fects of exercise on the periosteal or endocortical percentage double-labeled ... ing exercise sufficient to stimulate periosteal formation in the femur did not ...
JOURNAL OF BONE AND MINERAL RESEARCH Volume I,Number 9,1993 Mary Ann Lkkrt, lac., Publishers
Does Weight-Bearing Exercise Affect Non-Weight-Bearing Bone? LORRI J. TOMMERUP.,' DIANE M. RAAB,' THOMAS D. CRENSHAW,' and EVERETT L. SMITH1
ABSTRACT
It was previously reported that weight-bearing exercise increases femur periostenl formation in 3-yenr-old, 230 kg cross-bred sows. To determine if this weight-bearing exercise also stimulated non-weight-bearing bone, bone formation in the seventh rib from these same sows was measured blstomorphometrically on the periosted, endocortical, and osteonal envelopes. The sows were randomly assigned by body weight to basal (B, n = a), control (C, n = 7), or trained (T, n = 7 ) groups. After 3 weeks of exercise adaptation, T walked on a treadmill for 17 weeks at 5 km/h, 24 minutedday, 5 daydweek, at 5% grade. Groups were sacrificed initially (B) or after 20 weeks (Cand T). Periosted mineral apposition rate (MAR) was calculated over 136 days. Osteonnl and endocortical M A R were calculated over the 14 days prior to sacrifice. There were no effects of exercise on the periosteal or endocortical percentage double-labeled surface (dLS/BS), osteonal remodeling frequency (N.dL.On/B.Ar), or MAR in any bone envelope of the rib. In conclusion, weight-bearing exercise sufficient to stimulate periosteal formation in the femur did not activate formation in the rib. Bone response to weight-bearing exercise appears to be specific to the loaded bones.
INTRODUCTION
ing exercises could be expected to affect the weight-bearing spine and legs,'5,7-p)and upper body exercises could be exEIOHT-BEARINO EXERCISE INTERVENTION increases pected to affect exercise-loaded arms. (I4) bone mass or reduces its decline in both humans(1d) Whether weight-bearing exercise can systemically affect and other animals. 0-v Some cross-sectional human studies a non-weight-bearing bone that is nor directly loaded by also show that people with a history of exercise and physi- the exercise has not been resolved. The results of some cal activity have higher measures of bone Re- cross-sectional human studies suggest that a non-weightcently, Raab et al.C7)demonstrated that femoral periosteal bearing bone, such as the radius, may be affected by formation and mineral apposition rate were greater in weight-bearing exercise. ( 3 ~ 6 ~ 1 0 . 1 2Human ) and animal exertreadmill exercise-trained young adult sows than in seden- cise intervention studies have shown both no effcct(z.16' and a positive effect(16)of weight-bearing exercise on nontary sows. The skeleton is under both systemic and local regula- weight-bearing bones. tion. Systemic hormonal actions stimulate either bone reFew studies have attempted to investigate the potentially sorption or formation throughout the skeleton, but me- differential effects of exercise on weight-bearing and nonchanical loading, such as exercise, may stimulate only a weight-bearing bones from the same subjects. The results local or regional response.(13)Specific exercise loads are of exercise studies designed to load the skeleton are often likely to have local effects proportional to the forces pro- difficult to compare because of a lack of standardization duced by the attached muscles or to various compressive, in exercise frequency, type, duration, intensity, and meatensile, and bending forces. Thus, lower body weight-bear- surement techniques. To avoid these problems, both a
W
'Biogerontology Laboratory, University of Wisconsin-Madison. 'Center for Hard Tissue Research, Creighton University, Omaha, Nebraska. 'Department of Meat and Animal Sciences, University of Wisconsin-Madison.
1053
TOMMERUP ET AL.
1054
weight-bearing bone (the femur)(71and a non-weight-bearing bone (the rib) were analyzed from the same experiment. The purpose of the present investigation was to evaluate histomorphometrically any effect of the weight-bearing exercise on modeling or remodeling activity in the nonweight-bearing bone.
Histomorphometric analysis
Cross sections 1 cm thick from the vertebral end of the 5-6 cm midsections were dehydrated and embedded undecalcified in polymethyl methacrylate (modified by Kalscheur V, Comparative Orthopaedic Research Laboratory, Department of Surgical Sciences, School of Veterinary Medicine, University of Wisconsin-Madison, Madison, WI 53706).(17' The embedded blocks were sectioned on a MATERIALS AND METHODS milling machine at 600 pm. Sections were glued to glass slides and ground to 80-95 pm (Buehler Ecomet 11). Two A group of 20 adult cross-bred sows were ranked by sections at least 4 mm apart were analyzed from each rib. body weight and blocked into seven groups.(') Sows in The average area analyzed was 150 mm'/sow. each block were randomly assigned to a basal (B, n = 6), Preliminary observations suggested that there were difcontrol (C, n = 7). or trained group (T, n = 7). All sows fering amounts of intracortical remodeling and periosteal were maintained nongravid for 8 months prior to the study formation in different cortices of the rib. Five regions (Iin order to ensure steady-state bone cell activity. At the be- V) were approximated and delineated on each section (Fig. ginning of the study, the animals in the B group were sacri- 1). The following measurements were made in each region ficed and analyzed to determine baseline bone characteris- using a Zeiss Zidas digitizing system: tics. The C and T groups then continued in the exercise study for 20 weeks. 1. The percentage of forming surface (dLS/BS)(In) on The sows were kept in individual standard 2.1 x 0.6 m the periosteal and endocortical surfaces was defined as the pens in a specific pathogen-free facility. The pens provided percentage double-labeled surface relative to the total measpace for standing and lying; forward or backward move- surable bone surface. Double-labeled periosteal surface ment was limited to approximately 0.3 m. The corn-soy- had a tetracycline label and at least one calcein label. Doubean meal diet (2 kg per day) fed throughout the study met ble-labeled endocortical surface had two calcein labels, or exceeded recommended levels for all nutrients. Water since the initial tetracycline label had been almost entirely was available ad libitum. All animal care and experimental resorbed. procedures were approved by the University of Wisconsin 2. The number of forming osteons (N.dL.On/B.Ar)("' Research Animal Resource Committee. was determined by counting the number of calcein doublelabeled osteons relative to the area analyzed. 3. Interlabel width (IrL. Wi)(l81 was measured between Exercise protocol the midpoints of the fluorescent labels. On the periosteal The T sows walked on a motor-driven treadmill 5 days/ surface, IrL.Wi was measured from the midpoint of the week at 5% grade for 20 weeks.(7) Exercise time was in- tetracycline label to the center of the double-calcreased from 5 minutedday the first day to 20 minutes/ cein-labeled region. On the endocortical surface and the day on day 8. Exercise speed began at 2.5 km/h and was osteons, 1rL.Wi was measured between the midoints of the was progressively increased by 0.8 km/h in weeks 2, 3, and 4. two calcein labels. Mineral apposition rate Starting with week 4, the T sows walked at 5 km/h, 20 then calculated for the periosteal (IrL.Wi/136 days), endominutedday for the next 17 weeks. Relative exercise inten- cortical, and osteonal surfaces (IrL.WiIl4 days). sity declined from 60% of maximal heart rate capacity in Since the B group had not been given a tetracycline week to 50% of maximum heart rate capacity in week 18. label, all periosteal analyses were limited to the C and T groups.
Bone labeling Calcium binding fluorescent markers were infused into the jugular vein of all sows.(71The C and T sows were given a 2 day oxytetracycline injection (25 mg/kg of body weight) at the start of the experiment. All sows were given double calcein (Sigma, St. Louis, MO) injections (15 mg/ kg of body weight) beginning 21 days before sacrifice. The calcein protocol was 2 days injection, 12 days off, 2 days injection, and 5 days off before sacrifice.
Data collection At sacrifice, the left seventh rib was removed from each sow and cleaned of soft tissue. The midpoint of each rib was located and marked. A 5-6 cm section with end points equidistant from the rib midpoint was cut with a band saw. Sections were futed in 70% ethanol for later histomorphometric analysis.
Statistics Histomorphometric results were analyzed according to a split-plot two-factor randomized block design, with treatment group (B, C, and T) and body weight blocks as whole plots and rib regions I-V and group-region interactions as subp1ots.(lP1For all variables, the average value of the two rib sections taken from each sow was analyzed. All statistical analyses were done using the SAS General Linear Model program (SAS 1982). Residual plots from all analyses of variance were evaluated for homogeneous variance and normality. Dependent variables were transformed when appropriate with an arcsin square-root transformation for periosteal and endocortical dLS/BS and a squareroot transformtion for periosteal MAR. Between-group differences were tested (Tukey honest significant difference (HSD]) using preplanned comparisons to identify age (B versus C) and exercise (C versus T)
1055
EXERCISE AND NON-WEIGHT-BEARING BONE
effects. Pleural versus cutaneous and caudal \ersus cranial differences among regions were evaluated by contrasts in the following way: regions I1 and IV (combined pleural region) were contrasted with regions 111 and V (combined cutaneous region), and region I (caudal region) was contrasted with regions 1V and V (combined cranial region). All differences were considered significant at P < 0.05.
RESULTS All 14 sows completed the 20 week study with no apparent exercise-related problems in the T group.”’ The three groups did not differ with respect to initial age, body weight, number of prior litters, or months postpartum. At the end of the 20 weeks the C group was significantly older than the B group (Table 1). A cardiovascular exercise effect was demonstrated indirectly using the exercise and resting HR measurements. First, although absolute exercise intensity (treadmill walking speed and grade) remained the same from week 4 to week 20, the relative exercise intensity (expressed as a percentage of estimated maximal HR capacity) decreased from 60% in week 5 to 50% by week 18. Second, sows in the T group had a significantly lower resting HR than sows in the C group at week 10;at week 19. however, the resting HR difference between C and T was not ~ignificant.(~)
Rib cortical bone formation Treatment Group Effects (Table 2): Neither the periosteal nor the endocortical forming surface of the rib was
affected by exercise (C versus T , p > 0.9). nor was there any effect of age on the endocortical forming surface (B versus C , p > 0.9). There were no differences found among the three groups in the number of double-labeled osteons (P > 0.7). Periosteal MAR did not differ significantly between the T and C groups (p > 0.9). Osteonal MAR did not differ with respect to age or exercise (P > 0.9, all three groups). Finally, endocortical MAR did not differ among the groups (P > 0.6, all three groups).
Regional Effects (Figs. I and 2 and Table 3): Periosteal dLS/BS and MAR were significantly greater in the cutaneous and the caudal regions. Endocortical dLS/BS and MAR were significantly greater in the pleural and the cranial regions. Osteonal activation frequency was significantly greater in the pleural regions than in the cutaneous regions. There were no significant regional differences in osteonal MAR. There were no significant group-region interactions in any variable measured.
DISCUSSION This analysis provided an opportunity to investigate and compare exercise effects on both a weight-bearing and a non-weight-bearing bone from the same animals in the same experiment. The weight-bearing bone (femur) showed significantly greater modeling in T than C sows; periosteal dLS/BS was 27% greater and periosteal MAR was 76% greater in T compared to C.(7)However, the non-weight-bearing boine (rib) showed no significant dif-
TABLE1 . Sow SAMPLECHARACTERISTICS~
Age (years) Group (N) B (5)
c (7) T (7)
Body weight (kg)
Initial
Final
Initial
Final
3.06 f 0.22 3.07 f 0.10 2.88 f 0.18
3.06 f 0.22b 3.48 f 0.1Ob 3.29 f 0.18
235 f 10 230 f 8 228 f 8
230 f 1 1 229 f 10 224 f 1 1
.Final age and body weight for C and T are initial + 20 weeks. Final age and body weight for B are initial + 3 weeks. Values represent treatment group mean f standard error of the mean (SEM)of each group. Walues with the same superscript within a column are significantly different from one another (p < 0.05).
TABLE
2. GROUPEFFECTS~
Formation surjace
Mineral apposition rate
Ps
Ec
On
Ps
(dLS/BS)
(%)
(mm-’)
(NdaY)
B (5)
-
c (7)
33.47 f 3.70 41.40 f 5.33
30.71 f 7.30 31.41 f 4.43 28.88 f 4.07
0.69 f 0.21 1.02 f 0.15 0.98 f 0.16
0.74 f 0.23 0.67 f 0.12
Group (N)
T (7)
-
Ec (cLm/dayl
(cLm/daY)
1.26 f 0.12 1.10 0.04 1.18 f 0.06
1.33 f 0.10 1.30 f 0.08 1.35 f 0.06
*
On
aPeriosteal (Ps), endocortical (Ec), and osteonal (On) bone formation surfaces and mineral apposition rates of the seventh rib midsection pooled across regions. Values represent treatmeni group mean f SEM of each group (untransformed data). There were no significant differences among the groups in any of the variables measured. There were no significant group-region interactions.
TOMMERUP ET AL. predictions that muscle growth interacts with respiratory mechanics to move the rib laterally and expand the chest c a ~ i t y . ( ~ ‘Periosteal .~~’ rib formation may continue in the n U human throughout life, although the rate of modeling activII ity appears to decrease with age.(16’Epker and Frost postulate that lateral (cutaneous) rib drift is in part stimulated IV by the mechanics of breathing. They suggest that during inspiration, the cutaneous surface becomes less convex I (compression) and experiences high osteoblast activity and V the pleural surface becomes more convex (tension) and experiences high osteoclast activity. The net effect is expansion of the chest cavity. The consistent drift in all sows 111 suggests that mechanical loading during respiration influences formation of the rib of the young adult sow. FIG. 1. Periosteal, endocortical, and osteonal data were Respiration increases during exercise to meet energy collected in each region (I-V). For analysis of regional effects, regions were combined and contrasted in the follow- needs and to regulate body temperature. Mechanical forces ing way: cutaneous (111 and V) versus pleural (I1 and IV); on the rib cage may increase during weight-bearing activity caudal (I) versus cranial (IV and V). because of greater tidal volume, frequency, and total work of breathing. In this study, there were no differences among groups in rib formation. The cardiovascular exercise intensity was moderate; it did not exceed 60% of maximum heart rate reserve and had decreased to 50% of maximum heart rate reserve by the time the exercise program reached a steady state. Whatever greater respiratory PLEURAL muscle activity may have been required in the exercise group, it was of insufficient intensity to have a direct effect on rib modeling or remodeling.
0 53
m
CRANIAL
CAUDAL
Systemic effects of weight-bearing exercise Results of some human cross-sectional studies seem to imply that a long period of weight-bearing exercise or a very intense weight-bearing exercise program may systemically affect a non-weight-bearing bone.(l*lll The results CUTANEOUS of cross-sectional observations may be confounded by selfFIG. 2. Shading and arrows indicate regional formation selection bias, initial differences in bone mass, or favorand direction of modeling drift as measured by dLS/BS able genetics. In addition, many of these subjects particiand MAR. Circles represent the relative distribution of pated in activities that directly load the radius through double-labeled osteons. muscle contraction. Thus direct loading rather than a systemic effect of weight-bearing exercise could be responsible for the exercise effects at those sites. There is little eviferences as a result of weight-bearing exercise in periosteal, dence from longitudinal studies of a systemic effect of mechanical loading or exercise on bone modeling and remodendocortical, or osteonal modeling or remodeling. Two potential stimuli could have affected the rib in this eling. In studies in which both loaded and nonloaded study: (1) constant, repetitive, direct loading by the action bones are measured, responses are usually absent in the of respiratory muscles and increased mechanical forces in- nonoaded bones. (1.13.17) Weight-bearing exercise and muscle contraction create volved in the work of breathing, and (2) a systemic effect mechanical loads on bone and cause the bone to bend. via some circulating growth factor(s). Bone adapts to increases in strain (the change in bone length relative to its original length) by increasing its mass Direct mechanical loading of the rib and/or changing its architecture; this increases the bone’s A local effect of exercise or mechanical loading on the resistance to bending and returns strain levels to bone b e i i directly loaded is strongly ~ u p p o r t e d . ~ ~ . ~normal. , ~ ~ . ’(13.13) ~ . ~ ~The ’ difference in response between the feIn the present study, ribs from all three groups of animals mur and the rib in this experiment strongly supports the showed regional response to the mechanics of breathing existence of local rather than systemic skeletal adaptation and the pull of the intercostal muscles. The patterns of to exercise. There are two explanations for failing to detect a sysperiosteal and endocortical formation in all groups suggest a modeling drift, with greater formation in the cutaneous temic effect of loading - if one in fact exists - in this study: and caudal directions than in the pleural and cranial direc- (1) the exercise intensity was too low and/or (2) the experitions. This direction of modeling is consistent with earlier ment was too short. First, the exercise intensity war great
*
EXERCISE AND NON-WEIGHT-BEARING BONE
1057
TABLE3. REOIONN. EFFECTS~ Formation surface
Mineral apposition rate On (mm-l)
PSb
Ec
Region ( N )
(dLS/BS)
(W
Cutaneous (19) Pleural (19) Caudal (19) Cranial (19)
50.49 f 6.41c
20.32 f 2 . 6 3 ~
0.67
20.03 f 4.13
42.44 f 4.86
1.18 f 0.12
46.16 f 3.11d
19.76 f 3.03d
0.90
24.05 f 3.76
39.97 f 4.49
1.00 f 0.10
f
Ec (wlday)
PSb
0.08c
* 0.10
1.06
* 0.22~
0.19
f
0.03
1.00 f 0.27~ 0.51
f
0.08
On
1.02 f 0.05~
1.26 f 0.08
1.37 f 0.09
1.36 f 0.04
1.07 f 0.05d
1.40 f 0.05
1.25 f 0.07
1.38 f 0.04
aPeriosteal (Ps), endocortical (Ec). and osteonal (On) bone formation surfaces and mineral apposition rates of the seventh rib midsection pooled across groups. Cutaneous, regions 111 + V; pleural, regions I1 + IV; caudal. region I; cranial, regions IV + V. Values represent treatment group mean i SEM of each region clr combined regions (untransformed data). bFor the periosteal surface only, data from 14 sows were analyzed (C and T groups only). CCutaneous significantly different from pleural within a column (p < 0.001). dCaudal significantly different from cranial within a column (p < 0.001). Caudal significantly different from cranial within a column (p < 0.01).
enough to increase femur formation and decrease heart rate. A more rigorous exercise program may have increased systemic hormone levels, but greater respiratory demands may have increased direct mechanical loading on the ribs as well. Second, the duration of the experiment should have been sufficient to elicit systemic skeletal effects from this exercise program. Since the porcine bone formation period is approximately 2-3 months,") bone effects should have been evident after 5 months of stimulation. Systemic skeletal effects of adaptation I:O increased exercise would be anticipated to occur along with other systemic adaptations to a change in activity level. For example, cardiovascular adaptation (a significant decrease in resting and exercise heart rate) occurred by week 10 of the 20 week experiment. The systemic stress during the last 10 weeks of the experiment was significantly less than in the first 10 weeks. Further tests of systemic skeletal adaptation to exercise should be conducted, with systemic stress maintained constant by increasing the exercise intensity throughout the experimental period. It is also possible that a systemic skeletal effect of exercise does not occur during the adaptive phase but instead results from chronic (many years) steady-state weight-bearing exercise. To examine long-term systemic effects would likely require at least a 1 year training study of adult animals. Since any effects would be small, a large number of animals would be required, and longitudinal in vivo measures of bone mass changes within the same animal would be necessary.
Conclusion Weight-bearing exercise sufficient to produce a cardiovascular training response and stimulate periosteal formation in the femur did not affect modeling or remodeling activity in the rib. The existence of a systemic effect
of weight-bearing exercise on non-weight-bearing bone is not supported by these results. To evaluate the effects of exercise on bone, the sites measured should be those that are directly loaded by the exercise.
ACKNOWLEDGMENTS We thank Vicki Kalscheur, Mary Checovich, and Murray Clayton. This study was supported in part by the National Institute of Biogerontology. A portion of this work was previously presented in abstract form in J Bone Miner Res 65104. 1991.
REFERENCES 1. Smith EL, Reddan W, Smith PE 1981 Physical activity and
2.
3.
4.
5.
6.
7.
calcium modalities for bone mineral increase in aged women. Med Sci Sports Exerc 33(1):60-64. Krolner B, Toft B, Nielsen SP, Tondevold E 1983 Physical exercise- a prophylaxis against involutiond vertebral bone loss: A controlled trial. Clin Sci 64541-546. White MK. Martin RB, Yeater RA, Butcher RL, Radin EL 1984 The effects of exercise on the bones of postmenopausal women. Int Orthop 7:209-214. Chow R, Harrison JE, Notarius C 1987 Effect of two randomised exercise programmes on bone mass of healthy postmenopausal women. BMJ 292:607-610. Dalsky GP, Stocke KS, Ehsani AA. Slatopolsky E, Lee WC. Birge SJ 1988 Weight-bearing exercise training and lumbar bone mineral content in post menopausal women. Ann Intern Med 108:824-828. Smith EL, Gilligan C. McAdam M, Ensign CP, Smith PE 1989 Deterring bone loss by exercise intervention in premenopausal and postmenopausal women. Calcif Tissue Int 44: 312-321. Raab DM, Chenshaw TD, Kimmel DB. Smith EL 1Wl A histomorphometric study of cortical bone activity during in-
TOMMERUP ET AL.
1058 creased weight-bearing exercise. J Bone Miner Res 6(7):741-
749. 8. Woo SL, Kuei SC, Amiel D, Gomez MA, Hayes WC, Ake-
son WH 1981 The effect of prolonged physical training on the properties of long bone: A study of Wolffs law. J Bone Joint Surg [Am] 63:780-786. 9. Martin RK, Albright JP, Clarke WR, Niffenberger JA 1981 Load-carrying effects on the adult beagle tibia. Med Sci Sports Exerc 13(5):343-349. 10. Brewer V, Meyer BM, Keele MS. Upton SJ, Hagan RD 1983 Role of exercise in prevention of involutional bone loss. Med Sci Sports Exerc 15:445-449. 11. Jacobson PC, Beaver W, Grubb SA, Taft TN, Talmage RV 1984 Bone density in women: College athletes and older athletic women. J Orthop Res 2(4):328-332. 12. Stillman RJ, Lohman TG, Slaughter MH, Massey BH 1986 Physical activity and bone mineral content in women aged 30 to 85 years. Med Sci Sports Exerc 18(5):576-580. 13. Currey J 1984 The Mechanical Adaptations of Bones. Princeton University Press, Princeton, NJ. 14. Simkin A, Ayalon J, Leichter I 1987 Increased trabecular bone density due to bone-loading exercises in postmenopausal osteoporotic women. Calcif Tissue Int 4059-63. 15. Aloia JF, Cohn SH, Ostuni JA, Cane R, Ellis K 1978 Prevention of involutional bone loss by exercise. Ann lntern Med 89356-358. 16. Darby LA, Pohlman RL, Lechner AJ 1985 Increased bone calcium following endurance exercise in the mature female rat. Lab Anim Sci 35(4):382-386. 17. Dunneisen E 1987 An improved technique for methyl-methacrylate embedding and toluidine bluelbasic fuchsin staining of undecalcified bone. Trans Orthop Res Soc 12:180. 18. Parfitt AM, Drezner MK, Glorieux FH. Kanis JA, Malluche H, Meunier PJ. Ott SM, Recker RR 1987 Bone histomorphometry: Standardization of nomenclature, symbols, and units. J Bone Miner Res 2595-610.
19. Snedecor GW, Cochran WG 1980 Statistical Methods, 7th ed). Iowa State University Press, Ames, IA.
20. O'Connor JA, Lanyon LE. MacFie H 1982 The influence of strain rate on adaptive bone remodeling. J Biomech 15:767781. 21. Lanyon LE, Rubin CT 1984 Static versus dynamic loads as an influence on bone remodeling. J Biomech 17:892-905. 22. Rubin CT,Lanyon LE 1984 Regulation of bone formation by applied dynamic loads. J Bone Joint Surg [Am) 66:397402. 23. Rubin CT,Lanyon LE 1985 Regulation of bone mass by mechanical strain magnitude. Calcif Tissue Int 37:411-417. 24. Landeros 0,Frost HM 1966 Comparison of amounts of remodeling activity in opposite cortices of ribs in children and adults. J Dent Res 45(1):152-158. 25. Snellman 0 1987 Growth and remodelling of the ribs in normal and scoliotic pigs. Acta Orthop S a n d (Suppl) l(49):l85. 26. Epker BN, Frost HM 1966 Periosteal appositional bone growth from age two to age seventy in man. Anat Rec 154:
573-578. 27. Nelson ME, Fisher EC, Dilmanian FA, Dallal GE, Evans WJ 1991 A I-y walking program and increased dietary calcium in postmenopausal women: Effects on bone. Am J Clin Nutr
53:1304-13 11.
Address reprint requests to: Lorri Tommerup Biogerontology Laboratory 504 North Walnut Street Madison. WI 53705 Received in original form April 6, 1992; in revised form September 28, 1992; accepted February 11, 1993.
