Catheter-Related Complications among Children Treated with Hemodialysis: A Prospective Observational Study

Introduction

Renal replacement therapy (RRT) is increasingly utilized in children with acute kidney injury (AKI) refractory to conservative management and for maintenance therapy in end-stage kidney disease (ESKD).^1,2 Among RRT modalities, hemodialysis (HD) accounts for a substantial proportion of pediatric RRT and requires reliable vascular access via an arteriovenous fistula (AVF), a graft, or a central venous catheter (CVC). Although AVF is preferred for long-term access, CVCs are most frequently used for both acute and chronic indications because of their ease of insertion, the ability to initiate dialysis immediately, and feasibility in emergencies and resource-constrained settings.³

CVC use is, however, associated with infectious, thrombotic, and mechanical complications that may interrupt dialysis delivery and adversely affect patient outcomes.⁴ Reported catheter-related complication rates range from 0.9–22.4 episodes per 1,000 catheter-days.⁵

Most published evidence derives from adult cohorts or from pediatric studies conducted in high-income countries, where patient profiles, access strategies, and infection-control resources differ considerably from those in low- and middle-income settings. Indian pediatric data remain sparse; existing studies are largely retrospective, focused on a single outcome, or limited by small sample sizes.

Given these limitations, we conducted a prospective observational study at a tertiary referral center to systematically evaluate the incidence, spectrum, and determinants of catheter-related complications in children on HD, as well as factors associated with catheter longevity.

Materials and Methods

This prospective cohort study was conducted from July 2021 to December 2022 at the Pediatric Nephrology unit of a tertiary care teaching hospital in India. The study was approved by the Institutional Ethics Committee (INT/IEC/2021-21-1308) and conducted in accordance with the principles of Good Clinical Practice and the Declaration of Helsinki. Written informed consent was obtained from parents or legal guardians; assent was obtained from children aged ≥7 years.

Children aged 3–14 years weighing >15 kg who required HD for AKI refractory to conservative medical therapy or for maintenance therapy in ESKD were eligible for inclusion [Figure 1]. The weight threshold of >15 kg was applied because double-lumen HD catheters of appropriate caliber are not reliably available for smaller children at our institution, in whom peritoneal dialysis is preferred. Exclusion criteria were pre-existing tunneled catheters placed at another center and refusal of consent. Patients were followed from catheter insertion until catheter removal, hospital discharge, dialysis discontinuation, or death.

Close

Figure 1: Flow diagram for study participants. AKI: Acute kidney injury; ESKD: End-stage kidney disease; HD: Hemodialysis; CVC: Central venous catheter

Export to PPT

HD was performed using Fresenius 4008S or 5008S machines. Vascular access type was determined by clinical urgency and anticipated dialysis duration: non-tunneled double-lumen catheters for short-term or emergent access, and tunneled cuffed catheters (Permcath) for anticipated maintenance HD.

Catheter insertions were performed by consultant pediatric nephrologists or senior residents trained in central venous access. Real-time ultrasound guidance was used in all cases where equipment was available and patient stability permitted; the proportion of procedures performed under ultrasound guidance was recorded. A standardized bloodstream infection prevention bundle was applied to all insertions, comprising hand hygiene, maximal sterile barrier precautions, chlorhexidine gluconate skin antisepsis, and optimal catheter-site selection. Up to two insertion attempts per operator were permitted; the number of attempts was documented. A standardized heparin-locking protocol was used after every dialysis session to reduce thrombosis risk. Catheter characteristics, including site, type, size (Fr), length, and cumulative catheter-days, were recorded prospectively.

Catheter-related bloodstream infection (CRBSI) was classified as follows

• Definite CRBSI: Isolation of the same organism from the catheter hub and a peripheral blood culture with a colony count ratio of ≥3:1 (hub: peripheral) or a differential time to positivity of ≥2 hours.

• Probable CRBSI: ≥1 positive blood culture plus clinical signs of infection without an identifiable alternative source.

• Possible CRBSI: Resolution of symptoms following antibiotics or catheter removal in the absence of positive cultures or an alternative infectious focus.⁶

Mechanical dysfunction was defined as failure to achieve or maintain the prescribed blood flow rate (target ≥150 mL/min, or ≥3 mL/kg/min in smaller children) due to obstruction or catheter malposition, requiring clinical intervention.⁷ Catheter-related thrombosis was defined as ultrasound-confirmed venous thrombus associated with impaired catheter flow or dysfunction.⁸ Insertion-related bleeding was defined as peri-procedural hemorrhage requiring prolonged compression (>10 min) or exit-site intervention.⁹

Results

A total of 97 children (57.7% male; mean age 103.5 ± 35.0 months) underwent CVC insertion during the study, with a median dwell time per catheter of 14 days (IQR 7-20). The most common indications for HD were chronic kidney disease (CKD) (30.9%), nephrotic syndrome (NS) (22.7%), rapidly progressive glomerulonephritis (RPGN) (13.4%), and hemolytic uremic syndrome (HUS) (12.4%); 20.6% had prior HD exposure [Table 1]. Non-tunneled catheters were used in 91 cases (93.8%) and tunneled in 6 (6.2%), reflecting institutional practice in which tunneled access is reserved for established maintenance HD. The right internal jugular vein (IJV) was the most common insertion site (70/97; 72.2%), followed by the right femoral vein (21.6%), with left IJV and left femoral vein each accounting for 3.1% of insertions. The subclavian vein was avoided in all cases.

Complications

Thrombosis

Catheter-related thrombosis occurred in nine patients (9.3%; incidence 1.38/1,000 catheter-days), significantly reducing median catheter survival to 31.6 days (IQR 6.8–56.3) compared with 72.1 days (IQR 60.3–83.9) in patients without thrombosis (p <0.001) [Table 2]. Thrombosis was associated with adverse catheter outcomes (χ² = 17.83, p = 0.002): reinsertion was required in 6/9 cases (66.7%), dialysis continued on heparin in one (11.1%), and was discontinued in 2 (22.2%).

Catheter survival analysis

Kaplan-Meier analysis demonstrated significant differences in catheter survival by insertion site (log-rank p <0.001), insertion-related hematoma (p <0.001), CRBSI (p <0.001), and thrombosis (p = 0.008) [Figure 2]. Mean catheter survival was longest for right IJV catheters at 77.9 days (95% CI 65.3–90.5), followed by right femoral at 49.8 days (95% CI 25.6–73.9), left IJV at 9.0 days (95% CI 4.0–14.0), and left femoral at 3.3 days (95% CI 2.7–3.8). Insertion-related hematoma shortened mean survival to 3.6 days (95% CI 1.3–5.9) compared with 78.4 days (95% CI 66.8–89.5) without hematoma. Thrombosis similarly shortened catheter survival to 31.56 days (95% CI 6.78–56.34) compared with 72.14 days (95% CI 60.34–83.94; p = 0.008) without thrombosis.

Close

Figure 2: Kaplan−Meier catheter survival curves. (a) − Survival by the presence of insertion-related hematoma. (b) − Survival by catheter insertion site. (c) − Survival by the presence of catheter-related thrombosis.

Export to PPT

Discussion

This prospective cohort study, among the largest published from India evaluating CVC complications in pediatric HD, identified CRBSI as the leading complication (3.21 episodes per 1,000 catheter-days), followed by insertion-related bleeding/hematoma (1.99), mechanical dysfunction (1.84), and thrombosis (1.38). Access site, prior HD exposure, and insertion-related bleeding were independent predictors of catheter-related complications on multivariable analysis, while serum albumin and thrombosis showed trends toward association that did not reach statistical significance. Right IJV access demonstrated markedly superior catheter survival compared with femoral sites.

The predominance of right IJV use (72.2%) in our cohort reflects its well-established advantages: a straighter anatomical course to the right atrium, lower rates of malposition, fewer insertion attempts, and reduced exposure to groin colonization compared with femoral sites.^10,11 Femoral catheter use persists for emergencies despite a higher risk of infection and thrombosis from groin colonization and mechanical instability. Femoral catheters, though often unavoidable in emergencies or when upper-body access is exhausted, carry higher infection and thrombosis risks attributable to dense perineal flora and greater mechanical instability.¹⁰ The dramatically elevated hazard ratio for left femoral access (HR 9.71) should be interpreted with caution, given the very small subgroup size (n = 3). We avoided subclavian access entirely due to the risks of pneumothorax, hemothorax, and venous stenosis, which are particularly problematic in children who may require future AV fistula creation.

Non-tunneled catheters accounted for 93.8% of insertions and 87.5% of all complications, a proportion that primarily reflects their predominant use in the cohort rather than an inherently higher per-catheter risk. Non-tunneled catheters lack the subcutaneous cuff that impedes bacterial migration along the external catheter surface, a mechanism believed to account for the 2- to 3-fold higher complication risk reported for non-tunneled versus tunneled HD catheters by Weijmer et al.¹² Notably, however, the median functional duration of non-tunneled catheters in our cohort was 13 days (IQR 7–18), with 50% remaining functional until elective removal beyond 30 days, outcomes that compare favorably with the conventional recommendation for replacement within 14 days and suggest that carefully managed non-tunneled catheters can have acceptable longevity in resource-limited settings.

Our CRBSI rate of 3.21 episodes per 1,000 catheter-days exceeds rates reported in pediatric studies restricted to tunneled catheters, 1.72/1,000 catheter-days¹³ and 1.6 episodes/1000 catheter-days¹⁴, a difference likely attributable to our inclusion of predominantly non-tunneled catheters and the resource-constrained environment. The microbial profile showed a preponderance of Gram-positive organisms (MRSA, MSSA) and Enterococcus faecalis, consistent with hub contamination and skin colonization as the dominant pathogenic mechanisms in catheter-related infections. Gram-negative isolates (Acinetobacter baumannii, Pseudomonas aeruginosa) were also identified and may reflect nosocomial acquisition in an intensive care setting. This contrasts with the predominantly Gram-negative bacteremia reported by Shimon et al.,¹⁵ which was largely driven by urinary and abdominal sources rather than catheter-attributable infection.

Hypoalbuminemia is recognized as a marker of impaired host defense and has been associated with increased CRBSI risk and longer hospitalization in patients undergoing HD.¹⁶ In our cohort, serum albumin >3 g/dL at insertion was associated with a trend toward fewer complications and better catheter survival (HR 0.43, p = 0.065). While this did not achieve statistical significance, likely due to limited power, the clinical implication is that nutritional optimization before and during dialysis may be a modifiable strategy for improving catheter outcomes.

Catheter-related thrombosis occurred at 1.38 episodes per 1,000 catheter-days, closely approximating the rate of 1.3/1,000 catheter-days reported by Napalkov et al.¹⁷ Thrombosis rates reported in the literature range widely (0.6–33%), with higher rates observed in hematological malignancies and other prothrombotic states.¹⁸ The relatively low thrombosis incidence in our cohort may in part reflect routine post-dialysis heparin locking and proactive ultrasound surveillance, practices incorporated into our care bundle but not consistently adopted across all published series.

Peri-procedural complications were infrequent and clinically manageable. Insertion-related bleeding or hematoma occurred in 13 patients (13.4%) but resolved conservatively in all cases, comparable to rates reported by Kander et al.⁹ Ultrasound guidance was employed systematically whenever feasible and likely contributed to the absence of kinking, arterial puncture, and pneumothorax in our series. Accidental catheter removal occurred in 2 right IJV cases (2.1%), a substantially lower rate than the 9.3% reported by Kojima et al.¹⁹ and Levy et al.,²⁰ who used adhesive tape rather than sutures to secure catheters, underscoring the importance of a secure fixation technique.

Prior HD exposure emerged as an independent risk factor for complications (HR 0.49, p = 0.024). This finding may reflect underlying vascular changes from repeated catheterizations, altered immune competence in patients with prolonged or recurrent kidney disease, or selection bias toward more complex patients who have already required dialysis. Prospective characterization of vascular access history in future studies would help clarify this association.

Amid scarce prospective data on pediatric HD CVC complications in India, this study provides one of the largest prospective datasets from an Indian tertiary-care setting evaluating both infectious and non-infectious catheter-related complications and identifying predictors of catheter failure.

This study has several limitations. As a single-center prospective cohort conducted at a tertiary referral hospital, findings may not generalize to primary or secondary care settings, to infants and neonates, or to regions with different resource availability, patient profiles, or antimicrobial resistance patterns. The tunneled catheter subgroup was small (n = 6; 6.2%), precluding robust comparison between catheter types. The study period and follow-up duration were insufficient to capture late complications or transition to permanent access modalities. Formal a priori sample size calculation was not performed; accordingly, the study should be considered exploratory in its multivariable findings, particularly for subgroups with very small sample sizes (left IJV, left femoral).

A majority of CRBSI cases were not microbiologically confirmed, relying instead on clinical and surrogate criteria; this may have inflated the reported infection rate and limited the precision of CRBSI-related incidence estimates. Similarly, the markedly elevated hazard ratio for left femoral access (HR 9.71, 95% CI 2.70–34.89) is derived from a very small number of events and should be interpreted with caution, as wide confidence intervals reflect substantial statistical imprecision rather than a definitive effect size.

Unmeasured confounders, including inter-operator variability, antibiotic stewardship policies, and changes in practice over the study period, may further influence outcomes. Future multicenter studies with larger, more diverse cohorts, standardized microbiological confirmation protocols, and formal sample size calculations are warranted to validate and extend these findings.

In children undergoing HD via CVCs, CRBSI was the most frequent complication, and femoral access, prior HD exposure, and insertion-related bleeding were independently associated with adverse catheter outcomes. The right IJV placement was associated with the longest catheter survival. These observational findings are hypothesis-generating and warrant validation in larger multicenter pediatric cohorts, especially in low- and middle-income settings.