Fracture non-union: Can biomarkers predict outcome?

Fracture non-union: Can biomarkers predict outcome?

Injury, Int. J. Care Injured 44 (2013) 1725–1732 Contents lists available at ScienceDirect Injury journal homepage: www.elsevier.com/locate/injury ...

956KB Sizes 3 Downloads 28 Views

Injury, Int. J. Care Injured 44 (2013) 1725–1732

Contents lists available at ScienceDirect

Injury journal homepage: www.elsevier.com/locate/injury

Review

Fracture non-union: Can biomarkers predict outcome?§ I. Pountos a,b, T. Georgouli a, S. Pneumaticos c, P.V. Giannoudis a,b,* a b c

Academic Department of Trauma and Orthopaedic Surgery, School of Medicine, University of Leeds, UK Leeds Biomedical Research Unit, Leeds, UK Academic Department of Trauma & Orthopaedic Surgery, School of Medicine, University of Athens, Athens, Greece

A R T I C L E I N F O

A B S T R A C T

Article history: Accepted 9 September 2013

Delayed bone healing and non-union occurs in approximately 10–15% of long bone fractures. Both pathologies may result in prolonged period of pain, disability and repetitive operative interventions. Despite intense investigations and progress done in understanding the pathophysiologic processes governing bone healing, the diagnostic tools have not been altered. The clinical findings and radiographic features remain the two important landmarks of diagnosing non-union and even when the diagnosis is established there is debate on the ideal timing and mode of intervention. Emerging evidence suggest that there are certain molecules and genes that can serve as predictors of potentially unsuccessful fracture union. This article summarises the current evidence on the available ‘bio-markers’to predict fracture non-union. ß 2013 Elsevier Ltd. All rights reserved.

Keywords: Biomarkers Bone healing Non-union Genetic factors

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . Materials and methods . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genetic predisposition relevance . . Serum molecular mediator activity Discussion . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

Introduction Following a traumatic insult, the musculoskeletal system mounts both a local and a systemic reaction with primary aim being the prompt restoration of normal function [1,2]. This is one of the most complex physiological processes aiming to the repair of the fractured bone without the formation of scar tissue [3,4]. Approximately 5–10% of the 6.2 million fractures occurring annually in the United States are associated with impaired healing

§ No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study. * Corresponding author at: Leeds Biomedical Research Unit, Academic Department of Trauma & Orthopaedics, Leeds General Infirmary, Clarendon Wing Level A, Great George Street, Leeds LS1 3EX, UK. Tel.: +44 113 3922750; fax: +44 113 3923290. E-mail address: [email protected] (P.V. Giannoudis).

0020–1383/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.injury.2013.09.009

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

1725 1726 1726 1726 1727 1729 1730 1730

including delayed union or non-union [5]. Non-union is defined by some authors as the cessation of all reparative healing processes without bone union at 6 months following fracture, while the term delayed union is used when fracture union has not been achieved within an adequate period of time elapsed since the initial injury. According to the US Food and Drug Administration (FDA) requirements non-union is considered to be established when a minimum of nine months have elapsed since injury and the fracture site shows no visibly progressive signs of healing for a minimum of three months [6]. A number of risk factors that contribute to the development of fracture non-union have been identified over the years. These include local and systemic factors that could be related to the fracture personality as a result of the initial insult, the treatment provided or the patient’s specific characteristics and co-morbidities (Fig. 1) [7–25]. Fracture non-union is a dreaded complication with devastating outcomes for the patient. Not infrequently, a complex and

1726

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732

Risk factors for compromised bone healing

• • •

• • •



Patients dependant

Fracture dependant

Age and Sex Nutritional state Comorbidities

• Bone involved • Pattern of fracture • Comminution • Fracture gap • Extend of trauma • Soft tissue damage • Interposed soft tissue • Lack of cortical apposition • Surgical treatment • Mechanical stability • Infection

• • • • •

Diabetes Hypothyroidism Anaemia Osteoporosis Vascular disease

Alcohol abuse Smoking Drugs • • •

Steroids NSAIDs Chemotherapy

Radiation therapy

Table 1 Flow chart diagram of included studies.

Fig. 1. Risk factors for fracture nonunions [7–25].

expensive treatment is required, associated with multiple surgical procedures, prolonged hospital stay, pain and functional and psychosocial disability [26–28]. The implementation of treatment protocols can be complex and long-lasting [29–32]. Moreover, the resulting socio-economic burden should not be underestimated [33,7,34,35]. Being able to identify promptly patients at high risk of non-union, will allow early appropriate targeted treatment intervention leading to a successful outcome [36,37]. Such an approach would benefit not only the patient’s wellbeing but also the health care system in terms of the cost implications associated with long lasting treatment interventions. However, in order for such a strategy to be successful, the appropriate ‘biomarkers’ should be available to predict early the anticipated impaired fracture healing response. A biomarker, in order to be useful must possess some important properties including the ability to predict non-union, to have high sensitivity and specificity and to be easily obtainable [38–40]. The aim of this study therefore is to provide an up to date overview of the currently available ‘biomarkers’ that could be potentially used for the diagnosis of fracture non-union. Materials and methods We searched Medline with general keywords such as ‘nonunions’, ‘bone healing’, ‘bone turnover molecules’ both isolated or in combination to specific words including ‘predictor’, ‘genetic’, ‘predisposition’. For paper selection, the initial inclusion criteria were studies publishing results on potential molecules or genes involved in bone healing in vivo both in humans and animal models. Exclusion criteria included publications of non-English literature or studies having incomplete documentation. Manuscripts that fulfilled the inclusion criteria were subjected to further review of their references list to ensure that no relevant study was left out from the final analysis. Relevant papers were reviewed and data including the studied molecules/genes, methodology, model, results and conclusions were extracted and analysed. Results In total 44 papers met the inclusion criteria, Table 1 [41–84]. According to the content, papers of interest were grouped into either ‘genetic predisposition relevance’ or ‘serum molecular mediator’ activity. Genetic predisposition relevance A biologically insufficient osteogenic response due to an inherent genetic element has been proposed by a number of

studies, Fig. 2 [48–73]. In vivo knock-out animal models have shown that several mutations are sufficient to alter the biologic response either arresting healing or upregulating it. More specifically, Manigrasso et al. [59] reported that 5-lipoxygenase mutated mice healed fractures faster and the newly formed bone had substantially better mechanical properties. Similarly, Myostatin (GDF-8), smooth muscle calponin and recombination activating gene 1 (RAG1) ablated animals have shown an upregulated bone healing reponce [62,63,66]. In two recent studies, mice lacking the tumour suppressor Pten (phosphatase and tensin homologue) and sclerostin gene exhibited increased bone volume and amplified fracture healing rates with larger and more mineralised calluses [64,65]. In contrast to these studies, TNFalpha, chondromodulin-I and VEGF receptor deficient mice have shown impaired bone healing in similar animal studies [60,61,71]. Moreover, transgenic mice, in which the chondrocytes’ Wnt/bcatenin signalling pathway is inhibited have also shown an impaired fracture healing potential [73]. In addition, BMP-2 lacking mice could develop spontaneous fractures that do not resolve with time having blocked signals for the initiation of fracture healing [67]. Fibroblast growth factor receptor 3 (FGFR3) mutated mice showed deficiencies in both osteoblast and osteoclast functions demonstrating decreased bone mass and deficient mineralisation [77]. As far as other growth factors are concerned, the absence of locally produced BMP-7 has no effect on postnatal limb growth, articular cartilage formation, maintenance of bone mass, or fracture healing [69]. Similarly, BMP-4 ablation was found to have no effect on bone formation and function in the limb [68]. These findings might suggest that the presence of other BMPs might be sufficient to compensate for the absence of BMP-4 or BMP-7. In experimental animal models of induced atrophic non-unions, the local expression of BMP-2, 3, 3B, 4, 6, 7, GDF-5, 7, and BMP antagonists noggin, drm, screlostin, and bone morphogenetic protein and activin membrane-bound inhibitor (BAMBI) were found to be significantly lower in the non-union group [58]. In another study, insulin-like growth factors (IGF) and IGF binding protein-6 were significantly higher in the non-union group while the IGF binding protein-5 was significantly lower at several time points [76].

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732

1727

Genec Predisposion to Non-union Gene polymorphisms

Non-union site expression expression: • • • • • • • • • • • • • • •

Cysteine dioxygenase Myomegalin Cartilage oligomeric matrix protein Fibromodulin Clusterin Fibronectin 1 ACTA2 Actin TGF-b-stimulated protein TSC22 MGP Collagen III BMP-4 Drm/Gremlin Follistatin Noggin Insulin-like growth factors expression :

• • •

BMPs (2, 3, 3B, 4, 6, 7) BAMBI Screlostin

• • • •

Noggin • MMP-13 SMAD6 • TLR 4 • TGF-β PDGF ADAMTS18

Knockout Models Increased response: 5-lipoxygenase • • Myostatin (GDF-8) smooth muscle calponin • • RAG1 • Pten • Sclerostin Decreased response: • BMP-2 TNF-alpha • Chondromodulin-I • • VEGF receptor deficiecy Wnt/β-catenin • • FGFR3 No effect: • BMP-4 and BMP-7

Fig. 2. Genetic factors predisposing to non-union [58–73].

In humans, limited data exist to date linking a specific gene to a predetermined inability for bone healing. Zimmerman et al. [72] compared the expression of a large number of molecules from samples obtained from the non-union site to that of patients with uneventful healing. Their results showed significant upregulation of a number of molecules including Cysteine dioxygenase, myomegalin, cartilage oligomeric matrix protein, Fibromodulin, Clusterin, Fibronectin 1, ACTA2 Actin and TGF-b-stimulated protein TSC22. In two similarly designed studies, in normally healing fractures, matureosteoblasts isolated from woven bone, were negative for MGP and collagen III mRNA, while in non-unions, osteoblasts were positive for both MGP and collagen III mRNA [74,75]. Substantially elevated concentrations of BMP-4, Drm/ Gremlin, follistatin, and Noggin in non-union tissue was found after comparison with normal healing bone by Fajardo et al. [81] In a similar study, BMP-7 concentration was found to be significantly higher in the healing bone compared to non-unions. In cases of hypertrophic non-unions, the expression of matrix metalloproteinases-7 and -12 have been reported to be significantly upregulated [82]. It was also noted that both molecules were able to bind and degrade BMP-2 in vitro. Xiong et al. [70] in a broad study including people from the major human ethnic groups underlined the significance of ADAMTS18 and TGFBR3 in the determination of bone mass density. They have also highlighted the potential use of ADAMTS18 level determination which was significantly lower in subjects with non-union fractures as compared to subjects with normal-healing fractures. Advances in the field of genetic research, and more precisely the use of single nucleotide polymorphisms (SNPs), enable further determination of single nucleotide alternation in the DNA sequence. Using these tools, Dimitriou et al. [78] analysed fifteen SNPs within four genes of the Bone Morphogenetic Protein (BMP) pathway in 109 patients with long bone fractures (62 non-unions vs. 47 uneventful union). Polymorphisms on Noggin and SMAD6 genes were found to be associated with higher non-union risk. In another study, polymorphisms within the PDGF gene seem to be linked with the development of non-unions, while MMP-13

polymorphism was also reported as a potential genetic risk factor [79]. Finally, Szczenyet et al. [80] proposed that mutation of TLR 4 and TGF-b genes could be crucial for pathogen recognition and elimination leading to prolonged pathogen existence in the fracture gaps and arrest of the healing process. Serum molecular mediator activity Bone healing begins immediately after injury and requires the coordinated action of several cells types together with signals both acting systemically and locally [2,3]. These signals could be in a variety of forms, including inflammatory molecules, proteins, enzymes and osteoinductive growth factors. It has been shown that the levels of such molecules increase after fracture and possibly are affected by factors like surgery, medications or even fluid and blood administration [84]. Therefore, in theory, if a deficient response for the formation of non-unions exists then these molecules could be considered as markers of the healing process. Based on this principle, current evidence suggests that this approach could be feasible. In animal studies, serial determination of ALP activity in dogs during fracture healing showed a link between the increase of ALP and union process [50]. In the ‘union’ group ALP kinetics showed an initial increase, returning to the baseline within a two month period. In the delayed union group the levels remained elevated for longer period while in the nonunion group no significant changes occurred. In cases of potentially infected non-unions, the combination of serum osteocalcin, bonespecific alkaline phosphatase, and deoxypyridinoline concentrations provided an accuracy of 96% at 4 weeks in a New Zealand White rabbits [46]. In another study, Winger et al. [83] suggested that the urinary levels of MMP9 and MMP13 may have potential as metabolic markers to monitor the progression of fracture healing as found to corresponded to the mRNA expression and immunohistologic appearance of the proteins within callus tissues. In contrast, Carboxy-terminal propeptide of procollagen type I (PICP), skeletal alkaline phosphatase (sALP), and amino-terminal

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732

1728 Table 2 Biochemical markers for non-union: human studies. Author/year

Molecule

Study design

Results

Oni et al. (1989) [53]

ALP and OC

50 patients with closed long bone fractures

Emami et al. (1999) [54]

ALP, OC,crosslinkedtelopeptide

30 patients with tibial fractures

Kurdy (2000) [44]

BsALP, PICP, PIIINP, and ICTP

Twenty consecutive patients with isolated tibial shaft fractures (3 delayed unions)

Herrmann et al. (2002) [45]

OC, ALP, b-CrossLaps

14 patients long bone fractures (2 delayed unions)

Zimmermann et al. (2005) [43]

TGF-b1

Zimmermann et al. (2007) [42]

TGF-b1

10 non-unionvs 10 matched cases prospective study, 15 nonunionvs 15 successful union

Ohishi et al. (2008) [52]

Urinary Type I collagen C-terminal telopeptide, pyridinoline, deoxypyridinoline, serum C-terminal telopeptide, and N-midportion of osteocalcin (OC(N-mid))

33 patients with vertebral fractures (9 delayed or nonunion)

Sarahrudi et al. (2008) [49]

VEGF

114 consecutive patients with long bone fractures (11 non-unions)

Goebel et al. (2009) [48]

Fibroblast GF-23

55 patients with THR aseptic loosening

Wang et al. (2009) [57]

VEGF and BMP-2

42 established non-unions treated with extracorporeal shockwave treatment

Sarahrudi et al. (2010) [84]

Macrophage colony stimulating factor

Moghaddam et al. (2011) [51]

TRACP 5b and CTX

103 patients with normal vs. 10 patients with impaired fracture healing 15 non-unionvs 15 successful union

Sarahrudi et al. (2011) [41]

TGF-b1

Granchi et al. (2013) [47]

Fibroblast GF-2

Higher values of osteocalcin in normally uniting fractures No difference of ALP between groups Lower levels of bone specific alkaline phosphatase between 4 and 7 weeks in patients with delayed healing. Patients with delayed union had significantly lower PICP levels and marginally lower BsALP levels at twenty weeks. PIIINP levels were significantly higher at ten weeks. 1 month delay in OC increase in patients with delayed healing. ALP higher in the patients with delayed healing Lower TGF-b1 levels in the non-union group at 4 weeks Gradual decline in serum TGF-b1 concentration in patients with delayed union and lower TGF-b1 levels in the non-union group at 4 weeks. (p = 0.00006). Differences in kinetics of serum OC(Nmid) in the non-union group (started to increase at 2 weeks, peaked at 24 week The differences in the rest of the studied molecules were observed Serum VEGF concentrations in patients with impaired fracture healing were higher but their values failed to reach statistical significance. Patients with aseptic loosening had a >3-fold elevation of serum FGF-23 levels Serum ALP, serum phosphate, and phosphate clearance were within normal limits Patients with bony union showed significantly higher serum NO level, TGF-b1, VEGF and BMP-2 at 1 month after treatment as compared to patients with persistent non-union. No significant differences in the M-CSF serum concentration found between the two groups. In the first week, CTX decreased significantly in cases of delayed fracture healing TRACP 5b were significantly decreased at weeks 4 and 8 Non-union group had higher TGF-b1 concentrations at 6 weeks (p = 0.01) Lower FGF-2 serum levels in patients who did not heal after surgery

9 non-unionvs 9 matched cases 88 children undergoing surgical treatment for orthopaedic conditions

OC: osteocalcin; ALP: alkaline phosphatase; GF: growth factor; VEGF: vascular endothelial growth factor.

propeptide of type III procollagen (PIlINP) failed to serve as predictors of bone union [55]. Similarly, TGFbeta, PDGF, FGFb, and BMP 2/4 determination at the fracture site failed to serve as ‘predictor(s)’ since both the non-union and control groups demonstrated the same presence and distribution of growth factors [56]. In humans several molecules have been investigated to date Table 2 [41–45,47–49,51–54,57,84], Moghaddam et al. [51] studied the serum levels of tartrate-resistant acid phosphatase 5b (TRACP 5b) and C-terminal cross-linking telopeptide of type I collagen (CTX) in 15 patients with atrophic non-union which were matched to 15 patients with uneventful healing serving as the control group. In the delayed fracture healing group, the CTX serum

levels were significantly decreased in the first week and TRACP 5b was found decreased at weeks 4 and 8. Patients with delayed healing found to poses lower levels of bone specific alkaline phosphatase between 4 and 7 weeks than did patients with normal healing, although no such differences were seen for osteocalcin [54]. Granchi et al. [47] has evaluated the peripheral levels of fibroblast growth factors-2 (FGF-2) in 88 children undergoing surgical treatment fororthopaedic conditions. They reported significantly lower FGF-2 serum levels in patients who did not heal after surgery. On the other hand, FGF-23 was found elevated by more than 3-fold at day 3 in patients with aseptic loosening of total hip replacement implants [48]. In this study, although the authors did not specifically involve patients with fractures, one

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732

Invasiveness Easily obtainable/ minimally invasive

Specificity

Sensivity

True negave

True posive rate

rate

IDEAL BIOMARKER

Predictability

Robust

Indicator of the condion/risk, progress and treatment effect

Easy, simple, cost effecveness

Fig. 3. Attitudes of the ideal biomarker.

could hypothesise that the healing/osteointegration process is analogous to that seen in long bone fracture healing. Transforming growth factor b1 has been the subject of a number of studies as a potential marker of delayed fracture healing. Zimmermann et al. were first to depict that although TGFb1 levels increase after fracture, in patients with delayed bone healing a decline occurs earlier [43]. At 4 weeks, serum levels of TGF-b1 were significantly lower in the delayed fracture healing group. In a later prospective study involving 15 patients with nonunion with matched controls, the same authors, presented a gradual decline in serum TGF-b1 concentration in patients with delayed union and lower TGF-b1 levels in the non-union group at 4 weeks [42]. These results however, were contradicted by the study of Sarahrudi et al. [41] who in a similar study reported higher TGFb1 concentrations at 6 weeks for the non-union group. Discussion A biomarker is a measurable characteristic, which can be used as an indicator for a condition or a state of the organism. Hulka et al. [38] defined the biological markers as ‘‘cellular, biochemical or molecular alterations that are measurable in biological media such as human tissues, cells or fluids’’. Based on this definition biomarkers could range from simple observations i.e. temperature for fever, to cancer markers. Furthermore the rapid recent development of molecular biology and laboratory techniques together with advances in proteomics, genomics, and biotechnologies have brought in surface many molecules that could be used in the diagnosis, prognosis, prediction or therapy response of different conditions. The ideal biomarker would have to fulfil several criteria, Fig. 3 These include high specificity and sensitivity for the given condition with absent of false positive and false negative observations [31]. The biomarker should have predictive value with long half-life in bodily fluids/tissue but equally with levels proportionate to the extent of the condition in question [39]. It would have to be easily obtainable with minimally invasive techniques. Finally, its determination should be simple, accurate, rapid, and inexpensive [39,40].

1729

In this context we analysed the results of numerous studies dealing either with genetic markers of increased non-union susceptibility or circulating molecules closely related to the evolution of the fracture healing process. Such knowledge would enable clinicians to identify patients at risk early, modify treatment protocols and thus preventing long-standing morbidity and disability [86]. However, it seems that despite these studies indicating specific markers and covering a great variety of molecules and genes, in reality, we are far away from a definite answer. To make matters worse, as highlighted by the study of Brinker et al., [87] patients with non-unions often have complex biology and accompanying pathologies. It is of note that 31 out of the 37 patients suffering with unexplained or unexpected nonunions found to have an endocrine/metabolic disorder including Vitamin D deficiency, hypothyroid, hypogonadism and poor calcium intake [88]. With regard to serum cytokines and growth factors, it has been shown that in addition to the local release, systemic release of several important molecules takes place simultaneously and their kinetic pattern follows a specific trend [85,88]. It has been speculated that an insufficient systemic supply of such molecules could lead to reduced osteogenic differentiation of progenitor cells and loss of bone mass [89]. Therefore, it could be hypothesised that differences in kinetics could enable us to depict cases of inadequate fracture healing response. Several studies have proposed molecules to serve as ‘biological markers’ of fracture union. However, several draw-backs exist. Firstly, a number of factors can influence their release and systemic levels. Such factors include the patients’ age, race, gender, weight, bone mass, medications and concomitant pathologies [90,91]. For instance, Kaiser et al. [92] has pointed out that TGF-b1 serum concentration could be lower in smokers compared to nonsmokers at 8 weeks after fracture. The same study reported higher concentration in males and younger individuals at 24 weeks post injury as well as for chronic alcoholics and non-diabetic patients. It is also of note that even the sampling conditions or the technique used to measure the TGF-b1 could alter significantly the result [93]. In another study, it was highlighted that the age, gender, diabetes mellitus and cigarette smoking significantly influence the expression of M-CSF and VEGF after fracture of long bones in humans [94]. Of note, diabetic patients showed significantly elevated overall VEGF levels when compared to non-diabetic patients. Therefore, further studies with larger patient cohorts are desirable to better understand the influence of these endogenous and exogenous factors on the expression of the osteogenic response during human fracture healing [94]. Secondly, the extend of trauma and the size of the underlying bone could equally have a substantial impact on the release of circulating levels of growth factors [95]. Finally, it should become clear that the early local expression and concentration in bone cannot be mirrored by their circulating levels. From a clinician’s perspective, genetic testing to identify cases of aberrant bone healing could be a powerful tool. Data from several animal models have highlighted potential pathways responsible for the development of non-union. However, several variables exist including the variation reported within inbred animal strains which has been associated with significant differences in terms of healing rate, mineral density and regeneration rate [96,97]. Moreover, the potential discrepancies between animal and human bone healing governing processes should not be underestimated. In addition, atrophic non-union animal models are produced by periosteal cauterisation, which differs significantly from the clinical conditions in which atrophic non-unions develop in humans [58]. The above issues that one has to take into account when analysing data from animal studies, similarly, have to be

1730

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732

considered when analysing human data. The differences between individuals in terms of the amount and strength of callus together with the time to union could suggest genetic variation between populations. In this context, the gene–environmental interactions should be taken into account. For example, lessons learnt from other areas of bone biology like peak bone mass, show that a link between genetic variations and vitamin receptor alleles exist [98]. However, this association could be only reproduced in some populations while rejected in others [99,100]. Therefore, genetic heterogeneity is coupled with environmental factors, making any attempt of a simplistic solution rather difficult [100]. Bone healing biology is so well orchestrated that any defect between the haematoma formation and the remodelling could affect the final outcome. It should also be mentioned, as bone healing resembles embryonic development and bone formation, significant phenotypic alternation would lead to major defects sometimes not compatible with life. Therefore, shuttle fine genetic defects could be responsible that would require extensive research with large cohort of patients with control of the environmental and other contributing factors. Despite the substantial continuous progress in the identification of a robust marker for bony non-union, the presented available data could only serve as foundation work towards this goal. Limitations include the small number of patients recruited and analysed, the genetic heterogeneity and variation of the population used, the variability of fractures presented and the inherit difficulty in randomizing patients with non-unions. For instance, should cases, where a less than perfect surgical technique is used, be included in the study? Equally should patients with different risk factors be included in the same study? Despite the work that has been done up to now, no robust recommendations can be made with respect to which ‘biomarker’ should be used in the clinical setting. Future studies must include high number of patients to allow control of the heterogeneity of the population. A different approach with regard to the study design should also be considered allowing for the screening of molecules capable of inducing and upregulating osteogenesis, chondrogenesis and bone mineralisation. Moreover, combination of molecules at different time points rather than a single marker alone could offer more predictability with increased sensitivity and specificity. Finally, negative feedback loops should be taken into account as such complex mechanisms could be responsible for the conflicting data, i.e. TGF-b1 data. Conclusion Several molecules and genes have been investigated as predictors of fracture non-union. However the limited available data do not encourage the routine use of any of the existing markers to assess the progression of a fracture to union or being at risk of non-union. Further research is warranted to assess their potential role. Conflict of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper. References [1] Calori GM, Giannoudis PV. Enhancement of fracture healing with the diamond concept: the role of the biological chamber. Injury 2011;42:1191–3. [2] Marsell R, Einhorn TA. The biology of fracture healing. Injury 2011;42:551–5. [3] Giannoudis PV, Jones E, Einhorn TA. Fracture healing and bone repair. Injury 2011;42:549–50. [4] McKibbin B. The biology of fracture healing in long bones. J Bone Joint Surg Br 1978;60 B:150–62.

[5] Praemer A, Furner S, Rice DP. Musculoskeletal injuries. In: Musculoskeletal Conditions in the United States. Park Ridge, IL: American Academy of Orthopaedic Surgeons; 1992. p. 85–124. [6] Wheelers’ Online Orthopaedic Textbook. http://www.wheelessonline.com/ ortho/tibial_non_unions. [7] Day SM, DeHeer DH. Reversal of the detrimental effects of chronic protein malnutrition on long bone fracture healing. J Orthop Trauma 2001;15:47– 53. [8] Pountos I, Georgouli T, Blokhuis TJ, Pape HC, Giannoudis PV. Pharmacological agents and impairment of fracture healing: what is the evidence? Injury 2008;39:384–94. [9] Pountos I, Georgouli T, Bird H, Kontakis G, Giannoudis PV. The effect of antibiotics on bone healing: current evidence. Exp Opin Drug Saf 2011;10: 935–45. [10] Urabe K, Hotokebuchi T, Oles KJ, Bronk JT, Jingushi S, Iwamoto Y, et al. Inhibition of endochondral ossification during fracture repair in experimental hypothyroid rats. J Orthop Res 1999;17:920–5. [11] Heppenstall RB, Brighton CT. Fracture healing in the presence of anemia. Clin Orthop Relat Res 1977;123:253–8. [12] Lauing KL, Roper PM, Nauer RK, Callaci JJ. Acute alcohol exposure impairs fracture healing and deregulates b-catenin signaling in the fracture callus. Alcohol Clin Exp Res 2012;36:2095–103. [13] Ko SB, Lee SW. Do fibula nonunions predict later tibia nonunions? J Orthop Trauma 2013;27:150–2. [14] Ziran B, Cheung S, Smith W, Westerheide K. Comparative efficacy of 2 different demineralized bone matrix allografts in treating long-bone nonunions in heavy tobacco smokers. Am J Orthop (Belle Mead NJ) 2005;34:329–32. [15] Calori GM, Albisetti W, Agus A, Iori S, Tagliabue L. Risk factors contributing to fracture non-unions. Injury 2007;38(Suppl. (2)):S11–8. [16] Hayda RA, Brighton CT, Esterhai Jr JL. Pathophysiology of delayedhealing. Clin Orthop Relat Res 1998;355(Suppl):S31–40. [17] Dickson K, Katzman S, Delgado E, Contreras D. Delayed unions and nonunions of open tibial fractures. Correlation with arteriography results. Clin Orthop Relat Res 1994;302:189–93. [18] Claes L, Augat P, Suger G, Wilke HJ. Influence of size and stability of the osteotomy gap on the success of fracture healing. Orthop Res 1997;15: 577–84. [19] Rodriguez-Merchan EC, Forriol F. Nonunion: general principles and experimental data. Clin Orthop Relat Res 2004;419:4–12. [20] Pountos I, Georgouli T, Bird H, Giannoudis PV. Nonsteroidal anti-inflammatory drugs: prostaglandins, indications, and side effects. Int J Interferon Cytokine Mediator Res 2011;3:19–27. [21] Van Wunnik BPW, Weijers PHE, Van Helden SH, Brink PRG, Poeze M. Osteoporosis is not a risk factor for the development of nonunion: a cohort nested case-control study. Injury 2011;42:1491–4. [22] Gao YS, Ai ZS, Yu XW, Sheng JG, Jin DX, Chen SB, et al. Free vascularised fibular grafting combined with a locking plate for massive bone defects in the lower limbs: a retrospective analysis of fibular hypertrophy in 18 cases. Injury 2012;43:1090–5. [23] Weaver MJ, Harris MB, Strom AC, Smith RM, Lhowe D, Zurakowski D, et al. Fracture pattern and fixation type related to loss of reduction in bicondylartibial plateau fractures. Injury 2012;43:864–9. [24] Hardeman F, Bollars P, Donnely P, Bellemans J, Nijs S. Predictive factors for functional outcome and failure in angular stable osteosynthesis of the proximal humerus. Injury 2012;43:153–8. [25] Reverte MM, Dimitriou R, Kanakaris NK, Giannoudis PV. What is the effect of compartment syndrome and fasciotomies on fracture healing in tibial fractures? Injury 2011;42:1402–7. [26] Ashman O, Phillips AM. Treatment of non-unions with bone defects: which option and why? Injury 2013;44(Suppl. (1)):S43–5. [27] Giannoudis PV, Atkins R. Management of long-bone non-unions. Injury 2007;38(Suppl. (2)):S1–2. [28] Giannoudis PV, Kontakis G. Treatment of long bone aseptic non-unions: monotherapy or polytherapy? Injury 2009;40:1021–2. [29] Guyver P, Wakeling C, Naik K, Norton M. Judetosteoperiosteal decortication for treatment of non-union: the Cornwall experience. Injury 2012;43: 1187–92. [30] Calori GM, Mazza E, Colombo M, Ripamonti C. The use of bone-graft substitutes in large bone defects: any specific needs? Injury 2011;42:S56–63. [31] Argintar E, Edwards S, Delahay J. Bone morphogenetic proteins in orthopaedic trauma surgery. Injury 2011;42:730–4. [32] Schroeder JE, Mosheiff R. Tissue engineering approaches for bone repair: concepts and evidence. Injury 2011;42:609–13. [33] Papanna MC, Al-Hadithy N, Somanchi BV, Sewell MD, Robinson PM, Khan SA, et al. The use of bonemorphogenic protein-7 (OP-1) in the management of resistant non-unions in the upper and lower limb. Injury 2012;43: 1135–40. [34] Giannoudis PV, Atkins R. Management of long-bone non-unions. Injury 2007;38(Suppl. 2):S1–2. [35] Wilkins RM, Kelly CM. The effect of allomatrix injectable putty on the outcome of long bone applications. Orthopedics 2003;26:s567–70. [36] Axelrad TW, Einhorn TA. Use of clinical assessment tools in the evaluation of fracture healing. Injury 2011;42:301–5. [37] Gelalis ID, Politis AN, Arnaoutoglou CM, Korompilias AV, Pakos EE, Vekris MD, et al. Diagnostic and treatment modalities in nonunions of the femoral shaft: a review. Injury 2012;43:980–8.

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732 [38] Hulka BS, Griffith JD, Wilcosky TC, editors. Biological markers in epidemiology. New York: Oxford University Press; 1990. [39] Riley RD, Sauerbrei W, Altman DG. Prognosticmarkers in cancer: the evolution of evidence from single studies to meta-analysis, and beyond. Br J Cancer 2009;100::1219–29. [40] Minton JP, Chevinsky A. Present status of serum markers. Semin Surg Oncol 1989;5:426–35. [41] Sarahrudi K, Thomas A, Mousavi M, Kaiser G, Ko¨ttstorfer J, Kecht M, et al. Elevated transforming growth factor-beta 1 (TGF-b1) levels in human fracture healing. Injury 2011;42:833–7. [42] Zimmermann G, Moghaddam A, Reumann M, Wangler B, Breier L, Wentzensen A, et al. TGF-beta1 as a pathophysiological factor in fracture healing. Unfallchirurg 2007;110:130–6. [43] Zimmermann G, Henle P, Ku¨sswetter M, Moghaddam A, Wentzensen A, Richter W, et al. TGF-beta1 as a marker of delayed fracture healing. Bone 2005;36:779–85. [44] Kurdy NM. Serology of abnormal fracture healing: the role of PIIINP PICP, and BsALP. J Orthop Trauma 2000;14:48–53. [45] Herrmann M, Klitscher D, Georg T, Frank J, Marzi I, Herrmann W. Different kinetics of bone markers in normal and delayed fracture healing of long bones. Clin Chem 2002;48:2263–6. [46] Southwood LL, Frisbie DD, Kawcak CE, McIlwraith CW. Evaluation of serum biochemical markers of bone metabolism for early diagnosis of nonunion and infected nonunion fractures in rabbits. Am J Vet Res 2003;64:727–35. [47] Granchi D, Devescovi V, Pratelli L, Verri E, Magnani M, Donzelli O, et al. Serum levels of fibroblast growth factor 2 in children with orthopedic diseases: potential role in predicting bone healing. J Orthop Res 2013;31:249–56. [48] Goebel S, Lienau J, Rammoser U, Seefried L, Wintgens KF, Seufert J, et al. FGF23 is a putative marker for bone healing and regeneration. J Orthop Res 2009;27:1141–6. [49] Sarahrudi K, Thomas A, Braunsteiner T, Wolf H, Ve´csei V, Aharinejad S. VEGF serum concentrations in patients with long bone fractures: a comparison between impaired and normal fracture healing. J Orthop Res 2009;27: 1293–7. [50] Komnenou A, Karayannopoulou M, Polizopoulou ZS, Constantinidis TC, Dessiris A. Correlation of serum alkaline phosphatase activity with the healing process of long bone fractures in dogs. Vet Clin Pathol 2005;34:35–8. [51] Moghaddam A, Mu¨ller U, Roth HJ, Wentzensen A, Gru¨tzner PA. Zimmermann G. TRACP 5b and CTX as osteological markers of delayed fracture healing. Injury 2011;42:758–64. [52] Ohishi T, Takahashi M, Yamanashi A, Suzuki D, Nagano A. Sequential changes of bone metabolism in normal and delayed union of the spine. Clin Orthop Relat Res 2008;466:402–10. [53] Oni OO, Mahabir JP, Iqbal SJ, Gregg PJ. Serumosteocalcin and total alkaline phosphatase levels as prognostic indicators in tibial shaft fractures. Injury 1989;20:37–8. [54] Emami A, Larsson A, Petre´n-Mallmin M, Larsson S. Serum bonemarkers after intramedullary fixed tibial fractures. Clin Orthop Relat Res 1999;368:220–9. [55] Klein P, Bail HJ, Schell H, Michel R, Amthauer H, Bragulla H, et al. Are boneturnovermarkerscapable of predictingcallusconsolidation during bonehealing? Calcif Tissue Int 2004;75:40–9. [56] Brownlow HC, Reed A, Simpson AH. Growthfactorexpression during the development of atrophicnon-union. Injury 2001;32:519–24. [57] Wang CJ, Yang KD, Ko JY, Huang CC, Huang HY, Wang FS. The effects of shockwave on bone healing and systemic concentrations of nitric oxide (NO), TGF-beta1 VEGF and BMP-2 in long bone non-unions. Nitric Oxide 2009;20:298–303. [58] Niikura T, Hak DJ, Reddi AH. Global gene profiling reveals a downregulation of BMP gene expression in experimental atrophic nonunions compared to standard healing fractures. J Orthop Res 2006;24:1463–71. [59] Manigrasso MB, O’Connor JP. Accelerated fracture healing in mice lacking the 5-lipoxygenase gene. Acta Orthop 2010;81:748–55. [60] Gerstenfeld LC, Cho TJ, Kon T, Aizawa T, Tsay A, Fitch J, et al. Impaired fracture healing in the absence of TNF-alpha signaling: the role of TNF-alpha in endochondral cartilage resorption. J Bone Miner Res 2003;18:1584–92. [61] Maes C, Carmeliet P, Moermans K, Stockmans I, Smets N, Collen D, et al. Impaired angiogenesis and endochondralbone formation in mice lacking the vascular endothelial growth factor isoforms VEGF164 and VEGF188. Mech Dev 2002;111:61–73. [62] Kellum E, Starr H, Arounleut P, Immel D, Fulzele S, Wenger K, et al. Myostatin (GDF-8) deficiency increases fracture callus size Sox-5 expression, and callus bone volume. Bone 2009;44:17–23. [63] Yoshikawa H, Taniguchi SI, Yamamura H, Mori S, Sugimoto M, Miyado K, et al. Mice lacking smooth muscle calponindisplay increased bone formation that is associated with enhancement of bone morphogenetic protein responses. Genes Cells 1998;3:685–95. [64] Burgers TA, Hoffmann MF, Collins CJ, Zahatnansky J, Alvarado MA, Morris MR, et al. Micelackingpten in osteoblasts have improved intramembranous and late endochondralfracture healing. PLoS ONE 2013;8:e6385–87. [65] Li C, Ominsky MS, Tan HL, Barrero M, Niu QT, Asuncion FJ, et al. Increased callus mass and enhanced strength during fracture healing in micelacking the sclerostin gene. Bone 2011;49:1178–85. [66] Toben D, Schroeder I, El Khassawna T, Mehta M, Hoffmann JE, Frisch JT, et al. Fracture healing is accelerated in the absence of the adaptive immune system. J Bone Miner Res 2011;26:113–24.

1731

[67] Tsuji K, Bandyopadhyay A, Harfe BD, Cox K, Kakar S, Gerstenfeld L, et al. BMP2 activity, although dispensable for bone formation, is required for the initiation of fracture healing. Nat Genet 2006;38:1424–9. [68] Tsuji K, Cox K, Bandyopadhyay A, Harfe BD, Tabin CJ, Rosen V. BMP4 is dispensable for skeletogenesis and fracture-healing in the limb. J Bone Joint Surg Am 2008;90(Suppl. (1)):14–8. [69] Tsuji K, Cox K, Gamer L, Graf D, Economides A, Rosen V. Conditional deletion of BMP7 from the limb skeleton does not affect bone formation or fracture repair. J Orthop Res 2010;28:384–9. [70] Xiong DH, Liu XG, Guo YF, Tan LJ, Wang L, Sha BY, et al. Genome-wide association and follow-up replication studies identified ADAMTS18 and TGFBR3 as bone mass candidate genes in different ethnic groups. Am J Hum Genet 2009;84:388–98. [71] Yukata K, Matsui Y, Shukunami C, Takimoto A, Goto T, Nishizaki Y, et al. Altered fracture callus formation in chondromodulin-I deficient mice. Bone 2008;43:1047–56. [72] Zimmermann G, Schmeckenbecher KH, Boeuf S, Weiss S, Bock R, Moghaddam A, et al. Differential gene expression analysis in fracture callus of patients with regular and failed bone healing. Injury 2012;43:347–56. [73] Huang Y, Zhang X, Du K, Yang F, Shi Y, Huang J, et al. Inhibition of b-catenin signaling in chondrocytes induces delayed fracture healing in mice. J Orthop Res 2012;30:304–10. [74] Lawton DM, Andrew JG, Marsh DR, Hoyland JA, Freemont AJ. Expression of the geneencoding the matrixglaprotein by matureosteoblasts in humanfracture non-unions. Mol Pathol 1999;52:92–6. [75] Lawton DM, Andrew JG, Marsh DR, Hoyland JA, Freemont AJ. Matureosteoblasts in humannon-unionfracturesexpresscollagen type III. Mol Pathol 1997;50:194–7. [76] Koh A, Niikura T, Lee SY, Oe K, Koga T, Dogaki Y, et al. Differentialgene expression and immunolocalization of insulin-like growth factors and insulin-like growth factor binding proteins between experimental nonunions and standard healing fractures. J Orthop Res 2011;29:1820–6. [77] Su N, Sun Q, Li C, Lu X, Qi H, Chen S, et al. Gain-of-function mutation in FGFR3 in mice leads to decreased bone mass by affecting both osteoblastogenesis and osteoclastogenesis. Hum Mol Genet 2010;19: 1199–210. [78] Dimitriou R, Carr IM, West RM, Markham AF, Giannoudis PV. Genetic predisposition to fracture non-union: a case control study of a preliminary single nucleotide polymorphisms analysis of the BMP pathway. BMC Musculoskelet Disord 2011;12:44. [79] Zeckey C, Hildebrand F, Glaubitz LM, Ju¨rgens S, Ludwig T, Andruszkow H, et al. Are polymorphisms of molecules involved in bone healing correlated to aseptic femoral and tibial shaft non-unions? J Orthop Res 2011;29: 1724–31. [80] Szcze˛sny G, Olszewski WL, Zagozda M, Rutkowska J, Czapnik Z, SwobodaKopec´ E, et al. Genetic factors responsible for long bone fractures non-union. Arch Orthop Trauma Surg 2011;131:275–81. [81] Fajardo M, Liu CJ, Egol K. Levels of expression for BMP-7 and several BMP antagonists may play an integral role in a fracture nonunion: a pilot study. Clin Orthop Relat Res 2009;467:3071–8. [82] Fajardo M, Liu CJ, Ilalov K, Egol KA. Matrix metalloproteinases that associate with and cleave bone morphogenetic protein-2 in vitro are elevated in hypertrophic fracture nonunion tissue. J Orthop Trauma 2010;24: 557–63. [83] Wigner NA, Kulkarni N, Yakavonis M, Young M, Tinsley B, Meeks B, et al. Urine matrix metalloproteinases (MMPs) as biomarkers for the progression of fracture healing. Injury 2012;43:274–8. [84] Sarahrudi K, Mousavi M, Thomas A, Eipeldauer S, Ve´csei V, Pietschmann P, et al. Elevated levels of macrophage colony-stimulating factor in human fracture healing. J Orthop Res 2010;28:671–6. [85] Pountos I, Georgouli T, Henshaw K, Bird H, Giannoudis PV. Release of growth factors and the effect of age, sex, and severity of injury after long bone fracture. A preliminary report. Acta Orthop 2013;84:65–70. [86] Dimitriou R, Kanakaris N, Soucacos PN, Giannoudis PV. Genetic predisposition to non-union: evidence today. Injury 2013;44(Suppl. (1)): S50–3. [87] Brinker MR, O’Connor DP, Monla YT, Earthman TP. Metabolic and endocrine abnormalities in patients with nonunions. J Orthop Trauma 2007;21: 557–70. [88] Giannoudis PV, Pountos I, Morley J, Perry S, Tarkin HI, Pape HC. Growth factor release following femoral nailing. Bone 2008;42:751–7. [89] Manolagas SC, Jilka RL. Bone marrow, cytokines, and bone remodeling. Emerging insights into the pathophysiology of osteoporosis. N Engl J Med 1995;332:305–11. [90] Gundberg CM, Looker AC, Nieman SD, Calvo MS. Patterns of osteocalcin and bone specific alkaline phosphatase by age, gender, and race or ethnicity. Bone 2002;31:703–8. [91] Hoesel LM, Wehr U, Rambeck WA, Schnettler R. HeissC.Biochemical bone markers are useful to monitor fracture repair. Clin Orthop Relat Res 2005;440:226–32. [92] Kaiser G, Thomas A, Ko¨ttstorfer J, Kecht M, Sarahrudi K. Is the expression of transforming growth factor-beta1 after fracture of long bones solely influenced by the healingprocess? Int Orthop 2012;36:2173–9. [93] Zhao L, Wang L, Ji W, Lei M, Yang W, Kong FM. The influence of the blood handling process on the measurement of circulating TGF-b1. Eur Cytokine Netw 2012;23:1–6.

1732

I. Pountos et al. / Injury, Int. J. Care Injured 44 (2013) 1725–1732

[94] Ko¨ttstorfer J, Kaiser G, Thomas A, Gregori M, Kecht M, Domaszewski F, et al. The influence of non-osteogenic factors on the expression of M-CSF and VEGF during fracture healing. Injury 2013;44:930–4. [95] Stoffel K, Engler H, Kuster M, Riesen W. Changes in biochemical markersafter lowerlimbfractures. Clin Chem 2007;53:131–4. [96] Manigrasso MB, O’Connor JP. Comparison of fracture healing among different inbred mouse strains. Calcif Tissue Int 2008;82:465–74. [97] Jepsen KJ, Hu B, Tommasini SM, Courtland HW, Price C, Terranova CJ, et al. Genetic randomization reveals functional relationships among morphologic and tissue-quality traits that contribute to bone strength and fragility. Mamm Genome 2007;18:492–507.

[98] Ferrari S, Rizzoli R, Chevalley T, Slosman D, Eisman JA, Bonjour JP. Vitamin-Dreceptor-gene polymorphisms and change in lumbar-spine bone mineral density. Lancet 1995;345:423–4. [99] Looney JE, Yoon HK, Fischer M, Farley SM, Farley JR, Wergedal JE, et al. Lack of a high prevalence of the BB vitamin D receptor genotype in severely osteoporotic women. J Clin Endocrinol Metab 1995;80: 2158–62. [100] Fujita Y, Katsumata K, Unno A, Tawa T, Tokita A. Factors affecting peak bone density in Japanese women. Calcif Tissue Int 1999;64:107–11.