The Association Between Upper and Lower Extremity Deep Vein Thrombosis and Peripherally Inserted Central Catheters: Think Below the Waist

Background:

It is well known that peripherally inserted central catheters (PICCs) are associated with an increased risk of venous thromboembolism (VTE). While most PICC VTE studies have focused on upper‐extremity deep vein thrombosis (DVT), some suggest that PICCs may also be associated with lower‐extremity DVT. The extent of this association and its magnitude compared to upper extremity thrombosis is unknown. Using data from a multi‐hospital quality initiative, we examined the association between PICC use and upper and lower extremity DVT. We hypothesized that PICCs would be associated with an increased risk of thrombosis in both extremities.

Methods:

Since 2011, the Hospital Medicine Safety Consortium (HMS) has collected detailed demographic and clinical data related to VTE in hospitalized medical patients at 40 hospitals across Michigan. Using the HMS database, we defined a cohort of patients who had a PICC present by hospital day 2. The association between PICC placement and ultrasound confirmed, symptomatic upper and lower extremity DVT and/or pulmonary embolism within 90‐days of hospital discharge was then examined. We compared risk of upper extremity DVT, lower extremity DVT and pulmonary embolism among patients with PICCs to those without these devices. To account for confounding, multivariable, logistic regression models that included a priori specified covariates of interest (age>70, cancer, past history of venous thromboembolism, immobilization, leg swelling) were fit. Stata SE V.11 was used for all analyses.

Results:

Data were collected on a total of 52,989 patients. At 90‐days, the overall VTE rate was 0.56%. A total of 100 (0.19%) patients developed a pulmonary embolism, 56 (0.11%) developed a symptomatic upper extremity DVT, and 143 (0.27%) developed a symptomatic lower extremity DVT. 942 (1.8%) patients had a PICC in place by hospital day 2. Following multivariable adjustment, PICC presence by hospital day 2 was associated with a significantly increased risk of upper extremity DVT, (OR=10.1 [4.2 – 24.2], p <0.001), lower‐extremity DVT (OR=3.1 [1.3 – 7.5], p = 0.014) and all VTE (including PE) (OR=3.1 [1.8 – 5.5], p <0.001).

Conclusions:

In this analysis of a large, hospitalized general medical population, the presence of a PICC by hospital day 2 was independently associated with an increased risk of DVT in both upper and lower extremities. While observational in nature, this analysis strengthens the call for mechanistic studies to better understand this association. Longitudinal studies that measure biological markers of inflammation and thrombosis (e.g., d‐dimer, C‐Reactive Protein) in the presence of a PICC thus appear necessary. Because PICCs are highly thrombogenic and increasingly included in contemporary VTE‐risk stratification algorithms, hospitalists should be cautious with inserting PICCs in patients who are at risk of both upper and lower extremity thrombosis.