The Efficacy of Guy's Stone Score for Predicting the Stone-Free and Complication Rates in Children Treated by Percutaneous Nephrolithotomy

Introduction

Pediatric renal stone disease (RSD) mostly occurs secondary to metabolic abnormalities, recurrent urinary tract infections, and anatomic problems of the urinary system. RSD is a global health problem with an incidence of ∼1%–5% in developed countries. ¹

Due to the recurrent nature of RSD, minimally invasive surgical techniques as percutaneous nephrolithotomy (PCNL) is now preferred as a treatment strategy for protecting renal configuration and function even in large and/or complex renal stones. Although PCNL is defined as minimally invasive, the technique has the risk of serious complications and does not always render the patient stone free. ²

Tomas and colleagues developed Guy's stone score (GSS) by using preoperative radiologic findings for predicting outcomes of PCNL in adult population (Fig. 1). ³ GSS provides a systematic and quantitative evaluation of renal stones. This standardization enables us to compare the outcome of minimally invasive procedure with respect to other similar series. This scoring system is accepted as an effective tool for predicting success and complication rates after validation with large adult PCNL series. ^{3 –5} However, limited studies were presented and conflicting results have been reported in literature as to whether GSS could predict stone-free or complication rates in children. ^{6 –9} The aim of this study was to determine the efficacy of GSS for predicting stone-free and complication rates in children undergoing PCNL.

OPEN IN VIEWER

FIG. 1.

Guy's stone score; based on all stones seen not just those targeted in procedure. Abnormal anatomy is defined as abnormal renal anatomy, an abnormal collecting system, or a patient with difficult access. Stent encrustation does not affect the score (3).

Patients and Methods

After institutional approval, a total of 197 renal units (RUs) in 173 children who referred to our unit and consecutively underwent PCNL operation were retrospectively reviewed.

Demographic data as age and gender, clinical information as surgical history, preoperative radiologic evaluation, solitary kidney or kidney abnormalities, size and location of stone, information about operation, complications, stone-free status, and duration of hospitalization were received from standard forms for PCNL procedure, which was filled in for each patient just before and after operation. Preoperative CT or plain abdominal radiography (KUB)/intravenous urography was evaluated and classified according to the description of GSS. ³ All children were independently scored by two clinicians who were unaware of the results (two urology trainees—F.S.E. and O.Ö.). Stone burden was calculated by multiplying the longest two dimensions obtained from KUB or CT.

Operations were performed under general anesthesia at prone position and first or third generation cephalosporin was used for preoperative antibiotic prophylaxis. All percutaneous access was gained under fluoroscopy guidance by using Amplatz or balloon dilator in the operation room by the urology team. Pneumatic or holmium laser lithotripter was used to fragment stones. Ten to fourteen French nephrostomy tubes were placed into tract at the end of the operation. Operation time was defined as the time from insertion of ureteral catheter until suturing of the skin.

Patients were examined with KUB and ultrasound (USG) in second postoperative day for clamping the nephrostomy tube and tube was removed in patients without symptoms. This imaging is also helpful in evaluating the early results after PCNL. If the patient had symptoms such as pain after clamping the nephrostomy tube, further investigation for presence of residual stone was planned. The symptomatic patients with nonopaque stones and suspicion of residual stone in KUB were scanned with CT.

PCNL success was evaluated by radiologic imaging based on stone characteristics (USG and KUB or CT) at third postoperative month. Results were designated as stone free or having residual stones (any evidence of persistent stone fragments irrespective of size). The absence of any stone in radiologic evaluations was considered as stone free.

Preoperative complications were graded with Satava classification; postoperative complications were graded with modified Clavien classification. ^10,11

Groups were compared for factors associated with stone-free status and the risk of complication.

Statistical analysis was done by SPSS, version 15.0. Chi-square and Mann–Whitney U test were used for univariate analyses. Significant variables on univariate analyses were included in a multivariate logistic regression analysis, to evaluate factors associated with stone-free and complication rates.

Results

A total of 197 RUs in 173 children (71 girls, 102 boys) with median age of 6 years (1–17 years) were included to study.

Five patients (3%) had solitary kidney. Renal anomaly was identified in four patients (2%). Three of them had horseshoe kidney and one had a duplex system. Seventy-two patients had previous history of ipsilateral stone treatment (42%). Characteristics of the children, stone, and PCNL operation were summarized in Table 1. Generally multiple stone formations are high (44%) and single stones were mostly located in calyceal system (19%).

Univariant analyses of these characteristics according to stone-free status are detailed in Table 2. Localization of stone significantly differs between groups in univariant analyses. However, in multivariate analysis, stone location is not defined as a factor associated with stone-free status.

Median stone burden of 197 RU was 4 cm² (range 1–21). Stone burden was higher in residual stone group, but was not significantly different between other groups (p = 0.17). Single tract with a subcostal access location is preferred in most patients. Any significant difference was not detected in operation characteristics, including operation time and length of hospital stay between groups (0.15, 0.55, respectively).

All RUs were classified according to GSS before the intervention. Forty-seven RUs (24%) were grouped as 1, 90 RUs (46%) were grouped as 2, 32 RUs (16%) were grouped as 3, and 28 RUs (14%) were grouped as 4.

Overall, stone-free rate was 77% for groups. Stone-free rate was found to be 89% for GSS group 1, 78% for group 2, 75% for group 3, and 57% for group 4. GSS was significantly associated with stone-free status in univariate analysis (Table 3). When analyzed by using multivariate logistic regression, GSS was determined as the only significant factor predicting the stone free status (Table 4). Stone-free rates in GSS group 4 were eight and four times lower, respectively, when compared with those of GSS group 1 and 2.

Overall complication rate was 17%. Preoperative complications were mostly group 1 and postoperative complications were mostly grade 2. RUs were evaluated for an association between GSS and risk of complications. There is no significant association detected between GSS and the overall complication rate as well as complications requiring intervention (p = 0.24, p = 0.42, respectively) (Table 5).

Discussion

The main subject of pediatric PCNL series in the literature is mostly about the factors associated with complications and the stone-free status. However, few of them focused on predictive value of scoring systems and conflicting results have been reported about GSS association with the outcomes. ^{6 –9} In the present study, we aim to evaluate the efficacy of GSS for predicting the stone free status and risk of complication in pediatric patients undergoing PCNL. With respect to multivariate analysis, GSS was found to have a predictive ability for stone-free status, however, is not associated with the risk of complication. We believe that although GSS is limited, it might provide useful insight for clinicians, specifically while making a decision to choose treatment options and counseling to the parents of children.

PCNL monotherapy in pediatric population leads to complete stone removal in about 80%–94% of patients in all age groups. In addition, success rates were in downward trend in partial staghorn and complex staghorn stones. ^{6 –9} In our series, overall complete stone-free rate following a single session was 77%. When considering small residual stones, the success rate increases to 85%. However, in children, there is still no clear consensus regarding clinically insignificant residual fragments. Therefore, we prefer to report the complete stone-free rate. Hence, stone-free rate of our study is typical and likely to be comparable to other pediatric PCNL series.

Recently, GSS ≥2 was declared as a predictor of failure. ^5,7 Consistent with these studies, in our present study, GSS ≥2 was associated with prediction of the residual stone. Our multivariate analysis showed that stone-free rate in GSS group 4 was found to be eight and four times lower than those in GSS group 1 and 2, respectively.

Other factors associated with the success of the procedure defined in the previous studies were stone burden and multicaliceal involvement. ^7,9 Consistent with the previous studies, multicaliceal involvement is associated with lower stone-free rates in our univariate analysis, however, this association was not approved by multivariate analyses.

Also, our study did not define an association between stone burden and stone-free rates, in contrast to other PCNL series. However, according to the GSS, partial and complex staghorn stone are defined in group 3 and 4, respectively. One possible explanation is that partial or complete staghorn stone sizes in children may vary depending on age and kidney volume. As an example, although a 20 × 15 mm single stone in 10-year-old children is classified as GSS group 1 or 2 according to the GSS; in a 2-year-old child, the stone that is the same size, but partial or staghorn, is classified as group 3 or 4.

Increased use of the PCNL was associated with potential complications, despite the higher success rates. The effectiveness of the GSS in predicting the complication were validated in adult PCNL series, however, conflicting results have been reported in the pediatric population. Pediatric series reported that the complication rate was 29%–39% with PCNL monotherapy; however, multivariate analysis did not support the GSS and complication correlation. ^6,9 Most of these complications were minor, mostly Clavien 1 or 2.

Conversely, Utangaç and colleagues reported 12% for the rate of complication after first miniperc session and GSS ≥2, multicaliceal involvements, including partial staghorn and complex staghorn stone, were regarded as the risk factors for postoperative PCNL complications. ⁷ Our overall complication rate was 17%, in addition, we did not find a significant association between GSS and complication rate or GSS and intervention requiring complication rate. Our complication rate was quite comparable with that reported by Utangaç and colleagues. In contrast to the last mentioned study, we found similar complication rates in GSS 1 and 4 (13% vs 14%). This might be due to the fact that we did not insist on providing stone-free status in PCNL monotherapy in GSS 3 and 4 patients. This protocol may lead to less complication rates.

Operation technique and experience of the surgeon was reported as another factor that affected the PCNL outcome. ⁹ In this present study, all interventions were directed by a senior experienced endourologist (B.O.), and therefore, surgeon experience was not evaluated as a factor in our study results.

GSS provides an advantage of multifactorial evaluation, however; GSS was designed for the adult populations. Pediatric kidney and stone diseases are in many ways different from adult kidney and stone diseases. In pediatric populations, complex stone is not only associated with stone burden and also kidney size, and its configuration changes with age. In addition, stone density and composition differ according to underlying metabolic disorders. Congenital anatomic abnormalities limit the surgical intervention performance. Therefore, only one formal scale such as GSS may not fully reflect the complexity of stone in children.

When current literature and our study's results are interpreted together, GSS has a significant predictive ability for stone free status after pediatric PCNL. However, designing a new scoring system or modifying GSS in a large multicentric study according to the characteristics of pediatric population would ultimately provide a better prediction in RSD.

One main limitation is the retrospective design of our study and this may create a bias. Another limitation in this study is the fact that we use different radiologic modalities and this can affect the homogeneity of preoperative classification and interpretation of our outcome. To minimize radiation exposure (as low as reasonably achievable procedure), we avoid the CT evaluation after PCNL. However, stone-free status was determined by using two radiologic evaluation methods (USG and KUB). CT was used in conditions such as suspecting a residual stone, requirement of intervention, and nonopaque stones. Despite these limitations, our results conclude that GSS might be a helpful scoring system for predicting the stone-free status, but not a strong predictor of PCNL complications.

Conclusion

Current study revealed that GSS has a predictive ability for stone-free status, however; GSS is insufficient for predicting complications after pediatric PCNL. Even though GSS does not fully reflect the characteristics of the pediatric population, the scoring system might be a helpful tool for clinicians in counseling the parents of children with urolithiasis and in recommending treatment options. Further large-scale studies based on characteristics of pediatric renal anatomy and stone might overcome these disadvantages.