Table 1.
Clinical characteristics of subjects according to four eating behavior groups.
| Eating behavior groups | Normal eating behavior | LNDE alone | Skipping breakfast alone | Skipping breakfast concomitant with LNDE |
|---|---|---|---|---|
| N (% of total) | 34,971 (57.5) | 11,761 (19.3) | 7,423 (12.2) | 6,645 (10.9) |
| Men, n (%) | 20,010 (57.2) | 7,934 (67.5) | 5,174 (69.7) | 5,005 (75.3) |
| Age (years) | 45.2 ± 12.9 | 42.9 ± 12.0 | 39.2 ± 12.0 | 38.5 ± 11.0a,b |
| BMI (kg/m2) | 22.8 ± 3.2 | 23.2 ± 3.4 | 23.0 ± 3.5 | 23.3 ± 3.7b |
| Waist circumference (cm) | 80.2 ± 9.0 | 81.3 ± 9.5 | 80.3 ± 9.6 | 81.4 ± 10.0b |
| Systolic blood pressure (mmHg) | 121 ± 16.4 | 122 ± 16.2 | 121 ± 15.8 | 122 ± 15.8 |
| Diastolic blood pressure (mmHg) | 74 ± 12 | 75 ± 13 | 74 ± 12 | 75 ± 13b |
| Aspartate aminotransferase (IU/L) | 22 (19–26) | 22 (19–27) | 21 (18–25) | 21 (18–26)a,b |
| Alanine aminotransferase (IU/L) | 18 (13–25) | 19 (14–28) | 18 (13–26) | 19 (14–28)b |
| γ-glutamyltransferase (IU/L) | 22 (16–36) | 25 (17–43) | 23 (17–37) | 25 (18–43)b |
| Triglycerides (mg/dL) | 90 (61–136) | 91 (61–141) | 90 (61–140) | 91 (60–142) |
| HDL-cholesterol (mg/dL) | 62.6 ± 15.3 | 62.3 ± 15.4 | 60.2 ± 15.1 | 60.7 ± 15.1a |
| HbA1c (%, NGSP) | 5.5 ± 0.6 | 5.5 ± 0.6 | 5.4 ± 0.6 | 5.4 ± 0.6a |
| Past history of CVD, n (%) | 450 (1.3) | 159 (1.4) | 85 (1.1) | 67 (1.0) |
| Proteinuria, n (%) | 2,009 (5.7) | 741 (6.3) | 623 (8.4) | 568 (8.5) |
| Daily alcohol consumption, n (%) | 6,887 (19.7) | 3,761 (32.0) | 1,491 (20.0) | 1,999 (30.1) |
| Current smoker, n (%) | 8,611 (24.6) | 3,776 (32.1) | 3,747 (50.5) | 3,574 (53.8) |
| Having regular exercise, n (%) | 9,028 (25.8) | 2,741 (23.3) | 1,495 (20.1) | 1,316 (19.8) |
| MetSc, n (%) | 4,595 (13.1) | 1,610 (13.7) | 900 (12.1) | 849 (12.8) |
*Bonferroni P < 0.0083, (aLND alone versus skipping breakfast concomitant with LNDE; bskipping breakfast alone versus skipping breakfast concomitant with LNDE).
cMetS was determined according to the modified Adult Treatment Panel III criteria because of the lack of fasting plasma glucose measurement.
Concentrations of hepatic enzymes and triglycerides are expressed as median (interquartile range). Linear trends in clinical parameters across the four eating habit groups were examined by Pearson's correlation coefficient. Hepatic enzymes and triglycerides were log transformed before analysis. All P for trend values in continuous variables were <0.0001. Differences in categorical variables between the four dietary groups were examined by χ 2 tests. All P values were <0.0001, except for past history of CVD (P = 0.16) and MetS (P = 0.02, Cramer V value, 0.01).
BMI: body mass index; CVD: cardiovascular disease (including stroke); HDL: high-density lipoprotein; LNDE: late-night-dinner eating; MetS: metabolic syndrome; NGSP: National Glycohemoglobin Standardization Program.