Complications associated with the use of peripherally inserted central catheters and midline catheters in COVID-19 patients: An observational prospective study

Abstract

Background

Among the many interesting aspects of clinical care during the SARS-CoV-2 pandemic, vascular access still deserves some attention. Peripherally inserted central catheters (PICCs) and midline catheters (MCs) are venous access devices inserted by ultrasound-guided puncture of veins of the arm, which have been associated with the possibility of minimizing infectious complications in different populations of patients. We have investigated their performance in SARS-CoV-2 patients.

Methods

As the incidence of catheter-related bloodstream infections (CRBSI) in patients hospitalized for COVID-19 is still unclear, we have designed a single-center, prospective observational study enrolling all patients with established diagnosis of SARS-CoV-2 infection who were admitted to our hospital in the period between October 2020 and April 2021 and who required either a PICC or a MC.

Results

We recruited 227 patients. The cumulative incidence of CRBSI was 4.35% (10 cases), that is, 3.5 episodes/1,000 catheter days. Four CRBSI occurred in patients with PICCs (4.5/1,000 catheter days) and 6 in those with MCs (3.2/1,000 catheter days).

Conclusions

Our data suggest that COVID-19 patients may have a more pronounced tendency for the development of catheter-related infections compared to other populations of patients.

Background

COVID-19 is a new-onset respiratory infectious disease sustained by the etiologic agent SARS-CoV-2 (Severe Acute Respiratory Syndrome Coronavirus 2). The symptoms consist in a preponderant respiratory involvement, sometimes leading to severe respiratory failure and thromboembolic complications1, 2. During the COVID-19 pandemic, it became clear that many patients hospitalized with SARS-CoV-2 infection were potentially prone to other infectious complications, and several studies have addressed the incidence and prevalence of hospital-acquired infections in these patients3, 4.

Considering that all patients hospitalized with COVID infection usually require a venous access,5, 6 much attention has been focused on the incidence of catheter-related infections. Moreover, COVID-19 patients have also been widely recognized to be potentially prone to catheter-related thrombosis (CRT).7, 8 Though some retrospective studies have tried to define whether the incidence of such catheter-related complications is higher in COVID patients than in non-COVID patients, the results are controversial.9, 10, 11, 12 For example, some of them reported a global rise in frequency of catheter-related bloodstream infections (CRBSI) during the COVID-19 pandemic,9, 10, 11 while others focused on the incidence of bloodstream infections in critically ill COVID-19 patients,¹² including the presence of a central venous catheter as one of the major risk factors. Beyond all, what emerged from the very beginning, as well as the assumption shared by these studies, was the need to strengthen preventive measures in setting up and managing vascular accesses.

The goal of this observational prospective study was to assess the incidence of catheter-related complications—with special focus on CRBSI—in a cohort of COVID-19 patients with 2 different vascular access devices, peripherally inserted central catheters (PICCs) and midline catheters (MCs).

Methods

This study was designed as a single-center, observational, prospective study.

The primary end point was to evaluate the incidence of PICC-related and MC-related bloodstream infections in a cohort of COVID-19 patients admitted to the non-ICU wards of our University Hospital during an 8-month period (October 2020-May 2021).

The secondary end point was to assess the incidence of other relevant catheter-related complications.

Patients

After approval by the local Ethics Committee, we recruited all patients consecutively admitted in non-ICU wards who had a clinical-laboratory diagnosis of SARS-CoV-2 infection and required a PICC or a MC. In our hospital, most COVID-19 patients in ICU (Intensive Care Unit) required the insertion of multiple-lumen centrally inserted central catheters (CICCs) or femorally inserted central catheters (FICCs); however, in COVID-19 patients admitted to non-ICU wards, a MC or a PICC was considered more appropriate.5, 6

For each patient, we recorded the main comorbidities: Chronic obstructive pulmonary disease (COPD, recent pneumonia, pulmonary neoplasms), diabetes (measuring its prevalence in the population under study and noting the cases in which the disease was not well controlled, ie, when glycated hemoglobin was higher than 7.6%), hypertension, and other cardiovascular diseases (atrial fibrillation, heart failure, previous ST-segment (ST segment in electrocardiography) elevation myocardial infarction, or non-ST-segment elevation myocardial infarction). The patients’ body mass index (BMI) was also recorded.

Devices

The choice, insertion, and management of the devices were made in accordance with our hospital policies, which are consistent with the recommendations developed by GAVeCeLT, the Italian Group of Long-Term Venous Access.¹³

A central line was preferred whenever the infusion of irritant or vesicant drugs, or the need for hemodynamic monitoring, was anticipated. If the veins of the arm were of appropriate caliber, a PICC was the preferred central venous access device; CICCs were inserted if the arm veins were too small, or if the patient had chronic renal failure stage 3B or worse, or if a multiple lumen central line (3 or 4 lumens) was needed. FICCs were inserted in patients with contraindications to both CICCs and PICCs. All PICCs were 4 Fr single-lumen or 5 Fr double-lumen, power-injectable, polyurethane, nonvalved, and open-ended.

Peripheral access was considered in patients not requiring hemodynamic monitoring but requiring discontinuous or continuous infusion of solutions and drugs not associated with the risk of endothelial damage. In most situations, the preferred venous access device—as suggested by the GAVeCeLT recommendations5, 6—was a MC, as defined by the 2021 ERPIUP (European Recommendations for Proper Indication and Use of Peripheral venous access) consensus,¹⁴ that is, a 20-25 cm catheter inserted in the deep veins of the arm by ultrasound guidance. We used only 4 Fr single-lumen MCs, power-injectable, polyurethane, nonvalved, and open-ended.

All devices were inserted according to the GAVeCeLT insertion bundles for PICCs and MCs,13, 15 which include pre-procedural ultrasound scan of the veins, proper hand hygiene, skin antisepsis with 2% chlorhexidine, maximal barrier precautions, ultrasound-guided venipuncture at midarm, intraprocedural assessment of the position of the tip, catheter securement by subcutaneous anchorage, sealing of the exit site with cyanoacrylate glue, and coverage with a transparent semipermeable membrane. The tip of the PICCs was located at the junction between the superior vena cava and right atrium, using intracavitary ECG (Electrocardiography); the tip of the MCs was placed in the thoracic tract of the axillary vein, using real-time ultrasound.

All devices were managed according to the GAVeCeLT maintenance bundles for PICCs and MCs,13, 15 which include weekly dressing change, proper periodic replacement of the infusion lines, use of neutral displacement needle-free connectors and disinfecting caps (port protectors), and flushing the device with normal saline before and after use.

Each device was monitored by health care providers who reported through Trakcare (InterSystems Corp), by the information system adopted by our hospital, and by daily updates regarding possible complications, also closely monitoring catheter insertion and removal times, in order to accurately estimate the catheter days.

Complications

We recorded all relevant catheter-related complications, including the following:

• (1)Immediate or early complications directly caused by the insertion maneuver (intraprocedural bleeding, postprocedural hematoma, primary malposition, failure of insertion).
• (2)Catheter-related infections (infection of the exit site, CRBSI): we considered exclusively CRBSI and not all central line-associated bloodstream infections, since in any case of suspected infection the diagnosis of CRBSI would be established only if the culture of catheter blood would become positive at least 2 hours earlier than the culture of peripheral blood, according to the method of differential time to positivity.16, 17
• (3)Symptomatic CRT: in case of clinical signs suggesting CRT, the diagnosis was established by ultrasound scan of the veins of the arm and of the supraclavicular or infraclavicular area. No attempt was done to quantify the incidence of asymptomatic CRT or of the fibroblastic sleeve.
• (4)Accidental dislodgement or removal of the device: partial dislodgement (more than 2 cm of catheter withdrawal) and complete dislodgement were considered, either caused accidentally by the staff or self-inflicted by agitated, noncooperative patients.
• (5)Irreversible lumen occlusion: we considered only lumen occlusions resistant to dis-obstruction maneuver, eventually requiring removal of the device.

Statistical analysis

All relevant data concerning the patients and the devices were recorded on an electronic database, and further analyzed by a standard statistical software (Stata/IC 16.0, StataCorp LLC).

Normal distribution of continuous variables was assessed by the Shapiro-Wilk test. Continuous variables with normal distribution are presented as mean ± standard deviation and were compared using Student's t-test. Continuous variables with non-normal distribution are presented as median and interquartile range and were compared by the nonparametric Wilcoxon rank-sum test. Categorical variables were presented as number of patients, or percentages, with 95% CIs. No statistical sample size calculation was performed a priori, and sample size was equal to the number of patients treated during the study period. P ≤ .05 was considered statistically significant. The analyses have not been adjusted for multiple comparisons, and given the possibility of type I error, the findings should be interpreted as exploratory and descriptive.

Results

A total of 227 COVID patients were recruited in the study. Their characteristics are summarized in Table 1.

Table 1.

Features and main comorbidities of the patients' cohort

Mean age (years)	78.08 ± 13.49
Males (%)	95 (41.8)
Female (%)	132 (58.1)
Previous lung disease (%)	53 (23.3)
Diabetes (%)	47 (20.7)
HbA1C >7.6 (%)	3 (1.3)
Hypertension (%)	121 (53.3)
Cardiovascular disease (%)	140 (61.6)
BMI >30 (%)	51 (22.4)

NOTE. A total of 227 venous access devices were inserted: 69 PICCs (30.4%) and 158 MCs (69.6%); 68 devices (29.9%) were inserted on the left side and 159 (70.1%) on the right side.

As regards PICCs, 58 were single-lumen and 11 double-lumen. One PICC was tunneled, for the purpose of moving the exit site to a more favorable skin area.

BMI, body mass index; HbA1C, glycated hemoglobin; MC, midline catheter; PICC, peripherally inserted central catheter.

Regarding infective complications, we recorded 10 cases of CRBSI (4.4%), all diagnosed by differential time to positivity. Staphylococcus aureus was responsible for 50% of the infections, followed by Candida parapsilosis (30%), Candida glabrata (10%), and Candida albicans (10%).

Considering a total of 2,752 days/catheter, the incidence of CRBSI was 3.5/1,000 days/catheter. PICCs were the cause of CRBSI in 4 cases (5.8%); considering a total of 889 catheter days, the incidence of CRBSI was 4.5/1,000 PICC days.

MCs were the cause of CRBSI in 6 cases (3.8%); considering a total of 1,863 catheter days, the incidence of CRBSI was 3.2/1,000 MC days.

There was a higher incidence of CRBSI in patients with BMI>30; out of 10 CRBSI, 5 occurred in patients with BMI>30 (2 were PICC-related and 3 MC-related).

The mean time of CRBSI occurrence after insertion was similar for PICCs and MCs (see Table 2).

Table 2.

Total catheter days and incidence of CRBSI among PICCs and MCs

	PICCs	MCs
Total number	69	158
Catheter days	889	1,863
CRBSI (%)	4 (5.8)	6 (3.8)
CRBSI/catheter days	4.5	3.2
Days before onset of CRBSI (mean)	13.17 + 8.35	14.25 + 3.59
Days before onset of CRBSI (CI)	4.04-21.93	8.53-19.97

NOTE. There was no statistical difference in the incidence of CRBSI or in the time of their onset, comparing PICCs and MCs.

CI, confidence interval; CRBSI, catheter-related bloodstream infection; MCs, midline catheters; PICCs, peripherally inserted central catheters.

As regards noninfective complications, there were no relevant complications at insertion or in the first 48 hours after insertion. We recorded 11 dislodgements (4.84%), including 5 devices removed by the patient and 6 dislodged accidentally by the staff.

There were only 2 cases of symptomatic CRT (0.9%), both in MCs, one occurring very early, 24 hours after insertion, and the other on day 11 after insertion. Thus, the incidence of symptomatic CRT was 0% for PICCs and 1.3% for MCs.

We had no case of irreversible lumen occlusion requiring the removal of the device.

Discussion

In this study, we focused on PICCs and MCs, since their use was quite frequent in COVID patients hospitalized in our non-ICU wards. The indication to their use was based on our hospital policies and on the national recommendations issued during the COVID pandemic,5, 6 suggesting the advantage of PICCs as central lines and of MCs as the peripheral venous access in these populations of patients. In particular, the use of MCs was implemented in COVID patients, for the purpose of ensuring a reliable and durable peripheral venous access for hydration, support therapy, and blood sampling. MCs were considered contraindicated in patients requiring hemodynamic monitoring and/or infusion of irritant or vesicant drugs not compatible with the peripheral route.¹⁵

Our incidence of CRBSI was high compared to the rates commonly reported for non-COVID patients, both in the literature and in the clinical experience of our University Hospital.18, 19, 20

Interestingly, the incidence of CRBSI was similar in PICCs versus MCs, both in percentage and regarding time of onset. This result is in wide contrast to what has been repeatedly stated in the literature, namely that MCs might be associated with a lower infection risk compared to central lines. As a result of this hypothetical lower risk of infection, in some hospitals MCs have been systematically preferred to PICCs, sometimes forcing their indication.

As there is no logical etiopathogenetic explanation for such difference in terms of risk of infection, we are not surprised by our data that show, to our knowledge for the first time and clearly, that MCs are as prone to infection as PICCs, at least in COVID patients.

In our study, the time of onset of infection was approximately 2 weeks (13.6 ± 6.59 days); this value differs from what has been reported in the literature. For example, in a study recently published by Yong et al,²¹ the mean time of onset of infection for cuffed and uncuffed tunneled PICC catheters was 24 days and 19 days, respectively. Therefore, patients with COVID-19 enrolled in our study, in addition to facing a higher risk of CRBSI, demonstrated an earlier tendency to develop this complication.

In our study, insertion-related complications were not reported, probably because of the consistent adoption of well-designed insertion bundles and also because all insertions were performed by specifically trained clinicians of our PICC Team.

As regards the distribution of the pathogens responsible for the catheter-related infection (namely, 3 cases of CRBSI: 1 sustained by C parapsilosis, 1 sustained by C glabrata, and 1 by C albicans), this could be, in our opinion, partially motivated by the frequent use of parenteral nutrition in patients with severe respiratory failure or hemodynamic instability, like those hospitalized due to COVID-19 pneumonia.

One unexpected finding of our study was the very low incidence of symptomatic CRT, which occurred in none of the PICCs and in only 2 of the MCs. The actual incidence of CRT in PICCs has been discussed in many papers in the last decade,²² with some authors suggesting that PICCs may be associated with the higher incidence of CRT compared to CICCs23, 24 versus other authors arguing that, when PICCs are inserted according to the current international recommendations (choice of a proper ratio between catheter caliber and vein diameter, ultrasound-guided puncture, correct intraprocedural verification of the position of the tip, sutureless securement, and so on), the risk of PICC-related thrombosis may be very low or even 0.25, 26, 27 In this regard, the data collected in this study seem to confirm the latter view. Similar low rates of PICC-related thrombosis have been reported in recent clinical studies on oncologic patients.28, 29 Our finding is even more surprising considering the widely accepted thrombophilia of COVID patients. In this regard, probably, this factor was counteracted—and maybe lessened—by the systematic antithrombotic prophylaxis adopted in our population of patients. Since this study focused on COVID-19 patients, it should be pointed out that the finding related to thrombosis may have been mitigated by the concomitant administration of anticoagulant therapy (whether prophylactically dosed or not).

Thus, the most interesting issue is the increased susceptibility of COVID patients to CRBSI, which may be explained by several factors. First, patients affected by COVID-19 are elderly individuals (mean age of 78 ± 13.49 years) and tend to have multiple comorbidities. In other words, patients hospitalized for COVID-19 represent a selected group of individuals at high risk for complications. Also, we found a significant relationship between CRBSI and BMI: in patients with BMI>30, the incidence of CRBSI was 4 times higher than in patients with BMI<30 (2.79% vs 9.8%). The reason for this predisposition of obese patients to infective complications could be attributable to the greater severity of SARS-CoV-2 infection, and hence the increased criticality of pulmonary commitment. Furthermore, the higher incidence of CRBSI in COVID-19 patients could also be explained by a less effective management of the venous access device, due to a less frequent access to the patient for surveillance of the exit site, change of dressing, and replacement of infusion lines. As already stated in the literature,30, 31 due to the risk of exposure, the time spent in the rooms of COVID-19–positive patients was significantly lower than standard. This may inevitably imply the risk of neglecting some fundamental care practices in the daily management of vascular accesses, thus increasing the risk of infection.

Furthermore, the logistic issues related to the pandemic have brought an inevitable turnover of the nursing staff, so that operators with limited experience have been rapidly—and sometimes not fully—trained in the management of venous access devices.

In other words, the higher incidence of CRBSI in COVID patients may be explained not only by intrinsic factors related to the viral infection but also by the fact that this was an older and more obese population of patients, and that the management of the devices might have been suboptimal.

Limitations of the study

Our study has several limitations.

• (1)First, we considered only PICCs and MCs, and not other devices frequently used in COVID patients (such as CICCs and FICCs).
• (2)As a single-center one, this study lacks generalizability.
• (3)All our patients were hospitalized in non-ICU wards; ICU patients, characterized by more severe COVID infections and more comorbidities, may have a different incidence of infective and noninfective complications (including CRT).
• (4)Last, we made no effort to verify the actual compliance of the nursing staff to the maintenance bundles; as previously discussed, suboptimal care of the device may have increased the incidence of infective complications.

Conclusions

Our data show that—at least considering PICCs and MCs in non-ICU settings—the incidence of CRBSI is higher in COVID-19 patients than in the general population. Our study did not detect significant differences in the rates of infection between PICCs and MCs, undermining the literature assumption that MCs are associated with less infective complications than PICCs.

On the other hand, noninfective complications (including CRT) were minimal or absent in PICCs and MCs utilized in non-ICU COVID patients, suggesting that proper insertion bundles and management bundles are effective in minimizing such complications in this population, as much as in non-COVID patients.