Effect of slow-release sodium fluoride on cancellous bone histology and connectivity in osteoporosis

Bone, Vol . 15, No . 6, pp. 691-699,1994

Copyright ® 1994 Elsevier Science Ltd Printed in the USA . All rights reserved

Pergamon

8756-3282194 $6 .00 + .00

8756-3282(94)00042-5

Effect of Slow-Release Sodium Fluoride on Cancellous Bone Histology and Connectivity in Osteoporosis J . E . ZERWEKH, 1 H . K . HAGLER, 2 K . SAKHAEE,' F . GOTTSCHALK,' R . D . PETERSON,' and C . Y . C . PAK' ' Center for Mineral Metabolism and Clinical Research 2

Department of Pathology 3 Department of Orthopedic Surgery, University of Texas Southwestern Medical Center at Dallas, Dallas, TX, USA Address for correspondence and reprints : Joseph E . Zerwekh, PhD, Center for Mineral Metabolism and Clinical Research, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd ., Dallas, TX 75235-8885, USA .

Abstract

Introduction

We have previously demonstrated that a treatment regimen of slow-release sodium fluoride (SRNaF) and continuous calcium citrate increases lumbar bone mass, improves cancellong bone material quality, and significantly reduces vertebral fracture rate in osteoporotic patients . In order to assess whether such treatment also improves trabecular structure, we quantitated cancellous bone connectivity before and following 2 years of therapy with SRNaF in 23 patients with osteoporosis and vertebral fractures . In addition, we performed bone histomorphometry on the same sections used for connectivity measurements . There was a significant Increase in L2-L4 bone mineral density during therapy (0 .827±0 .176 g/cm2 SD to 0 .8721-0.166, p = 0.0004) . Significant histomorphometric changes were represented by increases in mineral apposition rate (0.6±0 .4 lLm/d to 1 .1±0 .7, p = 0 .0078) and adjusted apposition rate (0.4±0 .3 p.m/d to 0 .6±0 .4, p = 0 .016) . On the other hand, trabecular spacing significantly declined (from 1375±878 p.m to 1052±541, p = 0 .05) . Two-dimensional quantitation of trabecular struts on iliac crest histological sections disclosed significant increases in mean node number per mm' of canceBous tissue area (0 .22±0 .12 vs . 0 .39±0 .27, p = 0.0077), the mean node to free-end ratio (0 .23±0.21 vs . 0 .41±0.46, p < 0 .05), and in the mean node to node strut length per mm 2 of cancellous area (0.098±0 .101 vs . 0 .212±0.183, p < 0.01) . There were no significant changes in any of the measurements associated with free-end number or free-end to freeend strut length . When patients were divided into those with severe and mild-modest spinal bone loss (based upon initial lumbar bone density) the significant changes in connectivity occurred in patients with mild-moderate bone loss, but not in those with severe bone loss, suggesting that fluoride's effect is in part dependent on the presence of a certain critical amount of bone . This finding in combination with the previously reported increases in bone mass and bone material quality may explain the significant reduction in vertebral fracture rate observed with this particular fluoride regimen .

Trabecular bone structure is an important determinant of bone strength and may be useful in gauging risk of developing fractures and in assessing response to treatment in osteoporotic disease states . In recent years, a number of different approaches to the study of bone structure have been described, including measurement or calculation of trabecular width, separation, and density (Parfitt et al . 1983), assessment of the trabecular and marrow star volume (Vesterby 1990), trabecular bone pattern factor (Hahn et al . 1992), reflection ultrasound (Antich et al . 1991), and high resolution microfocus computed tomography (Feldkamp et al . 1989) . All of these techniques provide a measure of bone structure and/or quality but provide little or no information on trabecular connectivity . In addition, most of these methods require specialized equipment and facilities not readily found in all bone research laboratories . Recently, two methods have been described which permit quantitation of trabecular connectivity on standard histological sections from iliac crest bone biopsies . Both methods rely on the identification of different strut types as well as quantitation of trabecular nodes and free ends . In the original description (Garrahan et al . 1986), assessment of trabecular connectivity again required the use of sophisticated hardware and specialized software not generally available to all investigators . By refining this technique, Mellish et al . (Mellish et al . 1991) described a manual method for assessing two-dimensional cancellous bone connectivity which utilized commonly available equipment . In the present study, we have utilized this relatively simple manual method to assess the effect of 2 years of treatment with intermittent slow-release sodium fluoride (SRNaF) and continuous calcium citrate on trabecular connectivity in 23 osteoporotic patients . In addition, we compared the changes in cancellous bone connectivity to the biochemical, histological and clinical responses observed in these patients to further assess the effects of this therapeutic regimen on skeletal integrity . Materials and Methods Subjects

Key Words : Fluoride-Connectivity-Trabeculae-StrutsOsteoporosis-Bone .

The inclusion criteria for entry into this study was demonstration of at least one vertebral compression fracture for both women 691

692

and men . Women must have had a natural or surgically-induced menopause while the men must have been eugonadal . One hundred and ten such patients were recruited into a randomized placebo-controlled trial of SRNaF in the management of postmenopausal osteoporosis (Pak et al . 1994) . None of the patients had hyperadrenocorticism, primary byperparathyroidism, renal tubular acidosis, thyrotoxicosis, multiple myeloma, renal failure, or liver disease, and none had a prior history of taking anticonvulsants, glucocorticoids, estrogen, fluoride, calcitonin, excessive amounts of alcohol, or pharmacological doses of vitamin D preparations . None had active peptic ulcer disease . All patients were admitted to the General Clinical Research Center for biochemical as well as radiological assessment prior to procurement of bone biopsy . Trans-iliac biopsies were obtained in 26 of these fluoridetreated patients with spinal osteoporosis suffering from vertebral compression fractures as previously described (Zerwekh et at . 1992) . In three patients, bone biopsies were either broken or of insufficient quality so as to exclude cancellous connectivity measurements . Thus, in the present study, the patient population comprised 17 women and 6 men with a mean age of 55 years (range 28-77 years) . Fifteen women were postmenopausal and two had early oophorectomy : the six men had idiopathic osteoporosis . Treatment schedule

All 23 patients received cyclic fluoride treatment in the form of SRNaF (Slow Fluoride, Mission Pharmacal Co ., San Antonio, TX ; sodium fluoride incorporated in a wax matrix tablet to assure delayed release) . Seventeen patients took 25 mg of SRNaF twice daily, with 50 µg 25-hydroxyvitamin D, (25(OH)D,) twice weekly and sufficient calcium supplementation (as calcium citrate in divided doses) to bring the daily calcium intake to 1500 mg/day . This form of sodium fluoride has been previously shown to maintain serum fluoride within 95-190 ng/ml (believed to be the therapeutic window), while providing safety of usage (Pak et al . 1986a, 1986b, 1989, 1990) . Both the 25(OH)D, and calcium supplementation were administered continuously throughout the study . SRNaF was taken daily for 3 months followed by 8 weeks of a fluoride-free period to complete one 5-month cycle . Each patient completed four such cycles for a total treatment interval of 20 months . The remaining six patients took SRNaF, 25 mg, with calcium citrate, 400 mg calcium, before breakfast and at bedtime for 12 months . Fluoride was then discontinued for I to 2 months while maintaining calcium citrate supplementation . The cycle was repeated so that each patient received this treatment regimen for 26 to 28 months . Bone mass measurement

Bone mineral density of L2-L4 vertebrae was quantitated in each patient by dual photon x-ray absorptiometry or dual energy x-ray absorptiometry (Lunar Radiation, Madison, WI) . At baseline, the extent of spinal bone loss was determined by comparing the patient's value to the mean value for a 30-year-old woman or man established for the same instrument . In so doing, it was possible to divide the patient population into two groups . Patients with initial bone mineral density values less than 65% of the 30-year-old normal mean were considered to have "severe" bone loss . Patients with initial bone mineral density values greater than or equal to 65% of the 30-year-old normal mean comprised the group with a "mild-moderate" bone loss . Bone mineral density measurements were then repeated in each patient upon completion (approximately 2 years) of the SRNaF treat-

J . B. Zerwekh et al . : NaF improves trabecular connectivity ment regimen described above in order to calculate the percent change in bone mineral density . Qu antitation of spinal fractures

Lateral spinal films taken before treatment and at completion of approximately 2 years of fluoride therapy were compared in order to determine new and recurrent fractures occurring during treatment . For each vertebra from T3 to L5, landmarks (anterior and posterior corners and midpoints) were recorded using an electrostatic digitizing board (Scriptel Corporation, Columbus, OH) with a coefficient of variation of 1 .5% . A computer software program was used to compute changes in vertebral heights and area, as well as to calculate magnification error between the two sets of x-rays . After correction for magnification error if any, a reduction in height of more than 20% of anterior, middle, or posterior height, accompanied by a reduction in area of more than 10% in a previously unaffected vertebra, was considered a ''new" fracture (Watts et al . 1990), or if in a previously fractured vertebra to be a "recurrent" fracture . Biochemical analysis

Blood screen was performed as SMA-20 (Smithkline Lab, Dallas, TX) . Skeletal fluoride content was determined on 100-µm thick sections of the biopsy specimens after ashing as previously described (Boivin et al . 1988) . Bone biopsy analysis

Two full-thickness transcortieal iliac crest bone biopsies were performed under local anesthesia with a 7 .5-mm trochar needle . The first was done before the study, and the second (taken on the opposite side), following at least 30 days of fluoride withdrawal after receiving the drug intermittently for 20 or 26-to-28 months as described above . All patients received two courses of declomycin (250 mg three times/day for 2 days for the first label and again after a 10-day interval for 4 days) before the biopsy . Bone biopsies were obtained 5 days after the last dose of declomycin . The core biopsies were stained with Villanueva Osteochrome (Polysciences, Inc ., Warrington, PA) before dehydration and subsequent embedding in methylmethacrylate . All sections were cut at 10 p m on a Jung model K microtome (Jung, Heidelberg, Germany) . Sections for light microscopy were stained with toluidine blue . Static parameters were quantitated from these sections while dynamic indices were assessed from the Villanuevastained sections . Quantitative histomorphometry was performed on all samples, using cursor tracings automatically transmitted to an IBM PS/2 computer and analyzed by a program (R & M Biometrics, Nashville, TN) designed to calculate values for the desired area and perimeter parameters . The histomorphometric measurements and calculations were based on modifications of previously published methods (Frost 1983) using a correction for section obliquity . The nomenclature of the measured and calculated variables is that approved by the American Society for Bone and Mineral Research (Parfitt et al . 1987) . Trabecular number (Tb .N) and separation (Tb .Sp) were calculated according to Parfitt et al . 1983 . All bone biopsies were quantitated by the same observer who was blinded to the identity of the biopsy specimens . Twvo-dimensional analysis of trabecular structures

The toluidine blue stained sections were mounted on a light box and viewed by a CCD camera using a wide-angle lens, enabling

J . E . Zerwekh et al . : NaF improves trabecular connectivity the entire section to appear in a single field of view . The grey level image of the bone section was captured in a frame stored in a Macintosh II equipped with a data translation quick-capture video board (Fig . la). The image was converted to a binary image by interactive thresholding . After three erosions and three dilations of the binary image, most of the noncalcified marrow elements were removed (Fig . Ib) . Next, the cancellous area of interest was delineated by drawing lines along the upper and lower boundaries . This step also served to remove artefactual bone dust and any shattered pieces of bone from the measurement field . The corticomedullary junction was then delineated by hand using a mouse cursor. The resulting cancellous area of interest was then clipped from the remaining image and the area automatically determined (Fig . le) using NTH Image software (Version 1 .50, NIH, Bethesda, MD) . Once the cancellous area was outlined, the nodes and free ends within the area were identified and counted . A node (N) was defined as a point at which three or more trabeculae joinedA free end (F) was defined as the point at which a trabecula was not joined to any other trabecula . A trabecular strut was defined in the same way as previously reported (Garrahan et al . 1986 ; Mellish et al . 1991) . In short, any part of the trabecular bone that connected any combination of the node, free end, or cortex was counted as a trabecular strut . Struts intersecting either upper or tower boundary were excluded from the analysis . Using this criteria, the following strut types were defined : node to node (NN), free end to free end (FF), node to free end (NF) and cortex struts (CXS) . To overcome difficulties in defining a node and free end when the strut was short, the criterion defined by Mellish et al . (1991) was employed . Connectivity measurements were made on 3 serial sections from 18 patients and from 2 serial sections in the remaining 5 . Statistics Paired-t analysis was used for comparing the differences between baseline and post-therapy biopsies when data was normally distributed . For non-normally distributed data, the Wilcoxon signed rank test was used. The Wilcoxon rank-sum test was used to assess whether significant differences of connectivity measures existed between patients with severe (borderline L2-L4 bone density <65% of 30-year-old normal mean) and mild-moderate (a65%) bone loss prior to and following therapy . In addition, this same test was utilized for examining whether significant differences existed in the mean percent change of the histomorphometric and connectivity parameters between severe and mildmoderate bone loss groups . The percent change was calculated as the difference between post- and pretherapy values divided by the pretreatment value . Statistical analysis was performed using BMDPx, statistical software . Results

Biochemical The effects of 2 years of intermittent SRNaF and continuous calcium citrate therapy on serum parameters are summarized in Table I . Prior to therapy, all patients had normal serum calcium and phosphorus . Mean serum alkaline phosphatase was mildly elevated . There were no significant changes for any of the parameters following therapy, although mean serum alkaline phosphatase fell to within normal limits . Also shown in Table I are the bone fluoride results . Following therapy, there was a significant increase in skeletal fluoride content from 0 .10 ± 0 .06% to 0 .18 ± 0 .08% (p < 0 .01) .

69 3 Clinical Table II summarizes the changes in L2-L4 bone mineral density during 2 years of SRNaF and calcium citrate therapy . For all patients, there was a significant and positive increase in BMD (mean increase of 5 .4%) . When patients were divided into those with severe bone loss (pre-BMD <65% of the 30-year-old normal value) and mild-moderate bone loss (>65% of 30-year-old normal value), a significant increase in BMD was observed for both groups . Patients with severe bone loss demonstrated a greater mean percent change in BMD than those with mildmoderate bone loss (8 .0 vs . 3 .4%), but this difference was not significant . For all 23 patients, there were nine new vertebral fractures in four patients following therapy for a group vertebral fracture rate of 0 .196/patient cycle . When patients were divided into those with severe or mild-moderate bone loss groups, it was observed that all 9 fractures occurred in patients classified as severe bone loss (data not shown) . Histomorphometric For cancellous bone, there were significant increases in mineral apposition rate and in adjusted apposition rate following treatment (Table III) . For the derived structural indices, SRNaF therapy also caused a significant decrease in trabecular spacing . The remainder of the structural and dynamic parameters did not show a significant change during therapy . Examination of the sections under polarized light demonstrated lamellar bone with no woven bone observed, a finding consistent with a previous report utilizing backscattered electron imaging to examine cancellous bone architecture in fluoride-treated subjects (Zerwekh et al . 1992) . Furthermore, there was no prolongation of the mineralization lag time following SRNaF administration supporting the lack of a fluoride-induced mineralization defect . Comparison of histomorphometric parameters between severe and mildmoderate groups (Table IV) demonstrated no significant differences prior to or following therapy in all parameters, except mineral apposition rate, probably due to small patient numbers . There was a significant increase in mineral apposition rate for patients with mild-moderate bone loss following therapy (0 .6 0 .4 vs . 1 .2 ± 0 .6 .m/day, p p = 0 .008), but no significant changes for patients with severe bone loss when considered separately . Comparison of the mean percent change in each parameter following therapy between severe and mild-moderate patient groups demonstrated no significant differences in the response between the two groups except for osteoid thickness (17 .2% for severe vs . 1 .9% in mild-moderate, p < 0 .05) .

Connectivity Measurements of connectivity prior to and following SRNaF therapy for all 23 patients are summarized in Table V . There was a significant increase in node number per mm 2 of cancellous tissue area (N#/Area) as well as a significant increase in the node to node strut length (NN/TSL) . The ratio of node number to free-end number (N/F) was also significantly increased . There were no other significant changes . Table VI summarizes the connectivity data for the 23 patients when divided into severe and mild-moderate bone loss categories . Prior to therapy, patients with mild-moderate bone loss had a significantly greater node-to-node length relative to total trabecular length (NN/TSL) than the group with severe bone loss . The mild-moderate group had significant increases in N#/Area and NN/Area following approximately 2 years of fluoride therapy . However, patients with severe bone loss did not show sta-



J . E . Zerwekh et al . : NaF improves trabecular connectivity

694

B

.%

W

C

Fig . 1 . (A) Grey level image of a 10 µm section of bone from a transcortical iliac crest biopsy as stored in the computer . The bar at the bottom represents 4 mm . (B) The same image as in (A) but following thresholding and conversion to a binary image . Note the loss of most nonmineralized marrow elements . (C) The cancellous region of interest upon which connectivity measurements are to be made has been clipped from (B) by editing with a mouse . The area of the image is 33 mm' .

1.

E . Zerwekh et al . : NaF improves trabecular connectivity

695

Table I . Effect of slow-release NaF and continuous calcium citrate therapy on biochemical parameters in 23 osteoporotic patients Parameter Serum Ca (mg/dL) Serum P (mg/dL) Serum alkaline phosph .(a/L) Bone F (% ash weight)

Pre-

Post- °

Normal

9 .6 ± 0 .4 3 .5 ± 0 .6

9 .5 ! 0.4 3 .6 ± 0.5

8 .5-10 .5 2 .5-4 .5

130 ! 96

109 *- 52

<140

0 .10 '- 0 .06

0 .18 ! 0 .08*

0 .08 '- 0 .05b

'Post- samples represent approximately 24 months of therapy . °Taken from Boivin et al . (1988) . *Significantly different from pre- (p < .01) by Wilcoxon signed rank test . tistically significant improvement in any of the parameters of connectivity, even though mean values for N#/Area, NN/TSL, and NN/Area were substantially higher after treatment . When the mean percent change for each parameter of connectivity was assessed and compared between severe and mild-moderate groups, only the NN/Area parameter demonstrated significant differences (149% in severe vs . 93% in mild-moderate, p < 0 .05) . Discussion It is generally believed that two mechanisms of bone loss exist at the microanatomic level . They are represented by trabecular thinning, and penetration-erosion of trabeculae . The latter process leads to proportionately greater disruption of trabecular structure, but the two mechanisms are to some extent interdependent . They ultimately lead to osteoporosis, a disease characterized by low bone mass, by causing deterioration of bone tissue microarchitecture, enhanced bone fragility, and increased fracture risk . Thus, an effective therapeutic regimen for osteoporosis should restore the biomechanical competence of bone not only by increasing bone mass and quality but also by restoring normal bone structure . We have previously demonstrated that the use of a slowrelease preparation of NaF, given intermittently with continuous calcium citrate supplementation, could increase vertebral bone mass while reducing spinal fracture rate (Pak et al . 1989) . This regimen has been demonstrated to produce trabecular bone that is normally mineralized (Zerwekh et al . 1990) and of improved material quality (Zerwekh et al . 1991) . In addition, this regimen

caused no deleterious effects on cortical bone histology and quality (Zerwekh et al . 1992), unlike the experience with other fluoride preparations and regimens (Bang et al . 1978 ; Boivin et al . 1989) . More recently, in a placebo-controlled randomized trial it was reported that the slow-release-NaF treated group had a significantly lower individual new vertebral fracture rate (0 .057 vs . 0 .204/pt cycle, p = 0 .017), higher fracture-free rate (83 .3 vs . 64 .7%, p = 0 .042) and lower group fracture rate (0 .085 vs . 0 .239/pt cycle, p < 0 .006) than the placebo-treated group during an average treatment period of 2 .5 years (Pak et al . 1994) . In the present study, we observed that SRNaF treatment had no effect on serum calcium, phosphorus, or alkaline phosphatase . As expected, bone fluoride content increased significantly during fluoride therapy . The mean value of 0 .18% relative to bone ash weight was well below the reported toxic concentrations found in fluorotic bone of 0 .79 ± 0 .36% (Boivin et al . 1989), reflecting a modest dosage and slow-release nature of our fluoride preparation . Similarly, there was a modest but significant increase in L2L4 BMD of 5 .4% in the whole group of 23 patients during 2 years of therapy . When they divided into two groups representing severe bone loss (pre-BMD value <65% of the 30-year-old normal mean), and modest bone loss (a65% of the 30-year-old normal mean), the increases in BMD were still evident and significant for both groups . Histomorphometric analysis of bone biopsies also substantiated a beneficial effect of this therapeutic regimen on bone histology . The mean increase in cancellous bone volume was nearly 15% . This increase was not significant, however, due principally to large patient variability . As previously observed (Zerwekh et al . 1992), SRNaF significantly increased mineral apposition rate as well as adjusted apposition rate when all patients were grouped together . When patients were separated into severe and mild-moderate bone loss groups, no significant differences were observed between the two groups for either pre- or posttherapy values . Fluoride therapy did promote a significant increase in mineral apposition rate for the mild-moderate bone loss group . Further evaluation of the mean percent change in each histological parameter following therapy between the severe and mildmoderate bone loss groups demonstrated a significant difference in the osteoid thickness response between the two groups . The greater mean percent increase observed for the severe group probably reflects the lower initial mean value and thus a greater percent change . Despite this sole difference in response between the two groups, no patients demonstrated increased osteoid parameters or prolongation of mineralization lag time, supporting a

Table II . Effect of slow-release NaF and continuous calcium citrate therapy on spinal bone mineral density BMD (gm/cm') Group All patients (n = 23) Severe (n = 13) (Pre-BMD <65% of 30-yr-old normal) Modest (n = 10) (Pre-BMD 1-65% of 30-yr-old normal) Severe vs . modest

Pre-

Post-

Pre- vs . post-

0 .827 ± 0 .176'

0 .872 '- 0 .166

p = 0,0004-

0 .717 x 0 .090

0 .773 ±0 .106

p<0 .Olt

0 .971 ± 0 .157 p = 0 .0001$

p = 0 .0002t

values expressed as mean '- SD . *Significance assessed by paired-t analysis . }Significance assessed by Wilcoxon signed rank test . *Significance assessed by Student's t test. 'A11

1 .00 t 0 .140

p = 0 .026-



696

J . E . Zerwekh et al . : NaF improves trabecular connectivity Table III. Histomorphometric parameters for cancellous bone in 23 patients with osteoporosis prior to and following approximately 2 years of combination intermittent SRNaF and continuous calcium citrate therapy Parameter BV/TV (%) OV/BV (%) Th .Th (µm) OS/BS (%) Ob .S/BS (%) ES/BS (%) MAR (µm/d) AJAR ' d (µm/d) O .Th (µm) W .Th (µm) MS/BS' (%) MS/BS (%) Mlt' (days) Tb.N (mm - ') Th .S (pm)

Pre-

Post-

Normal'

11 .6 ± 5_8" 1 .6 .t 1 .5 138 x 31 9 .1 !- 5 .4 1 .8 _ 2 .2 3 .7- 1 .4 0 .6 ± 0 .4 0 .4 ± 0 .3 11 ! 4 41 ! 5 3 .0 ! 3 .2 1 .8 ! 1 .8 23 12 0 .88 ± 0 .44 1375 ± 878

13 .3 1 6 .4 1.6 w 1 .1 131 ± 26 10 .7 ± 4 .8 1 .8 s 2 .0 3 .2 ± l2 1.1 s 0 .7 (p = 0 .0078)` 0 .6±0 .4(p=0 .016) 11 i 4 41±7 3 .9 x 2 .9

20 .8 t 4 .4 2 .2 ± 1 .1 133 ± 34 16 .7 x 7 .0 6 .1 ± 3 .8 4 .1 r 2 .1 0-5 t 0 .04 0 .36±0 .01 9 .2 ± L9 47±13 11 .2 ± 5 .7 4 .3±28 17 _ 14 1 .39 ' 0 .28 605 3 151

2 .5-2 .5

24 - 21 1 .04 t 0 .47 1052 ± 541 (p = 0 .05)

Histomorphometric parameters are abbreviated according to the nomenclature scheme approved by the American Society for Bone and Mineral Research (Parfitt et al . 1987) . 'Normal subject data is derived from Kimmel et al . (1990) except for Tb .N and Th .S which are taken from Parfitt et al . (1983) . "All values expressed as mean , SD . `Post- value significantly different from pre- value by paired-t analysis . d Measurements derived from data using all tetracycline-labelled surfaces as the mineralizing surface is designated by a ('), while no such indication represents only double-labelled surfaces .

Note .

lack of a deleterious effect of this fluoride regimen on bone mineralization . For the derived structural indices, only trabecular spacing demonstrated a significant reduction following therapy and was most likely due to the increases in bone area and/or changes in trabecular width . Although trabecular number increased, the change was not significant . Nevertheless, these observations on the histological specimens suggested that SRNaF could have exerted a positive effect on cancellous hone structure or connectivity . To resolve this issue, we utilized a recently described technique (Mellish et al . 1991) for quantitating trabecular connectivity in bone histological sections from our patients undergoing intermittent SRNaF therapy with continuous calcium citrate supplementation . Among the parameters measured, node number, node-to-node strut length, and node to free-end ratio are measures of connectivity of the trabecular plates, white free-end to free-end strut length, as a percentage of total strut length, is an indicator of loss of connectivity (Compston et al . 1987) . For the entire group of patients, baseline assessment of connectivity demonstrated values for each of the parameters comparable to previously published values for osteoporotic patients using similar methods for measurement of connectivity (Compston et al . 1987 ; Garrahan et al . 1986) . We found that SRNaF caused a significant increase in node number and in the node to free-end ratio, supporting improved connectivity . Node-to-node strut length significantly increased when expressed relative to cancellous area but failed to reach significance (p = 0 .10) when expressed relative to total strut length . The failure to reach significance may have been due in part to the relatively large range of values measured in these sections . In order to further define the mechanism by which SRNaF and continuous calcium citrate therapy affected parameters of connectivity, we sought the dependence of indices of connectivity on the severity of spinal bone loss at entry . Prior to therapy, the group with mild-moderate spinal bone loss demonstrated a significantly greater ratio for the mean node-to-node length rel-

ative to total length, than the group with severe bone loss . Following therapy, there were statistically significant increases in the mean node number and in the mean node-to-node length relative to cancellous area only in the patients with mildmoderate bone loss . When a comparison of the mean percent change in each connectivity parameter was made between the severe and mild-moderate bone loss groups, only the NN/Area response was significantly different between the two groups, the response being significantly greater in the severe group . This is probably reflective of the lower pretherapy value in the severe group and thus a greater percent change for a given increase in the NN/Area . Additional patients in each category of bone loss will help to assess whether each group responds comparably to this form of fluoride therapy . The current findings suggest that the improvement in canceltons connectivity observed with the present mode of fluoride administration is more marked for patients in whom some cancellous structural organization is still present . This interpretation is consistent with the known effect of fluoride to stimulate bone formation upon existing bone surfaces and not by promoting endochondral ossification . Thus, an increase in cancellous bone volume, however produced, is believed to only occur by thickening of existing trabeculae . Although trabecular separation was significantly diminished for the whole group following 2 years of SRNaF treatment, there was no increase in trabecular thickness as quantitated from histological sections . Since trabecular thickness was measured directly, it represents an average of all trabeculae, including those arising from the endosteal surface of cortical bone . In general, these trabeculae tend to be thicker and, if present in sufficient quantity, might bias the assessment of trabecular thickness towards a higher value . On the other hand, there was little or no change in the remaining parameters of connectivity . Ideally, a decrease in the free-end to free-end strut length would be further supportive of improved connectivity . However, it has been previously demonstrated that the greatest coefficient of variation in connectivity



I . E . Zerwekh et al . : NaF improves trabecular connectivity Table IV . Histomorphometric parameters for cancellous bone in 23 patients with osteoporosis, divided into severe and mildmoderate bone loss groups, prior to and following approximately two years of combination intermittent SR-NaF and continuous calcium citrate therapy

Parameter BV,TV (%n) PrePostOVIBV (°s) PrePostTb .Th (Pm) PrePostOSBS (4e) PrePostOb .S/BS (9a) PrePostES BS PrePostMAR (µm) PrePostAJAR" PrePostO .Th (Pm) PrePostW .Th (µm) PrePostMIt (days) PrePostMS/BS' (%a) PrePostMSBS PrePostTh .N (mm - ') PrePostT6 -s (Pm) PrePost-

Mild-moderate bone loss (n = 10)

Severe bone loss (n = 13) 11 .136 .6' 11 .7 5 .4

12.7 15 .3

4 .7 7 .3

1 .5 2 1 .5 1 .5 3 0 .8

1 .8 2 1 .6 1 .7 ± 1 .5

135 3 23 131 2 33

145 ± 41 133 2 16

8 .1 !5 .1 10.7 3 4 .9

11 .03 5 .5 10 .7 4 .8

1 .3 ' 1 .6 1 .6 3 0 .9

2 .1 3 2 .7 2 .2 3 2 .9

4 .0 3 1 .6 3 .3 3 1 .3

3 .6 ± 1 .2 3 .2 2 1 .1

0 .6 0 .4 1 .1 3 0 .7

0 .620 .4 1 .2 ± 0 .6 (p = 0 .008)°

0 .330 .3 0 .6 ± 0 .4

0 .420 .3 0 .6 ± 0 .5

9 .723 .1 11 .0`34 .1

12 .035 .3 11 .7 4 .5

40 3 3 .3 39 ± 4 .7

42 3 5 43 3 9

23 ± 14 30 3 29

24 3 10 18 3 11

2 .1 ± 2-5 3 .932 .8

4 .0 3 3 .8 3 .733 .2

1 .1 ± 1 .1

2 .4 3 2 .4

2 .6 3 2 .1 2 .4 2 2 .8

0 .84 3 0 .48 0 .94 3 0 .43

0 .94 2 0.41 1 .17 3 0.52

1488 3 944 1148 2 527

1216 ± 811 927 2 562

'All values expressed as mean 2 SD . bPost- value significantly different from pre value by Wilcoxon signed rank test . `Measurements derived from data using all tetracycline-labelled surfaces as the mineralizing surface is designated by a ('), while no such indication represents only double-labelled surfaces .

measurements resided in the free-end to free-end strut length measurement (Garrahan et al . 1986 ; Mellish et al . 1991) . This problem is likely the result of the use of conventional histological sections, viewed in two dimensions, for strut analysis . In particular, an unknown proportion of the free ends will be created by the process of sectioning, increasing both the numbers of free ends expressed per mm2 and the strut length of these struts possessing one or two free ends .

697 Table V . Mean cancellous connectivity measurements in 23 patients with osteoporosis before and following 2 years of intermittent SRNaF and continuous calcium citrate therapy Pretreatment N#/Area (#/mm) F#/Area (#/mm2) N/F NN/TSL (% TSL) NN/Area (mm/mm2) NFITSL (46 TSL) NF/Area (mm/mm) FF/PSL (% TSL) FF/Area (mm/mm) CXS/TSL (% TSL) CXS/Area (mm/mm 2)

0 .22 1 .20 0 .23 12 .9 0 .098 31 .9 0 .237 33 .9 0 .227 21 .2 0 .159

3 3 ! 2 3 3

0 .12' 0 .59 0 .21 12 .0 0 .101 11 .5 3 0 .106 3 17 .4 3 0 .107 3 9 .6 3 0 .100

Posttreatment 0 .39 3 1 .14 3 0 .41 3 19 .4 3 0 .212 a 30 .2 2 0-269 3 31 .9 2 0 .239 3 18 .5 2 0 .160 2

0.27' 0.35 0.46** 17 .4 0.183', 13 .8 0.152 21 .8 0.137 11 .9 0-134

Note, Connectivity abbreviations are described in the methods section and are expressed relative to total trabecular stmt length (TSL) or to total cancellous tissue area (Area) . 'All values expressed as mean 3 SD . *Significantly different from pretreatment value by paired-t analysis, p = 0 .0077 . **Significantly different from pretreatment value by Wilcoxon signed rank test at p < 0 .05 . Significantly different from pretreatment value by Wilcoxon signed rank test at p < 0 .01 . Although the observed changes in connectivity were small, they most likely reflect a parallel change in skeletal structural competence . Following therapy, four patients with the lowest baseline L2-L4 BMD demonstrated new vertebral fractures . The patient with the lowest initial bone density developed five new fractures; the same patient had a fall in node number relative to area (-0 .147) and in the node-to-node length relative to total length (-0 .066) following therapy . Another patient with two new spinal fractures demonstrated a fall in node number of -0 .130 after fluoride . For the two remaining patients who developed one fracture each, positive changes in node number and node-to-node length were only minimal at best . Although the significance of this observation is limited by the few patients who developed fractures following therapy, it may point to an important association between increases in cancellous connectivity and improvement in bone strength during SRNaF therapyA word of caution must be made with regard to the assessment of connectivity of bone from two-dimensional sections . There is no theoretically conceivable way of estimating 3-dimensional connectivity of cancellous bone from single 2-dimensional sections . Certainly, the use of serial sections in assessing connectivity can improve the quantitation . However, until the above described method for assessing cancellous connectivity is rigorously tested against 3-dimensional techniques, the results should be assessed with caution (Odgaard & Gundersen 1993) . Despite these limitations, the application of this method to quantitation of trabecular connectivity has received increasing application (Parisien et al . 1992 ; Shen et al . 1993) . In addition, the application of strut analysis to bone biopsies from normal subjects has demonstrated greater age-related loss of structural integrity in women than in men (Compston et al . 1987), consistent with data obtained using other methods for the assessment of bone structure (Hahn et al . 1992 ; Vesterby et al . 1989 ; Weinstein & Hutson 1987) . To our knowledge, the current study represents the fast application of this technique to fluoride-treated osteoporotic patients . Our finding of an improvement in trabecular connectivity during fluoride therapy is consistent with the results of Vesterby et al . (1991) using marrow space star volume and with the his-

698

J . E . Zerwekh et al . : NaF improves trabecular connectivity

Table VI . Mean cancellous connectivity measurements in 23 osteoporotic patients divided into severe and modest hone loss groups before and following approximately 2 years of intermittent SRNaF and continuous calcium citrate therapy

Parameter N#/Area PrePostF#/Area PrePostNN/TSL PrePostNN/Area PrcPostNF/TSL PrePostNF/Area PrePostFF/TSL PrePostFF/Area PrePostCXSITSL PrePostCXS/Area PrePostN#/F# PrePost-

Severe bone loss

Mild-moderate bone loss

0 .216 ± 0.140' 0 .319 ± 0-222

0 .219 + 0 .092 0 .477 ± 0 .314 (p = 0 .038)°

135 ± 0.71 . 1 .19 + 0-34

1 .00_ 0 .31 1 .06 1 0 .36

0 .083- 0 .093 0 .16 a 0 .146

0 .190 ? 0 .128 (p = 0 .029)` 0 .233 _ 0 .207

0 .077 :~ 0 .100 0 .165 '- 0 .168

0 .125 + 0 .100 0 .271 '_ 0 .193 (p -- 0 .004) 1

0 .319 '- 0 .149 0 .309 + 0 .153

0 .322 + 0 .077 0 .293 + 0 .124

0.240 + 0 .136 0 .263 x 0 .166

0 .232 + 0 .054 0,275 + 0 .142

0 .376 x 0 .176 0 .318 ± 0 .195

0 .293 + 0 .170 0 .320 + 0 .257

0 .239 + 0 .100 0 .246 + 0 .134

0 .212 ± 0 .120 0 .231 ± 0 .148

0 .225 + 0 .121 0 .208 '- 0 .130

0 .195 ± 0 .104 0 .154 ± 0,104

0 .173 - 0 .129 0 .169 ± 0 .123 NS

0 .141 ± 0 .041 0 .148 ± 0 .152 NS

0.223 '- 0 .255 0.278 ± 0 .171

0 .250 x 0 .153 0 .586 + 0 .659

in vertebral fracture rate despite a marked increase in bone mass (Riggs et al . 1990), no quantitation of trabecular structure has been reported for that study . It is conceivable that large doses of plain NaF, as used in that study, may have promoted the formation of bone of poor material quality as well as stimulated bone resorption . thus negating improvement in trabecular connectivity .

Acknowledgments : The authors wish to express their appreciation to

Marian Soheili for bone biopsy tissue preparation and to Janice Smith for clerical assistance in the preparation of this manuscript . This study was supported by USPHS Grant R0l-AR16061 (CYCP) and MOl-RR00633 . References

Abbreviations are identical to those described in Table IV . 'All values expressed as mean + SD . 'Post value significantly greater than pre value by Wilcoxon signed rank test, 'Mild-moderate group value significantly different from severe group value by Wilcoxon rank-sum test . Note .

tological assessment of Mackie et al . (1989) . On the other hand, Aaron et al, (1991) have demonstrated no improvement in trabecular architecture during fluoride therapy despite significant increases in bone volume and trabecular thickness . Such differences between studies are probably the result of variations in the method used to assess trabecular connectivity as well as differences in the amount and duration of NaF therapy . Thus, in contrast to our study, the negative study of Aaron et al . (1991) reported significant increases in osteoid parameters, as well as a much higher bone fluoride content than in our patients . In summary, we have found that therapy with intermittent SRNaF with continuous calcium citrate promotes a small but significant increase in parameters of trabecular connectivity in patients with osteoporosis and vertebral fractures . This observation combined with previous findings of increased bone mass and improved bone material quality may partly explain our recent observation of a significant decline in vertebral fracture rate in patients while on this therapy (Pale et al . 1993, 1994) . Although a recently completed trial of plain NaF therapy in postmenopausal osteoporosis failed to demonstrate a significant reduction

Aaron, J . E . ; De Vemejoul, M : C .; Kanis, J- A . 'the effect of sodium fluoride on trabecular architecture . Bone 12:307-310 : 1991 . Antich, P . P . ; Anderson, J . A . ; Ashman, R . H . ; Dowdey, J . E . ; Gunzales, J . ; Marry . R . C . : Zerwekh, J . E . : Pak, C . Y . C . Measurement of mechanical properties of bone material in vitro by ultrasound reflection : Methodology and comparison with ultrasound transmission . J . Bone Miner . Res . 6:417-426 : 1991 . Bang . S . ; Hand . C . A . ; Boivin, G . ; Demeurisse, C . ; Gusto, M . ; Tonchon-Danguy, H . 1 . ; Verg, J . M . Morphometric and biophysical study of bone tissue in industrial flunmsis . Courvviser,, H . ; Ikmath, A : Baud, C . A ., it, Fluoride and bone . Bern: Hans Huber ; 1978 ; 168-175 . Boivin, G . : Chapuy, M . C . ; Haul, C . A . ; Meunier, P. I . Fluoride content in human iliac bone . Results in controls, patients with fluorosis . and osteoporotics treated with fluoride . J . Bone Miner . Re.e . 3 :492-502 ; 1988 . Boivin, G . : Chavassieux, P . ; Chapuy, M . C . : Baud, C . A . ; Meunier, P . J . Skeletal fluomsis. Histomorphometric analysis of hone changes and bone fluoride content in 29 patients- Bone W:89-99 ; 1989 . Compston, J . E . ; Mellish, R . W . E- : Garrahan, N- J . Age-related changes in iliac crest trabecular microanatomic bone structure in man . Bone 8 :289-292, 1987 . Feidkamp, L . A . : Goldstein, S . A . ; Partin, A . M . ; Jesion, G . : Kleerekoper, M . The direct examination of three-dimensional bone architecture in vitro by computed tomography. J . Bone . Miner . Res . 4 :3- 11 : 1989 . frost, H . M . Bone histomorphometry : Analysis of trabecular bone dynamics . Recker, R . R . . ed . Bone histomarphomerry : Techniques and imerpremtion. Bo,, . Raton, FL- CRC Press ; 1987 ; 595-610. Guardian . N . 1 . ; Mellish, R . W . E. ; Compston, J . E. A new method for the twodimensional analysis of bone structure in human iliac crest biopsies . J. Mi. 142 :341-349; 1986. , cros Home, M .; Vogel, M . ; Pompesius-Kempa, M .; Delling, G . Trabecular bone pattern factor-A new parameter for simple quantification of bone micmaeehitecture . Bone 13 :327- 330 ; 1992Kimmel, D. B . ; Recker . R . R . ; Gallagher, J . C . ; Vaswani, A . S . ; Alms, J . F . A comparison of iliac bone histomorphometric data in post-menopausal osteoporotic and normal subjects . Bone Miner . 11 :217-235 ; 1990 . Mackie . 1 . G . ; Rule, 7. . A . ; Leyshon, R . L . ; Lane, J . ; Watkins, G . : Berry, P . F. A . Treatment of bone weakness in patients with femoral neck fracture by fluoride, calcium, and vitamin D . J . Bone . /t . Sure . 71b :l 11-117 ; 1989 . Mellish . R . W . E . : Ferguson-Pell, M . W . ; Cochran, G . V . B . ; Lindsay, R . ; Dempster . D . W . A new manual method for assessing two-dimensional cancellous bone structure : Comparison between iliac crest and lumbar vertebra . J . Bone Miner. Res . 6 :689-696; 1991 . Odgaard, A . ; Gundersen, H . L G . Quantification of connectivity in eaneellous bone, with special emphasis on 3-D reconstruction . Bone 14 :173-189 ; 1993 . Pak, C . Y . C . ; Sakhaee, K . ; Antich, P . ; Zerwekh, 1 . ; Peterson, R . ; Piziak, V . ; Gonzales, L, Hatah, M . : Poindexter . 1 . Update in the treatment of osteoporosis with intermittent slow-release sodium fluoride and continuous calcium citrate . Christiansen, C . ; Riis, B ., eds . Fourth international Symposium on Osteoporosis and Consensus Development Conference . Aalborg, Denmark : Handel-Strykkeriet Aalborg Aps . ; 1993 ; 122-124 . Pak, C . Y. C . ; Sakhaee, K . ; Gallagher, C . ; Parcel, C . ; Peterson . R . ; Zerwekh, J . E . ; Lemke, M . ; Britton, F . ; Hsu, M . ; Adams, B . Attainment of therapeutic fluoride levels in scram without major side effects using a slow-release preparation of sodium fluoride in postmenopausal osteoporosis. J . Bone Miner . Res . 1 :563 571 : 1986x .

J . E . Zerwekh et al . : NaF improves trabecular connectivity

Pak, C . Y . C . ; Sakhaee, K . ; Parcel, C . ; Poindexter, 1 . ; Adams, B . ; Bahar, A . ; Beckley, R . Fluoride bioavailability from slow-release sodium fluoride given with calcium citrate . J. Bone Miner. Res . 5 :857-862 ; 1990 .

699

Vesterby, A . Star volume of marrow space and trabeculae in iliac crest : Sampling procedure and correlation to star volume of first lumbar vertebra . Bone 11 :149155 ; 1990 .

Pak, C . Y . C . ; Sakhaee, K . ; Piziak, V . ; Peterson, R . D . ; Breslau, N . A . ; Boyd .

Vesterby, A . ; Gundersen, H . 1 . G . ; Melsen, F . ; Masekilde, L . Normal postmeno-

P . ; Poindexter, I. R . ; Herzog, 1 . ; Heard-Sakhaee, A . ; Haynes, S . ; Adams-

pausal women show iliac crest trabecular thickening on vertical sections . Bone 10:333-339 ; 1989 .

Huet, B . ; Reisch, J . S . Slow-release sodium fluoride in the management of postmenopausal osteoporosis : A randomized controlled trial . Ann . Intern . Med . 120:625-632;1994 . Pak, C . Y . C . ; Sakhaee. K . ; Zerwekh, J . E . Effect of intermittent therapy with a slow-release fluoride preparation . J. Bone Miner . Res . 5 :5149-5155 ; 1986b . Pak, C . Y . C . ; Sakhaee, K. ; Zerwekh, 1 . E . ; Parcel, C . ; Peterson, R . ; Johnson, K . A safe and effective treatment of primary osteoporosis by intermittent application of slow-release sodium fluoride : Augmentation of vertebral bone mass and inhibition of fracture . J . CJin . Bndocrinol . Metab . 60 :150-159; 1989 . Parfrtt. A . M . : Drezner, M . K . ; Glorieux, F. H . ; Rams, 1 . A . ; Malluche, H . ; Meunier, P . J . ; Oh, S . M . ; Recker, R . R . Bone histomorphometry : Standardization of nomenclature, symbols and units . J . Bone Miner. Res . 2:595-610 ; 1987 . Partitt, A . M . ; Matthews, C . H . E . ; Villanueva, A . R . ; Kleerekoper, M . ; Frame, B . ; Rao, D . S. Relationship between surface, volume and thickness of iliac trabecular bone in aging and in osteoporosis : Implications for the microanammic and cellular mechanism of bone loss . J . Clin . Invest . 72 :1396-1409 ; 1983 . Parisian, M . ; Mellish, R . W . E . ; Silverberg, S . J . ; Shane, E . ; Lindsay, R . ; Bilezikian, 1 . P . ; Dempster, D . W . Maintenance of cancellous bone connectivity in primary hyperparathyroidism : Trabecular strut analysis . J . Bone Miner, Res . 7 :913-919 ; 1992. Riggs, B . L . ; Hodgson, S . F. ; O'Fallon, W . M . ; Chao, E . Y . S . ; Wahner, H. W . ; Muhs, 1 . M . ; Cedel, S . L . ; Melton, L. 1 . BI . Effect of fluoride treatment on

Vesterby, A . ; Gundersen, H . J . G . ; Melsen, F . ; Mosekilde, L . Marrow space star volume in the iliac crest decreases in osteoporotic patients after continuous treatment with fluoride, calcium, and vitamin D, for five years . Bone 12 :3337 ; 1991 . Watts, N . B . ; Harris, S . T . ; tenant, H . K. ; Wasnich, R . D . ; Miller, P . D. ; Jackson, R . D. ; Licata, A . A . ; Ross, P . ; Woodson, G . C ., III ; Yanove, M . 1 . ; Mysiw, W . 1 . ; Kohse, L . ; Rao, M. B . ; Steiger, P . ; Richmond, B . ; Chestnut, C. H., III . Intermittent cyclical etidronate treatment of postmenopausal osteoporosis . N . Bngi. J. Med. 323 :73-79 ; 1990 . Weinstein, R . S . ; Havant, M . S . Decreased trabecular width and increased trabecular spacing contribute to bone loss with aging . Bone 8:137-142 ; 1987 . Zerwekh,1 . E . ; Antich, P . P . ; Pak, C . Y . C . ; and collaborators . Slow release (SR) NaF with Ca citrate provides safety of usage and stimulates formation of normally mineralized bone of improved quality (abstract) . J . Bone Miner . Res . 5(Suppl . 2):S430 ; 1990 . Zerwekh, J . E . ; Antich, P . P . ; Sakhaee, K . ; Gonzales, J . ; Gottschalk, F . ; Pak, C . Y . C . Assessment by reflection ultrasound method of the effect of intermittent slow-release sodium fluoride-calcium citrate therapy on material strength of bone . J . Bone Miner . Res . 6:239-244; 1991, Zerwekh, J . E . ; Antich, P. P ., Sakhaee, K. ; Prior, L ; Gonzales, I . ; Gottschalk, F. ; Pak, C . Y . C . Lack of deleterious effect of slow-release sodium fluoride treatment on cortical bone histology and quality in osteoporotic patients . Bone Miner . 18 :65-76; 1992 .

the fracture rate in postmenopausal women with osteoporosis . N. East . J . Med . 322:802-809 ; 1990 . Shen, V . ; Dempster, D . W . ; Birchman, R ; Lindsay, R . Loss of cancellous bone mass and connectivity in ovariectomized rats can be restored by combined treatment with parathyroid hormone and estradiol . J . Clot . Invest . 91 :24792487 ;1993 .

Dare Received : February 22, 1994 Dare Revised: March 25, 1994 Dare Accepted: June I, 1994