The impact of comorbidities on post-operative complications following colorectal cancer surgery

Background Colorectal cancer surgery is complex and can result in severe post-operative complications. Optimisation of surgical outcomes requires a thorough understanding of the background complexity and comorbid status of patients. Aim The aim of this study is to determine whether certain pre-existing comorbidities are associated with high grade post-operative complications following colorectal cancer surgery. The study also aims to define the prevalence of demographic, comorbid and surgical features in a population undergoing colorectal cancer resection. Method A colorectal cancer database at The Prince Charles Hospital was established to capture detailed information on patient background, comorbidities and clinicopathological features. A single-centre retrospective study was undertaken to assess the effect of comorbidities on post-operative outcomes following colorectal cancer resection. Five hundred and thirty-three patients were reviewed between 2010–2018 to assess if specific comorbidities were associated with higher grade post-operative complications. A Clavien-Dindo grade of three or higher was defined as a high grade complication. Results Fifty-eight percent of all patients had an ASA grade of ASA III or above. The average BMI of patients undergoing resection was 28 ± 6.0. Sixteen percent of all patients experienced a high grade complications. Patients with high grade complications had a higher mean average age compared to patients with low grade or no post-operative complications (74 years vs 70 years, p = 0.01). Univariate analysis revealed patients with atrial fibrillation, COPD, ischaemic heart disease and heart failure had an increased risk of high grade complications. Multivariate analysis revealed pre-existing atrial fibrillation (OR 2.70, 95% CI 1.53–4.89, p <0.01) and COPD (OR 2.02 1.07–3.80, p = 0.029) were independently associated with an increased risk of high grade complications. Conclusion Pre-existing atrial fibrillation and COPD are independent risk factors for high grade complications. Targeted perioperative management is necessary to optimise outcomes.

Surgical department at TPCH provides care for colorectal cancer patients whose clinical background and comorbidities are captured in a detailed clinicopathological database. A retrospective study using The Prince Charles Hospital colorectal cancer database was undertaken to investigate the effect of patient comorbidities on post-operative outcomes. Patient charts were individually reviewed and data extracted by trained medical personnel. Data was obtained from previous and current admissions as well as correspondence letters, follow up documentation and outpatient/readmission notes.

Study population
Patients who underwent surgical resection of colorectal cancer at The Prince Charles Hospital between January 2010 and December 2018 were included in the study.
Patients were excluded if they had endoscopic resection of the malignant lesion without surgical intervention. Patients below the age of 12 were also excluded from the study.

Measures and definitions
Basic demographic data such as age, gender, date of birth, height, weight, body mass index (BMI) and American Society of Anaesthesiologists (ASA) grade were documented. Specific pre-operative conditions were grouped into comorbid categories: cardiovascular, respiratory, metabolic, autoimmune and renal disease.

Cardiovascular comorbidities
Cardiovascular diseases recorded included ischemic heart disease, previous coronary artery bypass grafting, previous percutaneous coronary intervention, pacemaker insertion, defibrillator insertion (AICD), previous valve repair, heart failure (any type), cardiomyopathy, hypertension, pulmonary hypertension and atrial fibrillation.

Respiratory comorbidities
Respiratory diseases recorded included asthma, COPD, bronchiectasis, cystic fibrosis and obstructive sleep apnoea.

Metabolic, autoimmune and renal disorders
Metabolic diseases recorded included type I diabetes, type II diabetes and hyperlipidaemia. Autoimmune conditions recorded included rheumatoid arthritis, psoriasis, polymyalgia rheumatica, and systemic lupus erythematosus (SLE). Baseline renal function was noted in each patient.

Surgical and pathological features
Surgical data included the operation performed, urgency of surgery, operative approach, discharge date and length of stay. Pathological data included tumour histopathology, histological grade and TNM stage of disease.

Post-operative outcomes
The database and patient charts were reviewed for specific post-operative complications and their treatment was documented. Complications were Graded from I to V according to the Clavien-Dindo Classification of Surgical Complications (S1 Appendix). Complication grades were divided into low grade complications (Clavien-Dindo grades I-II) and high grade complications (Clavien-Dindo grades III-V) for analysis. Post-operative complications were defined as those that arose up to 60 days post-operatively. Grading was performed by trained medical clinicians. Quality assurance of collected data was undertaken by an independently trained clinician.

Statistical analysis
Differences in demographic features, comorbidities and surgical/pathological features between the two complication groups were assessed using t-tests and Fisher's exact tests as appropriate. Statistically significant results were defined as those with p � 0.05.
The association between comorbidities and high grade post-operative complications was assessed using uni-and multivariate cox regressional analysis to assess odds ratios for each comorbidity. All comorbidities were assessed using univariate analysis, while only those that were found to be statistically significant were analysed in the multivariate model. Statistical review was performed by a qualified biomedical statistician. Data was analysed with Stata v14 software (StataCorp).

Ethics
Ethics approval for this database was granted by the Prince Charles Hospital Human Research Ethics Committee (Reference: HREC/17/QPCH/295). A waiver of consent was approved to allow access to confidential patient information without consent. Patients were not anonymized prior to data collection. However, patient names and certain other identifying data was not recorded in the database to help guard against confidentiality breaches. Patient data was accessed between January 2018 and June 2019.

Results
Five hundred and thirty-three patients who underwent colorectal cancer resection between January 2010 and December 2018 were reviewed.

Demographics
The mean age of patients undergoing colorectal cancer surgery was 67.4 years ± 15.6 SD. The mean BMI was 27 ± 6.0 SD. There was a 50:50 even distribution of males and females within the cohort and the majority of patients (n-= 309, 58%) had an ASA of �3 (indicating the patient is living with a severe disease process).
Twenty-eight percent of patients had respiratory disease with asthma (n = 58, 11%) and COPD (n = 62, 12%) being the most common lung conditions. Fifty-four percent of all patients had metabolic disease (n = 288) with hyperlipidaemia being the most common (n = 158, 30%) followed by type II diabetes (n = 96, 18%). Only 2% of patients (n = 11) had an autoimmune disease. The majority of patients had a preoperative estimated glomerular filtration rate (eGFR) >90 ml/min/1.73m2 (n = 304, 57%) which is indicative of no underlying renal disease. Thirty percent of the cohort had an eGFR between 60-89 while 9% had an eGFR between 45-59. Forty-three percent of patients had a degree of renal impairment prior to surgery. Fourteen percent of patients were current smokers at the time of surgery (n = 76). Table 1 outlines the demographic data and pre-morbid features of the patients.

Surgical and pathological features
The majority of cancers (n = 304, 57%) were located within the right side of the colon (from the caecum to the hepatic flexure). Left sided colon cancers (from the splenic flexure to the sigmoid) made up 30% of cases (n = 161) while rectal cancers made up 17% (n = 88).
Right hemicolectomies, extended right hemicolectomies, appendicectomies and caecectomies (n = 294, 55%) made up the majority of surgical cases and reflects the dominance of right sided colon cancers. The majority of surgical cases were performed laparoscopically (including laparoscopically assisted) (n = 410, 77%) while open cases (including laparoscopic converted to open) made up 23% of all operations (n = 121).
The most common histopathology was adenocarcinoma of no special type (n = 436, 82%). Most patients in both the nil/low and high grade complications groups had an early stage of caner (Stage I and IIA). The majority of our patients had stage I disease, with no difference in distribution of disease between the two groups (p = 0.73). The surgical and pathological features of the cohort are outlined in Table 2.

Complications
Complications were divided into surgical complications and medical complications.

Patient background and comorbidities associated with high grade complications
Patients who had Clavien-Dindo Complication Classifications graded III-V were classified as having high grade post-operative complications. The demographic features of high-grade post-operative complications are outlined in Table 1.
Out of the entire study cohort, 85 patients (16%) had high grade post-operative complications. Patients who had high grade post-operative complications were older than those with nil/low grade complications (median 74 vs 70 years old, p = 0.01).
Those with high grade complications had a higher proportion of patients with ASA III-V when compared to those with nil/low grade complications (73% vs 55%, p = 0.016).
Patients with high grade complications were also found to have double the incidence of COPD (21% vs 9%, p = 0.005) ( Table 1). There was no significant difference between the two groups with regards the incidence of type II diabetes, hyperlipidaemia and baseline renal function.

Univariate analysis of comorbidities associated with high grade complications
Patient background data, comorbidities and their association with high grade complications was analysed by univariate analysis and presented in  (Table 4).
Coronary artery bypass grafting was found to be just above the threshold for statistical significance (OR 2.14, 95% CI 1.02-4.49, p = 0.053). Types II diabetes, obstructive sleep apnoea, hyperlipidaemia, asthma and percutaneous coronary intervention were not found to be associated with high grade complications.
Multivariable logistic analysis (Table 5) was performed to allow for adjustment for relevant factors. Only comorbidities that were statistically significant in univariate analysis were analysed with multivariate analysis.

Discussion
We performed a retrospective study to determine the effect of comorbidities on surgical outcomes following colorectal cancer resection. Out study demonstrated that patients with pre-existing atrial fibrillation were almost three times more likely to have severe post-operative complications while patients with COPD were found to be twice as likely to have similar complications.
The independent association between atrial fibrillation and poorer surgical outcomes has never before been described in literature.
Our results demonstrate that patients with high grade complications following colorectal cancer surgery were more likely to be older, have significantly higher rates of ischaemic heart disease, previous coronary artery bypass grafting, atrial fibrillation, heart failure and COPD. This is of particular interest as previous studies have associated 'cardiac disease' with postoperative complications without delineating the effect of specific comorbidities. While ischaemic heart disease, heart failure and increased age were found to be associated with the  development of high-grade complications on univariate regression analysis they were not independently associated with high grade complications on multivariate analysis. Atrial fibrillation (AF) is a common cardiovascular comorbidity that frequently occurs within a similar demographic to that of CRC. Approximately 5% of Australia's population over the age of fifty-five have atrial fibrillation [14]. Although common, atrial fibrillation is by no means innocuous as it predisposes patients to thromboembolism, haemodynamic compromise and arrhythmogenesis [15]. Associated issues regarding ventricular rate control and anticoagulation in the peri-operative setting can further predispose patients with AF to haemodynamic compromise and post-operative bleeding [16,17].
Several studies have demonstrated an association between cardiovascular disease and an increase in post-operative complications [18,19], however, none of these studies have delineated the patient's specific type of cardiovascular disease. Cuthbert et al.'s 2018 paper demonstrated that patients with pre-existing cardiovascular disease had poorer overall median survival rates, with cardiovascular disease being the most common cause of non-CRC related mortality [20].  [21]. However, this study included patients who were found to be in AF either pre or post-operatively. Post-operative AF has been shown to be a marker of both systemic inflammation and anastomotic breakdown and thus, could account for the high all-cause mortality demonstrated in Walsh's study [22,23]. To date, there is no literature that specifically links the relationship between pre-existing atrial fibrillation and post-operative complications following CRC surgery.
COPD was also found to be independently associated with high grade complications. COPD is the second leading cause of disease burden in Australia in patients aged 65-74 [24]. In a 2004 population based study, patients with COPD had a 13-day mortality rate of 13%, compared with only 5.3% without COPD [25]. Lemmens et al. also demonstrated that COPD is associated with higher rates of pneumonia and post-operative bleeding following colorectal cancer resections [13].
One third of all patients with pre-existing atrial fibrillation and 29% of patients with COPD developed high grade post-operative complications. Amongst these patients, there was a higher proportion of medical complications compared to surgical complications. Patients with COPD who developed high grade complications had higher incidences of post-operative respiratory infections, ischaemic cardiac events and had an increased requirement for respiratory and ventilatory support. Patients with AF who developed complications were more likely to develop cardiac arrhythmias and respiratory complications.
Complications arising in those with AF and COPD seem to largely be due to an exacerbation of their underlying cardiorespiratory disease rather than a specific post-operative surgical aetiology. Further to this, atrial fibrillation and COPD also appear to be markers of poorer overall health and a higher comorbid burden in patients.
It is pertinent to note that across all patients within the study, there was a high burden of comorbid disease. Table 1 shows that 307 patients (58%) had an ASA grade of three or above, indicating a severe systemic disease that limits activity [26]. Data from the 2018 Bi-national Colorectal Cancer Audit [2] produced by the Colorectal Surgical Society of Australia and New Zealand (CSSANZ) showed that only 36% of colorectal surgical patients nationally had an ASA of 3 or greater. This indicates that patients in our hospital catchment area undergoing colorectal cancer resections have a higher burden of disease than those nationally which may be a contributing factor to the rates of post-operative complications following complex abdominal surgery.
The association between pre-operative atrial fibrillation and high grade complications following colorectal cancer resection is a novel finding. The association between COPD and poorer surgical outcomes is equally important in order to help optimise patients perioperatively and manage prospective outcomes. Medical optimisation of patients and their comorbidities in the perioperative period is also important in minimising high grade post-operative complications. Collaborating with physicians, anaesthetists, peri-operative medical teams and allied health is crucial in minimising the risk of high grade post-operative complications and improving patient outcomes.
This study highlights the importance of pre-existing atrial fibrillation and COPD as specific diseases which require careful perioperative optimisation, planning and management in order to minimise their impact on the incidence of high grade complications.