The Impact of Stone Density on Operative Complications of Retrograde Intrarenal Surgery: A Multicenter Study with Propensity Score Matching Analysis

Introduction

Retrograde intrarenal surgery (RIRS) is considered the gold standard in the treatment of renal stones up to 20 mm for all caliceal and, if anatomic factors make extracorporeal shock wave lithotripsy (ESWL) unfavorable, lower pole stones.^1,2 It is known that the success rate of RIRS depends on multiple factors, such as stone volume, renal caliceal anatomy, and stone density. ³ Among these, stone density has been shown to be an indicator of stone hardness and is measured on the Hounsfield unit (HU) scale.^4,5 The HU is obtained from a linear transformation of the measured attenuation coefficients during noncontrast computerized tomography (NCCT) reconstruction. With high stone densities, the urologist often has difficulty in stone fragmentation, which might result in higher complication rates due to prolonged operation times and higher intrapelvic pressure.^6–8 Although several studies have investigated the association between stone density and the success of surgical treatment of upper urinary tract stones, very few have examined the effect of stone density on the intraoperative and postoperative complication rates of RIRS.^9,10 To fill this gap, the presented study aimed to examine whether high stone density negatively affects the intraoperative and postoperative complication rates of RIRS.

Materials and Methods

This was a multicenter retrospective matched-pair study. The study protocol was approved by the Institutional Ethics Committee of the Faculty of Medicine of Bulent Ecevit University, Zonguldak/Turkey (approval number: 08.07.2020/2020-14). The data of 579 consecutive patients undergoing RIRS for upper urinary system stone disease from February 2016 to May 2020 at four referral centers in Turkey were analyzed. All operations were performed by surgeons with a minimum of 5 years' experience in RIRS. The patients' demographic characteristics, stone-related parameters (stone volume, location, density, number), and operative outcomes were retrospectively entered into our database. A total of 106 patients with incomplete records or renal anatomical abnormalities and patients who had previously undergone percutaneous nephrolithotomy on the same side were excluded from the study. The remaining 473 patients who had renal and/or upper ureteral stones and underwent RIRS were included. Using a cutoff value of 970 HU, patients were divided into two groups: patients with stone density of ≤970 HU were assigned to a low density (LD) group, and patients with stone density of >970 HU were assigned to a high density (HD) group. Afterward, propensity score matching was performed to reduce the risk of selection bias and control for confounding factors in our analyses. The matching criteria used were as follows: age, stone location, stone volume, stone number, use of a basket catheter, and a ureteral access sheath (UAS). Finally, 340 patients were matched in a one-to-one manner and included for further analysis.

All patients underwent preoperative evaluations that included a detailed anamnesis, a physical examination, and routine laboratory studies, including serum creatinine, urinalysis, and urine culture. All patients also underwent a preoperative radiological evaluation with NCCT. The stone volume was calculated based on the NCCT images using the formula as follows: length × width × height × π × 0.167 (mm³). For cases with multiple stones, the stone size was calculated as the sum of each stone. The stone locations were classified based on NCCT images as follows: renal pelvis, upper, middle, and lower calix, and upper ureter for stones located above the upper border of the sacroiliac joint. The mean HU was used to determine the attenuation value of the stone and was calculated using the elliptical region of interest incorporated into the largest stone area in an axial NCCT image. ¹¹ For cases with multiple stones, stone density was calculated as the largest stone's mean HU value. We used abdominal imaging window settings to evaluate NCCT images (width: 400; level: 40).

The patients' demographic characteristics, stone-related parameters, and operative outcomes were compared between the groups before and after propensity score matching. The primary objective was to examine whether the intraoperative and postoperative complication rates were higher in patients with stone density greater than 970 HU. Complications were classified as intraoperative and postoperative and graded according to the modified Satava classification system (MSCS) and the Clavien classification system (MCCS), respectively.^12,13 Postoperative complications were subdivided into two groups: grade 1 and 2 complications were classified as “minor,” and grade 3, 4, and 5 complications were classified as “major.” The secondary objective was to determine the stone-free rates according to stone density using the cutoff value of 970 HU. Stone-free status was defined as no evidence of residual stone or residual fragments of <3 mm on first-month postoperative NCCT.

Results

Before one-to-one propensity score matching, the clinicodemographic characteristics of 473 patients were analyzed. The LD group comprised 210 patients, and the HD group comprised 263 patients. The median age was significantly greater in the LD group, whereas the mean stone volume and mean stone number were significantly higher in the HD group. After one-to-one propensity score matching, which produced a total of 170 matched pairs, the dissimilarity between the groups was eliminated, and a good balance was achieved. There were no significant differences between the groups in terms of age, gender, body mass index, age-adjusted Charlson comorbidity index, preoperative urinary tract infection, hydronephrosis, and stone characteristics (volume, location, and number). The two groups were also similar in terms of preoperative Double-J stent placement and ESWL history. The baseline demographics and stone-related features before and after matching are presented in Table 1.

The operative outcomes before and after matching are displayed in Table 2. Before one-to-one propensity score matching, stone retrieval using a basket catheter was significantly more common in the HD group, and hospital stay was significantly shorter in the LD group. After one-to-one propensity score matching, the differences in all operative variables between the groups, including basket catheter use and length of hospital stay, were statistically insignificant.

Within the matched cohort, intraoperative complications were noted in 10 patients in the LD group and 15 patients in the HD group. In the LD group, six (3.5%) complications were classified as grade 1, two (1.2%) as grade 2a, and two (1.2%) as grade 2b according to the MSCS. In the HD group, 10 (5.9%) complications were classified as grade 1, 3 (1.8%) as grade 2a, and 2 (1.2%) as grade 2b. The most common intraoperative complication in both groups (n = 9, 2.6%) was failure to reach the stone, resulting in the need for retreatment. There were no grade 3 complications (complications requiring open or laparoscopic surgery) in either group. The intraoperative complication rates did not differ significantly between the LD and HD groups (5.8% and 8.8%, respectively; P = .29). The intraoperative complications are presented in Table 3.

Postoperative complications were noted in 18 patients in the LD group and 26 patients in the HD group (P = .14). In the LD group, most complications were minor (MCCS grade 2 or lower), and the rate of major complications was 1.8%. In the HD group, grade 1, 2, 3b, and 4b complications were observed in 2.9%, 8.8%, 1.8%, and 1.8% of the patients, respectively, while there were no grade 5 complications. There were no grade 4a complications in either group. The most common postoperative complication in both groups (Total number = 21, 6.2%) was fever requiring antibiotherapy. The most common major postoperative complications were steinstrasse (n = 3) in the LD group and steinstrasse (n = 3) and urosepsis (n = 3) in the HD group. The incidence of postoperative complications did not differ significantly between the LD and HD groups (10.6% and 15.3%, respectively; P = .14). The overall complication rates were also similar (15.2% and 21.1%; P = .16). Detailed descriptions of the postoperative complications are listed in Table 4.

After matching, stone-free status was achieved in 143 patients (84.1%) in the LD group and 148 patients (87%) in the HD group; the difference was not statistically significant (P = .27). Retreatment for residual stones was required in 18 (10.5%) patients in the LD group and 16 (9.4%) patients in the HD group; this difference was also statistically insignificant (P = .62).

Discussion

In this multicenter study, we focused on investigating whether there is an association between stone density and operative complications of RIRS. It has previously been shown that stone density is an important predictor of ESWL failure and the need for additional sessions, which might result in higher complication rates.^14,15 To avoid time-consuming and additional cost–incurring auxiliary procedures, the American Urological Association guidelines suggest that RIRS might be more suitable for the management of renal stones with densities greater than 1000 HU.^1,2 However, the success of RIRS in HD stones may be questioned since these stones often are not easily fragmented. A validated cutoff value affecting treatment outcomes of RIRS has not been reported in known studies. Therefore, we used a cutoff value of 970 HU for stone hardness since it was clearly shown in a prospective, high volume study that a stone density over 970 HU has a poor chance for stone clearance with ESWL. ¹⁶ Applying this cutoff value of 970 HU, our results, however, showed that higher stone densities had no significant effect on operative complications. As a secondary objective, we found that stone density did not affect the stone-free rates.

A large-series single-center study found that the stone volume and number were associated with higher operative complication rates. ¹⁷ Among demographic variables, age and gender have also been reported to be significant predictors of RIRS complications. ¹⁸ Before propensity score matching, we found that stone retrieval using a basket catheter during surgery was an independent predictor of postoperative complications. To minimize the effect of these confounding factors, we matched our cohorts in a one-to-one manner. To our knowledge, this is the first matched-pair study to investigate the association between stone density and RIRS operative complications. After matching, we found that higher radiological stone densities had no significant effect on either intraoperative or postoperative complications. The overall complication rate in the HD group was 21.1%, which is consistent with overall, nondensity-specific rates reported in the literature (8%–24.9%).^19,20

In this study, even though the intraoperative complication rate in the HD group was slightly higher than in the LD group, the difference was not statistically significant. The most common intraoperative complication in both groups was failure to reach the stone. However, most of these cases were reluctant to have a second surgery and were therefore classified as grade 1 according to the MSCS. Subgroup analysis showed no significant differences in terms of grade 1 and 2 complication incidence between the two groups. The overall intraoperative complication rate in the HD group (8.8%) is also comparable to overall, nondensity-specific rates reported in the literature (7%–30%).^19,20

A single-center retrospective study reported that most intraoperative complications were minor, with insignificant clinical consequences. ²¹ This finding is in line with our results. In that study, mild hematuria was the most common grade 1 complication, and failure to reach the stone resulting in the need for retreatment was the most common grade 2 complication. However, it is unclear whether stone density was an important factor for intraoperative complications, as it was not taken into account in that study. In this respect, the presented study could not show a correlation between higher stone density and a higher likelihood of intraoperative RIRS complications.

In this study, postoperative complications were categorized according to the MCCS, which is considered the most suitable approach for classifying surgical complications following upper urinary tract stone treatment. ²² Before matching, we found that the stone volume and the use of a ureteral basket catheter were predictors of postoperative complications after RIRS. After matching these parameters, we observed no statistically significant differences in the postoperative complication rates between the LD and HD groups. A recent retrospective study that used MCCS to classify postoperative complications of RIRS found that the most common complication was fever and reported a postoperative complication rate of 26%. The authors found that preoperative urine culture, irrigation time, and operation time were reliable predictors of postoperative complications and concluded that these factors should be carefully considered in the preoperative management of patients who are to undergo RIRS. ²³ In line with these findings, the most common postoperative complication in both groups in our study was fever. There were no significant differences between the groups in terms of major complication incidence. The postoperative complication rate in the HD group (15.3%) is also comparable to overall, nondensity-specific rates reported in the literature.^23,24

In our study, there were no statistically significant differences between the LD and HD groups in terms of stone-free and retreatment rates. More recently, the impact of stone density on the outcomes of RIRS was evaluated in a retrospective, single-center study which categorized patients into subgroups according to stone location. It was concluded that stone volume and number were the significant independent factors for treatment failure in kidney stones. However, the higher mean stone density was the only significant predictor of treatment failure in ureteral stones but not in kidney stones. ⁶ In the presented study, increased mean stone density was not a significant predictor of stone-free status in multivariate analyses both in kidney and upper ureteral stones. Moreover, the stone-free rate in the HD group (87%) is also consistent with previously reported nondensity-specific rates (70%–97%).^25,26 These findings suggest that RIRS should be confidently recommended to patients suffering from upper urinary tract stones with HD.

Certain limitations of this study should be noted. First, this was a retrospective study and was therefore subject to possible selection bias. To minimize this risk, we performed a one-to-one propensity score matching analysis and included previously published confounding factors as covariate-matching factors. Second, as this was a multicenter study, surgeons from the different centers may have influenced the clinical outcomes. However, all surgeons had at least 5 years of experience performing RIRS, and the surgical technique in the participating centers has been standardized since 2015. In addition, we manually measured the mean stone density using the mean attenuation values in axial NCCT images, which may have caused inter-reader variability. There is currently no consensus regarding the most appropriate stone density measurement technique. Standardization using a validated method could provide more accurate results. Finally, although it is a match-paired study, it has a limited number of patients that cause insufficient statistical power to detect definitive conclusions. If the study was designed as a high volume prospective study, there would have been a more concise conclusion for the results.

Conclusion

This multicenter study shows that higher stone densities have no significant effect on the operative complication and stone-free rates in patients undergoing RIRS. Although it has been demonstrated that higher stone densities are associated with a higher likelihood of ESWL failure, the results of this study suggest that HD stones can be safely and successfully treated by RIRS.