| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Joint Baylor College of Medicine-The University of Texas M. D. Anderson Cancer Center Endocrinology Fellowship Program (S.-C.J.Y., A.C.C.); and Section of Endocrinology (R.V.-S., R.F.G.), Department of Medical Specialties, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030
| Abstract |
|---|
|
|
|---|
| I. Introduction |
|---|
|
|
|---|
The goal of this review is to compile and organize a complicated and incomplete literature. In our review of the subject, we encountered several significant challenges. The first was the fragmented nature of the literature, composed of case and small series reports, and information provided by manufacturers. It was difficult in most cases to determine the incidence or prevalence of side effects associated with individual agents because most chemotherapeutic protocols involve multiple agents, and the protocols being used to treat individual malignancies are changing continuously. A second challenge was the lack of information on newer therapies in which endocrine complications may be delayed. As we contemplated the best medium to present this information, we decided to combine a written review with an interactive hypertext-based medium accessible on the Internet.2 The resulting document is intended to be a fluid medium that will, we hope, change and grow over time as more information becomes available. Our position in a major cancer center provides us with a unique opportunity to evaluate complications of new therapies and to observe the long-term effects of established therapies. The focus of this paper will be on endocrine side effects of nonhormonal therapies, excluding drugs with endocrine actions because of the general familiarity of these drugs within the community of endocrinologists. In cases where an endocrine agent is used in combination with other nonendocrine agents that have similar side effects, the relationship will be discussed.
This review will be organized according to endocrine system
component or side effect, which is the most likely point of access for
the practicing physician. A pharmacological classification of
nonhormonal chemotherapeutic agents and their toxic effects on the
endocrine system will permit rapid evaluation of individual drugs
(Table 1
), and a list of these endocrine
effects and the drugs that cause them will provide disease-oriented
information (Table 2
). These tables in
the on-line version will contain hypertext links for exploration of
specific drugs or complications to the endocrine system.
|
|
Several different mechanisms can be postulated in the pathogenesis of endocrine complications from nonhormonal cancer therapy. First, nonhormonal antineoplastic agents, in general, are cytotoxic, and the majority affect DNA replication, transcription, protein synthesis, or microtubule function, which are essential processes for cell division, growth, and regeneration. The subsequent death or injury of endocrine cells can result in glandular dysfunction. Second, an agent can interfere with the synthesis or normal processing of a hormone at the transcriptional, translational, or posttranslational levels. Third, an agent can enhance or inhibit secretion of a hormone through interaction with receptors or perturbation of second or third messenger metabolism. Fourth, an agent can perturb hormone delivery by changing the level of a carrier protein in serum or by competing for the binding sites on the carrier protein. Finally, a nonhormonal antineoplastic agent can potentiate or block hormone action at the end organ by interacting with the signal transduction pathway. Specific mechanism(s) of an endocrine effect will be discussed where available, although we will avoid a general discussion of the mechanism of antineoplastic effects of particular drugs, which is available elsewhere (3).
| II. Disorders of Glucose Metabolism |
|---|
|
|
|---|
1. Interferon-
2. Interferons are small peptide cytokines
(molecular masses of 1627 kDa) that interact with specific cell
surface receptors and modulate several cellular functions (7, 8, 9). The
endocrine effects of these cytokines have recently been reviewed (10).
In the United States, only recombinant forms of interferon-
2 have
been approved by the Food and Drug Administration for treatment of
certain types of malignancy, and we shall limit our discussion to this
drug agent.
Use of recombinant interferon-
2a and
2b to treat malignancies has
been associated with the development of hyperglycemia in previously
nondiabetic patients and worsening of glycemic control in known
diabetics (11, 12, 13, 14). Similar association of interferon-
and diabetes
mellitus is found in the treatment of viral hepatitis (15, 16, 17, 18).
Although the incidence of interferon-
-induced diabetes mellitus in
cancer patients is unclear, the incidence of diabetes mellitus is about
0.7% among patients who received 480800 MU of
interferon-
over 24 weeks for chronic active hepatitis C (19).
Diabetic ketoacidosis has been reported in a variety of conditions
treated with interferon-
2 (13, 14, 20). Several mechanisms have been
described for interferon-
-induced diabetes, the most important of
which is the effect of interferon-
to induce or enhance an existing
immune response (15, 21, 22, 23). Antiislet antibodies, or cell- or
antibody-mediated cytotoxicity, are perhaps involved in causing insulin
deficiency. A direct inhibition of preproinsulin synthesis in islet
ß-cells by interferons may also contribute to insulin deficiency
(24). Enhanced peripheral resistance to insulin has been suggested by
insulin clamp studies and by glucose tolerance test results (20).
Interferon-induced increase in cortisol, GH, and glucagon secretion may
contribute to the insulin resistance (20, 25). Induction of insulin
autoantibodies by interferon-
has also been reported (26, 27).
2. L-Asparaginase. L-Asparaginase is an enzyme derived from microorganisms. It inhibits protein synthesis by depletion of L-asparagine and is used primarily in the treatment of hematological malignancies. Hyperglycemia and glycosuria without ketonemia occurs in 114% of patients treated with L-asparaginase, an effect that is reversible upon discontinuation of the drug (28, 29, 30). Insulin therapy is frequently required, but close monitoring of blood glucose is necessary to avoid hypoglycemia after cessation of L-asparaginase. Hyperglycemia can be worsened by the concurrent administration of high-dose glucocorticoid in combination regimens (31, 32). Diabetic ketoacidosis has been reported (28, 33); after cessation of L-asparaginase, normalization of all laboratory parameters and no recurrence of diabetes or diabetic ketoacidosis in the absence of insulin therapy has occurred in at least one case (R. F. Gagel and R. V. Sellin, personal observation).
Inhibition of insulin (29, 30, 33, 34) or insulin receptor synthesis (35), leading to a combined insulin deficiency/resistance syndrome, is the presumed mechanism of the L-asparaginase effect. Paradoxically, hyperglucagonemia has been demonstrated in a single case of L-asparaginase-induced hyperglycemia (36). Why L-asparaginase targets insulin, insulin receptors, thyroid hormone-binding protein, and albumin synthesis, but not other proteins such as glucagon, is not clear.
Pancreatitis, which occurs in 12% of L-asparaginase-treated patients, provides yet another mechanism for the development of transient or permanent diabetes mellitus. The incidence of pancreatitis rises when L-asparaginase is combined with other drug therapy, i.e., 6.2% in the combined regimen of L-asparaginase, doxorubicin, vincristine, and prednisone (L-AdVP) (37).1
3. Streptozocin. Streptozocin (formerly called streptozotocin) is an N-nitrosourea derivative of glucosamide that is primarily used to treat malignant islet cell tumors and carcinoid tumors (38, 39). Streptozocin causes islet cell inflammation and destruction (40). This effect is species-specific and dose-related. Rats are most susceptible, and streptozocin has been used to create experimental models of diabetes mellitus (41, 42). The initial effect of streptozocin is degranulation of ß-cells (40), which can cause hypoglycemia in patients with insulin-secreting tumors. Human islet cells are less susceptible to the deleterious effects of streptozocin (43), and most streptozocin effects are reversible after cessation of therapy. Although the incidence of glucose intolerance is 660%, most cases are of mild to moderate severity and reversible upon discontinuation of the drug (38, 39, 44). In one study, about 1% of patients had persistent hyperglycemia during short-term follow-up of five months (38). The generally poor prognosis of metastatic islet cell carcinoma has made it difficult to determine whether streptozocin has deleterious long-term effects on human islet cell function, leading to diabetes mellitus.
B. Glycosuria and ketonuria
Some antineoplastic drugs affect the renal excretion of glucose
without actually affecting glucose metabolism. Drugs such as ifosfamide
or mercaptopurine can damage the renal tubules and cause glycosuria or
Fanconi syndrome (45, 46, 47, 48). A false positive reaction with the testing
agent for urinary ketones can be caused by mesna (2-mercaptoethane
sulfonate sodium) (49), an agent usually given together with ifosfamide
to decrease the incidence of hemorrhagic cystitis.
| III. Disorders of Free Water Clearance |
|---|
|
|
|---|
A. Syndrome of inappropriate ADH (SIADH)
Toxicity or damage to nerve terminals in the posterior pituitary
may cause inappropriate release of ADH [also known as arginine
vasopressin (AVP)], which leads to retention of free water and
hyponatremia. SIADH is characterized by hyponatremia, low serum
osmolality, and an inappropriately high urine osmolality in the absence
of diuretics, heart failure, cirrhosis, adrenal insufficiency, and
hypothyroidism. In cancer patients, SIADH may be caused by AVP
production by the tumors (e.g., up to 15% of small cell
lung cancer), abnormal secretory stimuli (e.g.,
intrathoracic infection, positive pressure ventilation), or
cytotoxicity affecting paraventricular and supraoptic neurons (50). It
is also possible that chemotherapy-induced lysis of AVP-containing
cancer cells leads to or worsens SIADH. Most of the literature
regarding SIADH caused by chemotherapeutic agents developed before the
availability of reliable AVP assays, and inappropriate elevation of
plasma AVP is not included in the definition of SIADH. Drug-induced
renal salt wasting [mediated by renal tubular toxicity,
e.g., cisplatin (51, 52)] or tumor-induced renal salt
wasting (mediated by atrial natriuretic peptides) (53, 54, 55) can also
cause hyponatremia, hypoosmolality, elevated urinary sodium, and
urinary osmolality. These SIADH-like syndromes are difficult to
distinguish from SIADH when signs and symptoms of fluid volume
depletion are subtle or absent. Nonetheless, there are convincing
reports that provide evidence of chemotherapy-induced hypothalamic or
posterior pituitary damage in the context of SIADH (56, 57, 58).
1. Vinca alkaloids. Vincristine and vinblastine are vinca alkaloids that inhibit tubulin polymerization. There have been at least seven reports associating vincristine use with SIADH (56, 59, 60, 61, 62, 63, 64), and some of these reports documented inappropriately high serum levels of ADH (61, 62). Further evidence for a drug effect is the recurrence of SIADH during subsequent therapy with vincristine (61). Vinblastine has also been reported to cause severe hyponatremia (65), and SIADH (66, 67, 68, 69). The incidence of SIADH induced by a vinca alkaloid alone has not been addressed in any large series. When vincristine is used in combination with cyclophos-phamide, doxorubicin, methotrexate, and prednisolone (CHOP-M), the incidence of hyponatremia is about 17% (70). In another series in which high-dose vinblastine, cisplatin, and bleomycin therapy were combined, eight of 12 patients developed hyponatremia and hypoosmolality (71). The presumed mechanism of vinca alkaloid-induced SIADH is paraventricular or supraoptic cell microtubular damage (56, 57). Another contributing factor, identified in rodents treated with vincristine, is gastrointestinal sodium and water loss leading to appropriate AVP secretion (72).
2. Cyclophosphamide. Cyclophosphamide therapy has been associated with hyponatremia (73, 74, 75) and SIADH (58, 76, 77, 78). Reversible SIADH has been reported in two patients treated with high-dose cyclophosphamide (50 mg/kg) (77, 78). A direct effect of cyclophosphamide on the hypothalamus is supported by autopsy findings of infundibular necrosis, decreased intraaxonal secretory granules, and depletion of posterior pituitary AVP in a case of fatal cyclophosphamide (1800 mg/m2)-induced hyponatremia (58). Other mechanisms have also been invoked. In a small series of 19 chemotherapy courses with lower dosages of cyclophosphamide, there was development of hyponatremia, plasma hypotonicity, and urinary hypertonicity without an increase in plasma AVP levels (79). This suggests that damage to the renal tubules and resultant defects of salt and water transport is the major cause of hyponatremia associated with low-dose therapy.
3. Cisplatin. There are many reports of cisplatin-induced hyponatremia due to renal salt wasting (51, 52, 80, 81, 82, 83, 84, 85). Several reports claim that cisplatin induces SIADH (86, 87, 88), and there is a case report in which the hyponatremia occurred because of combined effects of cisplatin, thiazide, and chlorpropamide (89). The mechanism of cisplatin-induced hyponatremia is unclear, but it has been suggested that renal toxic effects of cisplatin, causing a decrease of papillary solute content and maximal urinary osmolality (90, 91), is the major factor, rather than a direct effect, of AVP secretion. In a study of 70 patients, seven developed salt-wasting nephropathy and hyponatremia after receiving 342 mg/m2 (range: 200600 mg/m2) of cisplatin-based chemotherapy (85). Five of these patients had elevated levels of AVP that became suppressed with correction of hypovolemia. The stimulus for AVP release in these patients was probably hypovolemia caused by renal salt wasting.
4. Melphalan. Melphalan-induced hyponatremia and urinary sodium loss, thought to be SIADH, has been reported in two patients who received high-dose melphalan (2 mg/kg) in combination with vincristine and cyclophosphamide (92). Melphalan was implicated in the causation because challenge with the same regimen but a lower dose of melphalan (0.5 mg/kg) in one patient several weeks later did not cause hyponatremia. Whether melphalan potentiated the effect of the vincristine/cyclophosphamide combination or had a direct effect is unclear.
B. Nephrogenic diabetes insipidus
Nephrogenic diabetes insipidus can result from the effects of
ifosfamide or streptozocin on tubular reabsorption of water. Although
both drugs have cytotoxic effects on renal tubular cells, the cellular
mechanism of nephrogenic diabetes insipidus is unclear.
1. Ifosfamide. Ifosfamide has broad nephrotoxic effects, although tubular damage predominates (48, 93, 94, 95, 96, 97). In 11 young patients treated with ifosfamide alone, proximal renal tubular effects developed in all (96). Distal renal tubular defects developed in approximately one half of these, including one with nephrogenic diabetes insipidus.
2. Streptozocin. Streptozocin (also known as streptozotocin) is also nephrotoxic (98, 99). In addition to glomerular (proteinuria) and tubular (Fanconi syndrome) defects (99), two cases of nephrogenic diabetes insipidus have been reported after streptozocin therapy (100, 101). One case recovered spontaneously, and the other showed improvement after indomethacin therapy.
| IV. Disorders of Mineral Metabolism |
|---|
|
|
|---|
A. Hypocalcemia
1. Chemotherapy-induced tumor lysis syndrome. Massive cell
death resulting from cytotoxic chemotherapy can result in severe
hyperuricemia, hyperkalemia, hyperphosphatemia, and hypocalcemia
(102, 103, 104). Several factors contribute to the development of tumor
lysis syndrome, including the type of malignancy, the responsiveness of
the malignancy to therapy, and tumor burden. Hyperuricemia,
hyperkalemia, and hypocalcemia may contribute to the development of
cardiac arrhythmia, tetany, and sudden death. Tumor lysis syndrome
occurs most frequently in hematological malignancies (104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120). In a
series of 41 patients with acute leukemia, 25 developed tumor lysis
syndrome, three of which were moderate to severe (109). In another
retrospective review of 102 cases of high-grade non-Hodgkins
lymphoma, evidence of tumor lysis syndrome was present in 42% of
cases; the syndrome was severe and life-threatening in 6% (115). The
syndrome has been reported in nonhematological malignancies (121, 122, 123, 124, 125, 126, 127),
but these cases are relatively rare.
The occurrence of tumor lysis syndrome has been associated with a wide variety of therapeutic agents: fludarabine (105, 108, 110, 111, 114, 116), mitoxantrone (106), 6-mercaptopurine (128), methotrexate (129), chemotherapeutic conditioning for bone marrow transplant (112, 118) or total body irradiation for bone marrow transplant (130), and various combination chemotherapy regimens (115, 119)
Appropriate therapy to prevent complications of tumor lysis includes hydration, alkaline diuresis (131), inhibition of uric acid synthesis by allopurinol, and oral calcium or aluminum-based compounds to bind intestinal phosphate. In a hyperphosphatemic patient with hypocalcemia, the addition of an oral calcium-based compound will reduce phosphate and enhance calcium absorption. Intravenous calcium administration can potentially cause calcium phosphate precipitation in the presence of severe hyperphosphatemia and should be used cautiously. Dialysis (111, 132, 133) may be required for patients with symptomatic hypocalcemia and a serum phosphorus level > 3.3 mmol/liter (>10.2 mg/dl).
2. Platinum compounds. Hypocalcemia has been reported to occur in 620% of cisplatin-treated patients (134). Effects of cisplatin on renal tubular function (90, 91, 135), magnesium metabolism (see Section IV. C), bone resorption, and vitamin D metabolism may explain the hypocalcemia. Hypomagnesemia may cause a decrease in PTH secretion and a reduction in the calcium-mobilizing effects of PTH (136, 137, 138, 139, 140, 141). Hypomagnesemia also inhibits formation of 1,25-dihydroxyvitamin D (137, 142). In addition, cisplatin may have a direct inhibitory effect on bone resorption. In a prospective study of 13 cancer patients with hypercalcemia refractory to rehydration, treatment with a 24-h infusion of cisplatin (100 mg/m2) resulted in normalization of the serum level of calcium in nine patients (69%), with a mean duration of benefit of 38 days (143). There was no reduction in tumor size in any of these patients, and eight of the nine responders died of progressive cancer while remaining normocalcemic. An abnormality in vitamin D metabolism has also been observed in patients who received chemotherapy that included cisplatin (144). The level of 1,25-dihydroxyvitamin D decreased significantly after chemotherapy, whereas its precursors remained unchanged. The effect was attributed to inhibition of mitochondrial function in the kidney by cisplatin (144).
Carboplatin therapy, like cisplatin, is associated with a 1631% incidence of hypocalcemia, as reported by the manufacturer in the product package insert. There are, however, no reports in the medical literature corroborating this finding.
3. Antitumor antibiotics. Plicamycin (mithramycin) is an antitumor antibiotic that inhibits DNA-dependent RNA polymerase. This drug has a major effect on calcium metabolism (145, 146). It inhibits bone resorption at doses (25 µg/kg) that are ineffective for antineoplastic effects (147), resulting in a lowering of the serum calcium concentration within 2448 h. Plicamycin inhibits basal and PTH-stimulated osteoclast function by unclear mechanisms (148). The inhibitory effect of plicamycin on osteoclast function has made it a useful therapeutic agent for treatment of Pagets disease of bone and osteoclast-mediated hypercalcemia associated with malignancy (145). The hypocalcemic effect of plicamycin, as well as its hepatic and renal toxicity, has limited its usefulness as an antineoplastic agent.
Dactinomycin is another antitumor antibiotic that blocks DNA-directed RNA synthesis, causing hypocalcemia in animals. Dactinomycin abolishes the calcium-mobilizing activity of PTH, presumably by interfering with osteoclast-mediated bone resorption (149, 150, 151).
A syndrome of hypomagnesemia, hypocalcemia, and hypoparathyroidism has been reported with doxorubicin and cytarabine treatment in nine patients (152). Three of the nine patients had elevated phosphate levels of 1.341.92 mmol/liter (normal range: 0.71.25 mmol/liter). The exact mechanism for this effect is not clear.
4. L-Asparaginase. Thirty-seven to 60% of patients treated with L-asparaginase for more than 1 week developed hypoalbuminemia with an associated decrease of total serum calcium (32). The patients had no symptoms of hypocalcemia, and the corrected calcium concentration was normal.
B. Hypercalcemia
The incidence of hypercalcemia in cancer patients is 1.15% (153).
Approximately one fourth of these patients have identifiable bone
metastases. Hypercalcemia in the context of cancer is a poor prognostic
sign, associated with a shorter survival. Renal cell carcinoma,
non-small cell lung carcinoma, and myeloma are most commonly associated
with hypercalcemia. Despite the high prevalence of hypercalcemia in
cancer patients, we were unable to identify any reports of
hypercalcemia caused by an antineoplastic agent. However, the increased
interest in vitamin D analogs for treatment of cancer, and the
increased incidence of parathyroid tumors after head and neck
irradiation, may provide exceptions to this observation.
1. Radiation-induced hyperparathyroidism. Numerous studies have documented the association of low-dose (usually 27.5 Gy) external irradiation to the head and neck area and subsequent development of hyperparathyroidism (154, 155, 156, 157, 158, 159, 160, 161, 162, 163). There is an increased incidence (2.5- to 3-fold increase) of primary hyperparathyroidism after low-dose radiation exposure to the neck (162, 164). Among patients who develop primary hyperparathyroidism, 1430% had prior exposure to radiation (155, 158, 160). The mean interval from irradiation to development of hyperparathyroidism ranges from 2947 yr (154, 158, 165). An association of hyperparathyroidism and radiation exposure from radioactive iodine treatment has also been reported (166, 167). Experiments in animals support a causal relationship of irradiation and parathyroid tumors (168, 169). Concurrent thyroid cancer may be seen in more than 30% of patients with radiation-induced hyperparathyroidism (158, 164, 170). The clinical features of radiation-induced and non-radiation-induced hyperparathyroidism are similar (171, 172).
Hyperparathyroidism after high-dose irradiation is uncommon. In a report by Holten and Christiansen (173), there were no notable changes in the parathyroid function in the first 2 months after irradiation for head and neck cancers. In a subsequent prospective study, Holten and Petersen (174) noted a trend toward higher PTH serum levels 3 yr after irradiation, although there was no definite evidence of hyperparathyroidism (174). In a review of 220 cases of Hodgkins disease treated with neck irradiation, Nader et al. (175) were able to document only one case of primary hyperparathyroidism. In a case of breast cancer treated after mastectomy with multiple radiation sessions including the lower neck, a parathyroid adenoma was identified 21 yr after the last radiation treatment (176).
C. Hypomagnesemia
1. Platinum compounds. Cisplatin has toxic effects on the
kidney, the principal regulatory site for magnesium homeostasis.
Cisplatin causes morphological changes (91, 135) and necrosis (177) in
the proximal tubule, an important site of magnesium reabsorption.
Hypomagnesemia occurs in more than 90% of patients treated with
cisplatin (80, 134, 178, 179, 180, 181, 182, 183), and 10% are symptomatic with muscle
weakness, tremulousness, and dizziness. Vigorous hydration and use of
osmotic diuretics such as mannitol may prevent renal failure but has
little effect on renal magnesium wasting. Hypomagnesemia may persist
long after cisplatin has been discontinued (178, 179, 182).
Carboplatin may also cause hypomagnesemia. There are no large series in the literature, but the information available from the manufacturer (package insert) indicates that 60% of those taking the drug may be affected.
| V. Metabolic Bone Diseases |
|---|
|
|
|---|
In most cases it is not clear whether bone loss is due to the therapy, the underlying disease process (including the impact of cachexia, malnutrition, poor calcium and vitamin D intake), or a combination of the two (184). Bone loss is prominent in patients with disorders affecting hematopoietic cells, perhaps because of cytokine production and their intimate relationship with bone-forming cells (185). For instance, radiographic changes and bone pain are frequent presenting abnormalities of acute lymphocytic leukemia (ALL), a fact that suggests direct impact of ALL on bones (186). In the absence of an adequate control group and control for nutritional variables, it is difficult to ascertain the relative importance of the anticancer agents in causing osteoporosis in the series discussed below.
1. Methotrexate. Osteoporosis (generalized or localized) is observed in children receiving methotrexate therapy for ALL. In one series, five of 11 children treated with oral methotrexate for periods of 762 months developed severe osteoporosis, distal extremity pain, and associated fractures (187). In a second series, five children with ALL developed osteoporosis and lower extremity pain after treatment with methotrexate (30 mg/m2/week ip) for 1217 months. The bone pain resolved promptly upon discontinuation of methotrexate, but radiographic changes in bone persisted for longer periods (188). In a third series, 26 of 216 children treated with methotrexate for ALL developed osteoporosis, bone pain, and fractures (189). In general, the osteoporosis improves significantly after cessation of methotrexate therapy (189, 190). Because ALL can cause osteopenia and bone pain independently (185), perhaps by bone infiltration and cytokine secretion, it is difficult to assess the independent roles of disease and drug in these small series. A role for methotrexate is suggested by the fact that 20 of 37 children treated with methotrexate for leukemias developed osteoporosis and bone pain. In seven of these children, fractures and nonunion or delayed union of fractures were significant problems until methotrexate was discontinued (191).
The mechanism by which methotrexate causes osteoporosis is unclear. There are reports of increased urinary and fecal calcium excretion in some cancer patients treated with methotrexate (189, 192) and decreased bone formation with methotrexate administration in rats (193). These results suggest that methotrexate causes osteoporosis by a combination of decreased bone formation and increased resorption, although firm conclusions are difficult to draw.
2. Platinum compounds. Eleven of 16 ALL patients whose treatment included cisplatin or carboplatin had bone pain, limping, and fracture, with bone mineral densities that averaged 2.3 SDs below normal (194). The known effects of platinum compounds on calcium homeostasis include hypomagnesemia (see Section IV.C), hypocalcemia (see Section IV.A.2), and renal calcium wasting. These effects may cause enhancement of bone loss associated with ALL (186).
3. Combination chemotherapy. The majority of combination regimens for hematopoietic malignancies include high-dose glucocorticoids or methotrexate. Both have effects on bone formation and resorption. In one study, 29 men in complete remission after treatment for Hodgkins disease had reduced cortical and trabecular bone mineral density. There was no correlation in this group with the time since chemotherapy, the chemotherapeutic regimens, or the number of cycles of chemotherapy (195). Hypogonadism resulting from chemotherapy, high-dose glucocorticoid therapy, and Hodgkins disease per se were cited as possible causes of bone loss in these men. In another study of Hodgkins disease patients, no significant bone loss was found in men, but chemotherapy-induced menopause was associated with significant bone loss in women (196). In young women treated for lymphoma, chemotherapy-induced menopause was a more important cause of osteoporosis than cytotoxic agents and glucocorticoids (184).
Adjuvant chemotherapy for breast carcinoma (usually involving 5-fluorouracil, cyclophosphamide, and doxorubicin or methotrexate) is associated with low bone mass. Chemotherapy-induced premature menopause appears to be a major factor. Women with adjuvant chemotherapy-induced ovarian failure have lower cortical and trabecular bone mass than age-matched controls without breast carcinoma (197). The avoidance of estrogen replacement therapy in young women with breast carcinoma and premature ovarian failure compounds the problem and portends a major future health problem.
Bone marrow transplantation has profound effects on the marrow-bone interface. Patients receive large doses of cytotoxic drugs and irradiation designed to eradicate all cells in the marrow. It is therefore not surprising that bone-forming cells are also affected. In 24 patients undergoing bone marrow transplantation, profound effects on bone biomarkers were observed. The serum osteocalcin and alkaline phosphatase levels, thought to be indicators of bone formation, were low. N-telopeptides, bone collagen degradation products thought to be indicative of bone resorption, were increased in this group, which was studied over a 12-week period (198). There are no large-scale studies that address this issue.
B. Rickets or osteomalacia
Osteomalacia results from the failure of the organic bone matrix
to mineralize normally. Rickets is a unique variant in which there is
abnormal mineralization and maturation of the growth plate at the
epiphysis in children. The most common cause for these two conditions
is a decrease in the concentration of serum calcium and/or phosphorus
concentration. Nutritional deficiency and renal wasting of phosphorus
are common causes. Other contributing factors include systemic acidosis
and drugs such as anticonvulsants or aluminum.
1. Ifosfamide. Ifosfamide causes tubular damage leading to
renal phosphate wasting, hypophosphatemia, and rickets. The toxic
effects of ifosfamide on renal tubular function (see Section
III.B.1) include Fanconi syndrome in adults (45, 46) and children
(48). Tubular damage is seen most commonly with doses of ifosfamide
50 g/m2 or when used in combination with cisplatin.
Rickets is reported most commonly in the pediatric population (48, 95, 96, 97, 199, 200) with two of 11 children affected in one series (96).
The long-term outcome has not been reported.
2. Estramustine. In a prospective trial in prostate cancer metastatic to bone, which compared estramustine and clodronate to estramustine and placebo, both treatment groups developed hypocalcemia, secondary hyperparathyroidism, hypophosphatemia, and osteomalacia, with normal vitamin D levels (201). Bone resorption, quantitated by bone biopsy at 3 months, was decreased in both groups. The proposed mechanism for this striking clinical picture is suppression of bone resorption by estramustine, alone or in combination with clodronate, leading to hypocalcemia and secondary hyperparathyroidism, with renal phosphate wasting.
| VI. Thyroid Disorders |
|---|
|
|
|---|
A. Serum thyroid hormone-binding protein abnormalities
Thyroid hormone is bound preferentially in serum to thyroid
hormone-binding globulin (TBG) and albumin. The levels of these binding
proteins can be modified by sex hormones, glucocorticoids, narcotic
use, and nutritional and other factors. The observations presented
below describe changes in thyroid hormone binding garnered from small
published series. These reports describe thyroid hormone abnormalities,
but, in general, lack adequate controls for nutritional status, effects
of chronic illness, and non-antineoplastic drugs and do not provide any
sense of the true incidence of these effects.
1. L-Asparaginase. L-Asparaginase has been shown to affect serum thyroid hormone levels. It may inhibit synthesis of albumin and TBG, resulting in low total but normal free T4 values (202). This effect on TBG synthesis is reversible. Administration of glucocorticoids, frequently used in combination with L-asparaginase (31, 32), may also inhibit TBG synthesis, further lowering total T4 values. L-Asparaginase may also have effects on the pituitary or hypothalamus, causing hypothyroidism, which is discussed later (see Section VI.B.5).
2. 5-Fluorouracil. 5-Fluorouracil increases total T4 and T3 levels, but the patients are clinically euthyroid with a normal free T4 index and normal level of TSH, suggesting that this drug exerts its effect by increasing serum thyroid hormone-binding proteins (203).
3. Mitotane. Mitotane (o,p'-DDD) increases total T4 and T3 levels but without suppression of the TSH levels. In three patients with adrenocortical carcinoma treated with mitotane, the serum TBG level was found to be elevated (204).
4. Combined alkylating agent and podophylline therapy. In 27 patients treated with alkylating agents and podophylline derivatives, there were minor changes in thyroid hormone levels. The serum level of T4 and TBG decreased slightly over a 5-day period (still within the normal range). The serum levels of TSH, rT3, and thyroglobulin and the T4/TBG ratio were not changed (205).
B. Hypothyroidism
1. Irradiation. Irradiation to the head and neck is an
important cause of thyroid dysfunction. Radiation doses greater than 25
Gy to the thyroid gland can result in primary hypothyroidism; doses of
50 Gy or more to the hypothalamic-pituitary region are associated with
primary or secondary/tertiary hypothyroidism (206, 207, 208). The
probability of hypothyroidism after irradiation for lymphoma (209),
Hodgkins disease (210, 211, 212, 213), head and neck carcinoma (214, 215, 216, 217, 218, 219, 220, 221),
breast cancer (222), and bone marrow transplantation (223, 224) ranges
from 1520% (Table 3
). Doses greater
than 30 Gy were associated with a 2550% incidence of primary
hypothyroidism, with most cases identified 45 yr after the
irradiation. The relationship between the probability of hypothyroidism
and time of exposure or radiation dose is most clearly shown in Fig. 1
(210).
|
|
Factors that enhance the risk for development of hypothyroidism include a high radiation dose to the head and neck, duration since therapy, lack of shielding for midline structures, and combined radiation and surgical treatments. Other risk factors include hemithyroidectomy performed during the course of laryngectomy, or damage to the thyroid vascular supply during surgery (219).
There are also significant effects of brain or head and neck
irradiation on hypothalamic and pituitary regulation of TSH secretion.
In the series shown in Fig. 2
, there is
an almost linear time-related diminution in TSH secretion after
treatment with 3875 Gy for brain tumors (227) or nasopharyngeal
carcinoma (215, 228). Fifteen to 20% had diminished TSH secretion at 5
yr and approximately 35% at 10 yr (Fig. 2
). The combined effect of
irradiation on the thyroid and pituitary-hypothalamic area is so
striking that prudence suggests routine screening for this group of
patients. Measurement of both serum T4 and a TSH
concentration at 2- to 4-yr intervals is advisable.
|
3. Interferon-
2. Thyroid dysfunction is a recognized side
effect of interferon-
administration (15). Ten to 15% of
interferon-treated patients develop primary hypothyroidism; 50% have
detectable thyroid antibodies indicative of an underlying thyroiditis
(230, 231, 232, 233, 234). Patients with antithyroid antibodies before interferon
therapy are at higher risk for development of interferon-induced
thyroid dysfunction.
Two different mechanisms have been proposed for the effects of
interferon-
2 on thyroid function. In vitro studies have
shown an inhibition by interferon-
2b of thyroid follicular cell
proliferation and thyroglobulin release, suggesting a direct effect on
thyroid cell function. A second effect is stimulation of cell surface
expression of major histocompatibility class 1 and intercellular
adhesion molecules, with no apparent effect on major histocompatibility
class 2 molecules (235). These findings are thought to explain the
increase of thyroid autoimmunity found in patients with thyroid
effects.
Interferon-induced thyroid dysfunction is generally reversible. In a study of 68 patients receiving a 24-week course of interferon for hepatitis C, five developed primary hypothyroidism. Thyroid function test results in all normalized within 1.5 yr after cessation of interferon (234). The high prevalence of antithyroid antibodies in these patients, however, suggests the possibility of long-term development of thyroid dysfunction.
4. Interleukin-2 (Aldesleukin). Interleukin-2 causes thyroid
impairment in 2035% of treated patients. A common presentation is
the acute onset of painless thyroiditis with hyperthyroxinemia followed
by the development of primary hypothyroidism (236, 237, 238, 239). The
hyperthyroid phase is usually brief and may go unnoticed, whereas
hypothyroidism may last for months and may even become permanent in
some patients. Replacement thyroid hormone therapy is required in about
9% of these hypothyroid patients (238). Thyroid autoantibodies
developed in 28% of the interleukin-2-treated patients (240). Thyroid
dysfunction correlates with the duration and cumulative dose of
interleukin-2 (240). The combination of interferon-
and
interleukin-2 results in a similar clinical picture of thyroiditis
followed by hypothyroidism (236, 241).
5. L-Asparaginase. In addition to the inhibition of TBG synthesis discussed above, there is some evidence that L-asparaginase can cause hypothalamic or pituitary hypothyroidism (242). Nine of 14 children with ALL had a free T4 level less than 1 ng/dl (hypothyroid range) and a normal basal level of TSH after 3 weeks of treatment with L-asparaginase and prednisone, with return of the free T4 level to normal 23 weeks after cessation of L-asparaginase. In this same series, six patients who had a normal baseline TSH response to TRH stimulation before therapy showed no response to TRH after 3 weeks of combined treatment with L-asparaginase and prednisone. A control group of three patients with Hodgkins disease who received similar doses of prednisone plus other cytotoxic agents also showed blunted TSH response to TRH. These results suggest that L-asparaginase combined with prednisone or systemic illness (ALL) reduces the TSH response to TRH. This possibility should be considered in the differential diagnosis of a low serum T4 in this clinical setting.
6. Combination therapy. Several studies have documented an increased incidence of primary hypothyroidism in patients treated with multiple drug regimens with or without radiation. Specific combinations where substantial evidence supports a causative role are presented below.
a. Cisplatin, bleomycin, vinblastine, etoposide, and
dactinomycin.
Fifteen percent of patients who received a
combination of cisplatin, bleomycin, vinblastine, etoposide, and
dactinomycin developed primary hypoparathyroidism, whereas a control
group without chemotherapy did not (243). The correlation of
hypothyroidism and the cumulative doses of cisplatin and vincristine
implicates these two drugs in the causation (243).
b. Mechlorethamine, vinblastine, procarbazine, and
prednisolone.
Forty-four percent of patients who received six
cycles of this regimen for Hodgkins disease developed an elevated
serum TSH concentration; 6% had decreased T4 and free
thyroid index (244). None received irradiation. The length of follow-up
in this study was 164 months (median, 28 months).
c. Brain irradiation and vincristine, carmustine or lomustine, and
procarbazine.
Children with brain tumors (not involving the
hypothalamic-pituitary axis) who receive this chemotherapy combination
and brain irradiation have a 35% incidence of hypothyroidism, compared
with a 10% incidence for brain irradiation alone (245). The highest
incidence occurred when the thyroid gland was included in the radiation
field.
C. Thyrotoxicosis
1. Irradiation. Hyperthyroidism as a result of radiation
therapy is far less common than hypothyroidism (246), but several
studies have described this association (206, 246, 247). Two forms have
been described: radiation-induced thyroiditis and Graves disease.
a. Thyroiditis.
Silent thyroiditis with transient
thyrotoxicosis and low-to-absent radioactive iodine uptake has been
reported with an incidence of 0.6% in patients treated with
radiotherapy alone or in combination with chemotherapy for Hodgkins
disease (210, 247, 248). Most cases of thyroiditis-induced
thyrotoxicosis occurred within 2 yr of radiotherapy and were followed
by hypothyroidism several months later.
b. Graves disease.
Graves disease after radiation therapy
occurs most commonly after treatment of Hodgkins disease (210, 246, 249). Estimates of actuarial risk are 37%, with a relative risk of
7.220.4 in treated patients. The presence of human leukocyte antigen
B8 or DR5 is associated with an increased probability of Graves
disease in this context (249). The finding of the largest number of
affected patients among those treated for lymphoma has suggested a
relationship between the lymphoma and development of Graves disease,
although the syndrome has also been described in patients treated with
radiation for breast, laryngeal, and nasopharyngeal carcinomas (250).
2. Cytokines. Two cytokines, interferon-
2 and
interleukin-2, have been shown to cause hyperthyroidism. Three of 68
patients receiving a 24-week course of interferon for hepatitis C
developed hyperthyroidism; thyroid function tests normalized in all
three within 1.5 yr after discontinuing therapy (234). Graves disease
following transient thyrotoxicosis after interferon therapy has also
been reported (238).
Thyroiditis causing thyrotoxicosis has been reported after
interferon-
2 (251, 252) or combined therapy with interferon-
and
interleukin-2 (236, 241, 253). Interleukin-2 treatment by itself, in
the context of cancer therapy, causes transient hyperthyroidism in
67% (238, 239, 254). Hyperthyroidism was followed by hypothyroidism
in approximately one half of these patients (254). The mechanism of
interleukin-2-induced autoimmune thyroid dysfunction is unclear,
although interleukin-2-induced disruption of self-tolerance has been
suggested as a mechanism (255).
D. Thyroid neoplasm
1. Irradiation. Low-dose radiation exposure is a known cause
of thyroid neoplasia (206, 256, 257, 258, 259, 260, 261, 262, 263, 264). A pooled analysis of multiple
studies defined a dose-response relationship between radiation exposure
and relative risk of thyroid cancer for radiation doses of 05 Gy
(Fig. 3
) (265). Factors that enhance the
risk of radiation-induced thyroid cancer included the female gender
(a 40% higher rate), an age at radiation exposure of less than 15 yr,
and a period of 2030 yr since radiation exposure. Hypothyroidism
per se does not appear to increase the risk of thyroid
cancer, but experimental studies in a rat model have shown accelerated
progression of a radiation-induced thyroid carcinoma by an elevated TSH
level (266).
|
A variety of molecular abnormalities have been identified in papillary thyroid carcinomas (272). Two specific molecular defects have been identified in radiation-induced papillary carcinoma. The first is the presence of ras mutations in about 30% of radiation-induced tumors. There is disagreement about whether K-ras mutations are more common than mutations of the other ras genes in radiation-induced thyroid cancer (273, 274). The second oncogene involved is c-ret. Rearrangements of the c-ret protooncogene (the PTC rearrangement) have been identified in a high percentage of papillary thyroid carcinomas from children exposed to radiation after the Chernobyl nuclear disaster (275, 276, 277). The importance of these oncogene mutations and rearrangement in the genesis of these tumors is unclear.
a. Effects of low-dose radiation exposure to the neck from therapy
of other cancers.
Therapeutic irradiation of anatomical locations
other than the head and neck can result in unintended exposure of the
thyroid gland to low levels of radiation. In an international
case-matched control study of more than 4000 women with second cancers
after radiation treatment for cancer of the cervix, a 2-fold risk of
radiation-induced thyroid cancer was observed with an average dose of
0.11 Gy to the thyroid (278). However, this observation was not
significant statistically. In a Japanese study of women with second
cancers after chemotherapy and irradiation for breast cancer, the
relative risk of thyroid cancer was found to be 3.2, with chest and
supraclavicular exposure responsible for most cases (279).
b. Effects of high-dose radiation exposure to the neck from therapy
of other cancers.
Direct exposure of the thyroid gland to
tumoricidal doses of radiation increases the risk of papillary or
follicular thyroid carcinoma by 2.7- to 3.1-fold (264). The risk in
children is even higher, with a 13-fold increase seen in one series
(280). The development of thyroid carcinoma is most evident in
Hodgkins disease and non-Hodgkins lymphoma where long-term survival
is now common (210, 211, 280). In one series, 12 of 166 children who
had childhood neck irradiation for Hodgkins disease developed thyroid
cancer 719 yr later, compared with two of 750 adults treated
similarly (281). These findings point to a particular sensitivity of
children and indicate that long-term survival is a prerequisite for
development of thyroid cancer.
2. Chemotherapy. There is no convincing evidence that chemotherapy by itself is a risk factor for development of thyroid tumors, although this possibility has been suggested by two case reports, both in children with ALL (282, 283).
| VII. Disorders of Lipid Metabolism |
|---|
|
|
|---|
A. Hypertriglyceridemia
1. L-Asparaginase. L-Asparaginase
depletes asparagine, thereby affecting the synthesis of lipolytic
enzymes, lipoproteins, and lipoprotein receptors. Studies in adults by
Oettgen et al. (34) showed L-asparaginase to
have an initial hypolipidemic effect in most patients. Two adult
patients in this series subsequently developed hyperlipidemia with
increased levels of cholesterol, phospholipids, triglycerides, and
chylomicrons. In a study of 32 children with ALL treated with
L-asparaginase in combination with prednisone, vincristine,
and daunorubicin, hyperchylomicronemia with elevation of total
triglycerides was noted (285). This effect was reversible upon
discontinuance of the drug. The aforementioned ability of
L-asparaginase to inhibit insulin synthesis (see
Section II.A.2) may contribute to hypertriglyceridemia
although no studies have been performed to characterize this specific
effect. Another potential factor in the development of lipid
abnormalities is pancreatitis induced by L-asparaginase
(286, 287, 288, 289), although this relationship has not been proved.
2. Interferon-
2. Interferons cause hypertriglyceridemia by
increasing hepatic and peripheral fatty acid production (290) and by
suppressing hepatic triglyceride lipase (291). Long-term treatment with
interferon-
2 causes hypertriglyceridemia in approximately one third
of patients, most of whom have had previous serum lipid abnormalities
(292). Elevation of serum triglyceride levels to more than 1000 mg/dl
is not unusual. In one case, a therapeutic effect of diet and
gemfibrozil was observed in the presence of continued interferon-
treatment (293).
3. Retinoic acid derivatives. All-trans-retinoic acid (tretinoin) and other retinoic acid derivatives have been used in the treatment of several malignancies, most notably head and neck cancer and acute promyelocytic leukemia (294, 295). These compounds interact with one of several retinoic acid receptors to modify cell growth and induce terminal differentiation (296, 297). Their effects on lipid metabolism are well characterized (298, 299, 300, 301, 302, 303, 304, 305, 306, 307), although the mechanism of development of lipid abnormality is less clear. These abnormalities include hypertriglyceridemia (294, 295, 308) caused by elevated very-low-density lipoprotein (300, 306) and hypercholesterolemia associated with increased low-density lipoproteins (LDL) (306). Retinoid-induced hypertriglyceridemia has been reported to cause cerebrovascular infarction (308) and pancreatitis (304, 309). Hyperlipidemia associated with retinoid therapy has been treated with gemfibrozil (310, 311) or fish oil (312, 313).
B. Hypercholesterolemia
1. Mitotane. Several reports have documented new onset or
worsening of hypercholesterolemia in patients receiving mitotane for
treatment of adrenocortical carcinoma (314, 315). The etiological
mechanism may be related to inhibition of cholesterol oxidase (316).
Total cholesterol ranged from 368 mg/dl to about 560 mg/dl after 3
months or more of therapy, and the magnitude of the abnormality
appeared to be dose-related (314). The poor prognosis associated with
adrenocortical carcinoma suggests that a mild to moderate elevation of
cholesterol may have little importance in patients with aggressive
disease. However, in long-term (510 yr) survivors, continued
treatment with mitotane may result in the accelerated development of
atherosclerotic disease. Management of mitotane-induced lipid
abnormality in long-term survivors has not been addressed.
2. Cisplatin-based combination chemotherapy. A prospective study showed that cisplatin-based combination chemotherapy regimens for testicular cancer [cisplatin, etoposide, and bleomycin (PEB); alternating cisplatin, vincristine, methotrexate, and bleomycin with dactinomycin, cyclophosphamide, and etoposide (POMBACE)] increased the total cholesterol and LDL in 14 of 17 patients treated for 624 months (317). Seven of these patients had significant elevations of total cholesterol, although the long-term significance of these findings is not clear.
| VIII. Adrenal Dysfunction |
|---|
|
|
|---|