Sage Journals: Discover world-class research

Abstract

Objective

To identify factors associated with a false-positive result in a population-based colorectal cancer (CRC) screening programme with the faecal occult blood test (FOBT) in Catalonia between 2000 and 2010.

Methods

The study population consisted of participants of the Catalan CRC screening programme with a positive FOBT who underwent a colonoscopy for diagnostic confirmation from 2000 to 2010. A false-positive result was defined as having a positive test but detecting no high-risk adenoma or cancer in the follow-up colonoscopy. Multivariate logistic regression models were performed to identify sociodemographic and screening variables related to false-positive results. Adjusted odds ratios (OR) and their 95% confidence intervals (CI) were estimated.

Results

Over the screening period, 1074 (1.7%) of the 63,332 screening tests had a positive result in the Catalan CRC screening programme. The false-positive proportion was 55.2% (n = 546). Women were more likely to have a positive FOBT in the absence of CRC neoplasia than men (adjusted OR = 2.91; 95% CI: 2.22–3.28). During the first prevalence round, the proportion of false-positive results was higher than in subsequent rounds (69.5% vs. 48.9%; P < 0.05). Re-screening and having a bleeding pathology such as haemorrhoids or anal fissures were also associated with a false-positive result.

Conclusion

The proportion of false-positive results and the associated risks should be estimated to provide an eligible population with more reliable information on the adverse effects of screening.

Introduction

Screening for colorectal cancer (CRC) provides a simple and effective public health intervention to prevent and minimize the impact of CRC on the community. There is convincing evidence that CRC screening reduces mortality through detection and treatment of early stage cancers and by detection and removal of adenomatous polyps. Because large adenomas and cancers tend to bleed, presence of blood in the stool is considered to be a possible marker of the presence of either precancerous colonic adenoma or malignancy.¹ Most population-based screening programmes for CRC are based on faecal occult blood testing (FOBT), however, few health interventions are free of unwanted consequences.², ³ Although some randomized trials and pilot studies have already analysed the adverse effects of CRC screening^4–6 there is little research specifically addressing adverse effects of a long-term established CRC screening programme. One of the main adverse effects in a screening programme is a false-positive result. False-positive tests lead to discomfort, costs and risks from additional diagnostic and therapeutic procedures.⁷ We aimed to identify factors associated with false-positive results in the CRC screening programme in Catalonia (Spain) between 2000 and 2010.

Methods

Screening Procedure

A detailed description of the screening procedure is provided elsewhere.⁸ Briefly, a biennial screening programme for CRC using the guaiac faecal occult blood test (gFOBT) was implemented in Hospitalet de Llobregat. The target population (average of 65,000 individuals per screening round) included men and women aged 50–69 years who lived in the L'Hospitalet de Llobregat (screening area). Demographic data on this population was gathered from the Primary Healthcare Information System. Subjects were excluded according to the following criteria: history of CRC, adenomas or inflammatory bowel disease, criteria for hereditary CRC syndromes, colonoscopy in the previous five years or FOBT within the last two years, moving out of the screening area, terminal disease and severe disabling conditions.

From the first to third rounds, gFOBT was used as the screening test (Hema-screen^TM). In the fourth round, two tests were used, gFOBT and the immunochemical quantitative test (FIT) OC Sensor®. We compared both tests in terms of performance and participation in our screening population.⁹ The gFOBT was offered to 50,227 individuals and the FIT to 12,707. Six faecal samples (two samples from three separate bowel movements) were required to complete a gFOBT whereas only one sample was needed for the FIT. The presence of faecal occult blood in five or six samples (or in any sample after retesting if 1 to 4 positive samples) and a haemoglobin level of ≥ 100 ng/mL were used to designate a positive FOBT result. All participants with a positive test result were advised to have colonoscopy. If no lesions were found in the colonoscopy those subjects would be invited for screening again after 10 years.

Study Population

The study population consisted of participants of the CRC screening programme with a positive FOBT who underwent a colonoscopy for diagnostic confirmation from 2000 to 2010 (n = 989). A false-positive result was defined as a positive FOBT but no detected high-risk adenoma or cancer in the follow-up colonoscopy. Because there is ongoing debate concerning the target lesion in colorectal cancer screening we also performed a sensitivity analysis to examine the effect that including adenomas of any size would have on false-positive rates in our CRC screening programme.⁶

Data Collection And Analysis

The CRC screening programme information system included data on patient identification, age, sex, participation, appointment dates, screening test and colonoscopy results. We differentiated between initial and subsequent screening, and revised colonoscopy reports to identify sources of rectal bleeding such as haemorrhoids and anal fissures and severe complications during the diagnostic procedure (perforation, bleeding requiring transfusion, vagal syndrome or peritonitis-like syndrome).

Multivariate logistic regression models were performed to identify sociodemographic and screening variables associated with a false-positive result (0: true-positive result; 1: false-positive result). Adjusted odds ratios (OR) and their 95% CI were estimated.

Results

False-Positive Results in the Crc Screening Programme

Over the screening period, 1074 (1.7%) of the 63,332 screening tests had a positive result in the Catalan CRC screening programme. Our analysis was restricted to those individuals who underwent a colonoscopy (n = 989).

The sensitivity analysis showed the effect of including any adenoma on false-positive rates. The overall proportion of false-positive results ranged from 55.2% (n = 546) to 47.9% (n = 474) depending on the definition used (see Tables 1 and 2). The percentage receiving a false-positive result using the broader definition (any adenoma) changed from 44.0% to 36.9% in men, and from 69.9% to 62.4% in women.

Table 1

False-positive results in CCR screening in Catalonia (Spain), 2000–2010

	Positive FOBT^*		False-positive results^†
	n	(%)^‡	n	(%)	OR^‡‡	95% CI
Sex
Men	561	2.3	247	(44.0)	1
Women	428	1.5	299	(69.9)	2.91	(2.22–3.82)
Age group
50–60 years	382	1.4	219	(57.3)	1
60–69 years	607	2.2	327	(53.9)	0.98	(0.74–1.29)
Round
R1	334	3.4	232	(69.5)	1
R2	108	0.8	53	(49.1)	0.41	(0.26–0.65)
R3	180	1.1	94	(52.2)	0.45	(0.30–0.66)
R4	367	2.0	167	(45.5)	0.39	(0.28–0.53)
gFOBT^**	96	0.7	38	(39.6)	1
FIT^**	271	6.2	129	(47.6)	1.44	(0.88–2.37)
Type of screening
First screening	587	2.0	336	(57.2)	1
Successive screening	402	0.9	210	(52.2)	1.86	(1.31–2.65)
Total	989	1.8	546	(55.2)

Positive faecal occult blood test with diagnostic follow-up (colonoscopy);

†

False-positive results: not advanced neoplasia;

‡

Overall percentages of FOBT positive results among CRC screenees;

‡‡

Odds ratios adjusted for sex, age and round

gFOBT: guaiac test, FIT: immunochemical test;

Table 2

Sensitivity analysis: false-positive results in CCR screening in Catalonia (Spain), 2000–2010

	False-positive results^†
	n	(%)	OR^‡	95% CI
Sex
Men	207	(36.9)	1
Women	267	(62.4)	2.82	(2.16–3.68)
Age group
50–60 years	185	(48.4)	1
60–69 years	289	(47.6)	1.11	(0.85–1.46)
Round
R1	210	(62.9)	1
R2	46	(42.6)	0.43	(0.27–0.68)
R3	81	(45.0)	0.45	(0.31–0.66)
R4	137	(37.3)	0.37	(0.27–0.50)
gFOBT^*	32	(33.3)	1
FIT^*	105	(38.8)	1.28	(0.77–2.13)
Type of screening
First screening	295	(50.3)	1
Successive screening	179	(44.5)	1.72	(1.20–2.46)
Total	474	(47.9)

†

False-positive results: neither adenoma nor cancer;

gFOBT: guaiac test, FIT: immunochemical test;

‡

Odds ratios adjusted for sex, age and round

Table 1 shows factors associated with false-positive results derived from multivariate analyses. The logistic regression models were adjusted for age, sex and round. Women were more likely to have a positive FOBT in the absence of CRC neoplasia than men (OR = 2.91; 95% CI: 2.22–3.28). No differences according to age were found.

During the first prevalence round, 69.5% of the participants with a positive FOBT had a negative colonoscopy result. Participants in subsequent rounds had on average 48.9% false-positive results (P < 0.05).

In the fourth round, two screening strategies were compared (gFOBT and FIT). Among those individuals undergoing FIT, false-positive results were 8 percentage points higher than those who used the gFOBT, although the differences were not statistically significant.

Re-screening was associated with a false-positive result. Although the crude analysis suggested that a false-positive result was less likely in subsequent screenings than at first tests (52.2% vs. 57.2% of positive tests, respectively), after adjusting for potential confounding variables (age, sex and round) the direction of the effect was reversed from the crude estimate. Thus, those individuals who had previously participated in the CRC screening programme were more likely to have a false-positive result than those being initially screened (OR = 1.86; 95% CI: 1.31–2.65).

Factors Associated with Sex Differences in False-Positive Results

Table 3 shows the results from an analysis of factors associated with sex differences in false-positive results such as the screening strategy, haemorrhoids, anal fissures, menstrual bleeding or haemoglobin concentrations in blood. We reviewed the colonoscopy reports to determine whether the prevalence of bleeding pathology (haemorrhoids or anal fissures) could partly explain the sex differences in false-positive results. The prevalence of haemorrhoids and fissures among participants in our CRC screening was 24.5% (n = 242). Most of the screenees had only haemorrhoids (n = 239), both pathologies were reported in ten cases and finally anal fissures were identified in three individuals. Although having a bleeding pathology was a risk factor for false-positive results (sex-and age-adjusted OR: 1.84; 95% CI: 1.25–2.71), no sex differences regarding haemorrhoids and fissures were found (P = 0.96).

Table 3

Potential factors associated with sex-differences among CRC screenees with a flase-positive result

	Males			Females
	Total^*	FP	(%)	Total^*	FP	(%)	P value^‡
Screening strategy							0.5190
FIT	170	66	(38.8)	101	63	(62.4)
gFOBT^**	214	76	(35.5)	170	109	(64.1)
Bleeding pathology
Haemorrhoids or fissures							0.9555
No	422	170	(40.3)	325	218	(67.1)
Yes	139	77	(55.4)	103	81	(78.6)
Age^†							0.3041
≤56 yrs				93	69	(74.2)
≥57 yrs				335	230	(68.7)
Age^††							0.9648
≤66 yrs	419	189	(45.1)	338	239	(70.7)
≥67 yrs	142	58	(40.8)	90	60	(66.7)

Positive faecal occult blood test with diagnostic follow-up (colonoscopy); FP: false-positive results (not advanced neoplasia);

Data from the first round of screening not included;

†

Age stratification to study potential differences in false-positive results due to menstruation;

††

Age stratification to study potential differences in false-positive results due to haemoglobin concentration;

‡

Pvalue for sex interaction in the logistic regression model

We used age as a proxy indicator of menstrual bleeding and haemoglobin concentrations in blood, assuming that more than 97.5% of women at age 57 were already menopausal.¹⁰ We also considered that ≥ 10 years after the menopause the blood haemoglobin concentrations in women became similar to that of age-matched men.¹¹, ¹² No differences according to any of the age groups were found.

Results from the sensitivity analysis were nearly identical (see Table 4).

Table 4

Sensitivity analysis: potential factors associated with sex-differences among CRC screenees with a false-positive result

	Males FP	(%)	Females FP	(%)	p-value^‡
Screening strategy					0.3945
FIT	51	(30.0)	54	(53.5)
gFOBT^*	60	(28.0)	99	(58.2)
Bleeding pathology
Haemorrhoids or fissures					0.7699
No	141	(33.4)	194	(59.7)
Yes	66	(47.5)	73	(70.9)
Age^†					0.6314
≤56 yrs			60	(64.5)
≥57 yrs			207	(61.8)
Age^††					0.3069
≤66 yrs	157	(37.5)	218	(64.5)
≥67 yrs	50	(35.2)	49	(54.4)

FP: False-positive results (neither adenoma nor cancer);

Data from the first round of screening not included;

†

Age stratification to study potential differences in false-positive results due to menstruation;

††

Age stratification to study potential differences in false-positive results due to haemoglobin concentration;

‡

P value for sex interaction in the logistic regression model

Adverse Events After Colonoscopy

Through four screening rounds, the rate of severe complications after colonoscopy was 0.9%. Bowel perforation was reported in three cases and major bleeding was reported in six cases. All except one of these complications occurred in therapeutic colonoscopies following polypectomy.

Discussion

Half of the individuals with a positive FOBT had a negative finding on colonoscopy in our CRC screening programme. The first prevalent round produced a higher proportion of false-positive results. Although the screening programme is aimed at a healthy population, it is likely that in the first round some people with CRC symptoms or inflammatory bowel diseases were more likely to participate in the programme. A better selection of the eligible population in the subsequent rounds, in addition to the consolidation of the screening programme (leading to a better knowledge among health professionals and the target population) could be likely explanations for this finding. As has been found in other cancer screening programmes,¹³ the risk of a false-positive result increased with successive screening participations. No statistically significant differences were found regarding the screening test, although the proportion of false-positive results with FIT was higher than with gFOBT, probably because FIT detects blood at lower concentrations than gFOBT. Whilst improved analytical specificity reduces false-positive tests from dietary blood, its increased analytical sensitivity means that small losses from inflammatory diseases or other physiological sources will bring new false-positives.¹ The more striking finding was the gender-differences in false-positive results. Women showed more than a twofold likelihood of a false-positive result than men. Menstrual bleeding and sex-differences in blood haemoglobin concentrations made little, if any, contribution to the false-positive results in our study population, because the effect was independent of age. A quarter of the screened individuals had haemorrhoids, but no differences by gender were found.

Potential interference of drug intake such as ant-iplatelet and anticoagulation therapy and nonsteroidal anti-inflammatory drugs (NSAIDs) on FOBT performance has been reported.¹⁴ We did not collect information regarding medication in our screenees, although data from an unpublished study conducted among men and women aged 50–69 from the same screening area showed no statistical differences in the intake of NSAIDs in the overall group. However, women reported a significant higher intake of drugs for rheumatoid arthritis and osteoarthritis. Further research is needed to clarify the role of NSAIDS on false-positive results in order to recommend discontinuing NSAIDs, if clinically feasible, when collecting faecal samples.

There are no serious adverse health effects directly attributable to FOBT testing. The complications in a screening programme occur during follow-up of screened individuals with a positive test result, most of whom are disease-free. A high proportion of false-positive results leads to a number of unnecessary colonoscopies with associated costs and risks. Harm may arise from the preparation the patient undergoes before the procedure, the sedation used during the procedure, and the procedure itself. In randomized clinical trials, severe complications in FOBT screening programmes have been reported as 0.5% to 1.6%.¹⁵ The relatively low rate of adverse events after colonoscopy¹⁶ makes it difficult to differentiate acceptable from unacceptable incidence rates unless a large number of procedures are tracked over time. Another challenge is the accurate identification of those events presenting after dismissal from the endoscopy unit.¹⁴ Late complications may still be underestimated because of under-reporting.

Our study has several limitations: firstly, the estimated proportion of false-positive results could be biased although the adherence to those individuals adequately refered to colonoscopy among screen-positive individuals was quite high (93.7%).

Secondly, even though colonoscopy is the most reliable method available to date for assessing presence of colorectal neoplasms, it is not perfect, and missed adenomas at colonoscopy might have led to some overestimation of false-positive rates of the FOBTs.¹⁷

Thirdly, data concerning the follow-up of people with a false-positive result was not available at the time of the analysis. Follow-up data should be collected to identify interval cancers and to estimate the likelihood of testing positive again. Some studies suggested that the risk of CRC among individuals with a prior negative colonoscopy is very low for up to 20 years or longer.¹⁸, ¹⁹ The prevalence of advanced adenomas has also been reported to be very low for several years after a negative colonoscopy.²⁰, ²¹ However, those studies estimated risk of CRC following a negative colonoscopy performed among average-risk people. In a recently published paper, Brenner et al. showed an association between interval cancers and the performance of the preceding colonoscopy due to a positive FOBT (adjusted OR = 2.93; 95% CI: 1.13–7.61).²² Individuals aged under 60 when the false-positive result occurred will be invited to CRC screening 10 years after their false-positive result. Those individuals may have an increased likelihood of testing positive 10 years after the prior positive result. For example, haemorrhoids are a common condition in the target age range for CRC screening. Because this condition is often persistent, individuals with haemorrhoids who once tested positive at FOBT may also have a higher likelihood of receiving another positive result if they continue with FOBT.²³

Fourthly, we could not collect prospective data regarding potential factors associated with gender differences in false-positive results.

Finally, the study could not evaluate the adverse psychological effects associated with a false-positive result. To our knowledge only three studies to date have assessed the concern of screened individuals with a false-positive result.^24–26 Parker and colleagues measured anxiety scores in 100 gFOBT-positive individuals and found that these were highest after notification of a positive gFOBT and before investigation by colonoscopy. In patients with false-positive results, scores fell the day after colonoscopy and remained low one month later. In general cancer screening programmes, anxiety, distress and depression symptoms can be reduced by a brief orientation programme at the clinic visit.

In conclusion, the large gender differences in false-positive results was the most remarkable finding and merits more research. In addition, the proportion of false-positive results and the associated risks should be estimated to provide eligible populations with more reliable information on the adverse effects of screening.

References

Halloran

, Launoy

, Zappa

. Evidence for the effectiveness of colorectal cancer screening. In: Segnan

, Patnick

, von Karsa

, eds. European guidelines for quality assurance in colorectal cancer screening and diagnosis. 1st edn. Luxembourg: European Commission, Publications Office of the European Union, 2010

Muir Gray

, Patnick

, Blanks

RG.

Maximizing benefit and minimizing harm of screening.

BMJ 2008; 336: 480–3.

Marshall

KG.

Prevention. How much harm? How much benefit? III Physical, psychological and social harm.

Can Med Assoc J 1996; 155: 169–76.

Robinson

, Hardcastle

, Moss

SM.

The risks of screening: data from the Nottingham randomized controlled trial of faecal occult blood screening for colorectal cancer.

Gut 1999; 45: 588–92.

The UK CRC Screening Pilot Evaluation Team (2003) Evaluation of the UK Colorectal Cancer Screening Pilot Final Report. http://www.cancerscreening.nhs.uk/bowel/finalreport.pdf (last accessed 30 December 2011)

Croswell

, Kramer

, Kreimer

AR.

Cumulative incidence of false-positive results in repeated, multimodal cancer screening.

Ann Fam Med 2009; 7: 212–22.

Feldman

How serious are the adverse effects of screening?

J Gen Intern Med 1990; 5 (Suppl 5): S50–3.

Peris

, Espinas

, Muñoz

, Navarro

, Binefa

, Borras

JM.

Lessons learnt from a population-based pilot program for colorectal cancer screening in Catalonia (Spain).

J Med Screen 2007; 14: 81–6.

Garcia

, Binefa

, Milà

Evaluating colorectal cancer screening strategies (immunological test vs. biochemical test) in Catalonia, Spain 2008-2010.

Rev Esp Salud Publica 2011; 85: 593–602.

10.

de Bruin

, Bovenhuis

, van Noord

PA.

The role of genetic factors in age at natural menopause.

Hum Reprod 2001; 16: 2014–8.

11.

Brenner

, Haug

, Hundt

Sex differences in performance of aecal occult blood testing.

Am J Gastroenterol 2010; 105: 2457–64.

12.

Myers

, Saunders

CRG

, Chalmers

DG.

The haemoglobin level of fit elderly people.

Lancet 1968; 2: 261–3.

13.

Castells

, Molins

, Macia

Cumulative false-positive recall rate and association with participant-related factors in a population-based breast cancer screening programme.

J Epidemiol Community Health 2006; 60: 316–21.

14.

Petersen

BT.

Quality assurance for endoscopists.

Best Pract Res Clin Gastroenterol 2011; 25: 349–60.

15.

Moss

, Ancelle-Park

, Brenner

. Evaluation and interpretation of screening outcomes. In: Segnan

, Patnick

, von Karsa

, eds. European guidelines for quality assurance in colorectal cancer screening and diagnosis. 1st edn. Luxembourg: European Commission, Publications Office of the European Union: 71-102

16.

, Riffle

, Michaels

Serious complications within 30 days of screening and surveillance colonoscopy are uncommon.

Clin Gastroenterol Hepatol 2010; 8: 166–73.

17.

Van Rijn

, Reitsma

, Stoker

, Bossuyt

, van Deventer

, Dekker

Polyp miss rate determined by tandem colonoscopy: a systematic review.

Am J Gastroenterol 2006; 101: 343–50.

18.

Brenner

, Arndt

, Stürmer

, Stegmaier

, Ziegler

, Dhom

Long-lasting reduction of risk of colorectal cancer following screening endoscopy.

Br J Cancer 2001; 85: 972–6.

19.

Brenner

, Chang-Claude

, Seiler

, Stürmer

, Hoffmeister

Does a negative screening colonoscopy ever need to be repeated?

Gut 2006; 55: 1145–50.

20.

Brenner

, Haug

, Arndt

, Stegmaier

, Altenhofen

, Hoffmeister

Low risk of colorectal cancer and advanced adenomas more than 10 years after negative colonoscopy.

Gastroenterology 2010; 138: 870–6.

21.

Imperiale

, Glowinski

, Lin-Cooper

, Larkin

, Rogge

, Ransohoff

DF.

Five-year risk of colorectal neoplasia after negative screening colonoscopy.

N Engl J Med 2008; 359: 1218–24.

22.

Brenner

, Chang-Claude

, Seiler

, Hoffmeister . Interval cancers after negative colonoscopy: population-based case-control study. Gut 2012. Dec 26 [Epub ahead of print]

23.

Haug

, Knudsen

, Kuntz

KM.

How should individuals with a false-positive aecal occult blood test for colorectal cancer be managed? A decision analysis.

Int J Cancer 2012. Feb 5 [Epub ahead of print]

24.

Parker

, Robinson

MHE

, Scholefield

, Hardcastle

JD.

Psychiatric morbidity and screening for colorectal cancer.

J Med Screen 2002; 9: 7–10.

25.

Lindholm

, Berglund

, Kewwhnter

, Haglind

Worry associated with screening for colorectal carcinomas.

Scand J Gastroenterol 1997; 32: 238–45.

26.

Mant

, Fitzpatrick

, Hogg

Experiences of patients with false-positive results from colorectal cancer screening.

Br J Gen Pract 1990; 40: 423–5.

False-positive results from colorectal cancer screening in Catalonia (Spain),2000–2010

Abstract

Objective

Methods

Results

Conclusion

Introduction

Methods

Screening Procedure

Study Population

Data Collection And Analysis

Results

False-Positive Results in the Crc Screening Programme

Factors Associated with Sex Differences in False-Positive Results

Adverse Events After Colonoscopy

Discussion

References