Extraction-string ureteral stents and postoperative infection after flexible ureteroscopy: an IPTW-adjusted analysis

Abstract

Extraction-string ureteral stents facilitate non-cystoscopic removal after ureteroscopy, but concerns remain regarding postoperative infection and accidental dislodgement. We evaluated whether extraction-string stents were associated with postoperative infection after flexible ureteroscopic lithotripsy (fURSL) in a cohort restricted to procedures with the distal Double-J (D-J) stent portion positioned within the bladder. This retrospective observational study included 102 procedures in 92 unique patients: 58 extraction-string and 44 conventional non-string procedures. Postoperative infection was defined as fever > 38 °C or a positive postoperative urine culture during the stent-indwelling period. Propensity scores incorporated age, sex, positive preoperative urine culture, preoperative pyuria, and procedure laterality. Stabilized inverse probability of treatment weights were truncated at the 1st and 99th percentiles; effective sample size was 85.38. Crude postoperative infection occurred in 8/58 (13.8%) extraction-string and 6/44 (13.6%) non-string procedures. After weighting, estimated infection rates were 13.6% and 8.8%, respectively (OR 1.64, 95% CI 0.50–5.44; P = 0.416). Fever, positive postoperative urine culture, postoperative pyuria, and postoperative day 1 pain did not differ significantly. Sensitivity analyses accounting for repeated procedures and propensity-model specification were consistent with the primary analysis. Extraction-string stent use was not significantly associated with postoperative infection after fURSL; however, the limited number of events and wide confidence intervals preclude exclusion of a clinically meaningful difference.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00240-026-02071-5.

Introduction

Ureteral stents are frequently placed after ureteroscopy to maintain drainage and prevent transient obstruction. Extraction-string stents permit non-cystoscopic removal and may reduce removal-related pain and healthcare burden, but accidental dislodgement and ascending infection remain concerns [1–4]. Non-cystoscopic removal may also avoid an additional anesthetic procedure in pediatric patients [5].

Evidence regarding infectious risk remains heterogeneous. Comparative studies and meta-analyses have generally not shown a significant increase in urinary tract infection with extraction strings [2–4]. Recent reviews similarly suggest a small overall difference while emphasizing heterogeneity, low evidence quality, and imprecision [6, 7]. Infectious outcomes after ureteroscopy are also influenced by positive preoperative urine culture, diabetes mellitus, preoperative stenting, and prolonged operative time [8–10].

We therefore evaluated the association between extraction-string stent use and postoperative infection after fURSL using inverse probability of treatment weighting (IPTW), with separate analyses of fever, culture positivity, postoperative pyuria, and postoperative day 1 pain and sensitivity analyses addressing repeated procedures and propensity-model specification.

Methods

Study design and population

This retrospective observational study included fURSL procedures performed at Keelung Chang Gung Memorial Hospital from January through November 2025. The final cohort comprised 102 procedures in 92 unique adult patients (Fig. 1). All included procedures had the distal D-J stent portion/coil positioned within the urinary bladder. Procedures with the distal D-J portion retained in the ureter, premature extraction-string stent dislodgement, incomplete follow-up, or missing key data were excluded. Thirteen procedures with documented premature extraction-string stent dislodgement were excluded during cohort construction. Because the retrospective source data did not permit reliable reconstruction of the total number of extraction-string stents initially placed before application of all eligibility criteria, a valid dislodgement incidence could not be calculated.

Fig. 1.

Study flow diagram. D-J, double-J; fURSL, flexible ureteroscopic lithotripsy; IPTW, inverse probability of treatment weighting

Stent technique, fixation, and postoperative education

After fURSL, the D-J stent was placed with its distal portion/coil within the bladder. In the extraction-string group, a 3 − 0 nylon string was attached to the distal stent portion and secured externally with 3 M OPSITE transparent film dressing (Fig. S1). The string was fixed to the dorsal penis in male patients and the right medial thigh in female patients. The dressing was applied in two layers around the string, leaving sufficient redundant length to avoid excessive tension during routine activity or penile erection. All extraction-string patients received the same fixation protocol and were instructed to maintain hydration and avoid traction on the string. Premature dislodgement is a recognized complication [1, 11].

Antibiotic protocol

Patients with negative preoperative urinalysis (urine WBC ≤ 30/HPF) received prophylactic cefazolin within 30 min before surgery. Patients with positive preoperative urinalysis received cefazolin 1 g beginning one day before surgery unless culture results were available, in which case therapy was adjusted to susceptibility. Postoperatively, patients without positive preoperative urinalysis received oral cefadroxil 500 mg every 12 h for 3 days; those with positive preoperative urinalysis received 7 days of antibiotic therapy, adjusted to culture and susceptibility results when available.

Definitions and outcomes

Preoperative urine culture was positive only when an explicit organism was identified; “mixed,” “<1000,” and “nil” were classified as negative. Preoperative and postoperative pyuria were defined as urine WBC > 30/HPF. The primary postoperative infection outcome was fever > 38 °C or a positive postoperative urine culture during the stent-indwelling period; culture positivity counted as an event even without fever. Fever and culture positivity were also analyzed separately. Postoperative pain was assessed on postoperative day 1 using the same 0–10 Numeric Rating Scale (NRS) by a physician blinded to stent type.

Statistical analysis

Propensity scores for extraction-string stent use were estimated by logistic regression including age, sex, positive preoperative urine culture, preoperative pyuria, and procedure laterality (left, right, or bilateral). Stabilized IPTW weights were truncated at the 1st and 99th percentiles. Balance was assessed using absolute standardized mean differences (SMDs), with < 0.10 considered acceptable. Overall effective sample size after weighting was 85.38 and truncated weights ranged from 0.541 to 2.685. Weighted logistic regression with robust sandwich variance estimates was used for binary outcomes. Sensitivity analyses used patient-clustered robust standard errors, the original four-variable propensity specification, and the earliest eligible procedure per patient. No a priori sample-size calculation was performed because all eligible procedures in the study period were included. Given only 14 postoperative infection events, multivariable predictor modeling was considered exploratory and is reported in Online Resource 1. Two-sided P < 0.05 was considered statistically significant.

Results

The cohort included 58 extraction-string and 44 non-string procedures. Before weighting, the extraction-string group was younger (54.6 ± 10.4 vs. 61.5 ± 11.3 years; P = 0.002) and had less preoperative pyuria (39.7% vs. 63.6%; P = 0.027; Table 1). After IPTW, all prespecified covariates achieved absolute SMDs < 0.10 (Table S1 and Fig. S2).

Table 1.

Baseline characteristics of the study population

Variable	Extraction-string (n = 58)	No-string (n = 44)	P value
Age, years	54.6 ± 10.4	61.5 ± 11.3	0.002
Male sex, n (%)	34 (58.6)	28 (63.6)	0.684
Positive preoperative urine culture, n (%)	10 (17.2)	10 (22.7)	0.616
Preoperative pyuria (> 30 WBC/HPF), n (%)	23 (39.7)	28 (63.6)	0.027
Procedure laterality, n (%)			0.063
Left-sided procedure	18 (31.0)	23 (52.3)
Right-sided procedure	14 (24.1)	10 (22.7)
Bilateral procedure	26 (44.8)	11 (25.0)

Crude postoperative infection occurred in 8/58 (13.8%) extraction-string and 6/44 (13.6%) non-string procedures (Table 2). After IPTW, estimated infection rates were 13.6% and 8.8%, respectively (OR 1.64, 95% CI 0.50–5.44; P = 0.416). Weighted fever rates were 5.1% versus 4.5% (OR 1.13, 95% CI 0.23–5.55; P = 0.878), positive postoperative culture rates 13.6% versus 7.5% (OR 1.94, 95% CI 0.56–6.77; P = 0.298), and postoperative pyuria rates 15.0% versus 20.5% (OR 0.68, 95% CI 0.24–1.91; P = 0.466).

Table 2.

IPTW-adjusted postoperative outcomes

Outcome	Raw: String	Raw: No-string	IPTW String	IPTW No-string	Effect (95% CI)	P value
Postoperative infection	8/58 (13.8%)	6/44 (13.6%)	13.6%	8.8%	OR 1.64 (0.50–5.44)	0.416
Fever > 38 °C	4/58 (6.9%)	3/44 (6.8%)	5.1%	4.5%	OR 1.13 (0.23–5.55)	0.878
Positive postoperative U/C	8/58 (13.8%)	5/44 (11.4%)	13.6%	7.5%	OR 1.94 (0.56–6.77)	0.298
Postoperative pyuria	9/58 (15.5%)	13/44 (29.5%)	15.0%	20.5%	OR 0.68 (0.24–1.91)	0.466
POD1 pain (0–10 NRS)	1.71 ± 1.04	1.86 ± 1.30	1.60	2.17	MD -0.56 (-1.24 to 0.12)	0.106

OR, odds ratio; MD, mean difference; NRS, Numeric Rating Scale; U/C, urine culture

Postoperative day 1 pain was also similar. Crude NRS scores were 1.71 ± 1.04 versus 1.86 ± 1.30 (P = 0.515); weighted means were 1.60 versus 2.17, yielding a mean difference of -0.56 (95% CI -1.24 to 0.12; P = 0.106). The primary association remained non-significant with patient-clustered robust standard errors (OR 1.64, 95% CI 0.42–6.40; P = 0.474), the original propensity model (OR 1.64, 95% CI 0.50–5.37; P = 0.414), and one procedure per patient (OR 2.55, 95% CI 0.67–9.68; P = 0.169) (Fig. S3).

Discussion

In this IPTW-adjusted cohort restricted to procedures in which the distal D-J stent portion was positioned within the bladder, extraction-string stent use was not significantly associated with postoperative infection after fURSL. The primary IPTW analysis yielded an OR of 1.64 (95% CI 0.50–5.44), with weighted infection rates of 13.6% in the extraction-string group and 8.8% in the non-string group. Although the point estimate was above unity, the confidence interval was wide and compatible with both lower and substantially higher infection risk. Therefore, these findings should not be interpreted as demonstrating equivalence or non-inferiority, and a clinically meaningful difference cannot be excluded.

Our findings should be interpreted in the context of a heterogeneous literature evaluating extraction-string or tethered ureteral stents. Fröhlich et al. reported no significant increase in postoperative UTI among patients with extraction strings [2]. Dellis et al., in a systematic review and meta-analysis of comparative randomized studies, likewise found no difference in UTI rates between string and non-string stents, although stent dislodgement occurred more frequently with extraction strings [3]. A systematic review of non-cystoscopic stent removal in the pediatric setting also supported the feasibility of extraction-string stents while highlighting the distinct considerations associated with their use in children [5]. More recently, Osiński et al. reviewed 11 studies published between 2015 and 2023 and found that most did not demonstrate a statistically significant difference in UTI risk; however, one large multicenter retrospective study reported a small absolute increase in UTI with extraction strings, and the overall evidence was limited by heterogeneous infection definitions, antibiotic protocols, and generally high risk of bias [6]. Dong et al. subsequently synthesized five randomized controlled trials involving 598 patients and found no statistically significant difference in UTI between extraction-string and conventional stent removal (RR 1.45, 95% CI 0.48–4.42), but emphasized substantial heterogeneity and limited certainty of evidence [7]. Taken together, the available literature does not establish a clear increase in infection risk attributable to extraction strings, but the direction and imprecision of several estimates—including ours—mean that a clinically meaningful increase cannot be excluded. Differences in patient selection, stent dwell time, antibiotic protocols, and definitions of infection further limit direct comparison across studies.

Postoperative day 1 pain was also not significantly different between the two groups. After IPTW, mean NRS scores were 1.60 in the extraction-string group and 2.17 in the non-string group, corresponding to a mean difference of − 0.56 (95% CI − 1.24 to 0.12; P = 0.106). Extraction-string stents have been proposed to reduce discomfort associated with subsequent cystoscopic stent removal [1, 3, 4]; however, our pain assessment addressed early postoperative discomfort rather than pain during stent removal. Moreover, pain was assessed at a single postoperative time point and therefore does not capture changes in stent-related symptoms throughout the indwelling period. Accordingly, our findings should be interpreted only as showing no detectable difference in early postoperative pain within this cohort.

An important methodological feature of the present study was the standardization of distal stent position and external string fixation. The analysis was restricted to procedures in which the distal D-J stent portion/coil was positioned within the bladder, thereby reducing heterogeneity related to stent position. In addition, all extraction strings were secured using the same two-layer 3 M OPSITE fixation protocol, with sex-specific fixation sites and sufficient redundant string length to minimize traction during routine activity. Nevertheless, 13 procedures with documented premature extraction-string stent dislodgement were excluded during cohort construction. Premature dislodgement is a recognized complication of extraction-string stents [1, 11], and exclusion of these cases may introduce selection bias and limit generalizability. Furthermore, because the retrospective source data did not permit reliable reconstruction of the total number of extraction-string stents initially placed before all eligibility criteria were applied, a valid dislodgement incidence could not be calculated. The present findings therefore primarily apply to patients in whom the intended stent-indwelling period was maintained and should not be interpreted as an estimate of the overall dislodgement risk associated with extraction-string stents.

The sensitivity analysis restricted to one eligible procedure per patient deserves particular consideration. In this analysis, the estimated association between extraction-string use and postoperative infection was numerically stronger than in the primary IPTW analysis (OR 2.55, 95% CI 0.67–9.68 vs. OR 1.64, 95% CI 0.50–5.44). This increase in the point estimate indicates that the magnitude of the observed association was sensitive to the handling of repeated procedures. However, both estimates were imprecise, their confidence intervals were wide and substantially overlapping, and neither reached statistical significance. The wider confidence interval in the one-procedure-per-patient analysis is also consistent with the loss of observations and infectious events after restricting the dataset to a single procedure per patient. Accordingly, the higher point estimate should not be interpreted as evidence that extraction-string stents increase postoperative infection risk, but it reinforces that the present dataset cannot exclude a clinically meaningful increase in risk. The similarity of the primary estimate when patient clustering was instead addressed using patient-clustered robust standard errors (OR 1.64, 95% CI 0.42–6.40) provides some reassurance that repeated observations alone did not determine the primary result, although larger patient-level datasets are required to estimate the magnitude of any association more precisely.

Only 14 composite postoperative infection events occurred in the analytic cohort, limiting the ability to identify independent predictors reliably. The exploratory multivariable analysis should therefore be interpreted cautiously (Table S2). In particular, the absence of a statistically significant association between positive preoperative urine culture and postoperative infection in this dataset should not be interpreted as evidence against its clinical importance. Larger meta-analyses have consistently identified positive preoperative urine culture as an important risk factor for infectious complications or urosepsis after ureteroscopy [8–10]. The discrepancy likely reflects the limited number of events and consequent statistical imprecision in the present study rather than evidence of no association.

Several limitations merit emphasis. First, this was a retrospective, single-center study with non-random treatment assignment, no a priori sample-size calculation, and only 14 composite infection events, resulting in limited statistical precision. Second, although IPTW achieved acceptable balance in the measured covariates, residual confounding remains possible. Important potential confounders, including stone burden, operative time, diabetes mellitus, preoperative stenting, history of recurrent UTI, and ureteral access sheath characteristics, were not consistently available and therefore could not be incorporated into the propensity model. These factors have been associated with infectious complications following ureteroscopy [8–10] and could also influence the choice of stent strategy; imbalance could therefore bias the observed association in either direction. Third, exclusion of premature extraction-string stent dislodgement may introduce selection bias and limits extrapolation to the overall safety profile of extraction-string stents. Fourth, the sensitivity analysis restricted to one procedure per patient produced a higher point estimate than the primary analysis, underscoring the uncertainty surrounding the magnitude of the association. Finally, the composite infection endpoint combined fever > 38 °C and positive postoperative urine culture, which may represent different levels of clinical severity. Although these components were analyzed separately and neither showed a statistically significant between-group difference, heterogeneity within the composite endpoint remains a limitation.

Taken together, the present findings do not demonstrate a statistically significant association between extraction-string stent use and postoperative infection among patients who maintained the intended stent-indwelling period. However, the direction and imprecision of the primary estimate, together with the higher point estimate observed in the one-procedure-per-patient sensitivity analysis, indicate that a clinically meaningful increase in infection risk cannot be excluded. Larger prospective studies with standardized definitions of infection, systematic reporting of premature dislodgement, and patient-level analyses incorporating established infectious risk factors are needed to provide more precise estimates of the safety of extraction-string ureteral stents.

Conclusion

In this retrospective IPTW-adjusted analysis restricted to procedures with the distal D-J stent portion/coil positioned within the bladder, extraction-string ureteral stent use was not significantly associated with postoperative infection after fURSL. Sensitivity analyses did not identify a statistically significant association, although the one-procedure-per-patient analysis yielded a higher point estimate than the primary analysis. The limited number of events and wide confidence intervals preclude exclusion of a clinically meaningful difference. Larger prospective studies with standardized stent position, fixation, dislodgement reporting, and infection definitions are warranted.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

1. Conception and design: All authors. 2. Data acquisition: Jui-Ying Lin. 3. Data analysis and interpretation: Jui-Ying Lin and Cheng-Chia Lin. 4. Drafting the manuscript: Jui-Ying Lin and Cheng-Chia Lin. 5. Critical revision of the manuscript for scientific and factual content: Jui-Ying Lin, Cheng-Chia Lin, Chun-Te Wu and Kuan-Lin Liu. 6. Statistical analysis: Jui-Ying Lin and Cheng-Chia Lin. 7. Supervision: All authors.

Funding

No funding was received for conducting this study.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request, subject to institutional and ethical restrictions.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

This retrospective study was approved by the Chang Gung Medical Foundation Institutional Review Board and was conducted in accordance with the ethical standards of the institutional research committee and the Declaration of Helsinki and its later amendments.

Consent to Participate

The requirement for informed consent was waived by the Institutional Review Board because of the retrospective nature of the study. All patient data were anonymized before analysis.