Background: Colorectal cancer (CRC) is the third most common cancer in men and the second most common in women worldwide. Adenomas can be precursors to CRC, and inflammatory bowel disease (IBD) can present with the same symptoms as CRC. The majority of patients with CRC initially consult primary care. Symptoms associated with CRC are also common among primary care patients, but seldom caused by any significant disease. Reliable diagnostic aids would be helpful in deciding which patients to refer. Faecal immunochemical tests (FITs) are commonly used for this purpose in primary care in Sweden, but there is little evidence to support this use. Faecal calprotectin (FC) has been suggested as an additional test.
Aim: To explore how doctors in primary care investigate patients with suspected CRC, the value of FITs, symptoms and presence of anaemia in diagnosing CRC and adenomas in primary care, and whether FC tests could contribute to diagnosis.
Methods: Three studies (1-3) were carried out in Region Jämtland Härjedalen, Sweden. There was no screening programme for CRC. We used a point of care qualitative dip-stick 3-sample FIT with a cut-off of 25-50μg haemoglobin/g faeces, and a calprotectin enzyme-linked immunosorbent assay (ELISA) test with a cut-off of 100 μg/g faeces. 1: A retrospective, population-based study including all patients diagnosed with CRC or adenomas with high-grade dysplasia (HGD) during the period 2005-2009 that initially consulted primary care. Symptoms, FIT results, anaemia and time to diagnosis were retrieved from medical records. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated from FIT results at the region’s health centres 2008- 2009. (Paper I.) 2: A prospective cohort study including consecutive patients where primary care doctors requested FITs and/or FC tests, at four health centres, from 30 Jan 2013 to 31 May 2014. FITs, FC tests, haemoglobin and iron deficiency tests were analysed; patients and doctors answered questionnaires about symptoms. Patients were examined with bowel imaging or followed for two years. Findings of CRC, adenomas with HGD, adenomas with low grade dysplasia (LGD) ≥1 cm and IBD were registered. (Papers II and III.) 3: A qualitative study of interviews with eleven primary care doctors. We explored what made them suspect CRC, and their practices regarding investigation and referral with particular attention to their use of FITs. Qualitative content analysis with an inductive approach was used for the analysis. (Paper IV.)
Results: 1: Paper I: Of 495 patients 323 (65.3%) started the investigation in primary care. FITs were analysed in 215. In 23 cases with CRC, FITs were negative; 15 (65.2%) had anaemia. In 33 cases with CRC, FITs were performed due to asymptomatic anaemia; 10 (30.3%) had negative FITs. The time from start of investigation, to the diagnosis of CRC or adenomas with HGD, was significantly longer for patients with negative FITs.
2: 377 patients (9 diagnosed with CRC, 10 with IBD) were included. Paper II: Concordance of positive answers about symptoms from patients and doctors was generally low. Rectal bleeding (recorded by 43.5% of patients and 25.6% of doctors) was the only symptom related to CRC and IBD. The FIT showed a better PPV than rectal bleeding for CRC and IBD. When patients recorded rectal bleeding, the FIT had a PPV of 22.6% and a NPV of 98.9% for CRC and IBD. Paper III: The best test for detecting CRC and IBD was the combination of a positive FIT and/or anaemia with a sensitivity, specificity, PPV and NPV of 100%, 61.7%, 11.7% and 100% respectively. The FC test had no additional value to the FIT alone. The sensitivity, specificity, PPV and NPV of the FIT for CRC in study 1 was estimated at 88.4%, 73.3%, 6.2% and 99.7% respectively. In study 2, corresponding figures were 88.9%, 67.4%, 6.3% and 99.6% respectively.
3: Paper IV: We identified four categories: “Careful listening – with awareness of the pit-falls”, “tests can help – the FIT can also complicate the diagnosis”, “to refer or not to refer – safety margins are necessary”, and “growing more confident – but also more humble”. All doctors had found their own way to handle FIT results in the absence of guidelines.
Conclusion: The diagnostic process when suspecting CRC can be described as navigating uncertain waters with safety margins. FITs were often used by primary care doctors but with considerable variations in interpretation and handling of results. Rectal bleeding was the only symptom related to CRC and IBD, but the FIT showed a better PPV than rectal bleeding. The combination of a negative FIT and no anaemia may be useful as a rule-out test when CRC is suspected in primary care, and this potentially also applies when patients present with rectal bleeding. Further studies are needed to confirm this and to determine the optimal FIT cut-off value for this use.