36 boundper mole TBG,the expected ratio is 777000 #g/57000 mg = 13.6. The results of these analyses are summarized in Table 30 and Figure 40. There was no significant difference in TBG between the Rongelap and Utirik groups or between the exposed and unexposed Rongelap groups. Eight subjects had high TBG levels. These were all females, and five were known to be pregnant, a condition that can elevate the serum TBG. The results of these analyses provide no explanation for the difference in serum Ts concentrations between the larger groups of Rongelap and Utirik subjects; however, these results are based on highly selected sampling and therefore are not representative of the island groups as a whole. Examination of individual Utirik and Rongelap plasmas suggested that low T4 concentrations were associated with relatively low TBG values. To study the interrelationship of T4 binding and TBG levels in greater detail, the dialyzable frac- tions of T, (DFT4) were determined in a number of plasma samples with low, normal, or elevated TBG-binding capacities. The results are presented in Table 31. In subjects with low TBG, the mean DFT, was greater than in subjects with normal or elevated TBG. The absolute free T4 was not dif- ferent in the three groups, which indicated that the abnormal T, levels among these subjects reflect alterations in plasma hormonebinding rather than in thyroxine production rates. Since virtually all the subjects with low plasma T, concentrations in the larger Utirik and Rongelap groups had normal plasma TSH, we would expect a moresystematic study of TBG levels by RIA to show a higher frequency of low TBG levels in the Utirik group. Sucha studyis currently in progress. F. IODOPROTEIN STUDIES Analyses for protein-bound iodine in sera of Marshallese people revealed a level significantly higher than that seen in North America.!8.29 Subsequentstudies showedthat this elevation could be attributed to an increase in the iodoprotein fraction and indicated that this was responsible for incorrect evaluation of thyroid function during the early years of the surveys. This unusualfinding promptedusto determine serum iodoprotein levels in several other Pacific Islands groups (see Table 32). Except for a group of Americans living on Kwajalein Atoll, these groups also exhibited high serum levels of iodoprotein. The Maui group was comprised ofpersons of almost pure Hawaiian ancestry, some of whom ate seaweed in large quantty, but the iodoprotein level did not correlate with this dietary intake. In the Rongelap population, it 1s of interest that the iodoprotein levelre- mained high in two athyreotic boys; after subtotal thyroidectomy; and during thyroid suppression by thyroxine administration.?° This strongly implies an extrathyroidal (endogenous) or a dietary (exogenous) source for the iodoprotein. The cause of the elevation, however, has not been ascertained. One possibility is that it is produced in polymorphonuclear leukocytes. These cells are known to organify iodine during phagocytosis.9! Although the Marshallese do not have unusually high leu- kocyte counts, they do have a mild eosinophilia Table 32 Serum Iodoprotein Levels (4g %) in Pacific Island Populations and in Americans Group Rongelap exposed No thyroid nodules Thyroid atrophy Thyroid nodules, pre-op After partial thyroidectomy After total thyroidectomy Ailingnae Utirik Rongelap unexposed No. in group 10 Total I 7.22.4 Tal Iodoprotein 4021.0 2541.1 2 7 3 2.80.4 8.33.6 S.I+1.1 1.2405 3.61.1 44-18 1.300 4.239 2.62.0 4 5 24 7.81.2 16.1+89 8.742.8 3.7+0.8 3.8408 4.01.0 3.341.7 3.3417 4.63.5 4121.1 2.2408 3 6041.1 2.6+0.1 Kapingamarangi Pingalap 23 27 11.543.0 8.42.0 6.2%2.6 3.6--0.6 Americans: Kwajalein 12 6641.3 3.91.1 Hana(Maui) US. 13 9 6.71.3 6.44%1.3 4.1205 2.9=1.2 3.00.6 3.3419 1.905 1.11.0