Thromboembolic Events During Treatment with Cisplatin-based Chemotherapy in Metastatic Testicular Germ-cell Cancer 2000–2014: A Population-based Cohort Study

Background Cisplatin-based chemotherapy (CBCT) in testicular cancer (TC) is associated with elevated venous thromboembolism (VTE) risk, but trials evaluating the safety and efficacy of thromboprophylaxis are lacking. Objective To evaluate the arterial thromboembolism (ATE) and VTE incidence and risk factors during first-line CBCT for metastatic TC, and the effect of thromboprophylaxis on VTE and bleeding. Design, setting, and participants In a population-based study, 506 men administered first-line CBCT during 2000–2014 at three university hospitals in Norway were included. Clinical variables were retrieved from medical records. Outcome measurements and statistical analysis Patients with ATE and VTE diagnosed at initiation of or during CBCT until 3 mo after completion were registered. Age-adjusted logistic regression was performed to identify possible VTE risk factors. Results and limitations Overall, 69 men (13.6%) were diagnosed with 70 thromboembolic events. Twelve men (2.4%) experienced ATE. Overall, 58 men (11.5%) experienced VTE, of whom 13 (2.6%) were prevalent at CBCT initiation, while 45 (8.9%) were diagnosed with incident VTE. Age-adjusted logistic regression identified retroperitoneal lymph node metastasis >5 cm (odds ratio [OR] 1.99, 95% confidence interval [CI] 1.01–3.91), central venous access (OR 2.84, 95% CI 1.46–5.50), and elevated C-reactive protein (>5 mg/l; OR 2.38, 95% CI 1.12–5.07) as incident VTE risk factors. Thromboprophylaxis (n = 84) did not influence the risk of VTE (VTE incidence with or without prophylaxis 13% vs 8%, p = 0.16). The incidence of bleeding events was significantly higher among those who received thromboprophylaxis than among those without thromboprophylaxis (14.5% vs 1.1%, p < 0.001). Conclusions We found a high rate of thromboembolism incidence of 13.6%. Thromboprophylaxis did not decrease the risk of VTE but was associated with an increased risk of bleeding. Patient summary We found a high rate of thromboembolism (13.6%) during cisplatin-based chemotherapy for metastatic testicular cancer. Prophylactic treatment against thromboses did not reduce the thrombosis frequency, but it resulted in a high incidence of bleeding events.


1.
Introduction Testicular cancer (TC) treatment is a medical success story, with 10-yar overall survival approaching 90% even in metastatic disease [1]. The excellent prognosis in advanced TC was primarily achieved by the introduction of cisplatin in the late 1970s and with standardized diagnostics, treatment, and follow-up [2]. However, the general health of these young patients may be impaired by treatment-related morbidity, including thromboembolism.
Cancer patients have a four-to seven-fold higher risk of venous thromboembolism (VTE) than the general population [3]. VTE is among the leading causes of noncancer mortality among cancer patients [4]. The life-threatening potential of thromboembolism in TC patients was demonstrated by two recent large studies, reporting five-to sevenfold increased risks of death from cardiovascular disease (CVD), including VTE, during the 1st year after cisplatinbased chemotherapy (CBCT) [5,6].
The aims of this population-based cohort study were to evaluate ATE and VTE incidence and risk factors during primary CBCT for metastatic TC. Furthermore, we aimed to evaluate the effect of thromboprophylaxis, incidence of bleeding complications, and impact of thromboembolism on overall survival.

2.
Patients and methods

Patients
Treatment of metastatic germ-cell TC is centralized to four university hospitals in Norway, with treatment and follow-up according to the Swedish and Norwegian Testicular Cancer Group (SWENOTECA) proto-cols [16]. Thromboprophylaxis was not the standard treatment during the study period and was given at the discretion of the treating physician.
The use and type of thromboprophylaxis were registered. Only patients who received thromboprophylaxis for a minimum of 7 d were categorized as receiving such treatment [11]. Bleeding events throughout the study period were registered and classified as fatal, major (bleeding at a critical site and/or requiring transfusions with minimum two units of red cells and/or a fall in hemoglobin level of 2 g/dl) [19], or minor (clinically relevant nonmajor events).

Conclusions:
We found a high rate of thromboembolism incidence of 13.6%. Thromboprophylaxis did not decrease the risk of VTE but was associated with an increased risk of bleeding. Patient summary: We found a high rate of thromboembolism (13.6%) during cisplatin-based chemotherapy for metastatic testicular cancer. Prophylactic treatment against thromboses did not reduce the thrombosis frequency, but it resulted in a high incidence of bleeding events.
The longest axial diameter of RPLN metastasis was registered, and dichotomized with a 5-cm cutoff [10]. Khorana score was calculated based on the presence of TC and cutoff levels for BMI, hemoglobin, leukocyte, and thrombocyte count [20]. Creatinine clearance was estimated based on serum creatinine and age [21], with 90 ml/min/ 1.73 m 2 as the cutoff for normal kidney function [22]. Elevated CRP was defined as a value of >5 mg/l (upper normal limit).

Statistical analysis
Continuous variables are presented as median (interquartile range [IQR]), and categorical variables are presented as counts (proportion).
Groups were compared using the chi-square test. Cumulative survival was calculated with the Kaplan-Meier method.
The association between any thromboembolic events during treatment and overall mortality was assessed using age-adjusted Cox regression, presented as hazard ratio (HR) and 95% CI. Statistical analyses were performed using the SPSS 26.0 package (SPSS Inc., Chicago, IL, USA). Twosided p values of <0.05 were considered significant.

ATE incidence and risk factors
Overall, 12 men had ATE (2.4%). The majority (n = 11) occurred during chemotherapy ( Table 2). The median time from CBCT initiation to ATE diagnosis was 37 d (IQR 24-48). Five events were MI (1% of the total study population). Overall, 11 ATE patients were symptomatic. One was asymptomatic, identified at CT evaluation (renal infarction). There were no ATE-related deaths.
The median age at CBCT initiation for patients with ATE was 51 yr (IQR 40-53), considerably higher than for those without thromboembolic events (median age 32.4 yr, p < 0.001). Whereas 92% of patients diagnosed with ATE had IGCCCG good prognosis disease, 11 of 12 men had one or more CVD risk factors, most commonly smoking (n = 8) or obesity (n = 5). One patient had pre-existing CVD (stroke; Supplementary Table 1). Data are presented as n (%) unless otherwise specified. There are missing data for some of the variables (HCG, n = 1; AFP, n = 1; LD, n = 29). a Of 106 patients, 104 had stage I disease initially, of whom 92 relapsed while under surveillance. Two patients relapsed after radiotherapy for initially stage IIA disease. b According to the International Germ Cell Cancer Collaboration Group [15]. c None of which were PICC line.
The median age at CBCT initiation in 13 men with prevalent VTE was 46 yr (IQR 32-60; Overall, 196 men had none of the significant VTE risk factors identified in age-adjusted logistic regression mod-  els. Their incidence of VTE during chemotherapy was 4.6%, as compared with 13% among men with a minimum of one risk factor (p = 0.003).

Bleeding complications
The incidence of bleeding events in the study population was 4.2% (n = 21; Table 5). Overall, seven bleeding events occurred after the initiation of full-dose anticoagulation (10%). The incidence of bleeding events was significantly higher among those who received thromboprophylaxis (14%) than among those without thromboprophylaxis (1.1%; p < 0.001).
Bleeding was fatal (bleeding after RPLN dissection while on anticoagulation for pulmonary embolism) in one patient Data are presented as n (%) unless otherwise specified. All data based on laboratory and clinical examinations are at initiation of first chemotherapy cycle. There are missing data for some of the variables: HCG, n = 1; AFP, n = 1; LD, n = 29; hemoglobin, n = 35; leukocyte count, n = 43; platelets, n = 47; obesity, n = 1; Khorana score, n = 48; current smoker, n = 28, ECOG status, n = 89; creatinine clearance, n = 44; CRP, n = 129. a Only the 5 cm cutoff was associated with VTE risk [10]. The 3.5 cm cutoff was not significantly associated with VTE risk and is not reported [11]. b According to the International Germ Cell Cancer Collaborative Group [15]. c Khorana score was calculated based on the presence of testicular cancer, and cut-off levels for BMI, hemoglobin, leukocyte and thrombocyte count [20]. d Among 84 men with thromboprophylaxis, 81 men had low-molecular weight heparin (LMWH; n = 81), of whom 77 had low-dose LMWH (ie, enoxaparin 40 mg daily or dalteparin 5000 E daily) and four had LMWH in therapeutic dosage as prophylaxis (ie, enoxaparin 120 mg daily). Three received platelet inhibitors, for example, acetylsalicylic acid 160 mg daily. e Nine men were diagnosed with VTE while still on thromboprophylaxis and one after termination of thromboprophylaxis, and one had unknown disease. and major (none related to surgery) in five patients, of whom two (2.9%) were on full-dose anticoagulation, two (2.9%) were on thromboprophylaxis, and one (0.3%) was without thromboprophylaxis. Most bleeding events (n = 15) were minor.
In age-adjusted Cox regression, we observed a borderline significant association between prevalent or incident thromboembolism and overall mortality (HR 1.98, 95% CI 0.94-4.19). However, when including the prognosis group in the model, the association disappeared (HR 1.18, 95% CI 0.53-2.61).

Discussion
In this population-based cohort study, we found a thromboembolism incidence of 13.6% during primary CBCT for metastatic TC. Risk factors for incident VTE included RPLN >5 cm, central venous access, and elevated CRP. Importantly, thromboprophylaxis was not associated with a reduction in the VTE incidence, but with a high incidence of bleeding events, mostly minor. Our reported incidence rates of ATE (2.4%) and MI (1%) are considerably higher than the 0.3-1.2% ATE and 0.2-0.4% MI incidence rates reported previously [7][8][9]. In line with previous literature [23], men with ATE were older than those without thromboembolic events, and the majority had a minimum of one CVD risk factor. Still, these 12 men were considerably younger than the Norwegian general population at MI diagnosis (median age 51 vs 69 yr) [24], suggesting that CBCT-induced acute endothelial dysfunction might cause ATE [25].
The 11.5% VTE incidence rate confirms data from previous large studies [10][11][12]. In total, 2.6% of our patients had prevalent VTE, corroborating data from a large Spanish study [12] but lower than the rates of 4.4-6.5% reported by others [10,11]. Risk factors for prevalent VTE at TC diagnosis have not been reported previously. We found that prevalent VTE was more frequent in men with RPLN >5 cm, intermediate/poor prognosis disease, poor performance status, and elevated CRP. Consequently, we advise to examine these patients closely with regard to symptoms and/or radiologic findings, raising a suspicion of VTE.
Overall, 8.9% of our study patients were diagnosed with an incident VTE, supporting results from two large studies [10,11]. In line with previous large studies [8][9][10][11] and a recent literature review [26], we found that central venous access and large RPLN metastases were associated with an increased risk of incident VTE in age-adjusted analysis. Patients without any risk factors had 5% incidence of VTE, indicating a thrombotic potential of CBCT. We did not identify a high Khorana score as a risk factor for incident VTE, in contrast to two previous studies [10,11]. However, a high Khorana score (!3) was present in 39% of patients with prevalent VTE at TC diagnosis, probably reflecting advanced metastatic TC.
Elevated CRP at CBCT initiation was associated with an increased risk of incident VTE in our study, suggesting a proinflammatory state, rendering these men susceptible for the thrombotic potential of CBCT. To our knowledge, this is a novel finding in the TC patient population. Inflammation is important in the VTE pathogenesis in general [27,28] and among cancer patients [29]. A previous study among TC patients found elevated white blood cells to be associated with VTE [10], also reflecting the possible impact of inflammation. Age-adjusted and multivariable logistic regression. Overall, 13 men with prevalent VTE at initiation of chemotherapy were excluded. There are missing data for some of the variables: Khorana score, n = 48; performance status, n = 88; creatinine clearance, n = 44; CRP, n = 129. a Only the 5 cm cutoff was associated with VTE risk [10]. The 3.5 cm cutoff was not significantly associated with VTE risk and is not reported [11]. b According to the International Germ Cell Cancer Collaborative Group [15]. c Khorana score was calculated based on the presence of testicular cancer, and cutoff levels for BMI, hemoglobin, leukocyte, and thrombocyte count [20].
Recent randomized trials evaluating direct oral anticoagulants (DOACs) as thromboprophylaxis in ambulatory cancer patients given chemotherapy reported 60% risk reductions for VTE, with a double risk of bleeding [30,31]. Although the American Society of Clinical Oncology clinical practice guideline recommends thromboprophylaxis with DOACs or LMWH to selected high-risk ambulatory patients [32], no data from randomized trials support the routine use in TC patients, as the fraction of TC patients in recent trials was very small (<1%). Owing to the paucity of randomized data, a recent European Association of Urology guideline recommends balancing each patient's benefits and risk of thromboprophylaxis [33]. In line with previous reports [9,11], thromboprophylaxis did not reduce the incidence of VTE in our study, possibly due to the selection of patients with VTE risk factors for thromboprophylaxis. Although not statistically significant, Gizzi et al [8] reported 45% less VTE with thromboprophylaxis versus no thromboprophylaxis in a study among 151 TC patients with VTE risk factors (nine/97 vs nine/54, p = 0.23). However, a study reporting a 19% VTE incidence among 255 TC patients, of whom 93% received LMWH thromboprophylaxis, failed to show any effect of thromboprophylaxis [9].
Regarding thromboprophylaxis, the risk of bleeding complications must be taken into consideration. As many as 14% of our patients on thromboprophylaxis experienced a bleeding event. Even though these events were predominately minor, the proportion was considerably higher than among men without thromboprophylaxis (14% vs 1.1%, p < 0.001). In addition, the overall incidence of bleeding among men with thromboprophylaxis was considerably higher than the 2.5% reported in the Global Germ Cell Cancer Group (G3) study [34]. The 2.9% major bleeding incidence with thromboprophylaxis, mainly with LMWH, in our study was in line with the 2-3.5% reported in randomized trials using DOACs [30,31].
According to previous reports, cancer patients who develop thromboembolism, in particular VTE, have increased mortality during follow-up [3]. While some previous studies confirmed the adverse prognosis among TC patients with thromboembolism [9,12], neither our results nor the G3 study [11] confirmed this association when adjusting for the IGCCCG prognosis group.
Strengths of this relatively large study include the population-based design, homogeneous clinical practice across participating centers, and a predefined study population including only men administered first-line CBCT for metastatic TC. Data were extracted from medical records with a low risk of misclassification bias and a high likelihood of completeness. Limitations include missing data for some laboratory variables and skewness regarding selection of a low number of patients for thromboprophylaxis. To adjust for risk factors, a nonrandomized evaluation of thromboprophylaxis should ideally include a larger cohort than reported so far [8,9,11]. In our opinion, analyses on thromboprophylaxis reflecting clinical routine in this relatively large cohort is still important, given the absence of data from randomized trials and the rarity of metastatic TC.

Conclusions
In conclusion, although CBCT has a high thrombogenic potential, as demonstrated by the 5% incidence among men without any VTE risk factors, our study does not support the routine use of low-dose LMWH to prevent VTE. Given the high incidence of bleeding and the fact that VTE in this patient population did not influence survival, thromboprophylaxis should be considered only in selected patients. The most important risk factor for incident VTE seems to be central venous access use, which should be avoided in routine clinical practice [33].
Author contributions: Hege Sagstuen Haugnes had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. E U R O P E A N U R O L O G Y O P E N S C I E N C E 3 2 ( 2 0 2 1 ) 1 9 -2 7