Results
Our interviewers approached 7,430 individuals and 3,114 of them accepted the survey (rejection rate of 58.1 %). After 352 individuals suspended the interview, 2,695 subjects completed the survey (cooperation rate, 36.2 %; Fig. 1). The distributions of age, gender, size of residential area, and educational level were not significantly different from those of the general population of Korea (Fig. 1 and Table 1).
Fig. 1 Flow chart depicting the participation of subjects in the Korean Headache-Sleep Study. RLS: restless legs syndrome
Table 1 Sociodemographic distribution of all survey participants, the total Korean population, and of cases identified as migraine and restless legs syndrome
Sample number Total population p-value Migraine Restless legs syndrome
N (%) N (%) N (% , 95 % CI) N (% , 95 % CI)
Gender
Men 1345 (49.3a) 16,357,919 (50.6) 0.42b 36 2.7, 1.8–3.5 56 4.1, 3.0–5.1
Women 1350 (50.7a) 15,998,828 (49.4) 107 7.9, 6.5–9.4 86 6.4, 5.1–7.7
Age
19–29 542 (20.5a) 7,179,648 (22.2) 0.76b 25 4.5, 2.8–6.2 12 2.2, 1.0–3.4
30–39 604 (21.9a) 7,767,141 (24.0) 42 7.0, 5.0–9.1 34 5.6, 3.7–7.4
40–49 611 (23.1a) 8,012,377 (24.8) 39 6.5, 4.5–8.4 20 3.3, 1.9–4.7
50–59 529 (18.9a) 5,737,344 (17.7) 22 4.1, 2.4–5.9 35 6.5, 4.4–8.7
60–69 409 (15.6a) 3,660,237 (11.3) 15 3.9, 2.0–5.7 41 10.2, 7.3–13.1
Size of residential area
Large city 1,248 (46.3a) 1,5606,652 (48.2) 0.92b 76 6.1, 4.8–7.5 66 5.3, 4.1–6.5
Medium-to-small city 1,186 (44.0a) 1,4106,687 (43.6) 48 4.0, 2.9–5.1 58 4.9, 3.7–6.1
Rural area 261 (9.7a) 264,307 (8.2) 19 7.4, 4.2–10.6 18 6.8, 3.7–9.8
Educational level
Middle school or less 393 (14.9a) 6,147,782 (19.0) 0.75b 22 5.5, 3.3–7.7 41 10.5, 7.5–13.5
High school 1,208 (44.5a) 14,172,255 (43.8) 60 5.0, 3.8–6.3 62 5.1, 3.8–6.3
College or more 1,068 (39.6a) 1,2036,710 (37.2) 60 5.6, 4.3–7.0 38 3.6, 2.5–4.7
Total 2,695 (100.0a) 32,356,747 (100.0) 143 5.3, 4.5–6.2 142 5.3, 4.4–6.1
CI confidence interval
aAdjusted after age, gender, size of residential area and educational level
bCompared gender, age group, size of residential area, and educational level distributions between the sample of the present study and total population of Korea

Prevalence of migraine and RLS
Of the 2,695 participants, 1,273 (47.2 %) subjects had had at least 1 attack of headache, and 143 subjects (5.3 %) met the ICHD-2 criteria for migraine. So, 1,130 subjects were classified as having non-migraine headaches and 1,422 subjects were classified as non-headache controls. One hundred forty two (5.3 %) subjects positively responded at all 4 questions from the IRLSSG criteria and were classified as having RLS (Table 1).

RLS in subjects with migraine
Of 143 migraine subjects, 13 (9.1 %) were classified as having RLS. Of 1,130 non-migraine headache subjects and 1,422 non-headache control subjects, 78 (6.9 %) and 51 (3.6 %) subjects were classified as having RLS, respectively. A significantly greater prevalence of RLS was observed in both subjects with migraine (OR = 2.6, 95 % CI = 1.4–5.0, p = 0.003) and non-migraine headaches (OR = 2.1, 95 % CI = 1.4–2.9, p < 0.001) relative to non-headache control subjects. This pattern was consistent even after adjusting for sociodemographic variables (age, gender, educational level, and size of residential area) and sleep quality (PSQI score >5) (Table 2). There was no significant difference (9.1 % vs. 6.9 %, p = 0.339) in the RLS prevalence between migraineurs and non-migraine headache subjects.
Table 2 Prevalence and odds ratios of restless legs syndrome according to headache diagnosis
Age group, No of participants/ No. of RLS Non-headache control, N = 1,422 Non-migraine headache, N = 1,130 Migraine, N = 143
No. of RLS (%) OR No. of RLS (%) OR (95 % CI) P-value No. of RLS (%) OR (95 % CI) P-value
Aged 19–29, 542/12 4 (1.4) Reference 5 (2.2) 1.3 (0.4–5.0) 0.701 3 (12.0) 6.6 (1.2–36.8) 0.029
Aged 30–39, 604/34 6 (2.0) Reference 25 (9.3) 4.6 (1.8–12.0) 0.002 3 (7.1) 2.9 (0.6–13.4) 0.176
Aged 40–49, 611/20 3 (1.0) Reference 13 (4.7) 3.7 (1.0–13.5) 0.046 4 (10.3) 6.7 (1.5–33.5) 0.016
Aged 50–59, 529/35 17 (5.6) Reference 16 (7.8) 1.3 (0.6–2.7) 0.553 2 (9.1) 1.1 (0.2–5.6) 0.902
Aged 60–69, 409/41 21 (8.6) Reference 19 (12.8) 1.3 (0.6–2.5) 0.506 1 (6.7) 0.4 (0.1–4.0) 0.459
Total, 2695/142 51 (3.6) Reference 78 (6.9) 1.8 (1.3–2.7)a 0.002a 13 (9.1) 2.2 (1.1–4.3)a 0.018a
CI confidence interval, OR odds ratio, RLS restless legs syndrome
aAdjusted after sociodemographic variables and sleep quality

RLS prevalence according to age
We conducted logistic regression analysis for RLS prevalence according to headache diagnosis in 5 age groups. Subjects with migraine showed increased ORs for RLS in the 19–29 age group (OR = 6.6 95 % CI = 1.2–36.8, p = 0.029) and the 40–49 age group (OR = 6.7, 95 % CI = 1.5–33.5, p = 0.016) compared to non-headache controls. For subjects with non-migraine headaches, the 30–39 age group (OR = 4.6, 95 % CI = 1.8–12.0, p = 0.002) and the 40–49 age group (OR = 3.7, 95 % CI = 1.0–13.5, p = 0.046) showed increased ORs for RLS compared to non-headache controls. Migraineurs and non-migraine headache subjects failed to show increased ORs for RLS compared to non-headache controls in the 50–59 and 60–69 age groups (Table 2). We further divided our participants into 2 age groups (aged <50 years and aged ≥50 years) and compared the association between migraine and RLS. Migraineurs showed an increased OR for RLS compared to non-headache controls in aged <50 years (OR = 6.9, 95 % CI = 2.9–16.2, p < 0.001) but failed to show in aged ≥50 years (OR = 1.1, 95 % CI = 0.3–3.9, p = 0.883). For non-migraine headache subjects, an increased OR for RLS compared to non-headache controls was observed in aged <50 years (OR = 3.9, 95 % CI = 2.1–7.4, p < 0.001) but not in aged ≥50 years (OR = 1.5, 95 % CI = 0.9–2.4, p = 0.099).

RLS prevalence among migraineurs according to headache frequency
A significant increase in the prevalence of RLS was observed in both migraineurs with 1–10 attacks per month (12.5 %, OR = 2.0, 95 % CI = 1.3–3.2, p = 0.003) and migraineurs with >10 attacks per month (12.5 %, OR = 2.5, 95 % CI = 1.0–5.6, p = 0.038) relative to migraineurs with <1 attack per month (2.1 %) (Table 3).
Table 3 Prevalence and odds ratios of restless legs syndrome among migraineurs according to headache frequency
RLS N (%) OR (95 % CI)  P-value
Migraine <1 attack per month, N = 47 1 (2.1) Reference Reference
1–10 attacks per month, N = 80 10 (12.5) 2.0 (1.3–3.2) 0.003
>10 attacks per month, N = 16 2 (12.5) 2.5 (1.0–5.6) 0.038
CI confidence interval, OR odds ratio, RLS restless legs syndrome

Clinical characteristics of subjects having migraine with and without RLS
We investigated demographics, headache characteristics, associated symptoms, headache frequency per month, the visual analogue scale (VAS) score for pain intensity, and the Headache Impact Test (HIT-6) score of migraineurs according to the presence of RLS (Table 4). The only variable that showed a significant association with RLS was the HIT-6 score. The HIT-6 scores of migraineurs with RLS were significantly higher relative to migraineurs without RLS (59.1 ± 9.4 vs. 53.7 ± 9.2, p = 0.047).
Table 4 Clinical characteristics of migraine according to the presence of restless legs syndrome
Migraine with RLS, N =13 Migraine without RLS, N =130 P-value
Women, N (%) 10 (76.9) 97 (74.8) 0.867
Age, mean ± SD 39.8 ± 12.6 41.5 ± 12.5 0.640
Frequency per month, mean ± SD 5.8 ± 8.6 3.7 ± 6.0 0.251
Headache characteristics Unilateral, N (%) 9 (69.2) 72 (55.0) 0.323
Pulsating, N (%) 11 (84.6) 97 (74.8) 0.432
Moderate-to-severe intensity, N (%) 12 (92.3) 103 (78.6) 0.241
Aggravation by movement, N (%) 8 (61.5) 92 (70.2) 0.516
Associated symptoms Nausea, N (%) 11 (84.6) 114 (87.8) 0.742
Vomiting, N (%) 6 (46.2) 49 (38.2) 0.573
Photophobia, N (%) 8 (61.5) 76 (58.0) 0.806
Phonophobia, N (%) 10 (76.9) 91 (69.5) 0.575
Osmophobia, N (%) 7 (53.8) 61 (46.6) 0.616
Headache Impact test-6, mean ± SD 59.1 ± 9.4 53.7 ± 9.2 0.047
Visual analogue scale for pain intensity, mean ± SD 6.8 ± 2.2 6.1 ± 1.9 0.212
RLS restless legs syndrome, SD standard deviation