The diagnostic value of a change in bowel habit for colorectal cancer within different age groups

Abstract

Background

Change in bowel habit as a sole alarm symptom for colorectal cancer is disputed.

Objective

We investigated the diagnostic value of change in bowel habit for colorectal cancer, particularly as a single symptom and within different age groups.

Methods

This retrospective cohort study examined colorectal cancer fast track referrals and outcomes across four Swedish hospitals (April 2016–May 2017). Entry criteria constituted one or more of three alarm features: anaemia, visible rectal bleeding, or change in bowel habit for more than 4 weeks in patients over 40 years of age. Patients were grouped as having only change in bowel habit, change in bowel habit plus anaemia/bleeding or anaemia/bleeding only.

Results

Of 628 patients, 22% were diagnosed with colorectal cancer. There were no cases of colorectal cancer in the only change in bowel habit group under 55 years, while this was 6% for 55–64 years, 8% for 65–74 years and 14% for 75 years and older. Among subjects under 55 years, 2% with anaemia/bleeding had colorectal cancer, this increased to 34% for 55 years and older (P < 0.0001). Change in bowel habit plus anaemia/bleeding gave a colorectal cancer prevalence of 16% in under 55 years and increased to 30% for 55 years and older (P = 0.07).

Conclusion

Change in bowel habit as the only alarm feature has a low diagnostic yield for colorectal cancer in patients under 55 years.

Key summary

• Summarise the established knowledge on this subject:  Change in bowel habit as a sole alarm symptom for colorectal cancer is disputed.

• What are the significant and/or new findings of this study? Change in bowel habit as the only alarm feature has a low diagnostic yield for colorectal cancer in patients under 55 years.

Introduction

Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer death in Europe,¹ with a predicted 98,000 deaths in 2018.² Overall, CRC has a higher prevalence in men than in women, and the likelihood of developing the condition increases markedly after 50 years of age.³ Although 5-year survival rates for CRC are over 90% when diagnosed early, they decrease considerably with cancer stage progression.^4,5 In an attempt to establish earlier diagnoses, health authorities across Europe have implemented different national screening programmes and symptom-based investigative pathways for CRC. Although screening programmes are highly effective for early diagnosis and prevention,^6,7 the majority of CRC patients still encounter healthcare services only after symptom manifestation.⁸

Presenting features of CRC are often non-specific and vary considerably between individuals.⁹ The most common symptoms and clinical features include weight loss, anaemia, abdominal pain, rectal bleeding and change in bowel habit (CIBH).¹⁰ Unexplained visible rectal bleeding and iron deficiency anaemia are well-established CRC alarm features that often lead to prompt referrals for colonoscopy.^11–13 However, CIBH as a sole trigger for fast track colonoscopy evaluation is disputed: a 2008 meta-analysis and a 2011 systematic review found overall poor positive and negative likelihood ratios for CIBH, with a statistical heterogeneity of results between the included studies.^11,14 The American College of Gastroenterology does not recommend performing colonoscopies on non-bleeding patients with CIBH.¹³ In the United Kingdom, the National Institute for Health and Care Excellence recommends all patients aged 60 years or older with CIBH to get a prompt colonoscopy.¹⁵ However, in Sweden, Norway and Denmark, a lower threshold has been chosen with regard to CIBH; these Scandinavian countries recommend fast track colonoscopy for CIBH of over 4 weeks in patients over 40 years of age.^16–18 The discrepancies between country guidelines may be due to the lack of conclusive evidence for CIBH as an alarm symptom for CRC, as well as the non-specific nature of the symptom; indeed, CIBH can also be caused by common functional bowel disorders, such as irritable bowel syndrome (IBS).¹⁹ Moreover, given that IBS commonly affects young to middle-aged adults, the Scandinavian guidelines may lead to such individuals being unnecessarily fast-tracked for a colonoscopy.

With this in mind, we aimed to appraise the Swedish colonoscopy fast track referral system, in particular with regard to the diagnostic yield of CIBH for CRC within different age categories. We hypothesised that there would be a low diagnostic yield for CRC in patients presenting with CIBH only, i.e. without anaemia and/or bleeding, who were under 55 years of age.

Materials and methods

Study design

We performed a retrospective cohort study on all CRC fast track patients referred over a one-year period to the three endoscopy units at the Sahlgrenska University Hospital and to Kungälv Hospital (from May 2016 to May 2017). These hospitals receive patients from both urban and rural districts of Västra Götaland, a region in southwestern Sweden.

The study design was based on the Swedish national fast track pathway for CRC. Three alarm features prompt further investigation according to this pathway: anaemia, visible rectal bleeding and/or CIBH for over 4 weeks in patients over 40 years of age without other explanation. Of these three alarm features, only CIBH leads to an immediate referral for colonoscopy; in contrast, patients with only anaemia and/or visible rectal bleeding must first undergo outpatient clinic evaluation and are referred for fast track colonoscopy if deemed appropriate, i.e. if no obvious cause is found on, for example, rigid sigmoidoscopy (Figure 1). Patients can also enter the pathway through findings on medical imaging, rigid sigmoidoscopy or digital rectal examination suggestive of CRC;¹⁸ however, those patients who did not have anaemia, visible rectal bleeding or CIBH were not included in this study.

Figure 1.

Flowchart of the Swedish colorectal cancer (CRC) fast track pathway, with anaemia, visible rectal bleeding and change in bowel habit (CIBH) for over 4 weeks in patients aged over 40 years as the three entry criteria. High-risk patients are those with one or more of the following: have had CRC or are part of a surveillance programme for colorectal polyps, have had colitis for at least 20 years, have two first-degree relatives with CRC, have a relative who has been diagnosed with CRC or endometrial cancer when less than 50 years old. Not included in the flow chart or in this study are those patients who entered the pathway with only medical imaging, rigid sigmoidoscopy or digital rectal examination findings suggestive of CRC.

Study population and data collection

To be included in the study, patients had to have entered the CRC fast track pathway with one or more of these three alarm features. Furthermore, included patients also had to have undergone subsequent colonic investigations. For a thorough registration of symptoms, data were collected from diagnostic coding, a laboratory database, referral letters and medical records. Information on the following was documented: anaemia, visible rectal bleeding, change in stool form/frequency/colour, digital rectal examinations, rigid sigmoidoscopies, fatigue, drop in haemoglobin values albeit still within the normal range, positive fecal haemoglobin, palpable abdominal mass, unintentional weight loss, family history of CRC and abdominal/rectal pain. Subsequently, information was also collected on CRC diagnoses, including tumour locations, all registered non-CRC pathologies and histology results.

For the purpose of this study, all patients with any type of anaemia recorded – either mentioned in the referral or in the laboratory database within a month of referral as a haemoglobin count below the normal range (haemoglobin <130 g/L for men, <120 g/L for women) – were classified as anaemic. Furthermore, changes in stool form or frequency that had persisted for more than 4 weeks were categorised as CIBH. Although only patients over 40 years of age could be referred to the fast track pathway on the basis of CIBH alone, some patients referred with anaemia and/or visible rectal bleeding under 40 years of age had CIBH for more than 4 weeks recorded in referrals. These younger patients were, therefore, also registered as having CIBH for more than 4 weeks for this study.

To investigate CIBH as a trigger for CRC evaluation, the patients were categorised into three groups: (a) those with no CIBH, but with anaemia and/or visible rectal bleeding (anaemia/bleeding-only group); (b) those with CIBH in combination with anaemia and/or visible rectal bleeding (CIBH + anaemia/bleeding group); and (c) those with CIBH as the only criteria (CIBH-only group). The groups were also stratified according to age (<55, 55–64, 65–74 and ≥75 years) to determine its effect on the diagnostic value for CRC.

Data analysis

Statistical analysis was carried out using SPSS version 25.0 software (SPSS Inc., Chicago, IL, USA), with significance set at a P value of less than 0.05. Categorical variables were summarised by descriptive statistics, including total numbers and percentages, with comparisons between groups performed using the chi-square test.

Results

Patient identification

Over the one-year period, 698 individuals were referred to the fast track CRC pathway, of which 70 patients were excluded for various reasons, thereby leaving 628 patients eligible for further analysis (Figure 2).

Figure 2.

Flowchart of patients referred to the Swedish fast track colorectal cancer pathway.

Baseline characteristics

The mean age of the overall cohort was 67 years (range 22–95 years), with 55% being women. Of these, 62% (n = 390) had CIBH, 39% (n = 244) had anaemia and 46% (n = 291) had visible rectal bleeding. On further stratification according to the presence or absence of CIBH, the proportion of individuals with CIBH only was 28% (n = 174), with the CIBH + anaemia/bleeding group representing 34% (n = 216) and the anaemia/bleeding-only group representing 38% (n = 238). Full baseline characteristics of the overall cohort and individual subgroups are detailed in Table 1, including the proportion of individuals within each age category. Of interest, the proportion of individuals with CIBH only who were under 55 years of age was 17% (n = 30), which accounted for 5% (n = 30/628) of the entire fast track cohort, and the CIBH-only group had a higher proportion of women than the other groups.

Table 1.

The baseline characteristics of the overall cohort and clinical subgroups.

	Overall (n = 628; 100%)	CIBH only (n = 174; 28%)	CIBH + anaemia/ bleeding (n = 216; 34%)	Anaemia/bleeding (n = 238; 38%)
Women	347 (55.3%)	112 (64.4%)	110 (50.9%)	125 (52.5%)
Mean age (SD)	67 (13.3)	67.4 (12.1)	67.5 (14.3)	65.8 (13.1)
Age range, years	22–95	41–95	22–91	22–93
Age category
Under 55 years	117 (18.6%)	30 (17.2%)	43 (19.9%)	44 (18.5%)
55–64 years	120 (19.1%)	31 (17.8%)	44 (20.4%)	45 (18.9%)
65–74 years	205 (32.6%)	62 (35.6%)	75 (34.7%)	68 (28.6%)
Over 75 years	186 (29.6%)	51 (29.3%)	54 (25%)	81 (34%)

CIBH: change in bowel habit.

Prevalence of CRC

Following gastrointestinal investigations, 22% (n = 139) of patients were diagnosed with CRC. The prevalence of CRC under the various clinical presentations is detailed in Figure 3. For the clinical subgroups of interest, the highest prevalence of CRC was seen in the anaemia/bleeding-only group (n = 66/238; 28%) and, to a similar extent, the CIBH + anaemia/bleeding group (n = 59/216; 27%), while the CIBH-only group had the lowest prevalence of CRC (n = 14/174; 8%) (P < 0.0001).

Figure 3.

Colorectal cancer (CRC) prevalence by each symptom group – all ages. The combination of anaemia, bleeding and change in bowel habit (CIBH) had the highest prevalence of CRC (46%). For single symptoms, CRC prevalence was highest for anaemia (34%), followed by visible rectal bleeding (20%) and lowest for CIBH (8%).

We further analysed the prevalence of CRC in each clinical subgroup based on age category (Figure 4). While 2% of anaemia/bleeding-only patients under the age of 55 years were diagnosed with CRC, this increased to 34% for those aged 55 years and older (P < 0.0001). The presence of CIBH in association with anaemia/bleeding gave a prevalence of CRC of 16% in those under 55 years of age and increased to 30% for those aged 55 years and older (P = 0.07). There were no cases of CRC in the CIBH-only group who were under the age of 55 years, while there were 14 cases (10%) in those over 55 years of age (P = 0.075), the prevalence being 6% for those aged 55–64 years, 8% for those aged 65–74 years and 14% for those aged 75 years and older.

Figure 4.

Colorectal cancer (CRC) prevalence by symptom group and age category. There were no cases of CRC in the change in bowel habit (CIBH)-only group who were under 55 years, while this was 9.7% in those aged 55 years and older (P = 0.075). Moreover, while 2% of anaemia/bleeding subjects under 55 years were diagnosed with CRC, this increased to 33.5% for those aged 55 years and older (P < 0.0001). Finally, CIBH + anaemia/bleeding gave a CRC prevalence of 16% in those under 55 years, and increased to 30.1% for those aged 55 years and older (P = 0.07).

Of the 14 CIBH-only patients with CRC, the average age was 72 years, with 64% being women. The CRC was located in the left colon in 64% of cases (n = 9; five rectal, three sigmoid and one descending) and the right colon in 36% of cases (n = 5; four caecum and one transverse). Alongside the CIBH, the majority (93%) had one or more of the following clinical features: positive fecal haemoglobin, drop in haemoglobin within the reference range, palpable abdominal mass, abdominal/rectal pain, family history of CRC, palpable mass on digital rectal examination, unintentional weight loss, stools darker than usual, fatigue, palpable hepatomegaly and/or bloating; further details are provided in Table 2.

Table 2.

Clinical features of CIBH-only patients with CRC.

Age (years)	Sex	Non-CIBH symptoms	Cancer location
55	Male	None	Rectum
58	Female	Positive fecal haemoglobin, fatigue, bloating	Rectum
66	Female	>20 g/L drop in haemoglobin but not anaemic, palpable hepatomegaly	Transverse colon
67	Female	Anal pain, palpable mass on digital rectal exam	Rectum
67	Male	Positive fecal haemoglobin	Rectum
68	Female	Palpable left iliac fossa mass	Sigmoid colon
68	Female	Unintentional weight loss	Sigmoid colon
76	Female	Positive fecal haemoglobin, >10 g/L drop in haemoglobin but not anaemic, abdominal pain	Caecum
76	Male	Abdominal pain	Sigmoid colon
77	Male	Unintentional weight loss	Rectum
79	Male	Stools darker than usual	Descending colon
80	Female	Abdominal pain, stools darker than usual, brother and father had colon cancer, unintentional weight loss, fatigue	Caecum
82	Female	Unintentional weight loss, fatigue	Caecum
82	Female	Abdominal pain, weight loss	Caecum

CIBH: change in bowel habit; CRC: colorectal cancer.

Pathologies in CIBH-only patients without CRC

Of the 160 CIBH-only patients without CRC, 36% (n = 58) had no other organic pathology to account for their CIBH. The other 64% (n = 102) of patients had one or more of the following: adenomatous polyp of low-grade dysplasia, angiodysplasia, diverticulosis, haemorrhoids, microscopic colitis or unspecified colitis. The prevalence of these organic pathologies is illustrated in Figure 5. Of the 30 patients under 55 years of age with CIBH only, 53% (n = 16) had no evident organic pathologies to account for their symptoms, whereas the remaining 47% (n = 14) of patients had one or more: five (17%) patients had low-grade dysplastic adenomas, five (17%) had diverticulosis, five (17%) had unspecified colitis, four (13%) had haemorrhoids and one (3%) had collagenous colitis. Furthermore, in the CIBH-only patient group without CRC, two subjects were diagnosed with pancreatic cancer, one with kidney cancer and one with breast cancer within a year after entry into the fast track pathway – all were above the age of 55 years.

Figure 5.

Diagnostic outcomes in change in bowel habit (CIBH)-only patients without colorectal cancer (CRC) (all ages). In the CIBH-only patient group without CRC, diverticulosis (31%) and adenomatous polyps (25%) were the most prevalent pathologies, followed by haemorrhoids (10%) and unspecified colitis (8%). Microscopic colitis (3%) and angiodysplasia (2%) were less prevalent. Thirty-six per cent of these patients had no organic pathology to account for their symptoms.

Discussion

Through analysing CRC fast track referrals in Sweden, we evaluated the diagnostic yield of CIBH for CRC. We reported on the association between CRC and CIBH as a single symptom, or in combination with anaemia and/or visible rectal bleeding, within different age groups. We found that CIBH only had no CRC cases among patients under 55 years old, and that the likelihood of having CRC within the CIBH-only symptom group increased with age. Our results are in line with a UK case–control study that found a relatively low positive predictive value for CIBH in those aged under 60 years compared to older age groups.¹⁰

Our study brings into contention the current CRC pathways observed in Scandinavia, where patients over 40 years of age presenting with a CIBH only for more than 4 weeks are recommended to undergo fast track colonoscopies.^16–18 Given the lack of diagnostic yield for CRC seen in our study for this group, we would suggest the guidelines be appraised with a view to increasing the age limit. A higher age limit would also be more in line with guidelines of other national societies, such as the National Institute for Health and Care Excellence from the UK that advises fast track referrals for CIBH-only patients aged 60 years or older.¹⁵

Moreover, changing the age threshold from 40 years to 55 years may have beneficial implications on healthcare service provision. For example, 5% of all fast track referrals comprise subjects between 40 and 55 years of age presenting with CIBH only, which amounts to a considerable expenditure of colonoscopy workload if the results of our multicentre study are representative of all of Sweden; in the year 2017, 17,235 patients entered the national fast track pathway for CRC in Sweden,²⁰ of which 5% would be about 862 patients. The number of patients becomes even more substantial when applied to the populations of Norway and Denmark as well. Furthermore, unwarranted colonoscopies in the fast track pathway expose low-risk patients to unnecessary invasive examinations and the potential emotional distress of undergoing a cancer investigation.²¹ Importantly, it may also delay colonoscopies for patients who are actually at high risk of having CRC, such as patients with visible blood and/or anaemia and patients undergoing colonoscopy surveillance due to long-standing colitis or polyps.

Furthermore, one can speculate that Swedish-based physicians may be aware of the low diagnostic yield for CRC in patients under 55 years with CIBH only, and may not consistently be referring these patients into the fast track pathway. This assumption would be based on the fact that over the course of one year, and four hospitals, there were only 30 cases out of 628 fast track CRC referrals that comprised individuals with CIBH only under 55 years. Yet this value of 30 cases pales into insignificance when considering that CIBH only is a common symptom within that age group of the general population; in fact, change in stool form or frequency are among the diagnostic criteria for IBS, which affects about one in 10 people,²² predominantly young to middle-aged adults,²³ and is a frequent cause for seeking healthcare.^24,25 The Rome Foundation promotes a diagnosis of IBS based on symptom criteria and the exclusion of organic pathology in a judicious manner, with colonoscopy being reserved for those with alarm features (such as bleeding, anaemia, unintended weight loss or onset of symptoms above the age of 50 years) given its otherwise low diagnostic yield.²⁶ However, according to the Scandinavian CRC fast track guidelines, all those over 40 years of age who have IBS, as defined by the Rome IV diagnostic criteria,^18,19 and irrespective of alarm symptoms, should get an urgent referral for a colonoscopy.¹⁸ Hence, it would have been anticipated that the CRC fast track pathway would be inundated and overwhelmed with such referrals, although this was not the case. This suggests that Swedish clinicians are either not adhering to the guidelines, presumably because they are aware that IBS and other functional gut disorders are very common in younger patients, or are perhaps not fully aware of the current guidelines given that they were introduced shortly before the study was initiated.

Although CIBH-only patients aged less than 55 years may not be suited for the fast track colonoscopy referrals, they should not be ignored. Indeed, almost half of this patient group had other organic pathologies. Furthermore, some may have functional bowel disorders, which would not be apparent from colonoscopic or histological evaluations. In summary, patients who present with CIBH only aged less than 55 years do not appear to be suited for fast track colonoscopy, and instead should be considered for routine healthcare evaluation after seeing their family practitioner, which in certain cases may include a non-fast track colonoscopy, particularly if the patient has treatment-refractory diarrhea.^19,26,27

Moreover, previous studies have shown that CIBH in combination with alarm features, such as rectal bleeding, has a higher yield of CRC than either symptom alone.^11,28 We found that this was particularly true for patients aged under 55 years, for whom there was a significantly higher prevalence of CRC in the CIBH + anaemia/bleeding group than the anaemia/bleeding-only group or the CIBH-only group. This finding suggests that CIBH could be of particular importance in identifying young individuals at risk of CRC when the symptom presents in combination with alarm symptoms.

Furthermore, of the CIBH-only patients who were diagnosed with CRC, 93% had other clinical features suggestive of a possible malignancy, such as a family history of CRC, abnormal abdominal examination, positive fecal haemoglobin or a drop in haemoglobin, albeit still within the normal range. Indeed, only one patient with CIBH only who was diagnosed with CRC actually had CIBH as the only symptom. Of note, the CRC found in CIBH-only patients was spread throughout the lower gastrointestinal tract, with two-thirds located in the left colon and one-third in the right colon. This implies that when endoscopic evaluation is required to investigate further the symptoms of CIBH only then a colonoscopy should be preferred over a limited flexible sigmoidoscopy.

Finally, we note that the overall prevalence of CRC in the Swedish fast track pathway of 22% is far greater than that documented elsewhere within the international literature.^29,30 For example, a recent systematic review with meta-analysis of over 90,000 fast track referrals in the UK had a CRC diagnostic yield of only 7.7% (95% confidence interval 6.9–8.5%).³⁰ Our study was not designed to explain any international discrepancies in CRC prevalence when entering fast track pathways, and neither can it be explained by differences in age standardised incidence rates of CRC between Sweden (26.9/100,000) and the UK (32.1/100,000).³¹ We can only speculate that it may, in part, be due to Swedish family practitioners not being fully compliant in following the fast track referral pathway but rather being more selective in the cases being referred, which adds to our earlier comment on the unexpectedly low number of fast track referrals for CIBH alone in those aged less than 55 years.

A strength of this study was that we included all patients referred through the CRC fast track pathway to four different hospitals/units in Sweden, and believe that this multicentre study population is representative of Sweden as a whole and the neighbouring countries of Norway and Denmark. A limitation of this study was that not all factors associated with CRC were included in the analysis, such as weight loss, family history of CRC, fecal haemoglobin, abdominal mass or pain.^10,32–34 This limitation was due to us basing our study on the three primary entry criteria for the Swedish fast track pathway, and the fact that this information was not available in all patients in the medical records. Furthermore, our sample size of 628 patients only resulted in 30 patients who entered the pathway with CBH only at under 55 years, which is a relatively small number of patients, but nevertheless does corroborate with data from a large UK general practice registry reporting a low diagnostic yield of CRC in patients under 50 years with CIBH only, in which the positive predictive value for CRC following a single or repeated presentation with CIBH only was 0.5% and 1.2%, respectively.³⁵ Moreover, this study was initiated when the national guidelines were first implemented, and all physicians may not have been aware of the guidelines at the time.

In conclusion, CIBH, without anaemia or visible rectal bleeding, has a low diagnostic value for colorectal cancer in patients under 55 years of age. This would suggest that such patients should not be referred under the fast track CRC pathway for an immediate colonoscopy, as currently proposed by Scandinavian guidelines, but rather should be considered for routine healthcare evaluation instead.

Following the dissemination of our study findings within Sweden, the CRC fast track pathway has been altered during 2019, whereby the cut-off age for fast track colonoscopy for patients with CIBH only in Sweden has increased from 40 to 50 years.³⁶

Declaration of conflicting interests

The authors have no relevant conflict of interests to declare for this study.

Ethics approval

The regional ethical review board in Gothenburg approved the study protocol on 12 June 2017 (reference number 506-17).

Funding

This study was supported by the Swedish Medical Research Council (grants 13409, 21691 and 21692) and by the Faculty of Medicine, University of Gothenburg.

Informed consent

As this was a retrospective review of the medical records, obtaining informed consent from the included individuals was not deemed necessary and not requested by the ethical review board.