Perceptualand Motor Skills, 1998,86, 1061-1062. O Perceptual and Motor Skills 1998 SEASONAL VARIATIONS IN MELATONIN MAY MODULATE GLYCEMIC RESPONSE TO EXERCISE ' AND1 WEYDAHL ROBERT B. SOTHERN LENNART WETTERBERG Finnmark College, Nonuay University of Minnesota Karolinska Institrite Stockhohn, Sweden Sr(mmary.-After a 30-min. standardized bicycle exercise at two times of day in 5 mo. of the year, 62 subjects showed a yearly variation in their glycemic response that was about 5 months out of phase with the peak levels in the light sensitive hormone melatonin, suggotlng a reciprocal relationship between melatonin and glucose utiliza- tion. In studies done at 70' north, where the sun is below the horizon for several weeks during the winter and above the horizon for several weeks during summer, we found a rime-dependent response of (fingertip) blood glucose to exercise when 62 subjects (32 men, 30 women, ages 18 to 41 years) repeated a standardized 30-min. exercise at two times of the day (11:30 and 16:30 hr.) and in five months (Weydahl & Sothern, 1997). The glucose response during the 263 exer- cise spans was lowest for both sexes in December (winter) compared to other months, while the response was greatest for the women in September, and for the men in April and June. Ex- ercise beginning at 1130 hr. vs 16:30 hr. showed best recovery at all times of the year (p= .006) from analysis of variance, and women showed best recovery at each test time (p< .001). A circannual (yearly) rhyrhm was found both for baseline glucose (p=.02, acrophase [peak] = Feb. 21 [from least-squares fit of a 1 year cosine]), and for glucose responses during exercise [p < ,001, acrophase = Feb. 4 using original units (mmoL4); p= .05, acrophase = Dec. 19 using data as percent of starting value], and for glucose recovery if glucose values were nor corrected for the glucose starting value (p<.001, acrophase=Feb. 21 using mmoV1, p<.001, acrophase= Feb. 2 using data as % of the starting value). In spite of a significant seasonal time effect in re- covery, a correlation between amount of daylight and glucose recovery was not found. sign&- cant differences in recovery between months, times of day, and sexes were greatly influenced by glucose levels at the start of exercise, indicating that baseline glucose rhythmicity, both circa- dian and circannual, greatly influenced the response to exercise. Seasonality of melatonin has been implied (melatonin is important in Seasonal Affective Disorder, and its secretion is inhibited by light) and melatonin also d u e n c e s glucose utiliza- tion (Arendt, 1992; Ortega-Corona, Esparza-Avalos, Benitez-King, Prarz-Guitrado, & AntBn- Tay, 1991). Thus, we wanted to check on any relationship between melatonin rhychms and the glucose response to exercise. Saliva was collected up to six times from 19 women (185 samples) and 10 men (88 samples) near the days of exercise and measured for melatonin. We found a statistically significant circadian rhythm (p=.01, acrophase=06:08 hr.) with not much change in peak times between seasons; however, a circannual rhythm was found (p<.001, acrophase= Sept. 23), indicating significant seasonal changes in melatonin levels. The circannual rhythms for melaronin and glucose were about 5 mo. out of phase. close to the 6 mo. that would be expected were the glucose uchzation directly inhibited by melatonin. It is Wtely that, in addi- tion to other factors, melatonin interacts with and influences glucose urilization throughout the year. 'Address correspondence to Andi Weydahl, Division of Sports and Arctic Chronobiology, Finn- mark College, N-9500 Alta Norway or e-mail (andi@hifm.no).