You are on page 1of 3

Pathophysiology/Complications

ORI GI NAL ARTICLE

Alterations to Bone Mineral Composition as an


Early Indication of Osteomyelitis in the Diabetic
Foot
1 2
KAREN A. ESMONDE-WHITE, PHD MICHAEL D. MORRIS, PHD microbiology analysis, bone fragments
2 1
FRANCIS W.L. ESMONDE-WHITE, PHD BLAKE J. ROESSLER, MD
3 were stored in an ESwab Collection and
CRYSTAL M. HOLMES, DPM Transport system (Becton Dickinson,
Sparks, MD) and analyzed through stan-
dard hospital procedure. Fragments for
OBJECTIVEdOsteomyelitis in the diabetic foot is a major risk factor for amputation, but his-topathology were prepared by the
there is a limited understanding of early-stage infection, impeding limb-preserving diagnoses.
UMHS Tissue Procurement Core or the
We hypothesized that bone composition measurements provide insight into the early
pathophysi-ology of diabetic osteomyelitis. AAVA pa-thology laboratory. Only
otherwise-to-be-discarded bone fragments
RESEARCH DESIGN AND METHODSdCompositional analysis by Raman spectros-copy were used for research purposes.
was performed on bone specimens from patients with a clinical diagnosis of osteomyelitis in
the foot requiring surgical intervention as either a biopsy (n = 6) or an amputation (n = 11). Bone fragment preparation
Bone fragments for Raman spectroscopic
RESULTSdAn unexpected result was the discovery of pathological calcium phosphate min-
erals in addition to normal bone mineral. Dicalcium phosphate dihydrate, also called brushite, analysis were transported and stored in
and uncarbonated apatite were found to be exclusively associated with infected bone. gauze soaked with PBS enriched with
protease inhibitor (0.1% volume for vol-
CONCLUSIONSdCompositional measurements provided a unique insight into the patho- ume) and sodium azide (0.005% weight
physiology of osteomyelitis in diabetic foot ulcers. At-patient identification of pathological for volume) to prevent enzymatic or
min-erals by Raman spectroscopy may serve as an early-stage diagnostic approach. bacterial digestion of bone collagen and
stored at 2208C until examination. Most
Diabetes Care 36:3652–3654, 2013 specimens were examined by Raman

O
spectroscopy within 24 h of the biopsy
RESEARCH DESIGN AND or amputation surgery and thawed at
steomyelitis of the diabetic foot, herein
METHODS room temperature immediately before
called diabetic osteomyelitis, is a major
cause of lower-extremity analysis. The average size of the biopsy
3
amputation, yet an understanding of the Clinical study specimens was ,5 mm , and the average
pathophysiology and technologies en- This is an ongoing translational study size of the amputation specimens was .1
3
abling early diagnosis of this serious in- performed at the University of Michigan cm . Raman spectra were collected with
fection are lacking. Clinical and imaging Health System (UMHS) and the Ann microscopy instrumentation adapted for
tests show that whole-tissue properties of Arbor Veterans Affairs (AAVA) Hospital Raman microspectroscopy as described
bone, including hardness and minerali- and has been reviewed and approved by elsewhere (3).
zation, are directly affected by diabetic their respective institutional review
osteomyelitis (1,2). We hypothesized that boards. Bone was obtained from 17 pa-
RESULTSdTable 1 shows the clinical
compositional changes to bone mineral tients with a clinical diagnosis of diabetic
imaging, pathology, microbiology, and
and collagen matrix accompany clinically osteomyelitis requiring surgical interven-
Raman spectroscopy data for all study
observable alterations in bone hardness tion to collect a bone biopsy specimen (n participants. In most cases, multiple clin-
and mineralization. However, no studies to = 6) or to amputate (n = 11). No pa-tients ical imaging modalities (magnetic reso-
our knowledge have reported on the were treated with bone cements. Bone nance imaging, X ray, ultrasound, or bone
chemical composition of bone in diabetic fragments were prepared separately for scan) were used for preoperative
osteomyelitis. The objective of the present microbiological and histopathologi-cal identification of osteomyelitis. Pathology
study was to measure bone composition in analyses. All patients had bone cul-tures data on a range of pathophysiological
diabetic osteomyelitis with the use of performed, and some had additional soft states were reactive, active remodeling,
Raman spectroscopy. tissue and exudates cultured. For necrotic, or osteomyelitic bone. Addi-
tional histopathological findings of acute
ccccccccccccccccccccccccccccccccccccccccccccccccc
1 inflammation or fibrosis were found in a few
From the Department of Internal Medicine, Division of Rheumatology, University of Michigan Medical
2 participants in the amputation group. As
School, Ann Arbor, Michigan; the Department of Chemistry, University of Michigan, Ann Arbor,
3
Michigan; and the Department of Internal Medicine, Division of Metabolism, Endocrinology and expected, bone cultures revealed a mixed
Diabetes, University of Michigan Medical School, Ann Arbor, Michigan. population of gram-positive bacte-ria, with
Corresponding author: Blake J. Roessler, Staphylococcus, Streptococcus, or
roessler@umich.edu. Received 28 February 2013 and Enterococcus as the dominant species.
accepted 24 May 2013. DOI: 10.2337/dc13-0510
© 2013 by the American Diabetes Association. Readers may use this article as long as the work is Raman spectroscopy of the bone frag-ments
properly cited, the use is educational and not for profit, and the work is not altered. See revealed the presence of patholog-ical
http://creativecommons.org/ licenses/by-nc-nd/3.0/ for details. minerals in addition to normal bone

3652 DIABETES CARE, VOLUME 36, NOVEMBER 2013 care.diabetesjournals.org


Esmonde-White and Associates

Table 1dSummary of clinical data (imaging, histopathology, and microbiology) and only found in vivo under chronically
experimental Raman spectroscopy data for bone specimens obtained from biopsy acidic conditions, such as dental calculus,
specimens or amputations urinary stones, and chondrocalcinosis. To
the best of our knowledge, this is the sec-
Anatomic ond report of brushite in mature human
ID location Imaging Pathology Microbiology Raman bone. Brushite was identified by X-ray
B01 Great toe + + Staphylococcus + absorption and infrared spectroscopy in
fibrous dysplasia of the jaw (10). How-
B02 Medial arch + + Staphylococcus aureus + ever, this finding has not been repro-duced
B03 2nd toe +/2 2 Serratia marcescens, Enterococcus faecalis + in other studies, and results from only one
B04 Calcaneus + 2 E. faecalis, Stenotrophomonas maltophilia + patient were reported. Poorly carbonated
B05 + + S. aureus + apatite can be found in wo-ven, or
B06 + 2 Group G streptococcus + immature, bone and is less crys-talline
A01 + 2 Group B streptococcus + than mature bone mineral (11). By
A02 Great toe + 2 E. faecalis 2 contrast, the uncarbonated apatite found in
A03 2nd toe + 2 Escherichia coli, S. aureus 2 infected bone was more crystalline than
A04 2nd toe NA 2 S. aureus 2 immature bone mineral and suggests de-
A05 2nd toe NA + S. aureus 2 position of a pathological mineral.
A06 2nd toe +/2 + Staphylococcus + Normal serum calcium values in all the
A07 Great toe + 2 Staphylococcus 2 participants argue against the possi-bility
A08 3rd toe + 2 Staphylococcus 2 that we were observing brushite and
A09 2nd toe + + S. aureus, Group B streptococcus 2 uncarbonated apatite as a precursor in
A10 BKA + + E. faecalis, Pseudomonas putida 2 normal bone formation or as a nonbone
A11 2nd toe + 2 Candida sp. 2 precipitate resulting from systemic hyper-
Clinical evaluation of study participants also included age, sex, height, weight, disease duration, and calcemia. The likelihood that pathological
history of foot ulcers. Study participants were 41–87 years old. The biopsy cohort comprised two women minerals were formed by an inflammatory
and four men, and the amputation cohort comprised 11 men. In most cases, the affected foot was assessed response, immune response, or excessive
by X ray, magnetic resonance imaging, bone scan, or ultrasound imaging within 1 month of the biopsy or bone remodeling is not supported by our
amputation. If known, the anatomic location of the surgery or biopsy is included. In several cases,
observations and previous studies (12,13).
multiple clinical imaging modalities were used to ascertain the presence of osteomyelitis, and any
diagnostic radiology report is identified with a +. In a few cases, multiple clinical imaging tests did not Thus, we hypothesize that a bacteria bio-film
yield consistent or unambiguous preoperative identification of osteomyelitis. For those cases, the results is responsible for generating the acidic
are reported from the positive test. Inconclusive or ambiguous diagnostic radiology reports are identified environment necessary to form brushite. If
with a +/2. A + value for pathology results was reported only if the histopathological diagnosis was either the localized microenvironment cannot be
acute or chronic osteomyelitis. Positive histopathology reports included evidence of bone remodeling,
inflammation, necrosis, the presence of re-active bone, and osteolysis. As expected, Staphylococcus, adequately buffered, then acidic calcium
Streptococcus, and Enterococcus were the primary bacterial species recovered from bone cultures. phosphate minerals such as uncarbonated
Raman identification of abnormal minerals, either brushite or uncarbonated apatite, are also denoted with apatite and brushite may precipitate onto the
a +. Hypercalcemia and chronic metabolic acidosis were ruled out as a possible cause of pathological bone surface. This mechanism, al-though
mineralization because all participants had normal-to-low serum calcium levels and normal serum
bicarbonate levels. A, amputation; B, biopsy; BKA, below-the-knee am-putation; NA, not available. new in its application to diabetic
osteomyelitis, is the accepted pathway in
microbial degradation of bone postmor-
mineral. Two pathological minerals were and uncarbonated apatite. Composi- tem (14).
identified: brushite and uncarbonated ap- tional changes in bone currently cannot Associating Raman spectroscopy data
atite. A + for Raman spectroscopic results be identified by standard clinical imag- with anatomic location was an issue in the
was reported if brushite or uncarbonated ing or histopathology but are easily measurements and may have had an
apatite was detected. Raman spectra of measured by Raman spectroscopy. This impact on the rate of identifying patho-
control bone specimens were consistent study provides insight into the logical minerals. Biopsy specimens were
pathophys-iology of diabetic 3
with normal bone composition and did not small (,5 mm ) and taken directly from the
show evidence of pathological miner- osteomyelitis and identified a possible wound bed, so there was a greater as-
alization. Storage in enriched PBS did not early-stage marker of clinical disease. sociation between the spectroscopy data
affect induced compositional changes in a Many mechanisms of bone loss in and the anatomic location of the active
control study of healthy bone fragments. osteomyelitis have been proposed in the infection. Thus, we were able to identify
literature (4–6). Even though bacterial bi- pathological minerals in 100% of the bi-
CONCLUSIONSdIn this study, we ofilms are known to form in osteomyeli- opsy specimens. However, the amputated
applied Raman spectroscopy to measur- tis, direct bacterial attack on bone is tissue was large relative to the recovered
ing compositional changes in bone in- believed to be a negligible mechanism (7– fragments. Although we worked closely
fected by osteomyelitis of the diabetic 9). The present results suggest that with the pathology laboratory to obtain
foot. Bone fragments were examined from pathological mineralization accompanies bone specimens near the site of suspected
patients who underwent either sur-gical bacterial infection, providing insight into infection, obtaining precise anatomic in-
biopsy/debridement or amputation. An the pathophysiology of osteomyelitis. The formation was a challenge. This challenge
unexpected finding was Raman spec-tral presence of pathological minerals may was also apparent when we examined the
patterns corresponding to dicalcium also serve as a compositional marker of imaging and histopathology data. The lack
phosphate dihydrate, also called brushite, early-stage bone infection. Brushite is of correlation between imaging and

care.diabetesjournals.org DIABETES CARE, VOLUME 36, NOVEMBER 2013 3653


Osteomyelitis in the diabetic foot

histopathology data in the amputation K.A.E.-W. contributed to the data analysis 4. Chihara S, Segreti J. Osteomyelitis. Dis
cohort underscores the difficulty in iden- and drafted the manuscript. K.A.E.-W. and Mon 2010;56:5–31
tifying osteomyelitis across a large ana- F.W.L.E.-W. contributed to the data collec- 5. Henderson B, Nair SP. Hard labour: bac-
tomical unit, such as a digit or limb. We tion. K.A.E.-W., F.W.L.E.-W., C.M.H., M.D.M., terial infection of the skeleton. Trends
and B.J.R. contributed to the revision of the Microbiol 2003;11:570–577
suspect that incomplete sampling was
manuscript. K.A.E.-W., C.M.H., M.D.M., and 6. Wagner C, Kondella K, Bernschneider T,
primarily responsible for inconsistent B.J.R. contributed to the study design. K.A.E.-W., Heppert V, Wentzensen A, Hänsch GM.
Raman spectroscopic identification of C.M.H., and B.J.R. contributed to the study Post-traumatic osteomyelitis: analysis of
pathological minerals in amputated conduct. K.A.E.-W. and M.D.M. contributed inflammatory cells recruited into the site
bone. Future translational studies will to the data interpretation. K.A.E.-W. and B.J.R. of infection. Shock 2003;20:503–510
address developing enhanced anatomic approved the final version of the manuscript. 7. Gristina AG, Oga M, Webb LX, Hobgood
precision with respect to geographic K.A.E.-W. is the guarantor of this work and, CD. Adherent bacterial colonization in
analysis of dia-betic wounds. as such, had full access to all the data in the the pathogenesis of osteomyelitis. Science
It is intriguing to conceptualize an at- study and takes responsibility for the in- 1985;228:990–993
patient Raman spectroscopic measure- tegrity of the data and the accuracy of the data 8. Brady RA, Leid JG, Calhoun JH, Costerton
analysis. JW, Shirtliff ME. Osteomyelitis and the
ment of pathological mineralization.
Parts of this study were presented in abstract role of biofilms in chronic infection. FEMS
Intraoperative or transcutaneous Raman form at SPIE Photonics West, San Francisco, Immunol Med Microbiol 2008;52:13–22
spectroscopic identification of pathologi- California, 2–7 February 2013. 9. Nair SP, Meghji S, Wilson M, Reddi K,
cal minerals during biopsy or amputation The authors thank Jeff Kozlow and Jill White P, Henderson B. Bacterially in-
surgeries may distinguish bone infections Southwick for clinical collaboration at the duced bone destruction: mechanisms and
from noninfectious bone lesions. Point-of- AAVA Hospital. UMHS Tissue Procurement misconceptions. Infect Immun 1996;64:
care measurements are feasible because Core, UMHS Comprehensive Cancer Center 2371–2380
Raman spectroscopy is amenable to fiber- (grant CA49652), and Steven Chensue and 10. Yamamoto H, Sakae T. Brushite in fibrous
optic–based instrumentation. Our labora- Lynn St. Dennis, VA Pathology, prepared bone dysplasia of the jaw bone. Acta Pathol Jpn
tory has developed portable fiber-optic specimens. The authors also thank Kaiser 1987;37:1699–1705
Optical Systems for instrument support; Erin 11. Uthgenannt BA, Kramer MH, Hwu JA,
instrumentation for transcutaneous bone
Bigelow, University of Michigan Orthopaedic Wopenka B, Silva MJ. Skeletal self-repair:
measurements at bedside or in a surgical Research Laboratory, for preparation of ca- stress fracture healing by rapid formation
suite, and our ongoing human studies daveric bone specimens; and Haiping Sun, and densification of woven bone. J Bone
demonstrate in vivo feasibility and estab- University of Michigan Electron Microbeam Miner Res 2007;22:1548–1556
lish a basis for future translational Raman Analysis Laboratory, for technical assistance. 12. Takahata M, Maher JR, Juneja SC, et al.
studies of diabetic foot wounds (15). Mechanisms of bone fragility in a mouse
model of glucocorticoid-treated rheuma-
References toid arthritis: implications for insufficiency
AcknowledgmentsdK.A.E.-W. acknowledges 1. Devendra D, Farmer K, Bruce G, Hughes P, fracture risk. Arthritis Rheum 2012;64:
the training grant T32-AR-007080 from the Vivian G, Millward BA. Diagnosing osteo- 3649–3659
National Institute of Arthritis and Musculo- myelitis in patients with diabetic neuro- 13. Misof BM, Gamsjaeger S, Cohen A, et al.
skeletal and Skin Diseases and a career de- pathic osteoarthropathy. Diabetes Care Bone material properties in premenopausal
velopment grant from the Michigan Institute of 2001;24:2154–2155 women with idiopathic osteoporosis. J Bone
Clinical and Health Research (UL1RR024986). 2. Lavery LA, Armstrong DG, Peters EJG, Miner Res 2012;27:2551–2561
This work was supported by grants R01-AR- Lipsky BA. Probe-to-bone test for diagnos- 14. Herrmann B, Newesely H. Dekomposi-
055222 and R01-AR-047969 from the National ing diabetic foot osteomyelitis: reliable or tionsvorgänge des Knochens unter langer
Institute of Arthritis and Musculoskeletal and relic? Diabetes Care 2007;30:270–274 Liegezeit. 1. Die mineralishche Phase. An-
Skin Diseases (to M.D.M.) and R21-EB- 3. Dehring KA, Crane NJ, Smukler AR, thropol Anz 1982;40:19–31 [in German]
101026 from the National Institute of McHugh JB, Roessler BJ, Morris MD. 15. Esmonde-White FWL, Morris MD. Validat-
Biomedical Im-aging and Bioengineering (to Identifying chemical changes in sub- ing in vivo Raman spectroscopy of bone
B.J.R.), National Institutes of Health. chondral bone taken from murine knee in human subjects. In Proceedings of SPIE.
No potential conflicts of interest relevant joints using Raman spectroscopy. Appl Kollias N, Ed. San Francisco, CA, SPIE,
to this article were reported. Spectrosc 2006;60:1134–1141 2013, p. 85656K

3654 DIABETES CARE, VOLUME 36, NOVEMBER 2013 care.diabetesjournals.org