Predictors of Nephrostomy Catheter Dislodgement: Insights from a Retrospective Analysis

A retrospective review of 742 PCN procedures performed between June 2020 and June 2024 was conducted. Thirty-eight patients with documented evidence of spontaneous NCD formed the dislodgement group, and 38 patients without NCD served as the control group.

Control patients were selected using a 1:1 nearest-neighbor matching approach based on propensity scores. Age and sex were chosen as matching variables due to their clinical relevance and availability in the dataset. Age correlates with anatomical and physiological changes, such as reduced tissue integrity, which may influence catheter stability. Similarly, sex differences, such as muscle mass and fat distribution, could impact catheter dislodgment risks. These variables were normalized using Min–Max normalization to ensure equal weighting in propensity score calculations. Propensity scores were calculated via logistic regression, and each dislodgement group patient was matched to a control patient with the closest score. If no exact match was found, up to four nearest neighbors were considered.

Exclusion Criteria: Patients with prophylactically placed nephrostomy catheters, early post-procedure removal, infection, abscess, or hematoma in the nephrostomy tract, bedridden status, accidental trauma, or inappropriate mental status were excluded. Patients without a non-contrast CT scan within 3 months post-procedure or with a history of surgical intervention or radiotherapy involving the kidney were also excluded. Cases with imaging artifacts or renal axis anomalies were excluded.

Six different interventional radiologists, each with a minimum of 3 years of experience, carried out the PCN catheter placement. The procedure was carried out under ultrasound guidance, using an 18G needle to access the lower pole calyces. An 8F pigtail nephrostomy catheter (Argon Medical Devices, Plano, TX, USA) was placed after imaging with fluoroscopy. The catheter was secured to the skin using a non-absorbable 2/0 Silk surgical suture and tied with multiple knots in a braided pattern by the interventional radiologist who performed the procedure. Non-contrast abdominal CT scans were carried out using a 16-slice CT scanner (Somatom, Siemens, Germany). The evaluation of all non-contrast abdominal CT scans was conducted on a workstation using Fonet PACS software (Fonet Information Technologies, Ankara, Türkiye) with consensus between a radiology resident with 3 years of training and a radiologist with 10 years of experience in abdominal radiology, both of whom were blinded to the clinical data of the patients.

The total bilateral paravertebral muscle area at the level of nephrostomy catheter passage through the kidney was manually measured by drawing a region of interest (ROI). Additionally, the thickness of the paravertebral muscles at the same level was measured (Fig. 1). The total bilateral psoas muscle area (PMA) at the level of nephrostomy catheter penetration into the kidney was also manually measured by drawing a region of interest (Fig. 2) [17, 18]. The thickness of the subcutaneous fat tissue was measured at its thickest anterior point along the path where the catheter traversed from the skin to the kidney (Fig. 3).

The size of the kidney on the coronal plane where the nephrostomy was placed and the parenchymal thickness at the lower pole where the nephrostomy catheter was inserted were determined for all patients (Fig. 4). The distance from the renal cortex to the skin along the path of the nephrostomy catheter was also measured (Fig. 5). Additionally, the thickness of the posterior lateral abdominal wall muscle in the nephrostomy tract was measured (Fig. 6).

Factors affecting the overall body composition such as the total paravertebral muscle area, BMI (body mass index) subcutaneous fat tissue thickness, and total PMA were analyzed. Additionally, the factors affecting the tissues traversed by the catheter, such as kidney size, renal parenchymal thickness, posterior lateral abdominal wall muscle thickness at the catheter insertion site, the number of repeated nephrostomy procedures, and the cortex-to-skin distance along the catheter path, were also analyzed and compared between the two groups.

Data on blood urea nitrogen (BUN), creatinine (Cr), C-reactive protein (CRP), and white blood cell (WBC) levels at the time of nephrostomy or within 10 days post-procedure were retrospectively collected from the hospital information system and PACS. Hydronephrosis at the time of the procedure, right or left sidedness of the nephrostomy, the specific indication for nephrostomy, presence of malignancy, emergency or elective nature of the procedure (urosepsis patients underwent emergency procedures; elective procedures for other cases), patient age, gender, the number of days until catheter dislodgement, the number of previous nephrostomy procedures, fluoroscopy duration during the procedure, and total air kerma levels were also considered.

Statistical analysis was performed using SPSS for Windows version 18.0 (SPSS Inc., Chicago, IL, USA). The normality of data distribution was assessed through visual methods (histograms and probability plots) and analytical methods (Shapiro–Wilk test). Descriptive statistics for numerical data were reported as mean ± standard deviation or median (interquartile range), as appropriate. Categorical variables were summarized as frequencies and percentages.

For comparisons, the independent samples t-test was applied to normally distributed numerical data, while the Mann–Whitney U test was used for non-normally distributed data. Categorical variables were compared using the Pearson chi-square test.

Among 742 PCN procedures, the prevalence of catheter dislodgment was 5.1% (38/742). In the dislodged group, 68.4% were male compared to 44.7% in controls (p = 0.159). Other characteristics, such as hydronephrosis and the number of catheter placements, were not significantly different (p = 0.986 and p = 0.472). The median time to dislodgment was 35.5 days. Placement on the right side (55.3% dislodged vs. 57.9% controls, p = 1.0) and fluoroscopy times (p = 0.093) were comparable between groups. Malignancy was present in 57.9% of dislodged cases and 68.4% of controls (p = 0.476). BMI was not significantly different (24 vs. 27.5, p = 0.166) (Table 1).

The three most common diagnoses were kidney stones (28.9% total; 36.8% dislodged, 21.1% controls), bladder cancer (25%; 26.3% dislodged, 23.7% controls), and prostate cancer (14.5%; 13.2% dislodged, 15.8% controls) (Table 2).

The median subcutaneous fat thickness did not differ significantly between the groups (22.46 mm vs. 32.19 mm, p = 0.062). Other parameters, including PMA (p = 0.593), paravertebral muscle area (p = 0.138), paravertebral muscle thickness (p = 0.771), kidney coronal length (p = 0.839), lower pole renal parenchymal thickness (p = 0.644), and posterolateral muscle thickness (p = 0.740), were similar between the two groups. However, the cortex-to-skin distance was significantly shorter in the catheter dislodgment group compared to the controls (p = 0.001) (Table 3).

ROC analysis revealed an AUC of 0.67 for cortex-to-skin distance, with an optimal threshold of 46.65 mm, sensitivity of 92.1%, specificity of 39.5%, and PPV of 60.3% (Fig. 6).

This study has several limitations. The retrospective, single-center design and relatively small sample size may restrict the generalizability of our findings. The reliance on age and sex as matching variables, while justifiable due to their availability and clinical relevance, may overlook other potential confounders, such as comorbidities or catheter care practices. Although BMI would have been a useful variable for propensity-score matching, it was not included in our study due to data limitations, which may have affected the balance of unmeasured confounders. Additionally, the inclusion of imaging parameters, such as paravertebral and psoas muscle measurements, was exploratory in nature, and their lack of significance highlights the need for further validation in larger cohorts. Variability in procedural techniques among multiple operators and the absence of data on catheter care practices and patient education levels further limit the study’s scope. Addressing these limitations through prospective, multicenter studies with standardized protocols could provide a more robust understanding of the factors influencing catheter dislodgment.