[tommiegannert]

Figure 2 in the study says

  Adjusted for all covariates
  Odds
  1.09  None
  1.19  Decaf
  1.02  <1 cup/day
  1.00  1-2 cups/day
  1.02  3-4 cups/day
  1.13  5-6 cups/day
  1.53  >6 cups/day

Earlier they say

> We used light coffee drinkers (1–2 cups/day), instead of nondrinkers, as the reference group to avoid bias from individuals with poor health, who may avoid coffee due to their health status (49, 50).

[boomboomsubban]

Ahh, I could only see the truncated version on the site.

Though that's the reason they claim to have picked that group, as they were using biobank data they likely already knew the rates before the study started. Picking the lowest cohort from the get go is almost more sketchy.

[SAI_Peregrinus]

Given the decaf numbers, it looks like caffeine is having a protective effect on the negative effects from the rest of the coffee!