TEMPORAL CHARACTERISTICS OF TRANSMISSION-LINE LOADINGS IN THE SWEDISHCHILDHOOD-CANCER STUDY

A recent study conducted in Sweden reported that 1) leukemia risk in c hildren who lived near 220 or 400 kV electric-power transmission lines was associated with calculated historical magnetic field levels; 2) c hildren living within a distance of 50 m of transmission Lines had an elevated risk of leukemia; and 3) there was no association between leu kemia and residential magnetic fields measured many years after diagno sis. Subsequently, these investigators found through logistic regressi on analysis that disease was more strongly associated with calculated historical fields than with distance. Since the calculated historical fields in that study depended predominantly on distance and transmissi on-line load current, the logistic regression results suggest that his torical load current plays an important role in the epidemiological re sults. Thus, we studied hourly 1974 load-current data for six transmis sion Lines, and we examined 1958-1985 annual load-current data for 112 transmission lines. Most Lines exhibited marked diurnal load-current rhythms during 1974, and all six showed systematic weekday-weekend dif ferences. During 1958-1985, average loadings of Swedish 220 and 400 kV lines increased by about 1.3% year. Predictive-value and kappa-statis tic analyses indicated that Swedish transmission-line load currents we re not stable over long periods, so that contemporaneous load current (or a contemporary magnetic field measurement) was not a good surrogat e for historical load current (or historical magnetic fields). The res ults provide a potential explanation of the failure of the Swedish Stu dy to find an association between leukemia and contemporaneous magneti c field levels measured many years after the etiologic period, and sug gest that the inclusion of load-current data could significantly impro ve the quality of historical field calculations. (C) 1998 Wiley-Liss, Inc.