Magnetic Resonance Imaging of the Peritoneal Cavity among Peritoneal Dialysis Patients, Using the Dialysate as "Contrast Medium"
Friedrich C. Prischl*,
Thomas Muhr,
Eva M. Seiringer*,
Siegfried Funk,
Gert Kronabethleitner*,
Manfred Wallner*,
Wolfgang Artmann and
Reinhard Kramar*
*Third Department of Medicine/Nephrology and Second Department of Radiology, Krankenhaus der Barmherzigen Schwestern vom Hl. Kreuz, Wels, Austria.
Correspondence to Dr. Friedrich C. Prischl, Third Department of Medicine/Nephrology, Krankenhaus der Barmherzigen Schwestern vom Hl. Kreuz, Grieskirchnerstrasse 42, A-4600 Wels, Austria. Phone: +43-7242-415, ext. 2174; Fax: +43-7242-415, ext. 3993; E-mail: friedrich.prischl{at}khwels.at
ABSTRACT. The objectives of this study were to evaluate whetheradequate observation of abdominal pathologic features relatedto peritoneal dialysis (PD) was possible with magnetic resonanceimaging (MRI) under routine conditions, i.e., against the backgroundof the dialysate and without contrast medium. For 16 male andseven female patients (mean age, 51.8 ± 15.0 yr; meanduration of PD, 324 ± 542 d), 25 peritoneal MRI studieswere performed with the intraperitoneal dialysate as usual.Indications were symptoms or combinations of symptoms, suchas leakage or abdominal wall edema (n = 3), bloody dialysate(n = 4), suspected herniation (n = 1), suspected ultrafiltrationfailure (n = 2), and abdominal pain (n = 5), or routine assessmentafter initiation of PD (n = 12). The MRI protocol, which wasperformed with a 1.0-T scanner, consisted of breath-hold, coronaland transverse, T2-weighted, half-Fourier single-shot turbospin-echo sequences, using a standard body-array coil. MRI studieswere well tolerated and successfully completed for all excepttwo patients. Results indicated a leak along the catheter (n= 1), a leak in an umbilical hernia (n = 1), suspected leakage(n = 1), hernias (n = 5, in three patients), intraperitonealadhesions (n = 5, in four patients), a ruptured ovarian cyst(n = 1), and pleural effusions (n = 4). Pathologic findingsunrelated to PD or located extra-abdominally were observed in19 of the 25 studies. The catheter tip position was easily identifiedfor all patients. In conclusion, this first report on peritonealMRI using only dialysate as the "contrast medium" indicatesthat MRI permits detailed observation of all relevant, PD-related,abdominal pathologic features against the dialysate background,thus avoiding system contamination (and thus the risk of peritonitis).
Peritoneal dialysis (PD) is effective in replacing renal functionbut is associated with the risk of certain procedure-specificcomplications. The most common problems are of an infectiousnature (1). Other technique-related complications are herniation,leaks (with fluid accumulation adjacent to the site of the leak),intraperitoneal adhesions after peritonitis, and gradual deteriorationof dialytic efficiency (2). The corresponding clinical presentationsmay include symptoms such as bloody dialysate, pain or abdominaldiscomfort, visible or palpable hernias, dyspnea, soft-tissueswelling, and diminished ultrafiltration.
In the diagnostic evaluation of these PD patients, peritonealimaging may be necessary to confirm a clinically suspected diagnosis,to localize leaks, or to provide detailed information beforesurgical repair of the defect (3). The standard for imagingof peritoneal abnormalities among PD patients is computed tomography(CT) with intraperitoneal administration of contrast medium(4). According to a Medline search, there has been only onereport on magnetic resonance imaging (MRI) of peritoneal abnormalitiesamong PD patients with gadodiamide as the contrast agent (5).Intraperitoneal administration of contrast medium for both CTand MRI is a time-consuming procedure that must be performedby trained personnel under aseptic conditions. In addition,the procedure is associated with the potential risk of germinoculation, followed by peritonitis, and contrast media areexpensive.
In T2-weighted MRI scans, water appears hyperintense, i.e.,bright (white). Therefore, PD fluid, which consists of water,electrolytes, and glucose, may be excellently observed by MRIin a manner comparable to that for CT contrast media. The dialysateshould provide sufficient contrast between the peritoneal surfaceand abdominal organs and thus delineate any PD-related pathologicconditions. To prove this hypothesis, a study was undertakenwith 23 patients who were undergoing chronic ambulatory PD (CAPD),to examine whether adequate observation of the peritoneal cavityand any PD-associated abdominal pathologic features was possiblewith peritoneal MRI, using the dialysate as the only contrastmedium.
Patients
This open observational prospective study was conducted in accordancewith the Helsinki Declaration of Human Rights, and informedconsent was obtained from all patients. Overall, 16 male andseven female patients (mean age, 52.0 ± 15.3 yr; range,27 to 83 yr) were examined according to the protocol describedbelow. At the time of the MRI study, the patients had undergoneCAPD for a mean of 323.6 ± 542.1 d (median, 12 d; range,3 to 2039 d). End-stage renal failure resulted from diabetesmellitus type 2 in five patients, chronic glomerulonephritisand chronic interstitial nephritis in four patients each, vascularnephropathy and unknown causes in three patients each, polycystickidney disease in two patients, and diabetes mellitus type 1and multiple myeloma in one patient each.
Overall, 25 peritoneal MRI studies were performed for the 23patients. Peritoneal MRI was considered to be indicated forall patients with symptoms or combinations of symptoms suchas bloody dialysate, abdominal discomfort and/or pain, suspectedherniation, fluid overflow at the exit site, abdominal walledema, weight gain, or diminished ultrafiltration (Table 1).Twelve consecutive patients who just had begun CAPD and werewithout specific complaints were also asked to participate inthe peritoneal MRI study, to allow us obtain more experiencewith the technique. For four patients with complications ofPD, conventional CT with intraperitoneal contrast medium administration(CT peritoneography) and peritoneal MRI without contrast mediumadministration were performed in parallel, for comparison ofthe techniques.
Table 1. Findings for CAPD patients with PD-related clinical complaints and/or suspected intra-abdominal pathologic conditionsa
All examinations were performed independently by two experiencedradiologists, who were blinded with respect to the clinicalsymptoms of the patients. The results were then compared anddiscussed (in cases of differences). Finally, a diagnostic report,including information on the completeness and quality of thestudy, the position of the PD catheter, the presence or absenceof PD-related pathologic conditions, and non-PD-related abnormalfindings, was recorded for each patient.
Peritoneal MRI
No changes in the PD regimen were made for peritoneal MRI studies.Examinations were performed under daily-life conditions, withoutspecific preparations or extra dialysate exchanges and withthe patient’s usual intraperitoneal fill volume of 2000ml of dialysate.
MRI studies were performed by using a superconducting imagingsystem (Magnetom Impact Expert; Siemens Medical Systems, Erlangen,Germany) with a field strength of 1.0 T. The images were obtainedwith a standard body-array coil. The study protocol consistedof two breath-hold, interleaved, and two interleaved, transverse,T2-weighted, half-Fourier single-shot turbo spin-echo sequences(repetition time, 8.2 ms; echo time, 66 ms). The flip anglewas 160 degrees. The slice thickness was 5 mm and the matrixsize was 128 x 256, with fields of view of 400 mm (axial) and450 mm (coronal). When indicated, a T2-weighted, turbo spin-echosequence (repetition time, 5.520 ms; echo time, 128 ms) wasadded. The flip angle was 180 degrees. The slice thickness was3 mm and the matrix size was 184 x 512, with a field of viewof 400 mm. Assessment of the abdominal (and surrounding tissue)anatomic and pathologic features was made solely against thecontrast medium of the dialysate.
CT Peritoneography
For four patients with pathologic findings, CT peritoneographywas performed in parallel with the peritoneal MRI study, forcomparison. The study protocol for CT peritoneography was describedpreviously (6). In brief, the dialysate was drained completely,and non-contrast-enhanced spiral CT scans of the empty abdomenwere obtained with 10-mm sections from the diaphragm to thepubic symphysis. Next, the catheter extension was disconnected,followed by intraperitoneal instillation of 50 ml of iopentol(Imagopaque, 300 mg iodine/ml; Nycomed Imaging, Oslo, Norway)by a specially trained PD nurse or a physician, under sterileconditions. A new dialysate bag was then connected, and 1000ml of dialysate were instilled. The patient was encouraged tomove and walk about for 30 min, to achieve good distributionof the dialysate/contrast medium mixture. Again, spiral CT scansof 10-mm thickness were obtained from the diaphragm to the pubicsymphysis. Another dialysate exchange was performed immediatelyafter the contrast CT examination, to drain the dialysate/contrastmedium mixture.
The CT studies were also independently evaluated by the tworadiologists. The results of the CT scans and MRI studies werecompared only after evaluation and documentation of the latterresults had been completed.
In 23 of the 25 MRI studies, the examination was successfullycompleted and was well tolerated by the patient. The reasonsfor interruption of the examination were intolerable dyspneain a young female patient with marked pleural effusion (Figure 1),with an inability to remain in the supine position, andintolerable thoracic pain resulting from a splinter from WorldWar II trauma in a male patient.
Figure 1. Coronal, T2-weighted, half-Fourier single-shot turbo spin-echo scan for a 27-yr-old, female, chronic ambulatory peritoneal dialysis (CAPD) patient with dyspnea and overhydration. On the right side, marked pleural effusion can be observed. The arrowheads point to the intact right diaphragm; no leakage was observed during the study. On the left side, dialysate surrounds the spleen and no pleural effusion can be observed.
Comparison of contrast CT scans with the respective peritonealMRI studies with the dialysate as the contrast medium yieldedidentical results. In none of the cases could additional informationthat was not provided by peritoneal MRI be obtained from theCT scans. Examples are presented below.
The detailed indications for peritoneal MRI and the resultsof the studies are outlined in Table 1. Pathologic findingsto explain the clinical symptoms were observed in eight of the13 studies. Leaks were localized exactly for two patients, i.e.,along the catheter and in an umbilical herniation. As demonstratedin Figure 2, peritoneal MRI and contrast CT peritoneographyrevealed similar patterns of fluid distribution for the latterpatient and were able to localize the leak. For a 69-yr-oldfemale patient with abdominal wall edema and bloody dialysatein two consecutive bags, subcutaneous fluid accumulation wasdemonstrated, but no leak could be detected with either peritonealMRI or a contrast CT study. Because there were no other explanations,it was suspected that a small perforation of the peritoneumhad occurred but had spontaneously occluded during the overnightperiod of bed rest between the occurrence of edema and the imagingstudies.
Figure 2. (A) Axial, T2-weighted, half-Fourier single-shot turbo spin-echo scan for an 83-yr-old male patient with abdominal wall edema. At the level of the umbilicus, a small bowel loop can be observed, filled with air in an umbilical hernia (open star). The hernia is surrounded by hyperintense liquid emerging from the abdominal cavity to the subcutaneous tissue (arrow). A peritoneal leak within the umbilical hernia was diagnosed. It should be noted that the soft-tissue edema even reached the paravertebral region (closed stars). (B) Corresponding axial computed tomographic (CT) peritoneographic scan, obtained with the use of intraperitoneally administered iopentol, for the same patient. Contrast medium application revealed no additional findings, compared with magnetic resonance imaging (MRI). The arrowhead points to the leak, and the bubble-like structure within the subcutis corresponds to air within the small bowel loop herniation.
Bloody dialysate was observed for another three patients. Fora 46-yr-old female patient who had undergone two peritonealMRI studies within 10 mo, the bloody appearance of the dialysatecould be attributed to rupture of a simple, right-sided, ovariancyst. Figure 3 demonstrates the enlarged right ovary surroundedby hyperintense dialysate and no cyst (as observed 10 mo earlier).No pathologic conditions explaining the bloody dialysate wereidentified for a 53-yr-old female patient and a 65-yr-old malepatient. The bleeding resolved spontaneously and did not recurthereafter.
Figure 3. Axial, T2-weighted, half-Fourier single-shot turbo spin-echo scan of the lower pelvis for patient 4, obtained 10 mo after the first MRI study. On the right side, the ovary (arrow) is slightly enlarged, with an irregularly shaped, inhomogeneous, cystic structure, and surrounded by dialysate. Rupture of the simple, right-sided, ovarian cyst (detected 10 mo earlier) was diagnosed as the cause of the bloody dialysate.
For a 61-yr-old male patient with a clinically evident, left-sided,femoral hernia, peritoneal MRI revealed a large, left-sided,scrotal herniation containing sigmoid colon and dialysate anda small, right-sided, scrotal herniation. Again, MRI (Figure 4A)and CT peritoneography (Figure 4B) yielded similar results.Furthermore, an abdominal wall scar herniation (after tumornephrectomy) between the 11th and 12th ribs was observed onthe left side (not shown). Because of the risk of hernial ruptureattributable to the elevated intra-abdominal pressure producedby the dialysate in the upright position, both scrotal herniaswere surgically treated. The scar herniation was not treated,and it has not changed to date, under close clinical observation.
Figure 4. (A) Coronal, T2-weighted, half-Fourier single-shot turbo spin-echo scan for a 61-yr-old male patient, showing bilateral scrotal herniation. On the right side, a small hernia can be observed, only partially filled with dialysate (arrowhead). On the left side, a large hernia contains hyperintense dialysate at the bottom and a loop of the sigmoid colon (star). (B) Corresponding coronal, contrast-enhanced, CT reconstruction of the bilateral scrotal herniation. The right-sided hernia contained a larger quantity of dialysate mixed with contrast medium (arrowheads), compared with peritoneal MRI scans obtained the day before, but no additional information was obtained with contrast CT peritoneography.
Leaks could be definitely excluded for two patients with pleuraleffusions and generalized edema (an example is presented inFigure 1). For both patients, restriction of oral fluid intakeand the use of a high-glucose dialysate resulted in improvedultrafiltration, with disappearance of the effusions.
Abdominal pain of unknown origin was the reason for peritonealMRI for four patients. In a 50-yr-old male with recurrent peritonitis,adhesions were found around the spleen. For patient 10, adhesivestructures, possibly resulting from prior adnexitis, were observedin the lower pelvis (Figure 5). For a 56-yr-old female patientwith pain, leiomyoma of the uterus was the only pathologic finding.For the fourth patient, a pancreatic cyst was observed. Forall four patients, analgesic therapy was administered as needed,and the symptoms resolved with time.
Figure 5. Axial, T2-weighted, half-Fourier single-shot turbo spin-echo scan for a 45-yr-old female patient with uncharacteristic pain in the lower abdomen during dialysate exchanges. Small, right-sided, Y-shaped adhesions surrounded by dialysate can be observed (arrow) in the lower pelvis, between the rectum and the pelvic wall. The history of the patient revealed recurrent adnexitis in her twenties.
For the 12 patients without suspected PD-related problems whowere studied shortly after CAPD initiation, no relevant pathologicresults were noted. Only a small, asymptomatic, left-sided,femoral herniation, which has not been surgically treated, wasnoted for a 70-yr-old female patient. For a 41-yr-old male patientwho was markedly overhydrated at the initiation of PD, bilateralpleural effusions were observed.
The overall results for all 25 peritoneal MRI studies, withrespect to catheter localization, extraperitoneal fluid accumulation,intraperitoneal adhesions, hernias, leaks, and non-PD-relatedfindings, are presented in Table 2. The catheter tip could belocalized in all studies except the two that were not completed.The tip was located in the lower abdomen in all patients butwas regularly within the rectouterine/recto-bladder space (Douglas’space) in 16 only patients. Only six patients exhibited no additionalpathologic condition (unrelated to PD or located extra-abdominally).The findings were cholecystolithiasis (n = 6), liver cysts (n= 3), a liver hemangioma (n = 1), pancreatic cysts (n = 3) (observedtwice for patient 2), a hiatal hernia (n = 1), splenomegaly(n = 2), renal cysts (n = 2), nephrolithiasis (n = 1), adrenalneoplasia (n = 2), a right-sided ovarian cyst (n = 1), a uterineleiomyoma (n = 1), diverticulosis of the sigmoid colon (n =1), a small pericardial effusion (n = 1), and a cyst in thesacroiliac joint and an enlarged right bulbourethral gland (n= 1).
For PD patients, the occurrence of bloody dialysate, pain, soft-tissueswelling, or ultrafiltration failure may necessitate observationof the peritoneal cavity and surrounding structures during thediagnostic evaluation (4). Several methods, including sonography,scintigraphy (7), and CT, are used to evaluate intra-abdominal,PD-related problems. CT with intraperitoneal administrationof contrast medium, however, has been demonstrated to be thestandard method (8). Compared with conventional roentgenographyand scintigraphy, contrast CT provides the highest resolutionin the delineation of anatomic details and the demonstrationof extraperitoneal fluid (9). Recently, contrast CT peritoneographywas also used to evaluate the functional surface area of theperitoneum with stereologic methods, which might yield moreinformation on the solute-exchange capacity of the peritoneumin the future (10).
Unfortunately, CT peritoneography is rather time-consuming whenwidely accepted standard protocols are followed (6,8), withplain CT scans being obtained first, followed by instillationof contrast medium under aseptic conditions and acquisitionof another series of CT scans. Specially trained personnel areneeded to perform intraperitoneal contrast medium instillation,to prepare the patients for the examination (4). Recently, MRIusing contrast medium has been reported to offer multiplanarimaging capabilities for the evaluation of PD-related complications(5). However, the same limitations seem to apply to MRI as wellas CT peritoneography when intraperitoneal contrast media areused (5).
There has been little experience with CT peritoneography withoutintraperitoneal administration of contrast media, and the resultsseem to be poor (3). When MRI is used, water is observed ashyperintense with T2-weighted, turbo spin-echo techniques. Therefore,the high signal intensity of water or electrolyte solutionsshould highlight normal and pathologic anatomic features ofthe peritoneal cavity. The successful use of physiologic salinesolution as an intraperitoneal "contrast medium" for the evaluationof patients with peritoneal carcinomatosis has been reported(11).
It seemed reasonable to use T2-weighted, turbo spin-echo imagingwithout intraperitoneal contrast medium administration for theevaluation of PD-related complications among PD patients, becausethe abdomens of these patients are filled with an electrolytesolution to treat renal failure. This first report on peritonealMRI with the dialysis fluid as the only contrast medium clearlydemonstrates the effectiveness of this technique in elucidatingPD-related complications. As shown in 25 studies performed underdaily-life conditions, normal features, as well as all typesof known PD-related, intra- and extra-abdominal, pathologicconditions (Table 2), were excellently demonstrated. No differentor additional information was obtained by CT peritoneographywith intraperitoneal contrast medium administration when bothtechniques were performed for comparison (Figures 2 and 4).
In our still-limited experience, the major benefit of peritonealMRI is that no alterations in PD techniques are required forthe MRI examination. Therefore, there is no need for PD-trainednurses to spend time preparing the patients for contrast studies.When intraperitoneal dialysate serves the purpose of a contrastmedium, system contamination (and thus the risk of peritonitis)is avoided (12). Contrast medium costs are eliminated, and therelatively high x-ray load to which the patients are exposedduring CT is avoided. Although the costs of peritoneal MRI maybe considered a major drawback, this technique may be a valuablealternative to conventional contrast CT peritoneography forpatients with histories of allergic reactions to contrast media.
We conclude that peritoneal MRI using the dialysate as the onlycontrast medium may be a valuable alternative to CT peritoneographyfor effective observation of all noninfectious, PD-related,intra-abdominal complications among PD patients. Important resourcesmay be conserved, because this newly described technique ofperitoneal imaging is performed under daily-life conditions,on an outpatient basis.
Acknowledgments
We acknowledge the skillful help of the technical assistantsin the Radiology Department with the MRI and CT studies andthe dedicated work of our PD nursing team. This work was presentedin part at the 33rd Annual Meeting of the American Society ofNephrology, Toronto, Canada, October 13–16, 2000.
Khanna R, Nolph KD, Oreopoulos DG: Peritonitis and exit-site infection.In: The Essentials of Peritoneal Dialysis, Dordrecht, Kluwer Academic Publishers, 1993, pp 76–88
Bargman JM: Non-infectious complications of peritoneal dialysis.In: Textbook of Peritoneal Dialysis, 2nd Ed., edited by Gokal R, Khanna R, Krediet R, Nolph K, Dordrecht, Kluwer Academic Publishers, 2000,pp 609–646
Twardowski ZJ, Tully RJ, Ersoy FF, Dedhia NM: Computerized tomography with and without intraperitoneal contrast for determination of intraabdominal fluid distribution and diagnosis of complications in peritoneal dialysis patients. Trans Am Soc Artif Intern Organs 36: 95–103, 1990
Cochran ST, Do HM, Ronaghi A, Nissenson AR, Kadell BM: Complications of peritoneal dialysis: Evaluation with CT peritoneography. RadioGraphics 17: 869–878, 1997[Abstract]
Prokesch RW, Schima W, Schober E, Vychytil A, Fabrizii V, Bader TR: Complications of continuous peritoneal dialysis: Findings on MR peritoneography. Am J Radiol 174: 987–991, 2000[Abstract/Free Full Text]
Scanziani R, Dozio MB, Surian M, Rovere G, Magri F: Peritoneography and computerized tomography in CAPD.In: Peritoneal Dialysis,edited by LaGreca G, Ronco C, Feriani M, Chiaramonte S, Conz P, Milan, Wichtig Editore, 1991,pp 85–89
Johnson J, Baum S, Smink RD: Radionuclide imaging in the diagnosis of hernias related to peritoneal dialysis. Arch Surg 122: 952–954, 1987[Abstract]
Litherland J, Lupton EW, Ackrill PA, Venning M, Sambrook P: Computed tomographic peritoneography: CT manifestations in the investigation of leaks and abnormal collections in patients on CAPD. Nephrol Dial Transplant 9: 1449–1452, 1994[Abstract/Free Full Text]
Maxwell AJ, Boggis CRM, Sambrook P: Computed tomographic peritoneography in the investigation of abdominal wall and genital swelling in patients on continuous ambulatory peritoneal dialysis. Clin Radiol 41: 100–104, 1990[CrossRef][Medline]
Chagnac A, Herskovitz P, Weinstein T, Elyashiv S, Hirsh J, Hammel I, Gafter U: The peritoneal membrane in peritoneal dialysis patients: Estimation of its functional surface area by applying stereologic methods to computerized tomography scans. J Am Soc Nephrol 10: 342–346, 1999[Abstract/Free Full Text]
Magre GR, Terk M, Colletti P, Muggia F, Boswell W: Saline MR peritoneography. Am J Radiol 167: 749–751, 1996[Abstract/Free Full Text]
Received for publication March 28, 2001.
Accepted for publication June 29, 2001.
This article has been cited by other articles:
K. Herbrig, D. Reimann, T. Kittner, and P. Gross Dry cough in a CAPD patient
Nephrol. Dial. Transplant.,
May 1, 2003;
18(5):
1027 - 1029.
[Full Text][PDF]
Top
Abstract
Introduction
