The development of hypophosphataemic osteomalacia with myopathy in two patients with HIV infection receiving tenofovir therapy

Authors


Dr M. J. Parsonage, Department of Infectious Diseases & Tropical Medicine, North Manchester General Hospital, Delaunays Road, Crumpsall, Manchester M8 5RB, UK. Tel: 0161 7202733; fax: 0161 7202139; e-mail: ed.wilkins@pat.nhs.uk

Abstract

We report two cases in which osteomalacia developed in patients on tenofovir-containing highly active antiretroviral therapy (HAART) in the context of Fanconi syndrome with hypophosphataemia. Bone pain was the presenting feature and myopathy followed in one case. Disability was reversed with withdrawal of the drug and with mineral supplementation. The cases highlight the importance of considering the diagnosis of osteomalacia in patients treated with tenofovir. A possible association with incipient acute renal failure, particularly during nonsteroidal anti-inflammatory drug (NSAID) use, needs further investigation.

Introduction

Reduced bone mineral density has become increasingly important within the HIV-infected cohort as longevity has increased. It has been documented in a number of studies [1,2] and is possibly linked to the use of highly active antiretroviral therapy (HAART) [3], although, in studying this effect, it is difficult to separate the relative contributions made by length of time since diagnosis and HAART use, the two being interdependent.

Osteomalacia, however, has been documented in only a few cases [4] and the myopathy of osteomalacia not at all. Tenofovir causes recognized Fanconi syndrome and hypophosphataemia [5,6], although osteomalacia has rarely been linked to such treatment [7]. Osteomalacia has also occasionally been linked to cidofovir and adefovir, other nucleotide analogues [7]. In animal studies using high-dose tenofovir, osteomalacia was a frequent adverse event [8]. In other settings, however, Fanconi syndrome is known to cause osteomalacia and myopathy, which is reversible if the underlying cause can be corrected [9]. Such instances have included Fanconi syndrome secondary to monoclonal gammopathy [10] and ifosfamide treatment [11]. Hypophosphataemia caused by malabsorption syndrome [12], parathyroid hormone-related protein-producing tumours or other tumours (so called oncogenic osteomalacia) [13–15] also produces a reversible osteomalacia, mediated via the phosphaturic hormones fibroblastic growth factor 23 (FGF 23) and matrix extracellular phosphoglycoprotein (MEPE), themselves controlled by the membrane metalloprotease PHEX [16]. In such cases the hypophosphataemia can lead to pathological fractures and tetany [15]. Fractures or Looser's zones can be misdiagnosed as disseminated malignancy, leading to diagnostic delay and disability [17].

Case 1

A 47-year-old homosexual man was diagnosed with AIDS in 1989. His past medical history had included Kaposi sarcoma (treated in 1999) and hepatitis B infection (now surface antigen negative), but no abnormality of blood pressure, of renal function, or of phosphate metabolism and no history of a concomitant potentially nephrotoxic agent. He had been on antiretroviral therapy since 1991, with a triple class resistant virus.

In February 2002, as a result of virological failure on his previous salvage regimen, he was commenced on didanosine 250 mg daily, lamivudine 150 mg twice-daily (bid), tenofovir 300 mg daily and ritonavir/lopinavir 3 capsules bid. His CD4 count at the time was 200 cells/μL. His renal function at the time of the switch was normal, with a serum creatinine level of 64 μmol/L. For the next 2 years, the patient's HIV RNA was maintained below 200 HIV-1 RNA copies/mL and his CD4 count at 200 cells/μL. In December 2002, imidapril was added to his medication as treatment for lipodystrophy as part of the Ark-503 trial. In January 2003 he was found to have both borderline impaired glucose tolerance and hypogonadotrophic hypogonadism. Anabolic steroids and then transdermal testosterone replacement were added. A subsequent magnetic resonance imaging (MRI) scan (February 2004) showed a 4.5-mm pituitary adenoma. Screening urinalysis was performed 17, 21 and 23 months after commencement of tenofovir (but not at baseline), and proteinuria and glycosuria were demonstrated at all checks.

At month 16 of treatment with tenofovir, the patient first complained of leg pains. The left heel component of this pain was so severe that he saw a podiatrist. In month 23, ibuprofen was given, followed by rofecoxib for 1 week at month 25.

Results of biochemical profiles are recorded in Fig. 1. After 1 week of rofecoxib, impairment in renal function was observed, with a serum creatinine level of 169 μmol/L, a low normal serum phosphate level of 0.82 mmol/L, which continued to fall to 0.43 mmol/L, and a low normal serum potassium level. Serum alkaline phosphatase was raised at 358 IU/L. Serum calcium and albumin and serum alanine aminotransferase levels were normal throughout. Glycosylated haemoglobin (HbA1c) was normal and mean blood glucose was 6 mmol/L. Ultrasound of the renal tract was normal. X-ray of the lumbar spine showed changes consistent with osteopaenia with a t-value for bone mineral densitometry of −1.4 and a z-value of −1.1. An isotope bone scan performed 5 months after discontinuination of the tenofovir-containing regimen showed multiple areas of increased focal activity in the ribs, consistent with pseudofractures secondary to osteomalacia; there was no prior history of trauma.

Figure 1.

 Biochemical parameters for case 1. The vertical line shows the time at which tenofovir was stopped.

All nucleoside reverse transcriptase inhibitors (NRTIs) and nonnucleoside reverse transcriptase inhibitors (NNRTIs) were discontinued at month 25 and ritonavir-boosted lopinavir continued for 1 week until a new antiretroviral regimen was in place. Phosphate supplements were prescribed for 2 weeks. Over 8 weeks, bone pains reduced, no further analgesia or mineral supplements were required, and serum alkaline phosphatase and creatinine normalized. Proteinuria and glycosuria disappeared and all bony pains settled over the succeeding 8 months. A new antiretroviral regimen is now in place with no further problems.

Case 2

A 56-year-old Zambian female was diagnosed with AIDS in 2000. Past medical history included essential hypertension for 25 years, well-controlled on atenolol. Antiretroviral treatment had been given since July 2002, when she reached the UK. She was switched to her second regimen of tenofovir 300 mg daily, abacavir 300 mg bid, and ritonavir/lopinavir 3 capsules bid in February 2003 when her CD4 count was 200 cells/μL. Other medication included prophylactic cotrimoxazole. Baseline renal function was normal with a creatinine level of 88 μmol/L and a negative screening urinalysis. During treatment with this regimen, over the succeeding 11 months, HIV RNA was reduced from 2680 to 50 copies/mL; CD4 was maintained at 210 cells/μL.

Six months after commencement of tenofovir, renal function became mildly impaired (see Fig. 2). At 8 months of treatment the patient complained of right knee swelling and pain in both ankle joints. Swelling of the joints could not be confirmed on examination, but 14 days of treatment with ibuprofen was given. After 11 months of the tenofovir-containing regime, the patient had such pain and muscle weakness in her legs that a proximal myopathy was diagnosed and she was confined to a wheelchair. Loin pain was also reported for which no cause was found. Diclofenac was commenced, but after 1 week oliguria was reported and serum creatinine was found to be markedly elevated at 350 μmol/L. Both serum phosphate and serum bicarbonate levels were reduced and reached their nadir (0.47 mmol/L for phosphate) after all treatment was stopped, as in case 1 (see Fig. 2). Alkaline phosphatase was raised at 310 IU/L, but serum calcium and potassium and serum transaminases remained normal. Proteinuria was present at 2.59 g/24 h. A 24-h urinary phosphate excretion was performed 6 weeks after ceasing tenofovir treatment and was reduced at 2.4 mmol/24 h, reflecting an appropriate renal response to a low serum phosphate level, but was not performed at the time of the tenofovir usage. Serum 25- and 1,25-hydroxycholecalciferol were within the normal range. An ultrasound of kidneys and autoimmune and myeloma screens were normal.

Figure 2.

 Biochemical parameters for case 2. The vertical line shows the time at which tenofovir was stopped.

Treatment was with intravenous fluids and intravenous bicarbonate. Diclofenac, cotrimoxazole and antiretroviral therapy were discontinued and treatment with phosphate and 1-α-hydroxycholecalciferol given. After 2 weeks, serum phosphate was normal, alkaline phosphatase had reduced to 170 IU/L and supplements were discontinued. Myopathy and bone pains resolved over the succeeding 3 months. By 7 months after stopping the tenofovir-containing regimen, her 24-h urinary protein was normal at 0.17 g/24 h. She now remains well on a new antiretroviral regimen.

Discussion

The above-described patients both developed acute renal failure whilst on tenofovir. In case 2, the patient had an additional risk factor for renal disease, namely controlled essential hypertension. Although diabetes mellitus is a common cause of glomerular disease [18], impaired glucose tolerance such as in case 1 is not. Both patients took non-steroidal anti-inflammatory drugs (NSAIDs) only when bone pain was maximal. Rofecoxib [19], ibuprofen and diclofenac [20] have been linked with renal failure; commonly the mechanism is a reduction in glomerular blood flow [21] and more rarely an interstitial nephritis [22]; the latter could theoretically cause a tubulopathy.

Tenofovir is known to cause a renal tubulopathy [5,6] and, although not showing significantly different rates of renal impairment to stavudine in one study [23], has been linked to a reduced creatinine clearance in two others [24,25], a glomerular effect. Both the timing of the onset of the renal impairment (creatinine had doubled in both cases before any NSAIDs were given) and the presence of hypophosphataemia, which has never been reported with NSAID-induced interstitial nephritis, point to tenofovir being the cause. Moreover, both patients reported here were taking ritonavir/lopinavir concurrently with tenofovir. Ritonavir use has been shown to increase the concentrations of tenofovir by 30–40% [26,27] and has been anecdotally linked to tenofovir-related renal tubular disease [5]. NSAIDs appear to have simply resulted in a further deterioration in renal function in both cases, as would be expected in incipient renal failure [21,28]. It is of note that the product monograph for tenofovir advises caution with a list of nephrotoxic agents but not specifically with NSAIDs [26].

These patients were suspected of having hypophosphataemic osteomalacia because of normal vitamin D levels, a lowered serum phosphate level and a rising alkaline phosphatase level, associated with bone pain. Bone biopsy, which would have been necessary to confirm the diagnosis, was declined by both patients.

The bone scan appearances in case 1 are typical of the Looser's zones (pseudofractures) of osteomalacia and may often persist or worsen on isotope scanning when treatment is given [29].

Although the patient described in case 1 had concurrent pituitary adenoma and previous Kaposi sarcoma, the symptoms and the abnormalities of phosphate and alkaline phosphate levels were temporally related to tenofovir usage, and fully resolved, so that further mineral supplementation was not needed, on withdrawal of the drug. Moreover, pituitary adenoma has never been associated with hypophosphataemia despite its high incidence in humans [30]. Osteoporosis related to hypogonadism [31] might be a cause of a reduced bone density, but also seems unlikely because of the phosphate and alkaline phosphatase abnormalities. Moreover, the effects of anabolic steroids on bone mineral density in HIV may vary [32]. Biochemical markers of osteomalacia predated NSAID use.

The patient described in case 2 had been menopausal for 8 years and, whilst there are links between osteoporosis and the menopause [33], and between older age and osteomalacia [34], her menopausal state, even if contributory, was associated with a normal alkaline phosphatase and she was asymptomatic until tenofovir was added to her treatment.

In both patients, glycosuria, proteinuria and rising serum creatinine levels were evident prior to the clinical features of osteomalacia. A lowered serum phosphate level, however, was not present until the osteomalacia had developed and would seem to be a less effective screening tool for those at risk of tenofovir-related bone abnormalities, although screening for it is routinely recommended in tenofovir treatment [26]. Bone mineral densitometry studies in patients on tenofovir may be helpful, but are likely to be impractical.

Mineral supplementation and cessation of tenofovir appear to be the treatment of choice, and so far neither patient has had further hypophosphataemia off treatment and both are no longer dependent on mineral supplements, unlike those with other causes of osteomalacia who often suffer relapses when supplements are stopped, even when a causal tumour has been removed [35]. It should be stated, however, that further study is needed in the area of treatment of this condition.

Patients who have underlying renal abnormality, or who show features of the Fanconi syndrome during follow up, should discontinue tenofovir. Should patients develop bone pain or a myopathy on tenofovir treatment, hypophosphataemia should be looked for and treated and the drug stopped. Myopathy and osteomalacia are reversed with cessation of treatment. Nonsteroidal medication compounds the renal impairment and should be avoided.

Acknowledgements

We thank Dr M. Hammer and his team at the Department of Clinical Chemistry, North Manchester General Hospital for laboratory assistance.

Ancillary