Nonsteroidal Anti-inflammatory Drugs But Not Aspirin Are Associated With a Lower Risk of Post-colonoscopy Colorectal Cancer

Ka Shing Cheung; Lijia Chen; Esther W. Chan; Wai Kay Seto; Ian C.K. Wong; Wai K. Leung

Disclosures

Aliment Pharmacol Ther. 2020;51(9):899-908. 

In This Article

Results

Patient Characteristics and Risk of PCCRC-3y

Of the 234 827 patients who had undergone colonoscopy between 2005 and 2013, 187 897 (male: 91 961 [48.9%]) patients fulfilled the selection criteria (Figure 1), with a total duration of follow-up of 560 471 person-years. The median age at index colonoscopy was 60.6 years (IQR: 52.3–71.9).

In total, there were 854 PCCRC-3y cases including 707 (82.8%) distal and 147 (17.2%) proximal cancers with an overall incidence rate of 15.2 per 10 000 person-years. The median age of diagnosis of PCCRC-3y was 75.9 years (IQR: 65.5–83.8); and the median interval between index colonoscopy and PCCRC-3y was 1.2 years (IQR: 0.8–1.9).

Association Between NSAID use and PCCRC-3y

There were 21 757 NSAID users and the median duration of NSAID use was 0.7 years (IQR: 0.4–1.6) within 5 years preceding index colonoscopy. Among them, 55 (0.25%) patients were diagnosed with PCCRC-3y with an incidence rate of 8.4 per 10 000 person-years. For NSAID nonusers, the incidence rate of PCCRC-3y was 16.1 per 10 000 person-years.

On crude analysis, the HR of PCCRC-3y with NSAID use was 0.53 (95% CI: 0.40–0.69). On PS regression adjustment, the aHR of PCCRC-3y with NSAID use was 0.54 (95% CI: 0.41–0.70) (Table 2). Stratified analysis shows that NSAID use was associated with a lower PCCRC-3y risk in both proximal (aHR: 0.48, 95% CI: 0.24–0.95) and distal colon (aHR: 0.55, 95% CI: 0.40–0.74). As for aspirin, the aHR of PCCRC-3y was 1.01 (95% CI: 0.80–1.28; P = 0.92).

Effects of Individual NSAID on PCCRC-3y Risk

Of the 21 757 NSAID users, 3545 used more than one type of NSAIDs and were excluded in this analysis. Table 3 shows the effects of individual NSAIDs on PCCRC-3y risk. Diclofenac (n = 10 648) and naproxen (n = 2675) were the two NSAIDs found to be associated with a reduced PCCRC-3y risk (diclofenac, aHR: 0.48, 95% CI: 0.33–0.73; naproxen, aHR: 0.38, 95% CI: 0.16–0.92). There were no statistically significant association between PCCRC-3y and other NSAIDs (ibuprofen, mefenamic acid, indomethacin, sulindac, piroxicam and ketoprofen).

Duration- and Frequency Response Between NSAID use and PCCRC-3y

Table 4 shows that when compared with never user, a longer duration of NSAID use (>1 year) offers greater protection against PCCRC-3y (aHR: 0.42, 95% CI: 0.26–0.65) than shorter duration (≤1 year) of NSAID use (aHR: 0.53, 95% CI: 0.45–0.62; P trend < 0.001). Similar findings were observed for both proximal and distal cancers.

When compared with never user, more frequent NSAID use (≥weekly) also offers greater protection against PCCRC-3y (aHR: 0.46, 95% CI: 0.32–0.67) than infrequent (<weekly) NSAID use (aHR: 0.53, 95% CI: 0.45–0.61; P trend < 0.001). This finding was again observed for both proximal and distal cancers.

Subgroup Analysis

Table 5 shows that protective effect of NSAIDs was limited to patients aged ≥60 years (aHR: 0.48, 95% CI: 0.35–0.66) and patients without diabetes mellitus (aHR: 0.55, 95% CI: 0.41–0.73). The aHR of PCCRC-3y with NSAIDs was lower in females (aHR: 0.43, 95% CI: 0.28–0.66) than in males (aHR: 0.63, 95% CI: 0.44–0.91). NSAIDs were also associated with a significantly lower PCCRC-3y risk in those without the history of colonic polyps (aHR: 0.46, 95% CI: 0.32–0.67) but not in those with the history of colonic polyps (aHR: 0.67, 95% CI: 0.45–1.01).

Association Between NSAID use and Secondary Outcomes (PCCRC-all and PCCRC>3y)

We further looked into the effects of NSAIDs on PCCRC that developed in different time frames after index colonoscopy. There were a total of 1290 PCCRC-all cases (ie all CRC cases diagnosed >6 months after index colonoscopy) including 436 (0.2%) PCCRC>3y (median: 5.2 years; IQR: 3.7–7.2). While NSAID use was associated with a lower risk of PCCRC-all (aHR: 0.58, 95% CI: 0.50–0.76; P < 0.001), the benefit was not observed for PCCRC>3y (aHR: 0.78, 95% CI: 0.56–1.09; P = 0.149).

Results From PS Matching

Before PS matching, the majority of covariates were well balanced (ASD < 0.2), except for sex (Table S2). After PS matching, the cohort number was 64 806 including 21 650 NSAID users and 43 156 NSAID nonusers, with good balance of all covariates (ASD < 0.2). There were 246 (0.4%) PCCRC-3y cases in this matched cohort and NSAID use was also associated with a lower PCCRC-3y risk (HR: 0.57, 95% CI: 0.42–0.77).

Effects of Current or Past NSAID use on PCCRC

The aHR of PCCRC-3y with current and past NSAID use was 0.55 (95% CI: 0.43–0.71) and 0.58 (95% CI: 0.49–0.69) respectively (Table S3). The corresponding aHR of PCCRC-all with current and past NSAID use was 0.61 (95% CI: 0.50–0.74) and 0.65 (95% CI: 0.57–0.74) respectively. The aHR of PCCRC>3y with current and past NSAID use was 0.74 (95% CI: 0.53–1.02) and 0.80 (95% CI: 0.64–0.99) respectively.

Effects of NSAID use After Index Colonoscopy on PCCRC

The aHR of PCCRC-3y with post-colonoscopy NSAID use was 0.50 (95% CI: 0.28–0.91), while the aHR of PCCRC-all and PCCRC>3y was 0.40 (95% CI: 0.28–0.57) and 0.64(95% CI: 0.40–1.01) respectively (Table S4).

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....