In colorectal cancer screening, successful colonoscopy is related to the polyps detection rate and the percentage of complete colon examinations achieved with cecal intubation. Both factors depend on technical issues and the endoscopist's performance [1]. The quality standard for the polyps detection rate (set at 25% in men and 15% in women) is related to the withdrawal time and differences seem to be due more to endoscopists' individual sensitivity than to their technical expertise [2].

Completed colonoscopy rates of 97-99% have been reported and, although a rate of 90% is acceptable in routine clinical activity, it is best to aim for at least 95% of completed colonoscopies in screening programs [3]. Various technical factors play a part, such as the calibre [4] or stiffness [5] of the instrument used, and the time of day when the test is scheduled [6], but the endoscopists' experience and the number of endoscopies they have already performed have an important influence on the success of the test [7].

Some endoscopy units achieve good results because the activity is handled by a single, dedicated endoscopist [8, 9], while large teaching centres reach lower quality indicators for their colonoscopy activities [10]. In large hospitals, the overall colonoscopy completion rates are influenced by the procedure being implemented by trainee doctors and by endoscopists from different specialities (gastroenterological, surgical, or internal medicine). In such cases, strategies have been proposed to improve performance based on auditing programs, adjusting the various endoscopists' workload in the light of their results [11], or having trainees use auxiliary devices to facilitate cecal intubation [12].

In hospitals with a large number of practising colonoscopists, a rapid strategy to optimize global performance and achieve an acceptable number of complete colonoscopies with minimal patient discomfort could be to routinely use deep sedation with propofol. This is considered a procedure with an acceptable safety profile [13], but the extensive use of such a method to improve technical results is a debatable issue [14]. The aim of this study was to present the results of routine propofol sedation for colonoscopy at a large teaching hospital where a large number of doctors with differing levels of expertise are involved in colonoscopy activities. We also assessed the factors influencing the success rate of the procedure and ways to improve performance, analyzing the complementary aspects relating to the use of deep sedation.

We investigated the routine activities at an endoscopy centre at Padova University teaching hospital, consecutively enrolling all patients who had a colonoscopy under deep sedation in the study over a period of one year (February 2007 to February 2008).

Subject to patients' informed written consent, their clinical history was recorded to identify any anaesthesiological or procedural risks, collecting data on: a) non-gastrointestinal conditions; b) allergies or side effects of previously-used anaesthetics; c) ASA status [15]. Information was then recorded on the completeness of the colonoscopy, the propofol dosage used and any adverse events on a case report form (CRF) formulated specifically for the purposes of the study. The health professionals responsible for compiling the CRF were unaware that they were participating in the study. Patients were not enrolled in the study if they refused to sign the consent form or if they were willing to undergo colonoscopy without sedation, or if the endoscopist preferred for clinical or technical reasons to perform the colonoscopy without sedation (e.g. as a follow-up procedure in a patient who had already tolerated the test well, or in a patient with an operated colon).

There were 77 AAI during the procedures (5.2% among the complete colonoscopies and 10.9% among the incomplete colonoscopies; p-value = 0.02) to deal with 28 respiratory(solved using mask ventilation, without any need for intubation) and 39 hemodynamic problems, 8 episodes of regurgitation, and 2 of allergic reactions (the patients were simultaneously treated with antibiotic prophylaxis). We observed two complications of the endoscopic procedure, i.e. one perforation and one post-polypectomy hemorrhage. Polypectomy was performed during 231 procedures.

A statistically significant association emerged between ASA scores and completeness of colonoscopy (p-value = 0.002): 53.9% of patients with completed colonoscopies and 44.6% of those with incomplete colonoscopies had an ASA score of 2; the corresponding percentages among the patients with ASA scores of 3 or 4 were 7.8% and 18.5%.

The mean total doses of propofol administered differed statistically between the two groups with a p-value of 0.02 (191.2 mg vs 172.1 mg, respectively).

The endoscopists' levels of experience were unevenly distributed between the two groups, with a greater prevalence of highly experienced specialists associated with the completed colonoscopy group (24% vs 8.7%, p-value < 0.0001).

Sex, body weight, duration of the procedure, O2 l/m and SPO2 variables were not associated with the completeness of colonoscopy.

About 73% of patients with completed colonoscopies had adequately prepared bowels, 24% of patients had residual matter and 3% of patients had inadequately cleansed bowels; the corresponding percentages among patients with incomplete colonoscopies were: 47%, 26% and 27%, revealing a statistical association (p-value < 0.0001).

The need for AAI and age over 71 years coincided with a 2.7-fold and a 2.1-fold increase in the risk of a colonoscopy not being completed; weighing less than 60 kg also emerged as a risk factor (OR = 1.89).

The logistic model had an area under the ROC curve of 77.8% and the value of the Hosmer-Lemeshow goodness of fit statistic was 2.05, with a corresponding p-value of 0.91, indicating that the model seems to fit quite well. None of the other variables contributed to the predictive power of the model.

These findings are graphed in Figure 1 by patient's age and body weight categories (unmodifiable predictors) and bowel preparation.

The greatest variation (more than 85%) is recorded for a patient over 71 years old with a body weight below 60 kg and inadequate bowel preparation, needing AAI during a procedure conducted by a less experienced non-specialist.

It has already been emphasized that the colonoscopy completion rate is important for more than just academic reasons [8] and one important reason is that a completed colonoscopy reduces the likelihood of an advanced colorectal-cancer being detected later on [16].

In this report on a single centre's experience of colonoscopies systematically associated with deep sedation with propofol, the global success rate achieved by 21 endoscopists working at the same endoscopic centre with different levels of expertise and training was 93%.

The factors negatively influencing the success of the test were: a patient's small body size, age over 71 years, inadequate colonic cleansing and the need for AAI during deep sedation. For the purposes of this study, we did not correct the performance rate by excluding colonoscopies in which cecal intubation had been prevented by the presence of faecal material, because our aim was to establish the crude success rate for colonoscopy under deep sedation. It is worth adding here that more tests were interrupted due to poor cleansing issues when less expert endoscopists were involved.

Two complications of endoscopy (one perforation and one post-polypectomy haemorrhage) were recorded during the study period, a prevalence similar to that of other reports [17]. We used a broad definition of "adverse events" during colonoscopy under deep sedation, recording all action taken by the anaesthetist to restore adequate oxygenation or hemodynamics, even for short-lived monitoring problems, because such situations might influence the procedure or have to do with longer probe insertion times necessitating further propofol infusions. No hemodynamic resuscitation was needed and there were no major clinical sequelae after sedation.

One of the limits of this study is that we were not in a position to compare how different colonoscopic methods might improve the colonoscopy completion rate.

Although ours was an observational study on the controversial issue of the routine use of deep sedation for colonoscopies, it had the advantage of referring to a recognized target (i.e. 95% cecal intubations in colonoscopies for screening purposes). Other published series adopted a different approach, achieving a different gap between their results and the 95% target [18–20].

The performance measurement showed that the endoscopist's skill influences the success of colonoscopies even when deep sedation is used. In a clinical setting, this poses problems that have already been studied and overcome using quality improvement programs and auditing cycles [11]. All our endoscopists had gained experience during weekly sessions for more than a year, implementing 200 procedures alone (a number recognized as being sufficient to achieve a satisfactory performance in screening programs)[21]. When it comes to screening programs, the endoscopists' different levels of experience carry a different weight from the situation in routine clinical practice because the most important problem is often the shortage of operators. Solutions have to be found to deal with the problems of the burden of colonoscopies to perform and increasingly long waiting lists, which is why colonoscopies may be performed by gastroenterologists, clinical assistants, trainee gastroenterologists or nurses in some countries [22].

The aim of our study, however, was to ascertain whether using propofol sedation enables the standard to be reached in screening programs, when it is necessary to employ many endoscopists who are likely to have different levels of expertise.

Success rate has been seen as an expression of "technical machismo" [8] and some authors suggest that it should not be influenced by patient-related factors and that, in any case, a trained endoscopist should be able to complete 95% of colonoscopies successfully [23]. We considered it important to demonstrate that patients' weight and age, as already mentioned in other studies [24, 25] without deep sedation, are factors that also predict incomplete colonoscopies among series of colonoscopies performed under deep sedation at a busy teaching hospital service, where endoscopists of different abilities are at work.

Different strategies can be used to improve the quality of colonoscopic screening programs. We chose to consider the role of routine deep sedation. Propofol was preferred as a sedative because previous observational studies using other drugs for sedation during colonoscopy in almost 94% of the sample had still reported unacceptably low success rates [26]. In addition, it has recently been demonstrated that deep sedation enhances the polyps detection rate [27].

A weakness of our study is that we did not perform a thorough cost-benefit analysis, particularly as concerns the need for extra personnel to manage sedation. This issue depends largely on differences in the compensation awarded by national health systems and on their related organizational aspects, but relevant data could probably be obtained quite easily by adapting our findings to different situations in different countries. We also collected no details on patients' satisfaction with the procedure, because this information is only collected at our centre in the context of clinical trials and we wished to avoid any Hawthorne effect on the endoscopists' routine practice.

Our study has shown a possible strategy for further improving the rate of successful cecal intubations, i.e. by modifying certain organizational aspects when deep sedation is used.

If a "difficult patient" (aged over 71 and weighing less than 60 kg) has to undergo colonoscopy at the hands of a "less experienced non-specialist" (instead of a "highly experienced specialist"), then inadequate colon preparation reduces the chances of the colonoscopy being completed by 70%. So, special attention should be paid to colon-cleansing practices for colonoscopies that are to be handled by a less experienced endoscopist. Moreover, if any unwanted effects of propofol infusion demanding AAI can be avoided by a careful management of sedation, then 60% of incomplete procedures could be successful. In other words, if a difficult colonoscopy is programmed (due to the patient's characteristics), then containing the need for AAI and improving bowel cleansing enables the successful completion of 90% of incomplete colonoscopies whatever the expertise of the endoscopist involved.

In conclusion, where there are workforce shortages and at large endoscopic practices where colonoscopies are handled by doctors with different levels of training and experience, the extensive use of propofol enables the standard success rate for screening colonoscopies to be approached. An excellent result can be achieved, in terms of the completeness of the procedures, especially if deep sedation is combined with a careful preliminary assessment of the workload so that patients listed for endoscopic procedures who might prove particularly difficult to handle (a situation that is readily identifiable from simple, known features) are not assigned to endoscopists without a high level of specialist experience.