Background
The introduction of long term central venous catheters has been one of the most significant advances in medical care leading to its widespread use in a broad range of diseases, including infections and malignancies. Unfortunately, the use of these catheters can be associated with complications, most notably the development of a catheter-related deep vein thrombosis (DVT).
The reported incidence of catheter-related thrombosis (CRT) has been highly variable, ranging from 12-60% in various studies [
1,
2]. However, it has been commonly noted that patients with malignancy are at an increased risk of developing this complication [
3,
4]. Furthermore, the impact of a thrombotic event can be clinically significant, particularly in hemato-oncology patients, as they require long-term access for the administration of blood products, chemotherapy and stem cell rescue. However, data for catheter-related thromboses in patients with hematological malignancies has been limited [
5‐
7].
In 2010, Tran et al. were one of the first groups to examine the incidence of CRTs in hemato-oncology patients following insertion of either a long term skin tunneled (LTSTC) or peripheral (PICC) central venous catheter [
8]. They found that the incidence of thrombotic events was higher in those with a peripheral compared to central line, with the majority of patients managed by initiating anticoagulation. Several studies have supported these findings with reported rates of asymptomatic and symptomatic PICC-related thromboses of up to 50% and 22% respectively [
9‐
12].
Despite these data however this has rarely translated into clinical practice with the majority of UK hospitals electing to use PICCs as the procedure is considered less invasive. This is due to the reported increased incidence (approximately 3%) of serious mechanical complications in patients undergoing central venous catheter (CVC) placement, including arterial puncture (incidence range 3.1-15%) and pneumothorax (incidence range 0.1-3.1%) [
13,
14].
Nevertheless the risk of catheter-related thrombosis is not insignificant and can be associated with significant patient morbidity and potential mortality [
15]. Based on the current literature, it is likely that the incidence of thromboses will be higher in PICCs compared to LTSTCs, but more data is required [
16‐
18]. Accordingly we performed a retrospective cohort study comparing the cumulative incidence of thrombotic events following insertion of either a PICC or LTSTC line in hemato-oncology patients. We also analyzed secondary outcomes including post-thrombotic infection and pulmonary emboli.
Discussion
Over the years there has been an increasing use of PICCs due to the procedure being less invasive and the relatively low rate of line-related complications. Within our hospital trust, the current policy is to use PICCs first line, particularly for newly diagnosed patients who require urgent treatment, while LTSTCs are reserved for those undergoing stem cell rescue. However the risk of catheter-related thrombosis is often not considered. This is an important complication, especially in hemato-oncology patients, as it can potentially result in delayed administration of chemotherapy and expose them to the hazards of therapeutic anticoagulation [
20].
When examining the current literature, there is minimal data available regarding the incidence of catheter-related thrombosis in hemato-oncology patients. There are studies which have been performed in critical care patients, with one case-control study reporting a thrombotic incidence of 2.8% in PICCs compared to 0% in LTSTCs [
21]. Furthermore, a prospective randomized controlled trial in post-critical care patients who had routine surveillance for catheter-related thromboses also observed a higher incidence of thrombotic events in PICCs compared to short term CVCs (27.2% versus 9.6% respectively) [
22].
Ours is the largest most recent study which has focused on comparing the incidence of catheter-related thromboses in hemato-oncology patients. The data we have produced is similar to previous studies, suggesting that PICCs confer an increased risk of catheter-related thrombosis compared to LTSTCs. Furthermore the majority of our patients in both groups presented within the first 100 days of insertion, with the cumulative incidence of thrombosis increasing over time. This would suggest an association between dwell time of line and risk of thromboembolic event, although interestingly the dwell time of line was significantly longer in the LTSTC group compared to the PICC group.
When considering the association between dwell time and risk of thrombotic event there has been mixed data, with some studies showing an association between prolonged dwell time and increased risk of thrombosis, while others have not. Overall the data produced so far does seem to favor a lack of an association between dwell time and risk of thrombosis [
12,
22,
23]. However, within our own study, we were unable to exclude the possibility that longer dwell time contributed towards increased thrombotic risk as we were unable to perform an adjusted analysis.
There has been considerable interest in identifying other potential risk factors for thromboembolism, particularly in cancer patients, in order to prevent thrombotic events. These have been sub-divided into three categories: Patient-related, Disease-related and Treatment-related [
3,
19]. Given that we were examining hemato-oncology patients, we particularly focused on patient- and treatment-related risk factors, which included known inherited prothrombotic states, prior history of venous thromboembolic event and use of immunomodulatory drugs. Though it was unclear whether these risk factors would have the same impact in patients with central venous catheters, we assumed that these factors would contribute towards thromboembolic risk and therefore documented the incidence of each of these risk factors in both the PICC and LTSTC group. The use of IMiDs and anti-coagulant prophylaxis occurred with equal frequency in both groups, with none of our patients having a known inherited prothrombotic condition. However history of venous thromboembolism occurred more frequently in the PICC group compared to the LTSTC group, suggesting this is a contributing factor, despite it not being statistically significant.
Another point to consider in hemato-oncology patients is not only the risk of developing a catheter-related thrombosis, but the potential complications which can follow this. Our secondary outcomes examined the incidence of line-related infection, line removal, major hemorrhage and pulmonary embolism in hemato-oncology patients following development of a catheter-related thrombosis.
The risk of line-related infection has been previously shown to be significantly increased in those with a thrombosis, with colonization and sepsis rates reportedly doubling [
24‐
26]. This would be further enhanced in hemato-oncology patients as they frequently have severe and long-lasting neutropenia, placing them at greater risk of infective complications generally. Our data reflected this with 9 patients in the PICC group compared to only 1 in the LTSTC group developing line-related infection following a thrombotic event, with all of these patients subsequently having their line removed.
Given that our unit performs a large number of stem cell rescues (over 200 per year), we also examined whether development of a thrombosis had an impact on the patient’s treatment plan. Of those who developed a catheter-related thrombosis, 4 patients from the PICC group had their treatment delayed, which included 3 stem cell rescues. The median duration of treatment delay was 13.8 days (range 4-20 days), which ultimately will have resulted in prolonged hospitalization, although this was not formally analyzed within our study. Hence this highlighted not only the short-term impact of a catheter-related thrombosis, but the potential long-term consequences also with delays to treatment. Therefore, this is another factor which needs to be taken into consideration when determining the type of venous access for a hemato-oncology patient.
The high rate of thrombosis in cancer patients with central venous access has led to several studies examining the potential use of systemic anticoagulant therapy [
2]. Warfarin had previously been shown to successfully reduce the risk of catheter-related thrombosis in patients with solid tumors [
27]. However, this can be a risky strategy in hematological patients, due to the intermittent and prolonged periods of chemotherapy-induced thrombocytopenia.
Abdelkefi et al. were able to overcome this in a well-controlled hospital setting where hemato-oncology patients were randomized to receive either intravenous unfractionated heparin or normal saline infusions daily, with a significantly lower incidence of catheter-related thrombosis reported in the former compared to the latter group (1.5% versus 12.6% respectively) [
28]. Furthermore, there were no reported significant bleeding events, suggesting that low-dose unfractionated heparin could be safely administered in this group of patients. Our own data supports the safe use of therapeutic anticoagulation as we reported no major bleeding events, despite our patients being thrombocytopenic.
More recent studies, however, have shown no benefit from thromboprophylaxis in preventing catheter thrombosis [
29‐
31]. All of these were double blind trials, hence providing higher quality evidence compared to the earlier published randomized trials which were not blinded. Therefore, given these conflicting results, the current guidelines released by the American College of Chest Physicians (ACCP) do not recommend the routine use of anticoagulation in cancer patients with central venous lines until additional randomized controlled trials can confirm that the benefits of anticoagulation outweigh the risks [
32].
Given that our study was retrospective, the greatest limitation was the fact that we only identified symptomatic patients who underwent ultrasonographic examination. Recent studies which used prospective designs and screened for venous thrombosis in the absence of clinical symptoms suggest that catheter-related thrombosis is a complication that may be more prevalent than clinically perceived, with incidence rates as high as 64.5% [
18,
33].
In conclusion we observed a higher incidence of thrombotic events in hemato-oncology patients with PICCs compared to CVCs, which was associated with a higher rate of post-thrombotic complications, notably infection. Our findings have potential clinical implications where physicians have to weigh the risk and benefit of either PICC or CVC for central venous access, as this can potentially impact the long-term management of patients with hematological malignancies.
Future randomized controlled trials comparing PICCs against LTSTCs are required which are adequately powered to detect differences in both short-term and long-term outcomes (i.e. PE, hospitalization rates and mortality rates). Further data is also required to address the ongoing debate on the use of pharmacological thromboprophylaxis in patients with central venous access, with a particular focus on those with hematological malignancies.
Competing interests
BES has received honoraria and speakers’ fees from Therakos, a Johnson and Johnson company. All other authors declare that they have no competing interests.
Authors’ contributions
KD provided the data for PICC placements. RR and DC provided the data for LTSTC placements. NP, KW and PS analyzed the data. PS conceived of the idea for the study and participated in its design and coordination, and drafted the manuscript. All authors read and approved the final manuscript.