Increased cortical bone mineralization in imatinib treated patients with chronic myelogenous leukemia

Imatinib mesylate (Glivec, Gleevec™, Novartis International AG) and the second generation ABL tyrosine kinase inhibitors have markedly improved the outcome of patients with chronic myeloid leukemia (CML). More patients are receiving treatment with these inhibitors for prolonged periods of time and experience with imatinib now exceeds five years. The immediate side effects are usually mild and manageable. There have been recent reports on the long-term side effects associated with prolonged use of imatinib.1–3 Berman and coworkers found that imatinib treated patients had hypophosphatemia, lower osteocalcin levels and higher parathyroid hormone levels.2 Subsequent studies confirmed the observation of hypophosphatemia in patients receiving imatinib.4–8 The authors concluded that imatinib may affect bone remodeling and if left untreated, chronic hypophosphatemia may result in impaired bone mineralization, rickets, and osteomalacia.2

We, therefore, investigated bone mineral density (BMD) in imatinib treated CML patients and healthy controls. All imatinib treated CML patients at Sahlgrenska University Hospital were identified of whom 17 fulfilled the study inclusion criteria: (i) imatinib treatment duration ≥24 months, and (ii) first chronic phase of the disease with complete cytogenetic remission. The inclusion criteria were set to minimize the confounding effect of leukemia and to allow time for a possible bone remodeling effect of imatinib to take place. Sex and aged matched healthy individuals served as controls (Table 1). The imatinib dose was 400 mg per day. The treatment duration was 50±19 months (range 24–73). The study was approved by the local research ethics committee. Details of design and methods are available as an online supplement.

Table 1.Blood counts, serum chemistry and bone related variables in imatinib treated chronic myeloid leukemia patients (n=17), and age and sex matched healthy controls (n=17). The results are given as mean ± SD.

The CML patients had significantly lower levels of calcium, ionized calcium, phosphate and magnesium in serum compared with controls (Table 1). Two out of 17 patients had serum ionized calcium below the reference range (1.18–1.31 mmol/L), 2 out of 17 patients had serum phosphate below the reference range (0.7–1.6 mmol/L) while none of the patients had serum magnesium below the reference range (0.7–0.95 mmol/L). The excretion of calcium in the urine was also found to be significantly lower in the imatinib treated patients (Table 1). The markers of bone turnover, osteocalcin and bone-specific alkaline phosphatase were significantly lower in patients compared with controls, whereas the marker of bone resorption carboxyterminal cross-linked telopeptide of type I collagen was unchanged (Table 1).

Our data confirm and extend those of Berman and coworkers despite differences in the patient and control design.2 Berman et al. reported that imatinib treated CML and gastrointestinal stromal tumor patients with hypophosphatemia had a lower osteocalcin level, higher parathyroid hormone level and were treated with a higher imatinib dose (median 600 mg) than patients with a normal phosphate level. These findings were explained by an imatinib induced inhibition of bone turnover, which in turn, triggered a secondary hyperparathyroidism in an attempt to maintain calcium homeostasis. The authors raised concerns about an imatinib induced osteomalacia and suggested phosphate replacement therapy. In contrast to the hypothesized imatinib induced osteomalacia, we found that our imatinib treated CML patients had significantly higher areal BMD of the lumbar spine (+12%) and total hip bone (+12%) compared with controls (Table 2). Even compared with the standard reference, the patients had higher BMD than could be expected for age and gender. The patients’ T-score value was 0.07 and 0.38 for the total hip bone and lumbar spine respectively (Table 2). The patients’ Z-score value was 0.46 and 0.57 for total hip bone and lumbar spine (L1–4) respectively. A significant relationship was also observed between the imatinib treatment duration and areal BMD of the lumbar spine (L1–L4) (Spearman’s ρ=0.49, p=0.046). Similarly, for both radius and tibia the imatinib treated CML patients had significantly higher cortical volumetric BMDs than the controls. No difference was seen in trabecular volumetric BMD of either radius or tibia. The cortical cross-sectional area was also similar in the two groups (Table 2).

Table 2.BMD measurements in imatinib treated chronic myeloid leukemia patients (n=17) and controls (n=17).

Our results are also in line with previous in vitro studies and animal models. Osteoblasts are known to derive from mesenchymal stem cells, while osteoclasts are multinucleated giant cells of hematopoietic origin in the monocytic lineage. Both are targeted by imatinib and it has been shown that imatinib promotes osteoblast differentiation and inhibits osteoclastogenesis, presumably through its action on the colony-stimulating factor 1 receptor (C-FMS) and platelet derived growth factor receptor (PDGFR).9,10

Overall, biochemical findings suggest a suppressed bone turnover and the increased BMD indicates that bone formation and resorption are affected unequally. We suggest that imatinib uncouples bone formation from bone resorption in favor of the former, disturbing bone homeostasis and leading to a net increase in bone mineral density.

This is further supported by a recent study by Fitteri et al.,11 who reported an increased trabecular bone volume in imatinib treated patients. In contrast to our study, they did not analyze the bone mineralization, but measured the trabecular bone volume on decalcified iliac crest biopsies obtained before and during imatinib treatment. Therefore, bone mineralization, the most important measure of osteopenia, was not addressed.

The present report is the first demonstrating altered calcium and phosphate metabolism in imatinib-treated CML patients, concomitant with an increased cortical bone mineralization. Our study alleviates previous concerns about an accelerated osteomalacia, and measures to prevent this hypothesized long-term side effect seem unnecessary and might even be harmful.2,12

If imatinib is shown to suppress bone resorption without decreasing the bone quality, tyrosine kinase inhibitors could be novel antiosteolytic agents in skeletal disorders.

Footnotes

• Funding: the study was supported by grants from the Swedish Research Council (project K2006-71X-11630-07B and project 529-2004-6512), FoU Västra Götaland, the Swedish Society for Medical Research, “JK-foundation” Sahlgrenska University Hospital, “Volvo Assar Gabrielssons foundation”, the Foundations of the National Board of Health and Welfare, the Åke Wiberg Foundation, the Jeansson Foundation, the Tore Nilsson Foundation for Medical Research and the Magnus Bergvall Foundation.