Association between coffee and tea consumption and the risk of dementia in individuals with hypertension: a prospective cohort study

Abstract

Many studies have shown that drinking coffee and tea may be associated with the risk of hypertension and dementia. Limited research exists on their impact on dementia risk in hypertensive patients. This study aimed to determine the association between coffee and tea consumption and the risk of dementia development in hypertensive population by utilizing Cox proportional risk modeling with 453,913 participants from a UK biobank. Our findings reveal a J-shaped and U-shaped association between the risk of all-cause dementia and the consumption of coffee and tea respectively in hypertensive people. The hypertensive patients who drink 0.5–1 cup of coffee or 4–5 cups of tea per day have the lowest risk of dementia. A U-shaped relationship was observed between daily caffeine consumption and the risk of developing all-cause dementia and vascular dementia in the hypertensive population. Furthermore, the significant association between the amount of coffee and tea consumed and the risk of all-cause and vascular dementia were more likely to be found in hypertensive patients than in the non-hypertensive population.

Introduction

Hypertension and related complications are major public health issues with high morbidity and mortality, affecting a third of adults worldwide: In 2019, 1.3 billion people worldwide suffered from hypertension, with prevalence rates as high as 33 percent among people aged 30–79 years¹. Reducing systolic blood pressure to 110–115 mm Hg would prevent 11.8 million deaths in 2019¹. However, this goal has not yet been achieved, resulting in a huge health and economic burden on society. Therefore, preventing hypertension and related complications is paramount. Dementia commonly occurs in individuals with hypertension, with around 11 to 20 percent of dementia cases linked to abnormal blood pressure². Hypertension accelerates cognitive aging and increases the risk of Alzheimer's and other dementias^3,4. Preventing hypertension not only reduces the burden of hypertension but also has the potential to significantly decrease the occurrence of dementia in the medium to long term. For example, the incidence of dementia could be reduced by 40 percent through managing and controlling 12 risk factors such as hypertension⁴. It is therefore of great public health benefit to investigate measures to delay or prevent the onset and development of dementia in people with hypertension. Caffeine, the active ingredient in coffee, tea, and other beverages, is currently the most widely used stimulant in the world and plays a neuroprotective role in the fight against neurodegenerative diseases⁵. Therefore, high caffeine beverages such as coffee and tea are currently attracting significant interest in the field of dementia prevention. Longitudinal studies have shown that the type and amount of coffee and tea consumed are significantly associated with the risk of developing dementia in the future^6,7. Although the mechanism has not been fully elucidated, experts agree coffee and tea components such as caffeine may reduce dementia risk by decreasing neuroinflammation or providing neuroprotective benefits^8–10. Research in animal models demonstrates that drinking caffeinated beverages can not only lower the risk of hypertension, but also inhibit Aβ peptide production in the mice brain^11–13, reduce pro-inflammatory cytokine production¹⁴, and maintain blood–brain barrier function¹⁵. These benefits may slow down the progression of vascular inflammation, blood–brain barrier leakage, and β-amyloid microvascular deposition in hypertensive patients¹⁶, potentially delaying the onset of Alzheimer's disease. In addition, chlorogenic acid and caffeic acid in coffee and tea not only may improve symptoms of hypertension and reduce blood pressure^17,18, but also inhibit acetylcholinesterase (a key enzyme in the pathogenesis of Alzheimer's disease), reduce lipid peroxidation, prevent neuronal damage, and delay neurodegenerative disease¹⁹. While most experimental results are derived from animal studies and observational human experiments, these positive effects may also benefit humans in preventing or delaying dementia progression in hypertension patients. In addition, the association between coffee and the risk of developing dementia varies according to the type of coffee²⁰. Caffeinated coffee lowers dementia risk more than decaf coffee⁶.

To date, there have been limited studies on the relationship between coffee and tea consumption and the risk of dementia in people with hypertension, and it is unclear whether the benefits of these two drinks on dementia also apply to people with hypertension. To addressthese gaps, we employed information from a large and representative source of data, the UK Biobank (UKB), and performed a comprehensive analysis of the association between coffee and tea consumption and the risk of dementia among individuals with hypertension and compared these with the population without hypertension (Fig. 1).

Fig. 1.

Flowchart of participants included in the analysis.

Results

Baseline characteristics of the participants

Table 1 displays participant characteristics. Out of 453,913 participants, 54.62% had hypertension. The average age was 72.39 ± 8.11 years, with 54.33% being female and 94.13% being white. During an average follow-up time of 15.12 years, the population with hypertension appeared to have a higher incidence of all-cause dementia (3.11% vs 1.11%), Alzheimer’s disease (1.37% vs 0.50%), and vascular dementia (0.76% vs 0.18%), along with a shorter median survival compared to those without hypertension. There were statistically significant differences in age, sex, ethnicity, education level, occupation, TDI, BMI, smoking status, dietary patterns, alcohol consumption, physical activity, social relationships, duration of sedentary behavior, family history of dementia, APOE genotype, depression, history of cancer, history of cardiovascular disease, aspirin use status, lipid-lowering medications use status, dyslipidemia, hypertriglyceridemia, and serum 25(OH)D status between population with hypertension and without hypertension (P < 0.001). After adjusting those covariates, population with hypertension appear to be at a greater risk of developing all-cause dementia (HR = 1.149, 95% CI 1.059–1.247, P = 0.001), Alzheimer’s disease (HR = 1.137, 95% CI 1.006–1.286, P = 0.040), and vascular dementia (HR = 1.281, 95% CI 1.061–1.548, P = 0.010) compared to the population without hypertension (Table 2). Additionally, a multiple-adjusted restricted cubic spline with three knots was utilized to illustrate the relationship between coffee, tea, and caffeine consumption and all-cause dementia. Findings indicated that tea and caffeine intake exhibited a non-linear association with all-cause dementia (p for nonlinearity < 0.001), whereas coffee intake demonstrated a predominantly linear connection with all-cause dementia (P for nonlinearity > 0.001) (Figs. 2 and 3).

Table 1.

Characteristics of participants with or without hypertension.

Characteristic	Total population (N = 453,913)	Participants without hypertension (N = 205,991)	Participants with hypertension (N = 247,922)	χ2/F	P-value
Age, years, mean (SD)	72.39 (8.11)	69.67 (8.09)	74.65 (7.40)	46,683.806	< 0.001
Sex, n (%)				8348.035	< 0.001
Female	246,589 (54.33)	127,171 (61.74)	119,418 (48.17)
Male	207,324 (45.67)	78,820 (38.26)	128,504 (51.83)
Race/ethnicity, n (%)				310.136	< 0.001
White	427,270 (94.13)	193,394 (93.88)	233,876 (94.33)
Asian	10,505 (2.31)	5087 (2.47)	5418 (2.19)
Black	7577 (1.67)	3049 (1.48)	4528 (1.83)
Other	6965 (1.53)	3730 (1.81)	3235 (1.30)
Missing	1596 (0.35)	731 (0.35)	865 (0.35)
Occupation status, n (%)				19,256.755	< 0.001
Employed	263,718 (58.10)	140,752 (68.33)	122,966 (49.60)
Unemployed	34,760 (7.66)	16,214 (7.87)	18,546 (7.48)
Retired	151,148 (33.30)	47,005 (22.82)	104,143 (42.01)
Missing	4287 (0.94)	2020 (0.98)	2267 (0.91)
Education level, n (%)				3653.492	< 0.001
College or University	146,710 (32.32)	77,894 (37.81)	68,816 (27.76)
A level/AS level or equivalent	50,524 (11.13)	25,160 (37.81)	25,364 (10.23)
O level/GCSEs or equivalent	96,903 (21.35)	43,773 (21.25)	53,130 (21.43)
CSEs or equivalent	25,151 (5.54)	12,572 (6.10)	12,579 (5.07)
NVQ or HND or HNC or equivalent	29,610 (6.52)	11,361 (5.52)	18,249 (7.36)
Other professional qualifications eg: nursing, teaching	23,372 (5.15)	9295 (4.51)	14,077 (5.68)
Missing	81,643 (17.99)	25,936 (12.59)	55,707 (22.47)
TDI, n (%)				17.328	< 0.001
Low	87,330 (19.24)	39,375 (19.11)	47,955 (19.34)
Medium	135,265 (29.80)	62,012 (30.10)	73,253 (29.55)
High	230,778 (50.84)	104,347 (50.66)	126,431 (51.00)
Missing	540 (0.12)	257 (0.12)	283 (0.11)
BMI (kg/m2), n (%)				28,081.082	< 0.001
< 18.5	2373 (0.52)	1711 (0.83)	662 (0.27)
18.5–24.9	147,748 (32.55)	89,787 (43.59)	57,961 (23.38)
25–29.9	192,247 (42.35)	82,717 (40.16)	109,530 (44.18)
≥ 30	109,609 (24.15)	31,054 (15.08)	78,555 (31.69)
Missing	1936 (0.43)	722 (0.35)	1214 (0.49)
Dietary patterns, n (%)				657.363	< 0.001
Poor dietary pattern	66,814 (14.72)	27,654 (13.42)	39,160 (15.80)
Medium dietary pattern	234,272 (51.61)	106,052 (51.48)	128,220 (51.72)
Healthy dietary pattern	152,827 (33.67)	72,285 (35.09)	80,542 (32.49)
Smoking status, n (%)				2180.724	< 0.001
Never	247,786 (54.59)	117,644 (57.11)	130,142 (52.49)
Previous	157,209 (34.63)	64,113 (31.12)	93,096 (37.55)
Current	47,248 (10.41)	23,606 (11.46)	23,642 (9.54)
Missing	1670 (0.37)	628 (0.30)	1042 (0.42)
Alcohol consumption, n (%)				4063.858	< 0.001
Quartile 1	113,540 (25.01)	51,864 (25.18)	61,676 (24.88)
Quartile 2	113,541 (25.01)	56,615 (27.48)	56,926 (22.96)
Quartile 3	113,354 (24.97)	54,793 (26.60)	58,561 (23.62)
Quartile 4	113,478 (25.00)	42,719 (20.74)	70,759 (28.54)
Physical activity, n (%)				185.346	< 0.001
Low	64,696 (14.25)	28,971 (14.06)	35,725 (14.41)
Medium	178,177 (39.25)	85,064 (41.30)	93,113 (37.56)
High	108,952 (24.00)	50,311 (24.42)	58,641 (23.65)
Missing	102,088 (22.49)	41,645 (20.22)	60,443 (24.38)
Social conntact, n (%)				153.235	< 0.001
Active	309,765 (68.24)	142,438 (69.15)	167,327 (67.49)
Moderately	126,510 (27.87)	55,981 (27.18)	70,529 (28.45)
Isolated	17,638 (3.89)	7572 (3.68)	10,066 (4.06)
Duration of sedentary behaviour, n (%)				4707.333	< 0.001
< 2	25,813 (5.69)	14,027 (6.81)	11,786 (4.75)
2–3.9	130,518 (28.75)	66,999 (32.53)	63,519 (25.62)
4–5.9	164,966 (36.34)	72,658 (35.27)	92,308 (37.23)
≥ 6	132,616 (29.22)	52,307 (25.39)	80,309 (32.39)
Family history of dementia, n (%)				55.955	< 0.001
No	348,587 (76.80)	154,458 (74.98)	194,129 (78.30)
Yes	64,315 (14.17)	29,524 (14.33)	34,791 (14.03)
Missing	41,011 (9.03)	22,009 (10.68)	19,002 (7.66)
APOE Ƹ4 genotype, n (%)				5.490	0.019
No	336,593 (74.15)	152,519 (74.04)	184,074 (74.25)
Yes	107,001 (23.57)	48,923 (23.75)	58,078 (23.43)
Missing	10,319 (2.27)	4549 (2.21)	5770 (2.33)
Depressive status, n (%)				2009.694	< 0.001
No	345,946 (76.21)	150,592 (73.11)	195,354 (78.80)
Yes	107,967 (23.79)	55,399 (26.89)	52,568 (21.20)
History of cancer, n (%)				265.245	< 0.001
No	412,552 (90.89)	188,793 (91.65)	223,759 (90.25)
Yes	41,361 (9.11)	17,198 (8.35)	24,163 (9.75)
History of cardiovascular disease, n (%)				4220.394	< 0.001
No	435,404 (95.92)	201,901 (98.01)	233,503 (94.18)
Yes	18,509 (4.08)	4090 (1.99)	14,419 (5.82)
Aspirin use status, n (%)				11,625.82	< 0.001
No	391,443 (86.24)	190,101 (92.29)	201,342 (81.21)
Yes	62,470 (13.76)	15,890 (7.71)	46,580 (18.79)
Lipid-lowering medications use status, n (%)				23,460.885	< 0.001
No	375,165 (82.65)	189,709 (92.10)	185,456 (74.80)
Yes	78,748 (17.35)	16,282 (7.90)	62,466 (25.20)
Hypertriglyceridemia, n (%)				9586.886	< 0.001
No	255,608 (56.31)	131,761 (63.96)	123,847 (49.95)
Yes	170,386 (37.54)	61,851 (30.03)	108,535 (43.78)
Missing	27,919 (6.15)	12,379 (6.01)	15,540 (6.27)
Dyslipidemia, n (%)				686.504	< 0.001
No	312,253 (68.79)	144,665 (70.23)	167,588 (67.60)
Yes	77,924 (17.17)	32,032 (15.55)	45,892 (18.51)
Missing	63,736 (14.04)	29,294 (14.22)	34,442 (13.89)
Serum 25(OH)D status, n (%)				462.89	< 0.001
Severely deficient	52,186 (11.50)	22,460 (10.90)	29,726 (11.99)
Moderately deficient	169,510 (37.34)	75,655 (36.73)	93,855 (37.86)
Insufficient and above	185,483 (40.86)	88,062 (42.75)	97,421 (39.30)
Missing	46,734 (10.30)	19,814 (9.62)	26,920 (10.86)
Coffee intake, n (%)				36.184	< 0.001
None	100,199 (22.07)	46,151 (22.40)	54,048 (21.80)
0.5–1	124,475 (27.42)	56,294 (27.33)	68,181 (27.50)
2–3	140,858 (31.03)	63,713 (30.93)	77,145 (31.12)
4–5	60,530 (13.34)	27,046(13.13)	33,484 (13.51)
≥ 6	27,851 (6.14)	12,787 (6.21)	15,064 (6.08)
Coffee type, n (%)				1337.405	< 0.001
Decaffeinated coffee	68,152 (15.01)	31,205 (15.15)	36,947 (14.90)
Instant coffee	195,778 (43.13)	83,803 (40.68)	111,975 (45.17)
Ground coffee	80,658 (17.77)	40,630 (19.72)	40,028 (16.15)
Other type of coffee	6547 (1.44)	3030 (1.47)	3517 (1.42)
Missing	102,778 (22.64)	47,323 (22.97)	55,455 (22.37)
Tea intake, n (%)				77.339	< 0.001
None	66,366 (14.62)	30,459 (14.79)	35,907 (14.48)
0.5–1	51,973 (11.45)	24,345 (11.82)	27,628 (11.14)
2–3	133,057 (29.31)	60,410 (29.33)	72,647 (29.30)
4–5	116,158 (25.39)	52,020 (25.25)	64,138 (25.87)
≥ 6	86,359 (19.03)	38,757 (18.81)	47,602 (19.20)
Hot drink temperature, n (%)				1136.17	< 0.001
Very hot	77,034 (16.97)	38,886 (18.88)	38,148 (15.39)
Hot	301,347 (66.39)	134,819 (65.45)	166,528 (67.17)
Warm	70,581 (15.55)	29,823 (14.48)	40,758 (16.44)
Missing	4951 (1.09)	2463 (1.20)	2488 (1.00)
Caffeine intake, n (%)				133.322	< 0.001
Quintile1	93,929 (20.69)	44,123 (21.42)	49,806 (20.09)
Quintile2	95,511 (21.04)	42,706 (20.73)	52,805 (21.30)
Quintile3	89,546 (19.73)	40,792 (19.56)	49,254 (19.87)
Quintile4	87,779 (19.34)	39,275 (19.07)	48,504 (19.56)
Quintile5	87,148 (19.20)	39,595 (19.22)	47,553 (19.18)
Follow-up time, years, mean (P25, P75)	15.12 (14.45, 15.72)	15.11 (14.45, 15.71)	15.12 (14.45, 15.73)	12.066	< 0.001
All-cause dementia, n (CIR)	8690 (1.91)	2282 (1.11)	6408 (3.11)	1306.839	< 0.001
Alzheimer’s disease, n(CIR)	3855 (0.85)	1026 (0.50)	2829 (1.37)	552.422	< 0.001
Vascular dementia, n (CIR)	1931 (0.43)	374 (0.18)	1557 (0.76)	529.414	< 0.001
Survival time, M (P25, P75), year
All-cause dementia	14.50 (14.18, 15.64)	14.71 (14.32, 15.66)	14.32 (14.11, 15.62)	3196.166	< 0.001
Alzheimer’s disease	14.53 (14.20, 15.64)	14.73 (14.33, 15.66)	14.37 (14.13, 15.62)	2907.032	< 0.001
Vascular dementia	14.54 (14.20, 15.64)	14.74 (14.33, 15.66)	14.38 (14.14, 15.62)	2842.196	< 0.001

TDI Townsend deprivation index, BMI body mass index, APOE apolipoprotein E, CIR Crude incidence rate; M (P₂₅, P₇₅), median (Percentile₂₅, Percentile₇₅);

Table 2.

Risk of dementia in hypertensive patients compared to the general population.

	Adjusted HR (95%CI)
	All-cause dementia	Alzheimer disease	Vascular dementia
Model1	1.187 (1.116, 1.261)	1.175 (1.072, 1.288)	1.549 (1.339, 1.793)
P-value	< 0.001	0.001	< 0.001
Model2	1.118 (1.101, 1.266)	1.137 (1.023, 1.264)	1.565 (1.330, 1.841)
P-value	< 0.001	0.017	< 0.001
Model3	1.152 (1.070, 1.240)	1.124 (1.006, 1.256)	1.452 (1.225, 1.720)
P-value	< 0.001	0.039	< 0.001
Model4	1.149 (1.059,1.247)	1.137 (1.006,1.286)	1.281 (1.061,1.548)
P-value	0.001	0.040	0.010

Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI].

Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior].

Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease].

Model 4: adjusted for Model 3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Fig. 2.

A linear relationship between coffee consumption and incident dementia.

Fig. 3.

A non-linear relationship between tea and caffeine consumption and incident dementia.

Association between coffee consumption and dementia risk in the population with hypertension and without hypertension

After adjusting all covariates, a J-shaped relationship exists between coffee consumption and all-cause dementia risk in hypertensive patients (Fig. 4). Notably, hypertensive individuals consuming 0.5–1 cup daily had the lowest risk of all-cause dementia (HR = 0.733, 95% CI 0.613, 0.875), compared to those having 6 or more cups daily. No statistically significant connection was observed between coffee consumption and Alzheimer’s disease or vascular dementia risk in this group. Additionally, there was no correlation between coffee consumption and the risk of developing all-cause dementia, Alzheimer’s disease, and vascular dementia in the population without hypertension (Table 3).

Fig. 4.

A J-shaped relationship exists between coffee consumption and all-cause dementia risk in hypertensive patients. Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI]. Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior]. Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease]. Model 4: adjusted for Model 3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Table 3.

Association between coffee consumption and  risk of dementia in the participants.

			Coffee consumption, Cups/day					P-value
			None	0.5–1	2–3	4–5	≥ 6
			HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
All-cause dementia
Participants without hypertension		Model1	–	–	–	–	Ref.	0.304
		Model2	–	–	–	–	Ref.	0.650
		Model3	–	–	–	–	Ref.	0.608
		Model4	–	–	–	–	Ref.	0.945
Participants with hypertension		Model1	0.784 (0.680, 0.903)	0.688 (0.601, 0.788)	0.718 (0.628, 0.820)	0.753 (0.648, 0.874)	Ref.	< 0.001
		Model2	0.785 (0.665, 0.925)	0.729 (0.623, 0.854)	0.762 (0.653, 0.889)	0.788 (0.663, 0.937)	Ref.	0.003
		Model3	0.797 (0.670, 0.950)	0.749 (0.634, 0.885)	0.771 (0.655, 0.908)	0.782 (0.652, 0.939)	Ref.	0.017
		Model4	0.768 (0.637, 0.927)	0.733 (0.613, 0.875)	0.746 (0.626, 0.888)	0.763 (0.628, 0.929)	Ref.	0.014
Total population		Model1	0.817 (0.724, 0.923)	0.725 (0.646, 0.814)	0.762 (0.680, 0.853)	0.807 (0.711, 0.915)	Ref.	< 0.001
		Model2	0.838 (0.727, 0.965)	0.777 (0.679, 0.889)	0.822 (0.720, 0.938)	0.858 (0.741, 0.993)	Ref.	0.006
		Model3	0.863 (0.743, 1.004)	0.801 (0.694, 0.925)	0.836 (0.726, 0.963)	0.866 (0.741, 1.012)	Ref.	0.040
		Model4	0.825 (0.702, 0.970)	0.784 (0.672, 0.915)	0.804 (0.690, 0.935)	0.829 (0.700, 0.980)	Ref.	0.041
Alzheimer disease
Participants without hypertension		Model1	–	–	–	–	Ref.	0.695
		Model2	–	–	–	–	Ref.	0.403
		Model3	–	–	–	–	Ref.	0.584
		Model4	–	–	–	–	Ref.	0.949
Participants with hypertension		Model1	0.794 (0.637, 0.988)	0.701 (0.568, 0.864)	0.745 (0.607, 0.916)	0.750 (0.595, 0.946)	Ref.	0.018
		Model2	–	–	–	–	Ref.	0.184
		Model3	–	–	–	–	Ref.	0.568
		Model4	–	–	–	–	Ref.	0.393
Total population		Model1	0.834 (0.691, 1.008)	0.754 (0.630, 0.902)	0.807 (0.676, 0.963)	0.830 (0.682, 1.010)	Ref.	0.035
		Model2	–	–	–	–	Ref.	0.174
		Model3	–	–	–	–	Ref.	0.709
		Model4	–	–	–	–	Ref.	0.565
Vascular dementia
Participants without hypertension		Model1	–	–	–	–	Ref.	0.106
		Model2	–	–	–	–	Ref.	0.403
		Model3	–	–	–	–	Ref.	0.220
		Model4	–	–	–	–	Ref.	0.488
Participants with hypertension		Model1	0.711 (0.539, 0.937)	0.645 (0.496, 0.839)	0.590 (0.455, 0.766)	0.611 (0.453, 0.823)	Ref.	0.002
		Model2	0.720 (0.526, 0.986)	0.695 (0.516, 0.937)	0.603 (0.447, 0.812)	0.656 (0.469, 0.919)	Ref.	0.020
		Model3	0.712 (0.510, 0.994)	0.673 (0.490, 0.923)	0.615 (0.449, 0.842)	0.641 (0.449, 0.917)	Ref.	0.046
		Model4	–	–	–	–	Ref.	0.131
Total population		Model1	0.737 (0.575, 0.944)	0.654 (0.517, 0.828)	0.609 (0.482, 0.769)	0.693 (0.533, 0.900)	Ref.	0.001
		Model2	0.753 (0.569, 0.997)	0.693 (0.530, 0.905)	0.630 (0.483, 0.821)	0.731 (0.545, 0.982)	Ref.	0.014
		Model3	0.736 (0.547, 0.990)	0.671 (0.506, 0.889)	0.641 (0.485, 0.847)	0.722 (0.529, 0.985)	Ref.	0.032
		Model4	–	–	–	–	Ref.	0.164

Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI].

Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior].

Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease].

Model 4: adjusted for Model3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Association between tea consumption and dementia risk in the population with hypertension and without hypertension

After adjusting all covariates, a U-shaped association between tea intake and the risk of all-cause dementia in the hypertensive population was shown in Fig. 5. Tea drinkers had a lower risk of all-cause dementia compared to non-tea drinkers. The risk was lowest for hypertensive patients who drank four to five cups of tea daily (HR = 0.744, 95% CI 0.675, 0.887). In contrast, there was no association between tea consumption and the risk of all-cause dementia, Alzheimer’s disease, and vascular dementia in the population without hypertension (Table 4).

Fig. 5.

A U-shaped relationship exists between tea consumption and all-cause dementia risk in hypertensive patients. Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI]. Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior]. Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease]. Model 4: adjusted for Model 3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Table 4.

Association between tea consumption and risk of dementia in the participants.

		Tea consumption, cups/day					P-value
		none	0.5–1	2–3	4–5	≥ 6
		HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
All-cause dementia
Participants without hypertension	Model1	Ref.	0.862 (0.713, 1.042)	0.778 (0.666, 0.908)	0.734 (0.624, 0.862)	0.844 (0.712, 0.999)	0.003
	Model2	Ref.	–	–	–	–	0.105
	Model3	Ref.	–	–	–	–	0.091
	Model4	Ref.	–	–	–	–	0.061
Participants with hypertension	Model1	Ref.	0.846 (0.748, 0.957)	0.784 (0.710, 0.867)	0.762 (0.687, 0.845)	0.843 (0.756, 0.941)	< 0.001
	Model2	Ref.	0.871 (0.758, 1.000)	0.768 (0.685, 0.862)	0.763 (0.678, 0.859)	0.772 (0.679, 0.877)	< 0.001
	Model3	Ref.	0.866 (0.747, 1.004)	0.793 (0.703, 0.895)	0.768 (0.677, 0.870)	0.798 (0.698, 0.914)	< 0.001
	Model4	Ref.	0.826 (0.701, 0.972)	0.788 (0.691, 0.900)	0.774 (0.675, 0.887)	0.820 (0.709, 0.949)	0.003
Total population	Model1	Ref.	0.852 (0.768, 0.944)	0.784 (0.721, 0.853)	0.755 (0.692,0.824)	0.845 (0.771, 0.927)	< 0.001
	Model2	Ref.	0.878 (0.781, 0.986)	0.787 (0.715, 0.867)	0.769 (0.696, 0.849)	0.790 (0.710, 0.880)	< 0.001
	Model3	Ref.	0.862 (0.762, 0.976)	0.804 (0.726, 0.889)	0.765 (0.689, 0.850)	0.809 (0.722, 0.905)	< 0.001
	Model4	Ref.	0.819 (0.714, 0.940)	0.809 (0.724, 0.905)	0.762 (0.679, 0.855)	0.832 (0.736, 0.940)	< 0.001
Alzheimer disease
Participants without hypertension	Model1	Ref.	–	–	–	–	0.110
	Model2	Ref.	–	–	–	–	0.221
	Model3	Ref.	–	–	–	–	0.386
	Model4	Ref.	–	–	–	–	0.526
Participants with hypertension	Model1	Ref.	0.797 (0.657, 0.966)	0.839 (0.721, 0.976)	0.755 (0.644, 0.886)	0.867 (0.733, 1.026)	0.012
	Model2	Ref.	0.793 (0.635, 0.990)	0.831 (0.696, 0.990)	0.737 (0.613, 0.887)	0.788 (0.646, 0.960)	0.026
	Model3	Ref.	–	–	–	–	0.144
	Model4	Ref.	–	–	–	–	0.080
Total population	Model1	Ref.	0.799 (0.681, 0.938)	0.824 (0.726, 0.936)	0.752 (0.659, 0.859)	0.822 (0.713, 0.947)	0.001
	Model2	Ref.	0.791 (0.658, 0.951)	0.827 (0.714, 0.957)	0.769 (0.660, 0.896)	0.762 (0.645, 0.900)	0.007
	Model3	Ref.	–	–	–	–	0.061
	Model4	Ref.	0.722 (0.578, 0.902)	0.870 (0.733, 1.031)	0.800 (0.670, 0.955)	0.831 (0.687, 1.004)	0.038
Vascular dementia
Participants without hypertension	Model1	Ref.	–	–	–	–	0.429
	Model2	Ref.	–	–	–	–	0.796
	Model3	Ref.	–	–	–	–	0.485
	Model4	Ref.	–	–	–	–	0.261
Participants with  hypertension	Model1	Ref.	0.796 (0.620, 1.023)	0.683 (0.556, 0.838)	0.734 (0.596, 0.904)	0.784 (0.627, 0.980)	0.007
	Model2	Ref.	0.795 (0.598, 1.057)	0.664 (0.525, 0.841)	0.755 (0.596, 0.957)	0.769 (0.597, 0.992)	0.019
	Model3	Ref.	–	–	–	–	0.056
	Model4	Ref.	–	–	–	–	0.121
Total population	Model1	Ref.	0.798 (0.638, 0.997)	0.695 (0.580, 0.834)	0.729 (0.605, 0.877)	0.804 (0.660, 0.979)	0.002
	Model2	Ref.	0.783 (0.606, 1.010)	0.699 (0.568, 0.860)	0.748 (0.606, 0.923)	0.782 (0.625, 0.978)	0.017
	Model3	Ref.	0.781 (0.596, 1.024)	0.735 (0.592, 0.914)	0.730 (0.584, 0.913)	0.820 (0.648, 1.038)	0.047
	Model4	Ref.	–	–	–	–	0.125

Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI].

Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior].

Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease].

Model 4: adjusted for Model3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Association between interaction of coffee and tea consumption and dementia risk in the population with hypertension and without hypertension

After adjusting all covariates, the statistically significant associations were found between the interaction of coffee and tea consumption and the risk of all-cause dementia and Alzheimer’s disease in the hypertensive population (Fig. 6).Consuming four to five cups of coffee and six or more cups of tea daily is statistically significantly associated with the lowest risk of developing all-cause (HR = 0.541, 95% CI 0.311–0.939) dementia and Alzheimer’s disease(HR = 0.294, 95% CI 0.094–0.921) than those who drink six or more cups of coffee and don’t drink tea in hypertensive patients. Among them, we observed that with the increase in tea consumption, the relationship between coffee consumption and the risk of all-cause dementia in hypertensive patients gradually changed from a U-shaped to an inverted U-shaped (Fig. 7). In contrast, there was no statistically significant association between the interaction of coffee and tea and the risk of developing all-cause dementia, Alzheimer's disease and vascular dementia in the population without hypertension.

Fig. 6.

Association between interaction of coffee and tea consumption and the risk of all-cause dementia and Alzheimer’s disease in individuals with hypertension. We adjusted for socio-demographic characteristics (age, sex, ethnicity, education level, occupation status, TDI, BMI), lifestyle factors (smoking status, dietary patterns, alcohol consumption, physical activity, social contact, sedentary behavior duration), health-related issues (family history of dementia, APOE genotype, depression, cancer history, cardiovascular disease history), and serum markers and drug use (aspirin use, lipid-lowering medication use, dyslipidemia, hypertriglyceridemia, serum 25(OH)D status).

Fig. 7.

Correlation between interaction of coffee and tea consumption and risk of all-cause dementia in hypertensive patients. We adjusted for socio-demographic characteristics (age, sex, ethnicity, education level, occupation status, TDI, BMI), lifestyle factors (smoking status, dietary patterns, alcohol consumption, physical activity, social contact, sedentary behavior duration), health-related issues (family history of dementia, APOE genotype, depression, cancer history, cardiovascular disease history), and serum markers and drug use (aspirin use, lipid-lowering medication use, dyslipidemia, hypertriglyceridemia, serum 25(OH)D status).

Association between coffee type or hot drink temperature and dementia risk in the population with hypertension and without hypertension

After adjusting for all covariates, the statistically significant associations were found between coffee type and the risk of all-cause and vascular dementia in the population with and without hypertension, but no association with Alzheimer's disease. Ground coffee consumption revealed the lowest risk of the onset of all-cause dementia [(HR = 0.762, 95% CI 0.618–0.940) for non-hypertensive population and (HR = 0.682, 95% CI 0.588–0.791) for hypertensive population] and vascular dementia [(HR = 0.421, 95% CI 0.246–0.721) for non-hypertensive population and (HR = 0.655, 95% CI 0.478–0.898) for hypertensive population] compared with decaffeinated coffee consumption (Table 5). In addition, after adjusting for all covariates, no statistically significant association was identified between the temperature at which a hot drink was consumed and the risk of developing dementia (Table 6).

Table 5.

Coffee type on the risk of dementia in a population.

		Coffee type				P-value
		Decaffeinated coffee	Instant coffee	Ground coffee	Other type of coffee
		HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
All-cause dementia
Participants without hypertension	Model1	Ref.	0.800 (0.700, 0.914)	0.707 (0.605, 0.827)	1.297 (0.882, 1.909)	< 0.001
	Model2	Ref.	0.786 (0.674, 0.916)	0.773 (0.646, 0.924)	1.176 (0.729, 1.898)	0.004
	Model3	Ref.	0.789 (0.671, 0.926)	0.778 (0.644, 0.939)	0.933 (0.533, 1.632)	0.019
	Model4	Ref.	0.792 (0.663, 0.946)	0.762 (0.618, 0.940)	1.021 (0.569, 1.831)	0.017
Participants with hypertension	Model1	Ref.	0.918 (0.840, 1.003)	0.700 (0.626, 0.784)	1.055 (0.797, 1.395)	< 0.001
	Model2	Ref.	0.893 (0.805, 0.989)	0.713 (0.627, 0.812)	0.893 (0.628, 1.271)	< 0.001
	Model3	Ref.	0.902 (0.810, 1.005)	0.691 (0.603, 0.792)	0.810 (0.549, 1.195)	< 0.001
	Model4	Ref.	0.897 (0.798, 1.009)	0.682 (0.588, 0.791)	0.836 (0.554, 1.264)	0.021
Total population	Model1	Ref.	0.880 (0.818, 0.947)	0.706 (0.644, 0.773)	1.123 (0.895, 1.409)	< 0.001
	Model2	Ref.	0.862 (0.791, 0.938)	0.716 (0.645, 0.795)	0.973. (0.733, 1.292)	< 0.001
	Model3	Ref.	0.868 (0.794, 0.949)	0.713 (0.639, 0.796)	0.846 (0.615, 1.163)	< 0.001
	Model4	Ref.	0.866 (0.786, 0.955)	0.706 (0.625, 0.796)	0.884 (0.631, 1.239)	< 0.001
Alzheimer disease
Participants without hypertension	Model1	Ref.	0.888 (0.727, 1.083)	0.708 (0.558, 0.899)	1.368 (0.759, 2.465)	0.014
	Model2	Ref.	–	–	–	0.051
	Model3	Ref.	–	–	–	0.230
	Model4	Ref.	–	–	–	0.651
Participants with hypertension	Model1	Ref.	0.921 (0.806, 1.053)	0.720 (0.608, 0.853)	0.928 (0.590, 1.460)	0.001
	Model2	Ref.	0.956 (0.817, 1.119)	0.760 (0.623, 0.926)	0.941 (0.548, 1.616)	0.031
	Model3	Ref.	0.971 (0.822, 1.147)	0.768 (0.624, 0.947)	0.915 (0.511, 1.640)	0.049
	Model4	Ref.	–	–	–	0.067
Total population	Model1	Ref.	0.908 (0.813, 1.015)	0.709 (0.616, 0.814)	1.043 (0.729, 1.494)	< 0.001
	Model2	Ref.	0.936 (0.822, 1.066)	0.734 (0.625, 0.863)	1.073 (0.703, 1.638)	0.001
	Model3	Ref.	0.976 (0.851, 1.120)	0.763 (0.644, 0.905)	0.962 (0.597, 1.549)	0.005
	Model4	Ref.	0.963 (0.828, 1.119)	0.766 (0.636, 0.923)	0.991 (0.597, 1.644)	0.019
Vascular dementia
Participants without hypertension	Model1	Ref.	0.574 (0.421, 0.782)	0.337 (0.219, 0.517)	0.901 (0.329, 2.471)	< 0.001
	Model2	Ref.	0.594 (0.418, 0.845)	0.364 (0.226, 0.588)	0.320 (0.044, 2.313)	< 0.001
	Model3	Ref.	0.667 (0.463, 0.961)	0.424 (0.259, 0.695)	0.000 (0.000, 1.774^43)	0.004
	Model4	Ref.	0.685 (0.462, 1.015)	0.421 (0.246, 0.721)	0.000 (0.000,3. 225^140)	0.017
Participants with hypertension	Model1	Ref.	0.939 (0.780, 1.131)	0.673 (0.528, 0.857)	1.371 (0.818, 2.296)	0.002
	Model2	Ref.	0.842 (0.682, 1.041)	0.660 (0.502, 0.866)	1.188 (0.640, 2.205)	0.017
	Model3	Ref.	0.844 (0.676, 1.053)	0.639 (0.480, 0.853)	1.095 (0.554, 2.163)	0.019
	Model4	Ref.	0.912 (0.715, 1.165)	0.655 (0.478, 0.898)	1.420 (0.715, 2.818)	0.021
Total population	Model1	Ref.	0.823 (0.703, 0.964)	0.570 (0.462, 0.703)	1.213 (0.767, 1.918)	< 0.001
	Model2	Ref.	0.768 (0.641, 0.920)	0.580 (0.457, 0.735)	0.957 (0.533, 1.720)	< 0.001
	Model3	Ref.	0.778 (0.644, 0.940)	0.579 (0.451, 0.743)	0.798 (0.407, 1.564)	< 0.001
	Model4	Ref.	0.827 (0.672, 1.017)	0.574 (0.437, 0.755)	0.961 (0.489, 1.889)	0.001

Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI].

Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior].

Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease].

Model 4: adjusted for Model3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Table 6.

Association btween hot drink temperature and risk of dementia in the participants.

		Hot drink temperature			P-value for trend
		Very hot	Hot	Warm
		HR (95% CI)	HR (95% CI)	HR (95% CI)
All-cause dementia
Participants without hypertension	Model1	Ref.	1.126 (0.979, 1.296)	1.268 (1.060, 1.516)	0.034
	Model2	Ref.	–	–	0.296
	Model3	Ref.	–	–	0.338
	Model4	Ref.	–	–	0.242
Participants with hypertension	Model1	Ref.	–	–	0.051
	Model2	Ref.	–	–	0.140
	Model3	Ref.	–	–	0.154
	Model4	Ref.	–	–	0.366
Total population	Model1	Ref.	1.105 (1.021, 1.197)	1.192 (1.080, 1.314)	0.002
	Model2	Ref.	1.103 (1.007, 1.209)	1.151 (1.028, 1.289)	0.042
	Model3	Ref.	–	–	0.053
	Model4	Ref.	–	–	0.111
Alzheimer disease
Participants without hypertension	Model1	Ref.	1.257 (1.011, 1.564)	1.446 (1.098, 1.903)	0.029
	Model2	Ref.	–	–	0.128
	Model3	Ref.	–	–	0.304
	Model4	Ref.	–	–	0.416
Participants with hypertension	Model1	Ref.	1.208 (1.038, 1.406)	1.268 (1.056, 1.523)	0.026
	Model2	Ref.	1.270 (1.061, 1.521)	1.230 (0.988, 1.531)	0.034
	Model3	Ref.	–	–	0.116
	Model4	Ref.	–	–	0.223
Total population	Model1	Ref.	1.218 (1.075, 1.379)	1.311 (1.126, 1.527)	0.001
	Model2	Ref.	1.240 (1.073, 1.433)	1.249 (1.044, 1.493)	0.012
	Model3	Ref.	–	–	0.051
	Model4	Ref.	–	–	0.094
Vascular dementia
Participants without hypertension	Model1	Ref.	–	–	0.228
	Model2	Ref.	–	–	0.437
	Model3	Ref.	–	–	0.401
	Model4	Ref.	–	–	0.591
Participants with hypertension	Model1	Ref.	–	–	0.640
	Model2	Ref.	–	–	0.640
	Model3	Ref.	–	–	0.682
	Model4	Ref.	–	–	0.694
Total population	Model1	Ref.	–	–	0.358
	Model2	Ref.	–	–	0.306
	Model3	Ref.	–	–	0.351
	Model4	Ref.	–	–	0.497

Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI].

Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior].

Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease].

Model 4: adjusted for Model3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Association between caffeine intake and dementia risk in the population with hypertension and without hypertension

After adjusting for all covariates, a statistically significant association was found between caffeine intake and the risk of all-cause and vascular dementia in the hypertensive population, but no association with Alzheimer's disease. There was a U-shaped association between daily caffeine intake and the risk of all-cause dementia (Fig. 8) and vascular dementia (Fig. 9) in the hypertensive population. Compared to those who consume Quintile 1 of daily caffeine, individuals with higher caffeine intake generally faced a lower risk of all-cause and vascular dementia. Notably, hypertensive patients consuming Quintile 2 of caffeine daily have the lowest risk of both types of dementia [(HR = 0.764, 95% CI 0.669–0.872) for all- cause dementia and (HR = 0.596, 95% CI 0.446–0.798) for vascular dementia]. In addition, there appears to be a W-shaped association between daily caffeine intake and the risk of all-cause dementia in the population without hypertension (Fig. 10). Interestingly, in the population without hypertension, no statistically significant association was observed between caffeine intake and the risk of Alzheimer's disease and vascular dementia (Table 7).

Fig. 8.

A U-shaped relationship exists between caffeine intake and all-cause dementia risk in hypertensive patients. Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI]. Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior]. Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease]. Model 4: adjusted for Model 3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Fig. 9.

A U-shaped relationship exists between caffeine intake and all-cause dementia risk in hypertensive patients. Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI]. Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior]. Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease]. Model 4: adjusted for Model 3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Fig. 10.

Caffeine intake and the risk of all-cause dementia in the population without hypertension is W-shaped. Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI]. Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior]. Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease]. Model 4: adjusted for Model 3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Table 7.

Associaton between caffeine intake and risk of dementia in the participants.

		Coffeine consumption, mg/day					P-value
		Quintile1	Quintile2	Quintile2	Quintile4	Quintile5
		HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
All-cause dementia
Participants without hypertension	Model1	Ref.	0.734 (0.629, 0.856)	0.821 (0.708, 0.953)	0.788 (0.677, 0.918)	0.855 (0.734, 0.995)	0.001
	Model2	Ref.	0.758 (0.636, 0.902)	0.826 (0.697, 0.979)	0.730 (0.612, 0.872)	0.825 (0.693, 0.983)	0.004
	Model3	Ref.	0.767 (0.639, 0.920)	0.797 (0.667, 0.953)	0.729 (0.606, 0.876)	0.807 (0.671, 0.971)	0.007
	Model4	Ref.	0.783 (0.640, 0.958)	0.834 (0.685, 1.014)	0.675 (0.547, 0.832)	0.842 (0.688, 1.032)	0.006
Participants with hypertension	Model1	Ref.	0.888 (0.807, 0.978)	0.830 (0.751, 0.916)	0.816 (0.738, 0.904)	0.905 (0.817, 1.002)	< 0.001
	Model2	Ref.	0.860 (0.770, 0.960)	0.763 (0.679, 0.856)	0.785 (0.699, 0.882)	0.845 (0.752, 0.951)	< 0.001
	Model3	Ref.	0.845 (0.753, 0.950)	0.740 (0.655, 0.835)	0.759 (0.672, 0.858)	0.851 (0.754, 0.962)	< 0.001
	Model4	Ref.	0.870 (0.766, 0.989)	0.764 (0.669, 0.872)	0.799 (0.700, 0.913)	0.880 (0.770, 1.005)	0.001
Total population	Model1	Ref.	0.841 (0.775, 0.912)	0.825 (0.760, 0.896)	0.806 (0.741, 0.877)	0.888 (0.815, 0.966)	< 0.001
	Model2	Ref.	0.828 (0.754, 0.908)	0.780 (0.709, 0.858)	0.767 (0.696, 0.845)	0.837 (0.760, 0.923)	< 0.001
	Model3	Ref.	0.821 (0.745, 0.905)	0.757 (0.685, 0.837)	0.751 (0.678, 0.832)	0.839 (0.757, 0.929)	< 0.001
	Model4	Ref.	0.844 (0.757, 0.939)	0.786 (0.705, 0.877)	0.763 (0.682, 0.854)	0.868 (0.777, 0.970)	< 0.001
Alzheimer disease
Participants without hypertension	Model1	Ref.	0.722 (0.574, 0.908)	0.786 (0.629, 0.983)	0.750 (0.596, 0.945)	0.862 (0.686, 1.082)	0.037
	Model2	Ref.	–	–	–	–	0.172
	Model3	Ref.	–	–	–	–	0.226
	Model4	Ref.	–	–	–	–	0.292
Participants with hypertension	Model1	Ref.	–	–	–	–	0.202
	Model2	Ref.	–	–	–	–	0.142
	Model3	Ref.	–	–	–	–	0.274
	Model4	Ref.	–	–	–	–	0.777
Total population	Model1	Ref.	0.845 (0.747, 0.956)	0.828 (0.731, 0.938)	0.818 (0.720, 0.929)	0.880 (0.773, 1.001)	0.010
	Model2	Ref.	0.839 (0.727, 0.967)	0.800 (0.692, 0.926)	0.821 (0.708, 0.951)	0.894 (0.771, 1.037)	0.020
	Model3	Ref.	0.866 (0.745, 1.007)	0.829 (0.711,0.966)	0.807 (0.690, 0.943)	0.919 (0.785, 1.076)	0.049
	Model4	Ref.	–	–	–	–	0.328
Vascular dementia
Participants without hypertension	Model1	Ref.	–	–	–	–	0.211
	Model2	Ref.	–	–	–	–	0.346
	Model3	Ref.	–	–	–	–	0.432
	Model4	Ref.	–	–	–	–	0.690
Participants with hypertension	Model1	Ref.	0.953 (0.782, 1.160)	0.718 (0.580, 0.888)	0.818 (0.664, 1.007)	0.900 (0.730, 1.109)	0.019
	Model2	Ref.	0.924 (0.739, 1.154)	0.638 (0.497, 0.818)	0.791 (0.624, 1.002)	0.825 (0.650, 1.048)	0.006
	Model3	Ref.	0.881 (0.697, 1.114)	0.589 (0.453, 0.767)	0.781 (0.610, 1.001)	0.827 (0.644, 1.061)	0.002
	Model4	Ref.	0.945 (0.732, 1.220)	0.596 (0.446, 0.798)	0.800 (0.610, 1.050)	0.823 (0.624, 1.083)	0.006
Total population	Model1	Ref.	0.906 (0.762,1.079)	0.714 (0.593, 0.861)	0.774 (0.644, 0.931)	0.875 (0.729, 1.051)	0.004
	Model2	Ref.	0.875 (0.718, 1.066)	0.664 (0.536, 0.822)	0.747 (0.606, 0.921)	0.812 (0.659, 1.000)	0.002
	Model3	Ref.	0.847 (0.688, 1.043)	0.625 (0.499, 0.784)	0.748 (0.601, 0.930)	0.825 (0.664, 1.026)	0.001
	Model4	Ref.	0.917 (0.730, 1.153)	0.651 (0.506, 0.836)	0.784 (0.616, 0.997)	0.865 (0.680, 1.100)	0.011

Model 1: adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI].

Model 2: adjusted for Model 1 and lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior].

Model 3: adjusted for Model 2 and health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease].

Model 4: adjusted for Model3 and serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

Subgroup analysis and sensitivity analysis

When determining the association between the risk of all-cause dementia and coffee and tea consumption, types of coffee, and caffeine intake in hypertensive individuals, a stratified analysis was performed considering age, sex, and socioeconomic status. The findings suggest a higher risk of all-cause dementia in hypertensive patients aged over 65, male, and with a higher socioeconomic status (Supplementary Figs. 1–2). Overall, the subgroup analyses were consistent with the main results. We conducted four sensitivity analyses, and all associations remained significant and in line with the overall findings, demonstrating the robustness of the results, as presented in Supplementary Tables S4–S7.

Discussion

In this study, individuals with hypertension had a higher likelihood of developing dementia compared to the population without hypertension. There was an association between the risk of dementia and coffee and tea consumption in the total population, and the relationship between the risk of all-cause dementia and the consumption of coffee and tea in hypertensive people showed J-shaped and U-shaped patterns respectively. People who drink 0.5–1 cup of coffee or 4–5 cups of tea per day have the lowest risk of dementia. The significant association between the amount of coffee and tea consumed and the risk of all-cause and vascular dementia were more likely to be found in the hypertensive population than in the non-hypertensive population. Consumers of ground coffee have the lowest risk of all-cause dementia and vascular dementia. There is no correlation between the temperature of hot beverages and the risk of dementia. Furthermore, a U-shaped relationship was observed between daily caffeine consumption and the risk of developing all-cause dementia and vascular dementia in individuals with hypertension.

This study found that there was an association between the risk of dementia and coffee and tea consumption in the total population. However, previous studies on the relationship between personal coffee or tea consumption and dementia risk have mostly been observational studies, with no experimental research evidence, so the research conclusions have been controversial. For example, a cohort study spanning 25 years did not find a significant link between midlife coffee consumption and the risk of cognitive decline and dementia²¹. Regular tea consumption does not demonstrate a significant correlation with the risk of Alzheimer’s disease²². Short-term associations between coffee intake and cognition were observed, but no long-term effects were noted^23,24. Fortunately, our findings have been supported by several previous studies. The previous study found that the risk of dementia and Alzheimer’s disease was reduced by 60% by drinking one to two cups of coffee a day²⁵. In addition, drinking one cup of tea a day reduced the risk of dementia or mild cognitive impairment by 6% and two cups by 11%²⁶. There are several possible reasons for the inconsistency of these studies. Firstly, most studies primarily relied on self-reported coffee and tea consumption, which is susceptible to information bias. Secondly, the diagnosis of dementia varied among studies, and some level of misclassification of dementia may have occurred in most studies. Thirdly, inconsistent sample size and follow-up time in each study, along with excessive category combinations. Fourthly, there may be variability in the daily consumption of coffee and tea, making the exact amount and strength of coffee and tea consumption difficult to measure through questionnaires. Fifthly, the association between coffee and the risk of dementia varies depending on the type of coffee²⁰. Most of the studies did not distinguish between types of coffee and tea, and did not look at the consumption of a particular type of coffee or tea in isolation to see if it was associated with the risk of developing dementia. In addition, this study found that the relationship between the risk of all-cause dementia and the consumption of coffee and tea in hypertensive people showed J-shaped and U-shaped patterns respectively, which is consistent with previous studies^25,27,28. Among them, the study found that people who drink 0.5–1 cup of coffee or 4–5 cups of tea per day have the lowest risk of dementia, which is consistent with previous studies⁶. One possible reason is that coffee and tea ingredients, such as caffeine, may reduce the risk of dementia by decreasing neuroinflammation or providing neuroprotective benefits^8–10. However, excessive drinking of coffee and tea will lead to a large amount of caffeine intake in the body, disturb sleep patterns, and diminish internal antioxidant effects in the body²⁹. Excessive caffeine consumption can selectively and competitively bind to adenosine receptors, stimulating catecholamine secretion from the adrenal glands. This can lead to morphological changes in the brain⁷, increasing the risk of dementia and decreasing cognitive function³⁰. In addition, taking more than 500 mg of high caffeine into the body makes it more difficult for the body to eliminate caffeine³⁰. This can cause caffeine intoxication, leading to symptoms like anxiety, insomnia, and psychiatric disorders³¹, all of which are risk factors for dementia^32–34. The study also found that the association between caffeine intake and risk of dementia was U-shaped, suggesting that moderate coffee and tea drinkers are more likely to find a significant association between coffee and tea consumption and risk of dementia.

In this study, the statistically significant association between coffee and tea consumption and the risk of dementia was more likely to found in people with hypertension than in people without hypertension. The possible reason for this is that the active compounds in coffee and tea not only reduce the risk of hypertension (an independent risk factor for dementia)^17,18, but also inhibit acetylcholinesterase (a key enzyme in the pathogenesis of Alzheimer's disease), reduce lipid peroxidation and prevent neuronal damage¹⁹. Therefore, patients with hypertension who drink these drinks may get greater health benefits in delaying the development of neurodegenerative diseases such as dementia. Moreover, in addition to the separate links between coffee and tea consumption and the risk of dementia in hypertensive individuals, our study also found that there was a significant statistical significance between the interaction of coffee and tea the lower risk of developing all-cause dementia and Alzheimer's disease, aligning with previous studies⁶. This could be attributed to several factors. Firstly, coffee and tea share similar bioactive compounds like caffeine and chlorogenic acid, which work together to provide anti-inflammatory¹¹, antioxidant³⁵, blood pressure-lowering^17,18, and reduction of cerebral lipid peroxidation¹⁹ effects, thus helping prevent the onset of oxidative neurodegenerative diseases in the brain. Secondly, certain polyphenols in coffee and tea play a joint protective role in the pathogenesis of dementia⁶. Thirdly, coffee and tea help reduce the risk of cardiovascular metabolic diseases like diabetes³⁶ and hypertension³⁷, thus lowering the chances of dementia³⁸.

Prior research has shown that the link between coffee consumption and dementia risk differs based on the type of coffee consumed²⁰, which aligns with our discovery that the type of coffee has a notable impact on the risk of developing all-cause dementia and vascular dementia. The risk of developing all-cause dementia and vascular dementia was lowest among those who consumption of ground coffee. The likely reason for this is that different types of coffee have different levels of caffeine, with ground coffee having the highest caffeine content, instant coffee the second highest, and decaffeinated coffee the lowest³⁹. Caffeine is currently the most widely used psychostimulant in the world and has neuroprotective properties in the fight against neurodegenerative diseases⁵. Long-term consumption of caffeinated water not only slows the progression of memory deficits in mice with Tau pathology of Alzheimer's disease⁴⁰, but also reverses their memory deficits that occur with age⁴¹. Secondly, the preference for coffee varies among individuals⁴², and there are variations in chemical composition between different types of coffee^43,44. For example, ground coffee has the highest chlorogenic acid content⁴³. In addition, the benefits of chlorogenic acid in instant coffee may be offset by the harmful effects of added sugar, creamer and other ingredients⁴⁵. Therefore, we should pick the most suitable coffee and tea to drink in our normal life.

In this study, the statistically significant association between coffee and tea consumption, coffee type, caffeine intake, and the risk of dementia was more likely to be found in hypertensive patients over 65 years old, males, and those of high socioeconomic status. This may be because the prevalence of dementia rises sharply with age³⁸. While coffee²⁵ and tea²⁶ consumption can lower the risk of dementia, these rich caffeinated versions cannot offset age-related cognitive decline^46,47. Men seem to derive greater benefits from consuming coffee and tea than women, in line with previous research^27,28. Gender variations in brain structure and function could affect how coffee and tea consumption impact their protective advantages⁴⁸. In addition, regular tea consumption helped reduce Tauopathy and notably lowered the risk of dementia in males, while it did not have the same effect on Tauopathy in females²⁰. Higher income groups are more likely to drink coffee and tea, which leads to a more significant reduction in their risk of developing dementia. Lower-income groups might have more cases of undiagnosed dementia⁴⁹, while higher-income groups are more likely to receive a dementia diagnosis⁵⁰.

To our knowledge, few studies have looked at both coffee and tea consumption and caffeine intake in relation to the risk of dementia in people with hypertension. In addition, the reliability and validity of the study are strengthened by the large sample size, long-term follow-up and rigorously defined variables. However, this study has some limitations. Firstly, self-report of coffee and tea consumption at baseline may be subject to information bias. In addition, these kinds of measurements at baseline may not accurately reflect the long-term consumption since these behaviors are very likely to change with time. Secondly, estimates from higher intakes involved fewer participants, potentially resulting in imprecision in the dementia observed at these levels of coffee and tea consumption. Thirdly, UK Biobank is affected by a ‘healthy volunteer’ selection bias, as participants tend to be more health-conscious, resulting in an underestimation of dementia incidence and prevalence⁵¹. Fourthly, since the majority of UK Biobank participants were white UK citizens, our results may only be applicable to similar demographic groups. Finally, although we adjusted for many potential confounders, residual confounding from unmeasured or unknown variables may still affect our analyses.

Conclusions

There was an association between the risk of dementia and coffee and tea consumption in the total population, and the relationship between the risk of all-cause dementia and the consumption of coffee and tea in hypertensive people showed J-shaped and U-shaped patterns respectively. The significant association between the amount of coffee and tea consumed and the risk of all-cause and vascular dementia were more likely to be found in the hypertensive population than in the non-hypertensive population. Furthermore, a U-shaped relationship was observed between daily caffeine consumption and the risk of developing all-cause dementia and vascular dementia in individuals with hypertension.

Methods

Study design and population

The UK Biobank is a vast population-based study that recruited more than 500,000 participants (aged 39–74) at 22 assessment centers in England, Wales, and Scotland from 2006 to 2010⁵². Participants underwent physical examinations conducted by trained staff and completed touchscreen questionnaires. The cohort ensures continuous tracking of health-related outcome data of participant by establishing links with electronic records from primary care, hospitalization, and death registries. It has gathered a broad array of genetic and health data to explore genetic and lifestyle influences on various common diseases in middle-aged and older adults⁵³. All participants signed the informed consent form. In addition, UK Biobank adheres to the ethical principles of the 1975 Helsinki Declaration and has been approved by the North West Multi-Centre Research Ethics Committee (MREC) as a Research Tissue Bank (RTB), enabling researchers to operate within the approved scope without requiring additional ethical clearance. Among 502,370 UK Biobank participants collected at baseline, those with any of the following conditions were also excluded: (1) 4 people were excluded from the initial analysis for missing baseline data. (2) 45,536 participants who lacked baseline blood pressure measurement information were excluded. (3) 277 participants with secondary hypertension were excluded. (4) 2,428 participants who lacked baseline coffee and tea consumption information were excluded. (5) 212 subjects with a baseline diagnosis of dementia were excluded. A final cohort of 453,913 subjects who were followed until April 1, 2024, with an average follow-up of 15.12 years were included in this study (Fig. 1).

Assessment of Hypertension

The baseline information for the normal and hypertensive population was determined using a combination of self-reported data from the UK Biobank, hospital registry data, and the collection of two blood pressure measurements using an Omron HEM 7015-T automated sphygmomanometer. In this study, the baseline hypertensive population was defined by meeting one of the following criteria: (1) self-reported doctor diagnosis of hypertension; (2) self-reported use of antihypertensive medication; (3) measured systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg. The third criterion cut-off aligns with global hypertension definition⁵⁴.

Exposure assessment

Data on coffee and tea consumption were gathered through the Food Frequency Questionnaire (FFQ) to evaluate the type and intake of coffee and tea in an initial survey of the UK Biobank database⁵⁵. Participants were asked how many cups of coffee and tea they drank per day and what type of coffee and tea they typically consumed. We then looked at the relationship between different types of coffee and tea and consumption and the risk of developing dementia in people with diabetes. Caffeine intake (mg/d) was determined using self-reported coffee and tea consumption in the survey, with an assumption of 75 mg of caffeine per cup of coffee and 40 mg of caffeine per cup of tea^56–58 (See Table S1).

Assessment of outcome

We diagnosed dementia by utilizing health-related outcomes defined by the UK Biobank preprocessing algorithm⁵⁹, based on primary care, hospitalization, and death registry data during follow-up. Health-related outcome data were published in 2018, and data were dynamically updated to the end of this study in April 2024. Subsequently, dementia was classified according to the International Classification of Diseases (ICD) 9th and 10th editions. The dependent variable was dementia, including all-cause dementia, Alzheimer's disease and vascular dementia. The accuracy of UK Biobank in utilizing routinely collected healthcare datasets to detect dementia events showed high positive predictive value, sensitivity, and specificity^60,61 (See Table S2).

Assessment of covariates

The baseline survey obtained potential confounders such as corresponding socio-demographic, lifestyle and health-related issues through questionnaires. Confounders included age, sex (male and female), ethnicity (White, Asian, Black, and Other), education level (college or university degree, A level/AS levels or equivalent, O level/GCSEs or equivalent, CSEs or equivalent, NVQ or HND or HNC or equivalent, and other professional qualifications), occupation (employed, unemployed, and retired), Townsend Poverty Index(TDI) (low, medium, and high), body mass index (BMI) (< 18.5, 18.5–24.9, 25–29.9, and ≥ 30, kg/m²), smoking status (never, former, and current), dietary patterns (healthy dietary pattern, intermediate dietary pattern, and poor dietary pattern), alcohol consumption (IQRs, and quartiles), physical activity ((low, moderate, and high), social contact (active, moderately, isolated), Duration of sedentary behavior (< 2, 2–3.9, 4–5.9, and ≥ 6, hours/day), family history of dementia, APOE genotype, depression, history of cancer, history of cardiovascular disease, aspirin use status, lipid-lowering medications use status, dyslipidemia, hypertriglyceridemia, and serum 25(OH)D status (severely deficient, moderately deficient, and insufficient and above).

Socioeconomic status is evaluated through the Townsend Deprivation Index (TDI), where a higher TDI suggests a greater degree of deprivation^62,63. The UK Biobank database uses the APOE single nucleotide polymorphisms rs7412 and rs429358 for genotyping to identify carriers of the APOE ε4 allele^64,65. We establish a healthy dietary pattern based on the recommended intake of a heart-healthy diet in the United States⁶⁶. Physical activity intensity was determined by analyzing information from the International Physical Activity Questionnaire²⁰. Using the social connection index, determine the quality of the social relationship⁶⁷. Evaluate the duration of sedentary behavior based on activities like driving, computer usage, and television watching⁶⁸. Participants' serum 25(OH)D status was evaluated using baseline serum 25(OH)D data⁶⁹. Information on cancer, depression, cardiovascular, and other related diseases was gathered from medical records, self-reports, and death records. Definitions of each component of covariates in this study are described in Additional file: Table S3.

Statistical analyses

In the descriptive analyses of the normal and hypertension populations, a one-way ANOVA determined the mean (standard deviation (SD)) of continuous variables among the two groups, while a Pearson ${\chi }^2$ test identified any statistical differences in the proportions of categorical variables. The person-years at risk for each participant were computed starting from the recruitment date until the occurrence of dementia report, date of decease, loss to follow-up, or 01 April 2024, whichever event transpired first. The proportional risk assumption was validated with Schoenfeld residuals, and Cox proportional risk regression models were utilized to compute estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for the link between coffee and tea consumption and dementia risk in individuals with hypertension. Non-linear relationships between coffee, tea, and caffeine intake and all-cause dementia were assessed using restricted cubic spline curves with three knots. In Model 1, we adjusted for basic socio-demographic characteristics [age; sex; ethnicity; education level; occupation status; TDI; BMI]. In Model 2, we further adjusted for lifestyle [smoking status; dietary patterns; alcohol consumption; physical activity; social contact; duration of sedentary behavior]. In Model 3, we further adjusted for health-related issues [family history of dementia; APOE genotype; depression; history of cancer; history of cardiovascular disease]. In Model 4, we additionally adjusted for serum markers and drug use [Aspirin use status; lipid-lowering medications use status; dyslipidemia; hypertriglyceridemia; serum 25(OH)D status].

This study investigated whether the risk of developing dementia could be reduced in hypertensive and non-hypertensive populations by improving the consumption of caffeine-containing beverages such as coffee and tea. First, to explore the link between coffee and tea consumption and the risk of dementia in hypertension patients, risk ratios and 95% confidence intervals were computed separately. The relationship between different types of coffee (such as decaffeinated coffee, instant coffee, ground coffee, and other types of coffee) and the temperature of hot drinks (including very hot, hot, and warm) in relation to the risk of developing dementia in patients with hypertension was also examined. In addition, we also calculated the intake level of caffeine according to the intake of coffee and tea to further analyze the influence of caffeine on the risk of dementia in hypertension patients.

To assess whether age, sex, and socioeconomic status affected the association between coffee and tea consumption and dementia, subgroup analyses were performed by age (≤ 65 years and > 65 years), sex, and socioeconomic status (low, medium, high). In addition, to assess the robustness of our findings, we conducted four sensitivity analyses. First, we excluded participants who developed dementia within 2 years to minimize the chance of observing reverse causality in the associations. Secondly, we excluded outliers in coffee and tea consumption, like consuming 13 or more hot beverages daily. Thirdly, participants who had experienced a stroke at baseline were excluded. Furthermore, we also accounted for weight reduction and grip strength, taking into consideration the potential impact of frailty on dementia.

All statistical analyses were performed using Stata software version 17 and R version 4.0.2, and statistical significance (two-sided) was defined as a P value < 0.055.

Author contributions

B.W: Conceptualization, Methodology, Software, Formal analysis, Writing-Original Draft, Data Curation, Writing-Review & Editing. T.M: Investigation, Data Curation. L.Y: Investigation, Visualization. S.H: Methodology, Supervision. J.L:Methodology, Supervision. X.S: Writing-Review & Editing, Supervision, Funding acquisition, Project administration. All authors read and approved the final manuscript.

Data availability

The data underlying this article are available in UK Biobank, at https://www.ukbiobank.ac.uk/. This research has been conducted using the UK Biobank Resource under Application Number 98124. The datasets analyzed during the current study available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-71426-y.