Joint Bone Spine 77 (2010) 67–69

Case report

Vertebral osteitis adjacent to kyphoplasty

Daniel Wendlinga,∗, Michel Rungeb, Eric Toussirota, Ewa Bertolinia, Clément Prati a

a Service de rhumatologie, CHU Minjoz, 25030 Besancon, Franceb Service de radiologie B, CHU de Besancon, Besancon, France

a r t i c l e i n f o

Article history:Accepted 14 April 2009

Keywords:VertebroplastyKyphoplastySide effectsOsteitisVertebral

a b s t r a c t

Vertebroplasty and vertebral kyphoplasty are increasingly performed to treat vertebral fractures, mostnotably those related to osteoporosis. Adverse effects are uncommon and consist chiefly of cement leak-age out of the vertebral body and of vertebral fractures adjacent to the treatment site. We report twocases of vertebral osteitis adjacent to vertebroplasty sites, in a 60-year-old woman and a 79-year-oldman. Kyphoplasty to treat an osteoporotic vertebral fracture was followed by acute pain with an inflam-matory time pattern and laboratory evidence of inflammation. Time to symptom onset was 10 days and45 days, respectively. Magnetic resonance imaging showed changes consistent with inflammation in anadjacent vertebra (low signal on T1 images, gadolinium enhancement, and high signal on T2 images). Abiopsy of the lesion disclosed moderate nonspecific inflammation, with no microorganisms or evidence

of malignancy. Both patients recovered slowly. The male patient experienced a fracture at the site of the

is adjertebncais

t[ott

1

1

mtfrcotpHAer

1d

lesion. Few cases of osteitmay involve changes in v

© 2009 Société fra

Vertebroplasty and kyphoplasty are being increasingly used toreat vertebral fractures, most notably those related to osteoporosis1,2]. Adverse events are rare and consist chiefly of cement leakageutside the vertebral body and of vertebral fractures adjacent to thereatment site [2]. We report two cases of vertebral osteitis adjacento a kyphoplasty site.

. Case reports

.1. Case #1

This patient was a 60-year-old woman with an unremarkableedical history. In August 2006, a fall from the standing position led

o an osteoporotic fracture of T8. At the time, findings were normalrom the standard blood tests (erythrocyte sedimentation rate, C-eactive protein [CRP], serum protein electrophoresis, and levels ofalcium and phosphate). In January 2007, the fracture of T8 was thenly abnormality evidenced by MRI of the spine; the adjacent ver-ebras were normal. She continued to experience activity-limitingain and was therefore treated with kyphoplasty 4 months later.

istology showed a fracture site with no evidence of neoplasia. Inpril 2007, 6 weeks after the kyphoplasty procedure, she startedxperiencing back pain with an inflammatory time pattern. Labo-atory tests showed inflammation (erythrocyte sedimentation rate,

∗ Corresponding author. Fax: +33 381 668 686.E-mail address: dwendling@chu-besancon.fr (D. Wendling).

297-319X/$ – see front matter © 2009 Société francaise de rhumatologie. Published by Eoi:10.1016/j.jbspin.2009.11.004

acent to kyphoplasty have been reported. The underlying pathophysiologyral loading and cement leakage into the intervertebral disk.e de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.

78 mm/h; and CRP, 90 mg/L). In July 2007, 3 months after pain onset,MRI of the spine showed inflammatory changes in T9 (Fig. 1) withlow signal on T1 images, marked gadolinium enhancement, andhigh signal on T2 images from the entire vertebral body, disk, andepidural space. Findings by computed tomography (CT) consisted ofleakage of the cement into the suprajacent disk, intradiscal vacuumphenomenon, loss of height of the T8–T9 disk, and an erosion in theupper endplate of T9 (Fig. 2). Discitis was suspected and a biopsy ofthe T8–T9 disk and T9 body was therefore performed. The only find-ing was chronic nonspecific inflammation: the marrow spaces con-tained loose fibrotic tissue supplied by numerous capillaries and byarterioles; and there was a mild-to-moderate polymorphic inflam-matory infiltrate composed of lymphocytes, histiocytes, and fairlynumerous plasma cells. Siderophages were identified within theinfiltrate. No evidence of neoplasia or extraneous particulate mate-rial was found. No bacteria were recovered in 15-day cultures onaerobic media or anaerobic blood agar, and neither did 15-day cul-tures on Sabouraud medium show any fungal growth. Bisphospho-nate therapy was started. The clinical symptoms abated slowly andthe laboratory tests returned to normal. A follow-up MRI scan per-formed 2 months later showed low signal from the upper part of theT9 vertebral body on T1- and T2-weighted images with resolutionof the epidural abnormalities. The T9 vertebra was not fractured.

1.2. Case #2

This 79-year-old man had a history of osteoporosis diagnosedin 2001 and managed with intermittent bisphosphonate therapy.He had fractures of the L2 and L3 vertebras. In August 2007,

lsevier Masson SAS. All rights reserved.

68 D. Wendling et al. / Joint Bone Spine 77 (2010) 67–69

Fig. 1. Case #1: magnetic resonance imaging of the spine, lateral view, T1-weightedsequence after gadolinium injection. Note the high signal with overall gadoliniumenhancement of T9 immediately under the kyphoplasty site (T8), with involvementof the disk and epidural space.

Fig. 2. Case #1: computed tomography of the spine, coronal reconstruction. Leakageof cement into the suprajacent disk with vacuum phenomenon, loss of height of theinfrajacent disk, and erosion of the superior endplate of T9 (the vertebra infrajacentto the kyphoplasty site).

Fig. 3. Case #2: magnetic resonance imaging of the spine, lateral view, T2-weightedsequence. Note the high signal from T12 above the kyphoplasty site.

evaluation for worsening pain in the back and low back led tothe diagnosis of a fracture of T12. He had persistent incapaci-tating pain despite symptomatic treatment, and T12 kyphoplastywas therefore performed in October 2007. The preoperative MRIscan showed no abnormalities of T11 or L1. After some improve-ment over the first 10 days, he started to experience worseningpain with nocturnal exacerbations and laboratory evidence ofinflammation (CRP, 58 mg/L). An MRI scan visualized changes inthe anterior part of T11 (low signal on T1-weighted images andhigh signal on T2-weighted images) (Fig. 3), with no alterationsof the T11–T12 disk. The images were suggestive of anterosu-perior cement leakage from T12. A core needle biopsy of T11done in November showed nonspecific inflammation (a few frag-mented bone trabeculae edged with recent collagen fibrosis, witha meager inflammatory infiltrate and no evidence of malignancy).Findings were negative from 15-day cultures on aerobic mediaor anaerobic blood agar for bacteria and from 15-day cultureson Sabouraud medium for fungi. The clinical symptoms improvedsomewhat and the laboratory tests returned to normal. The bis-phosphonate treatment was restarted. An MRI scan performed 3months later showed improvements in the signal abnormalitiesfrom T11 and a wedge-shaped anterior crush fracture of the bodyof T11. A follow-up evaluation 9 months after the biopsy foundno further changes in the clinical manifestations or imaging studyfindings.

2. Discussion

Both our patients experienced pain with laboratory evidence

of inflammation shortly after undergoing kyphoplasty to treat anosteoporotic vertebral fracture. MRI showed signal changes consis-tent with inflammation of a vertebra adjacent to the kyphoplastysite, and the histological examination found evidence of nonspe-cific inflammation. These features are consistent with vertebral

t Bone

“i

qphfywop

vkkra[ypppttctittoh

ibr1aia

mtiwdod

[

[

[12] Urrutia J, Bono CM, Mery P, Rojas C. Early histologic changes following poly-

D. Wendling et al. / Join

osteitis”. In both patients, the abnormalities improved slowly andncompletely.

Vertebroplasty and kyphoplasty usually improve pain, function,uality of life, kyphotic angle, and vertebral height; kyphoplasty isarticularly effective on the last two variables [1,2]. Kyphoplastyas a good safety profile. The main complications are vertebral

ractures near the kyphoplasty site (16% of cases within the firstear) [3] and leakage of cement, which is more common in patientsith osteoporosis (7% of cases) [2]. Other complications (about 2%

f cases) include rib fractures, pulmonary cement embolism, com-ression of neurological structures, and spinal infection [2–5].

The vertebral alterations in our patients may reflect edema of theertebral body related to changes in mechanical loading near theyphoplasty site. This edema may presage a vertebral fracture. Thatyphoplasty or vertebroplasty may be associated with an increasedisk of fractures in the adjacent vertebras was suggested many yearsgo. The magnitude of the risk increase remains debated, however6]. Overall, an adjacent vertebral fracture occurs within the firstear in 20% of cases after vertebroplasty [6] and 16% after kypho-lasty [2]. In patients with osteoporosis, the risk associated with therocedure must be distinguished from the risk associated with theresence of prior vertebral fractures (fracture cascade). After ver-ebroplasty, adjacent fractures occur early and are more commonhan fractures of remote vertebras [7]. Factors associated with adja-ent fractures include the presence of intraosseous clefts within thereated fracture [8] and cement leakage into the disk [6]. Changesn loading of the adjacent vertebra related to increased rigidity ofhe treated vertebra may explain the occurrence of adjacent frac-ures [6]. This mechanism may have caused the fracture of T11 inur case #2 but cannot explain the laboratory test abnormalities oristological findings seen in both patients.

The manifestations in our patients might have suggested annfection. A few cases of local infection after vertebroplasty haveeen reported [9,10]. In a study of 22 patients who underwentevision surgery for new symptoms after vertebroplasty (among523 vertebroplasty patients), eight patients were found to haven infection [5]. In our patients, however, the negative bacteriolog-cal cultures and the histological features of the vertebral biopsiesrgued against an infection.

The cement used for vertebroplasty and kyphoplasty is poly-ethylmethacrylate, which can induce changes in neighbouring

issues. Furthermore, bone necrosis can occur at the bone–cement

nterface [11,12], perhaps promoted by the local heat produced

hen the cement polymerizes, and a foreign body reaction mayevelop [11]. However, this mechanism cannot explain the devel-pment of MRI and histological changes in the adjacent vertebra at aistance from the treatment site, except when the disk is involved

[

Spine 77 (2010) 67–69 69

also. In vitro, polymethylmethacrylate can accelerate the degen-eration of nucleus pulposus cells, suggesting that cement leakageinto the disk may decrease the flexibility of the disk [13]. In a case-series study of 1000 vertebroplasty procedures, cement leaked intothe disk in 12% of cases [3]. In a study of 641 vertebroplasty proce-dures, cement leakage into the disk was associated with a vertebralfracture in 10% of cases [4] and one patient had imaging findingssimilar to those in our two patients, at the interface with cementthat had leaked into the disk (although neither laboratory tests norhistological data were obtained).

In sum, the pathogenesis of vertebral osteitis complicating ver-tebroplasty or kyphoplasty is probably multifactorial, with cementleakage into the disk, even in tiny amounts, causing disk alterationsthat further modify the pattern of loading through the vulnerableadjacent vertebral body, leading to imaging study findings of edemaand increasing the risk of fracture.

References

[1] Deramond H, Saliou G, Aveillan M, et al. Respective contribution of vertebro-plasty and kyphoplasty to the management of osteoporotic vertebral fractures.Joint Bone Spine 2006;73:610–3.

[2] Bouza C, Lopez T, Magro A, et al. Efficacy and safety of balloon kyphoplastyin the treatment of vertebral compression fractures: a systematic review. EurSpine J 2006;15:1050–67.

[3] Layton KF, Thielen KR, Koch CA, et al. Vertebroplasty, first 1000 levels of asingle center: evaluation of the outcomes and complications. Am J Neuroradiol2007;28:683–9.

[4] Baumann C, Fuchs H, Kiwit J, et al. Complications in percutaneous vertebro-plasty associated with puncture or cement leakage. Cardiovasc Intervent Radiol2007;30:161–8.

[5] Yang SC, Chen WJ, Yu SW, et al. Revision strategies for complications and failureof vertebroplasties. Eur Spine J 2008;17:982–8.

[6] Trout AT, Kallmes DF. Does vertebroplasty cause incident vertebral frac-tures? A review of available data. AJNR Am J Neuroradiol 2006;27:1397–403.

[7] Trout AT, Kallmes DF, Kaufmann TJ. New fractures after vertebroplasty:adjacent fractures occur significantly sooner. ANJR Am J Neuroradiol2006;27:217–23.

[8] Trout AT, Kallmes DF, Lane JI, et al. Subsequent vertebral fractures aftervertebroplasty: association with intraosseous clefts. AJNR Am J Neuroradiol2006;27:1586–91.

[9] Schmid KE, Boszczyk BM, Bierschneider M, et al. Spondylitis following verte-broplasty: a case report. Eur Spine J 2005;14:895–9.

10] Vats HS, McKiernan FE. Infected vertebroplasty: case report and review of lit-erature. Spine 2006;31:E859–62.

11] Huang KY, Yan JJ, Lin RM. Histopathologic findings of retrieved specimensof vertebroplasty with polymethylmethacrylate cement: case control study.Spine 2005;30:E585–8.

methylmethacrylate injection (vertebroplasty) in rabbit lumbar vertebrae.Spine 2008;33:877–82.

13] Lazáry A, Speer G, Varga PP, et al. Effect of vertebroplasty filler materials onviability and gene expression of human nucleus pulposus cells. J Orthop Res2008;26:601–7.