Incidence of thromboembolic events in patients with locally advanced rectal cancer receiving neoadjuvant chemoradiotherapy

To the Editor,

Patients with rectal cancer showing locally advanced disease at the time of diagnosis require a multidisciplinary management. The CAO/ARO/AIO 04 trial showed that oxaliplatin, a third-generation platinum analogous, when combined with 5-FU and radiotherapy (RT), yielded an increased pathological complete response (pCR-rate) [1].

By contrast, the three phase III trials showed that the oxaliplatin addition did not significantly improve the pCR rates [2–4] and the impact of oxaliplatin on disease free and overall survival is still to be evaluated.

A comprehensive overview of the published studies on RT-related late adverse effects for rectal cancer evidenced an increased risk of thromboembolism (TE) in the long-term follow-up of the RT group of the Stockholm I and II trials, when they have been analysed together [5].

Cancer is one of the major risk factors for the development of Venous ThromboEmbolism (VTE) and colorectal cancer patients are associated with high risk of VTE. Both metastatic disease and comorbidity are strong predictors of VTE [6].

Several patients have multiple risk factors which work in a cumulative fashion [7]. Moreover, chemotherapy is also recognised as an independent risk factor for thrombosis and an analysis of four different triplet chemotherapy regimens has demonstrated a differential thrombogenic effect according to the use of platinum salts [8].

The aim of our analysis is to document the incidence and pattern of VTE and Arterial Thrombo Embolism (ATE) events that arise in locally advanced rectal cancer patients during preoperative radiochemotherapy and within six months after surgery. The main goal is to compare the incidence of thromboembolism among different chemotherapy regimens associated with radiotherapy and to evaluate the role that personal risk factors may play in the development of thromboembolism.

Material and methods

A retrospective analysis based on 77, histologically confirmed, locally advanced rectal cancer (stage IIA–III) (cT3 or Node positive) patients, who consecutively received chemoradiation in neoadjuvant settings between September 2007 and October 2010 (arm 1) and between 2000 and 2004 (arm 2), was performed.

The patients’ characteristics are listed in Table I.

Table I. Characteristics of the 77 studied patients.

Parameter	Total	Group 1	Group 2
Gender
Male	54	30	24
Female	23	9	14
Race
White	76	38	38
Black	1	1	–
Age
Mean	64	63	64
≤ 60	27	11	16
> 60	50	28	22
Tumour stage
IIA	30	11	19
IIB	2	–	2
IIIA	5	2	3
IIIB	27	21	6
IIIC	13	5	8
Tumour type
Mucinous	5	5	–
Non-mucinous	72	34	38
Tumour grading
G1	6	5	1
G2	45	28	17
G3	6	3	3
Missing	20	3	17

Baseline data regarding VTE specific risk factors [7] were collected from all patients and listed in Table II.

Table II. Risk assessment.

Parameter	Group 1	Group 2
BMI
Average	26	26
> 35 kg/mq	1	1
Performance status
0	27	23
1	9	15
2	1	–
Missing	2	–
Comorbidity
0	24	24
1	10	13
2	5	1
Smoker
Yes	19	21
No	20	17
Previous VTE
Yes	3	2
No	36	36
Platelet
Average	270	233
> 350	7	1
Leukocyte
Average	7.1	6.7
> 11 000	2	–
Haemoglobin
Average	13	12
< 10	1	3
CVC
Yes	39	38
No	–	–

According to ESMO clinical recommendation.

BMI, body mass index ≥ 35 kg/m²; comorbidity, recent surgery, diabetes, hormone replacement therapy; CVC, central venous catheter; haemoglobin, haemoglobin level < 10 g/dl and/or use of erythropoiesis-stimulating agent; leukocyte, leukocyte count ≥ 11 × 10⁹/l; platelet, pre-chemotherapy platelet count of 350 × 109/l or more; PS, performance status (ECOG).

The major exclusive criterion was VTE prophylaxis for any indications.

Radiation therapy was delivered with a three-dimensional (3D)-conformational multiple field technique to a dose of 45 Gy (in 25 daily fractions of 1.8 Gy given in 5 weeks) to the whole pelvis – PTV1-, plus 5.4–9 Gy (3–5 daily fractions of 1.8 Gy) to the tumour volume – PTV2-, with 6–15 MV energy photons. In all patients, the presacral, obturator and internal iliac lymph nodes were included. The external iliac lymph nodes were included if clinically positive or in case of T4 tumour. The inguinal lymph nodes were irradiated if there was major tumour extension to the anal sphincters.

The planned chemotherapy administered in association with radiotherapy was 5-fluorouracil (5-FU) (200 mg/m²/days by continuous i.v. infusion for 5 days per week on days of planned RT) plus oxaliplatin (50 mg/m²/days 1–8–15–22–29–35 i.v. once a week as a 2 hours i.v. concurrently with RT) in arm 1 and 5-FU (200 mg/m²/days by continuous i.v. infusion for 5 days per week on days of planned RT) in arm 2.

Surgery was carried out in all patients. Total mesorectal excision (TME) was the treatment of choice, when feasible. The operation was usually delayed for 6–10 weeks after the end of radio chemotherapy. All patients had prophylaxis with low molecular weight heparin (LMWH) for four weeks after hospitalisation.

Postoperatively, patients received adjuvant Folfox (Oxaliplatin 85 mg/m² days 1, folinic acid 400 mg/m² day 1, 5-FU mg/m² g1 bolus and 5-FU 1200 mg/m² i.c. days 1–2 every two weeks) or 5-FU based therapy for 4–6 months.

TE events were collected via chest and abdominal computed tomography (CT)-scan: i) at baseline; ii) before planned surgery; and iii) at the end of planned adjuvant treatment.

VTE and ATE events were classified and graded according to the National Cancer Institute Common Adverse Events version 4.1. No screening for polymorphisms was performed.

The association between drug exposure and principal outcomes were analysed using univariate analysis. Fisher's exact test was performed to determine statistical significance. This test is non-parametric and it does not rely on normality assumption. For this reason it appears to be the more suitable test for this type of data. The level of significance was 0.05 (one-sided).

Results

Seventy-seven patients with resectable locally advanced non-metastatic rectal cancer (23 women and 54 men, age range 38–83 years) were enrolled in the study.

Ten of the 77 patients experienced a TE during treatment with an overall rate of thromboembolism of 13%; 8/77 (10%) experienced VTEs and 2/77 (2.6%) experienced ATEs. No CVC-related thrombosis was reported. Sixty percent (6/10) patients experienced an asymptomatic VTE detected only on CT-scan, and two symptomatic VTE and two symptomatic ATE events were reported.

The majority of TE events (90%) occurred during neoadjuvant radiochemotherapy and were detected during the evaluation before surgery. Only one of these events occurred at the end of the planned adjuvant chemotherapy (arm 1).

Three events were considered Grade 3 and three Grade 2. One case of Grade 1 was reported. No cases of life-threatening events (G4) were reported in our series. A fatal PE (G5) was detected in the oxaliplatin group. In each group, we recorded one high grade and fatal (G5) ATE (cardiac ischemia and stroke, respectively). In the fluoropyrimidine group, the overall incidence of thromboembolism was significantly lower (1/38, 5%) compared with the oxaliplatin group (8/39, 21%, p = 0.047).

A univariate analysis of all parameters including the VTE specific risk factors, showed no significant association with VTE events in the two groups of treatment.

Discussion

In our study, we found that the incidence of TE is relevant. We reported an incidence of thromboembolism during treatment of 13%. According to Heit [9], such a relevant incidence can be explained by the presence of active cancer cells that directly induce a hypercoagulable state involving many complex interdependent mechanisms.

There is growing evidence that cancer cells can produce tissue factor and a cancer procoagulant, which activates the coagulation system.

In our series, the high incidence of TE is due to the record of symptomatic and asymptomatic TE as well.

The widespread use of the recently introduced multi-detector CT-scanner, has resulted in an increasing number of incidentally diagnosed PE in asymptomatic cancer patients. [10] In our series the majority of screened TE events were incidental (6/10), confirming that the incidence, prevalence and mortality rate of incidental TE are probably underestimated.

Our results confirmed that VTE is a common and potentially lethal disease in patients with active cancer.

Almost 30% of TE cases presented as sudden death. Two fatal ATE and one fatal PE were reported. Ninety percent of the TE events occurred during neoadjuvant radio chemotherapy. Only one event occurred after surgery.

We confirmed that chemoradiotherapy exposure is a strong risk factor for VTE. Chemotherapy and radiation are thought to act synergistically in causing VTE when administered simultaneously.

Cancer patients receiving cytotoxic chemotherapy have a 6.5-fold increased risk for VTE and chemotherapy is recognised as an independent risk factor for thrombosis because it may cause damage to the vascular endothelium, disequilibrium between procoagulant and anticoagulant molecules, tumour/endotelial apoptosis, cytokine activation and increased tissue factor activity [11].

According to Chau [12], we reported a significantly (p = 0.0473) higher thrombogenic effect of oxaliplatin exposure compared to fluoropyrimidine, during pelvic radiotherapy for rectal cancer according to the thrombogenic effect of platinum salts [8]. Oxaliplatin-associated vascular toxicity has been reported in hepatic non-tumour-bearing tissue in patients with metastatic colorectal cancer who underwent chemotherapy for potentially resectable liver metastases [13], and the vessel damage may be the cause of the thrombogenic potential of the drug.

Radical surgery doubled the risk of postoperative VTE, but the prophylaxis minimised that risk and no TE event has been reported postoperatively.

Our series, in contrast to the data provided in current reports, failed to show a relationship between all risk factors analysed and TE [11,14], but performing a test of significance on this type of data might be potentially misleading because small studies cannot rest on a normality assumption for that data. To this extent, performing a Fisher test is preferable as the test is non-parametric in nature and does not rely on normality of data.

However, the most important controversy regarding retrospective studies as opposed to randomised trials, is the fact that the test cannot account for the potential effect of other variables that could affect patients in the two arms. In this specific study, risk factor adjustments are not feasible given the small number of patients involved.

Thus, even if the one-sided Fisher test shows a significant p-value (0.047), great caution is required in interpreting this result. The most important recommendation emerging from this study is the clear need for a larger randomised clinical trial.

In conclusion, our results suggest VTE as one of the most important causes of morbidity and mortality in cancer patients, causing a negative impact on their quality of life. However, thromboprophylaxis for patients who are receiving chemotherapy still remains controversial. Our data suggests that the primary prophylaxis should be considered during the neoadjuvant arm, when oxaliplatin is associated with fluoropirimidine and radiotherapy, because the initial radiochemotherapy represents a critical period for TE incidence. Nevertheless, our data must be confirmed by a larger randomised clinical trial.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.