Background: Chemotherapy at the end of life (CEOL) is widely accepted as an indicator of aggressive care. However, the evidence is limited primarily to single-centre experiences, with no consensus regarding acceptable benchmarks for CEOL, nor how this may be changing over time and with novel treatment options. We describe ‘real world’ CEOL in two large, multisite Australian registries of metastatic colorectal cancer (mCRC) and both locally advanced and metastatic pancreatic cancer (PC). Methods: Data was analysed from the TRACC and PURPLE registries, two large prospective multisite Australian cancer registries collecting prospective demographic, tumour, treatment and outcome data for mCRC and PC respectively. We identified all decedents between May 2009 and November 2020, determined the proportion who died within 14 or 30 days of chemotherapy (14D / 30D), or within 30D after a new line of therapy, defined as the first cycle of a new treatment regimen. Using univariate analysis, we compared baseline demographic and clinicopathological variables and trends over time. Results: 1505 mCRC and 602 PC decedents were identified. 20.9% of decedents (21.6% mCRC and 19.3% PC) received chemotherapy within 30D, and 11.5% within 14D (12.3% mCRC and 9.6% PC). There were lower rates of 30D CEOL after the first cycle of a new line of therapy (4.3% mCRC and 5.7% PC). Rates of CEOL decreased over the study period for mCRC (median rate of initial 3-year period 28% versus 15% in last 3-year period), but remained largely static for PC (18.9% versus 17.9%). 30D CEOL was more likely with palliative than curative intent treatment (mCRC OR 1.6, 95% CI 1.14-2.25, p = 0.007, PC OR 5.3, 95% CI 1.6-17.8), and advanced rather than local disease in PC (PC OR 2.59, 95%CI 1.6-4.1, p < 0.001). There was a trend towards CEOL and poorer performance status (ECOG) across all groups, only significant for 30D CEOL mCRC (OR 1.51, 95% CI 1.04-2.2). There was no association between CEOL and patient age, gender, Charlson comorbidity score or lines of therapy. Conclusions: Real-world rates of CEOL are higher in our cohorts of mCRC and PC patients than historical benchmarks but comparable to contemporary reports, which may be due to a wider array of available active treatments. Overall rates are decreasing over time for mCRC but static for PC, which may reflect the poorer overall survival for PC and lack of new effective therapies. The lower rates of death after new lines of therapy may signify that CEOL is more likely with existing treatment regimens and that clinicians are less likely to initiate a new chemotherapeutic regimen at EOL.