The impact of incomplete colonoscopies: a single-centre retrospective study

Abstract

Background:

Screening for colorectal cancer reduces mortality by enabling early detection. In Quebec, follow-up within 60 days after an incomplete colonoscopy is recommended. In this study, we sought to assess the impact of delays in follow-up on patient outcomes.

Methods:

In this retrospective study, we included adults who underwent a colonoscopy following a positive immunochemical fecal occult blood test at the Centre intégré universitaire de santé et de services sociaux de l’Estrie–Centre hospitalier de l’université de Sherbrooke between Jan. 1, 2013, and Dec. 31, 2015. We verified colonoscopy adequacy and guideline adherence. We classified advanced polyps and colorectal cancer as clinically significant lesions (CSLs) to assess the clinical impact of incomplete or missing follow-up colonoscopies.

Results:

In 89 cases of incomplete colonoscopies, inadequate bowel preparation was the leading cause of exam interruption (61.8%). A total of 57 patients had a subsequent follow-up, and 23 colonoscopies were completed within the 60-day time frame. Six CSLs, including advanced polyps and cancer, were detected within the recommended 60-day time frame, and 4 were identified after 60 days. We found a statistically significant difference in the rates of colorectal cancer diagnosis (p < 0.001), the need for surgery (p < 0.02), and death (p < 0.001) between patients who had a complete colonoscopy diagnostic process and those who did not.

Conclusion:

The poorer prognosis associated with patients who had a delayed or missing follow-up highlights the importance of respecting provincial guidelines concerning follow-up after incomplete colonoscopies.

Colorectal cancer is the second leading cause of cancer-related deaths among men and the third among women in Canada, and represents an important burden for the health care system.¹ Beginning with the formation of polyps, the disease is known to have a progressive and asymptomatic growth. Given that the removal of early-stage polyps prevents their progression to a cancerous stage, the implementation of screening programs has been well justified. These programs have contributed to a consistent decline in the incidence of colorectal cancer in Canada since the early 2000s.² Specifically, a meta-analysis showed that colorectal cancer screening significantly decreases the risk of death as well as late-stage cancer diagnosis.³

The implementation of a colorectal cancer screening program in the province of Quebec established rigorous guidelines for clinicians who perform colonoscopies. To fully utilize the screening capabilities of a colonoscopy, a cecal intubation rate of 95% or higher as well as adequate bowel preparation are required. All incomplete or inadequately prepared colonoscopies must be followed by another evaluation to allow the complete visualization of the colon within 60 days of the first attempt, whether that be a second colonoscopy, a virtual colonoscopy, or a barium enema.⁴

A 2007 study found that around 13% of colonoscopies in Canada are incomplete.⁵ Of these, only 50% to 75% are followed by a complementary diagnostic exam, and the ideal 60-day time frame between the first and second exam is met in only half the cases. Bollart and colleagues suggested that an incomplete colonoscopy is associated with a fivefold increased risk of colorectal cancer compared with evaluations consistent with current recommendations (14.3% v. 2.9%).⁶ Notably, clinically significant lesions (CSLs) are detected in 10% to 20% of the patients who undergo a follow-up colonoscopy after an incomplete colonoscopy.^5–7 Despite these striking results, data are lacking to establish the clinical impact of a complementary exam omission following an inadequate colonoscopy. Furthermore, and specifically regarding Quebec’s screening program, data are lacking to evaluate the adherence to provincial guidelines or the consequences of exceeding the 60-day time frame.

This gap in the scientific literature prompted the present study. We sought to determine the impact on the patient’s prognosis of not completing a complementary exam within 60 days after an inadequate colonoscopy. Secondary objectives included assessing the colonoscopy completion rate and examining the timeliness of complementary exams at the Centre intégré universitaire de santé et de services sociaux (CIUSSS) de l’Estrie–Centre hospitalier de l’Université de Sherbrooke (CHUS). We also analyzed the causes of incomplete colonoscopies. Our working hypothesis was that incomplete colonoscopies would be associated with a worse prognosis.

Methods

Study design

We conducted a retrospective observational cohort study. We included all patients aged 18 years and older who underwent a colonoscopy following a positive immunochemical fecal occult blood test (iFOBT) at the CIUSSS de l’Estrie–CHUS between Jan. 1, 2013, and Dec. 31, 2015. We used the Infocentre database to start patient identification in July 2019. We excluded patients whose initial colonoscopy reports were not available.

Data collection

We extracted relevant data from the patients’ electronic medical records and compiled the data in an Excel database. When applicable, complementary exam reports were retrieved from other health care facilities affiliated to the CIUSSS de l’Estrie–CHUS. We collected follow-up data for patients with incomplete colonoscopies for a period of at least 2 years and 11 months following the initial colonoscopy.

We assessed adequate colonoscopy completion by evaluating the reported quality of bowel preparation and confirming the successful intubation of the cecum. The quality of the colonic preparation is commonly evaluated using the Boston Bowel Preparation Scale (BBPS), which uses both a qualitative and a quantitative grading system.⁸ Therefore, the bowel preparation was considered inadequate if the terms “inadequate” or “poor” preparation were used; 1 or more of the imaged colon sections was graded 0 or 1 on the BBPS; and the colonoscopy report indicated an excessive accumulation of stool in the bowel.

We considered the cecum intubated if the ileocecal valve or the appendiceal orifice were noted to be visualized, or an arrow indicated that the cecum was reached on the colonoscopy report.

We classified advanced polyps and colorectal cancer as CSLs. Advanced polyps included polyps measuring 1 cm or larger, polyps with high-grade dysplasia, and cancerous polyps. We defined poor prognosis by the late detection of advanced polyps, or colorectal cancer and death (all-cause death). For patients whose complementary exam was completed beyond the ideal 60-day time frame, we compiled consultations investigating signs and symptoms indicative of colorectal cancer, including abdominal pain, constipation, rectal bleeding, microcytic anemia, and weight loss.

Statistical analysis

We used IBM SPSS Statistics 25.0 to perform the statistical analysis. We used central tendency and dispersion measures corresponding to each type of variable to draw up descriptive statistics. We used the χ² and Fisher tests to compare the 3 subgroups of patients with incomplete colonoscopies: patients with no complementary investigation, and patients with complementary investigation either within 60 days or after 60 days. We considered statistical significance to be reached at a p value less than 0.05.

Ethics approval

The Research Ethics Board of the CIUSSS de l’Estrie–CHUS approved this study (project #2019–2996).

Results

Patient selection

We identified 868 patients who had a positive iFOBT result between Jan. 1, 2013, and Dec. 31, 2015 (Figure 1). Of these, 132 patients were excluded because colonoscopy reports were unavailable, leaving 736 patients. Among these, 646 underwent complete colonoscopies and 1 had a file that did not specify whether the colonoscopy was completed. A total of 89 patients had incomplete colonoscopies. Of these patients, 57 underwent a complementary investigation (second colonoscopy, virtual colonoscopy, or barium enema) and 32 did not receive the recommended follow-up investigation. Among the 57 patients with complementary investigations, 23 had their follow-up completed within the recommended 60-day time frame, and 34 experienced delays of more than 60 days.

Fig. 1.

Study flowchart. CHUS = Centre hospitalier de l’université de Sherbrooke; CIUSSS = Centre intégré universitaire de santé et de services sociaux; iFOBT = immunochemical fecal occult blood test.

Missing colonoscopy after positive iFOBT

The reasons justifying the 132 omissions of a standard colonoscopy after a positive iFOBT are detailed in Table 1, with some cases being associated with multiple justifications.

Table 1.

Justifications for absent colonoscopy following a positive iFOBT

Omission justification	No. (%) of patients
Patient refusal	27 (17.8)
Lost to follow-up	16 (10.5)
Recent colonoscopy (< 5 yr)	9 (5.9)
Sigmoidoscopy	8 (5.3)
Colonoscopy completed outside the covered area	5 (3.3)
Virtual colonoscopy	14 (9.2)
Cognitive impairment	5 (3.3)
Major comorbidities or palliative care	6 (3.9)
Negative second iFOBT	45 (29.6)
Death	3 (2.0)
Unknown	14 (9.2)
Total	152 (100)

iFOBT = immunochemical fecal occult blood test.

Incomplete colonoscopies

A total of 89 patients had incomplete colonoscopies. The endoscopists’ justifications for interrupting the colonic evaluation are listed in Table 2. The most common cause of noncompletion was inadequate bowel preparation (61.8%). We found that 72.2% of men and 47.1% of women had inadequate bowel preparation (p = 0.05).

Table 2.

Justifications for interrupting the colonoscopy

Interruption justification	No. (%) of patients n = 89
Inadequate preparation	55 (61.8)
Obstruction	11 (12.4)
Pain	2 (2.2)
Difficult anatomy	12 (13.5)
Other justification	3 (3.4)
Missing justification	6 (6.7)

Complementary investigations

To complete the colonic evaluation, 57 of 89 patients with incomplete colonoscopy (64.0%) underwent additional imaging. A total of 40 (70.2%) had a second colonoscopy, 15 (26.3%) had a virtual colonoscopy, and 2 (3.5%) had a barium enema.

Among the 89 patients with incomplete colonoscopy, 23 patients (25.8%) had previously undergone a colonic evaluation within 5 years of the incomplete colonoscopy. There was no difference in the proportion of previous colonic examination between patients with or without complementary investigation (p = 0.6) (Table 3).

Table 3.

Complementary investigation according to the type of previous colonic examination in patients who had an incomplete colonoscopy

Previous examination	No. (%) of patients with a complementary investigation
	Yes n = 57	No n = 32
Standard colonoscopy	12 (21.1)	6 (18.8)
Virtual colonoscopy	1 (1.8)	2 (6.3)
Barium enema	2 (3.5)	0
None	42 (73.7)	24 (75.0)

Patient age did not have an impact on whether a follow-up examination was performed after an incomplete colonoscopy (p > 0.9), as detailed in Table 4.

Table 4.

Age distribution among patients with a complementary investigation following an incomplete colonoscopy

Age, yr	No. (%) of patients with a complementary investigation*
	Yes n = 57	No n = 32
50–59	11 (73.3)	4 (26.7)
60–69	21 (61.8)	13 (38.2)
70–79	22 (64.7)	12 (35.2)
80–89	3 (60.0)	2 (40.0)

^*The percentages were calculated across the row to reflect that there was no association between age and whether an additional examination was performed.

Patients whose initial exam was incomplete due to an anatomic difficulty (n = 12) were more likely to have a follow-up complementary investigation (1 [8.3%] v. 11 [91.7%]). The segment of colon reached during the initial colonoscopy did not have an influence on the probability of having a complementary examination. When an advanced polyp was found during the incomplete colonoscopy, most patients had an additional colonic evaluation. Those variables have been linked to the execution or the omission of a complementary exam in Table 5. Some patients had more than 1 finding.

Table 5.

Characteristics of initial incomplete colonoscopies among patients who underwent additional examinations and those who did not

Incomplete colonoscopy characteristic		No. (%) of patients with a complementary investigation		p value
		Yes n = 57	No n = 32
Reason for noncompletion	Unknown	2 (3.5)	4 (12.5)	0.03
	Inadequate preparation	34 (59.6)	21 (65.6)
	Mechanical obstruction	6 (10.5)	5 (15.6)
	Pain	2 (3.5)	0
	Difficult anatomy	11 (19.3)	1 (3.1)
	Other	2 (3.5)	1 (3.1)
Furthest segment reached	Rectum	2 (3.5)	0	0.6
	Sigmoid	8 (14.0)	3 (9.4)
	Left colon	8 (14.0)	3 (9.4)
	Transverse colon	4 (7.0)	1 (3.1)
	Right colon	3 (5.3)	4 (12.5)
	Caecum	32 (56.1)	21 (65.6)
Findings	Nonadvanced polyp	22 (38.6)	12 (37.5)	0.02
	Advanced polyp	14 (24.6)	1 (3.2)
	Colorectal cancer	8 (14.0)	5 (15.6)
	Other	2 (3.5)	2 (6.3)
	None	11 (19.3)	10 (31.3)

Delay between the 2 evaluations

Among the 57 complementary exams that were analyzed, 23 (40.4%) were completed within the recommended 60-day time frame. The median time separating the 2 exams was 96 days. Depending on the chosen method of complementary evaluation, the time interval between the 2 evaluations varied (Figure 2). In fact, standard colonoscopy had a significantly higher probability to be completed after 60 days, and virtual colonoscopy was more likely to be completed within 60 days (p = 0.004).

Fig. 2.

Time intervals between evaluations based on the chosen method of complementary examination.

Findings and clinical impact

Among the 57 patients who received a complementary examination, nonadvanced polyps were discovered in 21 (36.8%), advanced polyps were found in 7 (12.3%), and colorectal cancer was diagnosed in 3 patients (5.3%). Other findings were detected in a total of 14 patients (24.6%). In other words, a CSL was noted in 10 patients (17.5%) whose initial exam was incomplete. Eight out of 10 of these lesions were new findings. Six of those CSLs were discovered within the recommended 60-day time frame, whereas 4 CSLs were found in a time frame longer than 60 days (2 of those were already present during the first exam). More nonadvanced polyps were found when the exam was completed after more than 60 days (17 after more than 60 days v. 4 within 60 days), but this difference was not statistically significant (p = 0.1). Findings for each follow-up time frame are listed in Table 6. No additional abnormalities were found in 17 patients (29.8%). In 32 of the complementary investigations, there was a new finding that had not been detected during the incomplete colonoscopy. All new findings are detailed in Figure 3.

Table 6.

Findings of complementary investigations performed within or outside the 60-day time frame

Variable	Less than 60 days	More than 60 days	p value
Findings	16	29	0.1
Nonadvanced polyp	4	17
Advanced polyp	4	3
Colorectal cancer	2	1
Other	6	8
None	9	8

Fig. 3.

Proportion of new findings during complementary investigations. CRC = colorectal cancer.

Patients’ files were examined to verify whether they sought medical assistance for symptoms that could be associated with the development of a colorectal lesion while waiting for a complementary evaluation, which was the case for 16 of the 89 patients with incomplete colonoscopies. Of these consultations, 14 (87.5%) were made when the delay to get the complementary examination exceeded 60 days or by patients who did not undergo a follow-up examination. The mean time between the incomplete colonoscopy and the consultation was 615.1 days.

Colorectal cancer was diagnosed in 17 patients whose initial colonoscopy was incomplete. This represents 19.1% of the patients with an incomplete colonoscopy. More than half of those (n = 13) were diagnosed during the first colonoscopy and 2 were diagnosed on a follow-up examination done within 60 days. The proportion of colorectal cancer is significantly higher in patients with incomplete colonoscopy (19.1% v. 6.2%, p < 0.001). The distribution of colorectal cancer stages in patients with complete and incomplete colonoscopies is shown in Figure 4.

Fig. 4.

Stage of colorectal cancer at diagnosis.

Among the 736 patients with available colonoscopy reports and completion status specified, 54 patients (7.3%) (38 of whom initially had a complete colonoscopy and 16 of whom initially had an incomplete colonoscopy) underwent colorectal surgery for the findings (Table 7). In other words, 18.0% of patients with incomplete colonoscopy later required a surgical intervention. This is a significantly greater proportion than for the patients with a complete colonoscopy (p = 0.02). Although this finding was not statistically significant, there appears to be a trend indicating that more patients who had additional investigations following their incomplete colonoscopy later underwent colorectal surgery (13 [86.7%] v. 2 [13.3%], p = 0.05).

Table 7.

Patient outcomes at follow-up

Outcome	No. (%) of patients		p value
	Complete colonoscopies n = 646	Incomplete colonoscopies n = 89
Colorectal cancer	39 (6.0)	17 (19.1)	< 0.001
Surgery	38 (5.9)	16 (18.0)	0.02
Death*	31 (4.8)	15 (16.9)	< 0.001

^*All-cause death.

A total of 6 patients (6.3%) had passed away (all-cause death) before the end of this study, and 15 of them were patients whose initial colonoscopy was incomplete. Proportionally, this represents 16.9% of patients with incomplete colonoscopy and 4.8% of those with complete colonoscopy (p < 0.001).

Discussion

In this study, we assessed the completion rate of colonoscopies following a positive iFOBT, to evaluate the risk of CSLs in patients with an incomplete diagnostic process and to examine how delays in follow-up influence outcomes. We also explored the association between incomplete colonoscopies and subsequent consultations for gastrointestinal symptoms.

Given the current literature, it was expected that 70% of the patients who had a suboptimal colonic evaluation would receive a complementary exam and that the omission of a complementary exam would worsen prognosis by missing 10% to 20% of CSLs.^5–7 It was also anticipated that if the complementary exam was performed after a prolonged delay, it would lead to a worse prognosis than for those who had their exam completed within 60 days. Similarly to what is seen in other Canadian provinces, such as 13.1% of incomplete colonoscopies in Ontario, the completion rate of colonoscopies at the CIUSSS de l’Estrie–CHUS was expected to be below the suggested standards. The main cause of noncompletion was expected to be poor bowel preparation.⁵ These results are particularly relevant in the context of improving colorectal cancer screening programs and optimizing the diagnostic pathway.

Consistent with literature from other Canadian provinces, the CIUSSS de l’Estrie–CHUS showed a completion rate that falls below the 95% benchmark.⁴ This result is comparable to the findings of Shah and colleagues that 13.1% of colonoscopies in Ontario are incomplete.⁵ In the present study, inadequate bowel preparation was the leading cause of noncompletion. However, when colonoscopies with poor preparation or mechanical obstruction are excluded, the adjusted completion rate rises from 87.7% to 96.9%, better reflecting endoscopist performance. These findings emphasize that improving bowel preparation is a key intervention point to ensure a complete diagnostic process. At the time, reporting the quality of bowel preparation was not mandatory. As a result, the BBPS was not documented in more than half (59.4%) of the colonoscopies that were still considered complete based on other elements in the report.

Suboptimal bowel preparation may result from a combination of patient-specific and systemic factors, including noncompliance, intolerance, cognitive limitations,⁸ or misunderstanding of instructions.

Furthermore, the possibility that obstructive lesions compromise both preparation and completion supports the need for heightened clinical vigilance. In 12.4% of cases, mechanical obstruction directly prevented completion, highlighting the importance of recognizing when poor preparation might mask a more serious underlying pathology. Use of precolonoscopy assessments and tailoring strategies, such as imaging, supervised preparation, or alternative protocols, could mitigate these challenges. Ness and colleagues have identified multiple independent predictors of inadequate bowel preparation, including a later colonoscopy starting time, a reported failure to follow preparation instructions, inpatient status, a procedural indication of constipation, taking tricyclic antidepressants, male gender, and a history of cirrhosis, stroke or dementia.⁹ Whereas male sex was also associated with poor preparation in our study, the result did not reach statistical significance (p = 0.05).

A subsequent investigation followed in 64.0% of incomplete colonoscopies. Some patients may have been managed outside our institution, but such cases are likely limited. Age and previous colonoscopy history were not associated with the likelihood of follow-up. In some clinical contexts, especially when malignancy was already suspected or diagnosed, patients may have been referred directly to oncology or surgery. Reasons for colonoscopy omission included a negative second iFOBT result or a previous colonoscopy within the past 5 years. However, repeating an iFOBT after a positive result or prescribing the test within 10 years of a prior colonoscopy is not a recommended practice. Data collection occurred during the initial implementation of the iFOBT in the province, which highlights the importance of ongoing training for health care providers who prescribe the test. This may have influenced the proportion of CSLs detected in patients with a positive iFOBT over time and introduces potential bias that could distort the results of studies on clinical practices. This underscores the importance of adhering to established screening guidelines.

The absence of a significant association between age and the omission of a follow-up examination does not refute the potential influence of systemic or patient-specific barriers, such as accessibility to care, individual health literacy, or the perceived importance of follow-up. Further studies remain necessary to explore how these factors and other determinants can affect the efficiency of follow-up care, specifically in the context of incomplete colonoscopies, although this has already been studied in other health care settings. Notably, lack of insurance and absence of a primary care provider have been linked to decreased follow-up visits after consultations in the emergency department.¹⁰

The detection of a CSL during the initial examination increased the likelihood of follow-up, particularly in the case of advanced polyps. However, this association was not observed for colorectal cancer, as 5 out of 13 patients with colorectal cancer identified during incomplete colonoscopy did not undergo further evaluation, possibly owing to advanced disease requiring immediate surgery or a palliative approach. These findings highlight the importance of clinicians applying individualized clinical judgment when determining the necessity of follow-up, considering the patient’s overall clinical context and prognosis.

Treatment of CSLs has been delayed, as only 40% of the complementary exams were completed within the recommended 60-day interval.² Repeat standard colonoscopies were associated with significantly longer delays than virtual colonoscopy or barium enema, likely owing to resource limitations. Since some patients may not have detectable abnormalities during follow-up, the use of virtual colonoscopy could be considered as an alternative when endoscopic resources are limited. Among the 57 patients who underwent a complementary examination, 10 (17.5%) had CSLs identified. Six of these were diagnosed within 60 days, and 4 were identified later. These results suggest that timely follow-up may facilitate earlier detection and treatment of CSLs.

Among the 89 patients with incomplete colonoscopies, 16 (18.0%) required medical visits for gastrointestinal symptoms during follow-up. Notably, 14 (87.5%) of these consultations occurred in those with delayed or absent follow-up exams. However, the average delay between the incomplete exam and the medical visit was over a year, which challenges the rigidity of the 60-day delay in order to effectively balance the use of limited health care resources while mitigating the risk of unrecognized disease progression. Colorectal cancer was diagnosed in 19.1% of patients with incomplete colonoscopies compared with 6.0% of those with complete evaluations (p < 0.001), and the colorectal cancer diagnosed in patients with incomplete colonoscopies was proportionally more advanced, emphasizing the potential impact of delays on disease progression. A high index of suspicion should be maintained in cases of incomplete colonoscopy with a suboptimal bowel preparation. Delays in diagnosis not only negatively affect clinical outcomes, but may also contribute to increased patient anxiety, repeated health care visits, and inefficiencies within the system.

These findings furthermore reinforce the importance of effective bowel preparation protocols.^11,12 Strategies such as doubling the preparation dose in patients with prior inadequate preparation, providing supervised preparation in a day medicine unit, and targeted education can improve compliance.^13,14 An educational approach could particularly help patients who hesitate to follow the preparation protocol owing to the high cost of bowel preparation solutions.¹⁵

At the CIUSSS de l’Estrie–CHUS, low-volume split-dose preparations, conscious sedation, and the use of pediatric colonoscopes in selected patients are already in place and have been associated with better compliance and comfort.^16–18 The need to implement these strategies must be underscored given the cost-effectiveness such strategies (e.g., patient education, pamphlets, calls) would have when compared with the cost associated with repeated procedures, additional consultations, and delayed diagnoses.

Limitations

The retrospective design of this study makes the findings dependent on the quality of data available in medical records, introducing the possibility of information bias due to missing or incomplete data. Notably, the bowel preparation quality was not specified in more than half of the colonoscopies that were considered complete based on other elements of the report. In addition, with only 89 incomplete colonoscopies included, the sample size was far smaller than what is necessary to achieve statistical power. Furthermore, the results rely on data from 2013 to 2015 and might not fully capture the current clinical reality or recent changes in guideline adherence. Lastly, as the cause of death was not included in the database, the study could not distinguish between deaths related to colorectal cancer and those from other causes, such as cardiac or pulmonary conditions.

Conclusion

This study provides a comprehensive evaluation of colonoscopy completion rates and their associated outcomes in the context of positive iFOBT results at the CHUS. Findings confirm that inadequate bowel preparation remains the leading cause of incomplete colonoscopies, which underscores the need for improved patient education and preparation protocols. The suspicion index for CSLs should remain high in the context of inadequate bowel preparation. Delays in follow-up examinations were associated with a higher likelihood of advanced colorectal lesions, emphasizing the importance of adhering to the 60-day time frame recommended in the provincial guidelines. Further studies remain necessary to explore whether demographic determinants act as barriers to good bowel preparation and to the completion of timely follow-up examinations. The key takeaway would be for public institutions to seriously invest in patient education and preparation support strategies.