Late-onset variant fibrodysplasia ossificans progressiva leading to misdiagnosis of ankylosing spondylitis^†

To the Editor:

Fibrodysplasia ossificans progressiva (FOP) is a rare (1 per 2 million) condition characterized by extra-skeletal, heterotopic bone formation in the soft tissues, leading to severe disability [Shore and Kaplan, 2008]. Patients classically have congenital malformation of the great toes (hallux valgus, malformed first metatarsals, and monophalangism) [Kaplan et al., 2009] and subsequently present in childhood with extra-skeletal bone formation involving muscle, tendon, ligaments, fascia, and aponeuroses [Cohen et al., 1993; Mahboubi et al., 2001]. Soft tissue non-inflammatory fibroproliferative masses often precede the formation of the osseous masses [Cohen et al., 1993]. The soft tissue masses usually appear in predictable locations, often commencing in the upper back and neck. Soft tissue injury at any stage can result in heterotopic bone formation at the site of injury [Shore and Kaplan, 2008].

The two defining features of classical FOP are the presence of congenital malformations of the great toes and progressive heterotopic ossification in characteristic anatomical patterns [Shore et al., 2006]. Atypical FOP patients have been described and have been subdivided into two classes, FOP-plus and FOP variants [Kaplan et al., 2009]. FOP-plus patients have the two defining features of classical FOP, but in addition have involvement of other systems that can include brain, eye, and bone marrow abnormalities [Kaplan et al., 2009]. Patients with variant FOP have major variations in the characteristics or severity of one or both of the two classic defining features of classical FOP.

Mutations in the ACVR1 gene, a gene encoding a type 1 bone morphometric protein transmembrane receptor, cause FOP [Shore et al., 2006]. A single recurrent heterozygous point mutation in the glycine-serine rich (GS) domain of exon 4 of ACVR1 (c.617G→A; p.R206H) is found in cases of classic FOP and most cases of FOP-plus. Other mutations have been associated with FOP variants and some cases of FOP-plus.

In this report, we describe a woman with variant FOP, characterized by normal great toes and late onset heterotopic ossification, who was misdiagnosed with ankylosing spondylitis for several years and treated intermittently with sulfasalazine and methotrexate.

The patient presented at 21 years of age with a “stiff back”. A plain X-ray of the lumbar spine was normal. She was referred to a rheumatologist who diagnosed non-specific spondyloarthropathy and sulfasalazine treatment was commenced. Treatment was continued for 12 months then discontinued by the patient because of a lack of improvement in symptoms.

Three years later, the patient presented with a painless mass on the inferior border of the mandible in the midline. Imaging with ultrasound revealed a heterogeneous mass measuring 4.3 cm × 3.6 cm, 1 cm below the skin. A fine needle aspirate was performed which demonstrated skeletal muscle with proliferative fibrous stroma suggestive of a reactive or inflammatory process. The mass resolved completely over 3 months with no remaining clinical or radiological evidence of the lesion.

During this time, the patient's lower back pain became severe and following further imaging including plain vertebral X-rays and a bone scan, a diagnosis of ankylosing spondylitis was made (Fig. 1). C-reactive protein, full blood count, electrolytes, calcium, phosphate, alkaline phosphatase, parathormone, anti double-stranded DNA, extractable nuclear antigen, anti-nuclear antibodies, rheumatoid factor, anti-myeloperoxidase, anti-proteinase 3 and monoclonal protein electrophoresis were done and were all normal. HLA-B27 was not detected. ANCA immunofluorescence was positive, suggesting an inflammatory process. The patient was re-started on sulfasalazine and methotrexate was added. No improvement occurred and the patient ceased the medication herself after several months. Following this, there were several episodes of pain centered on the right scapular region and an MRI of this region revealed changes consistent with myositis involving the rhomboid major, serratus anterior, and intercostal muscles on the right. This again resolved spontaneously over several months.

Three years later, the patient developed a painless mass in the region of the thyroid gland. The mass was non-ossified and measured 4.5 cm × 2.2 cm × 6.0 cm and extended from the hyoid bone to the lower pole of the thyroid. A fine needle aspirate was performed which showed inflammatory change. A plain X-ray and CT scan of the entire spine was performed at the same time, which revealed extensive ossification of the ligamentum flavum in the thoracic spine and interspinous ligament ossification in the lumbar spine (Figs. 2 and 3). Central canal narrowing due to ligamentum flavum ossification was noted between T4-6 and at T11. In addition, a thin layer of ossification was seen between the right trapezius and rhomboid, consistent with ossification of the previous inflammatory lesion. The diagnosis of FOP was suggested on the clinical and radiological findings and the patient was referred to the Genetics Clinic.

Examination at this time revealed significant limitation of neck and lumbar spine movement, particularly flexion and extension. Neurological examination was normal. Examination of the soft tissues of the back revealed two small, firm masses, one over the right scapula and one over the right lower lumbar spine, both consistent with ossified masses. Neither had been symptomatic at any stage. Clinical and radiological examination of the feet was normal (Fig. 4).

Sequencing of exon 4 of the ACVR1 gene was performed to confirm the clinical suspicion of FOP. The recurrent c.617G→A mutation seen in classical FOP and in some cases of variant FOP was not found, but, sequencing revealed heterozygosity for an adjacent mutation (c.605G→T), also in the GS domain. This mutation has been described only once before in a patient with similarly mild FOP but with the characteristically abnormal great toe on one side [Petrie et al., 2009]. Another patient with an ACVR1 mutation involving the GS domain (c.590_592delCTT; P197_F198delinsL) has been described with normal great toes and a later presentation (11 years of age) but within 6 months had ankylosis of all joints of the axial and appendicular skeleton [Patient 20; Kaplan et al., 2009].

Our patient represents an unusually late presentation of FOP together with normal great toes on clinical and radiological examination. Bilaterally unaffected great toes are rare in FOP, to the best of our knowledge having only been described in four previous patients [Kaplan et al., 2009]. Another atypical feature of this case is the natural history of the soft tissue masses. The neck lesions were painless and not associated with symptoms of inflammation. Although the biopsy results demonstrated fibro-proliferative change, the rapid production of heterotopic bone formation that might be expected to occur following this type of trauma (biopsy) did not occur [Lanchoney et al., 1995; Luchetti et al., 1996]. Complete resolution of the lesions after several months, without evidence of ossification, is also unusual. Ossification has been seen as soon as 6 weeks after inflammation or trauma [Gannon et al., 1998] but in this patient both neck lesions remain unossified, in one case 6 years after its appearance and biopsy. Only the areas over the right scapula and lower lumbar spine have demonstrable evidence of ossification on imaging. The radiological appearance of the thoraco-lumbar spine was eventually suggestive of FOP (Fig. 3).

This is an important case because it highlights the diagnostic difficulties associated with late onset, milder FOP, especially if the characteristic toe abnormalities are absent. Great toe abnormalities would normally draw attention to the diagnosis, but in this case, the presence of normal great toes contributed to the incorrect diagnosis of ankylosing spondylitis and treatment with sulfasalazine and methotrexate. Variant FOP caused by non-classical ACVR1 mutations should be considered in the differential diagnosis of ankylosing spondylitis, especially if resistant to treatment.