Efficacy and safety of extended duration to perioperative thromboprophylaxis with low molecular weight heparin on disease-free survival after surgical resection of colorectal cancer (PERIOP-01): multicentre, open label, randomised controlled trial.

OBJECTIVE: To determine the efficacy and safety of extended duration perioperative thromboprophylaxis by low molecular weight heparin when assessing disease-free survival in patients undergoing resection for colorectal cancer.
DESIGN: Multicentre, open label, randomised controlled trial. SETTINGS: 12 hospitals in Quebec and Ontario, Canada, between 25 October 2011 and 31 December 2020. PARTICIPANTS: 614 adults (age ≥18 years) were eligible with pathologically confirmed invasive adenocarcinoma of the colon or rectum, no evidence of metastatic disease, a haemoglobin concentration of ≥8 g/dL, and were scheduled to undergo surgical resection.
INTERVENTIONS: Random assignment to extended duration thromboprophylaxis using daily subcutaneous tinzaparin at 4500 IU, beginning at decision to operate and continuing for 56 days postoperatively, compared with in-patient postoperative thromboprophylaxis only. MAIN OUTCOME MEASURES: Primary outcome was disease-free survival at three years, defined as survival without locoregional recurrence, distant metastases, second primary (same cancer), second primary (other cancer), or death. Secondary outcomes included venous thromboembolism, postoperative major bleeding complications, and five year overall survival. Analyses were done in the intention-to-treat population.
RESULTS: The trial stopped recruitment prematurely after the interim analysis for futility. The primary outcome occurred in 235 (77%) of 307 patients in the extended duration group and in 243 (79%) of 307 patients in the in-hospital thromboprophylaxis group (hazard ratio 1.1, 95% confidence interval 0.90 to 1.33; P=0.4). Postoperative venous thromboembolism occurred in five patients (2%) in the extended duration group and in four patients (1%) in the in-hospital thromboprophylaxis group (P=0.8). Major surgery related bleeding in the first postoperative week was reported in one person (<1%) in the extended duration and in six people (2%) in the in-hospital thromboprophylaxis group (P=0.1). No difference was noted for overall survival at five years in 272 (89%) patients in the extended duration group and 280 (91%) patients in the in-hospital thromboprophylaxis group (hazard ratio 1.12; 95% confidence interval 0.72 to 1.76; P=0.1).
CONCLUSIONS: Extended duration to perioperative anticoagulation with tinzaparin did not improve disease-free survival or overall survival in patients with colorectal cancer undergoing surgical resection compared with in-patient postoperative thromboprophylaxis alone. The incidences of venous thromboembolism and postoperative major bleeding were low and similar between groups. TRIAL REGISTRATION: ClinicalTrials.gov NCT01455831. © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Introduction

Since the establishment of a link between anticoagulants and cancer survival,1 investigators have been evaluating the use of anticoagulation to improve cancer outcomes. Warfarin has been reported to decrease cancer recurrence in patients with melanoma.1 Heparins, in particular, low molecular weight heparin, tinzaparin,2 have shown antimetastatic properties in preclinical models by potentially inhibiting tumour cell invasion of the extracellular matrix, hindering protection of tumour cells in the circulation from immune mediated destruction, and impairing neovascularisation of micrometastases due to an antiangiogenic effect.3 An initial meta-analysis assessing the effect of low molecular weight heparin on cancer outcomes reported that the use of low molecular weight heparin improved overall survival.4 However, more recent analyses including all published studies to date have been unable to support an overall survival benefit for cancer patients.5 Nonetheless, most of these studies included patients with advanced malignancies or combined multiple tumour types, which can hinder the evaluation of the effect of low molecular weight heparin in preventing metastatic disease.

The perioperative period might be the ideal setting to study the antimetastatic effects of low molecular weight heparin.6 Surgery results in a hypercoagulable postoperative state and an inability to clear micrometastatic disease when the disease is present after surgical resection, leading to potential cancer recurrence and worse survival.7 In preclinical studies, perioperative administration of low molecular weight attenuated postoperative metastatic disease by preventing platelet and fibrin binding and thereby facilitating immune-mediated destruction of tumour cell microemboli by natural killer cells.8

Colorectal cancer is the third most common cause of cancer and about 35% of patients recur after surgical resection. The PERIOP-01 study was designed to investigate whether perioperative extended duration low molecular weight heparin (tinzaparin) could improve disease-free survival, when compared with postoperative administration in hospital of tinzaparin in patients with colorectal cancer without evidence of metastatic disease and who were scheduled to undergo surgical resection.

Methods

Study design and oversight

The PERIOP-01 trial was a multicentre, open label, randomised controlled trial comparing extended duration to in-hospital thromboprophylaxis using low molecular weight heparin. The members of the steering committee (see appendix) had final responsibility for the trial design, clinical protocol, and study oversight. The institutional review boards at each of the 12 participating sites in Quebec and Ontario in Canada approved the protocol, which is available online with the full text of this article. Data were collected at the sites and entered in an online database managed by the Methods Centre of the Ottawa Hospital Research Institute. A central adjudication committee, whose members were unaware of treatment assignment, reviewed all suspected outcome events. An independent data safety monitoring board periodically reviewed trial outcomes.

Study participants and setting

Patients with colorectal cancer, no evidence of metastatic disease, and scheduled to undergo surgical resection were potentially eligible. The inclusion criteria required a diagnosis of pathologically confirmed invasive adenocarcinoma of the colon or rectum, preoperative investigations that showed potential resectability without evidence of metastatic disease, a haemoglobin concentration of 8 g/dL or more, patients aged 18 years or older, and the capacity to provide written informed consent. Neoadjuvant treatment (chemotherapy or radiation or both) was allowed. Patients were excluded if they had carcinoma only present in a completely excised polyp; previous venous thromboembolism, including deep vein thrombosis or pulmonary embolism; a need for continuous anticoagulation; contraindication to heparin treatment (eg, previous history of heparin induced thrombocytopenia, thrombocytopenia (ie, platelet count <100×109/L), renal insufficiency with glomerular filtration rate <30 mL/min); and history of other cancers (except for adequately treated basal or squamous cell carcinoma or carcinoma in situ) within five years of the colorectal cancer diagnosis. Other exclusion criteria included pregnancy, breastfeeding, and unwillingness or inability to provide informed consent.

Randomisation and intervention

Eligible patients were randomly assigned in a 1:1 ratio at the time of decision for surgery, by use of a centralised, web based randomisation system, to extended duration thromboprophylaxis, within 24 h of randomisation (required a minimum of one preoperative dose to maximum of six weeks) and continuing for 56 days postoperatively, or in-hospital thromboprophylaxis, beginning on postoperative day 1 and continuing for the duration of stay in hospital. Tinzaparin (4500 IU, subcutaneously, daily) was used for thromboprophylaxis in both groups (extended duration and in-hospital thromboprophylaxis). Randomisation was permuted in blocks of two and four, prepared using random number tables by the trial’s statistician (RM). Randomisation was stratified by participating centre and tumour type (rectal or colon cancer). Patients were followed up for up to five years or to death, regardless of the duration of study drug.

Outcomes

The primary outcome was disease-free survival at three years in the intention to treat population. Disease-free survival at three years been shown to be an excellent predictor of overall survival at five years.9 10 11 Disease-free survival events were defined as survival without locoregional recurrence, distant metastases, second primary (same cancer), second primary (other cancer), or death.12 Secondary outcome measures were venous thromboembolism, major surgery related bleeding events, major bleeding events, clinically relevant non-major bleeding events, postoperative complications,13 14 transfusion requirements, overall survival at five years, wound infection, and anastomotic leakage. Venous thromboembolism was defined as objectively confirmed; symptomatic or incidental; proximal, lower, or upper extremity deep vein thrombosis; unusual site thrombosis (ie, cerebral, splanchnic, or renal vein thrombosis); or pulmonary embolism; occurring between the day of the surgery and postoperative day 56 (coincident with discontinuation of tinzaparin in the extended duration group). Major surgery related bleeding was defined as overt bleeding that was associated with a decrease in the haemoglobin level of 4 g/dL or more, led to transfusion of four or more units of packed red blood cells, required reoperation, or contributed to death during the initial seven days after surgery.15

Major bleeding was defined by the International Society of Thrombosis and Haemostasis as overt bleeding that was associated with a decrease in the haemoglobin level of 2 g/dL or more, led to transfusion of two or more units of packed red blood cells, occurred in a critical site (such as intracranial, intraspinal, intraocular, retroperitoneal, intra-articular or pericardial, or intramuscular with compartment syndrome), or contributed to death.15 Major bleeding events were captured between the day of randomisation to the day of the surgery and between postoperative day 7 and 56. All study outcomes were blindly adjudicated by an independent adjudication committee, consisting of oncology or thrombosis specialists. Compliance with study drugs was estimated using syringes counted by patients in a medication diary and reviewed by the study coordinator. Compliance was defined as high if 80% or more of the study drug was taken.

Statistical analysis

We expected a disease-free survival of 66% at three years in patients receiving in-hospital thromboprophylaxis. The trial hypothesis was that extended duration thromboprophylaxis with tinzaparin is superior to in-hospital thromboprophylaxis in providing at least an absolute reduction of 8.5% (hazard ratio of 0.75) in the primary outcome (disease-free survival at three years). To detect this hazard ratio using a significance level of two sided α=0.05 and a power of 80%, a total of 380 events needed to be observed before the final analysis. Based on the projected event rate for disease-free survival, we estimated that 1075 patients with colorectal cancer undergoing surgical resection would be sufficient to show this reduction.

The primary analysis was done in the intention-to-treat population that included all patients who had been randomly assigned to a group. The hazard ratios for the primary outcome were estimated using a Cox proportional hazard model controlling for age, sex, and centre. We also did a time-to-event analysis on overall survival at five years. Proportions of the other secondary outcomes were compared using χ2. The hazard ratios for the primary outcomes in prespecified subgroups were estimated using a Cox proportional hazard model controlled for age, sex, and centre. One sensitivity analysis including compliant patients (>80% of study drug) was planned for the primary outcome.

Two interim analyses for futility were planned. These analyses were expected to be done once 30% and 60% of enrolled participants have completed three year follow-up. The first interim analysis was done on 23 July 2020 using data from 591 patients, which constituted 55% of the total target sample size of 1075. The median follow-up was 2.2 years (interquartile range 1.0-3.6). Recurrent cancers, new cancers, or deaths occurred in 59 events in 298 patients receiving in-hospital thromboprophylaxis and 59 events in 293 patients receiving extended duration thromboprophylaxis. Based on these event rates, and the assumptions made in the original power calculation (380 events), the conditional power to detect a hazard ratio of 0.75 was determined to be 48%. On 6 November 2020, the data safety and monitoring board recommended to stop enrolment based on a low conditional power and difficulty in recruitment due to the covid pandemic. The final sample size was 614.

Patient and public involvement

Patient partners were not involved in the design or conduct of this study that began in 2011. Patient partners in the CanVECTOR network (www.canvector.ca) will be involved in dissemination/knowledge translation activities.

Results

From 25 October 2011 to 31 December 2020, 614 patients underwent randomisation at 12 hospitals in Canada and were included in the primary and secondary analyses (fig 1). The baseline characteristics of the patients were well balanced between groups (table 1). Overall, the mean age was 61 years, fewer than half of participants were women (41%), and most people were white (90%). Three hundred and eleven patients (51%) had primary rectal cancer whereas 303 (49%) had primary colon cancer. Two hundred and twenty nine (38%) patients received neoadjuvant treatment for rectal cancer and 287 (48%) patients with colon or rectal cancer received adjuvant treatment; these results were similar between groups. A total of 201 (67%) patients in the extended duration group and 203 (76%) patients in the in-hospital thromboprophylaxis group had laparoscopic surgery, respectively. Overall, 201 (33%) patients had node positive disease; 105 (34%) in the extended duration group and 96 (31%) in the in-hospital thromboprophylaxis group. The median duration of study drug, preoperatively, was four days (interquartile range two to nine) in the extended duration group; and postoperatively, 55 days (53-56) in the extended duration group and five days (three to seven) in the in-hospital thromboprophylaxis group. The study drug was discontinued as per participants’ wish in eight patients in each of the extended duration and standard thromboprophylaxis groups. Compliance was high in both groups at 93% and 97% in the extended duration and in-hospital thromboprophylaxis groups, respectively. The median duration of follow-up was 35 months (interquartile range 17-53) in the extended duration group and 35 months in the in-hospital thromboprophylaxis group (18-49).

Fig 1

Trial profile

Table 1

Baseline clinical characteristics. Data are number (%) of participants unless stated otherwise

Baseline characteristics	In-hospital thromboprophylaxis (n=307)	Extended thromboprophylaxis (n=307)
Personal
Mean (SD) age (year)	60.8 (12.6)	61.4 (13.2)
Women	121 (39)	128 (42)
Race or ethnicity:
White	273 (89)	278 (90)
Black or African American	6 (2)	4 (1)
Hispanic	3 (1)	3 (1)
Asian	11 (4)	15 (5)
Native American	9 (3)	5 (2)
Other	3 (1)	2 (1)
Mean (SD) body mass index*	5.5 (28.4)	5.7 (28.3)
Mean (SD) creatinine clearance (mL/min)	17.4 (74.8)	17.5 (76.7)
Tumour type
Colon	151 (49)	152 (49)
Rectum	156 (51)	155 (51)
Positive family history of colorectal cancer†	50 (17)	52 (18)
ECOG performance status
0 or 1	244 (80)	251 (82)
2	63 (20)	56 (18)
Concomitant antiplatelet drug treatments
Acetylsalicylic acid	42 (14)	46 (15)
Non-steroidal anti-inflammatory drugs	20 (7)	26 (9)
Neoadjuvant or adjuvant treatment‡
Neoadjuvant treatment (before surgical resection):	113 (37)	116 (39)
Pelvic radiation: short course (25 Gy in 5 fractions)	45 (15)	58 (20)
Pelvic radiation: long course (55 Gy in 25 fractions)	68 (22)	57 (19)
Adjuvant treatment:§	143 (53)	144 (54)
Capecitabine or fluorouracil	72 (27)	77 (29)
Oxaliplatin	63 (24)	61 (23)
Other	14 (5)	14 (5)
Operative data
Type of surgery:
Laparoscopy	203 (67)	227 (76)
Open	100 (33)	72 (24)
Procedure:
Right hemicolectomy	48 (16)	55 (18)
Left hemicolectomy	23 (8)	21 (7)
Transverse colectomy	7 (2)	8 (3)
Sigmoid colectomy	30 (10)	30 (10)
Low anterior resection	146 (48)	139 (47)
Abdominal perineal resection	31 (10)	29 (10)
Subtotal colectomy	3 (1)	4 (1)
Unknown or multiple procedures	15 (5)	13 (4)
Heparin administration in the operating room:
Given just before surgery	287 (95)	290 (97)
Given in the operating room	108 (36)	103 (35)
Estimated blood loss:
<200 mL	188 (63)	198 (68)
≥200-499 mL	91 (30)	75 (25)
500-999 mL	18 (6)	15 (5)
1000-2000 mL	3 (1)	5 (2)
>2000 mL	0	0
Cancer staging
Pathological T stage:**
0	27 (9)	22 (7)
1	25 (8)	23 (8)
2	76 (26)	81 (28)
3	149 (50)	145 (49)
4	22 (7)	23 (8)
Node positive††	96 (32)	105 (36)
Positive margins††	9 (93)	5 (2)
Intraoperative metastatic disease	2 (1)	4 (1)

SD=standard deviation.

Missing 37 values (18 in standard and 19 in extended duration).

Family history is defined as a diagnosis in one or more first degree relatives; missing 23 values (11 in standard and 12 in extended duration).

Excludes 12 patients who were randomly assigned but were withdrawn before surgery (four in standard and eight in extended duration groups).

Received within 12 months of surgery; missing 65 values (35 in standard and 30 in extended duration groups).

Missing nine values (three in standard and six in extended duration).

Missing nine values (four in standard and five in extended duration).

Missing eight values (four in standard and four in extended duration).

Primary and secondary outcomes

The primary outcome occurred in 235 (77%) of 307 patients in the extended duration group and in 243 (79%) of 307 patients in the in-hospital thromboprophylaxis group (hazard ratio 1.1, 95% confidence interval 0.90 to 1.33; P=0.4; table 2, fig 2).

Table 2

Primary and secondary clinical outcomes. Data are number (%) of participants

Outcomes	In-hospital thromboprophylaxis(n=307)	Extended thromboprophylaxis(n=307)	Hazard ratio (95% CI)	P value
Primary outcome
Disease-free survival at three years	243 (79.2)	235 (76.5)	1.1 (0.90 to 1.33)	0.4
Local recurrence*	42 (13.7)	44 (14.3)	—	—
Distal recurrence	9 (2.9)	7 (2.3)	—	—
New primary cancer	6 (2.0)	6 (2.0)	—	—
Death	3 (1.0)	7 (2.3)	—	—
Secondary outcomes
Venous thromboembolism	4 (1.3)	5 (1.6)	—	0.8
Pulmonary embolism	2 (0.7)	3 (1.0)	—	—
Deep vein thrombosis	2 (0.7)	1 (0.3)	—	—
Other (portal vein thrombosis)	0	1 (0.3)	—	—
Major surgery related bleeding	6 (2.0)	1 (0.3)	—	0.1
Major bleeding	2 (0.7)	0	—	0.5
Clinically relevant non-major bleeding†	6 (2.0)	9 (2.9)	—	0.5
Postoperative grade III-V complications (Clavian):
Cardiac	2 (0.7)	1 (0.3)	—	NS
Respiratory	3 (1.0)	2 (0.7)	—	NS
Neurological	0	1 (0.3)	—	0.5
Gastrointestinal	14 (4.6)	8 (2.6)	—	0.2
Renal	2 (0.7)	1 (0.3)	—	NS
Other	2 (0.7)	5 (1.6)	—	0.3
Wound infections	37 (12.1)	39 (12.7)	—	0.1
Anastomotic leakage (grade I-V)	13 (4.2)	6 (2.0)	—	0.8
Transfusion requirements	11 (3.6)	12 (3.9)	—	0.8
Overall survival at five years	280 (91.2)	272 (88.6)	1.12 (0.72 to 1.76)	0.1

NS=not significant; CI=confidence interval.

Both local and distant recurrences occurred in four patients in the in-hospital group and three patients in the extended thromboprophylaxis group.

Clinically relevant non-major bleeding was defined as overt bleeding that did not meet the criteria for major bleeding but was associated with the use of medical intervention, contact with a physician, interruption of the assigned treatment, discomfort, or impairment of activities of daily living.

Fig 2

Kaplan-Meier plot of cumulative event rates of disease-free survival

Table 2 shows the secondary outcomes. Venous thromboembolism occurred in five patients (2%) receiving extended duration and in four patients (1%) in the in-hospital thromboprophylaxis groups (P=0.8). Similarly, bleeding related to major surgery occurred in one patient (<1%) receiving extended duration group and in six patients (2%) in the in-hospital thromboprophylaxis group (P=0.1), and additional major bleeding events were reported in no (0%) patient and two (0.7%) patients in the extended duration and in-hospital thromboprophylaxis groups, respectively. Nine (3%) patients in the extended duration group and six (2%) patients in the in-hospital thromboprophylaxis group had a clinically relevant non-major bleeding event (P=0.5). Postoperative complications occurred in 136 (22%) of 614 patients and were similar in both groups, including 76 wound infections, 19 anastomotic leaks, and 41 Clavien grade III-V postoperative complications. Twenty three patients required transfusion of packed red blood cells: 12 (median 1; range 1-2) in the extended group and 11 (2; 2-4) in the in-hospital thromboprophylaxis group. Adverse events, unrelated to surgery, were reported in 102 patients in the extended duration group and 101 patients in the in-hospital thromboprophylaxis group, of which seven and two events were classified as being possibly related to study drug.

Overall survival at five years after randomisation was 272 (89%) patients in the extended duration and 280 (91%) patients in the in-hospital thromboprophylaxis group (hazard ratio 1.12, 95% confidence interval 0.72 to 1.76; P=0.1; table 2).

In the sensitivity analysis of compliant patients (>80% compliance), the primary outcome could be evaluated in 216 (78%) of 276 patients in the extended duration group and in 235 (80%) of 292 patients in the in-hospital thromboprophylaxis group (hazard ratio 1.13, 95% confidence interval 0.90 to 1.42; P=0.3).

Subgroup analyses for the primary outcome are shown in supplementary figure S1 in the appendix. The hazard ratios were not significant for the primary outcome in patients with colon cancer at 1.42 (95% confidence interval 0.88 to 2.30) and with rectal cancer at 0.88 (0.62 to 1.24). Patients who underwent surgery also had non-significant hazard ratios for laparoscopic surgery (0.96 (0.72 to 1.29)) and for open surgery (1.34 (0.96 to 1.87)). Similarly, no difference was reported for disease-free survival in patients who received pain controlled intravenous anaesthesia (1.15 (0.89 to 1.49) or epidural anaesthesia (1.13 (0.73 to 1.74)). No significant interactions between other subgroups and type of thromboprophylaxis were reported (supplementary figure S1).

Discussion

Principal findings

The PERIOP-01 trial showed that extended duration to perioperative thromboprophylaxis with tinzaparin did not result in an improvement in disease-free survival at three years compared with in-hospital thromboprophylaxis among patients with localised invasive colorectal cancer without evidence of metastatic disease who were eligible to undergo surgical resection. Furthermore, the rates of venous thromboembolism and postoperative major bleeding complications were low in this patient population, and similar between groups.

To our knowledge, this study is the largest randomised controlled trial assessing the role of low molecular weight heparin and disease-free survival in patients with cancer undergoing surgical resection. Recruitment was stopped prematurely because the predefined interim analysis reported a low power (48%) to reject the null hypothesis and difficulties in recruitment. The results of our study are consistent with the Tinzaparin in Lung Tumors (TILT)16 and the Adjuvant Chemotherapy with or without Nadroparin in Patients with Completely Resected Non-Small-Cell lung Cancer (NVALT-8) trials,17 evaluating the effect of adjuvant tinzaparin and nadroparin in patients with resected stage II-III lung cancer where low molecular weight heparin had no effect on overall or disease-free survival. Prophylactic dosing of tinzaparin was used in the PERIOP-01 trial because therapeutic dosing could have a higher risk of bleeding complications in the perioperative setting. A substantial number of patients with colorectal cancer are already bleeding from the underlying tumour before surgery, and major bleeding complications might increase after major abdominal surgery in patients receiving therapeutic dosing of an anticoagulant.18 Furthermore, other dosing regiments have already been assessed by TILT (intermediate dosing of tinzaparin 100 IU/kg, once a day for 12 weeks) and NVALT-8 (two weeks of therapeutic dosing then 14 weeks on intermediate dosing) and showed no benefit.16 17 The PERIOP-01 trial used tinzaparin as an early adjuvant treatment starting from the day of surgery until day 56. At day 56, patients with indications for adjuvant chemotherapy started treatment.

About 53% of patients in our trial received adjuvant chemotherapy (table 1). Although low molecular weight heparins have consistently shown antimetastatic properties in preclinical studies, this effect has not been translated into clinical benefit. The metastatic cascade and subsequent development of clinically detected recurrence is a complex process with a multitude of mechanistic interactions, and modifying one mechanism could simply be insufficient in any clinical setting. Nonetheless, the immediate postoperative period remains an important window to prevent the development of metastatic disease and an area to explore novel treatments. Importantly, the PERIOP-01 trial also assessed the role of preoperative thromboprophylaxis because the tinzaparin was initiated before the resection for four days (interquartile range two to nine) to ensure that if antimetastatic properties required preoperative administration, these could be abrogated before surgery and potential metastatic spread. Hence, the PERIOP-01 provides evidence that both preoperative and postoperative extended duration thromboprophylaxis with tinzaparin does not improve disease-free survival in patients with colorectal cancer undergoing surgical resection.

The rates of venous thromboembolism were low (<2%) without significant differences between the two groups. These low rates of venous thromboembolism are consistent with a recent meta-analysis of randomised trials reporting a 30 day incidence of clinical venous thromboembolism of 1.4% in patients receiving extended duration and 0.3% in patients receiving thromboprophylaxis in hospital after major abdominal or pelvic surgery.19 In our study, no significant reduction in symptomatic venous thromboembolism was reported between the extended duration and in-hospital thromboprophylaxis groups (2% v 1%). Several explanations might account for this discrepancy. Unlike previous randomised controlled trials, screening ultrasounds of the lower extremities were not done. The use of asymptomatic deep vein thrombosis of the lower extremities diagnosed on screening ultrasound or venography is often used as a surrogate outcome measure; however, its clinical relevance remains unclear. Outcomes important to patients, including symptomatic events, are preferred to establish clinical practice guidelines.20 Furthermore, the definition of venous thromboembolism used in the PERIOP-01 trial included upper extremity and unusual site deep vein thromboses. Finally, venous thromboembolism was reported over a 56 day follow-up period to coincide with discontinuation of tinzaparin in the extended duration group. Nonetheless, the PERIOP-01 trial is the largest randomised trial study of postoperative thromboprophylaxis with low molecular weight heparin at an extended duration in this population and provides additional data that will help clinicians to assess the risk (bleeding) to benefits (venous thromboembolism) ratio of using extended duration thromboprophylaxis in people with colorectal cancer undergoing surgical resection.

Limitations of this study

A limitation of our trial is the open label design, which could be associated with a risk of bias regarding the frequency of the outcomes, as compared with the frequency that might have been observed in a placebo controlled trial. However, the primary outcome measure (disease-free survival) in our trial is an objective outcome that means that bias is unlikely. Furthermore, all primary and secondary outcomes were adjudicated by a blinded, independent committee. The PERIOP-01 trial included patients with both rectal and colon cancers leading to some heterogeneity in the patient population. However, the trial was pragmatic and representative of current clinical practice of general and colorectal surgeons. Furthermore, all participating centres provided comprehensive colorectal cancer care in their respective regions and all patients undergoing a surgical resection were assessed for eligibility and enrolment, so we believe that the results are generalisable. Finally, the trial was stopped early after the recommendations from the data monitoring and safety board. However, given the lower power reported at the interim analysis, completion of the total sample size would be unlikely to change the interpretation and conclusions of the trial.

Conclusion

Extended duration to perioperative thromboprophylaxis with tinzaparin (given before surgery and for 56 days after surgery), as compared with in-hospital thromboprophylaxis, does not increase disease-free survival at three years in patients with localised invasive colorectal cancer without evidence of metastatic disease who were eligible to undergo surgical resection. The rates of clinically detected venous thromboembolism were low and extended duration thromboprophylaxis was not associated with a reduction in venous thromboembolism.

Low molecular weight heparin attenuates cancer metastases in preclinical models and has been reproduced in many animal models

This effect has not been translated into reduced metastases or improved cancer outcomes in patients

The perioperative period could be the ideal setting to study the antimetastatic effects of low molecular weight heparin because surgery results in a hypercoagulable state and an inability to clear micrometastatic disease, leading to potential cancer recurrence and worse survival

Extended duration to perioperative thromboprophylaxis with tinzaparin does not result in an improvement in disease-free survival at three years compared with in-hospital thromboprophylaxis among patients with localised invasive colorectal cancer without evidence of metastatic disease who were eligible to undergo surgical resection

Rates of venous thromboembolism and postoperative major bleeding complications were low in this patient population, and similar between groups