Statin and COX-2 Inhibitor Exposure Is Associated with Improved Survival in 4913 Newly Diagnosed DLBCL Patients: A Large Population Based Study

Background: In the era of chemo-immunotherapy, risk factors associated with survival in patients with diffuse large B-cell lymphoma (DLBCL) are largely limited to biologic characteristics of disease. Laboratory based studies have postulated that statins (through inhibition of the geranylgeranylation pathways) can induce apoptosis in DLBCL cells; pre-clinical data suggests an anti-cancer potential for metformin (through inhibition of cancer cell growth by activation of AMPK and inhibition of mTOR pathways) and cox-2 inhibitors (anti-proliferative effects through blockade of PI3K pathway, inhibition of angiogenesis and proliferation by blocking eicosanoid receptors). To date, these potential in-vitro benefits have not been demonstrated consistently in "real world" studies. Our objective was to assess the impact of medicines with a biologic potential on lymphoma outcome in the era of rituximab.

Methods: We performed a retrospective population-based study of adults ≥66 years diagnosed with DLBCL or transformed lymphoma treated in Ontario, Canada. Administrative databases held at ICES were used to assess the impact of select medications on patient outcomes. All patients treated, with curative intent, with a rituximab containing regimen between January 2005 and December 2015 were included. A 1-year lookback of medication exposure prior to commencing rituximab was used. Cox regression analyses were performed to determine the relationship between medication exposure and lymphoma outcomes. Additional analyses were completed to control for known confounders of survival, including the number of comorbid conditions.

Results: A total cohort of 4913 patients were treated with a rituximab containing regimen, most frequently R-CHOP, during the study timeframe. Median age was 75 years (IQR 70-80); 51% were male. The median number of cycles of chemotherapy was 6 (IQR 3-6). The median number of comorbidities was 11 (IQR 9-14). Sixty-nine percent had a high comorbidity score (≥10); 26.4% moderate (6-9); and 4.7% low (0-5). Where mortality data was available, 52.1% of the cohort died at a median of 1 year, of whom 67% died due to DLBCL. In the year prior to commencing lymphoma therapy 45.7% received statin therapy; 16.3% metformin; and 25.0% cox-2 inhibitor. In the univariate analysis, statin exposure (HR 0.88; 0.8 - 0.97) was associated with improved survival, but exposure to cox-2 inhibitor (HR 0.82; 0.65 - 1.04) and metformin (HR 1.11; 0.98 - 1.26) had a no impact, during this timeframe.

Adjusting for time varying exposure and demographic variables including socio-economic factors and comorbidities, we demonstrated that additional exposure to statin and cox-2 inhibitors in the 365 days prior to commencing lymphoma therapy was associated with a survival advantage, when compared to those who were never exposed. Statin exposure for 30 days (HR 0.97 [0.96-0.98]), 180 days (HR 0.84 [0.80-0.89]) and 365 days (HR 0.71 [0.63-0.79]) and cox-2 inhibitor exposure for 30 days (HR 0.95 [0.95-0.98]), 180 days (HR 0.76 [0.66-0.86]) and 365 days (HR 0.57 [0.43-0.74]) were independently associated with improved survival. In contrast metformin exposure had no impact on survival in this cohort.

Patients with moderate (HR 1.69; 1.25 - 2.29) and high comorbidity scores (HR 3.16; 2.36 - 4.21) had significantly higher risk of mortality (p<0.0001). Increasing age (p<0.001; HR 1.05 [1.05-1.06]) and male sex (p=0.003; HR 1.13 [1.04-1.23]) were also associated with increased mortality.