Osteoporosis of Rheumatoid Arthritis : Influence of Age, Sex and Corticosteroids By PAULD. SAVILLEAND OVEDKHARMOSH S The cortical thickness of the left radius just below the tuberosity was measured by the method of Meema.4 Standard lateral roentgenograms of the thoracic and lumbar spine were taken. All the roentgenograms were read by one of us without prior knowledge of the age or sex of the patient or any clinical data. The lateral lumbar spine was evaluated for bone density in 4 categories. In order to make these judgments, we took into consideration the quality of the trabecular pattern with increased vertical striation, loss of density compared to soft tissues, and thinness of the upper and lower plates of the vertebral bodies. Vertebral fractures were ignored on this assessment and were considered on a separate basis. The four categories are: Grade 0: normal bone. Grade 1: minimal loss of bone density; the end plates begin to stand out, giving a stencilled effect. Grade 2: vertical striation is more obvious; the end plates are now thinner. Grade 3: more severe loss of bone density than grade 2; the end plates are becoming less visible. Grade 4: ghost-like vertebral bodies; density is no greater than soft tissue. No trabecular pattern is visible (when the vertebral outline is covered, the body is invisible). Clinical data were recorded on self-coding data sheets from which cards were punched. The cards were sorted in various ways for purposes of the analysis. The statistical characteristics of the grading system used, its reproducibility among physicians and its properties consistent with the osteoporosis is an integral part of Cushing's syndrome,l it is commonly assumed that the osteoporosis and vertebral fractures seen in patients with rheumatoid arthritis who are treated with corticosteroids are mainly due to the drug. Actually, there are 2 separate bone entities which require assessment and which may not be directly related one to another; first, osteoporosis or decreased bone mass, and second, compression fractures of the spine. Either or both might be influenced by treatment. It is the purpose of this paper to assess the influence of age, sex, and severity of disease as well as of corticosteroid treatment on the osteoporosis of rheumatoid arthritis and on the incidence of compression fractures of the spine. INCE METHODS The patients were from a consecutive series who were selected from the Rheumatic Disease Clinic of The Hospital for Special Surgery because they had been diagnosed by the attending physicians as suffering from rheumatoid arthritis. There were 128 women and 36 men. Although the ARA criteriaa were not formally applied to each case, they were used as a basis for making the clinical diagnosis. Functional status was classified according to Steinbrocker, Traeger and Batterman? From The Hospitd for Special Surgery asliated with The New York Hospital-Cornell University Medical Center, New York. This investigation was supported in part by Public Health Service Research Grant No. 5 SO1 Fr-05495 from the General Support Branch Division of Research Facilities and Resources; 1 PO1 AM 09982-01 of the National Institutes of Arthritis and Metabolic Diseases and U.S.P.H.S. Graduate Training Grant TIAM-5414 of the National Institutes of Arthritis and Metabolic Diseases. PAUL D. SAVILLE,M.D.: Associate Professor of Medicine, Cornell University Medical Center; Associate Attending Physician, The Hospital for Special Surgery; Associate Attending Physician, New York Hospital; Senior Scientist; Chief, Metabolic Bone Disease Clinic and Associate Director of Research-The Hospital for Special Surgery. OVED KHARMOSH,M.D.: Orthopaedic Fellow. Present address: Department of Orthopaedics, Zchilov Municipal Hospital, Tel-Aviv, Israel. 423 ARTHRITIS AND RHEUMATISM, VOL. 10, No. 5 (October 1967) 424 SAVILLE AND KHARMOSH Table 1.-Distribution ______ of Corticosteriod Dose Among Rheumatoid Patients -__ Number of patients Dose of prednisone Females Zv5 - 10 mg. 12U.;. - M mg. 22?/1- 30 mg. >30 mg. 46 Males 7 2 1 For example, it is apparent that in both men and women there are significantly RESULTS more individuaIs in the severer grades of osteoporosis among those over 50 than Influence of Age and Corticosteroids Of the patients who were given corti- among those under 50. This is equally true costeroids, almost all took prednisone. Dis- for those treated with corticosteroids as for tribution of dose among the corticosteroid those untreated. This striking finding s u g treated patients is given in Table 1 in gests that age has a more important effect terms of prednisone. We have reported the on spinal osteoporosis than does corticosttypical dose schedule indicated in the pa- eroid treatment, as used in our clinics. It tient’s records. We have not taken into ac- accounts for the small number of younger count short periods of higher dose that may women and men with moderate or severe have been given on occasions, nor have we osteoporosis, as shown in Tables 2 and 3. The insufficiency of cases of osteoporosis considered that the patients may have taken amounts of the drug other than that among the younger age groups is thus a natural one, and makes it necessary to comprescribed and reported in the charts. Fig. 1 indicates that the average cortical bine mild, moderate and severe grades in thickness of the radius is similar in women order to compare men and women under who were given and those who were not 50 with respect to treatment with steroids. given corticosteroid treatment until the The result may be read from Tables 2 and sixth decade; at this time the cortical thick- 3 and is described as follows: For men no significant effect of steroid ness decreases in those treated; in subsequent decades both treated and untreated treatment on grades of porosity is found groups decrease, the decrease being great- for those under 50 years (P>O.2); but er in those receiving corticosteroid treat- such an effect is significantly present in ment. In men, aIthough average corticaI those over 50 (.Ol>P>.OOl). These tests thickness was less in the sixth decade, there are necessarily performed by combining was a decrease in cortical thickness in the grades, as described above. For women under 50 years, a steroid sixth decade in both treated and untreated groups (Fig. 2 ) . There was a tendency for effect on severity of grades is barely dis(grades being it to increase again in both groups with cernible (O.l>P.>OS) succeeding decades; differences between combined as above). In the case of women those men taking corticosteroids and those over 50, however, with sufficient cases present to test a division of grades into not were insignificant. Table 2 shows how spinal porosity va- normal, mild and moderate/severe, a ries in women given corticosteroids com- significant interaction between steroid pared to those not given them. Table 3 pre- treatment and grade of osteoporosis is sents the same breakdown for the men. found (.OZ>P.>OS). These tables may be read in several ways. It is of interest to note that adding all normal distribution ha:: been described elsewhere? 425 OSTEOPOROSIS OF RA: INFLUENCE OF AGE, SEX, CORTICOSTEROIDS Table 2.-Grades of Spinal Osteoporosis in Female Rheumatoid Patients Porosity N o steroids Steroids Under 50 years: Normal Mild Mod./sev. Over 50 years: 16 2 1 7 xam 7 0 P Norma1 2.99 > .2 I 8 24 20 Mild Mod./sev. 1= Xaw = 6.61 .02 P .05 17 > < 25 ~- 7 1 6 Fig. 1.-Shows the average cortical thickness of the radial shaft in 128 women at each decade from the third both for those treated with corticosteroids and those not so treated. 5 2 21-30 31-40 41-50 51-60 61-10 11-60 81-90 AGE GROUPS IN YEARS Fig. 2.-Shows the average cortical thickness of the radial shaft in 36 males at each decade from the f3th decade both for those treated with corticosteroids and those not so treated. P I - 50 51-60 61-10 AGE GROUPS IN YEARS 11-80 426 SAVILLE AND KHARAZOSH Table 3.--Grades of Spinal Osteoporosis in Male Rheumatoid Patients -- __.____-- ;Lo steroids Porosity _-Under SO ?tmr.s: Kormal Mild Lfotl.Jsev Orrrr 50 ?'c-</l..i: Normal Mild MOd./W\ __ ~ _____..____ Steroids ~ _ _ _ ~- ~- X",, = 2.57 P > .2 2 ~- 2 1 0 2 7 5 2 0 7 x2a,= 7.77 .01 1 __ .~ --__~ > P > ,001 ~~_____ Table 4.-Analysis of Covariance of the Regressions of Cortical Thickness of the Radius vs. Age in Female Rheumatoids Treated for Varying Lengths of Time with Corticosteroids ~- .~ P Value 2; Ratio scatter slope means scatter slope means * -__ >.2 >2 >.2 1.07 0.58 37 1.5 .97 7.34 .08 Significxit0+1>2. ,135 mni. Not significant .77 mm. .75 .05>P>.Ol P>.2 P>.2 scatter slope mcans 5.0 29 3.5; .OS>P>.Ol P>.2 .1 >P>.O5 1.06 9.22 - scatter slope means 1.34 Not significant. .85 mm. 5.O scatter slope means iiim. __ Comments .2>P>.1 >.2 <.01 P<.001 scattct slope means 1.16 - ______ Difference between means/slopes .01 .45 .32 Suggestive- P>.2 P>.OO1 - .06 P>.2 .2 .2 > > 4 > 2. Significant4 steeper than 3. .23 mni. .- 0 = no corticosteroid treatment 1 =less than 1 year 2 = 1-13 years 3 = 2-5 years 1=more than I: years. male and female cases under 50 years still fails to provide significant evidence for a steroid-porosity relationship in these younger cases. Durufion of Treatment In women, since cortical thickness of the radius decreases linearly with age after 50 (Fig. l ) , we have compared the regressions of cortical thickness versus age after 50; these regressions were then compared among the various treatment groups by analysis of covariance (Table 4).These analyses showed no significant difference hetween women not treated with corticosteroids compared to those treated for less than 427 OSTEOPOROSIS OF RA: INFLUENCE OF AGE, SEX, CORTICOSTEROIDS ment affects spinal osteoporosis in women. There were greater numbers in the higher grades of spinal porosity among women treated with corticosteroids for more than one year, compared to those not treated or treated with corticosteroids for less than one year. In men, the number of cases in each grade was insufficient for chi square testing. However, we were able to compare the cortical thickness of the radius in men not treated with corticosteroids com0 ' " " " " ' ' ~ 20 M 40 50 60 70 pared to those treated for more than one Age lyearsl year. Again, no significant differences were Fig. 3.-Regression of radial cortex thick- shown between the groups. ness versus age in women treated for one to five years compared to those treated for more Functional Disability than five years with corticosteroids. The lines Fig. 4 shows that in general the average best fitting these points were calculated by the cortical thickness of the radius decreases method of least squares. with increasing grades of functional disone year. There was, however, a significant ability both in men and in women. Table difference between these two groups com- 6 shows how spinal porosity grades genbined compared to those treated for 1 to eraLly increase with increasing functional 2 years, 2 to 5 years, or more than 5 years disability in women. 0 Females treated lor more than 5 years with corticosteroids * Females h a t e d for 1 t o 5 years with corticosteroids respectively. The three groups of patients treated for more than one year had thinner radial cortices than women untreated or treated for less than one year. There was no difference in cortical thickness when women treated for 1 to 2 years were compared to those treated for 2 to 5 years. In women treated for more than 5 years with corticosteroids, the rate of loss of cortical thickness with age was significantly greater than in those treated for less than 5 years. This difference resides in the decrement of cortical thickness with increment of age, which is greater for those women treated for more than 5 years. This slope difference, however, could be attributed not to the presence of a thinner radial cortex in older women treated with corticosteroids, but rather to a thicker radial cortex in the younger women treated for more than 5 years, suggesting that corticosteroids may have a protective effect on cortical bone in younger women (Fig. 3 ) . Table 5 shows how the duration of treat- Spine Fractures Compression fractures of the spine occurred in 16 women, all of whom were over the age of 50, and 3 men, one of whom was over 50. These overall figures may be broken down as follows: of 52 women who were not treated with corticosteroids, 9 (16.3 per cent) had compression fractures of the spine, while of 42 women given corticosteroids, 7 (16.6 per cent) had compression fractures of the spine; the difference is obviously not significant. DISCUSSION Bradley and Ansell,6 observing 216 patients with Still's disease, found that in 63 maintained on steroids for 3 months or longer, 5 developed vertebral fractures, compared with 2 out of 153 patients who had a shorter or no steroid therapy. On the other hand, Iong bone fractures were equally common in both groups. These authors concluded that patients with Still's 428 SAVILLE &ID, KHARMOSH Table 5.-The Effect of Treatment with Steroids on Spinal Osteoporosis in Female Rheumatoid Patients Steroids Spine porosity 1 2 3 6 26 9 6 1 1 1 1 4 4 1.5 7 8 3 3 0 24 None 1 year 1-2 years 2-5 years > j years - Xam = 8.045 .02 P <= .05 Xaw:=1.389 P >.2 Xam = 4.892 .1 P .2 I. None vs. all treated < < 1 year < 1 year vs. > 1 year 11. None vs. 111. 7 < < 3 - Significant. Not significant. Not significant. vertebral fractures, while none in the 36 not so treated sustained such fractures. These authors remarked that the difference could have occurred by chance alone. Our findings confirm that the incidence of spinal compression fractures in rheumatoid patients treated with corticosteroids in *... .... .... .... ... the dosage reported here is no greater than .... ... in patients who have not received corticosteroids. Furthermore, an overall inci.... .... ... .... ... dence of spine fractures of 16 per cent in .... ... .... ... arthritic women between SO and 75 years .... ... of age can be compared to the incidence .... ... .... ... found ... in normal women by Smith and .... ... .... Frame.8 These authors, examining 2,063 I 2 I women seen in routine health examinations, IUN:TIDNAL STATUS found an incidence of wedging or compresFig. 4.-Shows the average cortical thick- sion fracture of the vertebrae increasing ness of the radial shaft in men and women from 6.8 per cent in the groups from 60 to against grades of functional status. 64,and 11.5 per cent for those from 65 to 69, to 20.3 per cent for women 70 to 74 disease, especially females contracting the years old. The incidence found in our padisease before 5 years of age, are especially tients would seem to fall within this range. prone to osteoporosis and fractures and It is also clear that while corticosteroids inthat this is uninfluenced by corticosteroid crease the incidence and seventy of osteotreatment. McConkey and his associates7 porosis, they do so significantly in this examined 97 patients with rheumatoid series only in patients over the age of 50; arthritis and found osteoporosis in 28 per and in men, only the spine seems to be cent riot treated with corticosteroids and affected. We have failed to show an ad33 per cent treated with them. This difference was not significant. However, 5 cases verse effect of small doses of prednisone on out of 61 in the treated group sustained spine density in men and women under .::..... Itmale o . . . I.. , I . . I . . . 429 OSTEOPOROSIS OF RA: INFLUENCE OF AGE, SEX, CORTICOSTEROIDS Table 6.-Steinbrocker’s Classification on Spinal Osteoporosis in Female Rheumatoid Patients ~ Classification 1 Mild, no disability 2 Moderate, some disability 3 Severe, disabled 4 Crippled, confined lvs.2f-3i-1 Spine porosity 0 1 2 3 11 19 2 41 1 32 2 1 9 - the age of 50. Furthermore, we have shown that cortical thickness was greater in the sixth decade in women treated for more than 5 years with prednisone than it was in women treated from 1 to 5 years. This suggests that in women under 50, it is possible, with small disease-suppressing doses of prednisone, to steer the patient between the Scylla of rheumatoid-induced osteoporosis and the Charybdis of prednisone-induced osteoporosis. As regards the duration of treatment with corticosteroids, the spine data suggest, and the radial cortex data confirm, that steroids exert a measurable effect on bone density after more than one year of treatment and that this effect is not progressive with time. Castillo and associates9 evaluated hand radiographs for cystic changes and osteoporosis and found that cystic changes were more common and osteoporosis uncommon in men, especially manual workers, while the reverse was true for women; cysts were uncommon and osteoporosis more common. These authors attributed the greater degree of osteoporosis in women to lack of physical activity, while the cysts and lack of osteoporosis in men they ascribed to increased activity in laboring men. Our data show that both cortical thickness of the radius and spinal porosity are significantly related to the degree of disability both in men and women. Whether this effect is attributable to disuse or in some other way to the rheumatoid processlo is uncertain. 5 1 Xats>=22.843 P <.001 - - However, it is clear that women have more osteoporosis than men because of a sudden change, presumably the menopause, which occurs in the sixth decade. While small doses of corticosteroids have been shown to play a role in the development of osteoporosis, age, sex and physical disability are more important. SUMMARY Osteoporosis has been evaluated from roentgenograms of the lumbar spine and left radius in 164 patients suffering from rheumatoid arthritis. Osteoporosis was a common finding in men and women whether treated with corticosteroids or not, especially in women over 50. Corticosteroids given in the dosage schedule reported here increased the osteoporosis of the spine in both sexes, mainly over the age of 50; the radial cortex became thinner on corticosteroids in women over 50 but not men, and this effect was observed after more than one year on treatment, but it was not progressive. Spine fractures occurred in 16 per cent of women, all of whom were over 50; the incidence was not greater in those patients treated with corticosteroids nor was it much greater than in normal North American women of comparable age reported in the literature. ACKNOWLEDGEMENT We would like to thank Dr. Melvin S. Schwartz, Laboratory of Biornetrics, The Hospital for Special Surgery, for statistical advice and computational assistance. 430 SAVILLE AMJ KIIARMOSI-I SUMMARJOIN INTERLINGUA Esseva effectuate un evalutation del osteoporose in 165 patientes rheumatoide a base de radiographias spinal e cubital. Osteoporose esseva presente a plus alte nivellos de incidentia in feminas de plus que 50 annos de etate, sin reguardo a si o lion illas habeva recipite corticosteroides. U n augment0 significative de osteoporose esseva causate per corticosteroides solo in masculos e femininas d e plus q u e 50 annos de etate e post plus q u e u n anno de tractamento. Le incidentia de fracturas spinal esseva 16 pro cento in feminas tractate con corticosteroide e in feminas non tractate con corticosteroide. Le parametros del etate, del sexo, e del morbiditate mesme exerce u n plus importante influentia super l e osteoporose q u e micre doses de corticosteroide. REFERENCES 1. Plotz, C., Knowlton, A., and Ragan, C.: The natural history of Cushing's Syndrome. Amer. J. Med. 13:597, 1952. 2. Ropes, M. W., Bennett, G. A., Cobb, S., Jacox, R. and Jessar, A. R.: Diagnostic criteria for rheumatoid arthritis-1958 Revision. Ann. Rheum. Dis. 18:49,1959. 3. Steinbrocker, O., Traeger, C. H., and Batterman, R. C . Therapeutic criteria in rheumatoid arthritis. J.A.M.A. 140:659, 1949. 4. Meema, H., Meerna, S.: Measurable roentgenologic changes in some peripheral bones in senile osteoporosis. J. Amer. Geriat. Soc. 11:1170, 1963. 5. Saville, P. D.: A quantitative approach to simple radiographic diagnosis of osteoporosis: Its application to the osteoporosis of rheumatoid arthritis. Arthritis Rheum. 6. Bradley, B. W. and Ansell, B. M.: Fractures 7. 8. 9. 10. in Still's disease. Ann. Rheum. Dis. 19:135, 1960. McConkey, B., Frazer, G. M., and Blight, A. S.: Osteoporosis and purpura in rhenmatoid disease; prevalence and relations to treatment with corticosteroids. Quart. J. Med. 31:419, 1968. Smith, R. W. Jr. and Frame, B.: Concurrent axial and appendicular osteoporosis: Its relation to calcium consumption. New. Eng. Mecl. 273:73, 1965. Castillo, B. A., El Sallab, R. A., and Scott, J. T.: Physical activity, cystic erosions, and osteoporosis in rheumatoid arthritis. Ann. Rheum. Dis. 24:522, 1965. Duncan, H., Frost, H. M., Villanueva, A. R. and Sigler, J. W.: The osteoporosis of rheumatoid arthritis. Arthritis Rheum. 8: 943,1965.