A Novel Model to Predict the Risk of Readmission in Patients with Renal Colic

Abstract

Purpose:

To identify the clinical, laboratory, and imaging parameters that may increase the risk of readmission in patients with renal colic that is managed by active surveillance and to produce a novel model to predict the risk for this.

Patients and Methods:

We retrospectively reviewed patients with renal colic secondary to ureteral calculi admitted to our hospital from March 2009 until September 2010. The colic was managed with active surveillance for 6 weeks. Patients were divided into those who were not readmitted to the hospital within the follow-up period (group A) and those who were (group B).

Results:

From the 452 studied patients, 82 (18.1%) were readmitted to the hospital. Stone size (P<0.001) and location (P<0.001) and serum white blood cell count (P=0.009) were statistically significantly different between groups. These parameters were found to be independent predictors for readmission. A predictive model was produced to calculate the risk of readmission.

Conclusions:

Stone size and location and white blood cell count are independent predictors for potential readmission in patients with renal colic. Using these parameters, we may calculate the risk for readmission, and the latter may assist physicians in identifying the best treatment option.

Introduction

The prevalence of renal stone disease is estimated to be 2% to 20%, and the lifetime risk in Europe and United States is estimated between 5% and 12%.^1,2 Renal colic represents the most common clinical manifestation of urolithiasis causing pain and distress and influencing quality of life.^2,3 It constitutes a significant economic burden, with the annual expenditure of stone disease estimated to reach 2.1 billion dollars in the United States.⁴

Management of acute renal colic has been well established. When urgent intervention is not needed, active surveillance and medical expulsion therapy are options, because a great number of ureteral stones may pass spontaneously.⁵ Although active surveillance represents the most simple and cost effective treatment compared with invasive alternatives, it encrypts the risk of adverse events, such as acute renal failure, urosepsis, and recurrent renal colic resistant to analgesic treatment.⁴

While there in an ongoing interest in active surveillance of ureteral stones, there are no published data to exclusively examine the readmission rates within the follow-up period. Given the worldwide financial constraints in national medical systems, readmission represents an increasingly important topic. Identification of the parameters that can predict a higher risk of readmission may facilitate patient consultation. Furthermore, it may help urologists in the early identification of patients who will benefit most from a nonconservative treatment option, such as extracorporeal shockwave lithotripsy or ureteroscopy.

The aim of the present study is to analyze several clinical, imaging, and laboratory parameters in patients with renal colic to identify the ones that can predict hospital readmission because of adverse events. Besides this, we produced a predictive model consisting of independent predictors to calculate the readmission risk.

Patients and Methods

After the approval from the Scientific Board (Ethics Committee) of our institution, we retrospectively reviewed the medical records of a prospectively maintained database of patients referred to the emergency department with a complaint of renal colic secondary to ureteral calculi. The study period was between March 2009 and September 2010. In all cases, the diagnosis was based on clinical manifestations and confirmed by a noncontrast abdominal CT scan. A complete blood cell count and standard biochemical and urinary analyses were obtained at the acute phase. Acute renal failure was defined as any increase in serum creatinine levels above the upper normal limit (creatinine normal value=0.7-1.2 mg/dL).

The CT scan was used to extract information regarding stone side and location within the ureter, as well as stone maximal diameter. The sacroiliac joints were used to divide the ureter into an upper and lower part. The system of the Fetal Urology Society was used to classify the grade of hydronephrosis.⁶ Patients with fever, clinical symptoms of a urinary tract infection (UTI), or positive urinary culture were excluded from the study.

All patients were discharged with a prescription of oral diclofenac, 75 mg twice a day, for 3 days and tamsulosin, 0.4 mg once a day. A standard follow-up protocol was based on a new CT scan after 6 weeks and reevaluation in the outpatient clinic.

Based on the study concept, we evaluated the patients who were readmitted to the emergency department because of adverse events within the 6 weeks follow-up period. An adverse event was defined as the presence of any of the following: A new onset renal colic resistant to oral medical treatment or a febrile UTI. We divided patients into two groups. Group A consisted by patients who were not readmitted to the emergency department and group B consisted by the patients who were. A second CT scan was performed in all readmitted patients to prove the presence of the stone.

In an effort to identify independent predictors for hospital readmission in patients with renal colic, we evaluated several clinical, laboratory, or imaging parameters. We also produced cutoff values, and we analyzed their potential in prediction of adverse events. Based on the statistical analysis results, we produced a predictive model for the calculation of the probabilities for readmission.

Statistical analysis was performed by SPSS software version 17 (SPSS Inc, Chicago, IL). Numeric variables are present as mean±standard deviation. Categoric variables are described by their absolute number and percent frequency. Mann-Whitney U was used to compare means between numerical variables. Chi-square was used to compare categoric variables.

Cutoff values were obtained by receiver operating characteristic (ROC) analysis. A univariate analysis was performed to estimate the predictive value of several variables in identifying high-risk patients for hospitalization.

A univariate analysis, using logistic regression, was used to identify the predictive significance of stone size, stone location, and white blood cell (WBC) count in prediction of readmission. Multiple logistic regression analysis was then performed analyzing the variables that were found to be statistically important in univariate analysis. Sensitivity, specificity, positive and negative predictive values of the independent predictors were estimated by ROC analysis.

To produce a predictive model, we evaluated the beta “B” weights (produced by regression analysis), and we calculated the probabilities through the equation: probability=odds/1+odds.

All tests were two-tailed, and a P value<0.05 was considered statistically significant.

Results

The study cohort consisted of 361 (79.9%) males and 91 (20.1%) females with a mean age of 44.4 years. Mean stone size was 5.50±2.85 mm, and 231 (51.1%) calculi were identified on the right and 221 (48.9%) on the left side. Regarding stone location within the ureter, 213 (47.1%) stones were found in the upper ureter and 239 (52.9%) in the lower ureter. From 452 studied patients, 82 (18.1%) were readmitted within the 6 weeks of the follow-up period. The mean readmission time was 12.7±10.1 days.

ROC analysis was used to reveal the optimal cutoff values for age, serum WBC count, neutrophils count, and stone size defined as 60 years, 14.5×10⁹/L, 73%, and 7 mm, respectively.

Table 1 shows the characteristics of group A and B patients. There were statistically significant differences between the groups, regarding stone size (>7 mm), stone location, and serum WBC count (>14.5×10⁹/L). Spontaneous stone passage was statistically significantly higher (P<0.001) in patients of group A. Specifically, spontaneous passage was noticed in 202 (54.6%) patients of group A and in 18 (22.0%) patients of group B. Of the remaining 168 (45.4%) patients of group A who were not stone free after the surveillance period, 154 (91.7%) were treated with extracorporeal shockwave lithotripsy and 14 (8.33%) with ureteroscopy. Treatment modalities used for patients of group B were extracorporeal lithotripsy in 55 (85.9%) patients and ureteroscopy in 9 (14.1%) patients.

Table 1.

Study Cohort Characteristics

	Group A	Group B	P
Patients, n (%)	370 (81.9)	82 (18.1)
Sex, n (%)			0.172^a
Male	300 (83.1)	61 (74.4)
Female	70 (18.9)	21 (25.6)
Age (years), mean±SD	44.2±13.9	45.2±13.6	0.632^b
Size (mm), mean±SD	5.18±2.60	6.91±3.43	<0.001^* ^,b
Side, n (%)			0.641^a
Right	191 (51.6)	40 (48.8)
Left	179 (48.4)	42 (51.2)
Location, n (%)			<0.001^* ^,a
Upper ureter	160 (43.2)	53 (64.6)
Lower ureter	210 (56.8)	29 (35.4)
Hydronephrosis, n (%)			0.989^a
≤Grade 1	284 (76.8)	63 (76.8)
≥Grade 2	86 (23.2)	19 (23.2)
WBC (10⁹/L), mean±SD	11.1±3.34	12.1±4.54	0.168^b
Neutrophils (%), mean±SD	71.6±12.8	73.9±11.7	0.192^b
Cr (mg/dL), mean±SD	1.03±0.41	1.06±0.34	0.316^b
Urine analysis
RBC, n (%)	264 (71.4)	63 (76.8)	0.316^a
WBC, n (%)	60 (16.2)	19 (23.2)	0.134^a
Nitrates, n (%)	24 (6.49)	8 (9.76)	0.296^a

Statistically significant.

Chi-square test.

Mann-Whitney U test

SD=standard deviation; WBC=white blood cells; RBC=red blood cells; Cr=creatinine.

Thirty (36.6%) patients of group B were in acute renal failure at the time of readmission. Of them, 12 (40%) patients had been readmitted because of febrile UTI and 18 (60%) because of recurrent pain.

The presence of febrile UTI was the readmission cause in 38 (46.3%) patients, while 44 (53.7%) patients readmitted because of recurrent and resistant pain. Characteristics of patients who were readmitted, according to the cause of readmission, are seen in Table 2. A significant association was found between sex and the cause of readmission (P=0.008). The majority of female patients were readmitted because of febrile UTI while most of the male patients were readmitted because of recurrent pain. Furthermore, patients who were readmitted because of recurrent renal colic tended to return earlier in time after the first episode (P=0.003) in comparison with those patients who were readmitted because of febrile UTI. WBC count almost reached significance (P=0.057), with most of the patients who had increased levels during the first episode to be readmitted because of febrile UTI. There was no statistically significant difference in spontaneous stone passage rates and treatment modality used for stone removal between patients who were readmitted because of febrile UTI or recurrent pain (Table 2).

Table 2.

Characteristics of Patients Who Were Readmitted ^a

	Febrile UTI	Recurrent pain	P
Patients	38 (46.3)	44 (53.7)
Sex
Male	23 (37.7)	38 (62.3)	0.008^* ^,b
Female	15 (71.4)	6 (28.6)
Age (years)	43.8±13.8	46.3±13.6	0.331^c
Size (mm)	7.03±3.72	6.82±3.21	0.892^c
Side
Right	21 (52.5)	19 (47.5)	0.275^b
Left	17 (40.5)	25 (59.5)
Location
Upper ureter	22 (41.5)	31 (58.5)	0.236^b
Lower ureter	16 (55.2)	13 (44.8)
WBC (10⁹/L)	13.0±4.75	11.4±4.25	0.105^c
Neutrophils (%)	75.0±12.4	72.9±11.2	0.384^c
Readmission time (days)	16.1±10.9	9.75±8.41	0.003^* ^,b

Statistically significant.

Numerical variables are present as mean±standard deviation and categorical variables as absolute number and percent frequency.

Chi square test.

Mann Whitney U test.

UTI=urinary tract infection; WBC=white blood cell.

Table 3 shows the results of multivariate analysis. Stone size >7 mm, location in the upper ureter, and WBC count >14.5 were found to be the most significant independent predictors for readmission. Sensitivity, specificity, positive and negative predictive values of these independent predictors are seen in Table 4.

Table 3.

Univariate and Multivariate Analysis

	P	Odds ratio	B	95% CI
Univariate analysis
Size >7 mm	<0.001^*	3.451	1.239	2.020–5.893
Location (upper ureter)	0.001^*	2.399	0.875	1.459–3.944
WBC >14.5 (10⁹/L)	0.010^*	2.100	0.742	1.192–3.700
Multivariate analysis
Size >7 mm	0.001^*	2.655	0.976	1.493–4.722
Location (upper ureter)	0.041^*	1.751	0.560	1.023–2.996
WBC >14.5 (10⁹/L)	0.040^*	1.855	0.618	1.030–3.343
Constant	<0.001^*	0.114	−2.171

Statistically significant.

CI=confidence interval; WBC=white blood cell.

Table 4.

Sensitivity, Specificity, Positive and Negative Predictive Values of the Independent Predictors

	Sensitivity	Specificity	Ppv	Npv
Size >7 mm	36.6%	85.7%	36.1%	85.9%
Upper ureteral stone	64.6%	56.8%	24.9%	87.9%
WBC >14.5 (10⁹/L)	26.8%	85.1%	28.6%	84.0%

Ppv=positive predictive value; Npv=negative predictive value; WBC=white blood cell.

Table 5 presents the predicted probability of a patient to be readmitted to the hospital because of adverse events.

Table 5.

A Predictive Model for Readmission Because of an Adverse Event

Size >7 mm	Upper ureteral stone	WBC >14.5 (10⁹/L)	Probability
x	x	x	10.3%
✓	x	x	23.3%
x	✓	x	16.7%
x	x	✓	17.6%
✓	✓	x	34.7%
✓	x	✓	36.0%
x	✓	✓	27.1%
✓	✓	✓	49.6%

WBC=white blood cell.

Discussion

Approximately 65% of ureteral stones will pass spontaneously.^7,8 Thus, active surveillance of patients with ureteral stones and renal colic represents an acceptable therapeutic alternative.⁹ The addition of medical expulsive therapy in the conservative management of ureteral stones has added an increase in the rate of spontaneous passage, a shorter time to stone passage, and a decrease in the reported pain scores as well.¹⁰

A significant number of patients with renal colic whose colic is managed conservatively, however, will be readmitted during the surveillance period. There are no published contemporary statistical data to describe the characteristics of this category of patients. Based on our database, 18.1% of patients who were admitted to the emergency department of a tertiary hospital, were readmitted. Persistent pain, despite the administration of appropriate oral analgesia, may cause severe discomfort while urosepsis is an emergent medical situation. In these cases, urinary diversion or placement of ureteral stents may be needed. Apart from increased medical risk in these cases, financial cost is significantly increased.

Prompt and effective discernment of patients who are at high risk for adverse events would assist the therapeutic plan and management, and it would result in better patient care and decrease the total treatment cost. To the best of our knowledge, there are no published data on this issue. According to the results of the present study, we found that stone size, location, and WBC count are all independent predictors of readmission. Patients with an upper ureteral stone larger than 7 mm and serum WBC count >14.5×10⁹/L at the acute phase have an increased likelihood for readmission during the follow-up period.

Using these results, we produced a predictive model for calculating the risk of each patient to return to the emergency department with an adverse event. For patients with none of the above factors, the accumulative risk for readmission is 10.3%. In contrast, patients who have all of the above mentioned characteristics have a 49.6% risk for readmission. Apart from the above, we found that the majority of patients who were readmitted because of febrile UTI were women, and these patients had higher levels of WBC count during the first episode of renal colic. In contrast, we found that patients who were readmitted because of recurrent pain were mainly males, and they have the tendency to return sooner after the first episode.

Acute renal failure is a significant complication of urolithiasis that can lead to permanent renal damage if untreated for a long period. Based on the results of the present study, however, acute renal failure can remain unidentified in the absence of concomitant febrile UTI or recurrent pain; renal impairment was found incidentally during work-up in the surveillance period in patients who were readmitted to the hospital because of concomitant sepsis or resistant renal pain and serum creatinine elevation.

Stone size represents the most studied and most significant factor influencing ureteral stones passage. Smaller size stones, regardless of their location in the ureter, had higher spontaneous passage rates—46% to 85% for calculi <5 mm and 36% to 59% for stones 5 to 10 mm.^11,12 The relationship between size and spontaneous passage is roughly linear, with reporting rates to be of 87%, 72%, 47%, and 27% for stones measuring 1, 4, 7, and 10 mm, respectively.⁷

As expected, stone size was found to be a significant predictor for adverse events in our study. A cutoff value of 7 mm was found to be significant. The readmission risk for a patient with a stone >7 mm is 23.3%, and this probability is increased to 34.7% and 36% when this stone is located in the upper ureter or WBC count is >14.5×10⁹/L, respectively. Location of the calculus within the ureter represents another significant factor affecting the probability of a stone to pass spontaneously. This probability has reported to be 12% to 22% for proximal ureteral stones, 22% to 46% for midureteral stones, and 45% to 71% for distal stones.^11,13 In the present study, we found that patients with stones in the upper ureter had an increased risk for readmission during the surveillance period.

WBC count represents an acute phase factor associated with inflammation. The association between WBC and renal stone has been described in a recent study that reported increased levels may related to increased stone-free rates in patients with ureteral stones.¹⁴ The authors reported that this phenomenon may reflect the presence of a movable stone producing a greater inflammation area within the ureter rather than an impacted, nonmovable stone. Regarding our results, a WBC count greater than 14.5×10⁹/L represents a high risk factor for complications providing a 17.6% relative risk.

There are some limitations of the present study that should be addressed. The present study represents a work obtained by a single tertiary center. Because the active surveillance protocol is not the same in all healthcare centers, this may represent a limitation. Although collection of patient data was obtained prospectively, the study was made retrospectively. A prospective study evaluating the findings of the present study would justify and validate their importance. Because of the retrospective design of the study, we did not analyze a number of other laboratory parameters that could be potential predictors of the risk of readmission—for example, C-reactive protein. This issue can be evaluated by a prospective study. Based on the data of our cohort, all of the studied patients either came for consultation at the end of the surveillance period or they were readmitted because of an adverse event. Because the consultation was with different doctors, however, information regarding previous readmissions within the surveillance period may be missed, meaning that the readmission rate may be higher.

The present study produces a novel and simple predicting model calculating the risk of readmission from complications, in terms of urosepsis and resistant pain, in patients with renal colic managed by active surveillance. Based on our knowledge, this is the first study to provide information regarding the readmission rates and the risk for that based on clinical and laboratory parameters. Such a model may assist the identification of patients at high risk for adverse events. For these patients, a more aggressive treatment would save time, obviate severe medical conditions, and decrease the overall cost of therapy. To further evaluate the predictive performance of our model, a prospective internal validation will be initiated in our institution.

Conclusions

In patients with renal colic managed by active surveillance, there is always a risk for readmission. Using stone size, location, and WBC count, we can weigh that risk and consequently identify the best therapeutic management for each patient.

Footnotes

Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Picozzi

, Marenghi

, Casellato

, et al. Management of ureteral calculi and medical expulsive therapy in emergency departments. J Emerg Trauma Shock, 2011; 4:70–76.

Johri

, Cooper

, Robertson

, et al. An update and practical guide to renal stone management. Nephron Clin Pract, 2010; 116:159–171.

Mahmood

, Silbergleit

, Olson

, Cotant

. Urolithiasis: The influence of stone size on management. Nat Clin Pract Urol, 2007; 4:570–573.

Bensalah

, Pearle

, Lotan

. Cost-effectiveness of medical expulsive therapy using alpha-blockers for the treatment of distal ureteral stones. Eur Urol, 2008; 53:411–418.

Griwan

, Singh

, Paul

, et al. The efficacy of tamsulosin in lower ureteral calculi. Urol Ann, 2010; 2:63–66.

Fernbach

, Maizels

, Conway

. Ultrasound grading of hydronephrosis: Introduction to the system used by the Society for Fetal Urology. Pediatr Radiol, 1993; 23:478–480.

Preminger

, Tiselius

, Assimos

, et al. 2007 Guideline for the management of ureteral calculi. Eur Urol, 2007; 52:1610–1631.

Miller

, Kane

. Time to stone passage for observed ureteral calculi: A guide for patient education. J Urol, 1999; 162:688–690.

Hollingsworth

, Rogers

, Kaufman

, et al. Medical therapy to facilitate urinary stone passage: A meta-analysis. Lancet, 2006; 368:1171–1179.

10.

Bader

, Eisner

, Porpiglia

, et al. Contemporary management of ureteral stones. Eur Urol, 2012; 61:764–772.

11.

Hubner

, Irby

, Stoller

. Natural history and current concepts for the treatment of small ureteral calculi. Eur Urol, 1993; 24:172–176.

12.

Segura

, Preminger

, Assimos

, et al. Ureteral Stones Clinical Guidelines Panel summary report on the management of ureteral calculi. The American Urological Association. J Urol, 1997; 158:1915–1921.

13.

Morse

, Resnick

. Ureteral calculi: Natural history and treatment in an era of advanced technology. J Urol, 1991; 145:263–265.

14.

Sfoungaristos

, Kavouras

, Katafigiotis

, Perimenis

. Role of white blood cell and neutrophil counts in predicting spontaneous stone passage in patients with renal colic. BJU Int, 2012; 110:E339–E345.