Comparative evaluation of urolithiasis management options in patients with horseshoe kidney: A systematic review and meta-analysis

Abstract

Background

Horseshoe kidney is a rare congenital anomaly commonly complicated by urolithiasis. Extracorporeal shockwave lithotripsy (ESWL), ureteroscopy (URS), and percutaneous nephrolithotomy (PCNL) are treatment options for horseshoe kidney stones. The aim of this systematic review is to compare the benefits and risks of these management options.

Methods

MEDLINE, EMBASE, and Cochrane Library databases were searched from inception to February 2022. A total of 516 nonduplicate studies were screened against the inclusion and exclusion criteria. Studies comparing at least 2 interventions with ≥10 patients per intervention were included.

Results

Nine retrospective observational studies published from 2007 to 2021 with a total of 565 patients were included. Reported mean ± SD or mean (range) stone sizes ranged between 17.90 ± 2.43 mm and 27.9 ± 8.6 mm for PCNL, 8.4 (2–25) mm and 22.3 ± 9.1 mm for URS, and 11.9 ± 2.0 mm and 16.8 ± 4.4 mm for ESWL. There was no difference in single-session and overall stone-free rate (SFR) between PCNL and URS, with a risk ratio of 1.04 (95% confidence interval, 0.95–1.13; I² = 20.63%). Ureteroscopy had better stone clearance than ESWL, with an overall SFR risk ratio of 1.38 (95% confidence interval, 1.04–1.82; I² = 0%). There was no statistically significant difference in overall SFR between PCNL and ESWL. Most patients who underwent URS and ESWL experienced Clavien-Dindo (CD) grade I-II complications. Percutaneous nephrolithotomy was associated with the highest complication rates, including 5 CD grade III and 3 CD grade IV complications and a mean postoperative hemoglobin drop of 0.47 to 1.83 g/dL. There were no CD grade V complications across all studies.

Conclusions

There was no difference in SFR between PCNL and URS. Ureteroscopy was associated with a smaller stone burden and fewer and less severe complications. Ureteroscopy was found to be more effective than ESWL with a higher SFR and comparable safety profile. Further large-scale randomized controlled trials are needed to confirm these findings.

1. Introduction

Horseshoe kidney (HK) is the most common congenital renal anomaly, with a reported incidence ranging from 1/400 to 1/666 and a prevalence of 0.25% in the general population.^[1] Horseshoe kidney develops as a result of fusion of the lower poles of the kidneys during embryological development.^[2] An isthmus of functional renal parenchyma leads to arrest of normal embryological ascent and rotation of the kidneys, resulting in malrotation and anterior displacement of the collecting system.^[3] Patients with HK have associated ureteropelvic obstruction and impaired drainage of the collecting system, resulting in urine crystal aggregation and retention. These anatomic changes, frequently associated with concomitant metabolic abnormalities, increase the risk of kidney stone formation in these patients. In fact, urolithiasis is the most common complication of HK, with an incidence rate of up to 60%.^[3–5]

Urolithiasis in HK patients was historically managed with open surgical approaches. However, with advances in technology, minimally invasive surgical treatment options have been increasingly adopted. Several studies have reported the efficacy and safety of extracorporeal shockwave lithotripsy (ESWL), ureteroscopy (URS), and percutaneous nephrolithotomy (PCNL) in HK patients.^[1] The evidence for URS versus ESWL in HK urolithiasis has been summarized by Chen et al.^[6] and Yi et al.^[2] in their published systematic reviews (SRs). Both treatment options have been found to be effective and safe for HK stones <20 mm. However, URS offers a superior stone-free rate (SFR), up to 76%, and a lower risk of complications, likely due to high rates of retreatment associated with ESWL.^[2,6,7] On the other hand, several guidelines recommend PCNL as first-line treatment for stones greater than 20 mm, while counseling patients on the high potential rate of complications, up to 83%, associated with this procedure.^[7–9]

Overall, there remains a lack of clear consensus on the preferred treatment option for HK patients with urolithiasis. To our knowledge, there has been no published study that compares and summarizes all the evidence on PCNL, URS, and ESWL for HK stones. Therefore, the aim of this SR is to comparatively evaluate the benefits and risks of these 3 different therapeutic modalities to help inform patient and physician decision-making.

2. Evidence acquisition

2.1. Search strategy

This SR was conducted in accordance with the Cochrane review guidelines and the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement.^[10,11] Database searches were conducted in MEDLINE, EMBASE, and the Cochrane Library for English language articles published from database inception to February 8, 2022. Search terms included: (urolithiasis or nephrolithiasis or stone or calculi or calculus) AND (“fused kidney” or (fusion and kidney) or “horseshoe kidney” or HSK or HK). Searches were conducted independently by 2 review authors (Y.Q. and A.S.). After exclusion of duplicate citations, screening of titles and abstracts followed by review of full-text articles was performed independently by 2 review authors (Y.Q. and M.A.T.) for final inclusion, with any disagreements resolved by a third review author (N.D.). The authors of the included studies were contacted for any missing or incomplete data.

2.2. Study design criteria

We included only studies that were randomized controlled trials, nonrandomized comparative studies, or comparative observational (prospective or retrospective) studies that compared at least 2 different interventions with ≥10 patients per arm. Case reports, conference abstracts, letters, and animal studies were excluded. Studies lacking SFR outcome data specific to HK patients were excluded.

2.3. Study participant criteria

All HK patients undergoing an intervention for management of urolithiasis were included. No patient data were excluded based on the size or number of stones. Studies that included patients with anomalies other than HK were included in the analysis if HK-specific SFR was reported.

2.4. Types of interventions

Any study evaluating URS, ESWL, and/or PCNL for urolithiasis in HK patients was included. Any technique or lithotripsy modality was accepted for these interventions. Studies that included open surgery as the only comparator were excluded.

2.5. Outcomes

The primary outcomes of interest were initial postoperative SFR and overall or final SFR, measured at any point up to 3 months postoperatively and using any imaging modality. Secondary outcomes included the number of auxiliary procedures postoperatively and rate of retreatment. Data on the number of procedures, procedure duration, radiation exposure time, and length of hospital stay were also recorded.

Intraoperative and postoperative complication data were recorded according to the Clavien-Dindo (CD) classification system when sufficient data were available.^[12] Data on postoperative mean drop in hemoglobin (Hb) level were also recorded when sufficiently reported.

2.6. Risk-of-bias assessment

The included studies were evaluated for risk of bias (ROB) using the Newcastle-Ottawa Scale.^[13] Studies with a Newcastle-Ottawa Scale score ≤5 were designated as low-quality studies, 6 to 7 as intermediate-quality studies, and 8 to 9 as high-quality studies.

2.7. Data analysis

Meta-analysis of data from the included studies was conducted using Stata v17 (StataCorp, College Station, TX). Treatment effect was evaluated using the Mantel-Haenszel method and risk ratios (RRs) with 95% confidence intervals (CIs). Statistical significance was assessed using a 2-tailed test with p < 0.05 considered significant. Statistical heterogeneity was assessed using a χ² test and I² test, with p < 0.05 and I² > 50% indicating heterogeneity. Random-effects models were used. A formal assessment of publication bias was not possible because of the low number of included studies.

3. Evidence synthesis

3.1. Description of the included studies

Database searches yielded a total of 1347 studies. Following exclusion of duplicates, 516 citations were eligible for abstract screening. Subsequently, 27 full-text articles were assessed against eligibility criteria, resulting in 9 studies being included in the qualitative synthesis and meta-analysis (Fig. 1). All included studies were retrospective observational cohort studies, resulting in a total of 565 patients.^[13–22] Studies were published from 6 different countries, mostly between 2015 and 2021 except for 1 study (Viola et al.^[22]) published in 2007.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram of studies identified, excluded, and included. HK=horseshoe kidney.

3.2. Study characteristics

Characteristics and ROB assessment of the 9 included studies are summarized in Table 1. Baseline characteristics include the number of patients per intervention arm, age, gender, body mass index, stone size and location, and length of follow-up. In total, data from 565 patients were analyzed, of whom 267 underwent PCNL, 197 URS, 87 ESWL, and 14 PCNL with antegrade URS. Six studies compared 2 interventions. Only data from the 2 intervention arms containing more than 10 patients were included in the analysis from 1 study that compared 3 interventions^[19] and 2 studies that compared 4 interventions,^[16] as the remaining intervention arms contained fewer than 10 patients.

Table 1.

Characteristics of included studies.

Study author, country, publication date	Study design	Study duration	Intervention	No. patients	Gender, n (%)	Age, mean ± SD, range, yr		BMI, mean ± SD, kg/m2		Chief presenting complaint, %	No. stones	Single vs. multiple stones, n (%)	Stone size, mean ± SD or mean (range), mm	Stone location, n (%)	Length of follow-up	NOS score
Abdeldaeim et al.,[14] Egypt, 2021	Retrospective cohort	2016–2020	PCNL	30 (10 HK patients)	25 (83.3%) male, 5 (16.7%) female	42.53 ± 10.47, NR		29.3 ± 2.78		Pain, 66.7%	NR	All single stones	17.90 ± 2.43	Renal pelvis: 25 (83.3%) Lower calyx: 4 (13.3%) Upper calyx: 0 (0%) Middle calyx: 1 (3.3%)	12 wk	6
			URS	30 (10 HK patients)	23 (76.7%) male, 7 (23.3%) female	41.07 ± 10.71, NR		28.80 ± 3.64		Pain, 40%			14.97 ± 3.50	Renal pelvis: 17 (56.7%) Lower calyx: 9 (30%) Upper calyx: 1 (3.3%) Middle calyx: 3 (10%)
Kartal et al.,[15] Turkey, 2019	Retrospective cohort	2010–2017	PCNL	21	18 (85.6%) male, 3 (14.3%) female	41.5 ± 9.9, NR		25.9 ± 2.3		Pain, 52.4%	Mean ± SD: 1.8 ± 1.0	Single: 13 (61.9%) Multiple: 8 (38.1%)	24.5 ± 8.1	Renal pelvis: 6 (28.6%) Lower calyx: 5 (23.8%) Upper calyx: 2 (9.5%) Middle calyx: 2 (9.5%) Mixed calyx: 6 (28.6%)	3 mo	8
			URS	28	23 (82.1%) male, 5 (17.9%)	43.2 ± 8.4, NR		27.0 ± 2.6		Pain, 60.7%	Mean ± SD: 1.4 ± 0.6	Single: 18 (64.3%) Multiple: 10 (35.7%)	22.3 ± 9.1	Renal pelvis: 9 (32.1%) Lower calyx: 7 (25%) Upper calyx: 3 (10.7%) Middle calyx: 3 (10.7%) Mixed calyx: 6 (21.4%)
Singh et al.,[16] India, 2019	Retrospective cohort	1990–2014	PCNL	67	65 (76.5%) male, 20 (23.5%) female	All patients: 38.2, 7–67		All patients: 21.23		All patients: pain, 75%	NR	Multiple: 21 (31.3%)	DSA in mm2/mean (range): 595.5 (99–3 850)	Renal pelvis: 34 (44.7%) Calyx: 25 (32.9%) Pelvis + calyx: 17 (22.3%)	1 mo to 12.5 yr	6
			URS	10								Multiple: 3 (30%)	DSA in mm2/mean (range): 142.4 (28–290)	Pelvis: 3 (30%) Calyx: 4 (40%) Pelvis + calyx: 1 (10%) Ureter: 2 (20%)
Eryildirim et al.,[17] Turkey, 2018	Retrospective cohort	2007–2016	PCNL	38	56 (63.6%) male, 36 (36.4%) female	42.97	All patients: 41.16, 19–60	29.15	All patients: 28.27	NR	NR	NR	21.55	Renal pelvis: 12 (31.6%) Lower calyx: 9 (23.7%) Upper calyx: 3 (7.9%) Middle calyx: 3 (7.9%) Multiple calyces: 11 (28.9%)	3 mo	7
			URS	50		39.78		27.6					19.46	Renal pelvis: 18 (36.0%) Lower calyx: 13 (26.0%) Upper calyx: 5 (10.0%) Middle calyx: 6 (12.0%) Multiple calyces: 8 (16.0%)
Ergin et al.,[18] Turkey, 2017	Retrospective cohort	2005–2015	PCNL	60	39 (65%) male, 21 (35%) female	46.3 ± 18.6	All patients: 39, 1–72	NR		NR	NR	NR	27.9 ± 8.6	Renal pelvis: 18 (20.5%) Lower calyx: 36 (40.9%) Upper calyx: 14 (15.9%) Middle calyx: 20 (22.7%)	NR	6
			URS	36	24 (66.67%) male, 12 (33.33%) female	40.9 ± 15.9							17.8 ± 4.5	Renal pelvis: 8 (22.2%) Lower calyx: 11 (30.6%) Upper calyx: 7 (19.4%) Middle calyx: 10 (27.8%)
Blackburne et al.,[19] USA, 2016	Retrospective cohort	2002–2016	PCNL	33	All patients: 31 (69%) male. 14 (31%) female	51.2, 23–75		NR		NR	NR	NR	22.4 (10–57)	NR	Mean 20.5 (range 0–118) mo	7
			URS	22		48.1, 29–78							8.4 (2–25)
Gokce et al.,[20] Turkey, 2016	Retrospective cohort	2003–2014	URS	23	18 (78.3%) male, 5 (21.7%) female	44.2 ± 9.9	All patients: 42.5 ± 8.2, 16–78	NR		NR	Total: 32	Multiple: 9 (39.1%)	17.1 ± 5.1	Pelvis and upper pole: 17 (73.9%) Lower pole: 6 (26.1%)	6 wk	7
			ESWL	44	32 (72.7%) male, 12 (27.3%) female	42.8 ± 8.4					Total: 52	Multiple: 8 (18.1%)	16.8 ± 4.4	Pelvis and upper pole: 32 (72.8%) Lower pole: 12 (27.2%)
Ding et al.,[21] China, 2015	Retrospective cohort	2005–2014	PCNL	19	15 (78.9%) male, 4 (21.1%) female	44.3 ± 10.9	All patients: 43.4 ± 10.4, 24–66	NR		NR	NR	NR	27.1 ± 6.0	NR	1 mo	7
			PCNL with antegrade URS	14	11 (78.6%)male, 3 (21.4%) female	47.7 ± 8.5							28.5 ± 5.0
			URS	18	14 (77.8%) male, 4 (22.2%) female	42.9 ± 11.6							18.9 ± 3.6
			ESWL	11	9 (81.8%) male, 2 (18.2%) female	36.6 ± 8.2							11.9 ± 2.0
Viola et al.,[22] UK, 2007	Retrospective cohort	1987–2002	PCNL	19	All patients: 36 (81.8%) male, 8 (18.2%) female	52.5, 3–83	All patients: 50, 3–83	NR		Hematuria, recurrent UTI, pain	NR	Single: 5 Multiple: 14	DSA in mm2/mean (range): 197 (6–2400)	NR	Mean 42.3 (range 3–144) mo	5
			ESWL	25		48, 10–71						Single: 18 Multiple: 1	DSA in mm2/mean (range): 91 (10–1600)		Mean 36.5 (range 1–91) mo

BMI = body mass index; DSA = digitalized surface area; ESWL = extracorporeal shockwave lithotripsy; NOS = Newcastle-Ottawa Scale; NR = not reported; PCNL = percutaneous nephrolithotomy; URS = ureteroscopy; UTI = urinary tract infection. HK = horseshoe kidney.

Mean stone size was reported by all studies except 2, those of Singh et al.^[16] and Viola et al.,^[22] which instead reported digitalized surface area. The mean ± SD or mean (range) stone size for PCNL patients ranged from 17.90 ± 2.43 mm to 27.9 ± 8.6 mm and 8.4 (2–25) mm to 22.3 ± 9.1 mm in URS patients and 11.9 ± 2.0 mm to 16.8 ± 4.4 mm in ESWL patients. Ding et al.^[21] is the only study that included 14 patients who underwent PCNL with antegrade URS; the mean stone size for that intervention arm was 28.5 ± 5.0 mm. A majority of studies, except for 3, describe the specific location of treated stones, as shown in Table 1. The exact length of follow-up was not reported by all studies; therefore, it was extrapolated based on the date of final SFR evaluation.

3.3. Outcome data

Outcome data are summarized in Table 2. The primary outcome of postoperative SFR was reported in all included studies. However, as demonstrated in Table 2, there was a lack of consistency in the definition and reporting of both initial and first-session SFR and overall SFRs following retreatment or auxiliary procedures. All studies evaluated SFR using an imaging modality such as kidney, ureter, and bladder radiography, ultrasound, or CT, with the majority using kidney, ureter, and bladder radiography and ultrasound. Stone-free rate was defined by all studies as absence of residual stone fragments on imaging, with some studies specifying a less than 3-mm or less than 4-mm fragment size cutoff.

Table 2.

Intraoperative and postoperative data.

	Outcomes
Author study duration	Intervention: no. patients	No. sessions	Duration of procedure, mean ± SD, or mean (range), min	vDuration of radiation exposure, mean ± SD	Length of hospital stay, mean ± SD, or mean (range), d	SFR definition	SFR, %		Additional procedures to achieve stone clearance	Retreatment rate, %
Abdeldaeim et al.,*[14] 2016–2020	PCNL: 30 (10 HK) URS: 30 (10 HK)	NR	PCNL = 80.33 ± 15.42 URS = 56.43 ± 18.6	PCNL = 4.49 ± 0.80 min URS = 0.84 ± 0.41 min	PCNL = 1.27 ± 0.64 URS = 1.33 ± 0.71	Absence of any residual fragments ≥3 mm at 3 mo in CT for all patients (mixed anomalies). For HK patient subpopulation, postop day 1 SFR also reported	Single-session SFR	PCNL = 80%	PCNL group: 2 patients underwent ESWL	PCNL group: 6.7% Insufficient data available
								URS = 60%	URS group: 2 patients required second URS procedure, 1 patient underwent ESWL	URS group: 10% Insufficient data available
Kartal et al.,[15] 2010–2017	PCNL: 21 URS: 28	NR	PCNL = 86.6 ± 40.8 URS = 65.7 ± 29.9	NR	PCNL = 4.1 ± 2.2 URS = 2.07 ± 1.9	Assessed using KUB radiography on postop day 1 and US on 15th day postop (defined as the single-session SFR). CT performed 3 mo postop for final SFR	Single-session SFR (15th day postop)	PCNL = 81% URS = 71.4%	PCNL group: 3 patients underwent URS, one of which also underwent ESWL	PCNL group: 4.8%
							Final SFR (3 mo postop)	PCNL = 90.5% URS = 85.7%	URS group: 3 patients underwent ESWL and 1 had a ureteral stent placed in a separate session	URS group: 28.5%
Singh et al.,[16] 1990–2014	PCNL: 67 URS: 10	NR	PCNL = 87 (30–180) URS = 42 (30–110)	NR	PCNL = 4.7 (2–12) URS = 1.5 (1–3.5)	<4 mm of residual stone fragments or the absence of any stone fragments on plain abdominal radiograph. KUB radiography and US performed on postop day 1 and 30 Final SFR reported	Final SFR (up to 1 mo postop)	PCNL = 88%	PCNL group: 2 patients required URS	PCNL group: 22% of patients required second procedure, 3.5% required a third procedure, 1% required a fourth repeat procedure
								URS = 80%	URS group: 2 patients required Mini-PCNL	URS group: 10% of patients required a second repeat procedure
Eryildirim et al.,[17] 2007–2016	PCNL: 38 URS: 50	Mean no. sessions per case: PCNL = 1.00 URS = 1.22 Overall = 1.11	PCNL = 75.84 URS = 80.26	PCNL = 69.24 s URS = 20.62 s	PCNL = 3.15 (1–10) URS = 1.58 (1–7)	KUB radiography, US, and non-contrast CT were used for radiologic evaluation of stone clearance. CT used to determine final SFR. SFR after 1 wk and final SFR (3 mo postop) reported	Initial SFR (1 wk postop)	PCNL = 81.6% URS = 80.0%	PCNL group: 3 patients required URS	PCNL group: 7.9%
							Final SFR (3 mo postop)	PCNL = 84.2% URS = 82.0%	URS group: 3 patients underwent repeat URS session	URS group: 6.0%
Ergin et al.,[18] 2005–2015	PCNL: 60 URS: 36	NR	PCNL = 74.5 ± 19.3 URS = 40.5 ± 11.2	NR	PCNL = 2.2 ± 1.4 URS = 1.4 ± 0.7	<3 mm residual stone fragments. Only final SFR reported (not indicated at which timepoint postop)	Final SFR (not indicated at which timepoint postop)	PCNL = 90%	PCNL group: 14 patients required additional procedures (URS stated as an example but definite numbers were not presented in the study)	PCNL group: 23.3%
								URS = 72.2%	URS group: NR	URS group: NR
Blackburne et al.,†[19] 2002–2016	PCNL: 33 URS: 22
		Total for PCNL and URS = 56	NR	NR	NR	<4 mm residual fragments on postop imaging including KUB radiography, CT, or nephrostogram based on surgeon preference SFR reported per renal moiety treated Final SFR reported (not indicated at which timepoint postop)	Single-session SFR	PCNL = 70.3%URS = 100%PCNL = 94.6%URS = 100%	PCNL group: additional procedures included 3 PCNL, 5 URS, and 1 secondary PCNL + URS, and 1 secondary ESWL + PCNLURS group: none	PCNL group: 27%URS group: 0%
									Final SFR (not indicated at which timepoint postop)
Gokce et al.,[20] 2003–2014	URS: 23 ESWL: 44	Median no. sessions: URS = NR ESWL = 3 (1–6)	NR	NR	URS = 1.8 (1–3) ESWL = 0	No residual fragments ≥3 mm assessed using KUB radiography, US, or CT 2–6 wk after URS or 1–6 wk after ESWL Final SFR reported	Final SFR (up to 6 wk postop)	URS = 73.9%	NR	NR
								ESWL = 47.7% 22.7% of ESWL patients achieved stone-free status after single session
Ding et al.,[21] 2005–2014	PCNL: 19 PCNL with antegrade URS: 14 URS: 18 ESWL: 11	Mean ± SD no. sessions: PCNL = 1.4 ± 0.6 PCNL with antegrade URS = 1.3 ± 0.6 URS = 1.4 ± 0.5 ESWL = 1.7 ± 0.5	PCNL = 106.4 ± 16.6 PCNL with antegrade URS = 124.4 ± 15.1 URS = 93.1 ± 11.5 ESWL = 44.8 ± 5.3	NR	PCNL = 6.9 ± 1.4 PCNL with antegrade URS = 6.6 ± 0.8 URS = 1.1 ± 1.4 ESWL = 0	Stone clearance assessed using KUB radiography and US on day 1 and 4 wk postop Reported single-session and overall SFR	Single-session SFR	PCNL = 68.4% PCNL with antegrade URS = 92.9% URS = 88.9% ESWL = 27.3%	PCNL group: 5 patients required one additional session, 1 patient required 2 additional sessions PCNL with antegrade URS group: 4 patients required one additional session, 1 patient required 2 additional sessions URS group: 8 patients required repeat session ESWL group: 5 patients required additional repeat session	PCNL group: 31.6% PCNL with antegrade URS group: 35.7% URS group: 44.4% ESWL group: 45.5%
							Overall SFR (not indicated at which timepoint postop)	PCNL = 89.5% PCNL with antegrade URS = 71.4% URS = 55.6% ESWL = 72.7%
Viola et al.,[22] 1987–2002	PCNL: 19 ESWL: 25	NR	NR	NR	NR	Absence of residual fragments on day 1 postop KUB radiography. Reported day 1 postop SFR and final SFR (not indicated at which timepoint) for PCNL patients. SFR 3 mo postop reported for ESWL patients.	Postop day 1 SFR	PCNL = 75% ESWL = NR	PCNL group: 1 patient required second PCNL and then ESWL, 1 required repeat PCNL, 1 required ESWL, 1 required a percutaneous procedure and then was managed conservatively	PCNL: 26.3%
							Final FR (up to 3 mo postop)	PCNL = 90% ESWL = 31%	ESWL group: 7 patients required PCNL	ESWL group: 28%

ESWL = extracorporeal shockwave lithotripsy; HK = horseshoe kidney; KUB KUB = kidney, ureter, and bladder X-ray; NR = not reported; PCNL = percutaneous nephrolithotomy; postop = postoperative/postoperatively; SFR = stone-free rate; URS = ureteroscopy; US = ultrasonography.

*Abdeldaeim et al. report additional procedure data for all 60 patients in their study (including patients with HK as well as other anomalies) instead of HK-specific data.

†Blackburne et al. report SFR and retreatment rate data per renal moiety treated.

Other outcome data included the average or total number of sessions, procedure duration, duration of radiation exposure, additional procedures to achieve SFR, and retreatment rate. Of note, both the number of additional procedures to achieve overall SFR and retreatment rate were not always consistently reported. In addition, some studies estimated the retreatment rate based on the number of repeat procedures in an intervention arm as well as other additional procedures, whereas other studies separated these 2 procedure types. Abdeldaeim et al.^[14] reported auxiliary procedure data on all 60 patients in their study, which included patients with renal anomalies other than HK.

3.4. PCNL versus URS

Seven retrospective cohort studies compared PCNL with URS.^[14–19,21] Of these, only 4 studies specifically reported both initial single-session SFR, as well as overall SFR data.^{[15,17,19,21]} On the other hand, Abdeldaeim et al.^[14] reported single-session SFR only. Single-session SFR ranged between 68.4% and 81.6% for PCNL and between 60% and 100% for URS. There was no statistically significant difference between PCNL and URS for single-session SFR, with an RR of 1.00 (95% CI, 0.79–1.26; I² = 62.44%; Fig. 2A).

Figure 2.

(A) Forest plot of single-session SFR for PCNL versus URS. (B) Forest plot of overall SFR for PCNL versus URS. PCNL = percutaneous nephrolithotomy; SFR = stone-free rate; URS = ureteroscopy. CI = confidence interval; REML = random-effects models.

The overall SFRs ranged between 80.0% and 94.6% for PCNL and between 60% and 100% for URS. There was no statistically significant difference between PCNL and URS in terms of overall SFR, with an RR of 1.04 (95% CI, 0.95–1.13; I² = 20.63%; Fig. 2B). Mean procedure duration, radiation duration, and length of hospital stay were generally longer in patients undergoing PCNL compared with URS. Data on retreatment rate and auxiliary procedures were mixed between the 2 interventions (Table 2).

3.5. URS versus ESWL

Two retrospective cohort studies assessed URS versus ESWL.^[20,21] Ding et al. ^[21] reported both single-session SFR and overall SFR, whereas Gokce et al.^[20] report single-session SFR for ESWL patients only. Single-session SFR for patients undergoing ESWL was lower (up to 27.3%) than URS patients (up to 88.9%). Figure 3A and 3B show forest plots of single-session and overall SFRs for both interventions. There was a statically significant difference between URS and ESWL for overall SFR, with an RR of 1.38 favoring URS (95% CI, 1.04–1.82; I² = 0%; Fig. 3B). Ureteroscopy had a longer mean procedure duration and hospital stay as compared with ESWL.

Figure 3.

(A) Forest plot of single-session SFR for URS versus ESWL. (B) Forest plot of overall SFR for URS versus ESWL. ESWL = extracorporeal shockwave lithotripsy; SFR = stone-free rate; URS = ureteroscopy. CI = confidence interval; REML = random-effects models.

3.6. PCNL versus ESWL

Two retrospective cohort studies compared PCNL versus ESWL.^[21,22] Both single-session and overall SFRs were reported by Ding et al.^[21]; however, Viola et al.^[22] only reported single-session SFR for PCNL patients, with SFR for the ESWL cohort assessed at 3 months postoperatively. There was no statistically significant difference in overall SFR with an RR of 1.80 (95% CI, 0.80–4.05; I² = 80.86%; Fig. 4). Percutaneous nephrolithotomy was associated with longer mean procedure duration and hospital stay. Retreatment rates were similar between both groups (Table 2).

Figure 4.

Forest plot of overall SFR for PCNL versus ESWL. ESWL = extracorporeal shockwave lithotripsy; PCNL = percutaneous nephrolithotomy; SFR = stone-free rate. CI = confidence interval; REML = random-effects models.

3.7. Complications

Postoperative complications are summarized in Table 3 according to CD classification, including incidence rates of specific complications. Postoperative drop in mean Hb is summarized in Table 4. Across all studies, there were no cases of CD grade V complications, but there were 3 CD grade IV complications due to sepsis in patients undergoing PCNL. Kartal et al.^[15] reported 1 patient requiring postoperative intensive care unit admission and treatment with antibiotics. Clavien-Dindo grade I complications after PCNL included pain (3.03%–13.3%), fever not requiring antibiotics (1.8%–15.8%), hematuria/hemorrhage not requiring transfusion (1.49%–50%), pelvicalyceal system perforation in 1 patient, and prolonged nephrostomy tube placement in 2 patients. The incidence of hemorrhage/hematuria requiring blood transfusion (CD grade II) in PCNL patients ranged between 0% and 19.04%. In terms of CD grade III complications, there was 1 case of urinoma, 3 cases of postoperative clot retention (2.99%–4.76%), and 1 case of pyelovascular communication between the renal pelvis and vein. Postoperative drop in mean Hb among PCNL patients ranged between 0.47 and 1.83 g/dL.

Table 3.

Complications according to Clavien-Dindo classification.

			Clavien-Dindo I, n (%)							Clavien-Dindo II, n (%)			Clavien-Dindo III, n (%)						Clavien-Dindo IV, n (%)
													3a				3b		4b
Author study duration	Intervention: no. patients		Pain	Fever that did not require post-op antibiotics	Postop hematuria/bleeding not requiring blood transfusion	Ileus	Renal colic managed conservatively	Pelvicalyceal system perforation	Prolonged nephrostomy tube placement/poor drainage	Fever requiring antibiotics	Hemorrhage/hematuria Requiring blood transfusion	UTI requiring antibiotics	Ureteral stent placement	Clot retention	Urinoma	Perirenal hematoma	Endoscopic treatment for ureteral stone	Pyelovascular communication between renal pelvis and renal vein	Sepsis
Abdeldaeim et al.,*[12] 2016–2020	PCNL: 30 (10 HK) URS: 30 (10 HK)	PCNL	4 (13.3%)	2 (6.7%)	15 (50%)						1 (3.3%)
		URS	3 (10%)	6 (20%)	15 (50%)						0 (0%)
Kartal et al.,[13] 2010–2017	PCNL: 21 URS: 28	PCNL	NR	1 (4.76%)	NR						4 (19.04%)			1 (4.76%)	1 (4.76%)				1 (4.76%) urosepsis requiring antibiotics and ICU admission
		URS	NR	4 (14.3%)	NR						0 (0%)	2 (7.14%)	1 (3.57%) postop due to renal colic
Singh et al.,[14] 1990–2014	PCNL: 67 URS: 10	PCNL	NR	8 (11.94%)	1 (1.49%)			1 (1.49%)			2 (2.99%)			2 (2.99%)
		URS	NR	1 (10%)		1 (10%)					0 (0%)
Eryildirim et al.,[15] 2007–2016	PCNL: 38 URS: 50	PCNL	NR	6 (15.8%)	5 (1.8%)						3 (7.9%)		2 (5.2%) stent placement for ureteral stone				3 (7.9%)
		URS	NR	8 (16.0%)	3 (3.6%)						1 (2.0%)		1 (2.0%) stent placement for ureteral stone				3 (6.0%)
Ergin et al.,[16] 2005–2015	PCNL: 60 URS: 36	PCNL	NR		Observed but exact number NR.						NR
		URS			3 (8.3%)					2 (5.5%)	NR
Blackburne et al.,[17] 2002–2016	PCNL: 33 URS: 22	PCNL	1 (3.03%)							1 (3.03%)			1 (3.03%) due to failed nephrostomy clamping trial
		URS	0 (0%)										0 (0%)
Gokce et al.,[18] 2003–2014	URS: 23 ESWL: 44	URS		1 (4.35%)	3 (13.0%)		3 (13.0%)									0 (0%)
		ESWL		1 (2.27%)	6 (13.6%)		11 (25%)						13 (29.55%) of which 5 (11.4%) were due to renal colic			1 (2.27%)
Ding et al.,[19] 2005–2014	PCNL: 19 PCNL with antegrade URS: 14 URS: 18 ESWL: 11	PCNL		2 (10.5%)			0 (0%)				0 (0%)								1 (5.3%) urosepsis
		PCNL with antegrade URS		2 (14.3%)			0 (0%)				0 (0%)								0 (0%)
		URS		3 (16.7%)			0 (0%)				0 (0%)								0 (0%)
		ESWL		0 (0%)			1 (9.1%)				NR								0 (0%)
Viola et al.,[20] 1987–2002	PCNL: 19 ESWL: 25	PCNL			5 (27%)				2 (10%)									1 (5%)	1 (5%) septicemia managed with antibiotics
		ESWL

ESWL = extracorporeal shockwave lithotripsy; NR = not reported; PCNL = percutaneous nephrolithotomy; postop = postoperative/postoperatively; URS = ureteroscopy; UTI = urinary tract infection; VAS = visual analog scale. HK = horseshoe kidney; ICU = intensive care unit.

*Abdeldaeim et al. report complications for all 60 patients included in their study, which includes patients with HK as well as other anomalies. They recorded postoperative pain data using VAS, from which we report above the number of patients with at least moderate pain on VAS scale.

Table 4.

Postoperative hemoglobin drop.

Author, study duration	Intervention	Hb drop, mean ± SD or mean (range), g/dL
*Abdeldaeim et al.,[14] 2016–2020	PCNL	0.47 ± 0.34.
	URS	0.20 ± 0.14
Kartal et al.,[15] 2010–2017	PCNL	NR
	URS	NR
Singh et al.,[16] 1990–2014	PCNL	1.27 (0.2–5.0)
	URS	0.20 (0.0–0.4)
Eryildirim et al.,[17] 2007–2016	PCNL	1.55
	URS	0.13
Ergin et al., [18]2005–2015	PCNL	NR
	URS	NR
Blackburne et al.,[19] 2002–2016	PCNL	NR
	URS	NR
Gokce et al.,[20] 2003–2014	URS	NR
	ESWL	NR
Ding et al.,[21] 2005–2014	PCNL	1.83 ± 0.51
	PCNL with antegrade URS	1.43 ± 0.42
	URS	NR
	ESWL	NR
Viola et al.,[22] 1987–2002	PCNL	NR
	ESWL	NR

ESWL = extracorporeal shockwave lithotripsy; Hb = hemoglobin; NR = not reported; PCNL = percutaneous nephrolithotomy; URS = ureteroscopy.

There were no cases of CD grade IV or V complications in patients undergoing URS. Most CD grade III complications that occurred in URS patients involved postoperative endoscopic treatment or placement of a JJ stent due to obstruction, renal colic, or ureteral stones. Regarding CD grade II complications in URS patients, there were 2 cases of urinary tract infection requiring antibiotics and 2 cases of fever requiring antibiotics. Hemorrhage/hematuria requiring blood transfusion ranged between 0% and 2%. Common CD grade I complications after URS included pain, renal colic managed conservatively, and hemorrhage/hematuria not requiring blood transfusion. Postoperative drop in mean Hb for URS patients ranged between 0.13 and 0.20 g/dL.

Complication rates in patients who underwent ESWL were overall low. Of note, Gokce et al.^[20] reported that 29.5% of patients required postoperative ureteral stent placement and 1 case of perirenal hematoma, which was not observed in any other study. There were no cases of postoperative drop in Hb in ESWL patients.

4. Discussion

Horseshoe kidney is the most common congenital renal anomaly, but it remains a rare condition, affecting only 0.25% of the general population.^[2] Although up to a third of patients with HK remain asymptomatic, they are predisposed to a variety of different complications including kidney stone formation.^[23] The published literature on the management of urolithiasis in HK is varied and lacks clear consensus regarding the best treatment option for this unique cohort of patients. To our knowledge, this review provides the first detailed comparative summary and analysis of the evidence for PCNL, URS, and ESWL for management of HK stones. Our search criteria including all commonly utilized treatment modalities resulted in the inclusion of a larger pool of studies as compared with previously published reviews. One of the main findings of the present study is that PCNL and URS have equivalent single-session and overall SFRs; however, URS is less invasive and demonstrates a better safety profile. Furthermore, when comparing PCNL with ESWL, there was no statistically significant difference in overall SFR between the 2 treatment groups. Finally, similar to previously published analyses, our study identified a statistically significant difference between URS and ESWL for single-session and overall SFRs that favors URS.

Yi et al.^[2] and Chen et al.^[6] have previously published 2 SRs assessing the efficacy and safety of URS versus ESWL for management of HK stones. Similar to our study, both SRs included single-center retrospective cohort studies (4 studies included by Yi et al.^[2] and 5 studies included by Chen et al.^[6]). Neither SR had a minimum cutoff for the number of patients per intervention arm to be included in the final analysis. However, our cutoff of at least 10 patients per intervention arm reduced the number of studies comparing URS with ESWL that were ultimately included to 2 studies.^[20,21] Both previously published SRs found that URS was associated with greater initial and overall SFRs compared with ESWL.^[2,6] Only one of these studies reported a statistically significant difference in the retreatment rate between the 2 treatment modalities, which favored URS. As shown in Figure 3A and 3B, the results from our study are comparable to these findings, with an RR between the 2 interventions of 1.38 for overall SFR favoring URS (95% CI, 1.04–1.82; I² = 0%). In our analysis, patients who underwent ESWL generally had low single-session SFRs of up to 30% and required repeat procedures to achieve stone clearance. Ding et al.^[21] report that 5 of 11 patients required a second ESWL procedure, for a retreatment rate of approximately 45%. However, ESWL patients generally had shorter procedure duration and hospital stay compared with URS patients. In terms of complications, there were no major CD grade IV or higher complications associated with either modality. Most URS-related complications were CD grade I, whereas for ESWL, 11% of patients required postoperative stent placement for renal colic (CD grade 3a).

Percutaneous nephrolithotomy is the first choice of treatment for kidney stones 20 mm or larger.^[7,8] Despite higher complication rates associated with PCNL, some studies have reported no further increase in HK patients.^[24,25] Our SR included 8 studies that assessed PCNL as a treatment arm for HK urolithiasis for a total of 267 patients. The mean stone size ranged from 17.90 ± 2.43 mm to 27.9 ± 8.6 mm, which was similar to that of URS-treated patients (8.4 [2–25] mm to 22.3 ± 9.1 mm). Percutaneous nephrolithotomy had acceptable single-session and overall SFRs ranging between 68.4% and 81.6%, and 80.0% and 94.6%, respectively. When comparing PCNL with URS, no statistically significant difference was found between the 2 treatment modalities for single-session and overall SFRs (Fig. 2A). This was also the case when PCNL was compared with ESWL (Fig. 4). Furthermore, PCNL was associated with longer procedure duration, radiation exposure, and hospital stay compared with URS and ESWL. In terms of complications, PCNL was associated with significant major complications, including 5 cases of CD grade III complications (postoperative clot retention, urinoma, and pyelovascular communication) and 3 cases of CD grade IV (postoperative sepsis). Notably, postoperative bleeding requiring blood transfusion (CD grade II) ranged between 0% and 19.04% with a postoperative drop in Hb ranging from 0.47 to 1.83 g/dL.

Advancements in technology have broadened the use of flexible URS, making it a viable option for management of stones in challenging anomalous renal anatomy.^[26] As a result of significant morbidity associated with PCNL, URS has been increasingly used to treat larger stones because of its more favorable safety profile.^[15] In our study, which included a total of URS 197 patients, URS demonstrated satisfactory single-session and overall SFRs ranging between 60% and 100%. When evaluating the retreatment rate, URS had a slightly higher retreatment rate (0%–44.6%) than that of PCNL (4.8%–31.6%). In addition, the majority of complications associated with URS were CD grade I–II. Therefore, our results demonstrate that both PCNL and URS are effective treatments that can be utilized for larger HK stones; however, there is no statistically significant difference in efficacy between them. Furthermore, our data show that URS is a safer treatment modality.

Although the present review is the first to compare the benefits and risks of all available contemporary treatment options for HK, there are several limitations to be considered. All articles included in this study are retrospective, single-center studies with relatively small sample sizes. Furthermore, there was lack of standardized definitions and reporting of important outcomes such as SFR and intraoperative and postoperative complications across all studies. Because of the rarity of HK and considering that some centers may manage more HK patients than others, it is possible that the data are skewed by differing levels of technical skills possessed by each physician when treating HK stones. Finally, evaluation of HK anatomy, stone composition, and sources of energy used was not completed because of limited reporting of these parameters in the individual studies. The ROB assessment of the studies yielded mixed results; however, the evidence stems from retrospective observational studies, which have their own inherent biases; thus, caution is advised when interpreting the conclusions of any individual study. Despite this, the rarity of HK is evident in the number and quality of studies that are available. This is the most robust SR to date that encompasses international publications over a significant period and is therefore representative of the current globally available treatment options for HK stones. Additional experimental and larger-scale studies are needed to confirm the findings of the present study.

5. Conclusions

This SR demonstrates several notable findings when comparing PCNL, URS, and ESWL for the treatment of urolithiasis in HK patients. There was no difference in SFR between PCNL and URS, and URS was associated with fewer and less severe complications. However, the stone burden in the PCNL group was greater than that in the URS group. Ureteroscopy was found to be more effective than ESWL with greater stone clearance and comparable stone burden and safety profile. Further large-scale randomized controlled trials are needed to confirm these findings.

Acknowledgments

None.

Statement of ethics

Not applicable.

Conflict of interest statement

No conflict of interest has been declared by the authors.

Funding source

No funding was received related to the conduct of this study.

Author contributions

YQ: Conceptualization, investigation, data curation, writing—original draft, writing—review and editing, visualization;

MAT: Investigation, data curation, writing—original draft, writing—review and editing;

FB: Validation, formal analysis, resources;

AS: Formal analysis, resources;

ND: Conceptualization, writing—review and editing, supervision.