Renal Deterioration Index: Preoperative Prognostic Model for Renal Functional Outcome After Treatment of Bilateral Obstructive Urolithiasis in Patients with Chronic Kidney Disease

Abstract

Background and Purpose:

Patients who present with varying severity of obstructive urolithiasis behave differently after the treatment. Some patients recover with improved renal function while others progress to renal failure. Our objective was to objectively quantify which patients would progress to renal failure after treatment for obstructive urolithiasis.

Patients and Methods:

A prospective analysis of 167 patients with renal failure from bilateral obstructive urolithiasis who were treated and subsequently followed for at least 1 year was performed. Failure was defined as glomerular filtration rate (GFR) values less than 15 mL/min at 1 year follow-up. All patients had preoperative placement of a percutaneous nephrostomy tube for at least 5 days before treatment with either ureteroscopy or percutaneous nephrolithotomy. Multiple logistic regression analysis of affecting parameters was performed. A renal deterioration index (RDI) was constructed based on scores assigned to varying severity of multivariate significant factors and the receiver operating characteristic (ROC) curve was analyzed.

Results:

There were 48(28.7%) patients who progressed to CKD stage V at 1-year follow-up. Combined cortical width (≤0.001), proteinuria (0.01), positive urine culture (0.004), and nadir preoperative GFR postbilateral percutaneous nephrostomy (0.016) were statistically significant factors affecting renal deterioration on multivariate analysis. RDI has a high ROC curve (AUC=0.90) for predicting renal functional outcome. Combining these parameters in a prediction table yielded a RDI score ≥12 being associated with high odds risk (odds ratio=11.2) of treatment failure.

Conclusion:

RDI ≥12 is associated with renal deterioration after appropriate treatment of bilateral obstructive urolithiasis.

Introduction

Kidney stones and chronic kidney disease (CKD) are common, affecting 5% and 13% of the adult population, respectively.^1,2 Management of urinary stone disease in patients with CKD is often difficult.³

There are no tools at present that could explain the behavior of renal function after optimum treatment of obstructive nephrolithiasis in patients with CKD. A crude method is prediction of long-term renal function based on the presenting duration. Patients who present with obstructive urolithiasis and acute renal failure are likely to have a more favorable outcome than those with obstruction and chronic renal failure. In both scenarios, treatment of nephrolithiasis is likely to improve renal function. In patients with previously compromised renal function, this may be of temporary benefit in prolonging the ultimate requirement of renal replacement therapy (RRT).^2,4 Some patients do not improve after the treatment of nephrolithiasis. They alternatively progress to CKD stage 5 rapidly.^2,4
–6 Before embarking on extensive stone removal, it is pertinent to preoperatively prognosticate the category of patients who are not going to improve after treatment.

In this context, a routine, economical but comprehensive method to predict renal functional outcome would be helpful to promote early awareness, patient counseling, and better management of CKD among the population with obstructive urolithiasis. Evidence from earlier published studies indicates that clinical factors are already present at the time of initial clinical presentation.^{3,4

–11} Given the relative ease of measuring these risk factors, we attempted to develop a logistic regression model to establish their significance and then quantify in a scoring system to identify persons at greatest risk for adverse renal functional outcome.

Patients and Methods

Patients

The patients were recruited from the hospital database in a prospective manner from January 2009 to July 2010. After treatment, the patients were followed for disease progression assessment until July 2011. The ethics committee of the hospital approved this study. Patients who presented with altered renal function with bilateral obstructive renal or ureteral calculi and who were prospectively followed for at least 1 year after the treatment of the stone disease were included in the study. Patients with acute renal failure, pediatric patients, and those with solitary kidney with renal insufficiency were excluded.

Details on patient age, sex, weight, comorbidity, presenting clinical features, and significant medical and surgical history were recorded. Because hypertension and diabetes are major risk factors for development of CKD in the general population, a note of these comorbidities and preexisting CKD was made. Laboratory evaluation included urinalysis and culture, basic hematology, and serum biochemistry determinations. Proteinuria is a surrogate outcome in CKD, and there appears to be sufficient evidence to recommend changes in proteinuria as a surrogate for kidney disease progression. We estimated proteinuria in this study by the dipstick method. All patients were evaluated with plain radiography and ultrasonography (US) and/or CT of the kidneys, ureters, and bladder region. CT or US measured the maximum measurable cortical width in each kidney. The sum of cortical width of both kidneys constituted the combined cortical width.

The general principle of care is as in our previous articles.¹² Specifically, all patients underwent stabilizing hemodialysis if needed and percutaneous nephrostomy (PCN) placement in the appropriate calix. Our treatment strategy involves draining all hydronephrotic kidneys with PCN. The idea behind performing PCN in all patients was it is becoming increasingly safe, is being performed with increasing frequency on an outpatient basis, the anticipated dwell time was shorter, and that it may serve as a conduit for definitive antegrade treatment. The PCN tubes were placed strategically with careful planning, so that the matured tracts could be used for future percutaneous nephrolithotomy (PCNL) or antegrade ureteroscopy (URS). Multiple nephrostomy tubes were placed if needed for adequate drainage of all calices. Urine obtained at the time of PCN was sent for culture and sensitivity determinations, and the patient was treated with appropriate antibiotics. The help of a nephrologist was obtained for correction of fluid overload, electrolyte imbalances, acidosis, and anemia. Appropriate temporary RRT was also initiated if needed.

There is no guideline determining the exact duration to wait until the nadir creatinine level is achieved. To eliminate the element of acute renal obstruction as a cause of an elevated creatinine level, the patients with evidence of obstruction were drained via PCN for an adequate period until nadir serum creatinine level (a minimum of two equal lowest values) was reached. A minimum of 5 days of observation for creatinine improvement was allowed. All patients achieving nadir creatinine level below 1.5 mg% within 5 days were excluded from the prospective database to eliminate acute renal failure as a cause of obstructive uropathy. GFR estimates at presentation and nadir before intervention (minimum 5 days of adequate deobstruction) were performed by using the four-variable Modification of Diet in Renal Disease (MDRD) equation.¹³

Subsequently, patients were treated with either PCNL or URS in stages until rendered stone free. Bilateral Double-J stents were placed before discharge and removed at 1-month follow-up. Complete clearance was defined as nonvisualization of residual fragments on radiography and US at 1 month after the procedure. Patients were followed at 3-month intervals until 1 year and then at 6-month intervals. Selected patients with CKD received appropriate medical management. Treatment failure for the study was defined as the progression of renal function to CKD stage 5 or requirement of RRT within 1 year of treatment.

Statistical analysis

In the present study, all preoperative parameters were studied for the analysis. The relationship of preoperative parameters and treatment failure was analyzed by univariate and multivariate logistic regression, with odds ratios (OR) and confidence intervals (CI) calculated for each variable. A forward stepwise multivariate logistic regression analysis was used to build a prediction table (renal deterioration index [RDI]) for predicting the risk of renal deterioration. To evaluate the performance of the prediction table, the receiver operating characteristic (ROC) curves of significant multivariate factors and RDI were generated, respectively, and their areas under the ROC curve were compared with each other. OR for renal deterioration with different RDI scores were calculated. Analysis was performed using SPSS software version 15.0 (SPSS Inc, Chicago, IL). All P values were two-sided and a P value of <0.05 was considered statistically significant.

Results

Table 1 highlights the baseline clinical correlates and follow-up results. There were 48 (28.7%) patients who progressed to CKD stage 5 at 1-year follow-up. Thirteen patients and 3 patients had Clavien III and IV complications, respectively. The complications were febrile urinary tract infection (UTI) (6), bleeding necessitating multiple bladder washes (1), bleeding necessitating nephrostomy clamping (1), ureteral clot obstruction necessitating Double-J stent placement (1), prolonged urine leak (1), angioembolization (2), septic shock syndrome (1), and frank sepsis (2).

Table 1.

Patient Characteristics, Demography, Clinical Presentation, Intraoperative and Postoperative Parameters and Outcome

Patients (n)	169
Age; years (mean±SD)	48.06±14.09
Sex (male/female)	133/36
Serum creatinine at presentation; mg% (mean±SD)	7.26±4.42
Preoperative GFR (mL/min/1.73 m²)	13.9±11.5
Number of calculi units	338
Renal units
Partial staghorn	44
Compete staghorn	50
Other renal calculi	76
Ureteral units	168
Hemoglobin at presentation; gm% (mean±SD)	10.47±4.21
Serum bicarbonate at presentation; mEq/L (mean±SD)	17.95±5.96
Combined cortical width; mm (mean±SD)	23.04±8.52
Proteinuria (urine dipstick method)
0	64
1	69
>1	36
Positive preoperative urine culture (n%)	20.4%
Comorbidity (hypertension, diabetes, CKD)
0	103
1	49
2	15
3	2
Stablizing temporary pre-PCN dialysis (n%)	23.3%
CKD stage at presentation
Stage 3	22
Stage 4	33
Stage 5	114
Serum creatinine at 5 days of deobstruction; mg% (mean±S.D)	3.35±2.16
Endourologic management in units	338
PCNL - single tract	142
PCNL - multiple tract	32
URS	124
Mean operative time (minutes)
PCNL	62±34
URS	41±22
Perioperative complication
Clavien I	30
Clavien II	18
Clavien III	13
Clavien IV	3
Clavien V	0
Blood transfusion (n)
PCNL (preoperative)	20
PCNL (preoperative)	8
URS (preoperative)	6
URS (preoperative)	1
Serum creatinine at 1 year; mg% (mean±SD)	3.43±3.18
Treatment failure	49
GFR <15 mL/min	11
MHD	38

SD=standard deviation; GFR=glomerular filtration rate; CKD=chronic kidney disease; PCN=percutaneous nephrostomy; PCNL=percutaneous nephrolithotomy; URS=ureterostomy; MHD=maintenance hemodialysis.

Univariate analysis of risk factors and adverse outcome is tabulated in Table 2. Anemia, acidosis, cortical atrophy, positive urine culture, GFR at 5 days of deobstruction, and proteinuria were associated with treatment failure at 1 year. Table 3 shows the results of the multiple logistic stepwise regression analysis for the patients. Combined cortical width, proteinuria, positive urine culture, and nadir GFR postbilateral PCN were the combined significant factors.

Table 2.

Univariate Analysis of the Preoperative Parameters Determining Outcome

Variable	Odds ratio	95% confidence interval	P value
Age	0.9983	0.9749–1.0222	0.8877
Serum bicarbonate at presentation	0.9350	0.8803–0.9930	0.0242
Hemoglobin at presentation	0.7138	0.6094–0.8362	<0.0001
Combined cortical width	0.8143	0.7591–0.8734	<0.0001
Comorbidity	1.2417	0.7820–1.9716	0.3626
Proteinuria	2.8165	1.7943–4.4209	<0.0001
GFR at 5 days of deobstruction	1.8348	1.4844–2.2680	<0.0001
Urine culture	2.7205	1.2475–5.9331	0.0127
GFR at presentation	1.0681	1.2475–5.9331	0.0811
Renal or ureteric stone	1.7157	1.1074–2.6581	0.0139

GFR=glomerular filtration rate.

Table 3.

Multivariate Preoperative Significant Parameters Affecting Outcome

Variable	Coefficient	Standard error	Odds ratio	95% confidence interval	P value
Combined cortical width	−0.1764	0.0403	0.84	0.77–0.90	<0.0001
Nadir GFR after adequate deobstruction	0.3167	0.1320	1.37	1.06–1.78	0.016
Proteinuria	0.7261	0.2807	2.07	1.19–3.58	0.010
Urine culture	1.6019	0.5518	4.96	1.68–14.63	0.004
Constant	0.5182	-	0.00		<0.0001

GFR=glomerular filtration rate.

All the significant factors were then graded in increasing severity to provide a score for each variable (Table 4). The sum of each variable provided the RDI. For each patient, RDI ranged from 4 to 18. The sensitivity, specificity, positive and negative predictive value of RDI was compared with other individual significant factors by constructing ROC curves and calculating the area under the ROC curve (AUC). We performed ROC curves analysis for the regression model (Fig. 1). AUCs of cortical width, nadir GFR, proteinuria, and urine culture and RDI were 0.85±0.04, 0.83±0.03, 0.74±0.04, 0.59±0.04, and 0.90, respectively. AUC for RDI was 0.9 indicating very high accuracy for the prediction table. The OR table for various cutoffs of RDI is as in Table 5. The optimal threshold of RDI more than 12 was found in the regression model for high odds of treatment failure (Table 5).

FIG. 1.

Receiver operating characteristic curve analysis for regression model. GFR=glomerular filtration rate; PCN=percutaneous nephrostomy; RDI=renal deterioration index.

Table 4.

Renal Deterioration Index Scoring System

Score	1	3	5
Combined cortical width (mm)	>20	10–20	<10
GFR at 5th day (mL/min)	>60	30–60	<30
Proteinuria (dipstick method)	0	+1	>+1
Urine culture	Negative	Positive

GFR=glomerular filtration rate.

Table 5.

Odds Ratio Table for Various Cutoffs of the Renal Deterioration Index

RDI	Success	Failure	Odds ratio
<6	57	2	0.079
7–8	28	2	0.16
9–10	17	6	0.79
11–12	14	11	1.76
13–14	4	20	11.22
>14	0	8	37.95
Total	120	49

RDI=renal deterioration index.

Discussion

Most agree that stones in patients with CKD should be cleared. In patients with mild to moderate renal insufficiency, an aggressive approach is needed to render them stone free with improved renal function.^3,11 In a cohort of 171 patients with severe idiopathic calcium stone disease, Marangella and associates¹⁴ reported on patients with mild insufficiency and a mean GFR of 67 mL/min/1.73 m² at referral; a significant decline occurred during a mean follow-up of nearly 3.5 years. On the other hand, Worcester and colleagues¹⁵ reported that renal function determined by creatinine clearance decreases with age in stone formers at a higher rate than in nonstone-formers, and that patients with kidney stones do not have normal kidney function compared with healthy persons. Therefore, in clinical management, all efforts must be made to minimize renal injury while balancing the risks of obstruction from stones against those of urologic procedures.

PCNL and URS are the primary endourologic modalities for treating patients with CKD who have urolithiasis; however, they are not without complications.^16,17 Hypothermia, bleeding, metabolic acidosis, disturbances in serum electrolytes, urosepsis, and even deaths are the main complications.^8

–12 Despite technical advances, bleeding and urosepsis remain a major concern, even in patients with CKD who have sterile preoperative urine cultures.¹⁷ Because of the inhibition of cell-mediated immunity and humoral defense mechanisms, septicemia can easily develop in patients with CKD. Agrawal and coworkers⁹ reported a mortality rate caused by sepsis of 3.8% among 78 patients with advanced uremia who were treated for urolithiasis. Other authors have also reported higher overall complication rates in patients with CKD. Complex stones are associated with poorer results; necessitate longer operative time, multiple percutaneous tracts, and more ancillary procedures; and have a higher complication rate.

The overall benefit of PCNL is evident, although the complication rate and mortality risk are high in these patients. Therefore, one may expect that aggressive treatment of such patients could prevent the need for RRT, or at least significantly delay it in most patients with renal stones. Complications are therefore acceptable in patients where the treatment defers RRT significantly. On the other hand, there are some patients with severe renal insufficiency in whom despite aggressive stone removal, due to complications or even otherwise, RRT cannot be significantly delayed. There are also other important decision-making factors that need to be considered along with a decision-making algorithm such as desire to eradicate urine infection by relieving obstruction/removing stone, and preservation of “residual renal function” for easing water balance. RDI, however, may be an important preoperative tool to prognosticate the ultimate outcome.

There have been some attempts to identify predictors of prognosis and treatment outcome in patients with CKD who have urolithiasis. Kukreja and colleagues⁸ reported proteinuria (>300 mg/d), atrophic cortex (<5 mm), recurrent UTI, stone bulk (>1500 mm²), and pediatric age group as predictors of adverse renal outcome. Agarwal and associates⁹ found parenchymal thickness of 7 mm, clear urine in the collecting system, no renal sepsis, and recent onset azotemia as favorable predictors. Associated hypertension and diabetes are aggravators of CKD.¹⁸ In a study by Kurien and coworkers,¹² the majority of patients were in CKD stages 3 and 4. Improved renal function after PCNL was seen in one third of the patients. Atrophic cortex and postoperative complications, mainly infection and bleeding, predicted renal deterioration. Canes and colleagues¹⁹ found the improvement in postoperative eGFR to be predictive of improvement of renal function.

Our goal was to develop a relatively simple preoperative approach allowing prediction of treatment outcome in patients with altered renal function in bilateral urolithiasis. The article proposes to calculate RDI score in every patient presenting with chronic obstructive uropathy secondary to urolithiasis. We concentrate on four factors to determine the RDI score. CT or US measures the maximum measurable cortical width of each kidney. The sum of cortical width (in millimeters) of both kidneys constitutes the combined cortical width. Urinalysis determines proteinuria by dipstick method. Urine culture is estimated as either positive or negative. GFR estimates at nadir before intervention (minimum 5 days of adequate deobstruction) is performed by using the four-variable MDRD equation. All these parameters are then fed into Table 4 to calculate the RDI score. The RDI score ranges from a minimum 4 to a maximum 18. RDI ≥12 is associated with high odds of renal deterioration after appropriate treatment of bilateral obstructive urolithiasis. It can also be used to inform patients about their level of absolute risk to help them make an informed choice regarding the need for further intervention. Early identification of treatment failure may lead to more conservative intervention and early initiation of RRT.

We studied all the factors that were predicted by other authors to affect the renal functional outcome. We found in our multivariate logistic regression analysis that four factors were in combination associated with functional outcome. Our study data showed that treatment failure was relatively high in patients with reduced combined cortical width, proteinuria, low deobstruction GFR value, and positive culture. This is consistent with the results from previous studies. We constructed an easily measurable score (RDI) of the four relevant factors in a prediction table. Compared with the individual multivariate risk factors, RDI represented a good predicting test (AUC=0.90) and a useful aid for predicting adverse outcome. The prediction table represented the optimal combination of sensitivity and specificity at the optimal threshold.

The present regression model has some attractive features when compared with a traditional anticipating tool. Our regression model is based on routinely available preoperative data; therefore, the model can be applied easily in clinical practices. We think that the most important factor predicting renal recovery is the duration of obstruction. This variable was not included in the study because the actual duration of the obstruction cannot be predicted in most patients. It is more common for short duration of obstruction being evident as acute renal failure, however. Those patients whose creatinine level dropped down to nadir less than 1.5 mg% were excluded from the analysis. This ensured that patients had longer duration of obstruction.

Conclusion

There are no tools at present that could explain the behavior of renal function after optimum treatment of obstructive nephrolithiasis. Before embarking on an extensive stone removal, it is pertinent to preoperatively prognosticate the category of patients who are not going to improve after treatment. Our study data showed that the treatment failure was statistically high in patients with reduced combined cortical width, proteinuria, and low deobstruction GFR value at the fifth day and positive culture. We constructed an easily measurable score (RDI) of the four relevant factors in a prediction table. Compared with the individual multivariate risk factors, RDI represented a good predicting test (AUC=0.90) and a useful aid for predicting adverse outcome. RDI ≥12 is associated with renal deterioration after appropriate treatment of bilateral obstructive urolithiasis.

Footnotes

Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Coresh

, Selvin

, Stevens

et al. Prevalence of chronic kidney disease in the United States. JAMA, 2007; 298:2038–2047.

Jungers

, Joly

, Barbey

et al. ESRD caused by nephrolithiasis: Prevalence, mechanisms, and prevention. Am J Kidney Dis, 2004; 44:799–805.

Gupta

, Bolton

, Gupta

, Stoller

. Improved renal function following aggressive treatment of urolithiasis and concurrent mild to moderate renal insufficiency. J Urol, 1994; 152:1086–1090.

Teichman

, Long

, Hulbert

. Long-term renal fate and prognosis after staghorn calculus management. J Urol, 1995; 153:1403–1407.

Streem

, Geisinger

. Combination therapy for staghorn calculi in solitary kidneys: Functional results with long-term followup. J Urol, 1993; 149:449–452.

Singh

, Gupta

, Hemal

et al. Efficacy and outcome of surgical intervention in patients with nephrolithiasis and chronic renal failure. Int Urol Nephrol, 2001; 33:293–298.

Carvalho

, Martin

, Passos

, Riella

. Nephrectomy as a cause of chronic kidney disease in the treatment of urolithiasis: A case-control study. World J Urol, 2012E-pub ahead of print.

Kukreja

, Desai

, Patel

, Desai

. Nephrolithiasis associated with renal insufficiency: Factors predicting outcome. J Endourol, 2003; 17:875–879.

Agrawal

, Aron

, Asopa

. Endourological renal salvage in patients with calculus nephropathy and advanced uraemia. BJU Int, 1999; 84:252–256.

10.

Bilen

, Inci

, Kocak

et al. Impact of percutaneous nephrolithotomy on estimated glomerular filtration rate in patients with chronic kidney disease. J Endourol, 2008; 22:895–900.

11.

Paryani

, Ather

. Improvement in serum creatinine following definite treatment of urolithiasis in patients with concurrent renal insufficiency. Scand J Urol Nephrol, 2002; 36:134–136.

12.

Kurien

, Baishya

, Mishra

et al. The impact of percutaneous nephrolithotomy in patients with chronic kidney disease. J Endourol, 2009; 23:1403–1407.

13.

Levey

, Coresh

, Greene

et al. Chronic Kidney Disease Epidemiology Collaboration. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med, 2006; 145:247–254.

14.

Marangella

, Bruno

, Cosseddu

et al. Prevalence of chronic renal insufficiency in the course of idiopathic recurrent calcium stone disease: Risk factors and patterns of progression. Nephron, 1990; 54:302–306.

15.

Worcester

, Parks

, Josephson

et al. Causes and consequences of kidney loss in patients with nephrolithiasis. Kidney Int, 2003; 64:2204–2213.

16.

Lingeman

, Woods

, Toth

et al. The role of lithotripsy and its side effects. J Urol, 1989; 141:793–797.

17.

Gopalakrishnan

, Prasad

. Management of urolithiasis with chronic renal failure. Curr Opin Urol, 2007; 17:132–135.

18.

Saucier

, Sinha

, Liang

et al. Risk factors for CKD in persons with kidney stones: A case-control study in Olmsted County, Minnesota. Am J Kidney Dis, 2010; 55:61–68.

19.

Canes

, Hegarty

, Kamoi

et al. Functional outcomes following percutaneous surgery in the solitary kidney. J Urol, 2009; 181:154–160.