Early Changes in Bone Density, Microarchitecture, Bone Resorption, and Inflammation Predict the Clinical Outcome 12 Weeks After Conservatively Treated Distal Radius Fractures: An Exploratory Study

Authors

  • Ursina Meyer,

    1. Department of Human Movement Sciences, Maastricht University, Maastricht, The Netherlands
    2. Research School NUTRIM, Maastricht University, Maastricht, The Netherlands
    3. Centre on Aging and Mobility, University of Zurich, Zurich, Switzerland
    Search for more papers by this author
  • Joost J de Jong,

    Corresponding author
    1. Research School NUTRIM, Maastricht University, Maastricht, The Netherlands
    2. Department of Internal Medicine, Maastricht University Medical Center, Maastricht, The Netherlands
    • Address correspondence to: Joost de Jong, MSc, NUTRIM/Internal Medicine, Faculty of Health, Medicine and Life Sciences, Maastricht University, Universiteitssingel 40, 6229 ER, Maastricht, The Netherlands. E-mail: joost.dejong@maastrichtuniversity.nl

    Search for more papers by this author
  • Sandrine GP Bours,

    1. Department of Internal Medicine, Maastricht University Medical Center, Maastricht, The Netherlands
    Search for more papers by this author
  • András P Keszei,

    1. Department of Medical Informatics, RWTH Aachen University, Aachen, Germany
    Search for more papers by this author
  • Jacobus J Arts,

    1. Department of Orthopaedic Surgery, Maastricht University Medical Center, Maastricht, The Netherlands
    2. Research School CAPHRI, Maastricht University, Maastricht, The Netherlands
    3. Department of Family Medicine, Maastricht University, Maastricht, The Netherlands
    Search for more papers by this author
  • Peter RG Brink,

    1. Department of Trauma Surgery, Maastricht University Medical Center, Maastricht, The Netherlands
    Search for more papers by this author
  • Paul Menheere,

    1. Central Diagnostic Laboratory, Maastricht University Medical Center, Maastricht, The Netherlands
    Search for more papers by this author
  • Tineke ACM van Geel,

    1. Research School NUTRIM, Maastricht University, Maastricht, The Netherlands
    2. Research School CAPHRI, Maastricht University, Maastricht, The Netherlands
    3. Department of Family Medicine, Maastricht University, Maastricht, The Netherlands
    Search for more papers by this author
  • Bert van Rietbergen,

    1. Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands
    Search for more papers by this author
  • Joop PW van den Bergh,

    1. Research School NUTRIM, Maastricht University, Maastricht, The Netherlands
    2. Biomedical Research Center, Hasselt University, Hasselt, Belgium
    3. Department of Internal Medicine, Viecuri Medical Center Venlo, Venlo, The Netherlands
    Search for more papers by this author
  • Piet P Geusens,

    1. Research School CAPHRI, Maastricht University, Maastricht, The Netherlands
    2. Biomedical Research Center, Hasselt University, Hasselt, Belgium
    3. Department of Rheumatology, Maastricht University Medical Center, Maastricht, The Netherlands
    Search for more papers by this author
  • Paul C Willems

    1. Department of Orthopaedic Surgery, Maastricht University Medical Center, Maastricht, The Netherlands
    2. Research School CAPHRI, Maastricht University, Maastricht, The Netherlands
    Search for more papers by this author

ABSTRACT

Fracture healing is an active process with early changes in bone and inflammation. We performed an exploratory study evaluating the association between early changes in densitometric, structural, biomechanical, and biochemical bone parameters during the first weeks of fracture healing and wrist-specific pain and disability at 12 weeks in postmenopausal women with a conservatively treated distal radius fracture. Eighteen patients (aged 64 ± 8 years) were evaluated at 1 to 2 and 3 to 4 weeks postfracture, using high-resolution peripheral quantitative computed tomography (HR-pQCT), micro-finite element analysis, serum procollagen type-I N-terminal propeptide (P1NP), carboxy-terminal telopeptide of type I collagen (ICTP), and high-sensitive C-reactive protein (hsCRP). After 12 weeks, patients rated their pain and disability using Patient Rated Wrist Evaluation (PRWE) questionnaire. Additionally, Quick Disability of the Arm Shoulder and Hand (QuickDASH) questionnaire and active wrist range of motion was evaluated. Linear regression models were used to study the relationship between changes in bone parameters and in hsCRP from visit 1 to 2 and PRWE score after 12 weeks. A lower PRWE outcome, indicating better outcome, was significantly related to an early increase in trabecular bone mineral density (BMD) (β −0.96 [95% CI −1.75 to −0.16], R2 = 0.37), in torsional stiffness (−0.14 [−0.28 to −0.004], R2 = 0.31), and to an early decrease in trabecular separation (209 [15 to 402], R2 = 0.33) and in ICTP (12.1 [0.0 to 24.1], R2 = 0.34). Similar results were found for QuickDASH. Higher total dorsal and palmar flexion range of motion was significantly related to early increase in hsCRP (9.62 [3.90 to 15.34], R2 = 0.52). This exploratory study indicates that the assessment of early changes in trabecular BMD, trabecular separation, calculated torsional stiffness, bone resorption marker ICTP, and hsCRP after a distal radius fracture provides valuable information regarding the 12-week clinical outcome in terms of pain, disability, and range of motion and validates its use in studies on the process of early fracture healing. © 2014 American Society for Bone and Mineral Research.

Introduction

Distal radius fractures are among the most common fractures in elderly.[1] Although these fractures have a limited impact on general health-related quality of life,[2, 3] long-term complications are more commonly reported when using questionnaires especially developed to evaluate pain and function of the wrist: Persisting pain, stiffness, weakness of the wrist, and signs of the complex regional pain syndrome are the most mentioned complications that limit the patient's ability to manage daily life activities, such as household work or personal care.[4-8]

Considering the high variability in patient-rated pain and disability,[9] it would be of great clinical value to know to what degree specific bone parameters can predict long-term clinical outcome after a fracture. In studies examining baseline predictors for clinical outcome, patient factors, injury patterns, and fracture-related characteristics played only a limited role in predicting the clinical outcome after a fracture.[10-13] To our knowledge, there is no study available that evaluated long-term outcome in terms of pain and disability in relation to early changes in bone mineral density (BMD), local bone architecture, and serum bone markers and inflammation during fracture healing. Fracture healing is a complex process in which inflammation and bone remodeling are intensively involved. We, therefore, hypothesized that early changes in these mechanisms would predict clinical outcome at longer term. If so, influencing these changes early in the process of fracture healing could influence the clinical outcome.

The recent development of high-resolution peripheral quantitative computed tomography (HR-pQCT) allows in vivo assessment of 3D bone density and bone microarchitecture of the distal radius and calculation of biomechanical properties of the bone using micro-finite element analysis (µFEA).[14, 15] Therefore, the aim of this exploratory study was to assess the association between early fracture healing-related changes in bone mineral density, bone microarchitecture, calculated biomechanical properties, and serum markers of bone turnover and inflammation with wrist-specific pain and disability at 12 weeks postfracture in postmenopausal women with a stable, conservatively treated distal radius fracture.

Materials and Methods

Design and study population

Twenty women aged 50 years or older who presented with a stable distal radius fracture at the departments of orthopedic surgery and traumatology of the Maastricht University Medical Center were included in this study. All fractures were conservatively treated by cast immobilization, if necessary after manual reduction. Exclusion criteria were a history of previous fracture at the fractured side; known systemic or metabolic disorders leading to progressive bone deterioration; use of glucocorticoids; presence of an active inflammatory disease; presence of an active or suspected infection; and malignancy in the last 12 months pre-fracture. Patients were followed over a period of 12 weeks at four outpatient visits: at 1 to 2 weeks (visit 1), 3 to 4 weeks (visit 2), 6 to 8 weeks (visit 3), and 12 weeks (visit 4) postfracture. The protocol (registration no. NTR3821) was approved by the Medical Ethics Committee of the Maastricht University Medical Center, and all patients provided written informed consent before participation.

Measurements

Clinical outcome measure at 12 weeks postfracture

At visit 4, 12 weeks after the fracture, patients rated their pain and disability using the validated Dutch version of the Patient Rated Wrist Evaluation (PRWE) questionnaire. PRWE was chosen as the primary outcome measure because it is a simple and brief tool that was especially developed for evaluating the outcome in patients with distal radius fractures. It has previously been shown to be a reliable and valid measurement tool after distal radius fractures.[16, 17] The questionnaire consists of a pain subscale (5 items, contributing to 50% of the total score) and a disability subscale (10 items, 50% of total score). The total PRWE score ranged from 0 (no pain/no difficulty) to 100 (worst pain/disability).

In addition, we applied the Quick Disabilities of the Arm Shoulder and Hand (QuickDASH) questionnaire, which is the most widely used questionnaire for patients with general upper-extremity problems. It has been validated in patients with any or multiple disorders of the upper limb.[18-21] QuickDASH uses 11 items to measure physical function and symptoms on a 0 to 100 scale with higher scores representing greater disability. Further, patients were asked to rate their pain on a visual analogue scale (VAS) ranging from 0 (no pain) to 100 (most severe pain). Active range of motion was measured at the fractured side for total wrist dorsal and palmar flexion and total ulnar and radial deviation.

HR-pQCT measurements and micro-finite element analysis (µFEA)

At visit 1 (1 to 2 weeks postfracture) and visit 2 (3 to 4 weeks postfracture), the fractured radius was scanned by HR-pQCT (XtremeCT, Scanco Medical AG, Bruttisellen, Switzerland) using clinical in vivo settings by the manufacturer. The region of interest was based on anteroposterior and lateral radiographs of the fractured radius, in which the proximal edge of the lunate was used as reference. The scan length was set to 18 mm (Fig. 1), which means that this is twice the scan length of the standard protocol. This doubling was needed to make sure that the complete fracture was scanned. With an isotropic voxel size of 82 µm, each HR-pQCT measurement therefore resulted in 220 parallel CT slices with a scan time of 5.6 minutes and a radiation of <6 µSv. Because the patient's forearm was fixed in a cast, the forearm with the cast was placed in a custom cylindrical carbon holder with an inflatable cushion (Pearltec AG, Schlieren, Switzerland) to minimize patient motion. Scan quality was graded according to Pialat and colleagues,[22] a five-grade scheme ranging from grade 1 (no motion artifacts) to grade 5 (severe motion artifacts). Scans with insufficient quality (ie, grade 4 or 5) were repeated.

Figure 1.

Representative region of interest on the HR-pQCT image of the distal radius. The proximal edge of the lunate was used as the reference line. A scan length of 18 mm (220 slices) was selected to make sure that the complete fracture was scanned.

The HR-pQCT images were evaluated using the standard patient evaluation protocol provided by the manufacturer.[23] In short, the cortical and trabecular regions were separated first using a semiautomated contouring scheme in which the periosteal boundary surface of the radius was derived.[24] After contouring, a Laplace-Hamming filter (epsilon 0.5 and cut-off frequency 0.4) followed by normalization (range 0 to 1000) and global thresholding (threshold 400) was used to extract the voxels that represented mineralized bone and to create the segmented images. The following bone density parameters were calculated from the images: volumetric bone mineral density (mgHA/cm3) was assessed for the total region (D.tot), and the trabecular (D.trab) and cortical region (D.cort) separately. The microarchitectural parameters that were used in this study were the trabecular number (Tb.N; 1/mm), thickness (Tb.Th; mm), and separation (Tb.Sp; mm), which are all determined using a 3D ridge extraction method[25] and standard morphological relations. Bone geometry was expressed by the cortical thickness (Ct.Th; mm), which was calculated by dividing the cortical volume by the outer cortical surface.

µFE models were created from the segmented HR-pQCT images. Each voxel that represents bone tissue was converted into a brick element of the same size. In the standard µFE approach usually used in HR-pQCT,[14] equal properties are assigned to every bone tissue element. Whereas this approach has been well validated and successfully applied in clinical studies, its application to fractured bone is not trivial. The use of segmented images limits the models to represent only the mineralized phase. Particularly in the early stages of fracture healing, this approach might be inaccurate because it does not account for newly formed low-mineralized bone. Therefore, we additionally applied a more sophisticated approach, described earlier by Shefelbine and colleagues,[26] that distinguishes between three different tissue types (soft tissue, low-mineralized tissue, and high-mineralized tissue) with stiffness properties of the elements depending on its material composition. This approach has been validated for healing bone.[26] In the resulting µFE model, soft tissue was assigned a constant modulus of 50 MPa, whereas for mineralized bone a Young's modulus ranging from 5000 to 15,000 MPa results, depending on the material composition of the voxel. A Poisson's ratio of 0.3 was assigned to all elements, and all materials were assumed to be linear elastic and isotropic. For the standard µFE approach, a procedure similar to that described in earlier studies was used and all elements were assigned a Young's modulus of 10 GPa and a Poisson's ratio of 0.3.[14, 27] By subsequently applying different load cases, clinical relevant loading scenarios were simulated and the associated stiffness was chosen as outcome variable. The first load case represented a “high-friction” compression test with a prescribed displacement in the axial direction of 1% of the total length, from which the compression stiffness (S.comp; kN/mm) was calculated. The second load case represented a prescribed rotation of 0.01 rad around the longitudinal axis applied to the surface perpendicular to this axis from which the torsional stiffness (S.tors; kNmm/rad) was calculated. A third and fourth load case represented a prescribed rotation of 0.01 rad applied around the sagittal and transversal axes, respectively, onto the surfaces perpendicular to each axis, thus inducing a state of pure bending in two directions, from which the bending stiffness in each direction (sagittal and transversal) was calculated. Because the orientation of the patient's forearm was not the same during all visits, a quadratic mean bending stiffness was calculated (S.bend; kNmm/rad) from the sagittal and transversal bending stiffness.

In addition to HR-pQCT, and according to the Dutch standard care when a patient over the age of 50 years comes to a clinic with a fracture, BMD T-score at the femoral neck was assessed by dual-energy X-ray absorptiometry (DXA; Hologic Discovery A; Hologic Inc., Waltham, MA, USA). BMD T-score was assessed once between 4 and 8 weeks postfracture or was taken from a previous DXA scan if the patients already had had a DXA scan in the last year before the fracture.

Serum markers of bone turnover and inflammation, and vitamin D

At visit 1 (1 to 2 weeks postfracture) and visit 2 (3 to 4 weeks postfracture), venous blood samples were collected by puncture of the antecubital vein. Serum was separated from the blood clot within 20 minutes and freezed to −20°C until analysis. Serum bone markers representing bone resorption (carboxy-terminal telopeptide of type I collagen [ICTP]) and bone formation (procollagen type-I N-terminal propeptide [PINP]) were assessed by radioimmunoassay (UniQ PINP/ICTP; Orion Diagnostics, Espoo, Finland) in a certified laboratory. Additionally, fracture-induced inflammation was assessed by high-sensitive C-reactive protein (hsCRP) (BN ProSpec system; Siemens/Dade Behring, Liederbach, Germany). An additional blood sample was taken at visit 2 to analyze 25-hydroxyvitamin D (25(OH)D) by chemiluminescent immunoassay (IDS-iSYS Immunodiagnostic Systems GmbH, Frankfurt, Germany).

Statistical analysis

Patients' demographic characteristics, bone parameters, and clinical outcomes at 12 weeks postfracture were reported in means and their standard deviations. Changes in bone parameters between visit 1 and visit 2 were tested using a paired samples Student's t test. A linear regression was used to determine if the clinical outcome at 12 weeks was related to patients' or fractures' characteristics, such as fracture type (intra- versus extra-articular, comminuted versus noncomminuted, reduced versus nonreduced), fracture side (dominant versus nondominant), DXA T-score at femoral neck, or serum vitamin D level (≥75 nmol/L versus <75 nmol/L). The relationship between early changes in bone parameters and 12-week clinical outcomes was calculated using a linear regression model with the clinical outcome at 12 weeks postfracture as outcome, and the change in bone parameters between visit 1 and visit 2 as explanatory variable. The model was adjusted for the number of days between visit 1 and visit 2. Because adjusting the model for baseline bone values (ie, bone parameters at visit 1) did not improve the model according to the Akaike information criterion, we did not include baseline values in the model. In a second step, we additionally adjusted the model for fracture type (intra- versus extra-articular, comminuted versus noncomminuted, reduced versus nonreduced), fracture side (dominant versus nondominant), DXA T-score at femoral neck, or serum 25(OH)D level (≥75 nmol/L versus <75 nmol/L) to test whether these factors influenced the model. All analyses were performed using Stata 11.0 (StataCorp, College Station, TX, USA), and the significance level was set to α = 0.05.

Results

Patient characteristics

From the 20 patients who were initially included in the study, one patient declined participation at the first visit and another patient at the second visit. Thus, 18 women (aged 64.3 ± 8.4 years) completed the study. The characteristics of these 18 patients, including their fracture characteristics, are listed in Table 1. Mean DXA T-score at femoral neck was −2.2 (0.9) with 7 (44%) osteoporotic and 8 (50%) osteopenic women. Two patients had no DXA data. Serum 25-hydroxyvitamin D levels were 62.8 (29.3) nmol/L, with 10 (46%) patients having values below 75 nmol/L. The clinical outcome in terms of pain and disability, as assessed at visit 4 (85.4 (range 81 to 91) days postfracture), are also presented in Table 1. According to the PRWE classification of MacDermid and colleagues,[9] 5 (28%) patients reported experiencing no or minimal pain and disability, whereas 3 (17%) patients had mild, 5 (28%) patients moderate, 3 (17%) patients severe, and 2 (11%) patients very severe pain or disability, respectively, 12 weeks after the fracture. In 1 patient, VAS at 12 weeks postfracture is missing, and in 2 patients, the range of motion test could not be performed. HR-pQCT measurements were performed 9.1 (4 to 16) days (visit 1) and 23.8 (18 to 29) days after the fracture (visit 2) with the corresponding values shown in Table 2. Because of motion artefacts, 36% of the scans had to be repeated. Even if after repetition, 5 scans were still of insufficient quality, they were included into analysis. A representative image of a HR-pQCT and a FEA model is given in Fig. 2. One patient had missing ICTP, P1NP, and hsCRP data.

Table 1. Characteristics of 18 Postmenopausal Women With a Stable Distal Radius Fracture and Their Clinical Outcomes 12 Weeks After the Fracture
PatientAge (years)Fracture characteristicClinical outcomes at 12 weeks postfracture
Dominant sideIntra-articularComminutedReduction neededPRWE (counts)QuickDASH (counts)VAS (counts)Dorsal/palmar flexion (°)Radial/ulnar deviation (°)
  1. DXA = dual-energy X-ray absorptiometry; PRWE = Patient-Rated Wrist Evaluation Score; QuickDASH = Disabilities of the Arm, Shoulder and Hand Score; VAS = visual analogue scale of pain; NA = missing data.
  2. Total values are means (SD), respectively, n (%) “yes” for fracture characteristics. DXA T-score at femoral neck.
158.3NoNoNoNo10.527.50NANA
255.4NoNoNoNo59.543.2707545
364.5NoYesYesYes45.547.5417550
452.5YesNoNoNo0.00.009075
566.9YesNoNoYes42.57.5707035
672.0NoYesYesYes19.512.5208065
767.5NoNoNoNo31.534.1NA12085
850.1YesNoNoNo55.060.0507560
980.5NoYesYesNo68.047.5396040
1061.6YesNoNoYes64.550.0518550
1169.1NoNoNoYes21.036.4277035
1254.8YesYesNoYes99.097.7100NANA
1364.2YesNoNoYes84.077.3824535
1463.9NoNoYesYes27.929.597555
1568.7NoNoNoYes52.538.65013050
1675.9NoYesYesYes66.043.2282050
1757.9NoNoNoNo0.00.0017090
1874.5NoNoYesNo13.010.0010560
Total64.3 (8.4)6 (33%)5 (28%)6 (33%)10 (56%)42.2 (28.7)36.8 (25.9)38 (31)84 (35)55 (17)
Table 2. Bone Parameters at the Fracture Location Assessed by HR-pQCT, Calculated Biomechanical Parameters and Serum Bone Markers at 1 to 2 Weeks (Visit 1), 3 to 4 Weeks (Visit 2), and the Change Between Visit 1 and Visit 2 After a Stable Distal Radius Fracture
 Visit 1 (1 to 2 weeks postfracture)Visit 2 (3 to 4 weeks postfracture)Change between visit 1 and visit 2
  1. D.tot = total bone mineral density; D.cort = cortical bone mineral density; D.trab = trabecular bone mineral density; Ct.Th = cortical thickness; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; S.comp = compressional stiffness; S.tors = torsional stiffness; S.bend = bending stiffness; ICTP = carboxy-terminal telopeptide of type I collagen; PINP = procollagen type-I N-terminal propeptide; hsCRP = high-sensitive C-reactive protein.
  2. Values are expressed in means and their standard deviations for visit 1 and visit 2, and as mean (95% confidence interval) for the change between the two visits.
  3. aRefers to significant changes between visit 1 and visit 2.
Density parameters
D.tot (mgHA/cm3)232.2 (54.8)243.4 (50.4)11.2 (1.8 to 20.6)a
D.cort (mgHA/cm3)648.1 (123.4)629.3 (113.9)−18.8 (−32.7 to −4.8)a
D.trab (mgHA/cm3)144.8 (26)160.9 (29)16.2 (8.7 to 23.6)a
Microstructural parameters
Ct.Th (mm)0.408 (0.216)0.4 (0.193)−0.008 (−0.039 to 0.022)
Tb.N (mm-1)1.995 (0.455)2.029 (0.423)0.034 (−0.064 to 0.131)
Tb.Th (mm)0.063 (0.015)0.068 (0.013)0.005 (0.001 to 0.009)a
Tb.Sp (mm)0.468 (0.131)0.445 (0.099)−0.023 (−0.054 to 0.009)
Biomechanical parameters (standard protocol)
S.comp (kN/mm)14.8 (7.2)13.0 (7.8)−1.8 (−3.6 to 0.03)
S.tors (kNmm/rad)298 (135)272 (120)−26 (−69 to 18)
S.bend (kNmm/rad)430 (227)379 (213)−51 (−127 to 26)
Serum bone markers
ICTP (µg/L)3.6 (1.1)4.1 (1.1)0.5 (−0.1 to 1.0)
PINP (µg/L)48.9 (25.7)65.9 (39.1)12.9 (−3.7 to 29.5)
hsCRP (mg/L)3.8 (6.3)2.6 (4.5)−1.2 (−2.8 to 0.4)
Figure 2.

Representative image of a HR-pQCT and FEA model.

Relation between baseline characteristics and clinical outcome 12 weeks postfracture

PRWE total score or QuickDASH score 12 weeks postfracture were not related to baseline age, fracture side (dominant versus nondominant side) and type (intra-articular, comminuted, or reduced fracture), DXA T-score or 25(OH)D. VAS was higher in patients with a fracture at the dominant side (33.0 [95% CI 3.4 to 62.6] VAS counts; p = 0.031; adjusted R2 = 0.23) compared with the nondominant side. Lower palmar flexion range of motion was found after an intra-articular fracture (−23.3 [−46 to −0.7] angular degrees; p = 0.044; R2 = 0.21), and lower total ulnar and radial deviation range of motion after a reduced fracture (−17.8 [−233.7 to −1.8] angular degrees; p = 0.032; R2 = 0.24).

PRWE score

The relationship of PRWE at 12 weeks and the early changes in bone parameters is shown in Table 3 and Fig. 3. Higher early increases in D.trab were significantly related to lower, thus better, PRWE scores at 12 weeks postfracture (p = 0.022; adjusted R2 = 0.37). For the other density parameters (D.cort and D.tot), the relationship was not significant.

Table 3. Relationship Between the PRWE Score at 12 Weeks Postfracture as a Measure for Pain and Disability, and the Changes in Bone Parameters at the Fracture Location Assessed by HR-pQCT, Calculated Biomechanical Parameters, and Serum Bone Markers Between 1 to 2 Weeks (Visit 1) and 3 to 4 Weeks (Visit 2) After a Stable Distal Radius Fracture
 β-coefficient (95% confidence interval)p ValueAdj R2
  1. D.tot = total bone mineral density; D.cort = cortical bone mineral density; D.trab = trabecular bone mineral density; Ct.Th = cortical thickness; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; S.comp = compressional stiffness; S.tors = torsional stiffness; S.bend = bending stiffness; ICTP = carboxy-terminal telopeptide of type I collagen; PINP = procollagen type-I N-terminal propeptide; hsCRP = high-sensitive C-reactive protein; PRWE = Patient-Rated Wrist Evaluation Score.
  2. aRefers to a statistically significant relation.
Density parameters
D.tot−0.61 (−1.28 to 0.06)0.0730.27
D.cort−0.19 (−0.68 to 0.31)0.4380.13
D.trab−0.96 (−1.75 to −0.16)0.022a0.37
Microstructural parameters
Ct.Th−91 (−323 to 140)0.4120.13
Tb.N−48.8 (−116 to 18.4)0.1420.22
Tb.Th−352 (−2089 to 1386)0.6720.10
Tb.Sp209 (15 to 402)0.037a0.33
Biomechanical parameters (standard protocol)
S.comp−2.44 (−6.18 to 1.31)0.1860.19
S.tors−0.145 (−0.285 to −0.004)0.045a0.31
S.bend−0.064 (−0.15 to 0.022)0.1320.22
Serum bone markers
ICTP12.1 (0.0 to 24.1)0.05a0.34
PINP0.256 (−0.175 to 0.687)0.2230.21
hsCRP−0.263 (−5.02 to 4.495)0.9070.12
Figure 3.

Relation between the change in (A) trabecular density (Dtrab), (B) trabecular separation (TbSp), (C) torsional stiffness (Stors), (D) bone resorption (carboxy-terminal telopeptide of type I collagen [ICTP]) between visit 1 (1 to 2 weeks postfracture) and visit 2 (3 to 4 weeks postfracture) and patient-rated pain and disability (PRWE score) 12 weeks after a distal radius fracture.

On the microarchitectural level, lower PRWE was related to higher decreases in Tb.Sp (p = 0.037; R2 = 0.33) but not to changes in Tb.N or Tb.Th.

On the biomechanical level, lower PRWE score was related to higher increases in S.tors (p = 0.045; R2 = 0.31) but not to changes in S.comp nor for S.bend. When applying the Shefelbine method, the relation between S.tors and PRWE was −0.134 (−0.290 to 0.022; p = 0.086; R2 = 0.26).

On the serum bone marker level, higher PRWE was related to higher increases in ICTP (p = 0.050; R2 = 0.34) but not to changes in PINP or hsCRP.

Adjusting the analysis for fracture type (intra-articular fracture, comminuted fracture, or reduced fracture) did not influence the results; however, including fracture side (dominant versus nondominant hand) into the model increased adjusted R2 to 0.45 for the relation with early changes in D.trab (−1.03 [−1.78 to −0.27]; p = 0.011) with higher PRWE values for patients with a fracture at the dominant side (pfx-side < 0.1). The 25(OH)D levels did not influence the model, nor did the DXA T-scores. Adjusting the analysis for baseline bone parameter did not influence the results.

Excluding the 5 patients who had one scan with insufficient quality (subsample of 13 patients), beta-coefficients for early changes in D.trab, Tb.Sp, S.tors, and ICTP remained at the same level or became stronger. However, the associations were not significant anymore.

QuickDASH score

QuickDASH and PRWE score 12 weeks postfracture were highly correlated (r = 0.89). QuickDASH was related to changes in D.trab (−0.789 [−1.49 to −0.08]; p = 0.031; R2 = 0.40), Tb.Sp (174 [5.7 to 343]; p = 0.044; R2 = 0.37), ICTP (13.0 [3.4 to 22.6]; p = 0.012; R2 = 0.49), and S.tors applying the Shefelbine method (−0.138 [−0.266 to −0.010]; p = 0.036; R2 = 0.39).

Visual analogue scale

VAS was highly correlated to PRWE score and QuickDASH score 12 weeks postfracture (r = 0.88 and r = 0.76, respectively). VAS score was not significantly related to early changes in bone, biomechanical, or serum bone marker parameters. After correction for fracture side, VAS was significantly related to early changes in Tb.Sp (226 [26.5 to 425]; p = 0.029; R2 = 0.41).

Range of motion

Range of motion at 12 weeks was significantly correlated with PRWE score, QuickDASH, and VAS (all r < −0.57). The total palmar and dorsal flexion was related to early changes in hsCRP (9.62 [3.90 to 15.34]; p = 0.003; R2 = 0.52) (Fig. 4), but range of motion did not appear to be related to changes in bone density, architecture, and biomechanical parameters. The total radial-ulnar deviation range of motion was not related to early changes in bone parameters, including hsCRP. After correction for intra-articular or reduced fractures, the effects remained the same.

Figure 4.

Relation between the change in high-sensitive C-reactive protein (hsCRP) between visit 1 (1 to 2 weeks postfracture) and visit 2 (3 to 4 weeks postfracture) and total active dorsal and palmar flexion range of motion 12 weeks after a distal radius fracture. Two patients have missing range of motion data, and one patient does not have hsCRP data.

Discussion

The most striking finding in this study is that the changes in bone microarchitecture, calculated bone stiffness, bone resorption, and inflammation during the first weeks of fracture healing predicted functional outcome parameters assessed at 12 weeks after a distal radius fracture. PRWE scores 12 weeks after the fracture were explained for more than 30% by changes in D.trab, S.tors, T.Sp, and ICTP between days 9 and 23 after fracture. This indicates that early changes in bone structure and remodeling are important for clinical recovery; hence, a greater increase in D.trab and S.tors and greater decrease in bone resorption are related to a better long-term outcome for pain and function.

Each 10 mgHA/cm3 higher early increase in D.trab between visit 1 and visit 2 was related to a lower PRWE by 9.6 counts and a lower QuickDASH score by 7.9 counts, respectively. Because longitudinal changes of ≥14 points for PRWE and QuickDASH scores are considered minimal clinically significant changes,[28] the wide spectrum of early changes in D.trab ranging from −13.8 to +40.6 mgHA/cm3 may be considered as clinically meaningful. Reasons for smaller initial increases or even decreases in D.trab after a fracture, however, remain unclear. One possibility is that either these patients' healing process has already started with increased remodeling or that the process is delayed. The latter could explain persistent or enhanced pain and disability 12 weeks postfracture in these patients. In this respect, it would be interesting to see whether pain and disability will improve in these patients at later follow-up or whether this patient group will have substantially more persisting symptoms.

Each 1 µg/L higher early increase in ICTP was related to a higher, thus worse, PRWE by 12.1 counts at 12 weeks. During fracture healing, both markers of bone formation and resorption increase and can be used to reflect the stages of fracture healing and mineralization process.[29-34] It has been shown that serum bone resorption markers, such as ICTP, rise at an earlier stage than bone formation markers, indicating that during the very first weeks after a fracture, osteoclast activation–driven resorption occurs as the main process.[31, 32] This could be an explanation of why, in our study, the early increase in ICTP and not PINP was related to pain and disability at follow-up. It has been reported that patients with a delayed fracture healing tended to have higher increases in bone resorption markers in the first weeks postfracture.[33] Fracture extend and location may also increase ICTP response after a fracture with higher increases in more complex fractures or larger fracture areas.[32, 34] However, more complex fractures, in our study represented by intra-articular, comminuted, and/or fractures where a manual reduction was necessary, had no influence on the results.

Interestingly, early changes in hsCRP did not contribute to PRWE and QuickDASH score, suggesting that early changes in inflammation do not contribute to long-term pain and disability. In contrast, active range of motion was explained by 40% to 60% by early changes in hsCRP, indicating that early increase in inflammation resulted in better active motion at 12 weeks postfracture. Inflammation is considered as an important primary stimulus for fracture repair.[35, 36] The beneficial effect of inflammation during fracture healing is also supported by studies showing that anti-inflammatory drugs may delay fracture healing.[36]

On the other hand, no significant predictors of VAS were identified. Whereas PRWE and QuickDASH questionnaires both consist of questions about pain and functionality, VAS only includes information about patient-perceived pain. However, using the pain subscale of the PRWE questionnaire as outcome (data not shown), the results remained the same as for the total PRWE score. A possible explanation for the different results when using VAS could be that the VAS evaluates pain intensity, but not frequency, whereas PRWE also included frequency-related questions. Additionally, perceived pain is determined by many more factors than the healing process of the mineralized phase of bone after a fracture as assessed by HR-pQCT only. Further, differences in perceived pain could have influenced the clinical outcome at 12 weeks, and also the early changes in bone parameters, as patients with a higher pain tolerance might use their fractured arm more extensively. VAS score at the contralateral side could be an indicator of general pain tolerance. However, it did not influence our results (data not shown).

These results suggest that the integrated process of inflammation, regeneration, and remodeling that occurs during the early stages of fracture healing can be studied by the assessment of hsCRP, bone resorption markers, and bone structure by HR-pQCT to predict clinical outcome 3 months after a stable distal radius fracture.

More than half of all patients (56%) reported they still experienced moderate to very severe pain and disability 12 weeks after the fracture. A painless, full restoration of hand functionality is seen as the major goal after a distal radius fracture.[37] Knowing more about fracture-related predictors for clinical outcomes is important to identify patients at risk for having long-term complications. Once they are identified, it could become possible to treat them. Interestingly, patient and injury-related characteristics—as our data also confirmed—play only a small role in predicting the clinical outcome after a fracture.[10, 11] Although injury compensation,[10, 11] patients' education level,[10, 11] the presence of other medical problems,[11] and prereductional radial shortening[10] together explained 17% of the 1-year[11] and 25% of the 6-month[10] PRWE score, respectively, it is less consistently reported whether more injury-related parameters (ie, initial displacement of the fracture) predict clinical outcomes after a distal radius fracture.[13, 38] To our knowledge, no previous study evaluated the predictive values of the fracture healing process itself on pain and disability after a distal radius fracture. However, the reason why some patients show a greater increase in D.trab or decrease in ICTP, respectively, and whether this will lead to a more rapid consolidation and thus improved clinical recovery, remains unclear. The amount of structural changes and bone marker response could depend on the complexity of the fracture. However, in our study including conservatively treated fractures only, the associations between early changes in bone parameters and functional outcomes at 12 weeks were independent of articular involvement (intra- versus extra-articular fracture), comminution (comminuted versus noncomminuted fracture), and fracture reduction. Furthermore, the correlations between early bone changes and pain and disability were independent of the patient's vitamin D level or DXA T-score.

Our study has several strengths. Compared with other studies evaluating the clinical outcomes after distal radius fractures, the patients and fracture type in the current study are more homogeneous. We only included postmenopausal women with a stable, conservatively treated distal radius fracture. Of course, this reduces generalization of the results to other, more complex fractures, but it gives a clearer picture of what happens in conservatively treated fractures in mainly osteopenic or osteoporotic women. Further, this is the first study to our knowledge using HR-pQCT measures for the evaluation of the fracture healing process and its relationship to the clinical outcomes. The HR-pQCT measurements as well as the µFEA were obtained using the standard evaluation protocols provided by the manufacturer. Further, we used well-validated and widely used questionnaires that are recommended when evaluating clinical outcomes of distal radius fractures.[37]

Apart from these strengths, some limitations have to be mentioned. With 18 patients in the study, statistical power is limited, and therefore, our preliminary findings should be confirmed in larger studies. Because of this limited sample size, it was not feasible to analyze all bone parameters simultaneously to quantify the relative importance of predicators. Further, one should be careful with the interpretation of the derived structural parameters, such as the trabecular thickness, because the presence of woven bone between the trabeculae might affect the calculation of these parameters in an unexpected way. Additionally, our “baseline” value was assessed 9.1 days after the fracture during the first follow-up visit, which does not allow us to set these early changes in relation to bone density and structure before nor immediately after the fracture. Also, clinical outcomes were assessed at 3 months postfracture, whereas it would be interesting to know also the clinical outcomes at a longer period postfracture.

Once our results are confirmed in a larger population, future research should focus on whether certain interventions during fracture healing, for example, an immediate vitamin D or calcium supplementation, could improve fracture healing on a structural level and thereby maybe reduce patients' long-term pain and disability. To identify patients at risk, further research should also test whether HR-pQCT is needed to evaluate these early changes in D.trab after a fracture, or if a clinical QCT with poorer resolution, and/or other serum markers have the same power to predict clinical outcomes.

In conclusion, our explorative study indicated that early changes in trabecular BMD, calculated torsional stiffness, bone resorption marker ICTP, and hsCRP after a stable distal radius fracture are related with pain and disability after 12 weeks. If larger studies confirm these results, this suggests that assessment of these early changes during fracture healing could provide valuable information regarding the long-term clinical outcome of these fractures.

Disclosures

BvR is a consultant for Scanco Medical AG. All other authors state that they have no conflicts of interest.

Acknowledgments

The authors thank Liesbeth Jutten, Margareth Winants, and Kevin Görritz for their help in patient recruitment.

The study (Dutch trial register NTR3821) was supported by The Weijerhorst Foundation WH2. UM received a fellowship for prospective researchers by the Swiss National Science Foundation for this project.

Authors' roles: Study design: CA, PB, PM, BvR, JvdB, PG, and PW. Study conduct: JdJ, SB, CA, TvG, JvdB, PG, and PW. Data collection: JdJ, SB, CA, PM, and PW. Data analysis: UM, JdJ, AK, and BvR. Data interpretation: UM, JdJ, CA, JvdB, PG, and PW. Drafting manuscript: UM and JdJ. Revising manuscript content: SB, CA, PB, PM, TvG, BvR, JvdB, PG, and PW. Approving final version of manuscript: UM, JdJ, SB, AK, CA, PB, PM, TvG, BvR, JvdB, PG, and PW. UM takes responsibility for the integrity of the data analysis.

Ancillary