0022-5347/04/1722-0592/0 Vol.

172, 592–595, August 2004

THE JOURNAL OF UROLOGY® Printed in U.S.A.
Copyright © 2004 by AMERICAN UROLOGICAL ASSOCIATION DOI: 10.1097/01.ju.0000130510.28768.f5

Voiding Dysfunction

15-YEAR EXPERIENCE WITH THE MANAGEMENT OF EXTRINSIC

URETERAL OBSTRUCTION WITH INDWELLING URETERAL STENTS
STEVE Y. CHUNG,* ROBERT J. STEIN, DOUGLAS LANDSITTEL, BENJAMIN J. DAVIES,
DAVID C. CUELLAR, RONALD L. HREBINKO, TATUM TARIN AND TIMOTHY D. AVERCH
From the Department of Urology (SYC, RJS, BJD, DCC, RLH, TT, TDA), University of Pittsburgh Medical Center, and Biostatistics (DL),
University of Pittsburgh, Pittsburgh, Pennsylvania

ABSTRACT

Purpose: We assessed the success of retrograde placement of indwelling ureteral stents in the
management of ureteral obstruction due to extrinsic compression.
Materials and Methods: Between July 1987 and December 2002 adequate followup was
available for 101 patients who underwent primary retrograde ureteral stenting for extrinsic
ureteral obstruction. Mean age at presentation was 61.4 years (range 33 to 90). Chart review was
performed on all patients for primary diagnosis, symptomatology, degree of hydronephrosis,
creatinine levels (baseline, treatment and posttreatment), location of compression, size and
number of stents used, progression to percutaneous nephrostomy tube (PNT), stent failure, days
to stent failure, post-stent therapy and status at last followup.
Results: Mean length of followup was 11 months (range 1 to 127). In 101 patients 138
ureteral units (UU) were stented. Total stent failure occurred in 41 (40.6%) patients and 58
(42.0%) UU. A total of 40 (29.0%) UU required PNTs at a mean of 40.3 days (range 0 to 330)
with 18 PNTs placed in less than 1 week. Cases of stent failure that did not undergo PNT
placement included 18 (13.0%) UU at a mean of 52.4 days (range 3 to 128). A total of 90
(89.1%) patients had metastatic cancer at stenting with 32.2% dead at 5.8 months (range 1
to 32). Univariate and multivariate analyses identified cancer diagnosis, baseline creatinine
greater than 1.3 mg/dl and post-stent systemic treatment as predictors of stent failure.
Proximal location of compression and treatment creatinine greater than 3.11 mg/dl were
marginal predictors of failure on univariate analysis, while proximal location of obstruction
was also marginally significant on multivariate analysis. No predictors were identified for
early stent failure (less than 1 week).
Conclusions: At almost 1 year followup stent failure due to extrinsic compression occurred in
nearly half of treated patients. Analysis of data revealed a diagnosis of cancer, baseline mild
renal insufficiency and metastatic disease requiring chemotherapy or radiation as predictors of
stent failure. Managing extrinsic compression by retrograde stenting continues to be a practical
but guarded decision and should be tailored to each patient.
KEY WORDS: stents, ureteral obstruction, pressure

Proper management of extrinsic compression of the ureter
is a therapeutic dilemma for most urologists due to the na-
ture of the disease process. Historically the 2 options avail-
able for decompression of an obstructed collecting system
have been insertion of an internal ureteral stent or place-
ment of a percutaneous nephrostomy tube (PNT). The proper
timing and use of each modality are complicated by issues of
effectiveness and quality of life.
It has been well documented that stent failure in extrinsic
compression occurs in approximately 44% to 58% of cases.1–3
Previous reports have discussed various factors as predictors
of stent failure, but the relatively small number of patients
and short followup have not made the management issues of
extrinsic compression any clearer. By reviewing our 15-year
experience with patients who underwent retrograde ureteral
Accepted for publication February 27, 2004.
* Correspondence: 8635 W. 3rd St., Suite 1070, Los Angeles, Cal-
ifornia 90048 (telephone: 310-423-4700; FAX: 310-423-4711; e-mail:
chungsy@upmc.edu).
592
stenting to treat extrinsic obstruction, and performing uni-
variate and multivariate analyses, we identified predictors of
early and late failure, and provided a more definitive treat-
ment algorithm.
MATERIALS AND METHODS
In 15 years a total of 157 patients underwent primary
retrograde ureteral stent placement for the management of
compression at the University of Pittsburgh Medical Center
(Presbyterian Hospital, Montefiore Hospital, Shadyside Hos-
pital and Magee Woman’s Hospital). Ureteral obstruction
was defined as the presence of hydronephrosis confirmed by
computerized tomography, ultrasound or excretory urogra-
phy, and the presence of flank pain or increased serum cre-
atinine (crt), or both symptomatology and increased crt. Pa-
tients without documented hydronephrosis were excluded
from study. The causes of the extrinsic compression were
determined by history, and pathological and radiographic
 MANAGEMENT OF EXTRINSIC URETERAL OBSTRUCTION WITH INDWELLING URETERAL STENTS 593
TABLE 1. Distribution according to cause of extrinsic compression patients with pain) divided by the odds of stent failure in the
and success of treatment with ureteral stent unexposed category (eg asymptomatic). Lastly, a graphical
No. Stent No. Successful depiction of stent failures and successes over time was ana-
Extrinsic Compression
Failures Stents lyzed using the chi-square and Jonckheere-Terpstra tests.
Malignant cause:
Colon 7 13 RESULTS
Rectal 7 5
Breast 8 5 Of 157 patients who underwent attempted retrograde
Lymphoma 6 5 placement of internal stents 101 (57 women, 44 men) with a
Prostate 3 2 mean age of 61.4 years (range 33 to 90) had adequate fol-
Ovarian 3 2
Cervical 0 4
lowup data. A total of 64 patients had unilateral and 37 had
Uterine 3 1 bilateral ureteral involvement for a total of 138 extrinsically
Lung 0 2 obstructed ureteral units (UU). Of 138 UU cystoscopic inser-
Renal 1 1 tion of ureteral stents was successful in all but 7 (95%)
Pancreatic 0 2
Bladder 0 2
ureters. Patients with bilateral stents initially contributed 2
Gastric 0 2 observations to the data set but were found to have essen-
Post-transplant lymphoproliferative disease 0 1 tially repeat data for the 2 stents. Therefore we used data
Testis 0 1 from the first stent only for analysis (resulting in a total
Melanoma 0 1
Carcinoid 1 0
sample size of 101).
Small bowel 1 0 Of 101 patients 41 (40.6%) and of 138 UU 58 (42.0%) had
Peripheral nerve sheath 0 1 stent failure at mean of 11 months (range 0 to 127) of fol-
Total 40 50 lowup. Of the 58 stent failures 25 (43.1%) UU failed within 6
Nonmalignant cause: days. There were 27 patients (26.7%) who had 40 (29.0%)
Retroperitoneal abscess/bleed 0 4 PNTs placed at a mean of 40.3 days (range 0 to 330) due to
Retroperitoneal fibrosis 0 3
Uterine fibroid 1 1
stent failure. Of the 40 PNTs 18 were placed in less than 1
Vascular graft bifurcation 0 1 week. Patients with stent failure who did not undergo PNT
Abdominal aortic aneurysm 0 1 placement included 18 (13.0%) UU at a mean of 52.4 days
Total 1 10 (range 3 to 128). Reasons for not performing PNT procedures
included patient transfer to hospice care or family wishing no
further medical intervention.
Table 1 shows the frequencies of stent failures by specific
data were acquired by chart review (table 1). Other data causes of extrinsic compression. Individually none were sig-
obtained included patient age, sex, symptomatology, degree nificantly associated with stent failure (p values ranging
of hydronephrosis, location of extrinsic compression, stent from 0.09 to 0.65). Although individual causes did not lead to
size, number of stents placed, creatinine levels (baseline, any significant associations, malignancy as a whole (versus
treatment and posttreatment), post-stent therapy, stent fail- nonmalignant causes) showed a significant correlation with
ure, days to stent failure, PNT placement and status at last stent failure (p ⫽ 0.02). Only 56% (50 of 90) of the subjects
followup. with malignancy were treated successfully with stent place-
Stent failure was defined as persistent hydronephrosis ac- ment, whereas almost 91% of the nonmalignant cases were
companied by flank pain or persistently increased serum successfully treated with stents. Most patients with benign
creatinine in patients presenting with symptomatology or disease had intervention of the primary disease process.
renal failure, respectively. Stent failure in patients present- In identifying predictors of stent failure, patient and stent
ing with hydronephrosis, and renal failure and pain were specific variables were analyzed, and the results are depicted
evaluated individually depending on the clinical scenario.
Furthermore, in patients with bilateral ureteral obstruction
or obstruction of a solitary functioning kidney, azotemia and
anuria were considered stent failures. The impossibility of TABLE 2. Analysis of patient and stent specific categorical
retrograde stent placement due to severe external compres- variables for all stent failures
sion was also characterized as stent failure. In some cases Variable
No. No. Odds Ratio
renal scans or voiding cystograms were performed to confirm Failures (%) Successes (p value)
unremitting obstruction. Stent failure within 6 days but not Gender:
including unsuccessful stent placement was designated early Male 17 (38.7) 27 1.16 (0.88)
Female* 24 (42.1) 33
stent failure. Patients who underwent antegrade ureteral stent Diagnosis:
insertion after initial management with PNT placement were Ca* 40 (44.5) 50 8.00 (0.03†)
excluded from the study. No Ca 1 (9.1) 10
Associations between each outcome (total and early stent Symptom:
Pain* 11 (30.6) 25 0.51 (0.19)
failure) and each categorical predictor variable were assessed Asymptomatic 30 (46.2) 35
via contingency tables, chi-square tests and Fisher’s exact Hydronephrosis:
tests. For each continuous variable (age, baseline, treatment Mild 2 (25.0) 6 1.77 (0.35)
and posttreatment creatinine) summary statistics were given Moderate 19 (21.4) 33
Severe* 20 (39.2) 21
to describe the distributions separately within the failure Location:
(either any failure or early failure) and the nonfailure groups. Proximal 24 (50.0) 24 0.47 (0.10)
Significance tests were conducted using the likelihood ratio Distal* 17 (32.1) 36
test with logistic regression. Logistic regression was then Laterality:
Unilateral 25 (38.5) 40 1.28 (0.71)
used (for both outcomes) to form a multiple regression model Bilat* 16 (44.5) 20
of all variables that showed at least marginal significance Size (Fr):
(p ⫽ 0.15) in the previously described univariate analysis. 6 26 (36.1) 46 1.01 (0.82)
Fisher’s exact test was also used to test associations between 7* 8 (36.4) 14
Post-stent systemic treatment:
stent failure and each cause of extrinsic compression. To No 15 (30.0) 35 2.43 (0.05†)
quantify the strength of associations further the odds ratio Yes* 26 (51.0) 25
was calculated for each variable of interest. The odds ratio * Treated as the exposure category for calculating the odds ratio.
measures the odds of stent failure in the exposed category (eg † Significant.
594 MANAGEMENT OF EXTRINSIC URETERAL OBSTRUCTION WITH INDWELLING URETERAL STENTS
TABLE 3. Analysis of patient and stent specific continuous
variables for all stent failures
Variable/Failure Mean (SE) Odds Ratio (p value)
Age:
No 61.1 (1.58) 1.05 (0.75)*
Yes 61.9 (1.88)
Baseline crt:
No 1.00 (0.08) 2.07 (0.03†)
Yes 1.30 (0.12)
Treatment crt:
No 2.24 (0.31) 1.17 (0.07)
Yes 3.11 (0.36)
Posttreatment crt:
No 1.44 (0.18) 1.10 (0.55)
Yes 1.59 (0.17)
* Odds ratio for 10-year difference in age.
† Significant.
in tables 2 and 3. Only diagnosis of cancer, baseline creati-
nine greater than 1.30 mg/dl and presence of post-stent sys-
temic treatment (chemotherapy or radiation) were signifi-
cantly associated (p ⫽ 0.03, p ⫽ 0.03 and p ⫽ 0.05,
respectively) with failure status. Diagnosis of cancer showed
a particularly strong association with stent failure, yielding
an odds ratio of 8.0. Baseline creatinine and post-stent sys-
temic treatment yielded an odds ratio greater than 2 (2.07
and 2.43, respectively). Treatment creatinine greater than
3.11 mg/dl (p ⫽ 0.07) and proximal location of compression
(p ⫽ 0.10) showed a marginally significant association with
failure. Although posttreatment creatinine levels were
slightly higher in cases of failure, the association was not
statistically significant (p ⫽ 0.55). Furthermore age
(p ⫽ 0.75), gender (p ⫽ 0.88), symptomatology (p ⫽ 0.19),
degree of hydronephrosis (p ⫽ 0.35) and stent size (p ⫽ 0.82)
lacked any statistical association with failure status. Of the
11 ureters with 2 ipsilateral stents, 4 (36%) progressed to
failure and were statistically insignificant.
A multiple logistic regression was used to predict failure as
a function of diagnosis, location of obstruction, post-stent
systemic therapy, and baseline and treatment creatinine. In
this multiple regression model a diagnosis of cancer
(p ⫽ 0.02), post-stent systemic treatment (p ⫽ 0.02) and
baseline creatinine greater than 1.30 mg/dl (p ⫽ 0.004) were
all significantly associated with failure. Location of obstruc-
tion was only marginally significant (p ⫽ 0.06).
A total of 13 patients (7 women, 6 men) experienced early
stent failure (31.7% of total patients with failed stents). Uni-
variate analysis was performed but no predictive factors of
early stent failure were identified. The lack of statistical
associations for these variables may be due to the relatively
FIG. 1. Graph depicting stent failure and successes with time
small number of cases. Since none of the variables were
significantly associated with early failure, multiple regres-
sion models were not used for this outcome.
Significance of any changes in the success rate across 3 (5
to 6-year) calendar periods was assessed (fig. 1). The chi-
square test for any association between time periods and
success rates showed no significant results (p ⫽ 0.46). Simi-
larly the Jonckheere-Terpstra test for trend (across the or-
dered time periods) was not significant (p ⫽ 0.29).
DISCUSSION
Obstruction of urine flow can result from an intrinsic
source or extrinsic compression. It has been well documented
that stent placement relieves almost all obstruction due to
intrinsic pathology (eg stones) but only half of obstructions
due to extrinsic compression. The rationale behind this dif-
ficulty in bypassing an obstruction due to extrinsic pathology
is only partially understood. It is generally accepted that flow
through a stent is compromised when the stent lumen be-
comes plugged with mucus and debris, and periureteral stent
flow is impeded when an external mass compresses, kinks
and obliterates the space around a stent.4, 5 This combination
leads to the simultaneous obstruction of urine through and
around the stent, resulting in stent failure.
Experience with the ureteral stent was first reported in
1978 and, subsequently, stents were used for patients with
cancer with external compression of the ureter.6, 7 Success
was defined as cancer survival and improved quality of life,
and limited data prevented any analyses of stent failures. A
decade later Docimo and Dewolf reported on the first series of
patients with stent failure due to obstruction from extrinsic
compression and concluded that side holes and smaller stent
sizes were potential reasons for failure.1 Subsequently stent
failure was formally investigated in a prospective series with
3 months of followup.2 In this study the number of patients
was small (36) and a third underwent PNT placement weeks
before antegrade stent insertion. Using a multivariate anal-
ysis the authors reported the only predictor of stent failure
due to extrinsic compression was the degree of hydronephrosis.
We report a 40.6% primary stent failure rate in 101 pa-
tients with extrinsic compression after 11 months of fol-
lowup. Performing univariate and multivariate analyses we
determined that the 3 predictors of stent failure were cancer
diagnosis, baseline creatinine greater than 1.3 mg/dl and
post-stent therapy necessitating radiation or chemotherapy.
A diagnosis of cancer, and not the specific type of cancer, was
a strong predictor of failure. Patients who had malignancy
and post-stent systemic treatment (radiation or chemother-
apy) also had higher stent failure rates. We base our findings
on the selection of patients with advanced disease and the
generalized marginal response rates seen in systemic ther-
apy for many types of metastatic cancer. Ureteral obstruction
due to metastatic disease is an ominous sign8 and if extirpa-
tive surgery is not possible, a high likelihood of stent failure
exists. Benign diseases which were often transient and not
progressive were almost always successfully treated with
retrograde stenting. Often resolution of inflammation, drain-
age of abscess or resolution of hematoma contributed to the
success of stenting.
A baseline creatinine greater than 1.3 mg/dl was also
shown to be predictive of failure on univariate and multivar-
iate analyses. Patients with malignancy and weight loss have
many reasons for poor baseline renal function and after in-
sidious obstruction, the renal reserve may be insulted and
less likely to recover. This mechanism is alluded to in a
report on patients with malignant ureteral obstruction ben-
efiting from immediate nephrostomy tube placement for re-
covery of renal function and then antegrade stent insertion.9
We found no correlation with stent size and failure. This
finding is in agreement with an in vivo study of different
 MANAGEMENT OF EXTRINSIC URETERAL OBSTRUCTION WITH INDWELLING URETERAL STENTS
FIG. 2. Proposed algorithm using our predictors of failure for ef-
fective management of extrinsic compression.
stent diameters (eg 6 or 8Fr) not showing any appreciable
difference in pressure flow relationships.10 We also did not
find double stenting to be of value on long-term followup.
Most double stented ureters were performed with 2, 4.7Fr or
a 6Fr and a 4.7Fr stent. It is possible that with larger stents
success at long-term followup may be more likely.11, 12
Our data did not demonstrate a significant predictive role
for degree of hydronephrosis and demonstrated only a mar-
ginal role for proximal level of obstruction. These findings
may be attributed to the subjective nature of grading hydro-
nephrosis, especially moderate and severe cases in which our
data failed to expose any significance. However, the marginal
significance of proximal location of obstruction as a predictor
of stent failure agrees conceptually with the findings of
Mardis et al.13 They demonstrated that decreased flow was
found to a greater degree in ureters with simulated proximal
rather than distal obstruction. Resistance to flow is associ-
ated with diameter and length, and if the obstruction is
closer to its origin resistance to flow is increased.
With an 11-month followup we were able to identify 4
failures at more than 5 months after stent placement. These
cases demonstrated progression of metastatic disease. Al-
though most failures occur early, this disease progression
shows that stent failure may develop at any point throughout
the course of the disease process in patients with extrinsic
compression. After identifying patients in whom internal
stents are likely to fail, treatment options included place-
ment of PNT. However, nephrostomy tubes require cumber-
some external drainage bags, are prone to dislodgement and
decrease quality of life.7 Consideration of an alternative to
stents may be reasonable. The metallic wallstent with or
without a temporary internal stent has been successfully
used in patients with extrinsic compression.14 –16 The use of
more stiff “tumor” stents is an option, and although success
has been reported the size of the stent can make it difficult to
insert.17 An alternative to PNT placement is a subcutaneous
nephrovesical stent which provides drainage of urine from
the renal pelvis to the bladder without traversing the ob-
structed ureter.18
To our knowledge we report the largest series to date in the
595
management and outcome of retrograde ureteral stenting for
extrinsic compression. Weaknesses of our series include the
retrospective nature and difficulty in collecting data on stent
stiffness for analysis. It has been suggested that stiffness
may have a crucial role in maintaining ureteral patency and
may circumvent our predictors of failure. Contemporary
stents may also contribute to successful stent function as
suggested by our time graph, but individual stent character-
istics could not be evaluated.
CONCLUSIONS
Ureteral obstruction due to extrinsic compression should
be managed promptly with internal stent or PNT. Careful
identification of stable cases with good baseline renal func-
tion and malignancy that can be managed with extirpative
surgery are good candidates for stent drainage. Furthermore
all stable cases with benign disease should be stented regard-
less of renal function. With better selection the failure rate of
internal stents in extrinsic compression should decrease.
Several predictive factors for stent success are described in
this report and we propose an algorithm (fig. 2) for the
management of extrinsic compression.
REFERENCES
1. Docimo, S. G. and Dewolf, W. C.: High failure rate of indwelling
ureteral stents in patients with extrinsic obstruction: experi-
ence at 2 institutions. J Urol, 142: 277, 1989
2. Yossepowitch, O., Lifshitz, D. A., Dekel, Y., Gross, M., Keidar,
D. M., Neuman, M. et al: Predicting the success of retrograde
stenting for managing ureteral obstruction. J Urol, 166: 1746,
2001
3. Feng, M. I., Bellman, G. C. and Shapiro, C. E.: Management of
ureteral obstruction secondary to pelvic malignancies. J En-
dourol, 13: 521, 1999
4. Ramsay, J. W., Payne, S. R., Gosling, P. T., Whitfield, H. N.,
Wickham, J. E. and Levison, D. A.: The effects of double J
stenting on unobstructed ureters: an experimental and clinical
study. Br J Urol, 57: 630, 1985
5. Finney, R. P.: Double-J and diversion stents. Urol Clin North
Am, 9: 89, 1982
6. Finney, R. P.: Experience with new double J ureteral catheter
stent. J Urol, 120: 678, 1978
7. Hepperlen, T. W., Mardis, H. K. and Kammandel, H.: The pigtail
ureteral stent in the cancer patient. J Urol, 121: 17, 1979
8. Donat, S. M. and Russo, P.: Ureteral decompression in advanced
nonurologic malignancies. Ann Surg Oncol, 3: 393, 1996
9. Jenkins, C. N. and Marcus, A. J.: The value of antegrade stent-
ing for lower ureteric obstruction. J R Soc Med, 88: 446, 1995
10. Hübner, W. A., Plas, E. G. and Stoller, M. L.: The double-J
ureteral stent: in vivo and in vitro flow studies. J Urol, 148:
278, 1992
11. Fromer, D. L., Shabsigh, A., Benson, M. C. and Gupta, M.:
Simultaneous multiple double pigtail stents for malignant
ureteral obstruction. Urology, 59: 594, 2002
12. Rotariu, P., Yohannes, P., Alexianu, M., Rosner, D., Lee, B. R.,
Lucan, M. et al: Management of malignant extrinsic compres-
sion of the ureter by simultaneous placement of two ipsilateral
ureteral stents. J Endourol, 15: 979, 2001
13. Mardis, H. K., Kroeger, R. M., Hepperlen, T. W., Mazer, M. J.
and Kammandel, H.: Polyethylene double-pigtail ureteral
stents. Urol Clin North Am, 9: 95, 1982
14. Pauer, W. and Lugmayr, H.: Metallic wallstents: a new therapy
for extrinsic ureteral obstruction. J Urol, 148: 281, 1992
15. Lugmayr, H. and Pauer, W.: Self-expanding metal stents for
palliative treatment of malignant ureteral obstruction. AJR
Am J Roentgenol, 159: 1091, 1992
16. Barbalias, G. A., Siablis, D., Liatsikos, E. N., Karnabatidis, D.,
Yarmenitis, S., Bouropoulos, K. et al: Metal stents: a new treat-
ment of malignant ureteral obstruction. J Urol, 158: 54, 1997
17. Schlick, R. W., Seidl, E. M., Kalem, T., Volkmer, B. and Planz,
K.: New endoureteral double-J stent resists extrinsic ureteral
compression. J Endourol, 12: 37, 1998
18. Nissenkorn, I. and Gdor, Y.: Nephrovesical subcutaneous stent:
an alternative to permanent nephrostomy. J Urol, 163: 528,
2000