The effect of opposed versus unopposed hormone replacement therapy (HRT) on the risk of colorectal cancer (CRC).

Abstract

3619 Background: Separate studies suggest that estrogen alone (E) versus estrogen plus progesterone (EP) may have different effects on CRC risk, but data are conflicting. Prior meta-analyses did not address the role of opposed versus unopposed HRT specifically. We conducted a meta-analysis to summarize the relative risks (RR) of CRC due to E versus EP. Methods: From a total of 2,317 articles identified from PubMed and EMBASE, 5 randomized controlled trials (RCTs) and 13 observational studies were eligible for analyses. We considered these variables: study year/size/design/population, duration and recency of HRT use, method of exposure assessment, and outcome definition and ascertainment. RR were synthesized by random-effects models. Results: E was associated with a marginally significant decreased risk of CRC (RR 0.88, 95% CI 0.78-1.00). Adding progesterone significantly strengthened this association (RR 0.74, 0.67-0.82, p value for the difference between E and EP = 0.01). We found significant heterogeneity for E, but not for EP studies. Several factors were predictive of lower RR in the E arm and can potentially explain the heterogeneity, including current use, case-control design, and HRT exposure assessed by self-report (see Table). Recency was less important for the association between EP and CRC risk. Conclusions: There is consistent evidence supporting the association between EP and CRC risk reduction. While literature for the effect of E on CRC risk is inconclusive, our analyses suggest only current use of E is associated with a decreased CRC risk. HRT Variable Subgroups No. Pooled estimates Heterogeneity p values Meta-regression p values HRT E 12 0.88 (0.78-1.00) 0.003 0.01 EP 16 0.74 (0.67-0.82) 0.58 E Design RCT 1 1.08 (0.75-1.55) N/A <0.01 Cohort 4 0.99 (0.81-1.2) 0.14 <0.01 Case-control 6 0.76 (0.62-0.92) 0.01 Ref Recency Current user 3 0.73 (0.54-0.98) 0.03 <0.01 Former user 3 0.93 (0.62-1.41) 0.04 HRT assessed by self-report Yes 6 0.78 (0.63-0.97) 0.001 0.03 No 6 0.96 (0.84-1.09) 0.22 EP Design RCT 4 0.73 (0.57-0.94) 0.49 0.78 Cohort 5 0.76 (0.64-0.9) 0.33 0.85 Case-control 6 0.74 (0.63-0.87) 0.75 Ref Recency Current user 3 0.77 (0.57-1.04) 0.38 0.47 Former user 3 0.64 (0.46-0.89) 0.99 No significant financial relationships to disclose.