Tumor-induced osteomalacia: Lessons learned



Tumor-induced osteomalacia is a rare acquired metabolic disorder characterized by hypophosphatemia and inappropriately low serum levels of 1,25-dihydroxyvitamin D. Symptoms include chronic muscle and bone pain, weakness, and fatigue in association with a high risk of fragility fractures due to osteomalacia. The diagnosis is commonly delayed for years due to the nonspecific nature of the presenting symptoms, failure to include determination of serum phosphorus levels in blood chemistry testing, and difficulty in identifying the responsible tumor. The pathogenesis of tumor-induced osteomalacia involves tumor expression of fibroblast growth factor 23, a hormone that inhibits proximal renal tubular reabsorption of phosphate and down-regulates renal conversion of 25-hydroxyvitamin D to its active form, 1,25-dihydroxyvitamin D. The metabolic abnormalities may be partially or completely corrected with phosphate supplementation and calcitriol. A definitive diagnosis and treatment require excision of the responsible tumor.

In this report, we describe an unusual case in which the patient presented with bone and muscle pain, progressive muscle weakness, and fragility fractures due to an unrecognized metabolic bone disease. Despite evaluation with numerous laboratory tests and imaging studies by many specialists over a period of almost 2 years, the origin of his symptoms remained obscure. At various times he was treated, unsuccessfully, for piriformis syndrome, seronegative polyarthritis, severe osteoporosis, and polyostotic Paget's disease of bone. Following a hip fracture, further studies suggested the presence of tumor-induced osteomalacia, also called oncogenic osteomalacia. This is a rare paraneoplastic disorder characterized by renal phosphate wasting associated with hypophosphatemia and inappropriately low blood levels of 1,25-dihydroxyvitamin D (1,25-[OH]2D), resulting in a condition similar to hypophosphatemic rickets. The diagnosis of tumor-induced osteomalacia was confirmed by the dramatic improvement of symptoms and correction of metabolic abnormalities following excision of the responsible tumor.

This case illustrates the importance of including metabolic bone disease in the differential diagnosis of patients with persistent rheumatic symptoms of unknown origin and the perils of relying on standard laboratory panels that do not provide important diagnostic information.


The patient, a 46-year-old Hispanic male cigarette smoker, considered himself healthy and vigorous until April 2005, when he noted the onset of progressively severe pain in his buttocks, low back, arms, and legs. He was initially evaluated by his primary care physician and later by specialists that included an oncologist, an orthopedic surgeon, a physiatrist, and several rheumatologists. An injection of botulinum toxin type A in the right buttock for suspected piriformis syndrome provided no relief. A glucocorticoid injection in the left sacroiliac joint for apparent sacroiliitis did not help. Nonsteroidal drugs, including diflunisal, celecoxib, oxycodone, and pregabalin, did not relieve the pain.

In April 2006, a physical examination showed that the patient had great difficulty sitting or arising from a seated position. Getting up on the examination table caused much pain in his lower back, buttocks, and thighs, and moderate pressure produced pain over the ribs of the anterior chest and the left pectoral region. The patient had tenderness in both forearms and upper arms as well as in both quadriceps areas. His knees were painful on flexion and extension, with no sign of synovitis or fluid in any peripheral joint.

Laboratory tests showed that the white blood cell count and differential, red blood cell count, platelet count, and erythrocyte sedimentation rate (11 mm/hour) were normal. The comprehensive metabolic panel, a standardized set of blood chemistries performed at laboratories throughout the US, showed normal results for glucose, blood urea nitrogen, creatinine, sodium, chloride, CO2, the anion gap, total protein, albumin, globulin, the albumin:globulin ratio, calcium, aspartate aminotransferase, alanine aminotransferase, and total bilirubin, with an elevated level of total alkaline phosphatase of 205 IU/liter (reference range at outpatient laboratory 41–130). The level of bone-specific alkaline phosphatase was high at 164 IU/liter (normal 12–84). Other laboratory tests that had normal or negative results included tests for rheumatoid factor and antinuclear antibody, prostate-specific antigen level and C-reactive protein level, serum protein electrophoresis, and urinalysis. The clinical impression at that time was widespread metastatic disease because of the rib tenderness, findings of increased uptake of radioactive-labeled technetium in several ribs on a previous nuclear bone scan, and past elevation of the serum alkaline phosphatase level.

In May 2006, an attempted transiliac bone biopsy (without double tetracycline labeling) produced multiple tissue fragments that showed no evidence of malignancy in the marrow cells, and results of flow cytometry were normal. Unfortunately, the specimens were uninterpretable for evaluation of metabolic bone disease, due to fragmentation and lack of tetracycline labeling. A 1-week course of sulfasalazine and then a 2-week course of prednisone in dosages as high as 60 mg/day provided no relief of the patient's unrelenting symptoms. Over the following months, efforts to relieve pain with nonsteroidal drugs (including narcotic analgesics) continued, and additional studies were done. A total body nuclear bone scan with radiolabeled technetium, performed in August 2006, showed discrete foci of intense uptake in the right anterior third and seventh ribs, with increased uptake in both sacral alae, intense increased uptake on both sides of the knee joints (particularly on the right side), increased uptake at both ankle joints, and discrete focal increased uptake in the right mandible anteriorly.

In September 2006, evaluation by rheumatologists at a tertiary care medical center resulted in a diagnosis of Paget's disease. The patient was treated with oral alendronate at a dosage of 40 mg/day and then with 2 intravenous infusions of zoledronic acid (5 mg), first in November 2006 and again in December 2006. None of the treatments for Paget's disease resulted in an improvement in symptoms. In October 2006, a magnetic resonance imaging study of the right knee revealed multiple incomplete fracture planes paralleling the fused epiphysial plates in both distal femurs and proximal tibias medialis, corresponding to the areas of abnormal increased uptake on prior nuclear bone scans. The results of bone density testing by dual x-ray absorptiometry were interpreted as osteoporosis, based on a lumbar spine (L2–L4) T score of −3.0, using a female reference database. Later the same month, a computed tomography (CT) scan of the abdomen was read as normal, with no evidence of metastatic disease in the liver and no evidence of enlarged retroperitoneal lymph nodes. CT examination of the pelvis showed undisplaced fractures through both sacral alae and through the anteromedial aspects of the iliac bones, extending into the sacroiliac joints bilaterally, corresponding with areas of increased uptake on the previous nuclear bone scan.

In November 2006, the serum 25-hydroxyvitamin D (25[OH]D) level was 18 ng/ml (desirable level >30), and the serum C-telopeptide of type I collagen level was 254 pg/ml (reference range 60–700). The patient's pain and muscle weakness were now so severe that he used a wheelchair most of the time.

On February 1, 2007, the patient fell at home, sustaining an intertrochanteric fracture of the left hip. He was admitted to the hospital for surgical repair. Bone fragments obtained at the time of surgery were sent to the laboratory for analysis. There was no histologic evidence of malignancy or Paget's disease, although evaluation for metabolic bone disease was limited by the absence of tetracycline labeling. Upon admission to the hospital, the level of serum phosphorus, which was measured for the first time in the course of his protracted illness, was low at 1.5 mg/dl (normal 2.4–5.0), with values as low as 0.5 mg/dl prior to sufficient phosphorus replacement. Selected baseline hospital laboratory values are shown in Table 1. The renal tubular reabsorption of phosphate was low at 40% (normal 83–95%). The renal threshold phosphate concentration, calculated from the nomogram described by Walton and Bijvoet (1), was 0.6 mg/dl (normal 2.5–4.2). The serum FGF-23 level was markedly elevated at 262 RU/ml (reference range <180; Mayo Clinic Reference Laboratory, Rochester, MN).

Table 1. Results of selected hospital laboratory tests prior to tumor excision*
TestBaseline valueReference value
Serum calcium, mg/dl7.48.1–10.2
Serum ionized calcium, mg/dl4.64.6–5.4
Serum phosphorus, mg/dl1.52.4–5.0
Serum alkaline phosphatase, units/liter17720–150
Serum 25-hydroxyvitamin D, ng/ml1830–60
Serum 1,25-dihydroxyvitamin D, pg/ml615–60
Serum intact parathyroid hormone, pg/ml129.314–72
24-hour urinary phosphorus, gm2.050.9–1.3
24-hour urinary calcium, mg22.550–300

A presumptive diagnosis of tumor-induced osteomalacia was made on the basis of clinical and laboratory findings consistent with osteomalacia, hypophosphatemia, and elevated serum levels of FGF-23. A radiolabeled total body octreotide scan showed an intense focus of uptake in the shaft of the right mandible (Figure 1); this was the same region that showed increased uptake on a previous technetium bone scan. Several days later, the right mandible lesion was excised by an oral surgeon (EJU). The pathology report showed a benign collection of fibrils and mesenchymal tissue intertwined with apparent osteoid, with no histologic features of malignancy (Figure 2). Postoperatively, the patient was treated with calcitriol, large doses of oral phosphorus, and calcium supplementation. He began rehabilitation and gradual ambulation in the hospital and was discharged to home care. On a repeat assay performed 10 days after surgery, serum FGF-23 was undetectable. Four weeks after hospital discharge, the patient was walking with a walker, with great improvement in the chronic bone and muscle pain. A repeat radiograph of the left hip showed good position of the intrafemoral rod and radiographic signs of bone healing. Five months later, he continued to improve and was able to ambulate with the aid of a cane and work at his regular job.

Figure 1.

Radiolabeled octreotide scan showing intense uptake in the patient's right mandible (arrow).

Figure 2.

Microscopic appearance of fibrous tumor removed from the patient's mandible, showing numerous fibroblasts and spindle cells surrounding areas of osteoid deposits (original magnification × 125).


The diagnosis of tumor-induced osteomalacia is a challenge under the best of circumstances. Symptoms of osteomalacia are nonspecific and may be incorrectly attributed to fibromyalgia, inflammatory arthritis, or a functional disorder. Tumor-induced osteomalacia should be included in the differential diagnosis in patients with progressive weakness, bone and muscle pain, and fractures. Proximal myopathy, which is commonly reported in patients with osteomalacia, may cause a waddling gait and difficulty with ambulation. Proximal myopathy was particularly disabling in this patient, eventually resulting in the need for a wheelchair. The typical time from the onset of symptoms to a presumptive diagnosis of tumor-induced osteomalacia is often longer than 2.5 years (2), and the average time from then until the responsible tumor is identified is ∼5 years (3). The search for these tumors is confounded by their usual small size, slow growth, and difficulties with visualization on standard imaging studies.

Although in this patient the period of time from the onset of symptoms to diagnosis may have been shorter than average, he suffered for almost 2 years with his disease and its complications. His presentation and subsequent course provide an interesting learning experience that may lead to a faster diagnosis for other patients with tumor-induced osteomalacia. First, if serum phosphorus had been included in the original blood chemistry panels to evaluate his symptoms, the finding of a low level would likely have led to an evaluation of the causes of hypophosphatemia and phosphate-wasting syndromes. It is especially important to measure serum phosphorus levels in patients with elevated alkaline phosphatase levels and to obtain the blood specimen when the patient is in a fasting state. Unlike the tight control of serum calcium, phosphorus levels may fluctuate according to variables that include time of day and dietary intake. Serum phosphorus levels are maintained in the normal range due to complex interactions among intestinal absorption, exchange with intracellular and bone storage pools, and renal tubular reabsorption, with the kidney being the principal organ responsible for phosphorus homeostasis (4).

The differential diagnosis of hypophosphatemia includes autosomal-dominant hypophosphatemic rickets (ADHR), X-linked hypophosphatemic rickets (XLHR), and tissue-induced osteomalacia. Although ADHR and XLHR are biochemically indistinguishable from tissue-induced osteomalacia, the former two disorders typically present in childhood and are associated with lower-extremity deformities and short stature. A definitive diagnosis can be made by commercially available genetic testing of the FGF-23 and PHEX genes, which are defective in ADHR and XLHR, respectively, but not tissue-induced osteomalacia (4). Second, the numerous abnormalities found on imaging studies of the ribs, femurs, spine, and pelvis, in association with the absence of malignancy on bone marrow examination, should have suggested the possibility of osteomalacia or other types of intrinsic bone disease. Finally, although the performance of a bone biopsy was appropriate in the evaluation of this patient, its value was greatly diminished by the absence of double tetracycline labeling, inappropriate technique in obtaining the specimen, and failure to send the specimen to a center with experience in diagnosing metabolic bone disease.

The complex clinical presentation was punctuated by the patient's hip fracture, leading to hospitalization with appropriate metabolic evaluation. Because mesenchymal tumors that may cause tumor-induced osteomalacia commonly express somatostatin receptors (5), an octreotide scan using a radiolabeled somatostatin analog was performed. This clearly identified the small mandible tumor responsible for the elevated level of FGF-23. Interestingly, the mandibular abnormality was also seen, but not recognized as being clinically significant, on the previous technetium bone scan, which is considered to be an insensitive diagnostic tool for identifying these tumors (2).

Tumor-induced osteomalacia is typically caused by benign mesenchymal tumors but has also been associated with various carcinomas, sarcomas, neurofibromatosis, linear nevus syndrome, and fibrous dysplasia of bone. The tumors may occur in either bone or soft tissue and seem to favor the extremities and craniofacial region. The pathophysiology of tumor-induced osteomalacia is 2-fold: hypophosphatemia secondary to impaired proximal renal tubular reabsorption of phosphate, and inhibition of renal 25(OH)D–1-α-hydroxylase activity, which blunts the compensatory rise in serum 1,25-(OH)2-D in response to hypophosphatemia.

Since the earliest descriptions of tumor-induced osteomalacia (6, 7), continuing efforts have been made to identify humoral factors, or phosphatonins, responsible for the profound metabolic abnormalities. Although a variety of phosphatonin candidates are expressed by tumors associated with tumor-induced osteomalacia (8–10), accumulating evidence suggests that FGF-23 plays a key role in the pathogenesis of tumor-induced osteomalacia (8) and other disorders of renal phosphate wasting, such as ADHR (11) and XLHR (12). This is supported by an analysis of 32 cases of mesenchymal tumors associated with documented or suspected tumor-induced osteomalacia, in which FGF-23 was expressed in 17 of 21 cases that were evaluated by immunohistochemistry with polyclonal antibody and in 2 of 2 cases evaluated by reverse transcriptase–polymerase chain reaction (13). FGF-23 expression appeared to be localized to highly vascularized spindle cells that were characteristic of these tumors and likely responsible for tumor-induced osteomalacia.

In the patient described here, a high blood level of FGF-23 was detected before the mandible tumor was removed, and 10 days after the tumor was removed, FGF-23 was undetectable. In some patients with a presumptive diagnosis of tumor-induced osteomalacia, the responsible tumor is never found, and ongoing treatment with phosphorus supplements and calcitriol is required. The definitive treatment of tumor-induced osteomalacia is resection of the responsible tumor.


Dr. Lewiecki had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Lewiecki, Williams.

Acquisition of data. Lewiecki, Williams.

Analysis and interpretation of data. Lewiecki, Williams.

Manuscript preparation. Lewiecki, Urig, Williams.

Oral and maxillofacial surgery. Urig.


The image in Figure 1 was photographed by Mike Adam.