Table 5

Immunohistochemistry results

MarkerPlacebo (n = 13)Celecoxib (n = 22)P for treatment effects and interaction effect*
BaselinePosttreatmentP, pretreatment-posttreatment differenceBaselinePosttreatmentP, pretreatment-posttreatment differenceCelecoxib vs placeboPretreatment-posttreatmentNormal mucosa vs ACF
COX-2 IHC score
 Normal mucosa3.003.330.542.812.811.000.460.640.82
 ACF4.442.330.023.063.061.000.620.01
CD31 IHC score
 Normal mucosa17.6722.420.0920.8121.050.930.770.210.31
 ACF22.8923.780.7221.5624.380.340.910.39
Ki-67 IHC score
 Normal mucosa12.7411.510.3211.0814.060.0520.770.41<0.0001
 ACF28.9423.570.3230.7321.650.0450.980.04
SMAD4 (% intact stain)
 Normal mucosa38500.3236790.080.160.29
 ACF,§38750.1854620.650.620.08
ERα (% intact stain)
 Normal mucosa00000.611.00
 ACF,§17330.560230.710.91

Abbreviation: IHC, immunohistochemistry.

  • *Using repeated ANOVA.

  • ACF vs adjacent normal comparison is based on subjects with marker levels for pre- and post-magnification chromoendoscopy biopsies as well as for both normal and ACF tissue.

  • A total of 45 patients participated in this magnification chromoendoscopy substudy. Of these, 10 patients were taking cardioprotective doses of aspirin, and to avoid bleeding complications these individuals had rectal ACF counted but did not undergo biopsy of either ACF or normal mucosa. In addition, the numbers of cases studied by immunohistochemistry for SMAD4 and ERα were smaller still due to insufficient remaining tissue in some cases (numbers given in Table 6).

  • §SMAD4 and ERα are dichotomized as intact (SMAD4 = 3; ERα = 3) or deficient (SMAD4 <3; ERα <3). McNemar's test for paired proportions is used to assess whether there are significant baseline to 8-mo differences in the percent with SMAD4 intact or ERα intact.