Risk Factors and Clinical Outcomes of Upper Extremity Deep-Vein Thrombosis in Critically Ill Patients: A Retrospective Cohort Study

Abstract

Background

Upper extremity deep-vein thrombosis (UEDVT) is often under-recognized and perceived as less consequential than lower extremity DVT. This study evaluated the risk factors and outcomes of critically ill patients with UEDVT.

Methods

We evaluated patients in the adult ICUs of a tertiary-care hospital who had Doppler ultrasound for suspected acute upper extremity vein thrombosis between January 2022 and December 2023 and documented the presence of UEDVT and superficial-vein thrombosis (UESVT). We compared patients with and without UEDVT, specifically evaluating the incidence of pulmonary embolism and hospital mortality.

Results

Among 9605 ICU admissions, 501 patients underwent upper extremity ultrasound at a median of 13 days after hospital admission. These patients had a median age of 66 years and were predominantly male (60.3%). 143 patients (28.5%) had UEDVT and 169 (33.7%) UESVT (54 patients had both). UEDVT was significantly associated with central venous catheters, particularly hemodialysis lines. Pulmonary embolism was more common in patients with UEDVT (19.6% versus 10.3%, p=0.005; adjusted odds ratio, 2.08; 95% confidence interval, 1.16–3.71) with a higher rate in patients with multiple versus one affected vein (32.6% versus 13.4%, p=0.007). Hospital mortality was higher in patients with UEDVT (51.7% versus 31.6%, p<0.0001; adjusted odds ratio, 2.05; 95% confidence interval, 1.22–3.45). UESVT was not associated with pulmonary embolism or mortality.

Conclusions

UEDVT is an important complication in critically ill patients and may be associated with an increased risk of pulmonary embolism and mortality, underscoring the importance of prevention, early detection, and appropriate therapeutic approach.

Keywords

deep-vein thrombosis superficial-vein thrombosis venous thromboembolism intensive care pulmonary embolism upper extremity upper limb

Introduction

Upper extremity deep-vein thrombosis (UEDVT) accounts for approximately 10% of all deep-vein thrombosis (DVT) cases.^1,2 It is increasingly recognized in hospitalized patients. The reported incidence varies widely across studies, ranging from 1.6% to 22.8%.^3-5 UEDVT usually involves the brachial, axillary, subclavian and internal jugular veins.⁴ Due to several factors including central venous catheters, cardiac pacemakers/defibrillators, immobility, comorbidities and hypercoagulability,^5,6 UEDVT is more common in critically ill patients and represents a significant proportion of venous thrombotic events. A prospective cohort study of 3746 patients in the intensive care unit (ICU) found that 2.2% had non-leg thrombosis, predominantly UEDVT.⁴ Other studies revealed higher rates. In 113 patients with sepsis, venous thromboembolism (VTE) occurred in 37.2% and 14.2% had central venous catheter-associated UEDVT, accounting for one third of VTE events.⁷ In patients with traumatic brain injury in the ICU, UEDVT accounted for 46.5% of VTE events which occurred in 13.4% of patients.⁸

UEDVT can lead to multiple complications, including pulmonary embolism (PE), DVT recurrence, and post-thrombotic syndrome.^9,10 Studies evaluating the incidence of PE in patients with UEDVT have shown variable rates. One study found that only 2 of 200 patients with UEDVT (1%) had subsequent PE.¹¹ The same group later found a rate of 2% in a cohort of 300 patients diagnosed to have UEDVT based on inpatient or outpatient diagnostic Doppler ultrasound.¹² Another study found a slightly higher PE rate of 4.8%.⁵ In a nested prospective cohort of medical-surgical ICU patients, UEDVT was associated with an almost 12-fold increase in PE risk.⁴ Overall, UEDVT has been associated with lower rates of complications, including PE, compared with lower extremity DVT.^2,6,13,14 For instance, a comparative study found that UEDVT was associated with lower rate of clinically overt PE compared with lower extremity DVT (9.0% versus 29%).² This difference in complication rates may contribute to inconsistent use of therapeutic anticoagulation for UEDVT. In one study, only 73 out of 200 patients (36%) with UEDVT were put on therapeutic anticoagulation.¹¹

Despite increasing recognition, data on UEDVT in critically ill patients remain limited, particularly regarding its clinical significance and associations with PE and mortality. This study aims to address this gap by examining the risk factors for UEDVT and its associated outcomes in ICU patients.

Methods

Patients, Setting and Study Design

This study was a retrospective cohort study conducted in the adult ICUs at King Abdulaziz Medical City in Riyadh. The hospital was a tertiary-care center with > 1400 beds. The Intensive Care department had 8 different units that admit different types of critically ill patients. There are 6 medical teams in the ICU that perform daily multidisciplinary rounds. Each ICU medical team consists of one consultant, one staff physician/fellow, and 2 to 4 residents.¹⁵ All patients, aged 18 years and above who had upper extremity ultrasound between January 1, 2022 to December 31, 2023 for the clinical suspicion of acute DVT, based on clinical signs such as swelling, skin redness or warmth, were included in this study. Patients who had upper extremity ultrasound for reasons other than suspected acute DVT and those with length of stay in the ICU less than three days were excluded. This study was approved by the Institutional Review Board of the Ministry of National Guard – Health Affairs.

Data Collection

The collected data included demographics, comorbidities, admission service, findings of upper extremity ultrasound as read and reported by a radiologist (absence or presence of DVT in the deep veins [ulnar, radial, brachial, axillary, subclavian, internal jugular and innominate] or superficial veins [cephalic and basilic], findings of CT pulmonary angiography, presence of central venous catheter, peripherally inserted central venous catheters, permanent hemodialysis catheter and pacemaker at the time of ultrasound, vasopressor use, thromboprophylaxis at the time of ultrasound, pertinent laboratory tests on ultrasound day (hemoglobin, platelet count, international normalized ratio, activated partial thromboplastin time, and d-dimer) and infection with COVID-19 during the index hospitalization. We also noted anticoagulation practices after UEDVT diagnosis and reviewed the medical records to identify reasons for withholding therapeutic anticoagulation..

The primary outcome was the presence of UEDVT based on ultrasound findings. In this study, we considered DVT in the internal jugular vein, which is considered an axial vein, as UEDVT as this vein is very close to the upper extremity veins and as done by other studies.^3-5 The other outcomes studied were the presence of upper extremity superficial-vein thrombosis (UESVT), lower extremity DVT, PE, length of hospital and ICU stay, and hospital mortality.

Statistical Analysis

Patients were categorized into two groups based on ultrasound findings: patients with UEDVT and those without UEDVT. Continuous data were presented as median with interquartile range. Categorical variables were presented as frequency and percentage. To compare between-group differences, Student’s t-test or Mann Whitney U test was used for continuous data, as appropriate depending on the normality of data distribution, and the chi-square test or Fisher’s exact test for categorical data. We performed multivariate logistic regression analyses to identify the risk factors for UEDVT and UESVT in our cohort. The variables entered into the model were selected based on their clinical relevance and data availability and included baseline demographics, clinical characteristics (comorbidities, presence of venous catheters, vasopressor use, COVID-19 status) and time from hospital admission to perform Doppler ultrasound. We also performed a multivariable logistic regression analysis to assess whether UEDVT or UESVT were independent predictors of PE and hospital mortality. The additional variables entered in this model were baseline demographic and clinical characteristics (comorbidities, lower extremity DVT and vasopressor use). All variables entered into the regression models were abstracted from medical records, and none had missing values. The results of the regression models were presented as odds ratios (ORs) with 95% confidence intervals (CIs). We assessed model performance using Hosmer and Lemeshow goodness-of-fit test and Nagelkerke R² statistic. A two-sided p-value <0.05 was considered statistically significant. Data analysis was performed using SPSS version 15 (Statistical Package for the Social Sciences, IBM Corp., Armonk, NY, USA).

Results

Patient Characteristics

During the study period (January 2022 and December 2023), 9605 patients were admitted to the study ICUs. 501 of these patients (5.3%) had at least one ultrasound of the upper extremity and were eligible to be studied (Figure 1). The median time from admission to ultrasound was 13 days. Table 1 shows their characteristics. The median age was 66 years, the majority (60.3%) were males, and the median body mass index was 26 Kg/m². Most patients (80%) had chronic comorbidities with hypertension (62.7%), diabetes (53.5%), and chronic kidney disease (25.9%) were among the most prevalent conditions. Cancer and previous VTE were present in roughly 20% and 18% of patients, respectively. Patients were admitted across various services, with the majority (65.1%) under medical care. Lower extremity Doppler ultrasound for suspected DVT was performed in 72% of patients, and computed tomography pulmonary angiography for suspected PE was done in 39%.

Figure 1.

Flow diagram for the study cohort

Table 1.

Characteristics of patients

	All patients N=501	No upper extremity DVT N=358	Upper extremity DVT N=143	P-value
Age (years), median (Q1, Q3)	66.0 (48.5, 77.0)	67.0 (48.0, 77.0)	66.0 (51.0, 77.0)	0.67
Male sex, n (%)	302 (60.3)	214 (59.9)	88 (61.5)	0.72
Female sex, n (%)	199 (39.7)	144 (40.2)	55 (38.5)	0.72
Body mass index (kg/m ² ), median (Q1, Q3)	26.0 (22.1, 31.8)	25.9 (22.1, 32.0)	26.6 (22.1, 31.3)	0.77**
Obesity (BMI ≥ 30 kg/m ² ), n (%)	1158 (31.5)	111 (31.0)	47 (32.9)	0.69
Any chronic comorbidity, n (%)	401 (80.0)	284 (79.3)	117 (81.8)	0.53
Hypertension, n (%)	314 (62.7)	219 (61.2)	95 (66.4)	0.27
Diabetes, n (%)	268 (53.5)	183 (51.1)	85 (59.4)	0.09
Heart failure, n (%)	111 (22.2)	83 (23.2)	28 (19.6)	0.38
Chronic respiratory disease, n (%)	33 (6.6)	23 (6.4)	10 (7.0)	0.88
Chronic kidney disease, n (%)	130 (25.9)	88 (24.6)	42 (29.4)	0.27
Cancer, n (%)	101 (20.2)	72 (20.1)	29 (20.3)	0.97
Previous venous thromboembolism, n (%)	89 (17.8)	60 (16.8)	29 (20.3)	0.35
Admitting service, n (%)
Medical	326 (65.1)	230 (64.2)	96 (67.1)	0.44
Hematology/Oncology	53 (10.6)	38 (10.6)	15 (10.5
Obstetrics/Gynecology	6 (1.2)	6 (1.7)	0 (0)
Surgery	113 (22.6)	81 (22.6)	32 (22.5)
Trauma	3 (0.6)	3 (0.8)	0 (0)
Time to perform ultrasound from admission, median (Q1, Q3)	13.0 (6.0, 24.0)	13.0 (7.0, 21.3)	13.0 (5.0, 30.0)	0.29**
Cardiac pacemaker, n (%)	27 (5.4)	22 (6.1)	5 (3.5)	0.24
Permanent hemodialysis catheter (internal jugular), n (%)	78 (15.6)	45 (12.6)	33 (23.1)	0.003
Central venous catheter (internal jugular or subclavian), n (%)	201 (41.7)	138 (38.5)	71 (49.7)	0.02
Peripherally inserted central venous catheter, n (%)	151 (30.1)	101 (28.2)	50 (35.0)	0.14
Vasopressor use, n (%)	259 (51.8)	179 (50.1)	80 (55.9)	0.24
Tested extremity, n (%)
Right upper	237 (47.3)	164 (45.8)	73 (51.0)	0.001
Left upper	199 (39.7)	158 (44.1)	34 (26.6)
Bilateral	65 (13.0)	36 (10.1)	29 (20.3)
Pertinent Laboratory findings, median (Q1, Q3)
Hemoglobin (g/L)	102 (85, 120)	103 (85, 123)	99 (87, 115)	0.38**
Platelets x 10⁹/L	225 (144, 326)	237 (152, 338)	217 (134, 304)	0.25
Activated partial thromboplastin test (seconds)	30.4 (26.9, 35.5)	29.8 (26.6, 34.5)	31.6 (27.6, 39.0)	0.005**
International normalized ratio	1.2 (1.1, 1.3)	1.2 (1.1, 1.3)	1.2 (1.1, 1.4)	0.02**
D-dimer*	2.9 (1.5, 7.5)	2.6 (1.5, 7.0)	3.9 (1.5, 9.2)	0.54**
Positive COVID-19, n (%)	33/446 (7.4)	23/320 (7.2)	10/126 (7.9)	0.79
Anticoagulation practices at time of ultrasound, n (%)
No pharmacologic prophylaxis	121 (24.2)	94 (26.3)	27 (18.9)	0.08
Prophylactic UFH/LMWH	317 (63.3)	222 (62.0)	95 (64.4)	0.35
Therapeutic anticoagulation	63 (12.6)	42 (11.7)	21 (14.7)	0.81
Superficial-vein thrombosis present, n (%)	169 (33.7)	115 (32.1)	54 (37.8)	0.23
Doppler ultrasound of lower extremity performed for suspected DVT, n (%)	360 (71.9)	261 (72.9)	99 (69.2)	0.41
CTPA performed for suspected PE, n (%)	196 (39.2)	134 (37.5)	62 (33.4)	0.23

BMI: body mass index, CTPA: computed tomography pulmonary angiography, DVT: deep-vein thrombosis, LMWH: low-molecular weight heparin, Q1: first quartile, Q3: third quartile, PE: pulmonary embolism, UFH: unfractionated heparin.

*Data on D-dimer available for only 137 patients.

**Data are significantly skewed. The p-value is based on Mann Whiteny U test.

Upper Extremity Deep-Vein Thrombosis

143 patients (28.5%) had UEDVT (Figure 1), which represented 1.5% of all admitted patients. The right upper extremity was affected more commonly than the left (Figure 2). The most common affected veins were the internal jugular vein (n=69, 48.3% of patients with UEDVT), subclavian vein (n=35, 24.5%) and brachial vein (n=33, 23.1%) (Figure 3).

Figure 2.

Extremity affected by deep-vein thrombosis and superficial-vein thrombosis. 54 patients had concomitant thrombosis of deep and superficial veins

Figure 3.

Veins affected by thrombosis. 97 patients had one thrombosed deep vein, 30 patients had two thrombosed deep veins and 16 patients had three or more thrombosed deep veins

The characteristics of patients with and without UEDVT are shown in Table 1. Patients with UEDVT had similar age, sex, body mass index, and chronic comorbidities compared to those without UEDVT. Patients with UEDVT were significantly more likely to have a permanent hemodialysis catheter (23.1% versus 12.6%, p=0.003) and a central venous catheter (49.7% versus 38.5%, p=0.02). Patients with UEDVT had significantly higher partial thromboplastin time and international normalized ratio. D-dimer levels were higher in the DVT group but did not reach statistical significance. Testing for COVID-19 was performed for 446 patients and was positive at similar rate in both groups (around 7%). Anticoagulation practices varied, with prophylactic UFH being the most common regimen (46.7%). A smaller proportion of patients with UEDVT did not receive anticoagulant thromboprophylaxis (p=0.08).

On multivariable logistic regression analysis (Table 2), permanent hemodialysis catheter in the internal jugular vein was significantly associated with increased risk of UEDVT (OR, 1.926; 95% CI, 1.061–3.498). Days from admission to ultrasound had a borderline p-value (0.099). Other variables, including demographic factors, comorbidities, and clinical interventions, did not reach statistical significance. The model had acceptable calibration and moderate discrimination, but a limited explanatory power (Table 2).

Table 2.

Results of the multivariable logistic regression analysis for the predictors of upper extremity deep-vein thrombosis and superficial-vein thrombosis

	Upper extremity deep-vein thrombosis			Upper extremity superficial-vein thrombosis
Variable	Odds ratio	95% confidence interval	P-value	Odds ratio	95% confidence interval	P-value
Age per 1-yr increment	0.997	0.984-1.010	0.62	1.006	0.994-1.018	0.34
Sex female versus male	1.126	0.739-1.715	0.58	1.162	0.783-1.727	0.46
Body mass index per unit increment	0.998	0.972-1.024	0.86	0.992	0.967-1.017	0.51
Diabetes	1.390	0.829-2.329	0.21	0.997	0.619-1.606	0.99
Hypertension	1.023	0.553-1.894	0.94	0.818	0.465-1.442	0.49
Heart failure	0.813	0.480-1.376	0.44	0.853	0.520-1.401	0.53
Chronic respiratory disease	1.140	0.499-2.605	0.76	0.823	0.370-1.829	0.63
Chronic kidney disease	0.933	0.545-1.598	0.80	0.798	0.477-1.336	0.39
Cancer	1.080	0.645-1.810	0.77	0.909	0.556-1.484	0.70
Previous venous thromboembolism	1.215	0.724-2.041	0.46	1.214	0.743-1.986	0.44
Permanent hemodialysis catheter (internal jugular)	1.926	1.061-3.498	0.03	1.266	0.702-2.284	0.43
Central venous catheter (internal jugular or subclavian vein)	1.405	0.927-2.131	0.11	0.830	0.558-1.235	0.36
Cardiac pacemaker	0.600	0.206-1.749	0.35	1.376	0.574-3.303	0.47
Vasopressor use	1.068	0.705-1.619	0.76	1.060	0.717-1.566	0.77
Peripherally inserted central catheter venous catheters	1.178	0.752-1.845	0.47	1.253	0.822-1.909	0.29
Days from admission to ultrasound	1.008	0.998-1.018	0.10	1.001	0.993-1.008	0.82
Pharmacologic prophylaxis	1.392	0.841-2.305	0.20	1.147	0.731-1.798	0.55

Variables entered in each model: age, sex, body mass index, diabetes, hypertension, heart failure, chronic respiratory disease, chronic kidney disease, cancer, previous venous thromboembolism, cardiac pacemaker, permanent hemodialysis catheter (internal jugular), central venous catheter (internal jugular or subclavian), peripherally inserted central catheter venous catheter, and vasopressor use.

Model performance (upper-extremity deep-vein thrombosis): The model was statistically significant compared with the null model (chi-square value = 24.6, p=0.10) and had a correct classification rate of 72.4%. The Hosmer and Lemeshow goodness-of-fit test was acceptable (p=0.31). Nagelkerke R² = 0.069.

Model performance (upper-extremity superficial-vein thrombosis): The model was statistically significant compared with the null model (chi-square value = 7.3, p=0.98) and had a correct classification rate of 70.8%. The Hosmer and Lemeshow goodness-of-fit test was acceptable (p=0.09). Nagelkerke R² = 0.020.

Therapeutic anticoagulation was provided for the treatment for UEDVT in 103 out of 143 patients (72.0%) using intravenous heparin infusion for 69 patients (48.3%), therapeutic enoxaparin for 27 (18.9%) and oral anticoagulants for 7 (4.2%). Prophylactic dose anticoagulant was used in 28 patients (19.6%) and no anticoagulants in 12 (8.4%). For the 40 patients who did not receive therapeutic anticoagulation, the reasons were ongoing or recent bleeding in 16 patients (40.0%), bleeding risk factors (such as coagulopathy or severe thrombocytopenia) in 12 (30.0%), poor prognosis in 1 (2.5%) and unclear reasons in 11 (27.5%).

Upper Extremity Superficial-Vein Thrombosis

In the study cohort, UESVT was present in 169 patients (33.7%) (Figure 1); 54 patients had concomitant UEDVT. UESVT affected the right side more commonly than the left (Figure 2), involving the cephalic vein in 107 patients (63.3%) and the basilic vein in 75 patients (45.6%) (Figure 3).

Despite evaluating a broad range of clinical and procedural variables in the multivariable logistic regression analysis, no independent predictors of UESVT were identified (Table 2). The model also had acceptable calibration and moderate discrimination, but a limited explanatory power (Table 2).

Outcomes

Table 3 describes the outcomes of patients. Lower extremity DVT was diagnosed in 49 patients (9.8%); 29 patients of these patients had DVT diagnosis before upper extremity ultrasound. Lower extremity DVT was diagnosed in 10 patients with UEDVT (7.0%). Acute PE was diagnosed in 13.0% of the cohort and occurred more often in patients with UEDVT (19.6% versus 10.3%; p=0.005). In the patients with UEDVT, PE was diagnosed a median of 5 days before upper extremity ultrasound (interquartile range: -23, 3). Patients with thrombosis involving two or more veins had significantly higher rate of PE (15 out of 46 patients [32.6%]) compared to those with a single thrombosed deep vein (13 out of 97 [13.4%], p=0.007). PE occurred in 27 out of the 169 patients with UESVT (16.0%)—14.8% of the 115 patients with UESVT alone and 18.5% of the 54 patients with both UEDVT and UESVT (p=0.54). On multivariable logistic regression, previous VTE (OR, 4.18; 95% CI, 2.21–7.94) and UEDVT (OR, 2.08; 95% CI, 1.16–3.71) were significantly associated with acute PE (Table 4).

Table 3.

Outcomes of patients

	All patients N=501	No upper extremity DVT N=358	Upper extremity DVT N=143	P-value
Lower extremity DVT, n (%)	49 (9.8)	39 (10.9)	10 (7.0)	0.18
Pulmonary embolism, n (%)	65 (13.0)	37 (10.3)	28 (19.6)	0.005
Hospital death, n (%)	187 (37.3)	113 (31.6)	74 (51.7)	<0.0001
Stay in ICU (days), median (Q1, Q3)	20.0 (10.5, 35.0)	19.0 (9.8, 33.0)	24.0 (12.0, 37.0)	0.03**
Stay in hospital (days), median (Q1, Q3)	48.0 (25.3, 94.0)	48.0 (25.0, 95.5)	46.0 (26.0, 93.0)	0.89**

DVT: deep-vein thrombosis, ICU: intensive care unit, Q1: first quartile, Q3: third quartile.

**Data are significantly skewed. The p-value is based on Mann-Whiteny U test.

Table 4.

Results of the multivariable logistic regression analysis for the predictors of acute pulmonary embolism and hospital mortality

	Acute pulmonary embolism			Hospital mortality
Variable	Odds ratio	95% confidence interval	P-value	Odds ratio	95% confidence interval	P-value
Age per 1-yr increment	0.989	0.972-1.006	0.20	1.042	1.026-1.059	<0.0001
Obesity	1.383	0.775-2.469	0.27	1.553	0.948-2.545	0.08
Diabetes	0.649	0.320-1.318	0.23	0.713	0.396-1.284	0.26
hypertension	0.572	0.249-1.311	0.19	0.589	0.292-1.191	0.14
Heart failure	0.727	0.333-1.587	0.42	1.929	1.091-3.410	0.02
Chronic respiratory disease	0.728	0.216-2.448	0.61	0.587	0.219-1.571	0.29
Chronic kidney disease	1.263	0.610-2.614	0.53	1.263	0.724-2.201	0.41
Cancer	1.000	0.485-2.061	1.00	2.408	1.313-4.416	0.005
Previous venous thromboembolism	4.184	2.206-7.936	<0.0001	0.573	0.300-1.095	0.09
Vasopressor therapy	1.690	0.949-3.009	0.08	1.652	1.023-2.667	0.04
Deep-vein thrombosis of the lower extremity	0.689	0.271-1.748	0.43	0.836	0.517-1.352	0.47
Deep-vein thrombosis of the upper extremity	2.080	1.164-3.714	0.01	2.046	1.215-3.447	0.007
Superficial-vein thrombosis of the upper extremity	1.460	0.827-2.576	0.19	1.080	0.660-1.767	0.76

Variables entered in each model were age, obesity, diabetes (body mass index ≥ 30 kg/m²), hypertension, heart failure, chronic respiratory disease, chronic kidney disease, cancer, previous venous thromboembolism, vasopressor therapy, deep vein thrombosis of the lower extremity, deep-vein thrombosis of the upper extremity and superficial-vein thrombosis of the upper extremity.

Model performance (acute pulmonary embolism): The model was statistically significant compared with the null model (chi-square value = 49.5, p<0.001) and had a correct classification rate of 87.6%. The Hosmer and Lemeshow goodness-of-fit test was acceptable (p=0.66). Nagelkerke R² = 0.175.

Model performance (hospital mortality): The model was statistically significant compared with the null model (chi-square value = 63.2, p<0.001) and had a correct classification rate of 70.8%. The Hosmer and Lemeshow goodness-of-fit test was acceptable (p=0.90). Nagelkerke R² = 0.218.

Hospital mortality was notably higher in patients with UEDVT, with over half (51.7%) dying during hospitalization compared with 31.6% in the patients without UEDVT. The median time to death was 21 days (interquartile range: 9, 41) after the ultrasound. On multivariable logistic regression analysis, UEDVT was significantly associated with hospital mortality (OR, 2.05; 95% CI, 1.22–3.45) (Table 4). The other predictors of hospital mortality were age (OR per one-year increment, 1.04; 95% CI, 1.03–1.06), cancer (OR, 2.41; 95% CI, 1.31–4.42), heart failure (OR, 1.93; 95% CI, 1.09–3.41) and vasopressor therapy (OR, 1.65; 95% CI, 1.02–2.67). UESVT was not associated with mortality.

ICU length of stay was longer in patients with UEDVT (median 24 days versus 19 days, p=0.03), but the total hospital length of stay was similar.

Discussion

This study found that among 501 critically patients who had Doppler ultrasound for suspected upper extremity vein thrombosis, 143 (28.5%) had UEDVT and 169 (33.7%) UESVT. The presence of permanent hemodialysis catheter in the internal jugular vein was significantly associated with UEDVT. PE occurred more often in the UEDVT group (19.6% versus 10.3%), and involvement of multiple deep veins was associated with even higher PE rates. UEDVT was associated with an almost two-fold increase in the risk of PE and hospital mortality on multivariable analysis.

UEDVT is classified into two main categories: primary and secondary.⁶ The primary form, also known as Paget Schroetter Syndrome, is a manifestation of venous thoracic outlet syndrome.⁶ The secondary form is commonly diagnosed in patients with malignancies or, most commonly, central venous catheters and applies to patients who acquire DVT during hospitalization.⁶ In our study, patients diagnosed with UEDVT represented 1.5% of all admitted patients to the ICU. This is similar to the findings of a study where 2.2% of 3746 critically ill patients enrolled in a randomized controlled trial had non-leg venous thrombosis, 94.5% were UEDVT.⁴ These rates are lower than those reported in other studies likely due to differences in their methods, setting and patient characteristics.^7,8 For example, the study by Kaplan et al included 113 ICU patients with sepsis with upper extremity ultrasound performed if central venous lines were present.⁷ It found that 16 patients (14.2%) had UEDVT.⁷

In our study, upper extremity ultrasound was ordered by the ICU team for clinical suspicion of acute DVT; hence patients were symptomatic. UEDVT is usually diagnosed by Doppler ultrasound which has a sensitivity of 0.87 (95% CI, 0.73–0.94) and specificity of 0.85 (95% CI, 0.72-0.93) in a systematic review.¹⁶ We found that 51.5% of the 501 patients studied had either UEDVT or UESVT (17.8% DVT alone, 23.0% SVT alone and 10.8% both). The rate of DVT was similar to the findings of other studies. Among 3695 upper extremity Doppler ultrasound examinations, one study found that almost one third (27.0%) were positive for acute UEDVT.⁵

We found that permanent hemodialysis catheters were independent risk factors for UEDVT. Endothelial injury and venous stasis from such venous access devices are the key causes of thrombogenesis. Other studies found that central lines were associated with UEDVT.¹⁷ ICU patients with peripherally inserted central venous catheters had a VTE incidence of 10.6% in a meta-analysis, with a potentially higher VTE risk compared with central venous catheters.¹⁸ Hemodialysis catheters have larger diameters than most other central venous catheters, which likely contributes to more venous stasis. In our study, pharmacologic prophylaxis was surprisingly not protective against UEDVT. A notable proportion of patients (24.2%) were not receiving pharmacologic thromboprophylaxis at the time of ultrasound, which may be related to having a contraindication.

In the current study, patients with UEDVT had a higher rate of acute PE compared with those without UEDVT (19.6% versus 10.3%, respectively). Patients with two or more thrombosed deep veins had a higher rate of PE (32.6%), suggesting that greater thrombus load increases PE risk. We found that UEDVT was associated with almost two-fold increased risk of PE on multivariable logistic regression. Previous studies have shown variable PE rates in patients with UEDVT.^5,11,12 In a study in ICU patients, PE occurred in 14.9% of patients with non-leg DVT, which is comparable to our results.⁴ Multiple studies found that UEDVT was associated with a lower rate of PE compared with lower extremity DVT.^2,14

These findings highlight opportunities to improve UEDVT prevention strategies. These may include avoiding unnecessary jugular or subclavian catheters, ensuring timely removal of such catheters when no longer required, and minimizing mechanical trauma to veins during insertion, for example, through routine ultrasound guidance.¹⁹ The administration of vasopressors through peripheral intravenous catheters, when performed with appropriate precautions, has been shown in studies to be probably safe and effective.^20,21 Additionally, tailored pharmacologic prophylaxis for high-risk patients may offer further preventive benefit compared with standard dosing. These strategies warrant further investigation.

We also observed that 103 out of 143 patients (72.0%) with UEDVT received therapeutic anticoagulation, while the rest received either prophylactic dose of anticoagulants or no anticoagulants at all. In one study only 9 out of 67 patients (13.4%) with non-leg DVT received therapeutic anticoagulation within 3 days of diagnosis.⁴ In another study, 73 out of 200 patients (36%) with UEDVT were put on therapeutic anticoagulation.¹¹ As proximal UEDVT carries considerable risk of complications, therapeutic anticoagulation is usually provided with effectiveness extrapolated from studies of lower extremity DVT.^9,22 In our study, most patients who did not receive therapeutic anticoagulation for UEDVT had valid contraindications. However, no clear justification was identified in 11 patients. This finding may reflect the perception among some clinicians that UEDVT carries less clinical significance than lower extremity DVT and suggests the need for a better understanding of clinicians’ perceptions of UEDVT and clearer guideline recommendations for the management of UEDVT in critically ill patients.

We found that patients with UEDVT had a high mortality rate, reaching 51.7%. UEDVT was also associated with mortality on multivariable logistic analysis. Other studies have observed high mortality associated with UEDVT.^4,10,23 In patients with trauma, UEDVT was associated with a higher mortality rate (7.1% of 56 patients with UEDVT and 1.5% of 325 with no UEDVT, p=0.03).²³ Another study found that patients with UEDVT had significantly higher all-cause mortality than patients with lower extremity DVT (28.5% versus 4.9%, p=0.004).¹⁴ Another study found a hospital mortality of 28.4%.⁴ The high mortality observed in patients with UEDVT may be related to the severity of underlying diseases.^6,10

This retrospective cohort study has several strengths including the relatively large sample size, and studying critically ill patients, thus focusing on secondary UEDVT, most likely acquired during hospital stay. However, it also has notable limitations. The single-center design may limit the generalizability and external validity of the findings as patient mix, management, and ordering diagnostic imaging may differ across centers. Additionally, UEDVT rate among ICU may have been overestimated due to the applied eligibility criteria, including evaluating only patients with suspected UEDVT, introducing selection bias. Asymptotic UEDVT or symptomatic UEDVT in patients not considered candidates for anticoagulation may have gone undiagnosed. We did not assess severity of illness using established scores such as Acute Physiology and Chronic Health Evaluation II and Sequential Organ Failure Assessment. Because illness severity may influence the risks of UEDVT and mortality, the absence of these measures represents an important limitation of our analysis. We also did not examine other important factors that may affect the development of UEDVT or outcomes, including timing and duration of catheter placement, concurrent infections, or mechanical ventilation. While we evaluated central venous catheters present at the time of ultrasound, previously removed catheters were not considered. Additionally, our multivariable regression models had limited explanatory power, suggesting that unmeasured variables likely contributed to the observed associations. Therefore, the association between UEDVT and outcomes such as pulmonary embolism and mortality may have been influenced by residual confounding and should be interpreted as observational rather than causal.

Conclusions

This study underscores the clinical significance of UEDVT in critically ill patients. UEDVT was relatively common in our cohort with central venous catheters being an important risk factor. The protective effect of pharmacologic thromboprophylaxis against UEDVT also appeared to be limited. UEDVT also appeared to be associated with adverse outcomes such as PE and hospital mortality. These associations may partly reflect greater illness severity or other unmeasured confounders and therefore should be interpreted as observational rather than causal. Nonetheless, our findings may emphasize the importance of VTE prevention—particularly vigilant assessment of the need of central venous catheters and appropriate pharmacologic thromboprophylaxis, early detection and appropriate therapeutic interventions in the ICU setting. Our findings provide valuable observational insights that support the need for future prospective studies but are not sufficient on their own to support changes in clinical protocols.

Footnotes

Acknowledgment

The authors acknowledge the faculty of the Research Unit of the College of Medicine, King Saud bin Abdulaziz University for Health Sciences, for their support.

ORCID iD

Hasan M. Al-Dorzi

Ethical Considerations

The study was approved by the Institutional Review Board of the Ministry of National Guard Health Affairs, Riyadh, Saudi Arabia (Study number NRR24/050/7; IRB Ethical approval no: 000028424).

Consent to Participate

As the study was a retrospective with no direct contact with patients, informed consent was waived. This study was conducted in accordance with the guidelines of the Declaration of Helsinki (2000) and Good Clinical Practice E6 (R2).

Author Contributions

HMD contributed to the study concept and design, data collection, data interpretation and drafting the manuscript, and performed statistical analysis. RA, HM, RR, and SY contributed to the study concept and design, data collection, data interpretation and drafting the manuscript.

Funding

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

Declaration of Conflicting Interests

All authors declare that AI tools were not utilized in the writing of the manuscript, creation of images, collection and analysis of data. Authors also confirm that the manuscript was not published in any website or printed journal in any other language other than English.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.*

References

Joffe

Kucher

Tapson

Goldhaber

Deep Vein Thrombosis DVT FREE Steering Committee . Upper-extremity deep vein thrombosis: a prospective registry of 592 patients. Circulation. 2004;110(12):1605-1611.

Munoz

Mismetti

Poggio

, et al. Clinical outcome of patients with upper-extremity deep vein thrombosis: results from the RIETE Registry. Chest. 2008;133(1):143-148.

Espitia

Tissot

Miossec

, et al. Upper extremity venous thrombosis in hospitalized patients: a prospective epidemiological study. Thrombosis Research. 2024;233:174-180.

Lamontagne

McIntyre

Dodek

, et al. Nonleg venous thrombosis in critically ill adults: a nested prospective cohort study. JAMA Intern Med. 2014;174(5):689-696.

Cires-Drouet

Durham

Sharma

, et al. Prevalence and clinical outcomes of hospitalized patients with upper extremity deep vein thrombosis. J Vasc Surg Venous Lymphat Disord. 2022;10(1):102-110.

Heil

Miesbach

Vogl

Bechstein

Reinisch

. Deep Vein Thrombosis of the Upper Extremity. Dtsch Arztebl Int. 2017;114(14):244-249.

Kaplan

Casper

Elliott

, et al. VTE incidence and risk factors in patients with severe sepsis and septic shock. Chest. 2015;148(5):1224-1230.

Al-Dorzi

Al-Yami

Al-Daker

Alqirnas

Alhamadh

Khan

. The association of timing of pharmacological prophylaxis and venous thromboembolism in patients with moderate-to-severe traumatic brain injury: A retrospective cohort study. Ann Thorac Med. 2022;17(2):102-109.

Thiyagarajah

Ellingwood

Endres

, et al. Post-thrombotic syndrome and recurrent thromboembolism in patients with upper extremity deep vein thrombosis: A systematic review and meta-analysis. Thromb Res. 2019;174:34-39.

10.

Bleker

van Es

Kleinjan

, et al. Current management strategies and long-term clinical outcomes of upper extremity venous thrombosis. J Thromb Haemost. 2016;14(5):973-981.

11.

Levy

Bach

Fisher-Snowden

Pfeifer

. Upper extremity deep venous thrombosis: reassessing the risk for subsequent pulmonary embolism. Ann Vasc Surg. 2011;25(4):442-447.

12.

Levy

Albuquerque

Pfeifer

. Low incidence of pulmonary embolism associated with upper-extremity deep venous thrombosis. Ann Vasc Surg. 2012;26(7):964-972.

13.

Kucher

. Clinical practice. Deep-vein thrombosis of the upper extremities. N Engl J Med. 2011;364(9):861-869.

14.

Rokosh

Ranganath

Yau

, et al. High Prevalence and Mortality Associated with Upper Extremity Deep Venous Thrombosis in Hospitalized Patients at a Tertiary Care Center. Ann Vasc Surg. 2020;65:55-65.

15.

Arabi

Alshimemeri

Taher

. Weekend and weeknight admissions have the same outcome of weekday admissions to an intensive care unit with onsite intensivist coverage. Crit Care Med. 2006;34(3):605-611.

16.

Patel

Braun

Patel

, et al. Diagnosis of deep vein thrombosis of the upper extremity: a systematic review and meta-analysis of test accuracy. Blood Adv. 2020;4(11):2516-2522.

17.

Winters

Callas

Cushman

Repp

Zakai

. Central venous catheters and upper extremity deep vein thrombosis in medical inpatients: the Medical Inpatients and Thrombosis (MITH) Study. J Thromb Haemost. 2015;13(12):2155-2160.

18.

Puri

Dai

Giri

Huang

Zhao

. The incidence and risk of venous thromboembolism associated with peripherally inserted central venous catheters in hospitalized patients: A systematic review and meta-analysis. Front Cardiovasc Med. 2022;9:917572.

19.

Zawadka

La Via

Wong

, et al. Real-time ultrasound guidance as compared with landmark technique for subclavian central venous cannulation: a systematic review and meta-analysis with trial sequential analysis. Critical Care Medicine. 2023;51(5):642-652.

20.

Tian

Smyth

Keijzers

, et al. Safety of peripheral administration of vasopressor medications: a systematic review. Emergency Medicine Australasia. 2020;32(2):220-227.

21.

Aykanat

Myles

Weinberg

Burrell

Bellomo

. Low-concentration norepinephrine infusion for major surgery: a safety and feasibility pilot randomized controlled trial. Anesthesia & Analgesia. 2022;134(2):410-418.

22.

Valeriani

Di Nisio

Porceddu

, et al. Anticoagulant treatment for upper extremity deep vein thrombosis: A systematic review and meta-analysis. J Thromb Haemost. 2022;20(3):661-670.

23.

Walusimbi

Dossa

Faryar

Markert

McCarthy

. Upper Extremity Deep Venous Thrombosis Risk Factors, Associated Morbidity and Mortality in Trauma Patients. World J Surg. 2022;46(3):561-567.