Effect of 131I on the anemia of hyperthyroidism

0021-9681~83~050405-08~3.0~1.0 Copyright 0 1983 Pergamon Press Ltd

J Chron Dis Vol. 36.No. 5.pp.405-412.1983 Printed ~11Great Britain

All rights reserved

EFFECT

JEFFREY

F-DA/Bureau

OF 1311 ON THE ANEMIA HYPERTHYROIDISM A.

PERLMAN

and

of Radiological Health, HFX-150,

PHYLLIS

M.

OF

STERNTHAL

5600 Fishers Lane. Rockville. MD 20857. U.S.A

(Received 1 September

1982)

Abstract-Data from the National Thyrotoxicosis Therapy Follow-Up Study (NTTFS) are presented here to document the existence of anemia in hyperthyroidism. a mild and reversible anemia that is simultaneously ameliorated with reversal of the hyperthyroid state. Among 20,600 women entered into the NTTF study with no previous history of hematological disorders, the prevalence of anemia was found to range from l&15?
INTRODUCTION HEMATOLOGICAL findings of hyperthyroidism were first suggested by Kocker in 1908 [l]. Leukopenia, relative lymphocytosis, and neutropenia were noted as the major hematological abnormalities associated with the hyperthyroid state. To date, derangements of the red blood cells have not been similarly emphasized and are not included in the textbooks as sequellae of thyrotoxicosis. Since anemia can mimic or worsen the cardiovascular hyperdynamic state associated with hyperthyroidism, this paper focuses on the important concept that anemia co-exists with hyperthyroidism. The extent to which this anemia of hyperthyroidism is ameliorated with standard medicinal and surgical treatments is examined. The majority of cases of thyrotoxicosis in developed countries are due to Grave’s disease [a]. Autoimmune mechanisms underlying Grave’s disease are thought to be responsible for an increased incidence of pernicious anemia in women diagnosed with thyrotoxicosis. But pernicious anemia is typically reported to occur in less than 24; of Grave’s disease cases. There are now several reports that indicate a far greater prevalence of anemia in hyperthyroidism and thyrotoxicosis. Most recently, Hambsch [-iI has shown anemia to occur in 20% of hyperthyroid females. Perillie [4] has shown a similar 18”/0prevalence of anemia in hyperthyroid women and 20% in men. These figures suggest that some other mechanism, other than pernicious anemia, accounts for the high prevalence of anemia prior to treatment for hyperthyroidism. Hambsch has suggested a direct

THE

Address all correspondence to MS Sternthal. Dr Perlman is currently with NIH/NICHD. tlon Research. 7910 Woodmont Avenue, Bethesda, MD 20205, U.S.A. 405

Center

for Popula-

406

JEFFREYA. PERLMAN and PHYLLIS M. S~ERNTHAL

suppressive effect of thyroxine on the bone marrow, based on a small series of 100 women in whom anemia was more severe with increasing degrees of hyperthyroidism. The purpose of this paper is to document the extent and severity of anemia in hyperthyroid women and to test the consistency of the large National Thyrotoxicosis Therapy Follow-Up Study (NTTFS) data with the thyrotoxic suppressive mechanism suggested by Hambsch. In addition, the reversibility of this anemia is characterized. Whether women treated with radioactive iodine or surgery have historically exhibited a greater response, at least so far as hematological status is concerned, is explored. The nature of bone marrow exposure to radiation is characterized for recipients of 1311 therapy, providing a framework for understanding why patients receiving the greatest cumulative dosage, no matter how great, do not display “radiation effects”. MATERIALS

AND

METHODS

Data were derived from the National Thyrotoxicosis Therapy Follow-Up Study, a retrospective-prospective medical record review and follow-up study of over 35,000 persons diagnosed and treated for hyperthyroidism in 26 clinical sites in the United States during the years 1946-1968. The methodology of the NTTFS is documented by Saenger [24,25]. Persons entering the study were followed an average of 6.2 yr, prospectively and/or through the records, between 1961 and 1968. Medical records were abstracted for retrospective information, pertaining to all clinical information from 1946 to 1961. Physical examination forms, surgical reports, laboratory reports, treatment records, and personal and physician questionnaires were concurrently gathered to document all medical system interactions after 1961. These data were abstracted from the primary inpatient and outpatient records to yield a composite, uniform study record. Study records containing data on admissions to hospital, physical findings, hematology and serum biochemistry, medicines taken, radiation treatments received, surgery performed, therapeutic efficacy, complications, and death, were pre-coded for data processing. Analysis of these composite study records serve as the basis of the analyses for this paper. The master data tape contains 35,630 records. For the purposes of this project, the group of white females, the largest age-sex group in the National Thyrotoxicosis Study, served as the focus of our analysis. All white females were included, unless there was a previously documented hematological disorder in their medical history. White females without any clinical blood counts on record also were not included in the study. Patients were classified into two treatment groups: those treated with radioactive iodine at any time and those treated by surgery. Because the usage of drugs is similar among the 1311 and the surgery group, it is not considered to be a confounding factor. However, the small percentage of white females treated by drug therapy alone is not discussed in this paper. The surgery group included no one with a history of therapeutic radiation exposure. Thus, the surgery group is a pure comparison group. The radioactive iodine group does include some women who were surgically treated at some point in time. However, these women were placed in the radioactive iodine group, having been exposed to radiation at TABLE 1. CHARACTERISTICSOF STUDY PARTKIPANTS IN CROSS-SETTIONAL AND PROSPECTIVE ANALYSES

ACCORDING

TO ‘I-RtATMbNT

Treatment

Characteristics

GROUP

Radiation

group Surgery

12.477 50.6 14.8”,,

8116 38.9 10. l”;,

3353 49.0 I5.4”,, 9.3”,,

1111 37.6 8.9”,, 5.0”,,

Women

used in cross sectional analysis Size of group Mean age before therapy Prevalence of anemia before therapy

Women used in prospective analysis Size of subgroup Mean age before therapy Prevalence of anemia before therapy Prevalence of anemia post-therapy

‘l’I Effect on the Anemia

407

of Hyperthyroidism

least once. Table 1 shows two important characteristics of these groups. Note that all women were used in at least one of the longitudinal analyses, described below. However, only 3353 of the radiation group and 1111 of the surgery group had more than 1 battery of blood tests, i.e. both before and after treatment. The hematological record of only these women could be followed over time for improvements in CBC values. As the Table indicates, the prospectively followed surgery and radiation subgroups are respectively representative of the entire surgery and radiation groups. according to initial age and prevalence of anemia. Routine clinical blood counts (CBCs) contain data for the primary analysis. Protein bound iodine (PBI) assays are also analyzed to show time trends in hematologic status vs thyroid status. CBCs and PBIs were classified according to their relationship in time to the date of initial treatment; i.e. date of first radioactive iodine treatment or date of first surgery. CBCs are referred to as pre-treatment blood counts, first year post-treatment blood counts, or end of follow-up blood counts, usually taken more than 1 yr post treatment. For the rare patient with more than one blood test in any given time period. the last pre-treatment CBC, the last CBC in the first year post-treatment, and the last CBC on record were used in analysis. The following cross-sectional analyses were undertaken in the 12,477 women treated medically (by radiation) and the 8116 women surgically treated: (1) To establish

the extent and severity of anemia in hyperthyroid women prior to treatment. profiles of mean hemoglobin values by age and treatment group were compared to similar national white female hemoglobin distributions, prepared by the National Center for Health Statistics. Along with the means, the 5, 10, 25, 50, 75, 90 and 95th percentiles were generated. Kolmogorov-Smirnov statistics were used to determine differences in the distributions. (2) To establish the relationship between severity of hyperthyroidism, and the degree of anemia. pre-treatment hemoglobin values were arrayed according to pre-treatment protein bound iodine (PBI) values among the sum-total 20,593 women. A one-way analysis of variance model and regression model were fit to these data to seek an endogenous thyroid hormone effect on hemoglobin. In addition, the following prospective anal!,ses were undertaken 11 I1 surgery and 3353 radiation treated women:

in the subgroups;

i.e.

(1) To compare the hematologic response of radioactive and surgically treated patients, two way matched tables were derived. showing the joint frequency distribution of anemics among the prospectively followed subgroups. One such table was prepared to show comparative change in occurrence of anemia during the first year post-treatment, radiation vs surgery group. A similar table was prepared to show change in occurrence of anemia over the long run. A categorical 12 g”J, [S-7] was used as the criterion for determining anemia. (2) To test the consistency of the hypothesis that thyroxine excess suppresses erythropoiesis (and that correction of hyperthyroidism removes suppression) a regression model was applied to change in hemoglobin vs change in PBI assay using the 4464 women in the combined subgroups. The strength of the association between falling PBI and increasing hemoglobin offers insight into the plausibility of a thyrotoxic suppression hypothesis. (3) To demonstrate whether exposure to radioactive iodine compromises the ultimate hematological status, a linear regression model was applied to last known hemoglobin value on cumulative mCi received by the 3353 women. RESULTS

In Table 1, there appears to be a greater prevalence of categorical anemia among those selected for radioactive iodine treatment compared to those selected for surgery. Both prevalences exceed the 5% prevalence of anemia expected in the U.S. white female population [S]. However, when the radiation group hemoglobin distribution is com-

408

JEFFREYA. PERLMAN and PHYLLIS M. STERNTHAL

TABLE 2. MEAN SERUM HEMOGLOBINLEVELSBY AGE GROUP FOR U.S. WHITE FEMALESAND THYROTOXICOSIS TREATMENT GROUPS Pre-treatment Age group 25-34 35-44 45-54 55-64 65-74

‘3’I-groUp

12.85 12.90 12.88 12.95 13.05

hemoglobin Surgery group

U.S. white female [S]

13.09 13.26 13.00 13.24 13.14

13.8 13.8 14.0 14.2 14.1

pared to that of the surgery group using Kolmogorov-Smirnov statistics, there is no statistical difference @ > 0.05) between the two treatment groups. Similar comparison with the U.S. reference population reveals that both treatment groups have hemoglobin values distributed over a lower range than that of the U.S. white females. Table 2 reveals that the mean hemoglobin value in this series of hyperthyroid women is 12.9, compared to the U.S. mean of 14.0 for U.S. white females of all ages 2574. To understand the relationship of this anemia to the physiological state of hyperthyroidism, the pre-treatment hemoglobin values were arrayed according to the pretreatment protein bound iodine (PBI). Table 3 presents the mean serum hemoglobin values found for each of several PBI ranges. Analysis of variance (p < 0.001) revealed a significant difference in mean hemoglobin concentrations for the various PBI levels. Further, a linear regression model fit to the same data yielded a significantly negative slope of b = -0.054, (p < 0.001) indicating that as PBI went up, serum hemoglobin went down. However, the weakness of the association indicates that the association is indirect at best. Data from the prospective analym of 3353 radiation treated women and 1111 surgically treated women are presented in Tables 4 and 5. Each of these Tables contain frequencies of anemics among the subgroups at two points in time, yielding the prospective view that is most desirable in ascertaining time trends. In Table 4 are the results of participants with CBCs taken before treatment and at the end of their follow-up period. From this table, it can be derived that 15.4% of the radiation subgroup is anemic prior to treatment while only 9.3% of this subgroup is classified as anemic at the end of follow-up. Similar figures for the prospectively followed surgery subgroup are 8.9% prior to treatment and 5.0% after treatment. The Table further indicates that of those hyperthyroid patients known to be anemic before treatment, nearly 80% of the 13iI and nearly 90% of the surgery group were no longer anemic at the end of the follow-up period. Conversely, only 7.1% of those 13iI patients known to be non-anemic before treatment ended the follow-up period classified as anemics. A similar figure for the surgery group is 4.2%. Table 5 presents the same type of data for the more short term comparison. Not all of the prospectively followed women had laboratory values during the first year following their commencement of treatment. This accounts for the smaller marginal totals than in TABLE 3. MEAN SERUM HEMOGLOBIN VALUES BY PBI RESULTS AMONG FEMALE THYROTOXICOSIS PATIENTS WITH PRETREATMENTLAB TESTS

PBI Level

Mean serum hemoglobin

N

4.5-8.0 8X-8.9 9.0-9.9 lO.t&lO.9 11.0-I 1.9 12.G 12.9 13.0+ Overall

13.24 13.09 13.06 12.57 13.16 12.98 12.60 12.92

238 166 211 215 168 149 514 1661

13’1 Effect on the Anemia

TABLE 4. DISTRIBUTION OF ANEMICS AND

of Hyperthyroidism

NON-ANEMICS

BEFORE TREATMENT

409

AND

AT THE END OF

FOLLOW-UP

Treatment

subgroup Surgery

Anemics Non-anemics (at end of study)

Subgroup totals

Anemics (pretreatment)

109 (21.0”“)

409 (79.00:,)

518 (IOOU,)

Non-anemics (pretreatment)

204 (7.1”,,)

2631 (92.9”“)

2835 (loo”,,)

Non-anemics Anemics (at end of study) 12 (12.1”“) 43 (4.2”“)

Subgroup totals

(875”)

99 (1OO”J

969 (95.7”,,)

1012 (10011,,,

3353

1111

Table 4. Recovery rates in the first year post-treatment are similar to the long term recovery rates. On the other hand, it appears that 5.5% of the non-anemic radiation treated females, compared to 3.1% of the non-anemic surgically treated females, developed anemia during the first year. These figures are not statistically different, however. The linear regression analysis of change in PBI le~l vs change i~zhemoglobin (pretreatment vs last post-treatment laboratory tests) reveals a statistically significant negative association (p < 0.02) for the combined subgroups. This is compatible with the notion that improving thyroid function is related to improvement in hematologic status. However. the slope of the regression line, b = -0.0027, is not so negative as to suggest that there is a major stepwise improvement in hemoglobin with quantitatively greater reductions in serum thyroid hormones. Finally, using data from the 3353 radiation-treated women who had post-treatment laboratory values, a linear regression model of last hemoglobin value on cumulative mCi over the course of the study shows a significant (p < 0.04) association between increasing cumulative dose and final hemoglobin level. This in itself suggests the lack of any adverse effect of cumulative 1311 exposure on the bone marrow, corroborated by the erythroproliferative response noted in Tables 4 and 5. DISCUSSION

The erythropoietic effect of thyroid hormone in physiologic and subphysiologic serum concentrations has been well documented in the medical literature. Thyroid hormone enhances red blood cell formation through the initiation and completion of hemoglobin alpha chains [S]. Physiologic circulating levels of thyroid hormone are required for the production of an adequate red blood cell mass. Thus it is that anemia is a well known side effect of the hypothyroid state [9]. Much less is known about the erythropoietic effect of thyroid hormone in supraphysiologic serum concentrations; yet, the joint occurrence of these two conditions may have clinical ramifications [3]. The hyperkinetic circulation of hyperthyroidism may be mimicked or worsened by severe anemia, but it has not been verified to this point in time how often these two conditions are associated with each other [is, 161. TABLE 5. DISTRIBUTIONOF

ANEMICS AND NON-ANEMICSBEFORE

Treatment

AND ONE YEAR POST-TREATMENT

subgroup

Lj’I Anemics Non-anemics (during first year) Anemics (pretreatment) Non-anemics (pretreatment)

36

(22X?“)

Surgery Subgroup totals

122 (77.2”/,)

158 (100”‘)0

574 (94.596)

( loo”g

607

765

Non-anemics Anemics (during first year) 4 (12.9“‘)” 9

(3.1”,,1

21

Subgroup totals 31

(87.1”:,)

(loos,,,

284 (96.9’?;,)

(lOO”/,,)

293

324

410

JEFFREYA. PERLMAN and PHYLLIS M. STERNTHAL

Since 1980, a few articles have suggested the possibility of an anemia that represents a “suppressive effect” of the thyroid hormones in supraphysiological serum concentration [2,3, 13, 143. Earlier papers suggested the possibility of shortened red blood cell survival time, impaired utilization of iron by RBC progenitor cells, and low metabolic turnover of vitamin B-12 resulting in reduced proliferative activity of bone marrow erythroblasts [l&12]. Data from the National Thyrotoxicosis Therapy Follow-Up Study have been presented here to substantiate that an anemia of hyperthyroidism is prevalent in lO_lSO/;, of patients diagnosed as hyperthyroid. This is slightly lower than similar results recently published [2,3]. However, it is far higher than the anemia of Grave’s disease, estimated to be only 3% [15]. In this study, the slightly higher prevalence of anemia among women treated medically rather than surgically most likely reflects the selection of less chronically ill patients for the surgical treatment. The overall mean 12.9 go/r,hemoglobin in all 20,593 hyperthyroid women pretreatment is 1 goAbelow the expected mean hemoglobin for all U.S. white females. In our sample, 14.2% of these females are classified as anemic. Evidence has been presented in this paper showing that over 3/4 of anemics before treatment are no longer so classified during the first year following the commencement of therapy. This reflects the proximity of pretreatment hemoglobin concentrations to normal levels as well as the rapidity of the response to treatment. In this regard, the degree of improvement appears to be similar in both the surgically and medically treated hyperthyroid patients. Tables have been presented that indicate the permanence of anemia reversal. Patients treated by surgery and ’ 311 show the same magnitude of response to treatment; i.e. the percentage of anemics who are not longer so classified ranges from 79 to 88’4 after an average 6.2 yr following the initiation of treatment. In general, the anemia of hyperthyroidism can be characterized as a mild, reversible anemia responsive to the amelioration of the underlying pathophysiology by either surgery or 13’I. Our data do not indicate strong statistical associations between PBI levels and amount of hemoglobin in the blood prior to treatment. In addition, our data do not indicate a strong statistical association between improvements in PBI levels and improvements in hemoglobin levels after treatment. It thus appears that the hypothesis of Hambsch can not be substantiated by these data and that some other hyperthyroid-related derangement of physiology, other than a direct throtoxic suppression effect, underlies the initial hematological compromise. The similarity of hematological response of medically and surgically treated groups should disarm any worries of further reduction of RBC mass by irradiation. Evidence of radiation-related bone marrow suppression strictly arises from studies of high dose exposure [17, 181. To put this in perspective, Labedzki [19] has shown that, in humans, there is a decrease in circulating red blood cells, approaching a 40% reduction at 46 weeks following whole body irradiation with 70-100rad. Hibbs [20] has further suggested that one of the first signs of B- or y-radiation exposure to the bone marrow at the 50 rad level is a decrease in bone marrow stem cell proliferation as well as a reduction in the number of circulating lymphocytes. On the other hand, similar findings are not reported in the literature for the low dose exposure situation. In ’ 311 treated hyperthyroid patients, fl- and y-radiation absorption by the thyroid gland is high enough to ablate thyroid cells. Approximately 1300 rad are delivered to the thyroid gland for each mCi of 13’Z delivered [Zl] in euthvroid individuals.

Greater dosage is delivered to hyperthyroid glands. Because of the affinity of iodine tagged drugs to the thyroid gland, radiation exposure would be relatively low dose to the remainder of the body, including the gonads and the bone marrow. Only 0.68 rad are delivered to the whole bone marrow organ [18] for each mCi, given a 25% thyroid uptake of iodine. This includes systemic radiation and thyroid emitted radiation to the marrow in the spine and bones of the upper thorax. Even assuming a 50% iodine uptake by the hyperthyroid thyroid gland and subsequently more regional bone marrow exposure, one cannot assume more than 1 rad of exposure to the bone marrow organ per mCi administered. Therefore, one would estimate the typical bone marrow absorbed

*j’I Effect on the Anemia

of Hyperthyroidism

411

dose to be in the vicinity of 1-5 rad during the typical single treatment episode recorded in the National Thyrotoxicosis Therapy Follow Up Study. 3-5 mCi of radioactive iodine were administered during the typical treatment episode during the 1960s. Today, 1.25 mCi may be used as a first treatment [14,22]. Among the patients registered in the National Thyrotoxicosis Therapy Follow-Up Study, cumulative dosages of up to 250 mCi were documented. However, such high dose individuals had received this dosage in a highly fractionated fashion, over a long period of time. Cumulative dosage in such persons was high because up to 20 treatments were required. Still. in these unusual cases, no more than 10 mCi were given at any one time. Therefore, there are several reasons why bone marrow suppressive effects would not be seen in patients receiving therapeutic doses of radioactive iodine. even at the higher doses used in the 1940s: (1) The absorbed bone marrow dose, estimated to be between 0.5 and 7 rad, is low enough to constitute a minimal exposure to the bone marrow organ at each treatment episode. (2) Even where local bone marrow exposure might be high enough to effect erythropoiesis, functional recovery may well be complete enough throughout the system to blur any additive effect of multiple administrations of the radiation. Affected regions of bone marrow may be compensated for by extension of the active organ or by compensatory amplification of mature cell production in existing non-affected areas [23]. Functional bone marrow recovery in affected areas, if any. would be more complete than formerly thought [23]. (3) ’ “1 ahlatiorl cf the thyroid gland itself; with correction of’ thr hyperth~wid sttrto, reduces any underlying physiologic inhibitory or exhaustion effect that may suppress bone marrow erythropoiesis [2]. Our data from the NTTFS support these impressions. As expected, there is no evidence of hematopoietic system compromise as a result of exposure to ’ 311. The overwhelming positive response of anemics to treatment has been noted. In addition. in the short term prospective follow-up of women exposed to radiation (Table 5), only 5.5”,, of women showed a transition from a non-anemic to anemic, not statistically different than the 3.1”,, in the surgery group. Similarly, there is no long term cumulative effect evidenced from similar figures in Table 4. The fact that a statistically significant positive slope can be estimated from a regression model of last-known hemoglobin level on total mCi received indicates that cumulative exposure is related to degree of completeness of therapy, reversibility of hyperthyroidism, and amelioration of anemia-not to any long term damage to the bone marrow. REFERENCES I. 2. 3. 4. 5

6: 7.

x. 9.

IO. I I. 12.

13. 14.

Kockcr T: Blutuntcrsuchungen bei monbus basedowii mit beitragen zur fruhdugnose uni theoric der krankheit. Arch Klin Chir 97: No. 131. 1908 Reddy J. r’r al: The peripheral blood picture in thyrotoxicosis. NZ J Med. 93: 142-145. 1981 Hambsch K. Fischer H, Lanspeter D. et al: Hyper-thyroidism and anemia. Z Gesamte Inn Med lhre Grenzgeb 36: No. 6, 1981 Perillie PE. Fletcher J: Anemia in hyperthyroidism. Ann Int Med 70: 507-16. 1969 Wallach J : Interpretation of Diagnostic Tests. 3rd edn. Boston: Little. Brown and Company. I978 Ravel R: Clinical Laboratory Medicine. 3rd edn. Chlcago: Yearbook Medical Publishers, 197X National Center for Health Statistics: Hemoglobin and selected non-related findings of persons I 74 years of age. Vital and Health Statistics Series, Advance Data No. 46. DHEW Pub. No. (PHS) 79 1250. Public Health Service. Washington, DC January 1979 Fein HG. Rivlin R: Anemia in thyroid diseases. Med Clin North Am 59: 1975 Tudhope CR. Wilson GM: Anemia m hypothyroidism. Q J Med 29: 513-33, 1960 Grenfell RF, Sheeler LR: Association of Thyroid Dysfunction and Anemia. J Miss State Med Assoc 20: I979 Rivlin RS, Wagner HN: Anemia in hyperthyroidism. Ann Int Med 70: 504, 1969 Danowski T. Vester J. Sarah G: Hydrocortisone and/or dessicated thyroid in physiologic dosage. Metabolism 1393: 1393. 1964 Daae LN, c’t (11: The early follow-up of Iodine-131 treatment of thyrotoxicosis. Eur J Nucl Med 5: 1980 Blair AL. Lowe DC, et al: Long term follow-up of patients treated for hyperthyroidism with low dose radioactive iodine. Ulster Med J 49: 1980

412 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

JEFFREY A. PERLMAN and PHYLLIS M. STERNTHAL Wilkins RW, Lavinsky NG: Medicine: Essentials of Clinical Practice, 2nd edn. Boston: Little, Brown and co., 1978 Symons C: Thyroid heart disease. Br Heart J 41: 1979 Lilienfeld A: Epidemiological studies of the leukemogenic effect of radiation. Yale J Biol Med 39: 1966 Kereiakes 0. Rosenstein M: Handbook of Radiation Doses in Nuclear Medicine and Diagnostic X ray. Florida: CRC Press, 1980 Labedzki L: Hematological toxicity of total body irradiation. Strablentberapie 156: 1980 Hibbs CH, McClellan RO: Radiation and radioactive materials. In Doull J: Toxicology. Doull J (Ed) 2nd edn. New York: Macmillan, 1980 Wellman HN, et al: MIRD dose estimate report No. 5. J Nucl Med 16: 1975 Glanzmann C, et ul: Iodine-125 and iodine-131 in the treatment of hyperthyroidism. Clin Nucl Med 5: 1980 Tubiana N: Effects of radiation on bone marrow. Patbol Biol 27: 1979 Dobyns BM, et al: Malignant and benign neoplasms of the thyroid in patients treated for hyperthyroidism. J Clin Endocr Metab 38: 1974 Saenger EL, Thoma GE. Thompkins EA: Incidence of leukemia following treatment of hyperthyroidism. JAMA 205: 1968