sonography: a useful tool to detect the mechanical causes of renal transplant dysfunction

Sonography: A Useful Tool to Detect theMechanical Causes of RenalTransplant Dysfunction

Ronald H. Gottlieb, MD,1 Susan L. Voci, MD,1 Scott P. Cholewinski, MD,1 Donna F. Hartley, BA,1

Deborah J. Rubens, MD,1 Mark S. Orloff, MD,2 Oscar L. Bronsther, MD2

1 Department of Radiology, University of Rochester Medical Center, 601 Elmwood Avenue, Box 648,Rochester, New York 146422 Department of Surgery, University of Rochester Medical Center, 601 Elmwood Avenue, Box Surg.,Rochester, New York 14642

Received 1 October 1998; accepted 3 March 1999

ABSTRACT: Purpose. The purpose of this study was toevaluate the utility of sonography in distinguishingbetween mechanical and nonmechanical causes forrenal transplant dysfunction.

Methods. We reviewed all ultrasound examinationreports (n = 286) for 63 consecutive patients who re-ceived 64 renal transplants. We assessed the sen-sitivity and specificity of different degrees of hydro-nephrosis (mild, moderate, or severe) in detectingurinary tract obstruction; different volumes of new orincreasing peritransplant fluid in detecting urine leaks;different total volumes of peritransplant fluid in pre-dicting significant compression of the transplant; andDoppler vascular criteria for predicting arterial and ve-nous occlusion.

Results. All mechanical complications were de-tected (100% sensitivity) with specificities of 91.9% forureteral obstruction (criterion, moderate hydrone-phrosis), 83.4% for urine leaks (criterion, any new fluidor any increase), 91.4% for fluid collections that com-pressed the transplant (criterion, > 100 ml), and 100%for vascular occlusion (criteria, no flow for arterial oc-clusion; no venous flow and reversal of arterial flowduring diastole for venous occlusion).

Conclusions. Sonography is very useful in dis-tinguishing between mechanical and nonmechani-cal causes for renal transplant dysfunction. It hashigh sensitivity and acceptable specificity in this set-ting. © 1999 John Wiley & Sons, Inc. J Clin Ultra-sound 27:325–333, 1999.

Keywords: kidney, transplantation; kidney, ultraso-nography

The survival of transplanted kidneys has beenextended recently because of a number of fac-

tors. These include improved matching of donorsand recipients, more effective immunosuppres-sion regimens, and better surveillance of thetransplants following surgery, resulting in earliertreatment of complications.1 At our institution, abaseline sonographic examination is obtainedwithin 24 hours of renal transplantation. Follow-up sonography is done, as needed, to exclude amechanical cause for suspected transplant dys-function and to guide transplant biopsy once amechanical cause is excluded. Mechanical causesfor transplant dysfunction include disruptions ofthe vascular or urinary tract conduits to the graftand large surrounding fluid collections, whichcompress the transplant or its collecting system.Sonography must be extremely sensitive in de-tecting the mechanical causes for transplant dys-function to allow for prompt intervention. False-negative results could be catastrophic, resultingin loss of the graft or death of the patient second-ary to transplant infarction, urinary tract ob-struction, or infection. Our transplant surgeonsconsider lower specificity to be acceptable sincefalse-positive findings usually result only in addi-tional confirmatory imaging (eg, nuclear medicinerenal scanning or antegrade pyelography). Once amechanical cause for transplant dysfunction isexcluded, surgeons still rely on transplant biopsyto distinguish between the parenchymal causesfor deterioration in graft function. No imaging orlaboratory examination has been found suffi-

Correspondence to: R. H. Gottlieb

© 1999 John Wiley & Sons, Inc. CCC 0091-2751/99/060325-09

VOL. 27, NO. 6, JULY/AUGUST 1999 325

ciently accurate, before initiation of treatment, indistinguishing between the parenchymal causesof graft dysfunction (eg, rejection, acute tubularnecrosis, drug toxicity, and infection).2–6 The ob-jective of our study was to assess the accuracy ofsonography in detecting mechanical causes fortransplant dysfunction. Specifically, we evalu-ated whether sonography was sufficiently sensi-tive in identifying these complications to be usefulin this setting.

PATIENTS AND METHODS

We performed a retrospective evaluation of all re-nal transplants performed at our institution fromJanuary 1994 through October 1995. Sixty-threepatients (40 male and 23 female patients; meanage, 43 years; age range, 13–65 years) received 49cadaveric and 15 living related kidneys (1 patientreceived 2 kidneys). All diagnostic renal trans-plant ultrasound examinations (n 4 286) fromJanuary 1994 through February 1997 were con-sidered in the analysis. All patients underwentbaseline sonography (generally within 24 hours oftransplantation). Follow-up sonography was doneusually for suspected transplant dysfunction. Themean number of sonographic examinations perpatient was 4.5, with a range of 1–12 per patient.Sixty-two patients (98%) had more than 1 ultra-sound study. Sonographically detected urineleaks and vascular occlusions were confirmed sur-gically. Ureteral obstruction was documented byantegrade pyelography. Significant peritrans-plant fluid collections were those collections(other than those due to a urine leak) of suffi-cient volume to result in transplant dysfunctionthrough compression of the transplant. Drainageof those collections resulted in a return to normalor baseline acceptable renal function.

All ultrasound studies were performed with a128XP ultrasound scanner (Acuson, MountainView, CA) and 3.5–5.0-MHz sector and curved lin-ear phased array transducers using both duplexand color Doppler modes. The renal arteries andveins were traced to their origins at the iliac ves-sels. All peritransplant fluid collections weremeasured. The volume of peritransplant fluid wasdetermined by multiplying the 3 orthogonal di-mensions of each collection and summing the vol-umes of collections if there was more than 1. Theaccuracy of using different volume thresholds fordetecting significant peritransplant fluid was as-sessed. The criterion for urine leaks was increas-ing or new peritransplant fluid. We assessed theaccuracy of detecting leaks using different volumethresholds for increasing peritransplant fluid on

successive ultrasound examinations. The base-line ultrasound examinations were not includedin calculating the accuracy of detecting urineleaks since our criterion was new or increasingfluid on successive studies. Hydronephrosis wasgraded as mild (splaying of the renal pelvis), mod-erate (distention of the pelvis and calices), or se-vere (marked distention of the pelvis and caliceswith thinning of the parenchyma). We assessedthe accuracy of using different degrees of hydro-nephrosis as criteria for obstruction. The criterionfor arterial occlusion was no flow on duplex orcolor Doppler imaging despite maximizing Dopp-ler parameter sensitivity for flow. The criteria forvenous occlusion were no venous flow on duplexor color Doppler imaging and reversed diastolicflow on duplex evaluation of the intraparenchy-mal arteries. All ultrasound examinations wereinitially interpreted by board-certified radiolo-gists with special expertise in ultrasound. For ourstudy, the imaging findings, including measure-ments, were drawn from the reports of the exami-nations.

Only the most recent ultrasound examinationprior to the complication (within 48 hours) wasconsidered in the analysis (n 4 11) of urinarytract obstruction, vascular occlusion, or signifi-cant peritransplant fluid collection. All other ul-trasound examinations (n 4 59) in these patientswere excluded from the analysis of these compli-cations since there was no gold standard to con-firm whether these patients had an ongoing com-plication at the times of these ultrasound studies.We believed it would reduce our accuracy if weassumed these patients either had or did not havea mechanical complication at the times of theseexaminations, since we lacked proof of the pres-ence or absence of ongoing mechanical complica-tions. We had proof of the presence of mechanicalcomplications at the times of the ultrasound ex-aminations these patients had that we includedin our analysis.

Evaluation of the presence or absence of urineleaks depended on detecting increasing peritrans-plant fluid. This required assessing the volume ofperitransplant fluid on the examination within 48hours of documentation of a leak compared withthe volume of fluid on the preceding ultrasoundexamination (n 4 10).

Mechanical complications were assumed not tobe present at the times of the ultrasound studies(n 4 211) in the remaining patients. Transplantdysfunction was confirmed by biopsy to be due toother causes after 33% of these examinations (n4 69). For the remaining ultrasound examina-tions (n 4 142), patients were assumed to have

GOTTLIEB ET AL

326 JOURNAL OF CLINICAL ULTRASOUND

normal transplants at the times of these studies.The transplants were considered normal becausethe patients’ serum creatinine levels returned tonormal (< 1.3 ng/ml) or the patients did well clini-cally following their ultrasound examinationswithout intervention to correct a mechanical com-plication. Patients without mechanical complica-tions were followed for a mean of 144 days (range,1–586 days). Nine patients were followed for lessthan 1 month after their last ultrasound exami-nation. Three of these patients had transplantnephrectomies for significant post-biopsy hemor-rhage (all with severe rejection). Two patientshad transplant nephrectomies for severe rejec-tion. Three patients had transplant failure sec-ondary to severe rejection (n 4 2) or acute tubularnecrosis (n 4 1). One patient died in the imme-diate post-transplant period from causes otherthan transplant dysfunction.

Receiver operating characteristic curves weredeveloped (Epistat; Epistat Services, Richardson,TX) for volume thresholds in detecting urineleaks and significant peritransplant fluid collec-tions and for different degrees of hydronephrosisin detecting urinary tract obstruction. The sensi-tivity, specificity, and accuracy of ultrasound cri-teria in detecting each mechanical complicationwere determined by using patient examinationsas the unit of analysis. The presence or absence ofthe ultrasound parameters that maximized sen-sitivity when considering the examination as theunit of analysis was compared between the groupwith and the group without mechanical complica-tions. The Fisher’s exact test was used to deter-mine if the presence of 1 or more of the ultrasoundparameters on any examination was significantlydifferent in the 2 groups of patients. A p value ofless than 0.05 was considered significant.

RESULTS

There were 16 mechanical complications (25%) inthe 64 transplanted kidneys in our series. The

complications and outcomes are summarized inTable 1. Fourteen patients were affected. Two pa-tients had 2 mechanical complications each. Oneof these patients had drainage of a significantperitransplant fluid collection 13 days followingtransplantation and developed a ureteral obstruc-tion 300 days following transplantation. Theother patient had drainage of a significant peri-transplant fluid collection 11 days after trans-plantation and developed a ureteral leak 30 daysfollowing transplantation.

Five patients had urine leaks, all of which oc-curred early following transplantation (within 30days) and were at the ureter-bladder anastomo-sis. All of these patients required surgical repairof the ureter. Using any increase in peritrans-plant fluid or any new peritransplant fluid (> 0ml) (Figure 1) as a parameter for detecting a leakresulted in 100% sensitivity, 83.4% specificity,and 83.9% accuracy. Accuracy and specificitywere improved by increasing the threshold for de-tection to greater than 50 ml of fluid (Figure 2),but 1 of 5 leaks would have been missed (80%sensitivity, 90.2% specificity, and 89.9% accu-racy).

Ureteral obstruction occurred in 4 patients andwas due to an obstructing stone in 1 patient (Fig-ure 3) and a stricture in 3 patients. Using moder-ate or greater hydronephrosis as a criterion forobstruction resulted in 100% sensitivity with thehighest specificity (91.9%). Using mild hydro-nephrosis as a criterion resulted in a markeddecrease in specificity in detecting obstruction(Figure 4). Only 18 false positives (out of 227 ex-aminations) resulted from using moderate hydro-nephrosis as the threshold for obstruction,compared with 55 false positives using mild hy-dronephrosis as the threshold for obstruction.

Significant peritransplant fluid collectionswere present in 5 patients (4 lymphoceles and 1hematoma). Drainage of these collections resultedin restoration of the patients’ renal function tobaseline or normal. Using 100 ml as a threshold

TABLE 1

Mechanical Complications in Renal Transplants

ComplicationNo. Transplanted

KidneysTime of Occurrence

(Days Post-transplant) Outcome

Urine leak 5 2–30 All salvagedUreteral obstruction 4 15–330 All salvagedSignificant peritransplant

fluid collection5 11–45 All drained with

restoration of renalfunction

Arterial occlusion 1 1 day Loss of graftVenous occlusion 1 1 day Loss of graft

RENAL TRANSPLANT DYSFUNCTION


for detecting significant peritransplant fluid re-sulted in 100% sensitivity with the highest speci-ficity (91.4%; Figure 5). Sonography also detected1 abscess, which was drained successfully. How-ever, we did not include this case with the signifi-cant peritransplant fluid collections in our analy-sis since it did not compromise renal function bymass effect on the transplant. Peritransplantfluid collections (both significant collections andthose that did not compromise transplant func-tion) were generally detected early in the postop-erative period, with 50% of them detected within1 month, 75% detected within 2 months, and 95%detected by 6 months.

Using our criteria for arterial and venous oc-clusion resulted in 1 arterial and 1 venous occlu-sion (Figure 6) being detected, with no false posi-tives. Both patients lost their grafts despite rapidsonographic identification of the vascular occlu-sions.

The ultrasound criteria that resulted in 100%sensitivity in detecting each mechanical compli-cation are given in Table 2, with the accompany-ing specificity and accuracy. One or more of theseultrasound criteria were present on at least 1 ul-trasound study in all 14 kidneys (100%) with me-chanical complications compared with 28 (56%) of50 kidneys without mechanical complications.The presence of ultrasound criteria consideredpositive for a mechanical complication was sig-nificantly different (p 4 0.0013) between the 2groups of patients.

DISCUSSION

Sonography was very sensitive (100% sensitivity)with acceptable specificity (range, 83.4–100%) indetecting all mechanical causes for renal trans-plant dysfunction, justifying our surgeons’ confi-dence in using sonography for this purpose. Oursurgeons frequently use sonography to followtransplant recipients. They consider the cost ofsonography to be small (Medicare reimbursementis $263) compared with the devastating financialcost of the loss of a transplant and the resulting

<

FIGURE 1. Urine leak after renal transplant in a 36-year-oldman. (A) Sonogram shows moderate hydronephrosis andfluid (arrows) surrounding the transplant. The fluid collectionhad increased since the previous examination. (B) Early re-nogram shows a photopenic area (arrows) adjacent to therenal transplant. (C) Delayed renogram shows filling of thephotopenic area with tracer activity (arrows). The photopenicarrea was confirmed at surgery to be a urinoma secondary toa leak at the distal ureter.

GOTTLIEB ET AL


harm to the patient. This philosophy seems rea-sonable if one considers that in our institution theMedicare reimbursement for the hospitalization(including the cost of surgery) for a renal trans-plant is $38,000 and the Medicare reimbursementper hemodialysis session is $275.

Sonography detected all 16 mechanical compli-cations that could have led to loss of the graft (4urine leaks, 5 urinary tract obstructions, 5 signifi-cant peritransplant fluid collections, and 2 vascu-lar occlusions). Early detection of urine leaks andurinary tract obstructions by sonography enabledprompt treatment, and none of the patients withthese complications lost their grafts. Five pa-tients with significant peritransplant fluid collec-tions had the collections drained, restoring trans-plant function. The 2 patients with vascularocclusions (1 arterial and 1 venous) lost theirgrafts despite early detection by sonography.However, early sonographic detection of vascularocclusions has the potential to facilitate timelytreatment of these complications as well.

Documenting the usefulness of sonography indetecting mechanical complications in renaltransplants is important in this era of managedcare to ensure continued reimbursement for thisexamination in the follow-up of transplant pa-tients. Knowledge of the true-positive and false-positive rates of using different ultrasound crite-ria to detect ureteral obstruction, urine leaks,significant peritransplant fluid collections, andvascular occlusion is valuable in planning thesubsequent workup or treatment of these pa-tients. Our analysis quantitatively evaluated de-

fined ultrasound criteria for detecting mechanicalcomplications to assess the usefulness of the mo-dality. The specificity of sonography in detectingthe mechanical causes for transplant dysfunctionwas generally lower than its sensitivity. Our sur-geons, as discussed above, are willing to acceptthis, given the high potential for graft loss with amechanical complication versus false-positivesonographic findings usually resulting only in ad-ditional imaging.

Considering moderate or greater hydronephro-sis to indicate obstruction did not compromisesensitivity (100%) and increased specificity from75.3% to 91.9% compared with considering mildor greater hydronephrosis to indicate obstruction.We have found that mild hydronephrosis is morelikely to be due to nonobstructive causes (eg, over-hydration or bladder distention) than is moderatehydronephrosis, and far more false positives re-sult from using mild hydronephrosis as thethreshold for obstruction. False-positive sono-graphic findings of obstruction can result in ei-ther subsequent nuclear medicine renography orantegrade pyelography. We believe our specificityusing moderate hydronephrosis as a threshold isjustified given the alternative of loss of the trans-plant from urinary tract obstruction.

We evaluated all volume thresholds for signifi-cant fluid collections (large enough to compromisefunction) and found 100 ml to result in 100% sen-sitivity with a specificity of 91.4%. In 2 of the 5patients in whom fluid drainage resulted in im-provement in transplant function, no hydrone-phrosis was present. Renal dysfunction in these

FIGURE 2. Receiver operating characteristic curve evaluating using different volumes of fluid surrounding thetransplant to detect a urine leak. Using any new fluid or any increase in fluid [> 0 ml (cc)] surrounding thetransplant on serial examinations as our criterion for urine leak resulted in 100% sensitivity, 83.4% specificity,and 83.9% accuracy in the detection of urine leaks. We considered this the optimal threshold to detect leakssince the sensitivity was maximal with acceptable specificity. Increasing the threshold to 50 ml (cc) resultedin decreased sensitivity (80%) even though specificity (90.2%) and accuracy (89.9%) were slightly improved.The area under the curve is 0.928, p < 0.0001.



patients most likely resulted from the pressure ofthe fluid on the transplant or urinary tract ob-struction without hydronephrosis. When largefluid collections (> 100 ml) without hydronephro-sis surround a transplant, aspiration may be jus-tified before transplant biopsy to establish (orpossibly eliminate) the etiology for renal dysfunc-tion. When moderate or greater hydronephrosis ispresent with significant peritransplant fluid, anuclear medicine renal scan (if sufficient renalfunction) or antegrade pyelogram may be usefulto exclude urinary tract obstruction or urine leak.

All 5 urine leaks were detected if any increasein fluid or any new fluid surrounding the trans-plant on serial examinations was the criterion for

diagnosis, with a specificity of 83.4%. Increasingthe threshold to consider an examination positivefor urine leak to 50 ml resulted in a decrease insensitivity, with 1 leak being missed, althoughspecificity and accuracy improved. We thus con-sider the optimal ultrasound criterion for detect-ing leaks to be any increase in fluid or anyamount of new fluid surrounding the transplant.Using this criterion maximizes sensitivity withonly a small decrease in specificity in detectingurine leaks compared with using 50 ml as thethreshold (Figure 2). Nuclear medicine renogra-phy is a useful adjunct, with high specificity inconfirming the presence of leaks.7 If insufficientrenal function is present to allow for adequateevaluation by renography, aspiration of the fluidis helpful in detecting the presence of urine.

Sonography, with recent improvements inDoppler capabilities, is accurate in assessingtransplant perfusion.8 Both our cases of vascularocclusion were detected by sonography. No falsepositives for arterial occlusion were found usingthe criterion of no flow on duplex or color Dopplerimaging. No false positives for venous obstructionresulted from using the criteria of reversal of ar-terial flow during diastole and no detectable ve-nous flow. Confirmatory angiography can be ob-tained in patients when sonography identifiespotential arterial or venous occlusion. Our seriesdid not include any patients with renal artery ste-nosis in the transplant, despite the reported fre-quency of renal artery stenosis in up to 10% ofrenal transplant recipients,9–12 so we were unableto evaluate the accuracy of sonography in detect-ing this complication. However, previous workhas demonstrated that sonography is highly ac-curate in detecting arterial stenoses in renaltransplants.13

Our study was limited by its retrospective de-sign and relatively small number of patients.However, we believe that, despite these limita-tions, our analysis still adequately quantitativelyassessed the sensitivity and specificity of definedultrasound criteria in detecting the mechanicalcomplications we evaluated.

We also recognize that all ultrasound findingswere not confirmed by surgery, biopsy, or a sec-ond gold standard examination (eg, angiographyor antegrade pyelography). Confirming the re-sults of all ultrasound examinations would be im-possible in an analysis such as this since not allpatients could safely undergo surgery, biopsy, oran invasive radiologic study. Nevertheless, all ofour mechanical complications were documentedby surgery, antegrade pyelography, or percutane-

FIGURE 3. Obstructing stone in the distal transplanted ureter in a62-year-old woman. (A) Sonogram shows moderate hydronephrosis.(B) Antegrade pyelogram demonstrates a filling defect (arrow) due toan obstructing stone in the ureter.

GOTTLIEB ET AL


ous drainage. Patients without mechanical com-plications were confirmed by biopsy to have an-other cause for transplant dysfunction following33% of examinations. We assumed a mechanicalcomplication was not present at the times of theremaining examinations since these patients didwell clinically on follow-up.

In conclusion, we found sonography to be anextremely useful tool with high sensitivity andacceptable specificity in detecting the mechanical

causes for renal transplant dysfunction. We be-lieve that, owing to its being noninvasive andrelatively inexpensive, sonography should beused early and be the first imaging test to evalu-ate suspected transplant dysfunction. Sonogra-phy can effectively triage patients to undergoinvasive radiologic or surgical treatment of me-chanical complications as opposed to biopsy to es-tablish a parenchymal cause for transplant dys-function.

FIGURE 4. Receiver operating characteristic curve evaluating using different degrees of hydronephrosis todetect urinary tract obstruction. Use of moderate hydronephrosis as the threshold for obstruction resulted inmuch improved accuracy compared with using mild hydronephrosis as the threshold. The area under thecurve is 0.960, p < 0.0001.

FIGURE 5. Receiver operating characteristic curve evaluating using different volumes of fluid surrounding thetransplant to detect a significant peritransplant fluid collection. Significant peritransplant fluid collections wereconsidered those collections that, by mass effect, resulted in transplant dysfunction. Use of 100 ml (cc) as thethreshold maximized sensitivity (100%), with acceptable specificity (91.4%), in detecting significant collec-tions. The area under the curve is 0.971, p < 0.0001.



REFERENCES

1. Amante AJ, Kahan BD. Technical complications ofrenal transplantation. Surg Clin North Am 1994;74:1117.

2. Allen KS, Jorkasky DK, Arger PH, et al. Renal al-lografts: prospective analysis of Doppler sonogra-phy. Radiology 1988;169:371.

3. Don S, Kopecky KK, Filor RS, et al. Duplex Dopp-ler US of renal allografts: cause of elevated resis-tive index. Radiology 1989;171:709.

4. Genkins SM, Sanfilippo FP, Carrol BA. DuplexDoppler sonography of renal transplants: lack of

sensitivity and specificity in establishing patho-logic diagnosis. AJR Am J Roentgenol 1989;152:535.

5. Kelcz F, Pozniak MA, Pirsch JD, et al. Pyramidalappearance and resistive index: insensitive andnonspecific sonographic indicators of renal trans-plant rejection. AJR Am J Roentgenol 1990;155:531.

6. Perella RR, Duerincky AJ, Tessler FN, et al. Evalu-ation of renal transplant dysfunction by duplexDoppler sonography: a prospective study and re-view of the literature. Am J Kidney Dis 1990;15:544.

FIGURE 6. Thrombosis of the renal vein to the transplanted kidney in a 30-year-old man. Doppler arterialtracing shows reversal of diastolic flow (arrow). No venous flow from the transplanted kidney was detected.

TABLE 2

Best Thresholds for Detecting Mechanical Complications*

Complications (n = 16) Criteria Sensitivity Specificity Accuracy

Ureteral obstruction(n = 4)

Moderatehydronephrosis

100% 91.9% 92.1%

Urine leak (n = 5) ↑ > 0 ml† 100% 83.4% 83.9%Peritransplant fluid‡

(n = 5)> 100 ml 100% 91.4% 84.4%

Arterial occlusion (n = 1) No flow 100% 100% 100%Venous occlusion (n = 1) No venous flow

and reverseddiastolic flow

100% 100% 100%

*We considered the best thresholds in detecting mechanical complications to bethose that maximized sensitivity (100%) with acceptable specificity.

†Sensitivity, specificity, and accuracy were determined by using any new fluid orany increase in fluid on serial examinations as the criterion for a urine leak.

‡Significant peritransplant fluid that, by mass effect, compromised function bycompressing the transplant.

GOTTLIEB ET AL


7. Shah AN. Radionuclide imaging in organ trans-plantation. Radiol Clin North Am 1995;33:473.

8. Tublin ME, Dodd GD III. Sonography of renaltransplantation. Radiol Clin North Am 1995;33:447.

9. Hohnke C, Abendroth D, Schleibner S, et al. Vas-cular complications in 1200 kidney transplanta-tions. Transplant Proc 1987;19:3691.

10. Jordan ML, Cook GT, Cardell CJ. Ten years of ex-perience with vascular complications in renaltransplantation. J Urol 1982;128:689.

11. Pallesch J, Novick AC, Braun WE, et al. Vascularcomplications of renal transplantation. Urology1980;16:61.

12. Roberts JP, Ascher NL, Fryd DS, et al. Transplantrenal artery stenosis. Transplantation 1989;4:580.

13. Gottlieb RH, Hartley DF, Rubens DJ, et al. Diag-nosis of renal artery stenosis in transplanted kid-neys: value of Doppler waveform analysis of theintrarenal arteries. AJR Am J Roentgenol 1995;165:1441.



sonography: a useful tool to detect the mechanical causes of renal transplant dysfunction

Documents