Evaluation of preoperative ultrasonographic parameters to predict renal recovery in long-term survivors after treatment of feline ureteral obstructions: 2012–2019

Abstract

Objectives

The aim of this study was to determine whether preoperative ultrasound imaging characteristic(s) in cats suffering from unilateral benign ureteral obstructions are predictive of outcome after successful renal decompression with a subcutaneous ureteral bypass (SUB) device.

Methods

This was a retrospective study of 37 cats with unilateral, benign ureteral obstruction. Preoperative imaging characteristics (including renal pelvis diameter, parenchymal thickness [transverse plane], renal length and pelvic size:overall renal size) and biochemical data were evaluated for all cats diagnosed with a unilateral ureteral obstruction treated with a SUB device. Any patient with bilateral obstructions or documented bacteriuria/infection in the data collection period was excluded. All patients were followed between 3 and 6 months postoperatively to obtain postoperative biochemical data. Long-term outcome was defined as serum creatinine concentration at 3–6 months postoperatively.

Results

No preoperative imaging characteristics or biochemical findings were found to be significantly associated with long-term serum creatinine concentrations. The length of the kidney was found to be associated with change in blood urea nitrogen and creatinine with decompression but not with long-term renal values.

Conclusions and relevance

In this study, long-term renal function based on preoperative ultrasound imaging findings could not be predicted in cats with unilateral ureteral obstruction, regardless of the severity of the biochemical parameters, renal pelvic dilation (large or small pelvis), kidney size or thickness of renal parenchyma assessed.

Introduction

Benign ureteral obstruction(s) owing to ureterolithiasis, strictures or purulent plugs have been increasingly identified in feline patients over the past two decades.^1–3 Management options vary depending on etiology, severity of obstruction and treatment preference of the operator, and include medical management (fluid diuresis, mannitol, alpha-1 receptor antagonists), ⁴ surgical management (ureterotomy, neoureterocystostomy, ureteronephrectomy)^4–7 and interventional management (subcutaneous ureteral bypass [SUB] device, nephrostomy tube, ureteral stent placement).^8–14

With the growing use of SUB devices in veterinary practice, as well as the risk, expense and required follow-up associated with any intervention, determining preoperative predictors of long-term renal recovery is important, as they may be able to help in the decision faced by pet owners and clinicians in committing to surgical decompression. One study evaluated dozens of prognostic factors in cats undergoing decompression of ureteral obstruction by stent or SUB device placement for the treatment of benign ureteral obstruction. ¹⁵ The study found that no preoperative clinical, biochemical, imaging, historical or examination finding was predictive of long-term outcome, though abdominal ultrasounds were not evaluated in any great detail other than renal pelvis size in the transverse plane.

In our experience, renal recovery cannot be predicted based on the ultrasonographic appearance of the obstruction, and recovery – or the lack thereof – should not be predicted without encouraging decompression to relieve the parenchymal stretch. Therefore, the goal of the present study was to determine whether any preoperative ultrasound imaging characteristic(s) in cats with unilateral benign ureteral obstructions are associated with long-term renal recovery after renal decompression with a SUB device.

Materials and methods

Case selection

Medical records of all cats treated with a SUB device for a ureteral obstruction(s) at the Animal Medical Center from August 2012 to March 2019 were reviewed due to the availability of an electronic medical record. Cats were included for final analysis if they met the following criteria: benign unilateral ureteral obstruction; complete ultrasonographic imaging data available preoperatively; complete biochemical data and negative urine cultures both preoperatively and 3–6 months postoperatively; and documentation of renal pelvic and ureteral decompression based on an ultrasound report highlighting improvement in the size of the renal pelvis and ureter, as well as patency of the device documented during the flushing procedure.

Obstruction was diagnosed initially via abdominal ultrasound and confirmed via antegrade pyelocentesis with a fluoroscopic-guided ureteropyelogram, performed immediately prior to SUB device placement. All cats were treated with a standard perioperative management regimen, which has been described elsewhere. ⁹ Bilateral obstructions, evidence of bacteriuria or a positive urine culture were causes for exclusion in order to ascertain that successful decompression was the only known explanation for long-term renal biochemical changes.

Historical and laboratory data

Preoperative data evaluated included age, sex, weight, hydration status and duration of clinical signs. Preoperative biochemical data obtained included blood urea nitrogen (BUN), creatinine and symmetric dimethylarginine (SDMA; when available), potassium and packed cell volume (PCV), using IDEXX Laboratories, or, when after hours or otherwise clinically indicated, a point-of-care iSTAT CHEM 8+ device. Urine microbiologic culture was available for all cats pre-decompression via cystocentesis and/or pyelocentesis.

Imaging findings

Abdominal ultrasounds were reviewed by a single board-certified radiologist (ALR), and included the length of the kidney (sagittal plane), pelvic diameter (transverse plane), proximal ureteral diameter (maximum diameter in cross-section at the hilus), the parenchymal thickness (including the renal crest, in the transverse plane) and a ratio of the pelvic dilation to overall renal size in both planes (defined as pelvic size/[pelvic size + parenchymal thickness]) (Figures 1–3). Ultrasounds were further evaluated for the presence of irregular renal margination, renal asymmetry, the presence of a stricture or ureterolith, ipsilateral nephroliths, contralateral nephroliths or cystoliths.

Figure 1

Ultrasonographic image of the right kidney, in the transverse plane. The orange line denotes pelvic dilation

Figure 2

Ultrasonographic image of the left kidney, in the transverse plane. The green dashed line indicates pelvic diameter, whereas the blue dotted line notes parenchymal thickness. The orange dot-dash line represents the summation of the pelvic size and parenchymal thickness, which was used in calculating the ratio of pelvic dilation to overall renal size (defined as pelvic size/[pelvic size + parenchymal thickness])

Figure 3

Ultrasonographic image of the right kidney, in the transverse plane. The orange line denotes proximal ureteral diameter

SUB device placement

Confirmation of a ureteral obstruction was made in each cat by an antegrade ureteropyelogram using surgical assistance and fluoroscopic guidance immediately prior to SUB device placement (Norfolk Vet Products). Each device was placed in a routine manner, as described elsewhere. ⁹

Postoperative laboratory data

Renal values (BUN, creatinine, SDMA) were included prior to discharge from the hospital, at recheck examination for short-term (between 1 week and 1 month postoperatively) and long-term (3–6 months postoperatively) follow-up. The change in BUN and creatinine, preoperatively to both discharge and to long-term values, was calculated. Confirmation of long-term renal decompression was ascertained at follow-up based on a SUB device ultrasound and flushing report, the technique of which has been described elsewhere. ⁹ Follow-up time (in days) and cause of death (if applicable) were recorded for all cats.

Statistical methods

Baseline descriptive statistics are reported as mean ± SD for normally distributed variables and median (interquartile range) for non-normally distributed variables. Between-group analyses of variables related to the side of obstruction, irregular renal margination, renal asymmetry, the presence of a stricture or ureterolith, cause of obstruction (if identified ultrasonographically), the presence of nephroliths (ipsilaterally or contralaterally) or the presence of cystoliths were performed using ANOVA, as error residuals were normally distributed, generating least square means for analysis. Linear regression was also employed for the dependent nominal response variables related to BUN, creatinine, SDMA, PCV and parenchymal thickness (transverse) regressed by nominal covariates. The distribution of error residuals was assessed by visual inspection followed by the Anderson–Darling test. Analyses for proportions of categorical variables such as comparison of preoperative imaging characteristics with biochemical data, both prior to decompression, in the short and the long term were evaluated with a χ² test or Fisher’s exact test as appropriate. Dichotomous response variables related to death, margin irregularity and renal asymmetry were analyzed by a generalized linear model using a logistic link. All analyses were performed with SAS statistical software, where an unadjusted P value of <0.05 was deemed to be significant.

Results

Case selection

In total, 173 cats were treated with a SUB device between August 2012 and March 2019. Of these, 34 cats met the inclusion criteria for final analysis; exclusion criteria are outlined in Figure 4. Twenty-three cats (13.3%) died or were euthanized within the 3-month study period at a median of 17 days (range 1–82). Nine cats (5.2%) did not survive to discharge (range 1–7 days postoperatively). Fifteen cats (8.7%) died from renal or ureteral causes (related to the SUB device or refractory azotemia) during the 3-month study period, whereas six cats (3.4%) died of non-renal causes, including hepatic failure, respiratory or cardiac disease. In two cats (1.2%), the cause of death was unknown. Further evaluation of the 15 cats that were excluded and died of renal or ureteral causes was performed; of these, eight had incomplete medical databases (imaging or biochemical data), three had preoperative infections documented and one required subsequent surgery during the study period to the contralateral kidney. Three cats that died from renal causes within 3 months were then compared to the 34 cats that survived >3 months, where long-term creatinine concentrations are considered to plateau to baseline.

Figure 4

Inclusion and exclusion criteria for selected cats.

Historical and laboratory data

Descriptive statistics on included cats are summarized in Table 1. Twenty-nine cats were azotemic at the time of admission (85%), while five patients were non-azotemic (15%).

Table 1

Descriptive statistics on included cats

	Number of cats	Median	Range
Age (years)	34	9.5	0.5–17
Weight (kg)	34	3.725	2.03–5.88
Duration of clinical signs (months)	34	0.5	0.03–24
Preoperative BUN (mg/dl)	34	88	16–218
Preoperative creatinine (mg/dl)	34	6.3	1–17.5
Preoperative SDMA (µg/dl)	15	37	11–86
Discharge BUN (mg/dl)	34	34.5	17–112
Discharge creatinine (mg/dl)	34	2.2	0.9–6.1
Discharge SDMA (µg/dl)	11	16	10–29
Change in BUN, preoperatively to discharge (mg/dl)	34	25	−34 to 194
Change in creatinine, preoperatively to discharge (mg/dl)	34	1.85	−0.3 to 16
Short-term BUN (mg/dl)	34	40	19–79
Short-term creatinine (mg/dl)	34	2.1	1.1–5.1
Short-term SDMA (µg/dl)	14	17.5	8–38
Long-term BUN (mg/dl)	34	38	20–86
Long-term creatinine (mg/dl)	34	2.2	1.3–5.3
Long-term SDMA (µg/dl)	15	15	10–28
Historical BUN (mg/dl)	12	37.5	22–57
Historical creatinine (mg/dl)	12	2.55	1.2–3.7
Historical SDMA (µg/dl)	12	13	11–18
Overall change in BUN, preoperatively to long term (mg/dl)	34	37.5	−37 to 185
Overall change in creatinine, preoperatively to long term (mg/dl)	34	2.45	−0.6 to 15.6

BUN = blood urea nitrogen; SDMA = symmetric dimethylarginine

Median short-term BUN, creatinine and SDMA concentrations were 40 mg/dl (range 19–79), 2.1 mg/dl (range 1.1–5.1) and 17.5 µg/dl (range 8–38), respectively. Median long-term BUN, creatinine and SDMA concentrations were 38 mg/dl (range 20–86), 2.2 mg/dl (range 1.3–5.3) and 15 µg/dl (range 10–28), respectively. Of the 29 cats that were azotemic preoperatively, 28 had improvement in their long-term creatinine concentrations after decompression (97%) (Figure 5). Preoperative creatinine concentrations were significantly higher (P <0.001) than long-term creatinine concentrations. Twelve cats had both preoperative and long-term SDMA concentrations assessed; of these, 10 (83%) had elevated SDMA prior to decompression, and all had improvement in their SDMA levels with decompression (median preoperative SDMA 43.5 µg/dl vs postoperative SDMA 14.5 µg/dl). Preoperative SDMA concentration was significantly higher (P = 0.0002) than long-term SDMA concentration in all 12 cats.

Figure 5

Graph of serial creatinine measurements before and after subcutaneous ureteral bypass (SUB) placement in 34 cats presenting for unilateral benign ureteral obstruction

Imaging findings

Preoperative imaging data are summarized in Table 2. Twelve cats were obstructed on the right side (35%) and 22 on the left (65%). Irregular margination was identified in six (18%) cats. Renal asymmetry was identified in 25 (73.5%) cats. Ipsilateral nephroliths were identified in 10 (29%) cats; contralateral nephroliths in 16 (47%) cats. Cystoliths were identified in six (18%) cats.

Table 2

Ultrasonographic imaging data on obstructed kidneys in 34 cats presenting with unilateral ureteral obstruction

	Median	Range
Length of kidney (sagittal) (mm)	41	25–57
Pelvic size (transverse) (mm)	11	6–21
Proximal ureteral size (mm)	4	1–6
Parenchymal thickness (transverse) (mm)	9	4–15
Ratio of pelvic size to overall renal size (transverse)	5	3–8

The cause of the ureteral obstruction was identified on ultrasound as a ureterolith in 19 (56%) cats. The cause was not identified on ultrasound in 15 (44%) cats and was subsequently determined at the time of the antegrade pyelogram as dried solidified blood stones (DSBS; n = 5), a ureteral stricture (n = 6, two with a circumcaval ureter), ureterolithiasis (n = 2), flocculent material in the ureter (n = 1) and unknown (n = 1). Overall, 21/34 (62%) cats had a ureterolith, 5/34 (14.7%) DSBS, 6/34 (17.6%) a ureteral stricture, 1/34 (3%) debris and 1/34 (3%) was of unknown cause.

Two imaging characteristics on abdominal ultrasonography were significantly associated with preoperative bloodwork findings: length of the kidney in the sagittal plane and parenchymal thickness in the transverse plane. Renal length was positively associated with preoperative creatinine (P = 0.018) and SDMA (P = 0.001) concentrations; for every 10 mm increase in renal length, preoperative creatinine and SDMA concentrations increased by 2.53 mg/dl and 30 µg/dl, respectively. Parenchymal thickness in transverse was positively associated with preoperative BUN (P = 0.03). For every 10 mm increase in parenchymal thickness, preoperative BUN increased by 74.4 mg/dl.

No preoperative imaging characteristics were significantly associated with long-term BUN, creatinine or SDMA levels. However, statistical significance for overall change in renal parameters was found with decompression but was not related to long-term values. Larger length of the kidney in the sagittal plane was positively associated with a larger change in BUN and creatinine levels both in the immediate postoperative (preoperative period to discharge; BUN, P = 0.025; creatinine, P = 0.01) and long-term periods (BUN, P = 0.01; creatinine, P = 0.006) but was not associated with the absolute long-term BUN (P = 0.15) and creatinine (P = 0.32) concentrations. For every increase of 10 mm in renal length, the change in BUN, creatinine and SDMA concentrations from the preoperative period to discharge increased by 30 mg/dl, 2.59 mg/dl and 15.9 µg/dl, respectively. Moreover, for every 10 mm increase in renal length, the long-term change in BUN and creatinine increased by 31 mg/dl and 2.77 µg/dl, respectively. This finding is supported by the fact that the length of the kidney in the sagittal plane was associated with higher preoperative creatinine (P = 0.018) and SDMA (P = 0.001) concentrations but not long-term creatinine (P = 0.32) or SDMA (P = 0.95) levels.

Parenchymal thickness in transverse was positively associated with the change in SDMA levels, preoperatively to discharge (P = 0.019). For every 10 mm increase in parenchymal thickness, the change in SDMA preoperatively to discharge increased by 38 µg/dl. Thus, a larger kidney, in both length and parenchymal thickness, was associated with a larger change in renal values following decompression. However, this finding was not associated with long-term renal recovery, defined as long-term creatinine concentration.

Imaging data for the three non-surviving cats that died of renal causes were analyzed separately to evaluate if any imaging parameters differed between short-term non-survivors and the included cats that survived >3 months. The ratio of pelvic size to overall renal size in the transverse plane was found to be statistically significantly larger in the surviving than non-surviving cats (P = 0.0298), but the imaging parameters from which this ratio was derived, namely renal pelvic size in the transverse plane and renal parenchymal thickness in the transverse plane, were not found to be significantly different between included and excluded cats (P = 0.0869 and P = 0.0636, respectively).

Median duration of follow-up was 855 days (range 120–2430). At the time of writing, 18/34 cats (53%) were still alive and 16/34 (47%) were deceased. Causes of death were found to be progressive renal disease in eight cats (50%), neoplasia in two cats (12.5%), cardiac disease in two cats (12.5%), hepatic disease in one cat (6%) and unknown in three cats (19%).

Discussion

Lack of robust visible renal parenchyma, small renal length or severity of renal pelvic dilation can lead to a suspicion that the severity of azotemia may persist despite surgical intervention, and guide prognosis and medical decisions with assumptions of chronicity and severity. In this study, no preoperative imaging characteristics were significantly associated with long-term BUN, creatinine or SDMA concentrations after successful decompression with a SUB device. This indicates that, despite the size of the kidney or renal pelvis, degree of renal pelvic dilation, renal parenchymal thickness visible on ultrasound or ratio of pelvic diameter to overall renal size, surgical decompression should be considered in cats presenting with benign unilateral ureteral obstruction(s). Ninety-seven percent of azotemic cats had improvement in their long-term serum creatinine concentrations regardless of preoperative imaging parameters. As 15% of patients in this study were non-azotemic, improvement in renal function in this population may not be recognized by evaluating creatinine concentrations alone. A glomerular filtration rate study could be more sensitive in evaluating renal recovery (both for azotemic and non-azotemic cats), but is not widely used in clinical practice, and will delay the acute treatment needs of the patient. These findings are in agreement with those of Horowitz et al, ¹⁵ where limited imaging characteristics were explored and no single preoperative parameter was associated with renal recovery or overall prognosis.

In this study, there was no association between size (renal or pelvic) and outcome (Figure 6). The preoperative imaging characteristic found to be statistically significant to the change in BUN and creatinine with decompression was the length of the kidney. Similarly, parenchymal thickness in the transverse plane was positively associated with the change in SDMA with decompression. However, despite a larger change in renal values obtained with decompression, long-term renal recovery was not significantly different than for those with smaller renal size or parenchymal thickness. This supports the concurrent finding that cats with larger kidneys and more renal parenchyma measured on ultrasound may present with higher BUN, creatinine or SDMA concentration; therefore, there is more room to change postoperatively, but this does not predict the exact concentrations long-term in BUN and creatinine.

Figure 6

Differences in renal length prior to decompression for two cats with unilateral ureteral obstruction. For the patient shown in (a), renal length was measured at 25 mm. For the patient shown in (b), renal length was measured at 46 mm. Both cats had long-term International Renal Interest Society stage 3 chronic kidney disease after decompression (creatinine 3.1 mg/dl for the patient in [a], creatinine 2.9 mg/dl for the patient in [b]).

With the growing use of abdominal ultrasonography in the assessment of azotemic patients, caution must be exercised in both the diagnosis and assumption of prognosis in cats with suspected ureteral obstruction. Previous studies have evaluated abdominal ultrasonography in the diagnosis of ureteral obstruction, and have found that it may be inaccurate, with both false-positive and false-negative results, compared with the gold standard of antegrade pyelography. ¹⁶ Previous studies have found that renal pelvic diameters >13 mm were most consistently associated with ureteral obstruction,^17,18 but cats with ureteral obstruction may have less severe pelvic dilation, as did those in the present study and others.^{9,13,15,19,20} Furthermore, the definitive cause of obstruction is not always discernible from abdominal ultrasound. Forty-four percent of cats in the present study did not have the cause of their obstruction identified on ultrasound, and were diagnosed via a combination of pyelography and surgical exploration. Thus, although the use of abdominal ultrasound is essential in the work-up of azotemic feline patients, clinicians must exercise caution in the interpretation of results as the lack of an obstructive lesion, like a ureterolith, should not dismiss the diagnosis of a ureteral obstruction.

Interestingly, a larger proportion of cats were found to be obstructed on the left side (65%), than the right side. We suspect that this was owing to the relatively small population size, as previous studies have failed to show a significant difference in the number of cats obstructed on one side or the other.^6,13

There are several limitations to this study, including its retrospective nature and small number of included cases. Only unilateral obstructions were included to truly ascertain the effect of each imaging change with the change in renal functional parameters over time. This resulted in exclusion of a large number of cases, most importantly those who were bilaterally obstructed or had evidence of a urinary tract infection. For the evaluation of an appropriate time period to allow recovery, we focused on cats that survived at least 3 months after surgery, but we also assessed those that died, or were euthanized, prior to 3 months, to ensure bias was limited. This population was evaluated separately, as the scope of the current report was not to determine preoperative imaging characteristics that would predict overall survival, but instead to specifically examine cats’ baseline creatinine, which are seen to stabilize 3 months postoperatively, and how it relates to preoperative imaging. We have found in prior studies that serum creatinine plateaus to its lowest level at approximately 3 months after intervention;^13,15 therefore, we elected to focus on cats that had documentation of decompression and biochemical data within this time period. However, to ensure that there were not major imaging differences between survivors to >3 months and those that did not survive to 3 months, additional cats were evaluated to compare their imaging data to those that survived to the plateau period. Owing to this small number (n = 3), the statistical comparisons were not well powered. Regardless, only one parameter was considered to be statistically different between these two groups (ratio of pelvic size to overall renal size in the transverse plane). This ratio was derived from two imaging parameters found not to be statistically significant, so the true impact of this finding is unknown and likely of little clinical significance. Further studies are needed with larger sample sizes to evaluate if there are preoperative imaging characteristics that can predict short-term outcome in cats with ureteral obstruction, but that was not the scope of the present study, and other larger reports have shown that there is no association of short- or long-term survival or creatinine concentrations with any preoperative parameters.^13,15

Additional limitations include that the preoperative ultrasounds were performed by different radiologists at the same institution; however, inter-observer variability was minimized by one radiologist retrospectively evaluating all images. Furthermore, sagittal imaging evaluations were performed but ultimately not included in the report owing to variability in measurements with cortical asymmetry in this population. Standardized prospective studies are warranted to evaluate the prognostic significance of a larger array of ultrasonographic measurements on short- and long-term outcome in a greater number of cats treated with SUB devices.

Postoperative imaging data were not evaluated in the present study. While cats with SUB devices have serial ultrasounds as part of routine follow-up to confirm renal decompression and SUB device patency, at our institution these are typically performed by us on a different machine than the one used by the radiologists; thus, they were not included for direct comparison of renal imaging.

Not all renal values for these patients were measured on the same analyzer owing to the time and situation upon which they presented to the hospital. When multiple analyzers were used for the same patient, the machine used most frequently was evaluated to minimize machine variability for an individual patient. Moreover, SDMA measurement was not included for all patients, as it was added to biochemical panels midway through the study period at this institution.

Long-term muscle condition scores were not evaluated in these patients, and hydration scores were subjectively assessed by clinicians at follow-up; as creatinine is influenced by both parameters, this could have affected measurement of renal function. However, weight gain and presumably muscle mass, as frequently occurs in patients after successful renal decompression, would have been expected to cause an increase in creatinine and thus an apparent lack of renal recovery, which was not observed. Furthermore, dehydration, a known complication of chronic renal disease in cats, would be expected to cause a lack of apparent renal recovery at long-term follow-up, which was not observed.

Moreover, the causes of obstruction were dissimilar between cats, and chronicity of obstruction was unknown. Despite the lack of any previous studies demonstrating that cause or chronicity have prognostic significance in long-term outcome for cats with ureteral obstruction(s), a homogenous patient population would have been ideal to minimize bias.

Since only cats with unilateral obstruction were included, hypertrophy of the contralateral kidney cannot be ruled out as a contributing factor to long-term BUN and creatinine concentrations. This was thought unlikely to be a large factor in the changes seen in creatinine because 85% of cats in this study were concurrently azotemic at diagnosis while only unilaterally obstructed, suggesting that the non-obstructed kidney was not functioning normally, and therefore the renal hypertrophy mechanisms were likely exhausted prior to decompression.

Lastly, dietary therapy was dissimilar in cats after discharge, which may also have affected their long-term renal values, as diet is designed to slow the progression of chronic renal disease in residually azotemic patients, while cats with stone disease that are ultimately non-azotemic are encouraged to eat a stone-prevention diet.

Conclusions

Despite the aforementioned limitations, the lack of association between preoperative imaging characteristics and long-term renal recovery based on biochemical parameters warrants consideration by clinicians when recommending decompressive surgery for ureteral obstruction. Renal appearance on ultrasound alone should not guide decisions on whether decompression will improve long-term renal outcome. As hypothesized, no preoperative imaging characteristics were found to be clinically significantly associated with long-term renal recovery.