<p><b>BACKGROUND</b>There are few longitudinal studies on the associations of obesity with hypertension in young adults.</p><p><b>OBJECTIVES</b>To analyze longitudinally to what extent weight gain associates with hypertension in young adults.</p><p><b>METHODS</b>The subjects of this study consisted of 6,178 university students (male 4,098; female 2,080). The associations of hypertension with body type change were longitudinally examined by using the records of health examinations while at university. The prevalence ratios (PRs) for hypertension in their senior year were calculated on groups that changed toward obese against those that changed toward underweight. The logistic regression analyses were used to estimate odds ratios (ORs) for hypertension in their senior year of each factor. The analyses were conducted on (i) all subjects, (ii) non-hypertensive subjects in their freshman year, and (iii) by schools, in order to take into account physical activities.</p><p><b>RESULTS</b>The PRs of hypertension in subjects changed toward obese from their freshmen to seniors against ones toward underweight were 1.47 (95% CI; 1.00-2.15) for males and 3.50 (0.93-13.22) for females. In analyses limited to non-hypertensive subjects in their freshman year, results were similar to those of all subjects. The analyses by school also showed similar results to those including all subjects. In logistic regression analyses, although the factor most strongly associated with hypertension was body type in their senior year, the body type in their freshman still showed significant association with hypertension after the adjustment of senior year body type and hypertension in freshman year. The ORs for hypertension in obese subjects to normal weight ones in their senior year were 9.13 (95% CI; 5.77-14.45) for males and 22.59 (5.69-89.67) for females after adjusted by hypertension in freshman, body type in freshman and school.</p><p><b>CONCLUSIONS</b>These data suggest that the increase of BMI is linked to hypertension in university students.</p>