal and pelvic
abscesses were identified by the interventional radiology database, after
research ethics board approval. Patients without prior CT were excluded. The
imaging and clinical records were analyzed retrospectively. Intracollection
gas distribution was recorded as superficial bubbles, deep bubbles, or
air–fluid levels. Collections were classified accordingly: type 1,
air–fluid levels; type 2, superficial or deep bubbles and
air–fluid levels; type 3, superficial bubbles; type 4, deep bubbles; and
type 5, no gas.
RESULTS. One hundred five abscesses were examined in 61 patients,
ranging in age from 2–17 years. Eight of 8 of type 1, 16 of 16 of type
2, 19 of 21 of type 3, 8 of 13 of type 4, and 43 of 47 of type 5 were
drainable. The abscesses of all patients with an air–fluid level were
drainable. Of abscesses with deep bubbles, 61.5% were drainable, versus 90.5%
of those with superficial bubbles. Of those with superficial gas (superficial
bubbles or air–fluid levels), 95.6% were drainable. In comparison with
superficial gas, abscesses with deep trapped gas were associated with a longer
duration of drainage, longer hospital stay, lower percentage of successful
drainage, and higher percentage of residual collections. The difference is
significant for drainability (p = 0.0048; p = 0.0331 after
statistical adjustment for multiple testing).
CONCLUSION. Distribution of gas in an intraabdominal or pelvic
abscess is associated with drainability. Abscesses with superficial gas
(superficial bubbles or air–fluid levels) have a greater chance of being
drained successfully than do abscesses with deep trapped gas.
Introduction
Percutaneous drainage has become a widely accepted treatment for abdominal
abscesses
[1–6].
Previous studies in the literature have found that the success rate of
percutaneous abdominal abscess drainage is high
[1]. According to results
reviewed by vanSonnenberg et al.
[1], most studies showed a
success rate greater than 85%. In particular, the success rate was found to be
91% in a group of pediatric patients
[2]. A retrospective study by
Hemming et al. [3] found no
significant difference in mortality, morbidity, or length of hospital stay
between percutaneous drainage and surgical interventions. Similarly, Bufalari
et al. [7] reported that the
two methods are equally efficacious in postoperative intraabdominal abscesses.
These authors also stated that percutaneous drainage should be the treatment
of choice because it is less invasive and costly than surgical drainage.
However, few studies have been performed on potential predictors of
drainability of abdominal abscesses, and to the best of our knowledge, none
have studied children. Jaques et al.
[4] reported that only the site
of the abscess has a predictive value, with liver and subphrenic abscesses
having the highest success rates. Successful outcomes have been associated
with abscesses that are postoperative, non-pancreatic, and not infected with
yeast [5]. Benoist et al.
[6] reported that the only
significant predictors for failure of percutaneous drainage were an abscess
diameter of less than 5 cm and absence of antibiotic therapy.
At our center, most intraabdominal or pelvic abscesses in children are
treated by percutaneous drainage. The patients usually require sedation or
general anesthesia for drainage. Many of these children are toxic and ill, and
the sedation or anesthesia itself carries risk. Once the child is sedated,
aspiration is performed, which may or may not yield fluid despite a visible
collection. This problem prompted a study to determine whether the
distribution of gas within an abscess on CT may be associated with
drainability of an abscess. Our hypothesis was that distribution of gas within
the abscess reflects the nature of its contents. Gas may be trapped deep in
the collections as bubbles if the material is thick or viscous. Alternatively,
gas will rise to the surface if the material is thin, forming either an
air–fluid level or superficial bubbles. With this knowledge, we may be
able to inform parents and the referring team before the procedure on the
likelihood of successful drainage. In some patients, this knowledge also may
influence the management plan or the decision to proceed with attempted
drainage.
Materials and Methods
This retrospective study was drawn from patients referred to interventional
radiology for percutaneous abdominal drainage between June 2000 and June 2002
as identified in the Image Guided Therapy Database. Research ethics board
approval was obtained from our institution. Patients' medical records,
radiology reports, and images were reviewed. Patients with abdominal abscess,
prior CT, and attempted drainage were included. Patients without prior CT were
excluded. The study had no size criteria for inclusion of an abscess or for a
drainage procedure.
An abscess was defined as a walled-off collection of pus or infected fluid
with an enhancing wall as seen on CT. The primary outcome studied was
drainability of abscess contents. "Drainability" was defined as
the ability to aspirate liquid contents, either partially or completely, from
an abscess through the puncture needle at the time of the procedure. A
secondary outcome studied was reimaging results. Reimaging results included
resolution (no residual collection on imaging) or residual (any collection
still seen on last follow-up CT, even if small). The third outcome studied was
the need for repeated drainage. Other clinical parameters recorded were the
underlying diagnosis, length of history before procedure, site and size of
abscesses, characteristics of the drained substance, organisms cultured, time
that the catheter was indwelling, time to resolution by imaging, and length of
hospital stay. Time to resolution was recorded according to the date of
imaging showing no collection, even if later than clinical recovery and
improvement. The CT features studied were the distribution of gas, size, site,
and the presence of loculations. The distribution of gas was classified into
five types: type 1, single air–fluid level
(Fig. 1); type 2, deep or
superficial air bubbles with air–fluid level
(Fig. 2); type 3, superficial
bubbles with no air–fluid level (Fig.
3); type 4, deep bubbles with no air–fluid level
(Fig. 4); and type 5, no gas
(Fig. 5). A schematic of the
five types of collections is presented in
Figure 6.
View larger version (92K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —15-year-old boy with appendicitis. Axial CT slice
(contrast-enhanced) through pelvis shows large abscess that has superficial
bubble (arrowheads) and air–fluid level (asterisks)
(type 2).
|
|
View larger version (94K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4. —4-year-old girl with appendicitis. Axial CT slice
(contrast-enhanced) through abdomen shows crescent-shaped abscess on right
side, with several deep trapped bubbles (triple arrowheads) and
appendicolith (single arrowhead) (type 4).
|
|
All aspirations were performed under sonographic guidance, and drainage
catheters were placed under fluoroscopic control. Needle puncture was either
transabdominal for superficial collections or transrectal for deep pelvic
collections. Fluid aspiration was then attempted, and if possible, the abscess
was deemed drainable. The aspirate was sent to the laboratory for culturing.
If no fluid could be aspirated, that is, the abscess was not drainable, then a
drain was not placed. Contrast material was injected under fluoroscopy to
outline the collection. A guidewire was inserted, followed by an
appropriate-sized dilator; an 8–12 French all-purpose pigtail drainage
catheter (Boston Scientific) was then placed in the collection. The catheter
was connected to a drainage bag for further drainage. After the procedure, the
catheter was flushed with sterile saline regularly (5–10 mL 8 times,
hourly). The catheter was removed once the patient became asymptomatic and the
abscess was no longer draining. Patients were imaged again at variable
intervals for evaluation of the success of drainage (i.e., resolution).
Results
Sixty-one patients with a total of 105 intraabdominal and pelvic abscesses
were identified (38 boys [62%] and 23 girls [38%]; age range, 2–17
years; mean, 8.9 ± 4.1 [SD] years; weight range, 10.5–83.2 kg;
mean, 34.5 ± 10.8 kg). Major underlying diagnoses included perforated
nonoperated appendicitis (49%), postappendectomy peritonitis (28%),
postlaparotomy peritonitis (7%), peritonitis of unknown cause (8%), Crohn's
disease (5%), and miscellaneous (3%). The abscesses were in the pelvis (32%),
right lower quadrant (29%), right upper quadrant (14%), left lower quadrant
(12%), and miscellaneous locations (13%). Forty-two patients (68.9%) had a
single abscess, and 19 patients (31.1%) had multiple abscesses. The number of
abscesses per patient ranged from one to seven (mean, 1.7; median, 1). The
average size of the abscesses was 7.5 cm in the craniocaudal axis, 4.0 cm in
the transverse axis, and 4.5 cm in the anteroposterior axis.
The abscesses were categorized into different types
(Table 1). Most showed no gas
(47). In those with gas, the pattern of gas distribution most commonly was
superficial (21), followed by air–fluid level with bubbles (16), deep
bubbles (13), and single air–fluid level (8). All collections with an
air–fluid level were drainable (100%). Less than two thirds of those
with deep bubbles were drainable (61.5%). Cultures of aspirates in the
majority (69%) showed mixed bacterial growth, including Escherichia coli,
Klebsiella species, Staphylococcus aureus, Pseudomonas species,
and Streptococcus species. Single organisms accounted for the
remaining 31%. Drainage needed to be repeated for 25% (2/8) of type 1, 12.5%
(2/16) of type 2, 0% (0/21) of type 3, 23.1% (3/13) of type 4, and 12.8%
(6/47) of type 5. Only one patient in the whole sample required laparotomy
after failed drainage. This one patient corresponds to 7.7% (1/13) of type
4.
The needle used most commonly was 16-gauge (59%) or 18-gauge (29%); other
sizes (14-gauge, 20-gauge, and 22-gauge) accounted for the remaining 12%. In
our study sample, the catheter used was an all-purpose drain that was 8-French
(15%), 10-French (56%), or 12-French (29%). A percutaneous transabdominal
route was used in 69% and a transrectal route in 31%. The length of history
before drainage was 8.77 ± 3.12 days. The average time until removal of
the drain was 8.76 ± 4.61 days. The average resolution time as
determined by imaging was 11.43 ± 8.27 days. The average hospital stay
was 16.4 ± 5.59 days, and the average number of reimaging studies was
1.79 ± 1.18. No patients in the study sample died.
Because types 1–3 have superficial gas, the results from these three
groups were combined into one group, the superficial gas group. The
superficial gas group was then compared to type 4 (deep bubbles with no
air–fluid level). Type 5, with no gas, was not included in this
statistical analysis because gas distribution is irrelevant in these cases.
Using univariate analysis with the chi-square test and Fisher's exact test
(SAS, version 8.02, SAS Institute), the difference in drainability between the
superficial gas group and type 4 was found to be statistically significant
(odds ratio, 13.4375; p = 0.0048). The abscesses of all patients with
any form of air–fluid level were drainable (types 1 and 2). Similarly, a
statistically significant difference in residual collections was found between
the superficial gas group and group 4 (odds ratio, 0.2273; p =
0.0434). Using t tests, the superficial gas group was found to be
associated with a shorter drainage time (p = 0.0169) and a shorter
hospital stay (p = 0.0379). However, despite a trend observed in the
raw data, no statistically significant difference was found in the number of
days to CT-documented resolution (10.71 ± 9.54 for superficial gas
group vs 13.08 ± 9.90 for deep trapped gas group, p = 0.5481)
or the need for a repeated procedure (repeated drainage was not needed for
91.1% of superficial gas group vs 69.2% of deep trapped gas group, p
= 0.1537). No statistically significant difference existed in drainability
between unilocular and multilocular abscesses (odds ratio, 1.275; p =
0.71). The following variables were not associated with drainability: the size
of the abscess (p = 0.50), the site of the abscess (p =
0.64), and the size of the drainage catheter (p = 0.86). Detailed
statistical analysis cannot be applied to the relationship between
drainability and specific type of microorganism because the number of
abscesses with only a single species cultured was small, compared with the
number of abscesses with mixed bacteria. Direct comparison would be
inappropriate in this case. No specific trend was observed for any specific
microorganism. The statistical comparison between the superficial gas group
and the deep trapped gas group is presented in
Table 2. p values were
later adjusted for multiple testing. With p value adjustment, only
the drainability at the time of the procedure was statistically significant
(p = 0.0331). Even though the rest of the variables did not achieve
statistical significance, the results were clinically important and therefore
deserve further investigation.
Discussion
Despite advances in antibiotics and surgical techniques, intraabdominal and
pelvic abscess is still a serious complication of perforated appendicitis,
abdominal surgery, and peritonitis, resulting in morbidity, a prolonged
hospital stay, loss of school days, and considerable costs. Historically, a
mortality rate as high as 30% has been reported
[8–10].
The availability of percutaneous drainage of intraabdominal and pelvic
abscesses since the late 1970s has offered patients a minimally invasive
treatment. With percutaneous drainage, patients can avoid surgery during the
acute inflammatory phase, and elective surgery or interval appendectomy may be
undertaken later. Deveney et al.
[11] reported a decrease in
mortality from 39% to 21% after the introduction of percutaneous drainage.
Initially, only simple unilocular abscesses were drained percutaneously
[12]. Multiple, complex, and
multilocular abscesses are now drained percutaneously
[13]. More recently, tissue
plasminogen activator has been used to promote effective drainage.
Other authors have attempted to predict drainability. Features such as an
abscess diameter less than 5 cm, absence of antibiotic therapy, enteric
fistula, multiple or loculated abscesses, poor wall definition, large
abscesses, presence of necrotic tissues, pancreatic localization, and an Acute
Physiology and Chronic Health Evaluation II score greater than 15 have been
reported as predictors of drainage failure
[12–22].
To our knowledge, this study is the first to investigate the value of gas
distribution as seen on CT as a feature associated with drainability. To our
knowledge, this study is also the first to investigate this value in pediatric
patients. In this study, percutaneous drainage for abdominal abscesses was
highly successful overall. All abscesses with an air–fluid level were
drainable, compared with only 61.5% of those with deep bubbles; 95.6% of those
with any form of superficial gas were drainable, compared with 61.5% of those
with deep bubbles. These rates compare favorably with previously reported
results from vanSonnenberg et al.
[1], for whom most studies had
success rates of around 85–95%.
Deep gas bubbles were associated with a lower chance of successful drainage
(61.5% in group 4). This kind of collection may require repeated aspiration
after some time (23.1% in group 4), and surgical evacuation occasionally was
required if the collection remained undrainable (7.7% in type 4). Deep gas
bubbles with no air–fluid level required longer drainage, a longer
hospital stay, and a higher percentage of residual collections. Even though
these variables did not achieve statistical significance (because of the
relatively small sample size), the results were clinically important and
therefore deserve further investigation. No statistically significant
difference in length of history before drainage, loculations, need for
repeated procedures, or number of reimaging studies performed after a
procedure was found between the superficial gas group and the group with deep
bubbles. Despite a trend observed in the raw data, no statistically
significant difference was found in the need for repeated drainage and the
number of days to CT-documented resolution. In contrast to previous findings
[7,
21], multilocular and
unilocular abscesses did not differ significantly in drainability.
We acknowledge that only one patient in the study sample required
laparotomy after failed drainage. Therefore, it seems that relatively good
patient outcomes were achieved for all types of gas distribution. It is
possible that the current technique for abscess drainage has such a high
success rate that most abscesses can be drained, at least partially if not
completely, and that knowing the pattern of gas distribution does not
necessarily lead to a great increase in the need for laparotomy. However, this
information still is useful to relay to parents and to the referring team up
front. Furthermore, although not precluding attempted drainage, the presence
of deep bubbles may be helpful in deciding treatment options including
possible use of tissue plasminogen activator and in forewarning the
possibility of failed drainage. The reverse is also true—that for those
with an air–fluid level, one can be confident in telling parents there
is a higher chance of success.
The present study agrees with previously reported findings of a
polymicrobial presence in intraabdominal abscesses
[5]. Drainability did not
correlate with the size or site of abscesses or with the catheter size. No
specific trend was observed for any specific microorganism. We appreciate the
fact that only about 55% of abscesses in this study actually had signs of gas,
whether superficial or deep. The present study focused on abscesses in the
peritoneal cavity; one cannot extrapolate to solid-organ abscesses or empyema
in the pleural cavity.
The retrospective nature of the present study is associated with some
limitations. The follow-up reimaging schedule was not standardized. The time
to resolution was therefore a considerable overestimate because of late
imaging (the patient's condition clinically resolved before the follow-up
image was obtained). Therefore, the duration of drainage more accurately
reflected clinical resolution. The current study did not attempt a correlation
with sonographic characteristics because sonography is considered more
operator-dependent in detection of gas. A prospective study would be a logical
next step to correct these limitations. Repeating the study with a larger
cohort of patients from multiple centers to check for a statistically
significant difference in outcomes may be interesting.
In conclusion, the distribution of gas within abscesses as seen on CT is
associated with drainability. All abscesses with an air–fluid level were
drainable. Deep gas bubbles with no air–fluid level have, on average, a
statistically significant lower percentage of successful drainage. Abscesses
with deep gas bubbles also require longer drainage, a longer hospital stay,
and a higher percentage of residual collections than do those with superficial
gas. Even though these results did not reach statistical significance in this
study, the trends observed were clinically important and therefore deserve
further investigation. Interventional radiologists can use this information to
forewarn parents and the referring team of the likelihood of successful
drainage. In certain clinical situations, the information may affect
management, the decision to perform drainage, or the use of tissue plasminogen
activator.
References
- vanSonnenberg E, Wittich GR, Goodacre BW, Casola G, D'Agostino HB.
Percutaneous abscess drainage: update. World J Surg2001; 25:362
-369[Medline]
- Jamieson DH, Chait PG, Filler R. Interventional drainage of
appendiceal abscesses in children. AJR1997; 169:1619
-1622[Abstract/Free Full Text]
- Hemming A, Davis NL, Robins RE. Surgical versus percutaneous
drainage of intra-abdominal abscesses. Am J Surg1991; 161:593
-595[Medline]
- Jaques P, Mauro M, Safrit H, Yankaskas B, Piggott B. CT features of
intraabdominal abscesses: prediction of successful percutaneous drainage.
AJR 1986;146:1041
-1045[Abstract/Free Full Text]
- Cinat ME, Wilson SE, Din AM. Determinants for successful
percutaneous image-guided drainage of intra-abdominal abscess. Arch
Surg 2002;137:845
-849[Abstract/Free Full Text]
- Benoist S, Panis Y, Pannegeon V, et al. Can failure of percutaneous
drainage of postoperative abdominal abscesses be predicted? Am J
Surg 2002;184:148
-153[Medline]
- Bufalari A, Giustozzi G, Moggi L. Postoperative intraabdominal
abscesses: percutaneous versus surgical treatment. Acta Chir
Belg 1996;96:197
-200[Medline]
- Schechter S, Eisenstat TE., Oliver GC, Rubin RJ, Salvati EP.
Computerized tomographic scan-guided drainage of intra-abdominal abscesses:
preoperative and postoperative modalities in colon and rectal surgery.
Dis Colon Rectum1994; 37:984
-988[Medline]
- Fry DE, Garrison N, Heitsch RC, et al. Determinants of death in
patients with intraabdominal abscesses. Surgery1980; 88:517
-523[Medline]
- Skaut T, Nystrom PO, Carlsson C. Severity of illness in
intra-abdominal infection: a comparison of two indexes. Arch
Surg 1985;120:152
-158[Abstract/Free Full Text]
- Deveney CW, Lurie K, Deveney KE. Improved treatment of
intra-abdominal abscess: a result of improved localization, drainage and
patient care, not technique. Arch Surg1988; 123:1126
-1130[Abstract/Free Full Text]
- Gerzof SG, Robbins AH, Johnson WC, et al. Percutaneous catheter
drainage of abdominal abscesses: a five-year experience. N Engl J
Med 1981;305:653
-657[Abstract]
- Halasz NA, van Sonnenberg E. Drainage of intraabdominal abscesses:
tactics and choices. Am J Surg1983; 146:112
-115[Medline]
- Bernini A, Spencer MP, Wong D, et al. Computed tomography-guided
percutaneous abscess drainage in intestinal disease: factors associated with
outcome. Dis Colon Rectum1997; 40:1009
-1013[Medline]
- Levison MA, Zeigler D. Correlation of Apache II score, drainage
technique and outcome in postoperative intra-abdominal abscess. J
Am Coll Surg 1991;172:89
-94
- Malangoni MA, Shumate CR, Hollis AT, Richardson JD. Factors
influencing the treatment of intra-abdominal abscesses. Am J
Surg 1990;159:167
-171[Medline]
- Walters R, Herman CM, Neff R, et al. Percutaneous drainage of
abscesses in the postoperative abdomen that is difficult to explore.
Am J Surg1985; 149:623
-626[Medline]
- Lambiase RE, Deyoe L, Cronan JJ, Dorfman GS. Percutaneous drainage
of 335 consecutive abscesses: results of primary drainage with 1-year
follow-up. Radiology1992; 184:167
-179[Abstract/Free Full Text]
- Goletti O, Lippolis PV, Chiarugi M, et al. Percutaneous
ultrasound-guided drainage of intraabdominal abscesses. Br J
Surg 1993;80:336
-339[Medline]
- Brolin RE, Flancbaum L, Ercoli FR, et al. Limitations of
percutaneous catheter drainage of abdominal abscesses. Surg Gynecol
Obstet 1991;173:203
-210[Medline]
- Levison MA. Percutaneous versus open operative drainage of
intra-abdominal abscesses. Infect Dis Clin North Am1992; 6:525
-544[Medline]
- Heloury Y, Baron M, Bourgoin S, et al. Medical treatment of
postappendectomy intraperitoneal abscesses in children. Eur J
Pediatr Surg 1995;5:149
-151[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
