Bisphosphonate induced osteonecrosis of the jaws: Unravelling uncertainty in disease causality

To the Editor

The phenomenon of osteonecrosis of the jaws secondary to bisphosphonate therapy has been characterised by increased clinical recognition of the disorder, an evolving literature documenting its features and management but a prevailing uncertainty as to the causality of this unexpected side effect of bisphosphonate therapy. A report concerning jaw complications secondary to zoledronic acid in breast and prostate cancer patients has previously been published in Acta Oncologica [1]. Proposed causes have included chemotherapy, ionizing radiation, chronic osteomyelitis, inhibition of angiogenesis [2], microtrauma [3], failure of mineralization of newly formed osteoid [4] and impaired wound healing secondary to inhibition of matrix metalloproteinases [5], [6]. This letter concerns the clinical, cytological and clinical chemistry similarities between maxillomandibular osteonecrosis secondary to bisphosphonates and the hereditary condition osteopetrosis. Most described case series involve the administration of the nitrogen containing bisphosphonates, palmidronate and zoledronic acid. These medications inhibit fanesyly diphosphonate synthase a critical enzyme in the biosynthetic mevalonate pathway. They also cause apoptosis of osteoclasts and induce osteoprotegerin, which neutralises RANKL (receptor activator of nuclear factor-KB ligand) a receptor essential for osteoclast formation and activation [7].

The similarity between maxillomandibular osteonecrosis and the metabolic bone disorder osteopetrosis has been observed [8]. Osteopetrosis is a group of heterogeneous disorders characterised by increased bone density. At least nine forms have been described including osteopetrosis with precocious manifestations which usually results in death at an early age and a milder form with delayed manifestations known as osteopetrosis tarda or Albers-Schonberg disease [9]. Molecularly osteopetrosis is caused by cellular mutations in osteoclasts that cause failure of acidification of bones. The majority of cases of osteopetrosis are caused by defects in osteoclast vacuolar H⁺ATPase, alteration in an osteoclast specific chloride channel and carbonic anhydrase II dysfunction. Investigations in patients with this disorder reveal an increased creatinine kinase BB isoenzyme, serum acid phosphatase level and dual energy x-ray absorptiometry T and Z scores. Could this rare metabolic bone disorder provide the key to understanding an iatrogenic disease reaching epidemic proportions?

The first source of evidence is clinical and historical. In a retrospective review of the literature, case reports of maxillomandibular osteomyelitis in patients afflicted with osteopetrosis exist [10]. Additionally recalcitrant osteomylelitis of the jaw following tooth extraction has been described in a patient afflicted with malignant osteopetrosis [11]. This is a historical point frequently mentioned by the patient with iatrogenic osteonecrosis of the jaws who find that they have a non-healing wound after dental extraction. Furthermore bisphosphonate induced osteopetrosis has been described in a 12 year old boy [12]. The second hierarchy of evidence is cytological. The molecular alterations of osteopetrosis cause functional failure of the osteoclast ruffled border that normally form Howships laculae in resorbing bone. Osteopetrosis also can cause morphologic alteration in the ruffled border that can be assessed by electron microscopy. Absent ruffled borders have been shown in a case of autosomal recessive osteopetrosis in a 3-month old male infant [13]. A study of nine patients with infantile osteopetrosis showed increased osteoclast number and electron microscopy showed absence or marked diminishment of the ruffled border-clear zone complex in seven of the nine patients [14]. Osteoclasts in osteosclerotic oc/oc mice are devoid of ruffled borders [15]. Similarly osteoclasts that have been treated with bisphosphonates have been studied ultrastructurally. Apoptosis of osteoclasts secondary to bisphosphonates is well recognised. Initially they become devoid of ruffled borders, and later develop features of apoptosis with pyknotic nuclei, condensation of chromatin, degradation of cell organelles and DNA fragmentation [16]. DNA fragmentation in osteoclasts treated with bisphosphonates may be assessed by the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labelling (TUNEL) method. Aledronate studies in rats have shown that it stops membrane ruffling without destroying the osteoclast [17]. The evidence appears that osteopetrosis and bisphosphonate-induced osteopetrosis are disorders of dysfunction and morphologic alteration of the osteoclast-ruffled border. The cellular distinction between these disorders is that apoptosis of osteoclasts occurs with bisphosphonate treatment whereas this is not a feature of osteopetrosis.

The cellular and case lead clinical similarities of these two disorders are compelling but what about establishing causality. A recent publication albeit in abstract format described a patient with osteonecrosis of the mandible secondary to four years of sequential palmidronate and zoledronic acid. Investigations showed Creatinine kinase BB isoenzyme was 21% (upper limit normal 2%), acid phosphatase 4.9 U/l (N. 2.0 – 6.8), DEXA scan showed universally increased T and Z scores up to 191% of predicted values [18]. Creatinine kinase BB isoenzyme may be elevated in malignancies and spuriously high DEXA scan T and Z scores can be caused by osteosclerotic metastasis. The normal acid phosphatase, which is usually increased in osteopetrosis is an expected consequence of the effect of bisphosphonates on the relative abundance of osteoclasts compared with osteopetrosis. There are many potential confounders but the results are sufficiently compelling to warrant further investigation.

The mechanism for osteonecrosis secondary to bone thickening is most probably ischaemic. This may occur by extrinsic occlusion of the perforating regional circulation because of thickened bone cortices. There are two theories that deserve consideration. The first is thickening of the bone periostium/peridontium of the maxilla and mandible. In Caffey's disease (infantile cortical hyperostosis) a self limited disorder that causes periosteal inflammation and formation of subperiosteal immature lamellar bone there is a predilection for affecting the mandible, involving the bone in 90% of cases [19]. Seven percent of bone mass is recycled per week however the periostium remains in a dynamic physiological positive balance and is particularly liable to thickening in the presence of bisphosphonates. In immunohistochemistry studies in rats for the osteoclast differentiation factor RANKL, immunoreactivity was localised on spindle shaped mesenchymal cells around blood vessels near the bone surface in the peridontium [20]. Bisphosphonate treatment may cause failure of differentiation of these mesenchymal cells into osteoclasts with a dynamic failure to keep patent the spaces through which the microvaculature courses. Also deposition of subperiosteal bone may impair the nutrition of the underlying bone. A second ischemic theory is that of progressive constriction of the foramina through which the vasculature supplies the jaws. A correlative example in osteopetrosis of foraminal stenosis is optic atrophy secondary to stenosis of the optic foramen. Maxillomandibular osteonecrosis is not caused by large foraminal occlusion but probably by progressive occlusion of the microcirculation of the jaws. This is because the flow in a blood vessel is proportional to the reciprocal of the radius to the fourth power making small blood vessels more liable to potential occlusion. Also osteoclasts are present at the edge of resorption tunnels in Haversian canals implying a risk to the dynamic preservation of patency of Volkmann's canals. Finally it is not a requirement for absolute ischemia to occur. Dental extraction or exposed bone in the oral cavity represents an infected compound fracture that would normally never heal but for the overwhelming abundant regional circulation. A circumstance of comparative regional under perfusion is sufficient to lead to the chronic osteonecrosis problems that affect these patients.

Future investigations to determine the cause of this disorder should include performance of bone biochemistry including formation and resorption bone markers. Electon microscopy to assess absent ruffled borders in osteoclasts derived from bone excised from these patients should be undertaken. Immunohistochemistry studies looking at cathepsin-E that specifically localises at the ruffled border membrane of active osteoclasts may be informative [21]. The vascular injury hypothesis due to narrowing of canals penetrating the jaws may be investigated by performing multi slice computerised tomography of the mandible with orthoradial reconstructions to look for narrowing or absence of lateral lingual canals. These are small channels with a mean diameter 0.6mm±0.2 mm that principally occur in the premolar region [22]. Microscopic inspection of the Haversian systems of removed bone should also be performed. The association between bisphosphonates and the unexpected complication of maxillomandibular osteonecrosisis is now firmly established. An iatrogenic pseudo-osteopetrotic state is an emerging cause. Future publications need to move away from descriptive case series and establish the biochemical, cellular and radiographic reasons why this complication occurs.