The study design for the APC trial has previously been summarized here (Cardiac Risk of Cox2: Fact or Fiction?). The design of the PreSAP trial was similar to the APC trial, with inclusion and exclusion criteria, outcomes, study interval, and plan for stratification by low-dose aspirin use largely identical between the studies. Subjects with prior CVD were enrolled in both studies. The important differences in study design were that the APC trial investigated celecoxib 400 mg or 800 mg per day against placebo, while the PreSAP trial investigated only one dose of celecoxib, 400 mg per day, against placebo.
Subjects in both trials underwent repeat colonoscopy after 1 and 3 years of treatment. The recurrence of at least one colonic adenoma at either of these timepoints was the designated primary outcome.
2035 subjects were randomized into the APC trial (divided between the 3 study groups) and 1561 into the PreSAP trial (divided between the 2 study groups). In the APC trial, approximately 90 and 75 percent of randomized subjects underwent repeat colonoscopy at years 1 and 3, respectively. In the PreSAP trial, approximately 90 and 80 percent of randomized subjects underwent repeat colonoscopy at years 1 and 3, respectively.
Median age (59 61) and gender proportion (approximately 65 70 % male) were similar for both studies. The APC trial included a slightly higher percentage of both non-Hispanic white and black subjects compared to the PreSAP trial. In addition, subjects in the APC trial had a slightly higher body mass index (by approximately 1 BMI unit in both men and women), a higher percentage of subjects receiving therapy with low-dose aspirin (approximately 31 vs. 17 %), a slightly greater percentage of subjects reporting prior CVD events (approximately 14% vs. 12%), and a greater percentage of patients reporting hypertension (approximately 41% vs. 36%) compared to those in the PreSAP trial. The prevalence of diabetes was similar in both trials while the PreSAP trial had a higher percentage of current smokers (approximately 23% vs. 16%).
In both trials, celecoxib treated patients demonstrated lower incidence of recurrent adenomatous polyps at 3 years. In the APC trial, a 57% reduction in the celecoxib 400mg per day group (RR 0.43 (95%CI 0.31 0.61)) and a 66% reduction in the celecoxib 800 mg per day group (RR 0.34 (95% CI 0.24 0.50) on the risk of polyp recurrence compared to placebo was noted. In the PreSAP trial, a 36% reduction in the incidence of recurrent polyps was noted compared to placebo (RR 0.64 (95% 0.56 0.75)). Similar findings were found in both trials on the 3 year cumulative incidence of advanced adenomas. The chemoprotective effect of celecoxib was not modified by aspirin in either trial.
In both trials, celecoxib treated patients were found to have higher blood pressure than placebo treated patients. In addition, aspirin treated patients receiving celecoxib demonstrated a trend to higher rates of gastrointestinal ulceration and bleeding.
Composite CVD outcomes for the APC trial as published are slightly different than those reported previously. Cardiovascular events (any CVD death, MI, CVA, CHF) occurred in 18 (2.6%) subjects in the 400 mg per day group, 23 (3.4%) subjects in the 800 mg per day group, and 7 (1.0%) subjects in the placebo group. This translated to a relative risk of 2.6 (95%CI 1.1 6.1) for celecoxib 400 mg per day compared to placebo and 3.4 (95% CI 1.5 7.9) for celecoxib 800 mg per day compared to placebo. In the PreSAP trial, the risk of the adjudicated composite CVD outcome (different from the APC trial-in that only CVD was included) was not significantly higher in the celecoxib 400 mg per day subjects compared to placebo, with 23 (2.5%) of celecoxib treated patients and 12 (1.9%) of placebo treated patients reporting incident CVD events (RR 1.3 (95%CI 0.65 2.62). A separate, non-adjudicated CVD endpoint that included CVD death and non-fatal CVD outcomes was increased in celecoxib treated subjects in the PreSAP trial who were not using concomitant aspirin (RR 1.74 (95%CI 1.01 3.01)) but not in those who were using concomitant aspirin (RR 1.34 (95%CI 0.72 2.48))
Celecoxib at a dose of 400 or 800 mg per day is effective at reducing the incidence of recurrent adenomatous polyps. This protective effect is leveraged against a potential increase in the risk of fatal and non-fatal CVD events. The concomitant administration of low-dose aspirin may help abrogate this increased risk, at the potential expense of an increased risk for gastrointestinal ulceration and bleeding.
No pharmacologic agent has previously demonstrated chemoprotective effects against the recurrence of adenometous polyps to the magnitude demonstrated here. However, it remains to be shown the ultimate impact that prophylaxis with prolonged COX-2 inhibitor use would have on colorectal cancer morbidity and mortality. It is possible that if this benefit were large, it may outweigh the small but significant increase in CVD outcomes seen in these studies.
Several important issues are highlighted by the comparison between these two trials. For one, there may be a dose-response relationship in both chemoprotection and CVD risk. It remains to be shown whether a clinically significant chemoprotective benefit could be obtained with a smaller dose of celecoxib (i.e. 200 mg per day) in exchange for lower CVD risk. Second, CVD risk may be increased in patients with prior CVD disease or greater numbers of risk factors, as evidenced by the fact that lower risk was seen in the PreSAP trial, in which patients tended to have a lower prevalence of prior CVD disease and CVD risk factors, than in the APC trial.
What do these results mean for patients using celecoxib for reasons other than chemoprotection against recurrent adenometous polyps? First, there may be an added benefit against the emergence of colonic polyps and cancers. However, this has not yet been confirmed in patients without a prior history of polyps (primary prevention). It has also not been confirmed whether non-selective NSAIDs may provide the same benefits as COX-2 selective agents. Concomitant low-dose aspirin may help to reduce CVD risk. However, the gastroprotective effects of COX-2 inhibition will also be reduced when aspirin in added, limiting the benefit of using COX-2 inhibitors over non-selective NSAIDs for pain relief. Thus, current recommendations would still advise careful selection of patients to receive chronic therapy with COX-2 selective agents.