Central venous catheter‐related complications in hematologic patients: An observational study

The frequency of central venous catheter (CVC)‐related complications in hematologic patients has previously been studied but some uncertainty remains. Therefore, this observational cohort study was designed primarily to investigate mechanical and infectious complications related to CVC insertion in hematologic patients and secondarily to identify factors associated with these complications.

implemented worldwide. 3,4 These measurements have been shown to decrease the incidence of CVC-associated complications and overall hospital costs. 1,3,5 A few studies have been published outlining mechanical and infectious CVC-related complications in hematologic patients. However, those studies were either small, or the CVC insertion in the studies was mainly performed without ultrasound guidance and included only tunneled silicone CVCs. 2,6 Therefore, we designed this retrospective observational study with the primary aim of investigating the prevalence and incidence of CVC-related mechanical and infectious complications after insertion of nontunneled, noncoated CVCs in a cohort of hematologic patients. The secondary aim was to explore factors potentially associated with CVC complications. We hypothesized that hematologic patients have a high risk of both mechanical and infectious complications after CVC insertions and that several risk factors for complications can be identified.

| ME THODS
This study was approved by the Swedish Ethical Review Authority (dnr 2014/916 and 2018/866). The requirement for written informed consent was waived. The study was carried out at the Department of Intensive and Perioperative Care at Skåne University Hospital, Lund, Sweden. The manuscript was prepared according to the STROBE guidelines for observational studies. 7 Details on the study methods are presented in Figure 1. The subclavian vein was the preferred site of CVC insertion, since the patients were considered at high risk of infection and because the CVC was expected to remain in situ for more than a week.

| Outcomes
The primary outcomes were moderate and severe mechanical complications and infectious complications.
Mechanical complications included bleeding complications, arterial puncture, pneumothorax, arrhythmia, and nerve injury/rhizopathy within 48 h of insertion. Severity of mechanical complication was graded according to the Common Terminology Criteria for Adverse Events (CTCAE, Version 5.0) 8

| Statistical analyses
The sample size was based on the number of available patients during the study period. Results were expressed as a median (interquartile range) for continuous variables and a number (percentage) for categorical variables. When comparing binary variables, the chi-square F I G U R E 1 Flow charts describing the study methods. All inserted catheters together with the daily inspections and their management was documented in the patient's electronical file. Detailed data on catheter insertion and management, together with all microbiological data, were manually extracted and analyzed. Every patient had the right to opt out from the study. a Blood samples were obtained within 24 h prior to cannulation, and prophylactic coagulation enhancement was considered in patients with preprocedural coagulopathy, defined as platelet count <50 × 10 9 /L, prothrombin time (PT-INR) > 1.8, or APTT > 43 s (equal to more than 1.3× upper normal value). b CDC, central dialysis catheter; CVC, central venous catheter. Management of catheters is described in reference [13] (Thorarinsdottir et al.). c All catheter insertions and any preprocedural complication were documented in the electronic medical chart using a standardized CVC insertion template described in reference [12] (Björkander et al.). d Nonantimicrobial transparent catheter dressings (Tegaderm TM ; 3M) were changed every fifth day or more frequently if needed. Once the old dressing was removed, the insertion site was cleaned with 0.9% saline solution and washed with 0.5% chlorhexidine/70% alcohol (SCHA) solution. e The CVCs were removed after site treatment with 0.5% chlorhexidine/70% alcohol (SCHA) solution and the distal 5 cm of the catheter was cut and submerged in a culture tub, which was sonicated in 10 ml of broth. 0.1 ml of the broth was quantitatively cultured on blood agar plates. Growth of >10 2 CFU/catheter was considered significant colonization. The BACT/ALERT system (BioMérieux) was used for blood cultures. All bottles were incubated until microbial growth was detected or for a maximum of 5 days. g Electronic medical records of all patients eligible for inclusion were manually reviewed, and relevant data were extracted and entered in the research database (Microsoft Access, version 2016) test was applied. To evaluate the associations between independent variables and grades 2-4 bleeding and CRI, univariable and multivariable logistic regression analyses were applied. The selection of independent variables in the regression analyses was based on results from previous studies 12,13 and the significant results in the univariate analysis were further analyzed in a multivariable analysis.
The Hosmer-Lemeshow test was used to test the goodness of fit for multivariable testing. p < .05 was considered significant and all tests were two-tailed. All analyses were performed using SPSS 25 (SPSS Inc/IBM).

| RE SULTS
Baseline characteristics of patients and CVCs are presented in Table 2. In summary, a total of 589 catheter insertions were performed in 387 patients, during the study period of 71 months.
The majority of catheters inserted were single lumen 469/589 (80%) and most catheters were inserted in the subclavian vein 433/589 (74%). Of all patients (n = 387), 256 (66%) were men and the most common diagnosis was acute myeloid leukemia 153/387 (40%). TA B L E 1 Mechanical complications classified according to their severity, based on the terms defined by the Common Terminology Criteria for Adverse Events 8

| Outcomes
Detailed data on outcomes per insertion site are presented in Table 3.

| Mechanical complications
In Three cases (0.5%) presented with a pneumothorax on the same side as the catheter insertion, all verified with a plain chest X-ray. In 2/3 cases (67%), ultrasound-guided insertion was applied.

| Infectious complications
A detailed description of each case with sCRI can be seen in File S2.
In summary, 69 patients were diagnosed with sCRI (12%), yielding an incidence of 3.7 sCRI/1000 catheter days during the study period.
The incidence of sCRI in catheters placed in the subclavian vein was 3.3/1000 catheter days and 5.6/1000 catheter days in the internal jugular vein. There was no difference in the incidence of sCRI after subclavian compared with internal jugular insertions (p = .58). Out of all 69 sCRI, 12 were further classified as CRBSI, yielding a prevalence of 2.0% and an incidence of 0.64 CRBSI/1000 catheter days (Table 3).
Out of all 589 insertions, 373 (63%) cases were given antibiotics Detailed results of the uni-and multivariable regression analyses on sCRI are shown in Tables 4 and 5. High BMI (p = .031) and male gender (p = .002) were both independently associated with sCRI in the multivariable analysis.

| DISCUSS ION
This retrospective observational study on nontunneled and noncoated CVC insertions in hematologic patients demonstrated an overall high prevalence of moderate to severe mechanical complications and a high sCRI prevalence (12%) with an sCRI incidence of 3.7 sCRI/1000 catheter days. These results are compared with previously published studies on hematologic patients 2,6 and higher than earlier reports on general cohorts 12,13 and should therefore be carefully considered before cannulation of hematology patients.

| Mechanical complications
In a previous study on CVC insertions in an unselected cohort of   Since these terms differ, a more recent definition-suspected CRI-used in this manuscript, has been suggested in recent expert consensus-based clinical practice guidelines. 11 Many use CRBSI as a measure of CRI. However, as pointed out by others, one problem with the strict CRBSI definition in hematologic patients is that the majority of patients are treated with intravenous antibiotics both prior to catheter insertion and throughout its usage. 19 As many organisms in significant CRIs are hidden from antibiotics in a biofilm on the catheter but killed by antibiotics if released into the bloodstream, many blood cultures could yield false-negative results. 20 Therefore, we argue that a broader definition such as sCRI, despite lowering the specificity, increases the sensitivity for CRIs.
In a recent report from our group, Thorarinsdottir et al. studied CRIs in a general patient cohort, after the implementation of hygiene bundles, reported a CRI prevalence of 2.6% with an incidence of 1.9/1000 catheter days. 13 In, the present study on hematologic patients the prevalence was 12% and the incidence 3.7/1000 catheter days. These higher numbers underline the importance of daily inspections, evaluation of future needs, and extra precautions to avoid CRI in these susceptible hematologic patients.
Higher BMI and male gender were associated with sCRI in this cohort. Obesity has previously been identified as a risk factor for sCRI in critically ill patients. 21,22 In a prospective study on nonhematologic patients, Dossett et al. suggested that obese patients have an increased risk of sCRI because of longer severity-adjusted ICU stays, increasing their risk of nosocomial infections. 22 Another theory is that it could be due to increased perspiration, as bandages are more

F I G U R E 2
The Venn Diagram shows the relationship between preprocedural coagulation defects and the correction of hemostasis. The total number of isolated coagulopathies are presented separately, whereas the number of cases receiving any pro-coagulative treatment before insertion are presented within brackets, as explained in the figure above. Two hundred and two cases presented isolated platelet count <50 × 10 9 L, two cases presented isolated PT-INR > 1.8, and three cases had an isolated APTT > 43 s. For more details on the preprocedural correction of coagulopathy, see File S1. APTT, activated partial thromboplastin time; PT-INR, prothrombin time-international normalized ratio.
prone to detach leading to a risk of wound infection. 23 Further studies are needed to determine the association between high BMI and sCRI in hematologic patients.
Reports on male gender being a risk factor for sCRI in hematologic patients are scarce but has been described as a risk factor for sCRI in nonselected patients. 24 However, in one randomized TA B L E 4 Univariable regression analyses for each outcome variables. Independent variables were chosen based on previous studies. 12,13 Highlighted variables were further analyzed in a multivariable regression analysis (

| Limitations
One of the major limitations of this study is its retrospective nature and the risk of missing data. Although there is a well-established routine for documenting catheter-related complications and securing infection-suspected catheter tips with simultaneous peripheral blood cultures in the studied departments, adherence to the routines cannot be guaranteed. Furthermore, the lack of uniformity in defining CRIs is an issue when reporting data. Moreover, this study was performed in a selected cohort with low incidences of some outcomes, making it impossible to include some risk factors in the multivariable regression analyses.

| Conclusions
Patients with hematologic malignancies have a high risk of both grades 2-4 bleeding and sCRI after CVC insertion.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.

AUTH O R CO NTR I B UTI O N S
Data was collected and analysed mainly by MMR with contributions from HRT and TK. VL contributed with hematology specific data.