In these experiments, I used a home blood pressure monitor (upper arm cuff), to measure my blood pressure in the morning, on an empty stomach, after drinking a large cup of caffeine coffee.
To prepare for each set of readings, I sat down for 5 minutes before taking the first reading.
The coffee used in this experiment was strong filtered coffee. After taking a few baseline readings, I drank the coffee in one go. The amount was equivalent to 4 small expresso coffees. I then continued to take my blood pressure every few minutes. I then sat as still as possible, taking my blood pressure every three minutes.
Here are my results:
|I DRANK THE COFFEE AT THIS POINT|
Here is the data as a graph showing the effect of caffeine coffee on blood pressure:
You'll see that there is an almost immediate increase in systolic pressure, followed by a drop. This then increases again about 10 minutes after drinking the coffee to a maximum measured blood pressure of 138/89 about 23 minutes after the coffee was drunk. The blood pressure then starts dropping again, until normal blood pressure is restored at around 30 minutes after drinking the coffee.
What we see happening is a temporary increase in blood pressure in response to the coffee, which the body then corrects. How caffeine raises blood pressure seems to be unclear, but what is certain in my case is that it did raise my blood pressure, but only for a short period of time.
For those with high blood pressure, it would appear that coffee could be a problem due to the high spike after drinking the coffee. However, since my blood pressure is normal, I wont be switching to decaf.
My next experiment is to see what effect drinking filtered decaffeinated coffee has. Is it the caffeine in the coffee that raised my blood pressure, or something else? You can see the results of my experiment on the does decaffeinated coffee affect blood pressure page. The results may surprise you!
Want to Buy a Blood Pressure Monitor?
I recommend you get an upper arm monitor rather than the wrist variety, as my own testing shows them to be more accurate. The model I recommend is this one: