A decision tree model of hypertension among college students in Yunnan Province, China

Abstract

Purpose

Previous studies have indicated that the prevalence rate of hypertension in adolescents is high, but it has not received much attention and the influencing factors are unclear, especially in Yunnan Province, China.

Materials and methods

A cluster sampling method was used to investigate 4781 freshmen in a college in Kunming, Yunnan Province from November to December. Demographic and lifestyle data were collected using questionnaires, and height, weight and blood pressure were measured. Decision tree model of hypertension in college students was established by Chi-square automatic interactive detection method.

Results

Prevalence of prehypertension of systolic blood pressure (SBP) and diastolic blood pressure (DBP) were detected in 33.9% and 32.1%, respectively. Prevalence of hypertension of SBP and DBP was detected in 1.2% and 7.2%, respectively. The hypertension and prehypertension decision tree of SBP has gender (χ² = 728.64, p < .001) at the first level and body mass index (BMI) (boys: χ² = 55.98, p < .001; girls: χ² = 79.58, p < .001) at the second level. The hypertension and prehypertension decision tree of DBP has gender (χ² = 381.83, p < .001) at the first level, BMI (boys: χ² = 40.54, p < .001; girls: χ² = 48.79, p < .001) at the second level, only children (χ² = 6.43, p = .04) and red wine consumption (χ² = 8.17, p = .017) at the third level.

Conclusions

The present study suggests that gender, BMI, only children and red wine consumption were the main factors affecting hypertension in college students in southwest border areas of China.

Plain language summary

• Hypertension in Chinese adolescent is generally ignored. This study first reports the prevalence of hypertension in adolescents in Yunnan Province, China.

• Four thousand seven hundred and eighty-one freshmen were surveyed and height, weight and blood pressure were measured. A decision tree model was used to analyze the predictors of hypertension.

• The study demonstrated that gender, BMI, only children and red wine predict hypertension in adolescents.

Keywords:

• College students
• hypertension
• prehypertension
• influencing factors
• decision tree model

Introduction

Hypertension is one of the leading causes of death worldwide and a major risk factor for cardiovascular disease and related diseases [1]. Hypertension is a chronic disease with an increasing global prevalence. Although improvements in hypertension control have been reported in high-income countries over the past three decades, hypertension remains a major public health issue in low-income countries [2]. Several epidemiologic studies have shown that only a third of people in low - and middle-income countries understand their hypertension status, and only about 8% make it under control [3].

Hypertension is widely believed to be a more common condition among the elderly. The findings of some studies show that the incidence of hypertension has significantly increased in adolescents, while the common risk prediction models for hypertension have not been validated in adolescents [4]. U.S. National Health and Nutrition Examination Survey (NHANESN) reported that the prevalence rate of hypertension in children and adolescents aged 5–18 increased by 14.2% during 1999–2014 [5]. Hypertension in Chinese adolescents has received insufficient attention. Limited studies have shown that the blood pressure levels and prevalence rate of hypertension in Chinese children and adolescents are increasing. China Health and Nutrition Survey (CHNS) reported that systolic blood pressure (SBP) increased from 96.1 mmHg to 102.7 mmHg and diastolic blood pressure (DBP) increased from 62.6 mmHg to 67.4 mmHg in children and adolescents aged 7–17 during 1991 to 2015 [6]. A nationwide survey in China in 2014 showed that 6.4% of children aged 7–18 suffered from hypertension [7]. What’s more, adolescence shows higher blood pressure responsiveness [8]. Therefore, hypertension in adolescents should be regarded as a serious health problem and not ignored.

Essential hypertension is the most common cause of hypertension in adolescents [9]. Essential hypertension is increasingly recognized as a syndrome rather than a disease, which is affected by multiple factors [10–12]. A systematic review and meta-analysis showed that being aged 30–60, married, living in an urban area, high income, smoking, family history of hypertension, overweight, obesity, diabetes and emotional symptoms (anxiety, stress and anger) were positively associated with hypertension; education level, physical activity and low dietary salt are negatively associated with hypertension [13]. Reviewing the worldwide literature, lifestyle (smoking, drinking, dietary behavior, drug abuse, physical inactivity, mobile phone dependence, poor sleep quality, etc.), genetic factors (family history of hypertension, etc.), environmental factors (living environment, maternal growth hormone levels during pregnancy, etc.) and demographic variables (gender, age, etc.) and health status (obesity, basal metabolic rate, mental health status, etc.) are the main risk factors for hypertension [13–17].

Lifestyle is an influential factor in hypertension. First, it is widely recognized that smoking and drinking are risk factors for hypertension [3,17,18]. For instance, a previous study has found that smoking and drinking are factors influencing hypertension in Hungarian adolescents [19]. However, a large sample Mendelian randomized meta-analysis showed no association between smoking and changes in blood pressure [20]. Second, studies have shown that eating behavior greatly affects blood pressure levels. It is well known that high sodium, low potassium, and low fruit and vegetable intake affect levels of blood pressure [3,21]. Similarly, dietary behavior under the influence of globalization has been linked to hypertension. CHNS reported that fast food, sugar-sweetened beverages (SSBs) and salty snacks increased the risk of hypertension (HR = 1.23–1.42) [22]. A national study in Iran also found that the intake of junk food (take-away fast food, salty snacks, sweets and SSB) was significantly correlated with the blood pressure level of adolescents aged 6–18 [16]. Third, physical inactivity is considered as a risk factor for hypertension [3,17]. A cohort study covering nine provinces in China showed that martial arts, gymnastics, table tennis and other sports can reduce the risk of hypertension in people over 18 years old (HR = 0.79–0.88) [22]. Moreover, poor sleep quality (OR = 4.24) and smartphone addiction (OR = 2.21) were significantly and independently associated with adolescent hypertension [23].

Genetics is a crucial predictor of hypertension. There are hints from previous studies that positive genes are associated with nearly 50% of essential hypertension [10]. College students with two parents with hypertension have a significantly increasing risk of hypertension (OR = 4.5) [24]. A Hungarian study showed that history of hypertension in parents was an independent risk factor for hypertension in adolescents [19]. Chinese scholars believe that parental hypertension is an essential predictor of hypertension phenotype in normal weight adolescents [25]. Also, the etiology of hypertension involves the complex interplay of environmental, pathophysiological factors and genetic predisposition [26].

Obesity is associated with hypertension. First of all, obesity is considered an independent predictor of hypertension [3,19]. A study in Guangxi, China, found that the prevalence of hypertension in obese adolescents was significantly higher than that of those of normal weight [25]. An Italian study reported that the risk of hypertension in obese adolescents was four times higher than that in adolescents of normal weight [27]. Furthermore, the low metabolic rate (BMR) of Bangladeshi college students was positively correlated with SBP and DBP [28]. What is clear is that a low BMR is likely to cause obesity because a low BMR consumes extremely little energy, which can easily lead to fatty build-up and excess energy.

The environment, including the maternal environment during pregnancy and the living environment after birth, is associated with hypertension. It has been shown that maternal growth hormone levels during pregnancy are associated with the risk of prehypertension and hypertension in their offspring [29]. Additionally, post-birth living environmental factors such as environmental pollution (air, water, noise and artificial light sources), urbanization and low greening are also associated with hypertension [3]. Interestingly, a study found that the greening degree of living environment is associated with the shift of blood pressure, and the higher the greening degree, the lower the incidence of hypertension [30].

Gender and age are essential demographic variables in predicting hypertension. According to the epidemiological survey in Hungary, gender and age are the influencing factors of hypertension in adolescents [19]. Worldwide, it is widely believed that the prevalence of hypertension in boys is higher than that in girls [15]. A recent study found that age ≥24 years old is a risk factor for hypertension, and the incidence of hypertension increases with the increase of age [17].

While a considerable number of studies have explored the factors responsible for hypertension globally, data in adolescents are very limited. Only a few studies have discussed hypertension in Chinese adolescents [25,31]. College students have a higher prevalence of hypertension. For instance, the prevalence rates of hypertension in college students in the United States, Congo and Iran were 30.3% [24], 26.4% [17] and 6.9% [32], respectively. What’s more, college students have a low awareness of hypertension. Surveys in the United States showed that only 35% of college students understand about hypertension [33]. To date, there are few data available on hypertension among adolescents in Yunnan Province, China. Yunnan lies in the southwestern border of China, and evidence on the burden of hypertension and associated factors is lacking among college students. Consequently, this study investigated the prevalence and influencing factors of hypertension among college students in Yunnan Province. Most previous studies have used regression models to explore the factors that influence hypertension. Considering the interaction mechanism among the factors and the effectiveness of the decision tree model, we use the decision tree model to analyze the predictors of hypertension in this study. The results can provide data support for the prevention and improvement of hypertension in college students.

Materials and methods

Study participants

From November to December 2021, a questionnaire survey and a physical examination were conducted on freshmen at a college in Kunming, Yunnan Province, China. After excluding invalid questionnaire and incomplete physical examination data, 4781 university students were included in the analysis. The average age of college students was (19.7 ± 1.6) years. Among them, 1586 (33.2%) were boys and 3195 (66.8%) were girls. There were 3531 (73.9%) Han and 1250 (26.1%) ethnic minorities. There were 3787 (79.2%) from rural areas and 994 (20.8%) from urban areas. There were 997 (20.9%) underweight, 3065 (64.1%) normal, 519 (10.9%) overweight and 200 (4.2%) obesity. More demographic variables are listed in Table 1.

Table 1. Comparison of frequency of prehypertension and hypertension in college students with different demographic variables.

Variables	Categories	SBP					DBP
		Normal	Prehypertension	Hypertension	Z/χ^2	p	Normal	Prehypertension	Hypertension	Z/χ^2	p
Gender^a	Male	615	922	49	−26.91	.000	671	693	222	−19.27	.000
	Female	2488	697	10			2229	843	123
Age (years)^b	17–18	738	359	11	1.98	.372	695	335	78	2.33	.313
	19–20	1552	833	26			1445	802	164
	≥21	813	427	22			760	399	103
Major^b	Engineering	345	320	17	111.87	.000	334	267	81	68.54	.000
	Management	384	167	6			334	182	41
	Pedagogy	575	252	6			545	244	44
	Science	380	221	15			378	203	35
	Agronomy	264	185	6			253	166	36
	Literature	333	147	1			312	144	25
	Medicine	366	158	3			332	161	34
	Art	351	126	5			320	120	42
	Other	105	43	0			92	49	7
Nationality^a	Han	2309	1169	53	−0.95	.344	2163	1120	248	−1.48	.139
	Minority	794	450	6			737	416	97
Residence^a	Rural	2445	1297	45	−0.90	.368	2284	1232	271	−0.80	.421
	City	658	322	14			616	304	74
Only children^a	Yes	544	311	15	−1.71	.087	503	295	72	−2.00	.046
	No	2559	1308	44			2397	1241	273
Family type^b	Two-parent family	2659	1414	53	5.09	.165	2498	1322	306	1.93	.588
	One-parent family	250	110	4			224	114	26
	Combined family	155	68	2			141	77	7
	Other	39	27	0			37	23	6
Self-perceived socioeconomic status^b	Worse	260	179	7	14.47	.006	268	140	38	4.24	.375
	Poor	728	393	11			686	378	68
	Medium	1969	957	37			1806	938	219
	Good	120	68	4			118	58	16
	Better	26	22	0			22	22	4
Academic stress^b	Very heavy	164	102	6	7.25	.124	157	92	23	8.58	.073
	Heavy	1058	550	23			1016	503	112
	General	1830	931	29			1682	912	196
	Easy	40	23	1			32	21	11
	Very easy	11	13	0			13	8	3
The number of close friends^b	0	35	24	1	14.42	.002	36	21	3	3.76	.288
	1–2	460	207	10			420	214	43
	3–4	615	271	6			562	267	63
	>5	1993	1117	42			1882	1034	236
Father’s education level^b	Illiteracy	581	318	14	1.41	.842	547	297	69	2.06	.725
	Elementary school	717	377	10			666	356	82
	Secondary school	1123	571	24			1064	537	117
	High school	425	231	5			392	224	45
	University	257	122	6			231	122	32
Mother’s education level^b	Illiteracy	987	518	28	1.79	.775	919	501	113	1.82	.768
	Elementary school	740	371	10			698	349	74
	Secondary school	868	476	11			818	433	104
	High school	310	161	5			284	166	26
	University	198	93	5			181	87	28
Father’s occupation^b	Civil servant	194	112	8	4.06	.541	178	110	26	6.94	.225
	Worker	93	57	1			90	46	15
	Staff	170	88	1			163	79	17
	Merchant	1839	964	32			1708	920	207
	Farmer	410	213	10			380	212	41
	Other	397	185	7			381	169	39
Mother’s occupation^b	Civil servant	160	69	7	0.76	.980	147	70	19	2.64	.756
	Worker	98	51	0			87	50	12
	Staff	157	85	1			149	75	19
	Merchant	1947	1020	36			1803	978	222
	Farmer	257	138	5			241	139	20
	Other	484	256	10			473	224	53

^a Mann–Whitney’s test.

^b Kruskal–Wallis test.

Procedures

This study is part of the behavior and sub-health study in Yunnan adolescents (BSSYA). The research was designed to explore the effects of adolescent lifestyle and physical and mental sub-health through cross-sectional studies and cohort studies. The data included in this analysis are part of a cross-sectional survey. Participants were selected by cluster sampling, and all the sampled freshmen were included in the survey. Electronic questionnaire was produced using WenJuanXing (www.wjx.cn), which was verified by experts and tested by pre-investigation. College students were gathered in classrooms for the survey. After the investigator explained the purpose and precautions of the investigation, the college students scanned the QR code of the electronic questionnaire with their mobile phones and filled in the questionnaire. It takes about 15 min to complete the questionnaire. Demographic variables, lifestyle (smoking, drinking, eating, etc.) and family history of hypertension were investigated. College students were asked to fill out the questionnaire independently and without discussing it with their classmates. The survey was administered by trained interviewers, who were on hand to answer questions posed by participants. Participants who completed the questionnaire volunteered to have their height, weight and blood pressure measured. None of the participants underwent antihypertensive therapy prior to the blood pressure measurement. This was an anonymous survey in which participants provided informed consent prior to participation. Participants were also made aware that they could withdraw from the survey at any time. The study was approved by the Ethics Committee of School of Medicine, Kunming University (approval number: 2021001).

Measures

Sociodemographic variables

Demographic variables assessed in this study included gender, ethnicity, age, residence, only child, parental occupation, family type, parental educational background, the number of close friends, self-evaluation family economic conditions, academic pressure and family history of disease.

Height and weight measurement

The height and weight of participants were measured using an Ultrasonic Height and Weight Meter (Beryl, BYH01BT, Zhongshan, China). Participants took off their shoes and coats and stood upright with their heels together. The heel, sacrum and scapulae were in contact with the column in a three-point line standing position. The height measurement is accurate to 0.1 cm, and the weight measurement is accurate to 0.1 kg. The formula for calculating body mass index (BMI) is BMI = weight (kg)/height (m)². There are four categories of BMI [34]: low weight (BMI < 18.5 kg/m²), normal weight (18.5 kg/m² ≤ BMI < 24 kg/m²), overweight (24 kg/m² ≤ BMI < 28 kg/m²) and obesity (BMI ≥ 28 kg/m²).

Blood pressure measurement

Blood pressure was measured by an upper arm medical electronic sphygmomanometer (OMRON, J760, Kyoto, Japan). After 15 min of completing rest, the participants were seated and their blood pressure was measured using a sphygmomanometer at the same level as the heart and right arm cuff. The unit of measurement for blood pressure is mmHg. Participants diagnosed with hypertension had their blood pressure measured twice. Blood pressure includes SBP and DBP. Hypertension diagnostic criteria were recommended by the Chinese Guidelines for Hypertension Prevention and Treatment (2018 Revision) [35]: SBP < 120 mmHg and DBP < 80 mmHg was normal blood pressure, 120 mmHg ≤ SBP < 140 mmHg and/or 80 mmHg ≤ DBP < 90 mmHg was prehypertension, SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg was hypertension. Blood pressure was measured by a specialist nurse.

Statistical analysis

An Excel database was exported from the questionnaire system and was checked. Statistical analysis was performed using SPSS 23.0 (SPSS Inc., Chicago, IL) software. Descriptive statistics, Mann–Whitney test, Kruskal–Wallis test and decision tree model were performed in the SPSS software (SPSS Inc., Chicago, IL). Descriptive statistics were used to calculate the proportions of demographic variables and the prevalence of hypertension. The Mann–Whitney test and the Kruskal–Wallis test were used to compare the differences in the prevalence of prehypertension and hypertension in college students with different demographic characteristics. A Chi-square automatic interaction detection method was used to build decision trees for prehypertension and hypertension. In the decision tree model, the dependent variables were grade data (1 = normal blood pressure, 2 = prehypertension and 3 = hypertension). Independent variables in the model included demographic variables with statistical significance in univariate analysis and confounding variables that may affect hypertension in college students: gender (1 = boy, 2 = girl), only children (0 = no, 1 = yes), number of friends (1 = 0, 2 = 1–2, 3 = 3–4, 4 = >5), self-evaluation family economic conditions (1 = worse, 2 = poor, 3 = medium, 4 = good, 5 = better), major (1 = engineering, 2 = management, 3 = pedagogy, 4 = science, 5 = agronomy, 6 = literature, 7 = medicine, 8 = art, 9 = other), cigarette smoking (0 = no, 1 = yes), smoke electronic (0 = no, 1 = yes), exposure to passive smoking (0 = no, 1 = yes), drink white wine (0 = no, 1 = yes), drinking beer (0 = no, 1 = yes), drinking red wine (0 = no, 1 = yes), family history of hypertension (0 = no, 1 = yes), keep exercising (0 = no, 1 = yes) and BMI (1 = low weight, 2 = normal weight, 3 = overweight, 4 = obese). The maximum growth depth of the decision tree was 3, the minimum number of cases in the parent node was 100, and the minimum number of cases in the child node was 50. Using cross validation, the number of sample folds was 10. The importance values were adjusted by the Bonferroni method. Statistical significance was set at p < .05.

Results

Comparison of the prevalence of hypertension and prehypertension in college students

The prevalence of SBP prehypertension and hypertension was 33.9% and 1.2%, respectively, in college students. The prevalence rates of DBP prehypertension and hypertension were 32.1% and 7.2%, respectively. The total prevalence of hypertension in all participants, boys and girls was 8.5%, 5.7% and 2.8%, respectively. Table 1 shows the prevalence of hypertension and prehypertension in college students with different demographic variables. There were significant differences in the prevalence of SBP hypertension and prehypertension in college students with different gender (Z = −26.91, p < .001), major (χ² = 111.87, p < .001), self-evaluated family economic conditions (χ² = 14.47, p = .006) and number of friends (χ² = 14.42, p = .002). There were significant differences in the prevalence of DBP hypertension and prehypertension in college students with different gender (Z = −19.27, p < .001), major (χ² = 68.54, p < .001) and only children (χ² = −2.00, p = .004). However, there was no significant difference in the prevalence of hypertension and prehypertension among other demographic variables (p > .05).

Decision tree model

Figure 1 shows the SBP decision tree for predicting hypertension and prehypertension. The number of nodes in the SBP decision tree was 9, the number of terminal nodes was 6 and the depth was 2. The prediction accuracy of the model was 71.9%. In the first level, gender (χ² = 728.64, p < .001), the prevalence of prehypertension (58.1%) and hypertension (3.1%) in boys was higher than that in girls (21.8% and 0.3%, respectively). In the second level was BMI (boys: χ² = 55.98, p < .001; girls: χ² = 79.58, p < .001). The prevalence rates of prehypertension and hypertension in boys (64.5% and 10.9%, respectively) and girls (54.4% and 2.2%, respectively) with obesity were the highest.

Figure 1. SBP decision tree for Yunnan university students.

Figure 2 shows the DBP decision tree for predicting hypertension and prehypertension. The number of nodes in the SBP decision tree was 13, the number of terminal nodes was 8 and the depth was 3. The prediction accuracy of the model was 61.7%. In the first level, gender (χ²=381.83, p < .001), the prevalence of prehypertension (43.7%) and hypertension (14%) in boys was higher than that in girls (26.4% and 3.8%, respectively). In the second level was BMI (boys: χ² = 40.54, p < .001; girls: χ² = 48.79, p < .001). The prevalence rates of prehypertension and hypertension in boys (54.5% and 25.5%, respectively) and girls (35.6%% and 15.6%, respectively) with obesity were the highest. In the third level was only children (χ² = 6.43, p = .04) and red wine consumption (χ² = 8.17, p = .017). The prevalence of prehypertension (42.4%) and hypertension (23.5%) was highest in boys who are the only children with overweight BMI. Furthermore, the prevalence of prehypertension (43.8%) and hypertension (12.3%) was highest in boys who did not consume red wine with a normal weight and low weight.

Figure 2. DBP decision tree for Yunnan university students.

Discussion

Our findings indicate that the prevalence rates of SBP and DBP prehypertension were 33.9% and 32.1%, respectively, in Yunnan college students. Our data were similar to that of college students in Jilin Province, China (31.5%) [36]. Our results are compatible with previous findings indicating that the total prevalence of hypertension in college students was 8.5%, slightly higher than that of college students in Anhui, China (7.2%) [37], and greatly higher than that of college students in Wuhan, China (4.3%) [15]. Similarly, the prevalence of hypertension among college students in Yunnan is higher than that in Ethiopia (7.7%) [38], Iran (6.9%) [32] and Bangladesh (6.5%) [28]. Therefore, hypertension in Yunnan college students should be taken seriously.

We found that gender was at the top of the SBP and DBP decision trees for hypertension and prehypertension. The prevalence of SBP and DBP prehypertension and hypertension was substantially higher in boys than in girls. The total prevalence of hypertension in boys (5.7%) was twice as high as that in girls (2.8%). Our data were concordant with previous results: the prevalence rate of hypertension in boys (12.9%) was about four times that of girls (3.2%) in Anhui, China [37]. The prevalence rate of hypertension in boys (7.9%) was about five times that of girls (1.6%) in Wuhan, China [15]. In addition, studies of college students in Ghana [39], Thailand [40] and the United States [41] all showed that the prevalence of hypertension in boys was more than twice that of girls. The mechanisms that gender difference in the prevalence of hypertension in college students are not completely characterized, possible explanations are as follows: first of all, hypertension is higher in boys than in girls, probably because boys have more hypertension risk factors such as obesity, smoking and alcohol consumption [11]. Next, there are differences in the regulation of girls and boys essential arterial hypertension by sex chromosomes and sex hormones, including the occurrence and development of blood pressure in the course of life and the pathophysiological mechanism of blood pressure regulation [42]. Furthermore, it is related to the pattern of trajectories of blood pressure changes across the sexes. Gender differences in hypertension begin early in life and change throughout life [43]. Evidence suggests that girls SBP is lower than boys SBP from the age of 13; from the age of 16, DBP decreased in both sexes, but more rapidly in boys [44].

Our data suggest that BMI is at the second level of the SBP and DBP decision trees for hypertension and prehypertension. The highest prevalence of prehypertension and hypertension in SBP and DBP was observed in obese college students. Our results support previous findings. The World Health Organization refers to obesity as abnormal or excessive fat accumulation that presents a health risk [45]. Numerous population-based studies have documented obesity as a clear risk factor for hypertension [12,22]. Obese adolescents in Italy are four times more likely to develop hypertension than adolescents of normal weight [27]. Hypertension was associated with higher BMI among college students in Thailand and Bangladesh [28,40]. Obesity (OR = 4.03–10.47) was significantly and independently associated with hypertension in adolescents in China and Ethiopia [15,23,38]. It has been found that compositional disorders and metabolic syndrome constitute an intermediate phenotype of essential hypertension. Metabolic risk factors determined by body composition disorders and visceral obesity increase the risk of hypertension [46]. Moreover, there is several evidence supporting that the oxidative stress, genetics, epigenetics and mitochondrial factors play an important role in the relationship between obesity and hypertension [45]. In brief, our study supports the view that adolescents can prevent and improve hypertension through lifestyle changes [9]. Maintaining normal BMI through intake of low-sodium diet, increasing vegetarian diet, reducing alcohol use and sticking to physical activity can effectively prevent hypertension and achieve long-term health effects [17,47,48]. It is recommended that hypertension should be routinely screened in overweight and obese adolescents [25].

Interestingly, red wine consumption was at the third level of the DBP decision tree for hypertension and prehypertension. The prevalence of prehypertension and hypertension was highest in normal-weight and low-weight boys who did not consume alcohol. No effect of red wine consumption on DBP was observed in girls. It is widely recognized worldwide that alcohol use is a risk factor for hypertension and that abstaining from alcohol can help lower blood pressure [19,49,50]. In contrast, there is some evidence that red wine consumption has a protective effect against hypertension. Red wine polyphenols (RWPs), such as resveratrol, catechin, epicatechin, quercetin and anthocyanin, have a positive effect on hypertension, dyslipidemia and metabolic diseases [51]. Possible mechanisms for the protective effect of RWPs on hypertension may include the following. First, RWPs changes lipid profiles, decrease in oxidative stress, improve the functionality of β1-adrenoceptors and so on [52,53]. Second, RWPs decrease the expression of adhesion molecules and growth factors, involved in migration and proliferation of vascular smooth muscle cells, and inhibit platelet aggregation [54]. Third, RWPs enhance levels of high-density lipoprotein cholesterol, and inhibit lipoproteins oxidation, macrophage cholesterol accumulation and foam-cell formation [55]. Fourth, RWPs decrease arterial damage, decrease the activity of angiotensin-II, and increase nitric oxide [56]. Fifth, animal studies have shown that RWPs exert antihypertensive effects and improve endothelial function by reducing the plasma levels of ET-1 (endothelin-1) and the subsequent vascular production of O(2)(•^–) (superoxide anion) in rats [57]. RWPs can prevent Ang II-induced hypertension and endothelial dysfunction [58].

Unexpectedly, only children are also on the third level of DBP decision tree for hypertension and prehypertension. The prevalence of prehypertension and hypertension was highest among non-only boys who were overweight. No effect was observed on DBP of only child girls. This is the first study to our knowledge to find differences in hypertension in only children. It is widely believed that only children have further health problems. However, limited studies have shown that the only children is healthier than people think, both physically and mentally [59]. The CHNS showed an increase of standard deviation in age height Z score among only children compared with non-only children [60]. In particular, only children reported fewer symptoms of loneliness and depression [61,62]. Under the COVID-19 epidemic, non-only children aged 12–18 in China are more likely than only children to suffer from anxiety and depressive symptoms, and more likely to be affected by COVID-19 [63]. To be sure, only children receive more healthy investment from their parents, more love, attention, care and protection, and they have less financial and life stress. The lower prevalence of hypertension in only children may be related to the aforementioned advantages, which we will further focus on in the cohort study.

There are several strengths and limitations of our study. A main strength of our study is that this is the first study to investigate the prevalence of hypertension and its influencing factors in college students in Yunnan Province, China, filling the gap of relevant research data. In addition, a decision tree model was used to analyze the predictors of hypertension. The predictive ability of decision tree model was higher than that of logistic regression analysis model commonly used in influencing factors analysis [64]. The results provide effective data support for the prevention and control of hypertension in college students in Yunnan. There are a number of limitations to this study that should be noted. First, while height, weight and blood pressure were measured, behavioral data were collected through questionnaires, inevitably with some recall bias. Second, hypertension was measured only twice in college students with high blood pressure and only once in participants with normal blood pressure. The blood pressure data were meaningful when the participants were rested for 15 min before their blood pressure was measured. Third, some of the new findings, such as the lower rate of hypertension among only children, cannot be fully explained and will be further clarified in cohort studies.

Conclusions

In conclusion, our study found that the prevalence of hypertension in college students in Yunnan Province is higher than that reported in domestic and foreign literature. Gender, BMI, only children and red wine consumption were the main predictors of hypertension in college students. Yunnan college students should pay attention to the harm of hypertension and develop diet and physical activity plans to effectively prevent it. Colleges should pay attention to the burden of hypertension in college students. Health education programs on hypertension have been developed and implemented for different genders, BMI and non-only children. Additionally, hypertension should be screened regularly among obese college students. The results of this study have positive implications for developing strategies to address hypertension and promoting healthy lifestyles among adolescents.

Author contributions

Xu Honglv designed the study, took primary responsibility for writing the manuscript, managing the literature analysis, and undertook the statistical analysis. Wang Wei, Wang Min and Chen Junyu searched and read literature. Yu Chunjie and Li Min compiled the database. Liu Dehui, Wang Jiai and Jiang Yinghong collected the data. All authors reviewed the manuscript.

Ethical approval

The study was approved by the Ethics Committee of School of Medicine, Kunming University (approval number: 2021001).

Disclosure statement

The authors declare that they have no conflicts of interest.

Additional information

Funding

This work was supported by the Special Basic Cooperative Research Programs of Yunnan Provincial Undergraduate Universities’ Association (202101BA070001-117) and Introduced Talents Scientific Research Project of Kunming University (YJL2103).