In June 2003, a 50-year-old white woman who was 6 months postmenopausal and had a 19-year history of systemic lupus erythematosus treated chronically with prednisone sought a second opinion for back pain at the Columbia Presbyterian Medical Center. Her pain had begun acutely in December 2002 while shoveling snow. Although a DXA in October 2001 had revealed osteopenia (L-spine T-score = −1.99, total hip T-score = −1.65), she had previously been free of back pain or fractures. Intranasal calcitonin was prescribed by her primary care physician for presumptive fracture-associated pain in early January 2003, but her pain persisted, prompting an MRI in March 2003 that revealed a subacute partial compression fracture of the L2 vertebral body involving the superior endplate (Fig. 1A); no other fractures were identified (T11 was not fully visualized on this image because it was primarily a lumbar spine MRI; therefore, a minor T11 fracture cannot be definitively excluded). Her orthopedist recommended kyphoplasty of L2, which was performed in April 2003. Fluoroscopy during the procedure identified the single L2 fracture, which was reduced using a Kyphon balloon and injected with 3 ml of bone cement. The patient tolerated the procedure well and was discharged the following day with dramatic pain improvement. Forty-eight hours later, however, without any increase in activity level, her low back pain recurred and became incapacitating, confining her to bed. Six days later, an MRI was performed to assess the recurrence of severe pain and revealed stable kyphoplasty changes within L2 and interval development of several new vertebral fractures, including two acute, severe compression fractures of T11 and T12, with marked anterior wedging (although T11 was not fully visualized on the initial MRI, the follow-up clearly shows a dramatic change), as well as fractures of the L1 inferior endplate, L3 and L4 superior endplates, and the L5 inferior endplate (Fig. 1B). Calcitonin was discontinued and teriparatide was started in May 2003, with a significant reduction in pain over the next month.
When we first evaluated her in June 2003, examination revealed a Cushingoid-appearing woman with mild kyphosis and some tenderness to palpation in both the lower thoracic and lumbar spine. Renal, hepatic, hematopoietic, and thyroid function tests, as well as serum calcium, parathyroid hormone (PTH), N-telopeptide, and vitamin D levels, were within normal limits. Bone-specific alkaline phosphatase was elevated at 37.8 μg/liter (normal, 6.4-24.4 μg/liter), consistent with a healing fracture or possibly early teriparatide effect. The erythrocyte sedimentation rate (ESR) was 4 mm/h. DXA indicated L3 and total hip T-scores of −2.88 and −1.56, respectively.
The temporal relationship between the kyphoplasty procedure that eliminated her pain and the subsequent recurrence of pain with documented fractures of six adjacent vertebrae on the follow-up MRI is highly suggestive of causality. It seems unlikely that manipulation during the surgical procedure itself produced these fractures or that these new fractures occurred spontaneously some time after the March MRI and before the April kyphoplasty; the patient's pain disappeared immediately after the single L2 kyphoplasty, and the postprocedure fluoroscopy revealed no acute fractures. Although new endplate fractures might be undetected by fluoroscopy, the marked changes at T11 and T12 would be difficult to miss.
Traditionally, acute vertebral fracture has been medically managed, but percutaneous vertebroplasty and kyphoplasty are becoming increasingly frequent treatments. Percutaneous vertebroplasty refers to the injection of bone cement into a fractured vertebra using fluoroscopic guidance, with the goal of providing pain relief. Kyphoplasty includes inflation of a bone tamp within the vertebral body before cement injection, thereby slightly re-expanding the fractured vertebrae. Kyphoplasty offers the potential advantage of low pressure cement injection and fewer cement leaks, as well as improvement in vertebral height deformity and kyphosis.(1) Despite the increasing popularity of these two procedures, there have been no randomized controlled trials completed to evaluate the efficacy or safety of either procedure. Although prospective kyphoplasty case series suggest a rapid and lasting reduction in pain postprocedure, control groups are lacking.(2–4) In the only nonrandomized, controlled study of either procedure, there was no significant difference in pain scores between vertebroplasty and medical management at 6 weeks or 1 year.(5) No kyphoplasty study has formally reported the incidence of postprocedure vertebral fractures. Although it is clear that any prevalent vertebral fracture increases the risk of subsequent vertebral fracture,(6) there is some evidence that vertebroplasty may further elevate this already high risk in vertebrae adjacent to those injected. Grados et al.(7) reported an odds ratio of 2.27 for fracture in adjacent vertebra compared with 1.44 for nonadjacent vertebrae. Although the mechanism by which future fracture might occur postprocedure is currently unclear, it is postulated that vertebroplasty and kyphoplasty might shift the normal load transmission through the spine, predisposing it to fracture. One in vitro study by Berlemann et al.(8) reported that the failure load for cadaveric two-vertebrae spine units after cement injection of a single vertebra within the unit was lower than for untreated vertebral units. To our knowledge, similar studies have not been performed using kyphoplasty, and they may be helpful in delineating the possible pathogenesis of fracture after this procedure.
We report the case of a recently postmenopausal patient with steroid-induced osteoporosis who sustained multiple adjacent vertebral fractures in the days immediately after an L2 kyphoplasty. Although this temporal relationship is not proof of causation, it strongly suggests this possibility. If the adjacent fractures were a complication of kyphoplasty, the additional risk of subsequent fracture might be limited to those with steroid-induced osteoporosis. This case emphasizes the critical need for randomized controlled studies of vertebroplasty and kyphoplasty in patients with osteoporotic fractures.