Do Volume of Injection and Learning Curve Influence the Results of Endoscopic Correction of Vesicoureteric Reflux Using Dextranomer/Hyaluronic Acid? A Single Surgeon Experience

Introduction

Endoscopic treatment (ET) by injection of dextranomer/hyaluronic acid copolymer (Dx/HA) is an established method for correction of vesicoureteric reflux (VUR). The success rate of ET varies widely and is considered to be inferior to ureteric reimplantation surgery.^1,2 Several technical factors have been considered to improve the results of ET. The two common technical variables presumed to influence the success of ET are the volume of injection and surgeon experience.^3–8 Previous reports studying the impact of these variables were confounded by participation from multiple surgeons with varying experience, and inclusion of all grades of VUR.^4,7,9 It is well known that the results of ET are different for low- and high-grade VUR and the results are different in the hands of trainees and expert surgeons. Thus, these reports produced contrasting results, without any consensus. In this article, we tried to evaluate the influence of injection volume and surgeon experience on the success of treating moderate-grade primary VUR by a single surgeon.

Materials and Methods

This was a retrospective review of patients treated with endoscopic injection of Dx/HA for grades 2–4 primary VUR by a single surgeon. Patients with duplicated collecting systems, paraureteric diverticulae, secondary VUR, a history of previous antirefluxing procedures and patients without follow-up imaging were excluded. The standard indications for antireflux intervention included breakthrough febrile infections while on antibiotic prophylaxis or documented febrile urinary tract infection (UTI) after the cessation of prophylaxis. In those children with voiding dysfunction and constipation, our policy is to treat them adequately with urotherapy before offering endoscopic or surgical correction of VUR. Success was defined as the complete resolution of VUR on follow-up imaging performed 3–6 months after the injection. The ureters where VUR persisted after injection were defined as failures. Downgrading of VUR was not considered a success.

The endoscopic injection was performed using the standard STING technique ¹⁰ using a 3.7 F metal needle. Hydrodistention of the orifice was done before and after injection. Dx/HA was injected into the subureteric space at 6 O' clock position until a volcano appearance was achieved with a slit-like ureteric orifice on top of it. For large ureteric orifices (typically corresponding to H3 grade ¹¹ ), a combination of initial intraureteric injection followed by STING was done. The aim was to achieve a classic mound with good orifice coaptation and an orifice that remained closed with hydrodistention (grade H0). This was taken as the end point of injection (Fig. 1). Multiple punctures were avoided where possible. All children received antibiotic prophylaxis for at least 3 months after the injection, which was generally continued until reflux resolution was confirmed. Follow-up ultrasound (USG) was done at 1 month and voiding cystourethrogram (VCUG) or radionuclide cystogram (typically in girls >3 years old) was done at 3–6 months after the injection to assess VUR. The outcome measured was the success of injection, which was defined as the complete resolution of VUR after a single endoscopic injection. A yearly follow-up was advised after VUR resolution was confirmed.

OPEN IN VIEWER

FIG. 1.

Pre- and postinjection appearance of ureteric orifice with grade 4 VUR. VUR, vesicoureteric reflux.

The first step in the analysis was to assess the impact of volume of injection on success of Dx/HA injection by comparing the mean volume of injection between success and failure groups. Then, a receiver operating characteristic curve (ROC) was drawn and area under curve (AUC) was calculated to assess the cutoff value of injection volume that had the best discriminative ability to predict success. We also compared the success independently for two different cutoff volumes (0.8 and 0.9 mL) of injection. Finally, the patients were divided into two groups based on surgeon experience; the patients treated in the first 3 years were considered group 1 and those treated in the last 3.5 years were considered group 2. The two groups were compared with respect to patient age, mean grade of VUR, mean volume of Dx/HA injected per ureter, success, and complications. Statistical analysis was done using statistical software (IBM SPSS version 23); Student's t-test, Fischer exact test, ROC, and AUC with 95% confidence interval (CI) were used, and a P value <.05 was considered significant.

Results

Fifty-eight children (87 ureters, 23 boys) met the inclusion criteria. Most children (45/58, 77%) were evaluated after they presented with febrile UTI, whereas 13 children had antenatally detected hydronephrosis. Every child had at least one ureter with dilating reflux (grades 3 or 4), and most (80%) ureters in both groups had grade 3–4 VUR. Overall, 66/87 ureters (76%) were rendered reflux-free after one injection of Dx/HA. Of these, Data regarding volume of injection was available for 67/87 (77%) ureters, comprising 51 (76%) successes and 16 failures. The mean volume of Dx/HA injected per ureter was significantly higher in the successful cases than in failed cases (0.84 ± 0.15 mL versus 0.72 ± 0.17 mL, P = .008, Fig. 2).

OPEN IN VIEWER

FIG. 2.

Box-whisker chart showing significantly higher volume of injection in successful cases versus failed cases (P = .008).

On further exploring volume of injection as a continuous variable using ROC analysis (Fig. 3), the volume of 0.75 mL had a good discriminative ability to predict success of the ET procedure with maximum sensitivity and lowest false positive rate (sensitivity 72.5%, specificity 68.7%) with an AUC of 0.74 (95% CI: 0.60–0.89). Based on these findings, significantly greater success of injection was demonstrated when a cutoff of 0.8 mL volume was used (success 14/25 ureters for <0.8 mL and 36/42 ureters for 0.8 mL or more volume of injection, P = .0097). There was no difference in success using 0.9 mL as a cutoff volume (success 29/42 ureters for <0.9 mL and 21/25 ureters for 0.9 mL or more volume of injection, P = .2480).

OPEN IN VIEWER

FIG. 3.

Receiver operating characteristic curve analysis of volume of injection. Volume of 0.75 mL had the best discriminative ability to predict success of endoscopic injection.

Considering surgeon experience, there were 28 patients (41 ureters, 10 boys) in group 1 and 30 patients (46 ureters, 13 boys) in group 2. The mean patient age (4.4 and 3.8 years, P = .09) and mean reflux grade (3.3 and 3.2, P = .46) were comparable in both groups. The median volume of Dx/HA injection was 0.8 mL in both groups, which was larger than the 0.75 mL volume obtained as the best cutoff value for success using ROC analysis. However, the success of injection by ureter was higher in group 2 (38/46, 84%) than group 1 (27/41, 65%, P = .04, Fig. 4). There were no procedure-related complications in group 1, whereas there were three Clavien grade 2 complications (three ureters with transient ureteric obstruction) in group 2 (P = .24). All 3 ureteric obstructions occurred within 2 weeks after the injection, improved with observation and expectant management, with no persistent obstruction on follow-up. The follow-up in this study ranged from 6 months to 7 years (median 2 years). After the initial postoperative cystogram to confirm success, further follow-up included only yearly ultrasound and pediatric nephrology consultation. Our policy is to repeat VCUG again only if the child developed febrile UTI during follow-up, which did not happen in any child in the success group.

OPEN IN VIEWER

FIG. 4.

Bar diagram showing success and failures between the two groups. Significantly higher success was obtained in group 2 (P = .04).

Discussion

Ever since the introduction of Dx/HA for treatment of VUR, there has been a rapid increase in the use of ET as the first-line treatment for primary VUR, owing to its minimally invasive nature, acceptable success, and very low complication rate.^12,13 The success rate after endoscopic Dx/HA injection for primary VUR varies widely, between 59% and 95% by refluxing renal unit. ⁴ In this study, the overall success was 76% (66/87 ureters), which is comparable with the 77% success reported in a systematic review. ¹⁴ However, the success of ET still continues to be inferior when compared with those of ureteric reimplantation^1,2; hence there has been a constant effort to improve the success of ET, including several modifications of the injection technique, and the use of higher volume of injection. Still, it is unclear which technical variables are reliable predictors of success of ET. The wide variation in success rates reported in literature may be related to several technical factors, such as shape of the mound, volume of injection, and surgeon experience (learning curve). Some authors found mound morphology to be a significant predictor of success.^5,7 But others found that a good shape of the mound as assessed by the surgeon might not be a good predictor of success. ¹⁵ This is not surprising; being a two-dimensional qualitative subjective assessment by the surgeon, there is an inherent bias in assessing mound morphology. ¹⁶

A more objective variable is the minimum volume of Dx/HA injection required to achieve success, which has been investigated in a few studies with varying results. Kirsch et al. found no statistical difference in injected volume between successes and failures using a mean of 0.83 mL. ³ Watters et al. found that a moderate injected volume (0.8–1.0 mL) was associated with better success, whereas volume >1.0 mL was a significant predictor of treatment failure. ⁵ Paradoxically, on multivariate analysis Yucel et al. ⁷ observed that an injected volume of <0.5 mL was significantly associated with success compared with a volume of >0.5 mL; they speculated that a higher volume of Dx/HA implied a technically more difficult injection, resulting in a poorer outcome. In this report, the mean volume of Dx/HA injected per ureter in successful ureters was significantly higher than the mean injection volume in failed ureters (P = .008). Furthermore, when volume of injection was studied as a continuous variable using ROC analysis, we found that a cutoff volume of at least 0.75 mL Dx/HA was needed to produce success. Since 0.75 mL injection volume is difficult to measure during injection, for practical purposes, 0.8 mL may be taken as minimum volume required for achieving success. We confirmed this when a cutoff volume of 0.8 mL demonstrated significant difference between success and failure. When volume was measured as a continuous variable, Dave et al. ⁴ also found that using a volume of at least 0.8 mL Dx/HA was a potentially important factor to improve success. This optimal volume may ensure optimal fixation of the Dx/HA, prevent its displacement, and may also compensate for the volume loss expected with time. However, unlike other reports where higher injected volumes resulted in failures,^5,7 in this study, volumes in excess of 0.75 mL did not reduce the chance of success; when we used a higher cutoff injection volume of 0.9 mL, we found no statistical difference between successes and failures. We believe that what is important to achieve is an optimal orifice coaptation and grade H0 configuration ¹¹ at the end of injection. The volume of Dx/HA required to achieve this in an individual ureter is variable, and depends on the initial configuration of the ureteric orifice.

Several earlier reports found a definite learning curve associated with the success in ET.^3,4,6 A simple definition of surgical learning curve ¹⁷ is “the number of procedures an average surgeon needs to be able to perform a procedure independently with a reasonable outcome.” Although there is no clear definition of the minimum number of procedures required for a surgeon to achieve optimal results in ET, a review of literature reveals that it may be between 20 and 40 procedures.^3,6 In this study, superior success was achieved in the last 46 ureters (group 2), where the surgeon experience had already exceeded 40 procedures. The key technical points gained with increasing experience appear to be the ability to place the needle to the proper depth and assessment of the pressure and volume of injection material required to create the classic mound. ⁸ Since there is very little tactile feedback in ET, most of the learning is based on visual impression, which explains the significant number of procedures required to achieve optimal results. ⁶

Several authors also observed an increase in the volume of injection with increasing surgeon experience.^4,6,18 This is probably not due to a conscious effort by the surgeon to inject larger volume of Dx/HA; it is likely that with increasing experience, the surgeon overcomes certain inhibitions such as cost of second syringe, fear of wasting the material, and the fear of inducing ureteric obstruction.^4,6 The critical question that remains to be answered is whether the improved success with increasing surgeon experience is because of improved surgeon learning or the increased volume of injection with time. It is likely that both factors play a role. In this study, although the median volume of injection was similar, the success rate was higher in the last 46 ureters (group 2) than the first 41 ureters (group 1). This finding might strongly favor the role of surgeon experience and learning curve in the success of ET. In an interesting study, Dalkilic et al. ⁶ also demonstrated that the improved success was not only related to increased volume of injection, but also due to increasing surgeon experience. They categorized Dx/HA injections into three groups based on increasing experience by the same surgeon. Group 1 was early experience; it had the lowest volume of injection with the poorest success. Although the mean volume of injection was comparable in groups 2 and 3, group 3 with highest surgeon experience had significantly better success than group 2 with intermediate surgeon experience. ⁶

Although this study was retrospective, since our case selection criteria for ET have not changed over the years, the early and recent experience groups were matched for age and mean reflux grade. Similar to others,^3,5 our preference with ET is only for grades 2–4 VUR, and the majority of ureters (80%) treated in this series had grade 3–4 primary VUR. Our technique of injection also has remained constant. Throughout our experience, we used the STING technique primarily, and combined it with intraureteric injection only in large ureteric orifices. We think that these factors, and being a single surgeon experience, gives credence to the results of this study. The preoperative grade of VUR is probably the single most important nontechnical factor influencing the success of ET. ⁸ Even in expert hands, grade 5 VUR has been reported to have poor results with a single Dx/HA injection,^4,19 and may require multiple injections to achieve reasonable success. Multiple injections not only increase the number of general anesthesia sessions, but also increase the cost of the ET significantly. Considering these factors, ureteral reimplantation surgery has been our first choice for older children with grade 5 VUR requiring surgical intervention.

This study has several limitations. Being a retrospective study, there is no randomization, and there may be an inherent case selection bias. Furthermore, being a single surgeon series, the numbers are relatively small, and the records of volumes of injection were available only for about 77% of renal units. However, we believe this analysis offers some insights for improving the success of ET for VUR.

In conclusion, we found that a minimal volume of 0.75 mL Dx/HA injection may be required to achieve success of ET in moderate grade primary VUR. Success also improved with increasing surgeon experience.