Family Practice Advance Access originally published online on November 30, 2006
Family Practice 2007 24(1):3-6; doi:10.1093/fampra/cml059
Clinical features of colorectal cancer before emergency presentation: a population-based case–control study
Academic Unit of Primary Health Care, Department of Community Based Medicine, University of Bristol The Grange, 1 Woodland Road, Bristol BS8 1AU, UK
Correspondence to Dr William Hamilton, Email: w.hamilton{at}bristol.ac.uk
Received 25 July 2006; Accepted 17 October 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Objective. To identify the clinical features of colorectal cancer presenting as a surgical emergency.
Design. Population-based case–control study.
Setting. All general practices in Exeter Primary Care Trust, Devon, UK.
Participants. 349 patients with colorectal cancer, 62 of these having an emergency presentation. Five randomly selected controls matched by age, sex and general practice for each case.
Data. The entire primary care record, from 24 months to 30 days before diagnosis, was coded using the International Classification of Primary Care-2.
Main outcome measures. Symptom reporting by patients with emergency presentation of colorectal cancer compared with matched controls and non-emergency presentations.
Results. Eight features of colorectal cancer were associated with the 62 emergency presentations of colorectal cancer. 39 (63%) of patients had reported at least one symptom to their doctors a minimum of 30 days before the diagnosis. In multivariable analysis, three features remained independently associated with cancer: abdominal pain, odds ratio 6.2 (95% CI 2.8–14), P < 0.001; loss of weight 3.4 (1.3–8.5), P = 0.01; and diarrhoea 3.4 (1.2–5.7), P = 0.02. When emergency presentations were compared with elective cases, abdominal pain was more common [interaction odds ratio 2.3 (1.6–3.3); P = 0.047] and rectal bleeding less common [0.30 (0.08, 1.0); P = 0.040].
Conclusion. The majority of patients destined to have an emergency presentation of colorectal cancer have reported symptoms of their cancer to their doctor well before the emergency. Some emergency presentations should therefore be preventable.
Keywords. Colorectal cancer, diagnosis, emergency admissions, primary health care.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Colorectal cancer is common worldwide, with over 30?000 new cases in the UK annually.1 Between 17 and 29% of patients present with a surgical emergency, usually bowel obstruction.2–5 Mortality is higher in these patients.6 Earlier diagnosis may be beneficial by preventing emergency presentations. This was observed in a screening study.7
Three studies have interviewed patients presenting as an emergency about their recent symptoms: in two, most patients were asymptomatic before the emergency, and the median duration in those who did have symptoms was 3 months.8,9 Most patients in the third study described symptoms, with a median duration of 10 weeks, but these symptoms had not necessarily been presented to primary care.10 Furthermore, symptom reporting in all three studies was dependant upon the patient's memory and not validated by other means, leading to concerns about possible recall bias.10
We examined two clinical questions. First, have patients reported symptoms of their cancer to their GP at least a month before their emergency presentation? Second, do emergency patients have a characteristic pattern of symptoms?
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Participants
Our previously described population-based case–control study identified 349 colorectal cancer cases diagnosed in 1998–2002 in patients aged 40 years or more from Exeter Primary Care Trust, Devon, UK. These were ascertained from the local cancer registry and computer searches at all 21 general practices.11 Each case was randomly matched by age, sex and practice to five controls.
For the present study, cases were categorised into emergency or elective presentations by JC and a research assistant. Emergency presentations were defined as requiring surgical admission for suspected bowel obstruction or perforation, with colorectal cancer diagnosed during the admission, almost always at laporotomy. Ambiguous cases (n = 6) were reviewed by WH.
Data collection
The primary care record for the 2 years preceding diagnosis was coded using the International Classification of Primary Care.12 The aim of this study was to identify factors possibly allowing earlier diagnosis, so the 30 days before the diagnosis was excluded from all analyses.
Analyses
The first analysis was of the 62 emergency cases compared with their matched controls, using univariable and multivariable conditional logistic regression. Another multivariable model was then constructed using all 349 cases and their controls. Differences between emergency and elective cases were examined by fitting interaction terms for each variable with the emergency/elective status of the case. Analyses were performed using Stata version 8.13
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Of the 349 patients with cancer, 62(18%) were emergencies, and 287(82%) elective. The age and sex profiles were similar: median age (inter-quartile range) in emergencies 75 (65,84) and in elective cases 73 (65,80) years—Mann–Whitney P = 0.65; 27 (44%) emergencies and 150 (52%) of the elective cases were male—chi-square P = 0.21.
Comparison of emergencies with matched controls
The eight variables present in more than 5% of cases or controls were all more common in emergencies than in controls (Table 1). Of the 62 emergencies, 39 (63%) had reported at least one of the five main symptoms of cancer (diarrhoea, constipation, abdominal pain, rectal bleeding or loss of weight) to their GP at least 30 days before diagnosis.
|
Comparison of emergency with elective cases
In the multivariable analysis comparing all 349 cases with their controls, six of the eight features were independently associated with colorectal cancer, constipation and abdominal tenderness being the exceptions. Rectal bleeding had an interaction odds ratio of 0.30 (95% CI 0.08–1.0) P = 0.040, meaning it was less common in emergencies; conversely, the interaction odds ratio for abdominal pain was 2.3 (1.6,3.3), P = 0.047, as it was more common in emergencies.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
This study shows that most patients with an emergency presentation of colorectal cancer had reported symptoms to their GP at least 30 days before the crisis. Three symptoms (loss of weight, abdominal pain and diarrhoea) were independently associated with this mode of presentation. Furthermore, abdominal pain was more common—and rectal bleeding less common—in emergency compared with elective presentations.
Symptoms
The most prominent finding was the prevalence of abdominal pain. The association with emergency presentation was strong and consistent, with odds ratios of 6 in the two main analyses, and the positive interaction in the analysis of emergency against elective cases. The risk of colorectal cancer in an adult reporting unexplained abdominal pain to primary care is 
1%.11 Identifying which patient has an underlying cancer is a difficult task for primary care clinicians, let alone which will have a surgical emergency. Nonetheless, our results imply that a GP should consider colorectal cancer in patients with unexplained abdominal pain. If the risk of colorectal cancer is deemed high enough to warrant referral, then patients with abdominal pain should receive the highest priority in investigation, as they are most at risk of developing a surgical complication.
Diarrhoea and weight loss were also strongly associated with cancer in emergencies. Delayed investigation by doctors has been reported with diarrhoea in particular.14 Our findings take this one step further: the diagnostic delay may worsen the prognosis in allowing an emergency to occur. The weight loss findings may represent an association with advanced disease, which is itself associated with an increased risk of obstruction.5,10
One symptom—rectal bleeding—occurred more often in elective cases. Emergency presentation is less common in rectal cancers than colonic cancers, rectal bleeding would be less frequent too. Additionally, rectal bleeding is a well-recognised feature of colorectal cancer, and so referral may be made without delay. Conversely, anaemia was more common in emergencies, reiterating the importance of prompt investigation of this finding.15
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
The key finding is that
60% of emergency surgical admissions with colorectal cancer have described at least one symptom of their cancer to their GP a month or more before the crisis. In theory, therefore, some of these emergencies are preventable. Prevention may be by earlier recognition of the possibility of cancer, especially in patients with abdominal pain, weight loss or diarrhoea. If cancer is deemed to be a possible cause, patients with these symptoms should be investigated rapidly. |
Acknowledgments |
|---|
Project funding from the Department of Health. The funding source had no role in study design, data collection, analysis or writing of the report. All authors had full access to all data, and take final responsibility for publication. WH was funded through his research practice (Barnfield Hill, Exeter) and RCGP/BUPA and NHS Fellowships. JC was funded by an academic registrarship made possible by the Severn and Wessex Deanery. The views expressed in the publication are those of the authors and not necessarily those of the Department of Health.
Ethical approval: North and East Devon Local Research Ethics Committee.
We wish to thank all 21 general practices in Exeter, the Dendrite personnel, Jackie Barrett, and the Patients and Practitioners Service Authority.
|
Notes |
|---|
Cleary J, Peters TJ, Sharp D and Hamilton W. Clinical features of colorectal cancer before emergency presentation: a population-based case–control study. Family Practice 2007; 24: 3–6.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
1 Cancer Research UK. (2004) CancerStats Monograph 2004(Cancer Research UK, London).
2 Barrett J, Jiwa M, Rose P, Hamilton W. (2005) Pathways to the diagnosis of colorectal cancer: an observational study in three UK cities. Fam. Pract. Advance Access published on November 14, 2005; doi:10.1093/fampra/cmi093.
3 Mella J, Biffin A, Radcliffe AG, Stamatakis JD, Steele RJ. (1997) Population-based audit of colorectal cancer management in two UK health regions. Colorectal cancer working group, royal college of surgeons of England clinical epidemiology and audit unit. Br J Surg 84:1731–1736.[CrossRef][Web of Science][Medline]
4 Trickett JP, Donaldson DR, Bearn PE, Scott HJ, Hassall AC. (2004) A study on the routes of referral for patients with colorectal cancer and its affect on the time to surgery and pathological stage. Colorect Dis 6:428–431.
5 Cuffy M, Abir F, Audisio RA, Longo WE. (2004) Colorectal cancer presenting as surgical emergencies. Surg Oncol 13:149–157.[CrossRef][Web of Science][Medline]
6 Tekkis PP, Kinsman R, Thompson MR, Stamatakis JD. Association of Coloproctology of Great Britain I. (2004) The Association of Coloproctology of Great Britain and Ireland study of large bowel obstruction caused by colorectal cancer. Ann Surg 240:76–81.[CrossRef][Web of Science][Medline]
7 Scholefield JH, Robinson MH, Mangham CM, Hardcastle JD. (1998) Screening for colorectal cancer reduces emergency admissions. Eur J Surg Oncol 24:47–50.[CrossRef][Web of Science][Medline]
8 Hargarten SW, Richards MJ, Anderson AJ, Roberts MJ. (1992) Cancer presentation in the emergency department: a failure of primary care. Am J Emerg Med 10:290–293.[CrossRef][Web of Science][Medline]
9 Mulcahy HE and O'Donoghue DP. (1997) Duration of colorectal cancer symptoms and survival: the effect of confounding clinical and pathological variables. Eur J Cancer 33:1461–1467.[CrossRef][Web of Science][Medline]
10 Olsson L, Bergkvist L, Ekbom A. (2004) Symptom duration versus survival in non-emergency colorectal cancer. Scand J Gastroenterol 39:252–258.[CrossRef][Web of Science][Medline]
11 Hamilton W, Round A, Sharp D, Peters T. (2005) Clinical features of colorectal cancer before diagnosis: a population-based case-control study. Br J Cancer 93:399–405.[CrossRef][Web of Science][Medline]
12 WONCA. (1998) ICPC-2. International Classification of Primary Care. Prepared by the Classification Committee of WONCA(Oxford University Press, Oxford).
13 StataCorp. (2001) Stata Statistical Software: Release 8.0. (Stata Corporation, College Station, TX).
14 MacArthur C and Smith A. (1984) Factors associated with speed of diagnosis, referral, and treatment in colorectal cancer. J Epidemiol Community Health 38:122–126.
15 Yates JM, Logan ECM, Stewart RM. (2004) Iron deficiency anaemia in general practice: clinical outcomes over three years and factors influencing diagnostic investigations. Postgrad Med J 80:405–410.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||