The impact of age on rectal cancer treatment, complications and survival

The number of older patients with rectal cancer is increasing. Treatment outcome discrepancies persist, despite similar treatment guidelines. To offer the oldest patients optimal individually adjusted care, further knowledge is needed regarding treatment strategy and outcome. The present study aimed to evaluate treatment, postoperative complications, and survival in older patients treated for rectal cancer. This retrospective study included all 666 patients (n=255 females, n=411 males) treated for rectal cancer at Levanger Hospital during 1980-2016 (n=193 <65 years, n=329 65-79 years, n=144 ≥80 years). We performed logistic regression to analyse associations between complications, 90-day mortality, and explanatory variables. We performed a relative survival analysis to identify factors associated with short- and long-term survival. Despite a similar distribution of cancer stages across age-groups, patients aged ≥80 years were treated with a non-curative approach more frequently than younger age groups. Among patients aged ≥80 years, 42% underwent a non-curative treatment approach, compared to 25% of patients aged <65 years, and 25% of patients aged 65-79 years. The 90-day mortality was 15.3% among patients aged ≥80 years, compared to 5.7% among patients aged <65 years, and 9.4% among patients aged 65-79 years. Among 431 (65%) patients treated with a major resection with curative intent, the 90-day mortality was 5.9% among patients aged ≥80 years (n=68), compared to 0.8% among patients aged <65 years (n=126), and 3.8% among patients aged 65-79 years (n=237). The rate of postoperative complications was 47.6%. Pneumonia was the only complication that occurred more frequently in the older patient group. The severity of complications increased with three factors: age, American Society of Anaesthesiologists score, and >400 ml perioperative blood loss. Among patients that survived the first 90 days, the relative long-term survival rates, five-year local recurrence rates, and metastases rates were independent of age. Patients aged ≥80 years were less likely to undergo a major resection with curative intent and experienced more severe complications after surgery than patients aged <80 years. When patients aged ≥80 years were treated with a major resection with curative intent, the long-term survival rate was comparable to that of younger patients.


Introduction
The incidence of rectal cancer in Norway is among the highest in the world [1]. Moreover, the aging of the population has led to a high number of older patients. Questions remain to be resolved regarding rectal cancer Open Access *Correspondence: oystein.hoydahl@gmail.com treatment in this heterogeneous group of patients to offer optimal individualized care.
Treatments for rectal cancer have evolved over the last four decades. Diagnostic tools have become more available. The diagnostic work-up is performed according to standardized protocols that apply across age-groups [2]. The implementation of total mesorectal excision (TME) and the addition of preoperative radiotherapy for locally advanced tumours have improved survival [3,4]. Minimally invasive surgery has reduced surgical trauma, and protocols to enhance recovery after surgery have become a standard part of treatment [5,6]. The modern principles in rectal cancer treatments, including TME and preoperative radiotherapy, have been applied to all patients treated at Levanger Hospital since 1980. A prospective protocol for the operative strategy, radiotherapy, and surveillance was established, and excellent results were reported after the first ten years [7].
The fundamental treatment for rectal cancer includes a resection of the tumour-bearing segment of the rectum. This procedure is associated with substantial postoperative morbidity [8]. Complications may be fatal, particularly in aged, vulnerable patients with low capacity to withstand physiological stress. This risk has had a heavy impact on the choice of treatment for aged patients, and thus, it may adversely affect both functional results and survival. The number of older patients with rectal cancer is expected to increase in the years to come [9]; thus, deeper knowledge is needed to pursue individually optimized care.
The present study aimed to evaluate treatment, complications, and survival in patients with rectal cancer during 1980-2016, with a special focus on the older patients.

Methods
This study included all patients treated for rectal cancer at Levanger Hospital during 1980-2016. Levanger Hospital was the primary hospital of 10 municipalities in Norway, and the catchment area remained unchanged throughout the study period. The population increased by 18%, from 83,890 inhabitants in 1980, to 99,566 inhabitants in 2016.
We identified patients through the hospital administrative system and reviewed health records for all patients discharged with diagnosis codes for rectal cancer, based on the International Classification of Diseases, 8 th revision (ICD-8) codes 154, ICD-9 codes 154, and ICD-10 codes C20. To ensure a complete cohort, the retrieved data were crosschecked and confirmed with data recorded in the Norwegian Cancer Registry during 1980-2016. We retrieved data on demographic variables, comorbidities, treatment, tumour characteristics, histopathology, postoperative complications, and shortand long-term survival.
During the study period, 51 patients in our catchment area were referred for treatment to other hospitals; 33 patients were referred to the nearest university hospital for preoperative radiotherapy and underwent surgery there; and 18 patients chose to receive treatment in another hospital. These patients were not included in our cohort. The characteristics of these patients are presented Table 9 in Appendix. This study included a total of 666 patients treated for rectal cancer.
Rectal cancer was defined as a tumour located within 15 cm of the anal verge, measured with a rigid proctoscope. The rectal sections were defined according to the distance above the anal verge: the proximal rectum was at 12-15 cm; the middle rectum was at 6-11 cm; and the distal rectum was at 0-5 cm.
Disease stages were classified according to the TNM classification, sixth edition [10]. Signs of residual tumour after surgery were classified as R0 -no microscopic residual tumour; R1 -microscopically involved resection margin; and R2 -macroscopic residual tumour. A major resection with curative intent was defined as a resection of the tumour-bearing segment of the rectum, including R1-resections and tumour perforations, with no radiological or preoperative signs of metastases. Major resections were performed according to TME principles [3,11]. In four patients, the resections were performed laparoscopically, all in the last part of the study period (2010-2016). A histopathological verification of cancer was missing in 20 of 666 patients (3%); however, the rectal cancer diagnosis was evident from other examinations. Among these 20 patients, 7 underwent non-resection procedures, and 13 underwent best supportive care, without resection.
Preoperative radiotherapy was recommended for patients with fixed, locally advanced tumours, according to national guidelines established in 1993 [12]. During 1980-1999, referrals for radiotherapy were based on proctoscopy and digital examinations. In 2000, magnetic resonance imaging of the rectum became available at our hospital, and it was used as the decisive diagnostic modality for evaluating tumour resectability. All patients selected for radiotherapy were referred to the nearest university hospital. The majority of patients received 2 Gy ×25, but in selected cases, patients received 5 Gy ×5.
For this study, patients were categorized into groups according to treatment intent. The curative intent group included patients with (i) a major resection (R0 and R1) or (ii) a polypectomy. The non-curative intent group included patients with (iii) a major resection, (iv) a bypass/stoma, or (v) best supportive care.
Comorbid conditions were classified according to the American Society of Anaesthesiology (ASA) score and the Charlson Comorbidity Index (CCI) [13,14]. Emergency surgery was defined as a surgery due to evidence of large bowel obstruction or large bowel perforation.
Postoperative complications included any deviation from the normal postoperative course during the same hospital admission, and were noted in the patient records. The severity of postoperative complications was graded according to the Clavien-Dindo (CD) classification of surgical complications [15].
Clinical follow-up was initially conducted according to local guidelines. Starting in 1993, follow-ups were conducted according to similar national guidelines [2]. Normal follow-ups lasted for 5 years, but they were extended in selected cases. The follow-up period for this study ended on December 31 st , 2018. The mean follow-up time was 6.12 years (range: 0.02 -34.04, SD: 5.78).
Survival time calculations started from the date of admission and ended at the last known date that the patient was alive or the date of death. Patients that were alive December 31 st , 2018, were counted as censored cases. The mean follow-up time with regard to survival was 6.89 years (range: 0.01 -37.88, SD: 7.49).

Statistical analysis
The Cochran Armitage exact trend test was performed to test for trends in proportions; for example, the proportion of Hartmann's procedures (HPs) performed per decennium.
The Joncheere-Terpstra test was performed to test for the distribution of blood loss volumes (dependent variable) across decennium periods (independent variables). Kaplan-Meier analyses were performed to estimate the 5-year rates of local recurrences and metastases.
Logistic regression analysis was performed to test for associations between the 90-day mortality (dependent variable) and different explanatory variables. Ordinal logistic regression was performed to test for associations in doubly ordered r × c contingency tables; for example, the ASA scores in different age groups. The resulting odds ratio (OR) was a common OR estimate for any 2 × 2 contingency table that would occur, if the r ×c table were collapsed to a 2 × 2 table, based on any cut-off threshold, along the columns and rows. Multinomial logistic regression analysis was performed in singly ordered r x c contingency tables; for example, the type of treatment in different age groups.

Relative survival analysis
Relative survival is a measure of mortality compared to the general population. The observed survival in the group with cancer was divided by the expected survival of a comparable group in the general Norwegian population, matched with respect to age, sex, and the calendar year of investigation. Relative survival was estimated with the Ederer II method and analysed with STATA 16 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC) [16]. Multivariable analyses were performed with a full likelihood approach. We retrieved data on Norwegian population survival probabilities for every year, starting from 1980, calculated for groups divided by sex and age, from the Human Mortality Database [17]. Two

All patients
The characteristics of all 666 patients treated for rectal cancer in 1980-2016 are presented in Table 1. Patients were predominantly male (61.7%), and the mean age was The CCI and ASA score increased with increasing age. Distal tumours were more prevalent in the oldest age group. The rate of patients with stages I and II tumours increased throughout the study period, but the rate of patients with unknown stages declined; only one patient had an unknown stage in the last time period (2010-2016). Overall, 17% of patients aged ≥80 years had an unknown tumour stage. This rate declined from 41.7% in 1990-1999 to 0% in 2010-2016. The distribution of tumour stages did not differ between age groups.
The overall rate of patients treated with a major resection with curative intent was 65%, and this rate remained consistent throughout the study period. The rate varied across age-groups; it was 65% among patients under 65 years old, 72% among patients 65-79 years old, and 47% among patients ≥80 years old. The distribution of treatment intent categories differed across age groups. The proportion of patients that underwent best supportive care was higher among patients ≥80 years old.
Among patients treated with a non-curative intent, 27.1% (52/192)  Radiotherapy was administered to 33% (63/192) of the patients in the non-curative treatment intent group. Among these, 13 patients underwent radiotherapy as a part of a curative treatment plan, and 50 patients underwent palliative radiotherapy. Among the patients treated with palliative radiotherapy, 29 had metastases at diagnosis.
The 90-day mortality, overall survival, and relative long-term survival rates for all patients are presented in Table 2. The 90-day mortality after admission was 9.6%, and it increased significantly with age. In addition, the five-year overall and relative survival rates in patients that survived the first 90 days decreased with age. Prognostic factors associated with 90-day mortality are presented in Table 3. Age was not a significant factor, but the calendar year of treatment, a high ASA score, and the treatment intent category were significant independent variables. The five-year relative survival rates among all patients were 62

Patients with stages I-III disease treated with a major resection with curative intent
The characteristics of 431 (64.7%) patients with rectal cancer stages I-III that were treated with a major resection with curative intent (R0 and R1) are presented in Table 4. These patients were predominantly males (63.1% Ordinal multivariable logistic regression analyses of risk factors associated with the CD severity of postoperative complications are presented in Table 5. Independent risk factors were: increasing age, increasing ASA scores, and perioperative blood loss >400 ml.

Short-and long-term survival among patients that underwent a major resection with curative intent
The 90-day mortality, overall survival, and relative longterm survival rates in patients with rectal cancer stage I-III that underwent a major resection with curative intent are presented in Table 6. The 90-day mortality rate after admission was 3.2%.
The five-year relative survival rates decreased insignificantly with age in this patient group (    Multivariable analyses identified several factors associated with 90-day mortality (Table 7). Mortality increased with increasing age and ASA scores, and decreased over time (i.e., calendar year).
Prognostic factors associated with long-term relative survival are presented in Table 8. Age was not significantly associated with relative survival. However, CCIs ≥3, increasing ASA scores, emergency surgeries, and stage III disease were significantly inversely associated with long-term survival.

Local recurrence and metastasis among patients that underwent a major resection with curative intent
Local recurrence was diagnosed in 7% (29/417) of patients with rectal cancer stage I-III that underwent a major resection with curative intent. The overall estimated five-year local recurrence rate was 7.3% (95% CI: 4.5 to 10.1). The estimated five-year local recurrence rates after an R0 resection, an R0 resection with tumour perforation, and an R1 resection were 4.9% (95% CI: 2.5 to 7.3), 29

Discussion
The present study showed that the TNM stage at presentation was equally distributed across age groups. The overall rate of patients treated with a major resection with curative intent was 65%, but the rate varied across age groups: it was 47% among patients aged ≥80 years. One or more postoperative complications occurred in 47.6% of patients. The rates of postoperative complications were independent of age, except for pneumonia, which was more common in patients aged ≥80 years. The severity of postoperative complications, based on the CD score, increased with patient age, ASA score, and perioperative blood loss. The 90-day mortality rate was 3.2%, and the rate increased with age: it was 5.9% among patients aged ≥80 years. In patients that survived the first 90 days, the rates of five-year relative survival, local recurrences, and metastases were independent of age.

All patients
The incidence of rectal cancer has increased since the 1980s, at both the global and national levels. The main reasons for this increase are an increasing human development index [18], an aging population [19], and an age-independent approach to the diagnostic work-up of suspected cancer. We observed a successive increase in the rectal cancer incidence during the study period and a trend towards an increase in the rate of patients aged ≥80 years. Despite scarce evidence and a demand for knowledge, older patients are frequently excluded from clinical trials [20]. The present study included an unselected consecutive series of all patients treated for rectal cancer at a local hospital during nearly four decades, with a focus on patients aged ≥80 years. It has been well documented that inequities concerning rectal cancer treatment occur across age groups [21]. The optimal treatment for an individual patient is based on a complete staging of the disease. In the present series, the rate of patients with an unknown stage declined over time, and it was low compared to other series [22]. Tumour stages were evenly distributed across age groups, consistent with previous reports [23]. Although the disease stage is typically the defining determinant in treating younger patients, factors associated with increasing age highly influence treatment options in older patients [24].
In Norway, a standardized diagnostic work-up applies to all patients with rectal cancer [2]. It culminates in a summary meeting of a multidisciplinary team (MDT), where treatment options are considered in detail, based on diagnostic findings and the defined stage of disease. A thorough, objective evaluation of the patient's functional and physiological status and the patient's personal preferences regarding treatment are not emphasized in routine care; however, adding these features to routine care would constitute a major improvement in guidance for making decisions for these patients [25].
A non-curative treatment approach was applied to 28.7% of the patients in this study, consistent with previously reported 25-30% rates for incurable disease at diagnosis [1]. Despite a similar stage distribution between age groups at diagnosis, the rate of patients that underwent a non-curative treatment increased with age. Among patients aged ≥80 years, 42% underwent non-curative treatments. Only 17% of the older patients had verified stage IV disease at diagnosis, but 25% underwent non-curative treatment with an unknown stage of disease or a potentially resectable disease. Limitations regarding treatment in older patients are related to the coinciding peak incidences of co-morbid diseases, cognitive impairments, and physical impairments [26]. In the present study, objective measures of co-morbidity, ASA, and CCI scores increased significantly with age.
Palliative resection procedures were more common in younger age groups; the older patients more frequently underwent best supportive care. The overall rate of chemotherapy was 27.1%, and it declined substantially with patient age. Individualized treatment regimens may be well tolerated in older patients with good performance status, hence chronological age should not preclude these patients from chemotherapy [27,28].
The overall rate of procedures with curative intent was 71.3%. The rate of major resections was 65.1% and the rate of polypectomies was 6.2%. These rates were comparable to the major resection rates of 59.9-70.8% reported recently in an evaluation of Scandinavian and English patients with rectal cancer during 2010-2012 [22]. The present study found a resection rate of 66.3% in our Norwegian population. Resection rates decline consistently with increasing age, and they have varied substantially between countries, despite comparable treatment guidelines. In the present study, 47.2% of patients aged ≥80 years underwent a major resection with curative intent. In comparison, Swedish patients aged >75 years had resection rates of 61.9%, and English patients aged >75 years had resection rates of 45.7%.
The overall 90-day mortality rate was 9.6%. It increased with age, but decreased significantly throughout the study period. Among patients that survived the first 90 days in this series, the five-year relative survival rate was 66.3%. The rate decreased from 72.9% in patients aged <65 years to 48.3% in patients aged ≥80 years. The overall five-year relative survival rates for Norwegian patients with rectal cancer have increased successively over the years, from 43.8% during 1980-1984 to 72.4% during 2016-2020 [1,29]. Comparable rates during 2012-2016 have been reported in the other Nordic countries and the Netherlands [30,31].
Selecting the appropriate individualized treatment for rectal cancer is a major challenge in efforts to reduce morbidity and increase survival. The adverse effects of over-treatment may cause unnecessary harm, but under-treatment may reduce survival. Older patients with reduced physiological reserves are particularly prone to the adverse effects of cancer treatment, regardless of whether the approach is curative. This dilemma is reflected by differences in treatment rates, and it underlines the need for additional improvements in the treatment selection process for this group of patients.

Patients with stages I-III disease treated with a major resection with curative intent
Overall, an anterior resection was the most common procedure, with a rate of 64%. Patients aged ≥80 years had lower anterior resection rates (47%) and were more frequently treated with an HP (34%). This observation was consistent with previous findings [32,33]. The main advantage of a HP is that it avoids an anastomosis, which eliminates the potentially fatal effects of an anastomotic leak. Among older patients with reduced tolerability for surgical complications, the HP stands out as a safe choice. In this series, only three patients underwent surgery with a laparoscopic approach. Minimally invasive surgery should be considered for older patients as previous studies have demonstrated comparable postoperative outcomes as in younger patients [34].
A Swedish study that examined the postoperative outcome of an HP for rectal cancer found an overall HP rate identical to that found in the present study (11%). They reported that the HP was performed predominantly in older patients (mean age 79 years) with increased comorbidities, elevated ASA scores, and a poor WHO performance status [35]. In fragile patients, HP has the benefits of a shorter operative time, less bleeding, and a lower rate of serious complications, compared to other treatments. However, we lack evidence that clearly favours either the anterior resection or the HP [36]. A substantial number of patients that undergo anterior resections experience low anterior resection syndrome (LARS) [37]. However, the adverse effects of an endcolostomy due to a HP are well-documented [38].
Throughout the study period, in older individuals, the procedure of choice was increasingly an HP. The frequency of HPs increased from 3.7%, in the first observation period, to 22.8% in the last observation period. The increasing use of an HP over time was also observed previously by other authors [39,40]. The increasing proportion of older patients with comorbidities over time in our cohort might partly explain this observation. However, because the HP rate increased faster than the increasing proportion of older patients over time, our findings also indicated that there was a general trend towards an increased use of HP.
Following rectal cancer surgery, older patients are encumbered with considerable morbidity, ranging from acute infectious complications to permanent functional derangements [41]. Nevertheless, a limited number of studies have addressed complications in this group of patients [42]. Complication rates depend on patient selection for the study, the rate of emergency surgery, the stage of disease, and the level of the institution. The rate of complications in the present study was considerable (47.6%). Although high, this rate was comparable to rates found in previous studies (21-61%) [32,43]. Rates of major complications (CD ≥3) and anastomotic leaks were also comparable to those found in previous studies [42]. Complication rates did not differ significantly between patients under and over age 80 years, except for pneumonia, which was more common in older patients. Measures to prevent pneumonia in patients that require surgery have been shown to be effective [44], and should be considered routine care in older patients that undergo rectal cancer surgery. The elevated rate of severe complications that we observed in older patients highlighted their reduced capacity to withstand adverse postoperative events.
The rate of reoperations increased significantly during the study period, from 7.3% during 1980-1989 to 18.4% during 2010-2016. In comparison, a 2011 report of nearly 250,000 English patients observed a reoperation rate of 7.4% [45]. The increasing number of older patients with high ASA categories in recent years might have contributed to this observation. The increasing use of HP was likely an attempt to counterbalance the risk of severe complications that might require reoperations. The number of surgeons that performed rectal cancer surgeries increased throughout the study period; this factor may have adversely impacted the rate of postoperative morbidity, due to the complexity of these procedures. Rectal cancer surgery should be applied by highly experienced teams and in concordance with the latest knowledge. Previous studies that evaluated associations between complication rates and treatment volumes have shown conflicting results [46,47].
The 90-day mortality in patients aged ≥80 years that underwent a major resection with curative intent was 5.9%, compared to an overall 90-day mortality of 3.2%. These rates were low, compared to rates reported previously [48,49]. These relatively low rates could indicate that the selection of individuals fit for surgery was appropriate in this series. The five-year relative survival rate for all patients that underwent a major resection with curative intent was 84.2%. In comparison, the Norwegian national relative five-year survival rates for localized, regional, and metastasized rectal cancer during 2016-2020 were 98.2%, 81.5%, and 22.4%, respectively [1]. In contrast, the relative survival for patients that underwent surgery for stages I-III disease was 88.5% during 2016-2020 [50]. Previous reports that compared relative survival across age groups of patients treated with curative intent have shown acceptable long-term survival rates among older patients [32,[51][52][53]. Therefore, resection surgery should not be withheld based on chronological age.
In the present study, the overall five-year local recurrence rate was 7.3%. During the 1990s, treatment guideline violations, reflected by a low (0.9%) rate of preoperative radiotherapy, resulted in a high local recurrence rate (18.5%) and an adverse relative survival rate (72.0%) for that period [4]. During the two later timeperiods, local recurrence rates declined in parallel with an increase in relative survival, as the rate of preoperative radiotherapy increased [54]. The increasing use of radiotherapy observed in the present study was also observed at a national level [55]. The estimated five-year metastasis rate after an R0 resection was 19.5%. This rate was comparable to the national rates (20.2-22.1%) during 2006-2020, for patients that underwent resections for stages I-III disease [50].

Strengths and weaknesses
The main strength of the present study was the inclusion of a large number of consecutive patients treated for rectal cancer at a local hospital, in accordance with evidence-based guidelines. Another strength was the longterm observation period of 37 years. Our institution was the primary hospital for a stable population throughout the observational period, and the population was suitable for evaluating trends over time [56]. Complications for each patient were retrieved from hospital records. Norwegian referral policies are practically age-independent; hence, we believe only a small number of patients was not included in the scope of the current report.
The main limitation of the study was its retrospective design. Unknown or unrecorded confounders might have affected decisions regarding patient selection and treatment. An unknown number of complications may have gone unnoticed, and thus, were not included in the hospital records, especially during the earlier years of the study period. Consequently, the numbers of complications presented in this report must be viewed as minimums. A number of patients (n=51) in our catchment area underwent treatment at other institutions and were excluded from this study. In the excluded group, the 90-day mortality and the five-year relative survival rate among those that survived the first 90 days were nearly identical to those observed in the cohort included in this study. However, the addition of these 51 patients to our study cohort might have altered some of the results.

Future perspectives
The number of older patients with rectal cancer is predicted to escalate in the years to come. This escalation will increase the burden on healthcare systems, at both the national and global levels. Improvements in selecting and treating older patients with rectal cancer might enhance results and optimize the utilization of healthcare resources.
Prehabilitation is gaining interest in the surgical milieu and aims to enhance the individual patient's starting point prior to surgery. Currently, studies have been investigating the potential of prehabilitation in patients with rectal cancer, and the results may impact the future treatment of older patients [57,58].
Studies on the effect of age on morbidity have produced conflicting results [59,60]. Our observation that more severe complications occurred with increasing age may partly be explained by a higher proportion of frail patients in the oldest age groups compared to younger age groups. Frailty may be present in the absence of comorbid conditions, and it could be a factor in 25-46% of patients over 65 years old that undergo surgery for colorectal cancer [61]. The impact of frailty in patients undergoing surgery for colorectal cancer has been investigated, and it should be emphasized in future clinical practice [62,63]. Due to the increasing proportion of older patients with rectal cancer, we believe that a comprehensive geriatric assessment should be included as part of the routine work-up.
Physicians may be forced to reconsider treatment aims in older patients, because this group of patients is likely to choose functional status above survival [64]. This choice interferes with one of the most fundamental principles in treating patients with cancer. Moreover, as patients approach the end of life, their personal preferences regarding medical treatment might be more decisive than ever before.

Conclusion
This study showed that patients aged ≥80 years were less likely to undergo a major resection with curative intent compared to younger patients, despite comparable disease stages. The rate of complications following rectal cancer surgery was high across all ages, but the severity of complications increased with age. Patients aged ≥80 years that underwent a major resection with curative intent had long-term survival rates comparable to their younger counterparts. The future care of older patients with rectal cancer demands highly specialized teams that can focus on the distinctive demands in this specific group of patients.  Best supportive care 2 0 (0) 0 (0) 2 (13) 0 (0)