|
|
||||||||
Clinical Sciences |
Division of Nephrology,1 Department of Medicine, and Department of Radiology,2 Academic Medical Center, University of Amsterdam, Amsterdam; Department of Radiology,3 Gemini Hospital, Den Helder; Dianet Foundation,4 Utrecht–Amsterdam, The Netherlands
Correspondence to: A. Vlijm, Department of Nephrology, Academic Medical Center, Meibergdreef 9, Room A01-111, 1105 AZ Amsterdam, The Netherlands. a.vlijm{at}amc.uva.nl
| ABSTRACT |
|---|
|
|
|---|
Background: Computed tomography (CT) is often used to
confirm the diagnosis of encapsulating peritoneal sclerosis (EPS) but there is
no consensus on specific CT abnormalities. To establish CT findings
characteristic for EPS, we compared CT findings between EPS patients and
long-term peritoneal dialysis (PD) patients without EPS.
Methods: We included as cases all EPS patients in our
center from 1996 to 2008 that underwent a CT scan at the time of diagnosis.
Controls were all other long-term PD patients (PD duration
4 years)
without EPS that had a CT scan for different reasons. The CT scans were
blindly and independently reviewed by 3 radiologists: 2 abdominal radiologists
with PD knowledge (Observers 1 and 2) and 1 radiologist without PD experience
(Observer 3).
Results: We included 15 EPS patients and 16 controls.
Observer 1 found 6 CT findings that were significantly more often present in
EPS than in controls (p
0.05): peritoneal enhancement,
thickening, and calcifications; adhesions of bowel loops; signs of
obstruction; and fluid loculation/septation. Observer 2 scored almost
identically but Observer 3 scored differently. The sensitivity and specificity
of a combination of specific CT findings were, respectively, 100% and 94% for
Observers 1 and 2, and 79% and 88% for Observer 3.
Conclusion: CT scans showed characteristic
abnormalities that were significantly more often present in EPS patients
compared to long-term PD control patients. CT can be used to confirm the
diagnosis of EPS when experienced radiologists apply a combination of specific
CT findings.
KEY WORDS: Encapsulating peritoneal sclerosis (EPS); computed tomography (CT); case-control study.
Long-term peritoneal dialysis (PD) can lead to functional and morphological changes in the peritoneum (1–4). The most severe complication is encapsulating peritoneal sclerosis (EPS), a rare condition in which the bowel loops are entrapped in a dense cocoon of fibrous tissue (5). The exact mechanism leading to the development of EPS has not yet been identified (6). Risk for EPS increases with duration of PD (7,8) and it frequently occurs after discontinuation of PD (7,9,10). The consequences for patients are enormous: progressive malnutrition due to repeated bowel obstruction, sepsis, and eventually even death in a substantial number of cases (7–12). In addition to duration of PD, exposure to dialysis solutions with high glucose/glucose degradation product concentration is considered a risk factor for the development of EPS (13). Kidney transplantation was recently suggested as another possible risk factor for developing EPS (14). The diagnosis of EPS is based on clinical symptoms in combination with pathological findings and abdominal imaging (15), but these techniques do not allow unambiguous diagnosis.
Although abdominal imaging is used to confirm the diagnosis of EPS, there is no consensus on specific radiological abnormalities. Ultrasonography shows signs of adhesions, small bowel dilatation, and typical thickening of the intestinal wall (16), but interpretation of the images is very dependent on the radiologist and there are no data on reproducibility. Use of computed tomography (CT) to diagnose EPS was introduced in 1988 (17) and the technique has been markedly improved since then (18). Although some case reports describe CT findings in EPS patients (19–23), there are only a few studies that describe larger patient groups (24) and only two studies in which CT scans of EPS patients are compared to those of other PD patients (25,26). In a case-control study by Stafford–Johnson et al., CT findings of 10 EPS patients were compared to those of 71 control patients on PD from 1 month to 7 years (25). These authors concluded that peritoneal calcifications, peritoneal thickening, fluid loculation, and tethering of small bowel loops should be considered diagnostic for EPS. Tarzi et al. recently studied abdominopelvic CT scans of 27 patients with EPS and compared them with CT scans of 15 hemodialysis and 20 PD patients using a scoring system. A strongly significant difference between total CT scan scores at diagnosis of EPS and total CT scan scores of controls was found (26).
Encapsulating peritoneal sclerosis occurs mainly in long-term PD patients. Little is known about the occurrence of CT abnormalities in long-term PD patients without EPS; therefore, a control group should consist of long-term PD patients exclusively, since we cannot exclude that chronic PD per se might cause abnormal CT findings. The aim of the present study was to compare CT findings in EPS patients with CT findings in long-term PD patients without EPS. Our research question was, Are there CT abnormalities that are characteristic for EPS? To answer this question, we designed a retrospective case-control study and included all EPS patients as cases and all other long-term PD patients with an available CT scan as controls.
| METHODS |
|---|
|
|
|---|
REVIEW OF THE CT SCANS
Cases and controls were put in random order and 3 experienced radiologists
were asked to independently and blindly review the CT scans. The radiologists
were not aware of the prevalence of EPS in this study. Two of the radiologists
were abdominal radiologists from our center with, respectively, 15 and 13
years of clinical experience (Observers 1 and 2). Each had seen a small number
of the CT scans previously but the interval to the current reading was at
least 2 years. Therefore, the chances of recall bias were considered
negligible. The third observer was a radiologist with 10 years of clinical
experience but from a small general hospital without PD patients (Observer
3).
Based on literature and in cooperation with these radiologists, we developed a score form on which eight items could be scored as present, absent, or unable to evaluate. The items were peritoneal enhancement, peritoneal thickening, peritoneal calcifications, large bowel wall thickening, small bowel wall thickening, adhesions of bowel loops (e.g., indrawing to the center of the abdominal cavity), signs of bowel obstruction (e.g., bowel dilatation), and fluid loculation, including septation.
STATISTICS
Clinical characteristics of both patient groups are presented as means and
standard deviations. Ages are presented as means with ranges. Differences in
clinical characteristics between the cases and controls were analyzed with
independent sample t-tests. Differences in CT findings were analyzed with
chi-square statistics or Fisher's exact probability tests. To analyze
interobserver variability we determined means of agreement between the 3
radiologists. We combined specific CT findings and calculated sensitivity and
specificity.
| RESULTS |
|---|
|
|
|---|
We were able to include 16 patients as controls: 7 men and 9 women, mean
age 54 (range 32 – 78) years. Mean duration of PD was 62 ± 14
months, which was significantly lower compared to the EPS group (p
0.001). The underlying reasons for the CT scan were peritonitis in 5,
suspicion of abscesses or hematomas in 4, and leakage of peritoneal fluid,
diverticulitis, ultrafiltration failure, abdominal aortic aneurysm, gastric
malignancy, bowel ischemia, and liver ischemia in 1 each. None of the control
patients developed EPS during follow-up to June 2008. Nine patients died due
to other causes, 6 patients were successfully transplanted, and 1 patient was
transferred to hemodialysis.
REVIEW OF THE CT SCANS
A total of 31 CT scans (15 EPS patients and 16 controls) were reviewed.
Almost all scans were spiral CT scans with a slice thickness of 5 – 5.5
mm. Two thirds of all patients had received intravenous contrast medium. The
results of the 3 observers are summarized in
Table 1. The scores are
expressed as the number of positive findings for each specific item, followed
by the number of scans that could be evaluated according to the radiologist.
Figures 1(a) and 1(b) show the
percentage of positive findings for each specific item per patient group and
per observer.
|
|
In the readings by Observer 1, six CT findings were significantly more often present in EPS patients compared to controls: peritoneal enhancement, thickening, and calcifications; adhesions of bowel loops; signs of obstruction; and fluid loculation and/or septation. Observer 2 confirmed most of these findings but did not find a significant difference in peritoneal calcifications between the patient groups. Observer 3 had three CT findings that were significantly more often present in EPS: peritoneal thickening, calcifications, and adhesions. Agreement of scoring when all the separate items were taken into account was 90% between Observers 1 and 2, 75% between Observers 1 and 3, and 75% between Observers 2 and 3. Table 2 shows the percentage of agreement between the 3 observers per item. Differences in scoring among the observers were found mainly for peritoneal thickening and small bowel wall thickening. Localization of the thickened peritoneum was sometimes parietal, sometimes visceral, and, in some cases, both.
|
No single item could distinguish definitely between EPS and controls with the exception of fluid loculation and/or septation according to Observers 1 and 2 [Figures 2(a) and 2(b)]. However, this was present in only one third of the EPS patients. None of the observers found significant differences in large and small bowel wall thickening between the groups; therefore, we combined the remaining six items (peritoneal enhancement, thickening, and calcifications; adhesions of bowel loops; signs of obstruction; and fluid loculation/septation) to calculate sensitivity and specificity of the CT scan for diagnosing EPS. For this calculation, a gold standard was required. Since the diagnosis of EPS was not doubtful in any of the cases and was macroscopically confirmed in 73%, this "certain diagnosis" was used as the gold standard. When the cutoff point for a positive CT scan was arbitrarily set at positively scoring for at least three of six items (and two of five items when no contrast enhancement was used), the sensitivity and specificity were 100% (15/15) and 94% (15/16) for Observers 1 and 2, and 79% (11/14) and 88% (14/16) for Observer 3.
|
| DISCUSSION |
|---|
|
|
|---|
The CT findings that are considered characteristic for EPS were also found in some of the control patients. The "positive predictive value," the most important measure of a diagnostic method, could not be calculated because its value depends on the prevalence of the disease. Therefore, we developed a combination of CT findings and calculated sensitivity and specificity of the CT scan for diagnosing EPS. This combination was associated with very high sensitivity and specificity; however, it was developed in our EPS population and has not yet been tested in other populations.
According to the abdominal radiologists, peritoneal enhancement was very specific to EPS (p < 0.001); therefore, CT with contrast enhancement should be preferred. This enhancement, caused by perfusion of peritoneal vessels, might be due to an increased vascular peritoneal surface area (3). An inflammatory component as a cause of enhancement is not necessarily present in EPS (15).
The weakness of this study is that our control patients had a significantly shorter duration of PD than our EPS patients, although an inclusion criterion was having at least 4 years of PD. This is not surprising since long-term PD is a risk factor for developing EPS (7,8). However, none of the control patients developed EPS during follow-up. Another limitation of this study is that four cases of EPS were not macroscopically confirmed but these patients suffered from severe symptoms that could not be explained by another underlying disease. In 2000, an International Society for Peritoneal Dialysis ad hoc committee emphasized that the value of clinical symptoms of EPS should not be underestimated (15). Summers et al. described in their single-center experience of EPS that there is a spectrum of severity in this condition. The EPS patients were divided into a severe group, in which patients developed any standard surgical indication for laparotomy, and a mild/moderate group (12). In our study we did not find it necessary to divide our EPS patients into different groups because we found CT findings specific to EPS with very high sensitivity and specificity without this distinction.
It is important to have a proper tool to diagnose EPS because it can enable guidance of treatment. We also believe that delayed diagnosis limits the success of therapy and that patient outcome will improve when therapy is given at an early stage. There is some evidence that patients may benefit from immunosuppressive therapy (27,28) and from surgical treatment by surgeons with expertise in adhesiolysis in EPS (29). There is also some evidence that patients can be treated successfully with tamoxifen (12,30–32) and with supportive therapy such as total parental nutrition (8).
To our knowledge, no case-control studies have been published with a similar long-term control group (at least 4 years of PD); therefore, previous studies could not exclude the possibility of having CT abnormalities in long-term PD patients without EPS. In the present study, comparison between EPS patients and long-term PD patients without EPS was made: CT findings characteristic for EPS were found. We conclude that contrast-enhanced CT has high sensitivity and specificity for diagnosing EPS, especially when the scans are evaluated by experienced abdominal radiologists. Future studies should focus on validating the CT scan as a diagnostic tool in other populations of EPS patients. Also, the value of regular CT scanning in long-term PD patients can be studied using the combination of criteria developed in the present study.
| DISCLOSURE |
|---|
|
|
|---|
Received 17 July 2008; accepted 15 October 2008.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
|
|
D. E. Sampimon, A. Vlijm, S. S.K.S. Phoa, R. T. Krediet, and D. G. Struijk Encapsulating Peritoneal Sclerosis in a Peritoneal Dialysis Patient Using Biocompatible Fluids Only: Is Alport Syndrome a Risk Factor? Perit. Dial. Int., March 1, 2010; 30(2): 240 - 242. [Full Text] [PDF] |
||||
|
|
D. E. Sampimon, M. R. Korte, D. L. Barreto, A. Vlijm, R. de Waart, D. G. Struijk, and R. T. Krediet EARLY DIAGNOSTIC MARKERS FOR ENCAPSULATING PERITONEAL SCLEROSIS: A CASE-CONTROL STUDY Perit. Dial. Int., March 1, 2010; 30(2): 163 - 169. [Abstract] [Full Text] [PDF] |
||||
|
|
E. A. Brown COMPUTED TOMOGRAPHIC SCANNING AND DIAGNOSIS OF ENCAPSULATING PERITONEAL SCLEROSIS Perit. Dial. Int., September 1, 2009; 29(5): 502 - 504. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |