Retrospective Study of Bleeding Risk with Concomitant Vascular Endothelial Growth Factor Receptor Tyrosine Kinase Inhibitor and Anticoagulation

Abstract

Background

Patients with cancer are at higher risk for venous thromboembolism (VTE) and bleeding, in turn complicating anticoagulant therapy. An added complexity is the toxicity profile of agents used to treat certain cancers, namely the vascular endothelial growth factor receptor tyrosine kinase inhibitors (TKIs), which are associated with both thromboembolism and hemorrhages. The purpose of this study was to evaluate whether patients taking concurrent VEGF TKI and therapeutic anticoagulant were at higher risk for bleeding compared with patients taking VEGF TKI alone.

Materials and Methods

This was a single-center, retrospective chart review of patients who underwent treatment with a VEGF TKI with or without anticoagulant. The primary outcome included comparison of major bleeding rates between groups. Secondary outcomes included comparison of composite major and minor bleed rates and assessment of VTE incidence and recurrence among patients treated with a VEGF TKI and VEGF TKI plus anticoagulant, respectively.

Results

A total of 184 and 74 patients were included in the VEGF TKI alone and VEGF TKI + anticoagulant groups, respectively. Major bleeding events occurred in 6 of 184 patients (3.3%) and 6 of 74 patients (8.1%), respectively (p = .095). Composite major and minor bleeding events occurred in 22 of 184 (13.6%) and 17 of 74 (23%), respectively (p = .026). A total of 26 of 258 patients (10.1%) experienced a VTE event while taking a VEGF TKI, and 1 of 26 (3.8%) experienced a recurrent VTE event while taking a VEGF TKI plus anticoagulant.

Conclusion

Patients who received concomitant VEGF TKI plus anticoagulant had increased incidence of bleeding, although prospective studies are needed to further explore this association.

Implications for Practice

The current study showed that use of concomitant vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGF TKI) and therapeutic anticoagulation was associated with an increased risk of composite bleeding events, although at comparable rates between patients treated with VEGF TKI plus direct oral anticoagulant (DOAC) versus VEGF TKI plus non-DOAC anticoagulant. This suggests that when anticoagulation is indicated in a patient receiving a VEGF TKI, the novel DOACs may be a safe alternative to historical anticoagulants (warfarin and enoxaparin). These results fill a gap in the literature and will help guide treatment decisions for patients requiring concurrent VEGF TKI and anticoagulation.

Introduction

In the general population, venous thromboembolism (VTE), including deep-vein thrombosis and pulmonary embolism, is estimated to occur in 1–2 people per 1,000. However, incidence of VTE among the oncology patient population is significantly higher, with some estimates suggesting nearly a fivefold increased risk [1]. This is thought to be due to induction of a hypercoagulable state by factors such as prothrombotic tumors, multiple comorbidities, and patient-specific risk factors such as demographics [1, 2]. On the contrary, it is estimated that approximately 10% of all patients with advanced cancer have at least one bleeding episode throughout their lifetime. Factors such as local tumor invasion, abnormal tumor vasculature, or tumor regression contribute to this increased risk of spontaneous bleeding [3].

Given the paradoxical risks of both clotting and bleeding in these patients, treatment of comorbid conditions requiring anticoagulation such as VTE or stroke prevention in atrial fibrillation requires careful risk-benefit assessment. A recent study evaluated incidence of bleeding among patients with cancer treated with anticoagulation and reported higher incidence compared with patients without cancer regardless of anticoagulant used [4].

An added complexity to anticoagulant therapy in patients with cancer is the possibility of drug–drug interactions due to concomitant medications. One such example is with bevacizumab, a monoclonal antibody that targets the vascular endothelial growth factor (VEGF) to inhibit tumor angiogenesis. One study evaluated concomitant use of anticoagulation for VTE treatment and bevacizumab in patients with glioma and reported higher rates of intracranial hemorrhage of any grade and serious hemorrhages compared with patients who received bevacizumab monotherapy [5]. As such, the concurrent use of bevacizumab with anticoagulants may present an additive bleed risk due to disruption of the VEGF pathway [6, 7]. VEGF exerts protective effects on endothelial cells and plays a key role in endothelial cell regulation [8]. Interference of this pathway has been implicated in vascular complications, including both thrombosis and hemorrhage [7].

Another class of medications used in solid tumor malignancies that target the VEGF pathway are the VEGF tyrosine kinase inhibitors (TKIs). As with bevacizumab, the VEGF TKIs carry a risk for hemorrhagic adverse effects [9]. A recent meta-analysis reported all-grade bleeding and serious bleeding-related adverse event rates of up to 18% and 7%, respectively, with sunitinib [10]. The study also found that axitinib had the lowest reported incidence of serious bleeding-related adverse events with a 1%–4% occurrence rate. Other VEGF TKIs such as cabozantanib, pazopanib, and sorafenib ranked somewhere in between, with all-grade bleeding and serious bleeding-related adverse event rates ranging from 4% to 15% and 2% to 4%, respectively.

Given that both bevacizumab and the VEGF TKIs interfere with signal transduction of the VEGF pathway and documented evidence of additive bleed risk with concurrent bevacizumab and anticoagulant therapy, we hypothesize that a similar correlation exists with concurrent VEGF TKI and anticoagulant therapy. The purpose of this study was to assess whether patients who received concomitant VEGF TKI and anticoagulation were at higher risk for bleeding events compared with those who received a VEGF TKI alone.

Materials and Methods

Study Design

This was a single-center, institutional review board–approved retrospective chart review of patients who underwent cancer treatment with a VEGF TKI at Winship Cancer Institute of Emory University between December 1, 2005, and December 31, 2019. An electronic database query of all patients with a VEGF TKI on their medication profile was initially performed. Individual patients were then reviewed through Emory’s electronic medical record (EeMR) for study inclusion.

Inclusion and Exclusion Criteria

Patients were divided into two groups: (a) VEGF TKI alone and (b) VEGF TKI plus therapeutic anticoagulation. Patients were included in group (a) if they were 18 years of age or older and prescribed a VEGF TKI for cancer treatment. Patients were included in group (b) if they were 18 years of age or older, prescribed a VEGF TKI for cancer treatment, and were concomitantly receiving therapeutic anticoagulation. Patients were excluded if they were prescribed anticoagulation for VTE prophylaxis (as documented in clinical notes), they received <1 month of concomitant VEGF TKI and anticoagulation, or information was not readily retrievable from EeMR because of transfer of care or death.

Primary, Secondary, and Exploratory Outcomes

The primary outcome was to compare rates of major bleeding between patients treated with VEGF TKI plus therapeutic anticoagulation compared with those treated with a VEGF TKI alone. Major bleeding was defined as a bleed from a critical site (intracranial, intraspinal, intraocular, retroperitoneal, pericardial, or gastrointestinal), bleed associated with a hemoglobin reduction of ≥2 g/dL, bleed resulting in transfusion of ≥2 units of blood, or a bleed contributing to death [11].

Secondary outcomes were to compare composite rates of major and minor bleeding between the two groups and to quantify VTE incidence rates among all patients who received a VEGF TKI and recurrence rates among all patients who received a VEGF TKI plus anticoagulant. Minor bleeding was defined as a bleed that did not meet the criteria for a major bleed but prompted notification to a provider during a clinic visit or via telephone call.

As an exploratory outcome, minor bleeding rates between groups were also compared.

Statistical Analyses

SAS Software version 9.4 (SAS Institute, Inc., Cary, NC) was used for all data analyses. The significance level (α) was set at .05. Baseline characteristics between the two groups (VEGF TKI alone versus VEGF TKI plus anticoagulant) were compared using χ² or Fisher’s exact tests for categorical covariates and the ANOVA or Kruskal-Wallis tests for numerical covariates. For the primary outcome, major bleeding rates between groups were compared using Fisher’s exact test. Multivariable logistic regression was performed for the primary outcome and secondary outcome of composite major and minor bleeding rates between groups using backward variable elimination strategy with a stay criteria of p < .20. Firth’s bias-reduced, penalized likelihood approach was employed to account rare events [12]. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to evaluate the strengths of observed associations. Two sample t tests were used to compare the composite rates of major and minor bleeding between groups.

Results

Patients

Among 358 patients prescribed one or more VEGF TKI ± anticoagulant(s) within the previously specified study time frame, 258 patients met inclusion criteria. Reasons for exclusion are shown in Figure 1. Although there were statistically significant differences in cancer diagnosis, cancer treatment approach, and VEGF TKI agent, baseline characteristics were otherwise similar between groups. The majority of patients were male, were diagnosed with renal cell carcinoma (RCC), and received multiple lines and/or combinations of treatment. Additional characteristics are depicted in Table 1.

Figure 1

Consort diagram. This is a diagram that depicts the total number of patients included in each group (VEGF TKI alone versus VEGF TKI plus anticoagulant) as well as those excluded from the study based on exclusion criteria. Abbreviations: AC, anticoagulant; TKI, tyrosine kinase inhibitor; VEGF, vascular endothelial growth factor; VTE, venous thromboembolism.

Table 1. Baseline characteristics

Characteristics	VEGF TKI alone (n = 184)	VEGF TKI + AC (n = 74)	p value
Male, n (%)	132 (71.7)	48 (64.9)	.277
Age at cancer Dx, yr, mean ± SD	58.9 ± 12	59.3 ± 13.3	.808
Cancer Dx, n (%)			<.001
RCC	183 (99.5)	40 (54.1)
Sarcomas	0 (0)	13 (17.6)
HCC	0 (0)	10 (13.5)
Thyroid carcinoma	0 (0)	7 (9.5)
Neuroendocrine tumor	1 (0.5)	2 (2.7)
Adrenal cortical carcinoma	0 (0)	1 (1.4)
GIST	0 (0)	1 (1.4)
Comorbid conditions, n (%)			.636
HTN	128 (69.6)	47 (63.5)
DMII	38 (20.7)	17 (23)
Aneurysm	1 (0.54)	0 (0)
Other	6 (3.3)	1 (1.4)
None	43 (23.4)	21 (28.4)
ABW at VEGF TKI initiation, mean ± SD (n), kg	88.3 ± 23.1 (159)	86.8 ± 22.4 (63)	.664
Ht at VEGF TKI initiation, mean ± SD (n), cm	173.4 ± 12 (157)	172.1 ± 17.3 (62)	.508
CrCl at VEGF TKI initiation, mean ± SD (n), mL/min	67.5 ± 25.5 (154)	74 ± 32.2 (60)	.308
AST at VEGF TKI initiation, mean ± SD (n)	22.1 ± 14.5 (153)	29.3 ± 25.4 (60)	.179
ALT at VEGF TKI initiation, mean ± SD (n)	22.9 ± 26.4 (153)	28.7 ± 42.5 (60)	.664
Cancer treatment approach, n (%)			.006
VEGF TKI, then other	74 (40.2)	20 (27)
VEGF TKI alone	40 (21.7)	17 (23)
VEGF TKI, then immunotherapy	29 (15.8)	12 (16.2)
VEGF TKI + immunotherapy	11 (6)	1 (1.4)
Immunotherapy, then VEGF TKI	7 (3.8)	1 (1.4)
Other	23 (12.5)	23 (31.1)
VEGF TKI agent, n (%)			<0.001
Pazopanib	51 (27.7)	18 (24.3)
Cabozantanib	29 (15.8)	13 (17.6)
Sunitinib	21 (11.4)	6 (8.1)
Axitinib	11 (6)	3 (4.1)
Lenvatinib	2 (1.1)	4 (5.4)
Sorafenib	0 (0)	10 (13.5)
Multiple	70 (38.0)	20 (27)
Anticoagulant agent (n, %)	N/A		N/A^a^a Statistical analysis was not performed.
Apixaban		21 (28.4)
Rivaroxaban		20 (27)
Enoxaparin		16 (21.6)
Warfarin		8 (10.8)
Dabigatran		1 (1.4)
Multiple		8 (10.8)
Anticoagulant indication (n, %)	N/A		N/A^a^a Statistical analysis was not performed.
DVT		19 (25.7)
AF		19 (25.7)
PE		17 (23)
>1 of the above		9 (12.2)
Other		10 (13.5)

• ^a Statistical analysis was not performed.
• Abbreviations: ABW, actual body weight; AC, anticoagulant; AF, atrial fibrillation; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CrCl, creatinine clearance calculated using Cockcroft-Gault formula; DMII, diabetes mellitus type II; DVT, deep-vein thrombosis; Dx, diagnosis; GIST, gastrointestinal stromal tumor; HCC, hepatocellular carcinoma; Ht, height; HTN, hypertension; N/A, not applicable; PE, pulmonary embolism; RCC, renal cell carcinoma; TKI, tyrosine kinase inhibitor; VEGF, vascular endothelial growth factor.

Major Bleeding

Over a median follow-up duration of 4.4 months and 10.6 months in the VEGF TKI plus anticoagulant and VEGF TKI alone groups, respectively, major bleeding events occurred in 6 of 74 patients (8.1%) in the combination group and 6 of 184 patients (3.3%) in the monotherapy group (p = .095). Univariate (OR, 1.43; 95% CI, 0.38–5.35; p = .595) and multivariable (OR, 1.65; 95% CI, 0.44–6.64; p = 0.478) logistic regression analyses confirmed lack of a statistically significant difference between the combination versus monotherapy groups with respect to major bleeding events. Gastrointestinal bleeding was the most common site of major bleeding in both groups (Table 2). Table 3 shows the impact of tumor type (RCC vs. other) on major bleed incidence. A greater proportion of patients with a non-RCC diagnosis (hepatocellular carcinoma and sarcoma) had major bleeding events, although this was not found to be statistically significant and was likely due to small sample sizes.

Table 2. Major bleeding event types

Treatment group	Major bleeding event (n = 12)			Total
	Intracranial	Gastrointestinal	Perinephric hematoma
VEGF TKI alone, n (%)	2 (33.3)	4 (66.7)	0 (0)	6
VEGF TKI + AC, n (%)	2 (33.3)	3 (50)	1 (16.7)	6
Total, n (%)	4 (33.3)	7 (58.3)	1 (8.3)	12

• Abbreviations: AC, anticoagulant; TKI, tyrosine kinase inhibitor; VEGF, vascular endothelial growth factor.

Table 3. Impact of tumor type on major bleeding incidence

Major bleed (n = 12)	Tumor type			p value
	RCC (n = 223)	Other (n = 23)^a^a Hepatocellular carcinoma and sarcoma.	OR (95% CI)
Yes, n (%)	8 (3.6)	4 (17.4)	1.57 (0.36–6.89)	.552
No, n (%)	215 (96.4)	19 (82.6)

• ^a Hepatocellular carcinoma and sarcoma.
• Abbreviations: CI, confidence interval; OR, odds ratio; RCC, renal cell carcinoma.

Univariate and multivariable logistic regression analyses were performed to assess impact on the outcome of major bleeding of baseline characteristics, which differed between the two study groups. VEGF TKI agent (single vs. multiple), cancer diagnosis (RCC vs. other), and cancer treatment (VEGF TKI as first line of therapy vs. VEGF TKI plus immunotherapy as first line of therapy vs. other treatment as first line of therapy) were the baseline characteristics that were assessed using univariate and multivariable logistic regression analyses. On univariate analysis, cancer treatment was nonstatistically significant as evidenced by a p = .331 (other treatment vs. VEGF TKI as first line of therapy [OR, 10.86; 95% CI, 0.46–257.63; p = .140]; VEGF TKI plus immunotherapy vs. VEGF TKI as first line of therapy [OR, 8.08; 95% CI, 0.28–230.96; p = .222]) and VEGF TKI agent (multiple versus single) was also nonstatistically significant (OR, 3.01; 95% CI, 0.77–11.74; p = 0.113). Cancer diagnosis and cancer treatment were removed from the multivariable logistic regression model by backward selection method. However, VEGF TKI agent was included in the model and confirmed a nonstatistically significant difference between multiple versus single VEGF TKI agent (OR, 3.13; 95% CI, 0.85–12.77; p = .106).

The median time to first major bleed after VEGF TKI initiation was 217.5 days (interquartile range [IQR], 69.25–304.25) and 116.5 days (IQR, 43–159.25) in the VEGF TKI alone and VEGF TKI plus anticoagulant groups, respectively.

Of the 12 patients who suffered a major bleed in this study, 3 of 12 patients (25%) permanently discontinued treatment and 2 of 12 patients (16.7%) temporarily discontinued treatment (limited documentation precluded determination of permanent discontinuation). The remaining 7 of 12 patients (58.3%) continued VEGF TKI therapy after bleed resolution with seemingly no further bleeding complications.

Minor Bleeding

Over a median follow-up duration of 4.4 months and 10.6 months in the VEGF TKI plus anticoagulant and VEGF TKI alone groups, respectively, minor bleeding events occurred in 16 of 184 patients (8.7%) in the combination group and 11 of 74 patients (14.9%) in the monotherapy group (p = .143). Epistaxis was the most common type of minor bleeding in the VEGF TKI alone group, and gingival hemorrhage was the most common type of minor bleeding in the VEGF TKI plus anticoagulant group (supplemental online Table 1). Supplemental online Table 2 shows the impact of tumor type on minor bleed incidence. As shown, there was a statistically significant difference in minor bleed incidence (p = .01) based on tumor type.

The median time to first minor bleed after VEGF TKI initiation was 250 days (IQR, 57.75–326.75) and 118.5 days (IQR, 48.25–169.75) in the VEGF TKI alone and VEGF TKI plus anticoagulant groups, respectively. Of the 27 patients who experienced a minor bleed while taking a VEGF TKI, 3 of 27 (11.1%) discontinued VEGF TKI treatment (2 patients were switched to alternate therapies and 1 patient was transitioned to hospice care). Of the remaining 24 patients with a minor bleed, 14 of 24 (58.3%) continued VEGF TKI treatment with seemingly no further bleeding complications, 9 of 24 (37.5%) continued VEGF TKI treatment with 1 or more minor bleed recurrences, and 1 of 24 (4.2%) did not have sufficient documentation to determine bleed recurrence.

Composite Major and Minor Bleeding

Over a median follow-up duration of 4.4 months and 10.6 months in the VEGF TKI plus anticoagulant and VEGF TKI alone groups, respectively, the composite endpoint of major and minor bleeding events occurred in 22 of 184 patients (13.6%) in the combination group and 17 of 74 patients (23%) in the monotherapy group (p = .026). Univariate (OR, 2.54; 95% CI, 1.28–5.02; p = .007) and multivariable (OR, 2.73; 95% CI, 1.36–5.46; p = .005) logistic regression analyses confirmed higher overall bleeding incidence in the combination group.

On univariate analysis, cancer diagnosis (other vs. RCC) was statistically significant (OR, 2.52; 95% CI, 1.11–5.72; p = .028), cancer treatment was nonstatistically significant as evidenced by a p = 0.657 (other treatment vs. VEGF TKI as first line of therapy [OR, 1.49; 95% CI, 0.61–3.60; p = 0.3800]; VEGF TKI plus immunotherapy vs. VEGF TKI as first line of therapy [OR, 1.22; 95% CI, 0.43–3.45; p = 0.710]), and VEGF TKI agent (multiple versus single) was also nonstatistically significant (OR, 1.61; 95% CI, 0.82–3.17; p = .165). Cancer diagnosis and cancer treatment were removed from the multivariable logistic regression model by backward selection approach. However, VEGF TKI agent was included in the model and confirmed statistical insignificance between multiple versus single VEGF TKI agent (OR, 1.82; 95% CI, 0.91–3.64; p = .09).

VTE Incidence and Recurrence

Among all patients treated with a VEGF TKI in this study, 26 of 258 (10.1%) experienced a VTE event. Twenty of 26 (76.9%) of these events occurred when patients were taking a VEGF TKI alone, whereas 6 of 26 (23.1%) events occurred when patients were taking a VEGF TKI plus anticoagulant.

Among the 26 patients who experienced a VTE event, 1 of 26 (3.8%) experienced a recurrent VTE event while receiving a VEGF TKI plus anticoagulant. Median follow-up duration for thrombotic events among patients who received VEGF TKI plus anticoagulant versus VEGF TKI alone were 7 months and 10.6 months, respectively.

Concomitant Use of Medications Associated with Bleed Risk

To acknowledge the potential impact of concomitant medications on bleeding outcomes, use of medications known to increase bleed risk (P2Y12 inhibitors, selective serotonin reuptake inhibitor or serotonin-norepinephrine reuptake inhibitor, acetylsalicylic acid [ASA] 81 mg vs. 325 mg, non-ASA nonsteroidal anti-inflammatory drugs, fish oil, herbal products, and/or proton pump inhibitors) was collected. Figure 2 and supplemental online Figure 1 depicts concomitant use of these medications at the time of major and minor bleeds, respectively. Of note, at the time of major bleed, 2 of 12 patients (16.7%) were not taking any of the above medications and information was unknown for 1 of 12 patients (8.3%). At the time of minor bleed, 14 of 27 patients (51.9%) were not taking any of the above medications.

Figure 2

Concomitant medication use at the time of major bleed. This is a pie chart that shows concomitant medications associated with bleed risk that patients were taking at the time of a major bleeding event. The most common of these was a PPI. Abbreviations: ASA, acetylsalicylic acid; PPI, proton pump inhibitor; SNRI, serotonin-norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor.

Bleed Occurrence Based on Anticoagulant

As shown in Table 1, patients in the VEGF TKI plus anticoagulant group were treated with a variety of anticoagulants, although enoxaparin and direct oral anticoagulants (DOACs), including apixaban, rivaroxaban, and dabigatran, were more common than warfarin. A breakdown of major and minor bleeding events based on type of anticoagulant is shown in Figure 3. Bleeding rates based on anticoagulant (DOAC versus non-DOAC) are summarized in supplemental online Table 3.

Figure 3

Bleeding event based on anticoagulant. This is a bar graph that illustrates major vs. minor bleeding event incidence based on anticoagulant. Major and minor bleeding events were most common in patients receiving enoxaparin and rivaroxaban, respectively.

Bleed Occurrence Based on VEGF TKI

As shown in Table 1, patients received a variety of VEGF TKIs for oncologic treatment, although cabozantanib and pazopanib were the most common. Rates of major and minor bleeding events based on type of VEGF TKI are shown in Figure 4.

Figure 4

Bleeding event based on VEGF TKI. This is a bar graph that illustrates major vs. minor bleeding event incidence based on VEGF TKI. Major and minor bleeding events were most common in patients receiving cabozantanib and pazopanib, respectively. TKI, tyrosine kinase inhibitor; VEGF, vascular endothelial growth factor.

Discussion

This study found that patients who received VEGF TKI plus anticoagulant had a numerically, although nonsignificant, higher incidence of major bleeding but a significantly higher incidence of composite major and minor bleeding compared with those who received VEGF TKI alone. Based on existing data, the average rate of serious bleeding events (all-grade central nervous system and gastrointestinal hemorrhages—components of major bleeding as defined in our study) with VEGF TKIs approximates 4% [9]. The average rate of major bleeding events with low-molecular-weight heparin (LMWH), warfarin, and DOACs approximates 4%–7% depending on the agent [13-15]. In the current study, combination treatment with a VEGF TKI plus anticoagulant resulted in an 8.1% incidence of major bleeding events, suggestive of an additive bleed risk. However, it should be noted that a large percentage of patients in this study received anticoagulation for atrial fibrillation rather than for VTE, and this could have contributed to the differences observed in major bleeding rates compared with the aforementioned historical rates.

In this study, among all patients treated with a VEGF TKI, approximately 10% experienced a VTE event. In 2012, Qi and colleagues observed a VTE incidence rate of 3% with VEGF TKIs approved at the time (pazopanib, sunitinib, sorafenib, and vandetanib) [16]. The higher VTE incidence rate observed in our study may be due to a smaller overall sample size and inclusion of patients treated with other VEGF TKIs (axitinib, lenvatinib, and cabozantanib) as well. Additionally, among the patients who had a VTE event, 3.8% experienced a recurrent VTE while taking a VEGF TKI plus therapeutic anticoagulant. Based on previous studies, the average rate of recurrent VTE in oncology patients treated with therapeutic anticoagulation ranges from 4%-11% depending on the anticoagulant agent [13-15]. However, it should be noted that a large percentage of patients in this study received anticoagulation for atrial fibrillation rather than for VTE, and this could have contributed to the differences observed in recurrent VTE rates compared with the aforementioned historical rates.

This retrospective study provides valuable information regarding incidence of bleeding in patients who received VEGF TKI plus anticoagulant versus VEGF TKI alone and adds to the limited data in the literature on this topic. A recently published retrospective review evaluated incidence of clinically significant bleeding events among patients treated with a VEGF TKI plus factor Xa inhibitor (DOAC or LMWH) during the concurrent treatment period as well as factor Xa inhibitor alone treatment period [17]. Results revealed a greater than twofold higher risk of bleeding events during the concurrent treatment period over a median follow-up of 63 days. Furthermore, a 6-month subgroup analysis exhibited a similar trend for patients who received concurrent VEGF TKI plus LMWH, although a small sample size limited interpretation for those who received concurrent VEGF TKI plus DOAC. Our study validates these results by demonstrating higher overall bleeding incidence in the VEGF TKI plus anticoagulant group and adds to the literature the impact of anticoagulant agent (DOAC vs. other) on bleeding outcomes over a longer median follow-up duration. As shown in supplemental online Table 3, a comparison of composite bleeding event rates between patients treated with VEGF TKI plus DOAC versus VEGF TKI plus non-DOAC anticoagulant did not achieve statistical significance, suggesting that if concomitant VEGF TKI and anticoagulant therapy is warranted, the novel DOAC agents are a reasonable alternative to historical anticoagulants (warfarin and enoxaparin) from a safety perspective.

There are several limitations to this study. One limitation includes the possibility of confounders because of statistically significant differences between groups in certain baseline characteristics, including cancer diagnosis, cancer treatment approach, and VEGF TKI agent. To characterize this further, multivariable Firth logistic regression analyses were performed for major bleeding and composite major and minor bleeding outcomes using these covariates. VEGF TKI agent was found to be nonstatistically significant for both outcomes, and cancer type and treatment approach were removed from both multivariable regression analyses by backward selection approach because of p ≥ .20. Of note, univariate logistic regression of cancer type revealed a statistically significant increase in composite bleeding events in patients with a non-RCC diagnosis, although this was likely due to small sample size. Another limitation is that some patients were initiated on a VEGF TKI and/or anticoagulant prior to transfer of care to our institution, limiting our ability to evaluate for bleeding and thrombotic complications prior to care transfer and thereby impacting our assessment of median follow-up duration while on VEGF TKI and/or anticoagulant therapies. An additional limitation is that the majority of patients included received a VEGF TKI for treatment of RCC, whereas much smaller numbers of patients who received a VEGF TKI for other cancer diagnoses were included. Initially, the target population for this study consisted of patients who received a VEGF TKI for treatment of RCC. However, the sample size for patients with RCC who received concomitant VEGF TKI plus anticoagulant was small, and the data pool was thus expanded to include patients with all cancer types to allow for inclusion of additional patients in the combination group. Thus, the impact of tumor type on bleeding incidence cannot be excluded (as confirmed by Table 3 and supplemental online Table 2). In addition, several patients who experienced a bleeding event while on a VEGF TKI with or without an anticoagulant were also taking other medications associated with bleed risk, a possible confounder to the observed bleeding events. However, more than half of patients who experienced a minor bleed were not receiving any of these additional medications, strengthening the possibility that bleeding was associated with either the VEGF TKI and/or anticoagulant. Additionally, there was a component of subjectivity in determination of minor bleeding events, as some patients who were documented as having experienced a minor bleed had the same type of bleed present at baseline. However, in many cases, the preexisting bleed was further exacerbated by the addition of a VEGF TKI and/or anticoagulant. Finally, the possibility of reporting bias cannot be excluded and is an inherent limitation of a retrospective chart review.

Overall, this study observed a higher incidence of composite major and minor bleeding events with use of concomitant VEGF TKI and anticoagulant. Although this may be an expected finding given inherent risks of bleeding with either agent alone, this study quantifies the additive risk. In addition, patients who received combination therapy appeared to have a faster time to first major or minor bleed (approximately 120 days). However, neither major nor minor bleeds were statistically significant when evaluated as individual endpoints, and more than half of patients who experienced major or minor bleeds continued VEGF TKI ± anticoagulant therapy with seemingly no further complications. Therefore, concurrent use of VEGF TKI and anticoagulant therapy, when clinically indicated, is reasonable, and use of a DOAC is comparable to a non-DOAC in this scenario. However, close monitoring of signs and symptoms of bleeding is recommended for the first few months after combination therapy initiation. In addition, careful consideration of patient-specific factors and thrombotic versus hemorrhagic risk is recommended to optimize clinical outcomes while minimizing adverse effects when combination therapy is required.

We believe our findings are important and applicable to clinical practice. Our study included patients treated with newer VEGF TKI agents such as cabozantanib and explored outcomes among patients who received concurrent DOACs and VEGF TKI, a practice that has limited existing data to support its use. In addition, our study reflects important real-world data through inclusion of patients with multiple medical comorbidities, and there are likely to be differences in patient populations in clinical trial versus standard use settings. As such, our findings are relevant and hypothesis generating for future prospective studies.

Conclusion

This single-center, retrospective chart review found that patients who received concomitant VEGF TKI plus anticoagulant had increased incidence of bleeding compared with patients who received VEGF TKI alone when major and minor bleeding were evaluated as a composite endpoint. However, independent risk of major and minor bleeds was not significantly increased with use of combination therapy. In addition, when combination therapy is indicated, a DOAC is a reasonable alternative to a non-DOAC. Prospective studies to further explore these associations are necessary to establish definite treatment recommendations.

Acknowledgments

Research reported in this publication was supported in part by the Biostatistics Shared Resource of Winship Cancer Institute of Emory University and National Institutes of Health/National Cancer Institute under award number P30CA138292. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author Contributions

Conception/design: Sarah E. Caulfield, Mehmet A. Bilen, Tyler J. Beardslee

Provision of study material or patients: Sarah E. Caulfield, Mehmet A. Bilen, Tyler J. Beardslee

Collection and/or assembly of data: Akhilesh Sivakumar

Data analysis and interpretation: Chao Zhang, Subir Goyal

Manuscript writing: Akhilesh Sivakumar, Sarah E. Caulfield, Chao Zhang, Subir Goyal, Mehmet A. Bilen, Tyler J. Beardslee

Final approval of manuscript: Akhilesh Sivakumar, Sarah E. Caulfield, Chao Zhang, Subir Goyal, Mehmet A. Bilen, Tyler J. Beardslee

Disclosures

Mehmet A. Bilen: Exelixis, Bayer, Bristol-Myers Squibb, Eisai, Pfizer, AstraZeneca, Janssen, Calithera Biosciences, Genomic Health, Nektar, Sanofi (C/A, SAB), Xencor, Bayer, Bristol-Myers Squibb, Genentech/Roche, Seattle Genetics, Incyte, Nektar, AstraZeneca, Tricon Pharmaceuticals, Genome & Company, AAA, Peloton Therapeutics, and Pfizer (RF—Institution for work performed as outside of the current study). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board