Postoperative Deep Vein Thrombosis in Patients Undergoing Surgery With Curative Intent for Colorectal Cancer: A Prospective Cohort Study.

Introduction
Deep vein thrombosis (DVT) is a well-known postoperative complication, particularly among patients with cancer [1,2]. DVT is a significant cause of morbidity and mortality among patients with colorectal cancers [3,4] due to the increased risk associated with undergoing pelvic surgery of long duration, positioning during surgery, comorbid conditions, and neoadjuvant and adjuvant therapies. In addition, due to other risk factors, the colorectal cancer patients are categorized as high risk for developing postoperative DVT with a score of five or more when assessed using the Caprini risk assessment model [5,6].
To prevent postoperative DVT, several international guidelines, such as Enhanced Recovery After Surgery (ERAS) Society, the National Institute for Health and Care Excellence (NICE), the American Society of Colorectal Surgeons (ASCRS), and the American College of Chest Physicians (ACCP), have recommended postoperative DVT prophylaxis in the form of pharmacological as well as mechanical prophylaxis. The Caprini risk assessment model for venous thromboembolism (VTE) recommends a postoperative protocol of pharmacological thromboprophylaxis with low-molecular-weight heparin (LMWH), unfractionated heparin (UFH) or fondaparinux along with mechanical thromboprophylaxis like graduated compression stockings (GCS) for high-risk patients [7-10]. The recommended durations of thromboprophylaxis as per these guidelines are (i) standard duration of at least seven days for major abdominal surgeries; (ii) extended duration of 28 days for patients undergoing major open or laparoscopic abdominal or pelvic surgery for cancer including surgery for colorectal cancer; and (iii) duration of in-hospital stay recommended only for patients with a Caprini Risk Score of up to four.
Studies have indicated that there is a lower reported incidence of DVT/VTE in Asian populations [11,12]. A retrospective study over a 10-year period conducted in India among all patients diagnosed with VTE between 1996 and 2005 showed a VTE incidence of 17.46 per 10,000 admissions, and those following surgery were five per 10,000 surgeries, most of them (40.3%) being after general surgery [13]. A multicentric observational study conducted in India among 298 cancer patients who underwent abdominal or pelvic surgery found that none of the patients had symptoms of DVT [14]. However, duplex ultrasound studies were not done on any of them, and the duration of thromboprophylaxis given varied across centers.
The reported incidence of postoperative DVT in colorectal cancer patients ranges from 3.3% to 17.7% [1,4,15-18]. This risk remains high up to six months after surgery [18]. Studies have shown a reduction in VTE and its consequences following the extended thromboprophylaxis regimen [15,19-21]. However, other studies [22-24] have shown that only 8-27% of surgeons followed the recommended 28-day pharmacological thromboprophylaxis after colorectal cancer surgery. In a national survey, the majority (68%) of surgeons reported that they started thromboprophylaxis on admission and stopped it at discharge [23].
There is limited data on the incidence of postoperative DVT among colorectal cancer patients in the Indian context. A randomized study conducted in India with 99 patients who underwent curative surgery for colorectal cancer showed no instances of DVT on the 6 ± 1 postoperative day as confirmed by Doppler ultrasonography, regardless of whether the patients received LMWH prophylaxis (n=51) or not (n=48) [25]. The aim of our study was to estimate the frequency of postoperative DVT in Indian patients with colorectal cancer who have undergone curative intent surgery followed by postoperative thromboprophylaxis along with GCS during in-hospital stay.

Materials and methods
Study design and setting
A prospective observational cohort study was conducted from February 2019 to September 2020 by the colorectal unit of a tertiary care teaching institute (Christian Medical College, Vellore) in the state of Tamil Nadu, in Southern India. The study protocol for patients undergoing curative-intent surgeries for colorectal cancer included (a) risk assessment for developing postoperative DVT using the Caprini risk assessment model [5,6], (b) pre-operative assessment of DVT by a duplex ultrasound, performed by a consultant radiologist, to establish baseline status of the deep venous system as well as to identify patients with pre-existing DVT. Doppler ultrasonography was of both lower limbs and internal deep veins (the inferior vena cava, and the internal, external, and common iliac veins) to check for compressibility, phasicity, augmentation and flow in both lower limbs, and to assess flow in the internal deep veins, (c) one dose of LMWH or UFH at 6 PM on the evening prior to the surgery to all patients, (d) postoperative thromboprophylaxis with LMWH or UFH with GCS during the hospitalization period for all patients, (e) neoadjuvant and adjuvant chemo and radiotherapy was based on the TNM staging, the depth of penetration (T stage), involvement of lymph nodes (N stage) and the presence or absence of distant metastasis (M stage), as per the American Joint Committee on Cancer (AJCC) classification [26] (f) postoperative assessment of DVT in one to six months by a duplex ultrasonography doppler study done by a consultant radiologists who did the pre-operative assessment. Both open and laparoscopic surgeries were performed by the colorectal unit.
Inclusion and exclusion criteria
Inclusion Criteria
Inclusion criteria include: (i) Age ≥18 years; (ii) Diagnosis of colon or rectal cancer; (iii) Undergoing curative-intent surgery in the institute’s colorectal unit.
Exclusion Criteria
Exclusion criteria are as follows: (i) Previously diagnosed DVTA and (ii) symptomatic DVT detected on pre-operative duplex ultrasonography screening.
Caprini risk assessment model
The Caprini risk assessment model is a validated method of assessing the risk of development of VTE (Table 1) [5]. The Caprini Risk Score, the sum of the points based on the risk factors present, determines the risk levels and is categorized as low risk (score of 0-1), moderate risk (score of 2), high risk (score of 3-4), and highest risk (score of >5).
Postoperative follow-up
The patients were asked to return to the outpatient services of the unit for a follow-up with duplex ultrasound to investigate for asymptomatic DVT 28 days to six months after surgery. Those who could not return to the outpatient services were followed up with tele-consultation to assess if they had any symptoms of DVT.
Data collection
A survey questionnaire was developed to collect data on patients’ background characteristics, risk factors such as body mass index (BMI) and co-morbidities, clinical and radiological findings, operative and postoperative details along with complications, if any, and follow-up data 28 days after surgery. The information was collected by interviewing the patients and from their hospital records.
Statistical analysis
The data were processed and analysed using STATA version 18.0 (STATA Corporation, College Station, TX, USA). The primary outcome was the development of postoperative DVT. Descriptive analysis of frequencies (mean, median, and standard deviation (SD)) and percentages of patients’ characteristics such as age, sex, body mass index (BMI), co-morbidities, and other risk factors, and Caprini Risk Score were carried out. Bivariate analysis of postoperative complications and duration of surgery, with the type of surgery performed, was carried out. Fisher's exact test and the Mann-Whitney U test were performed to determine associations of risk factors for developing postoperative DVT among those who had and did not have a postoperative Doppler sonography and were considered significant at a p-value of <0.05.
Ethical considerations
Before the commencement of the study, the study was reviewed and approved by the Institutional Review Board at Christian Medical College, Vellore, India (IRB Min. No.11710 [OBSERVE] dated 03-12-2018). The consent form was prepared as per the institutional guidelines and translated into Tamil, Hindi, Telugu, and Bengali. Written informed consent was obtained from patients undergoing curative-intent surgery for colorectal cancer for participation in the study.

Results
Demographic characteristics, comorbidities, and risk factors
Between February 2019 and September 2020, 43 patients were recruited for the study. One patient was excluded from the study as his preoperative DVT screening had shown asymptomatic DVT. Therefore, 42 patients were included in the analysis. Most of them were males (78.6%), were between the ages of 20 - 74 years, with a mean age of 47.3 years (SD:13.04) and a median age of 48.5 years (Table 2). None of them had a family history of DVT or pulmonary embolism; 16 (38.1%) were smokers, chewed tobacco, or had alcohol, and 14 (33.0%) were obese (BMI>25kg/m2). Overall, 19 patients (45.2%) had one or more comorbidities, and hypertension in 12 patients (28.6%) and diabetes in 10 patients (23.8%) were the main comorbidities.
Cancer site, histopathological grading, TNM staging, and neoadjuvant therapy
A higher number of patients had rectal cancer (59.5%, n=25) than colon cancer (40.5%, n=17). Among the 17 patients who had colon cancer, the most common site was right colon (n=8), and others were synchronous adenocarcinoma of the right and sigmoid colon, metachronous recurrence at the ileo-transverse anastomosis after a right hemicolectomy was done for adenocarcinoma caecum seven years ago and had received three cycles of adjuvant chemotherapy, and multiple malignant colonic polyps in the sigmoid colon, transverse colon, and the hepatic flexure (Table 3). Most of the colorectal patients had moderately differentiated adenocarcinoma (66.7%, n=28), and were in TNM stage IIIB (45.2%, n=19) (Table 3). The Caprini Risk Score ranged from 4-7 (mean: 5.2, SD: 0.79), 19.0% had a score of four, 38.1% had a score of five, 26.2% had a score of six, 4.8% had a score of seven, and 81% (n=34) had a score of 5 or more. Twenty-two out of the 25 patients who had rectal cancer, mostly in TNM stage IIIB, received neoadjuvant chemoradiotherapy. They were operated between four and 30 weeks (mean: 9.6 weeks, SD: 5.84) of neoadjuvant chemoradiotherapy.
Surgical details and follow-up
Twenty-three (54.8%) surgeries were open procedures, two of which were started laparoscopically but were converted to open procedures, and 19 (45.2%) were laparoscopic surgeries. Of the 25 patients with rectal cancer, the common procedures were 11 (44%) abdominoperineal excision and nine (36%) low ultralow anterior resections. Of the 17 colon cancer procedures, the most performed procedure was six (35.3%) right hemicolectomies since the right colon was the most common location of colonic tumors. The duration of surgery available for 26 patients ranged from 60 to 300 minutes (mean: 195 minutes, SD: 57 minutes) for open procedures, and from 180 to 300 minutes (mean: 225 minutes, SD: 51 minutes) for laparoscopic surgeries. Overall, the maximum duration of any type of surgery was 300 minutes.
Postoperative complications occurred in 23 patients (54.8%), including 12 who had open surgery, 10 who had laparoscopic surgery, and one who had conversion from laparoscopy to open surgery. The most common complications following laparoscopic surgery were surgical site infection (n=7) and anastomotic leak (n=3), while the common complications after open procedures were urinary tract infection (n=4), and surgical site infections (n=3). All 42 patients received postoperative thromboprophylaxis along with GCS to prevent DVT. The study protocol included postoperative thromboprophylaxis with GCS during their in-hospital stay, and the duration of thromboprophylaxis was longer in patients with complications (mean: 10.3 days; SD: 5.06; range: 0-21 days) compared to those without complications (mean: 6.2 days; SD: 1.72; range: 1-9 days).
All patients were advised to visit the hospital for a Doppler ultrasonography four to 24 weeks after surgery to detect asymptomatic DVT. Due to travel restrictions during the COVID-19 pandemic, only 13 patients (31.0%) were able to visit the hospital for a Doppler ultrasonography at a mean duration of 9.85 weeks (range: 4-24 weeks) postsurgery and were found to have no DVT. The others who were followed up through teleconsultations for symptomatic DVT did not report any symptoms of DVT.
Among patients who underwent Doppler ultrasonography (n = 13), the mean duration of thromboprophylaxis was 9.08 days (SD: 4.75). Of these, 76.9% (n = 10) had a Caprini risk score ≥5 (mean: 5.08; SD: 0.76); 53.8% (n = 7) had at least one comorbidity; 69.2% (n = 9) experienced postoperative complications; 15.4% (n = 2) had advanced-stage colorectal cancer (Stage IIIA/B/C); and 15.4% (n = 2) had a history of addiction. Fisher’s exact test revealed no statistically significant association between undergoing follow-up doppler ultrasonography and the following factors: Caprini risk score ≥5 (p = 0.6861), presence of comorbidities (p = 0.5159), BMI >25 (p = 0.1587), postoperative complications (p = 0.3166), advanced-stage cancer (p = 0.4525), and addiction history (p = 0.0835). There was also no significant difference in mean thromboprophylaxis duration between those who did and did not undergo Doppler ultrasonography (Mann-Whitney U = -0.7482, p = 0.4533) (see Table 4). However, there was a statistically significant association between undergoing Doppler ultrasonography and receipt of neoadjuvant or adjuvant therapy, or both (p = 0.0414) (see Table 4).

Discussion
The aim of this study was to estimate the frequency of postoperative DVT in Indian patients with colorectal cancer who had undergone curative intent surgery and had postoperative pharmacologic plus mechanical thromboprophylaxis for the hospitalization period. One patient had a 30-week gap as the surgery had to be postponed because of COVID-19 infection; all the rest were operated between 4 and 16 weeks of neoadjuvant chemoradiotherapy. The results showed that none of the patients who had duplex ultrasound had asymptomatic DVT, and none of those who followed up through teleconsultation had symptomatic DVT.
For patients undergoing surgeries for colorectal cancer, the current guidelines recommend a 28-day postoperative pharmacological thromboprophylaxis along with mechanical thromboprophylaxis like GCS [7-10]. Choi et al. (2011) [3] and Khorana et al. (2011) [2] showed that patients with cancer are at a higher risk of DVT than those without. RCTs done by Kakkar et al. [15], Vedovati et al. [19] and several other studies have shown the benefits of extended 28-day thromboprophylaxis. A Cochrane review by Rasmussen et al. [27] reviewed randomized controlled trials studying the efficacy of prolonged thromboprophylaxis with LMWH in the prevention of postoperative VTE in patients undergoing abdominal or pelvic surgery. The incidence of 0.2% symptomatic VTE in patients receiving prolonged thromboprophylaxis with LMWH was significantly lower compared to 1.7% in the control group who received in-hospital thromboprophylaxis and either placebo/no-treatment during the extended period. However, other studies [22-24] have shown that only 8-27% of surgeons followed the recommended 28-day pharmacologic plus mechanical thromboprophylaxis after colorectal cancer surgery. In a national survey, the majority (68%) of surgeons reported that they started thrombo-prophylaxis on admission and stopped it at discharge [23].
Choi et al. (2011) [3] showed that the risk increased with age, BMI, the presence of one or more medical comorbidities, the site and stage of the cancer, and the administration of chemotherapy. Our study patients had multiple risk factors: Caprini Risk Score of >5 (81%), one or more comorbidities (46.2%), addictions (59.5%), advanced cancer (Stage IIIA/B/C) (35.5%), long duration surgeries of more than 45 minutes, a high-risk factor included in the Caprini Risk Score, and postoperative complications (69.2%). Comparing the patients who had follow-up duplex ultrasound and those who did not, they had similar risk profiles except that the group who had follow-up DVT had more exposure to neoadjuvant or adjuvant therapy, which increased their risk profile. Though the study patients had high Caprini Risk Scores and multiple risk factors, and though the duration of the postoperative pharmaco-mechanical thromboprophylaxis was only for the in-hospital period, there was no recorded instance of DVT during the follow-up scans. 
Studies have shown that Asian populations [11,12] and some from India [13,14,25] have a lower incidence of DVT in abdominal and pelvic surgeries in general and in colorectal cancers specifically. This may partly explain the absence of DVT in our study patients. Overall, our study indicates that postoperative pharmacological plus mechanical thromboprophylaxis for the in-hospital period was adequate to prevent DVT in Indian patients undergoing curative-intent surgery for colorectal cancer.
Limitations
The major limitations of this study were a small sample size and restricted follow-up Duplex ultrasound, both due to the COVID-19 pandemic travel restrictions. Although no symptoms of DVT were reported among those who were followed up through teleconsultations, their asymptomatic DVT could not be ruled out. As per the Caprini risk assessment model, surgeries lasting more than 45 minutes are considered a risk factor for developing DVT. Although data on the duration of surgery were available for only 26 patients, all surgeries with curative intent for colorectal cancer at our center exceeded 45 minutes in duration. Therefore, the absence of duration data for the remaining cases is unlikely to have affected the assessment of DVT risk.

Conclusions
Internationally, DVT is a significant cause of postoperative morbidity and mortality, especially among colorectal cancer patients having surgery. The patients included in this study did not have any symptomatic or asymptomatic DVT, though they had multiple risk factors for developing DVT.
While an extended course of pharmaco-mechanical thromboprophylaxis is recommended for patients undergoing surgery with curative intent for colorectal cancer, this study observed that pharmaco-mechanical thromboprophylaxis administered only during the hospital stay was sufficient to prevent postoperative asymptomatic DVT. These findings suggest a potentially lower risk of postoperative DVT in this patient population; however, larger prospective, multicentric studies are required to establish context-specific recommendations for practice in India.