Baseline characteristics of participants
The demographic characteristics of the 4427 participants included in this study in the wave 1 investigation are presented in Table 1. The total CH incidence rate was 654/4427 (14.77%), and CH incidence increased from Class 1 to Class 4. The average age of the participants was 62.28 ± 8.02 years, and 54.73% were male. There were differences in age, gender, and smoking history between the classes. It is also worth mentioning that there were differences in kidney disease status, stomach disease status and chronic lung disease status between the classes. The incidence of these diseases clearly increased with class, suggesting that the number of ACEs may be associated with these diseases.
Table1 Baseline characteristics of participants
Variables
|
Total (n =4427)
|
Class 1 (n = 1093)
|
Class 2 (n = 1405)
|
Class 3 (n = 1012)
|
Class 4 (n = 917)
|
Statistic
|
P
|
|
|
CH incidence
|
654 (14.77)
|
118 (10.80)
|
174 (12.38)
|
157 (15.51)
|
205 (22.36)
|
χ²=62.41
|
<.001
|
|
Age, years
|
62.28 ± 8.02
|
61.60 ± 7.66
|
62.08 ± 7.81
|
62.79 ± 8.15
|
62.86 ± 8.53
|
F=5.90
|
<.001
|
|
SBP, mmHg
|
127.71 ± 19.07
|
127.27 ± 18.12
|
128.29 ± 19.32
|
128.09 ± 19.47
|
126.92 ± 19.36
|
F=1.28
|
0.278
|
|
DBP, mmHg
|
75.94 ± 11.54
|
76.26 ± 10.96
|
76.20 ± 11.57
|
75.87 ± 11.79
|
75.27 ± 11.85
|
F=1.56
|
0.198
|
|
BMI, kg/m2
|
24.55 ± 14.41
|
24.56 ± 11.63
|
24.36 ± 6.65
|
24.53 ± 12.26
|
24.87 ± 24.65
|
F=0.23
|
0.878
|
|
Gender, n(%)
|
|
|
|
|
|
χ²=9.51
|
0.023
|
|
Male
|
2423 (54.73)
|
557 (50.96)
|
771 (54.88)
|
574 (56.72)
|
521 (56.82)
|
|
|
|
Female
|
2004 (45.27)
|
536 (49.04)
|
634 (45.12)
|
438 (43.28)
|
396 (43.18)
|
|
|
|
Hukou, n(%)
|
|
|
|
|
|
χ²=6.27
|
0.394
|
|
1 Agricultural
|
3330 (75.22)
|
818 (74.84)
|
1062 (75.59)
|
754 (74.51)
|
696 (75.90)
|
|
|
|
2 Non-Agricultural
|
655 (14.80)
|
178 (16.29)
|
203 (14.45)
|
154 (15.22)
|
120 (13.09)
|
|
|
|
3 Unified
|
442 (9.98)
|
97 (8.87)
|
140 (9.96)
|
104 (10.28)
|
101 (11.01)
|
|
|
|
Marriage status, n(%)
|
|
|
|
|
|
χ²=3.83
|
0.280
|
|
Out of marriage
|
3833 (86.58)
|
953 (87.19)
|
1223 (87.05)
|
881 (87.06)
|
776 (84.62)
|
|
|
|
In marriage
|
594 (13.42)
|
140 (12.81)
|
182 (12.95)
|
131 (12.94)
|
141 (15.38)
|
|
|
|
Drinking status, n(%)
|
|
|
|
|
|
χ²=7.93
|
0.048
|
|
No
|
1727 (39.01)
|
397 (36.32)
|
546 (38.86)
|
428 (42.29)
|
356 (38.82)
|
|
|
|
Yes
|
2700 (60.99)
|
696 (63.68)
|
859 (61.14)
|
584 (57.71)
|
561 (61.18)
|
|
|
|
Smoking status, n(%)
|
|
|
|
|
|
χ²=7.25
|
0.064
|
|
No
|
1887 (42.62)
|
430 (39.34)
|
601 (42.78)
|
449 (44.37)
|
407 (44.38)
|
|
|
|
Yes
|
2540 (57.38)
|
663 (60.66)
|
804 (57.22)
|
563 (55.63)
|
510 (55.62)
|
|
|
|
History of chronic diseases, n (%)
|
|
|
|
|
|
|
|
|
Hypertension
|
918 (20.74)
|
211 (19.30)
|
313 (22.28)
|
212 (20.95)
|
182 (19.85)
|
χ²=3.86
|
0.277
|
|
Dyslipidemia
|
411 (9.28)
|
103 (9.42)
|
129 (9.18)
|
96 (9.49)
|
83 (9.05)
|
χ²=0.15
|
0.985
|
|
Disabetes or HBS
|
233 (5.26)
|
65 (5.95)
|
69 (4.91)
|
52 (5.14)
|
47 (5.13)
|
χ²=1.44
|
0.696
|
|
Chronic Lung Diseases
|
347 (7.84)
|
57 (5.22)
|
89 (6.33)
|
104 (10.28)
|
97 (10.58)
|
χ²=32.67
|
<.001
|
|
Liver Disease
|
174 (3.93)
|
35 (3.20)
|
50 (3.56)
|
44 (4.35)
|
45 (4.91)
|
χ²=4.83
|
0.184
|
|
Stroke
|
68 (1.54)
|
10 (0.91)
|
20 (1.42)
|
22 (2.17)
|
16 (1.74)
|
χ²=5.89
|
0.117
|
|
Kidney Disease
|
287 (6.48)
|
49 (4.48)
|
81 (5.77)
|
72 (7.11)
|
85 (9.27)
|
χ²=20.81
|
<.001
|
|
Stomach Disease
|
968 (21.87)
|
182 (16.65)
|
316 (22.49)
|
218 (21.54)
|
252 (27.48)
|
χ²=34.70
|
<.001
|
|
Abbreviations: CH: chronic headache; SBP: systolic blood pressure; DBP: diastolic blood pressure; BMI: body mass index; BHS:High Blood Sugar
Notes: Continuous variables were expressed as mean±standard deviation (SD) in case of normal distribution and compared between two groups by ANOVA test. If the count variable had a theoretical number<10, Fisher’s exact probability test was used. Categorical variables are presented as counts (percentages) and compared by Chi-square test
In addition, we also provided a baseline description of the missing variables before interpolation, which is similar to the above results (Table S3).
Associations between ACEs and CH incidence in different classes
To investigate the association between ACEs and the risk of developing CH, we constructed three covariate models with different adjustments for logistic regression analyses, as shown in Table 2. According to univariate binary logistic regression (crude model), the ORs (95% CIs) for incident CH were 1.52 (1.17 ~ 1.96), with P = 0.001 for Class 3, with Class 1 as a reference, and the ORs (95% CIs) for incident CH were 2.38 (1.86 ~ 3.04), with P < 0.001 for Class 4, with Class 1 as a reference. After adjusting for age and gender in Model I or adjusting for all covariates in Model II, the results for Class 3 and Class 4, with Class 1 as a reference, were similar to those of the crude model. Both classes had significantly greater ORs (95% CIs) for incident CH. Model I: Class 3 1.63 (1.26 ~ 2.11) with P < .001, Class 4 2.60 (2.02 ~ 3.35) with P < .001/Model II: Class 3 11.68 (1.30 ~ 2.19) with P < .001, and Class 4 2.62 (2.04 ~ 3.38) with P < .001). These results indicated that a higher CH incidence was significantly associated with the number of ACEs exposed.
Table2 Odds ratios for incident chronic headache occurs in different logistic regression model
Cluster
|
Case
|
Crude
|
Model Ⅰ
|
Model Ⅱ
|
|
|
|
|
Total
|
654(14.77%)
|
|
|
|
|
Class1
|
118(10.80)
|
1.00 (Reference)
|
1.00 (Reference)
|
1.00 (Reference)
|
|
Class2
|
174(12.38%)
|
1.17 (0.91 ~ 1.50) 0.221
|
1.22 (0.95 ~ 1.57)0.121
|
1.23 (0.96 ~ 1.59)0.102
|
|
Class3
|
157(15.51%)
|
1.52 (1.17 ~ 1.96) 0.001
|
1.63 (1.26 ~ 2.11))<.001
|
1.68 (1.30 ~ 2.19)<.001
|
|
Class4
|
205(22.36%)
|
2.38 (1.86 ~ 3.04)<.001
|
2.60 (2.02 ~ 3.35)<.001
|
2.62 (2.04 ~ 3.38)<.001
|
|
Model I, adjusted for age, gender.
Model II, adjusted for age, gender, marital status, Hukou, smoking status, drinking status, systolic blood pressure, diastolic blood pressure, body mass index.
Associations between different types of ACEs and CH incidence
Additionally, we conducted multivariable logistic regression between each of the 12 different types of ACEs and CH incidence, as shown in Fig. 2. A higher CH incidence was significantly associated with 7 different types of ACEs, including physical abuse, household mental illness, domestic violence, bullying, an unsafe neighbourhood, parental death and parental disability. Emotional neglect, household substance abuse, parental separation or divorce, an incarcerated household member and sibling death were not linked to the risk of developing CH.
Subgroup Analyses
To further explore the relationship between ACEs and the risk of developing CH, we chose ACEs = 2 as the cut-off value for further subgroup analyses, as shown in Fig. 3. There were no interactions among the demographic characteristics. Compared to those with ACEs < 2, participants with ACEs ≥ 2 had ORs (95% CIs) more than 1-fold greater for incident CH. In conclusion, the findings of this study are widely applicable across population subgroups.