Reply to Ishaq et al.

We appreciate the interest shown by Ishaq et al. in our study [1].

Regarding pain at discharge as a measure of discomfort during colonoscopy, we used two different methods of pain measurement. The primary end point was procedural pain, but the patients were asked after the procedure about how painful the procedure had been. Thus, we did not ask about post-procedural pain experienced at the time of discharge, which we agree would not be a reliable measure of procedural discomfort. We chose our approach to ensure that the outcome assessor was blinded to the intervention and, therefore, to minimize the risk of bias.

As Ishaq et al. appropriately point out, the memory of pain during sedated colonoscopy can be modified by leaving the colonoscope without movement in the rectum before completing the procedure [2]. It is not clear, however, whether the same intervention lessens patients’ memory of pain during unsedated colonoscopy, as used in our study. Although we did not intentionally extend the procedure in this way, this source of bias cannot be entirely ruled out. We believe, however, that measuring pain in real time during colonoscopy is prone to similar risks of bias.

Intraprocedural measurement of pain can be performed in several ways. Commonly reported methods of obtaining pain scores involve the endoscopy assistant asking the patient about their level of pain at 2 – 3-minute intervals or at the discretion of the assessor. Because the endoscopy team cannot usually be blinded to an intervention such as the water exchange technique, and because pain during colonoscopy is usually fluctuating and high pain levels may be brief, these nonstandardized approaches involve a considerable risk of bias. Moreover, in the same manner as the endoscopists may influence recalled pain by prolonging withdrawal from the rectum, they may also influence pain scores by altering insertion force, speed, and adjunct maneuvers, such as gas suctioning and loop reductions. In an attempt to overcome these sources of bias and to obtain the true value of maximum pain during insertion, the endoscopy assistants in our trial asked the patients to score the highest level of pain experienced during and after negotiating each of the five colonic segments during insertion. This approach, we propose, involves a higher level of standardization than time intervals or the individual judgment of endoscopy assistants.

Ishaq et al. claim that the 4-point verbal rating scale (VRS-4) lacks ratio properties and is therefore less suitable for evaluating pain magnitude than a numerical rating scale or a visual analog scale. We acknowledge that the possibility to extrapolate pain scores from one rating scale to another is limited. However, this does not reduce the internal validity of our results. Pain levels reported in colonoscopy trials vary widely, probably as a result of differences in patient selection, pain measurement methods, and sedation practice. We argue that the VRS-4 used in our trial provided valid information about the differences between the intervention arms. We also argue that our approach to not exclude patients who had previously undergone colonoscopy, as well as the population-based design with few exclusions, served to reduce the overall risk of selection bias.

Finally, we would like to highlight that pain measurement during colonoscopy should serve a greater purpose than research alone. Measuring procedural pain post-procedure can only benefit future patients if the results are monitored, analyzed, and used to improve performance. Furthermore, pain recalled by the patient after the procedure is probably a better predictor of patient return rates for follow-up colonoscopy than real-time assessment [2]. Real-time pain measurement, on the other hand, can be used during the procedure to adjust insertion technique and guide sedation levels. Because the vast majority of patients in our trial were unsedated, the results reflect the true levels of discomfort provoked by colonoscopy. However, the best method of evaluating patients’ experiences of colonoscopy remains to be established.

• References

• 1 Garborg K, Kaminski MF, Lindenburger W et al. Water exchange versus carbon dioxide insufflation in unsedated colonoscopy: a multicenter randomized controlled trial. Endoscopy 2015; 47: 192-199
• 2 Redelmeier DA, Katz J, Kahneman D. Memories of colonoscopy: a randomized trial. Pain 2003; 104: 187-194