Coffee

What it is

Coffee is an infusion prepared from the roasted beans of Coffea plant species, consumed worldwide as a source of caffeine and bioactive polyphenols [1]. Typical brewed coffee provides 70 to 140 mg of caffeine per 200 mL serving, alongside chlorogenic acids and micronutrients [2] [3].

How it might work

Proposed mechanisms, not yet proven, include thermogenic, antioxidant and anti-inflammatory effects, adenosine receptor signalling, and changes to the gut microbiome [4].

What the research says

Cohort meta-analyses link coffee with lower all-cause death. One pooled analysis of 40 studies with 3,852,651 people found a relative risk of 0.85 at about 3.5 cups a day against no coffee, with no further drop at higher intake [5]. A meta-analysis of 17 cohorts found a relative risk of 0.87 for 3 to under 5 cups a day, stronger in women [6].

Heart outcomes follow the same pattern. An umbrella review found a relative risk of 0.85 for cardiovascular disease and 0.81 for cardiovascular death at 3 to 4 cups a day against none [1]. Decaffeinated coffee showed similar inverse associations for all-cause mortality [7].

Thirty prospective studies with 1,185,210 people linked the highest intake (median about 5 cups a day) with a relative risk of 0.71 for type 2 diabetes; each extra cup a day was linked with 0.94 [4]. Nine studies linked coffee with a lower risk of cirrhosis, with a relative risk of 0.56 per extra two cups a day, but results varied widely [8].

These findings are observational and cannot show that coffee causes the difference. Two trials on heart rhythm disagree in direction: a six-month open-label trial of 200 people with atrial fibrillation found fewer recurrences with about one cup a day (hazard ratio 0.61, 47% against 64%) [9], while a 14-day trial of 100 adults found more premature ventricular beats on coffee days [10].

Doses used in studies

The lowest death risk in pooled cohorts was at about 3.5 cups a day [5]. The atrial fibrillation trial asked the coffee group to drink at least one cup a day, averaging about one cup [9]. Caffeine challenge studies of panic used 400 to 750 mg, roughly five cups [11].

Safety and interactions

• Side effects: Common side effects of excess intake include insomnia, nervousness, restlessness, irritability, gastrointestinal upset, rapid heart rate, and muscle tremors [3] [12].
• Upper limits and toxicity: Major health authorities suggest an upper limit of 400 mg of caffeine daily for healthy adults, roughly equivalent to 3 to 4 standard cups of coffee [3]. Acute toxicity is rare with brewed coffee, but life-threatening caffeine overdose has been reported at acute doses around 10 grams [3].
• Medication interactions: Caffeine is primarily cleared by hepatic CYP1A2 enzymes; potent CYP1A2 inhibitors such as the antidepressant fluvoxamine and fluoroquinolone antibiotics like ciprofloxacin dramatically reduce caffeine clearance, amplifying side effects [12]. Caffeine also competitively interacts with theophylline and clozapine, and alters renal excretion of lithium [12].
• Pregnancy and breastfeeding: Pregnant women are advised to limit caffeine intake to no more than 200 to 300 mg per day, as higher consumption has been linked to increased risks of low birth weight, preterm birth, and pregnancy loss [3] [1]. Small amounts of caffeine pass into breast milk, where high doses may cause irritability and wakefulness in nursing infants [3].
• Pre-existing conditions: Individuals with severe cardiovascular arrhythmias, uncontrolled hypertension, panic or anxiety disorders, gastroesophageal reflux disease, or elevated bone fracture risk should consult a doctor before consuming high amounts of caffeinated coffee [3] [1].
• Heart rhythm, sleep and panic (trial evidence): In a 14-day randomised trial of 100 adults, caffeinated coffee days had more premature ventricular beats (rate ratio 1.51), and 36 fewer minutes of sleep a night; premature atrial beats did not change significantly [10]. In placebo-controlled studies, 400 to 750 mg caffeine (roughly five cups) caused panic attacks in 51% of 237 people with panic disorder [11].

Caveats and open questions

Most of the mortality, heart, diabetes and liver findings come from observational cohorts, which cannot show cause and effect. The umbrella review authors call for randomised trials to test whether these links are causal [1]. Most of the sleep studies reviewed were in male adults in Western countries [13]. The trials are short or small: 200 people in the atrial fibrillation trial [9] and 100 adults over 14 days in the rhythm trial [10]. A 2022 benefit-risk analysis was written by consultants whose firm received consulting fees from the National Coffee Association [14].