Digestive Diseases and Sciences, Vol. 37, No. 9 (September 1992), pp. 1396-1403

Low Bone Mineral Density in Patients with Inflammatory Bowel Disease FRAN(~OIS PIGOT, MD, CHRISTIAN ROUX, MD, STANISLAS CHAUSSADE, MD, DANIEL HARDELIN, MD, ODETTE PELLETER, THIERRY DU PUY MONTBRUN, MD, VERONIQUE LISTRAT, MD, MAXIME DOUGADOS, MD, DANIEL COUTURIER, MD, and BERNARD AMOR, MD

To assess the prevalence and risk factors for low bone mineral density in inflammatory bowel disease, we studied 61 consecutive patients, mean age 36 + 11 years. Twenty-seven had a Crohn' s disease and 34 ulcerative colitis (including 13 with ileoanal anatomosis). Three patients, two women and one man (32, 70, and 45 years old, respectively) had vertebral crush fractures. Bone mineral density measured by dual energy x-ray absorptiometry at spine and femoral level was more than 2 so below normal values in 23% o f the patients, all o f them having received steroid therapy. Eighteen patients (29%) had never received steroid therapy; their bone mineral density was not different than those who had. Univariate analysis showed a positive correlation between bone mineral density and body weight or oral calcium intakes, and a negative correlation with steroid daily dose. After ileoanal anastomosis, bone mineral density was not different from other groups and showed a positive correlation with time elapsed since coloproctectomy. We concluded that bone mineral density is low in patients with inflammatory bowel disease and exposes them to the risk o f bone fracture. Bone mineral density after ileoanal anastomosis may increase with time after surgery. KEY WORDS" bone mineral density; osteoporosis; inflammatory bowel disease; ileoanal anastomosis.

Patients with inflammatory bowel disease (IBD) are at risk for low bone mineral density because of reduced calcium intake, malabsorption of vitamin D, reduced activity, and amenorrhea. These physical and life-style factors act on bone mineral content (1), and it is well known that bone loss can be modified by corticosteroid therapy (2). It has been demonstrated that a decrease of one standard deviation (sD) in the bone density at peripheral sites gives a relative risk of hip fracture of 1.66 (3). Manuscript received October 1, 1991; revised manuscript received February 24, 1992; accepted February 27, 1992. From the Service d'hrpato-gastroentrrologie, Clinique de Rhumatologie, Hrpital Cochin, Paris; and Hrpital Lropold Bellan, Universit6 Ren6 Descartes, Paris, France. Address for reprint requests: Dr. Stanislas Chanssade, Service de Gastroentrrologie, Hrpital Cochin, 27, rue du Faubourg Saint Jacques, 75674 Pads Cedex 14, France.

1396

Osteoporosis has recently been reported to be frequent among patients with IBD, mostly after small bowel resection (4, 5). Patients with IBD have a quasinormal expectancy of life. Among these young patients, vertebral crush fractures or fractures of the appendicular skeleton may represent a serious handicap. It is now possible to measure bone mineral content with noninvasive methods. Thus, it seems logical to screen these patients in order to detect early decrease of bone mineral content. In previous studies, bone mass measurements were based on CT scanning of lumbar spine and single photon absorptiometry at peripheral site. The relative lack of precision of these methods and the important radiation exposure induced by CT scanning make dual energy x-ray absorptiometry (DEXA) a safer and more relevant diagnostic tool. Digestive Diseases and Sciences, I1ol. 37, No. 9 (September 1992)

0163-2116/92/0900-1396506.50/09 1992PlenumPublishingCorporation

BONE MINERAL DENSITY IN INFLAMMATORY BOWEL DISEASE F u r t h e r m o r e , D E X A allows a better evaluation of femoral neck bone density, which is a clinically relevant site. It is of importance that interpretation o f bone density values could be based on comparison with normal data obtained from a control population explored with the same device, and of similar geographic origin for w o m e n (6). The aims o f this prospective study were: (1) to assess the prevalence o f low bone mineral density in unselected outpatients with IBD using D E X A , and (2) to evaluate potential risk factors from clinical data, nutritional study, and dietary inquiry.

MATERIALS AND METHODS All consecutive patients with small and/or large bowel inflammatory disease referred to the gastroenterological unit between January to July 1990 were prospectively included in the study. All patients had a diagnosis of Crohn's disease or ulcerative colitis established on histological, endoscopic, radiological, and clinical criteria (7) before the study. Patients were in clinical remission or had active disease according to the Crohn's disease activity index (8) and Truelove's or endoscopic criteria for ulcerative colitis (9, 10). Patients with ulcerative disease limited to the rectum were excluded. For patients with prior coloproctectomy and ileoanal anastomosis, the diagnosis of pouchitis was based on the presence of endoscopic lesions, as previously described (11). Duration of the disease (time elapsed since the first typical clinical manifestation to the date of the study), maximal digestive extension during evolution (assessed by radiological and/or endoscopic examination), site and length of eventual digestive resections (established from preoperative data) were recorded. Cumulative life time steroid dose was expressed in milligrams of prednisone. The daily dose represented the ratio between cumulative steroid dose and duration of the disease and was expressed in milligrams of prednisone per day. Use of steroid enemas was noted. Other therapies such as azathioprine, vitamin D, and calcium supplements were documented. Clinical nutrition status was assessed by physical examination on the day of the study as normal nutritional status, mild or severe malnutrition (12). Body weight was expressed in kilograms and percent of ideal body weight (13). Body mass index (BMI), which is the heightadjusted weight, was expressed in kilograms per square meter. Physical activity during the month before the study was classified as: 0 = fully active, I = ambulatory, capable of light work, II = in bed less than 50% of time, capable of self-care but not of work activities, III = in bed more than 50% of time, capable of only limited self-care, and IV = completely bedridden (14). In women hormonal status was clinically assessed by menstrual history; eventual hormonal therapy was mentioned. Menopause was defined as amenorrhea for more than six months. Dietary Inquiry. A dietary inquiry retrospectively estimated the total calories, proteins, fat, carbohydrates, Digestive Diseases and Sciences, Vol. 37, No. 9 (September 1992)

fibers, and calcium oral intakes as previously described (15). Basal metabolic expenditure was calculated according to Harris and Benedict (16). Spinal Radiography and Bone Mineral Density. Vertebral crush fractures were assessed on dorsal and lumbar spine standard x-ray examination. The use of ultrasensifive films allowed minimal x-ray exposure. Bone mineral density was measured by DEXA (Hologic QDR 1000) at lumbar spine level (second, third and fourth lumbar vertebrae). At the proximal left femur level, four sites (femoral neck, trochanter and intertrochanter area, and Ward's triangle) were measured; an average (total femur) was obtained from the first three sites, Bone mineral density results were expressed as the number of standard deviations from normal values corrected for sex and age (Z score). Normal values for men were supplied by Hologic. As for women, geographical origin is of importance (6); normal values were obtained from the study of a group of 397 normal women living in the Pads area, using the same device (17). The reproducibility of the method in our hands for normal and osteoporotic patients at lumbar spine levels is less than 1%. Total duration of DEXA never exceeded 20 min. Biochemistry. Serum concentrations of calcium, albumin, and alkaline phosphatase were measured by standard methods. Statistical Analysis. Results were expressed as mean +SD (range). Differences between groups were analyzed using Student's unpaired test, comparison of incidences was made using chi-square test. Correlations were examined by linear regression analysis. RESULTS

Patients. Of 73 patients referred to the unit with a diagnosis of inflammatory bowel disease, 61 patients (26 male, 35 female), mean age 36 --- 11 (19-70) years agreed to participate to the study (Table 1). In four patients, measurements of vertebral bone density were not interpreted because of spondylarthropathic or osteoarthritic lesions. T w e n t y - s e v e n patients had Crohn's disease (12 patients with small bowel involvement, including five with a previous > 15-cm small bowel resection), 34 had ulcerative colitis [including 13 patients with prior c o l o p r o c t e c t o m y and ileoanal anastomosis performed 31 (5-53) months before the study]. At the time of the study, the Crohn's disease activity index was more than 130 in 17 patients, while l0 patients with Crohn's disease were in clinical remission. Six patients with ulcerative colitis were in clinical remission, while 15 had active disease using T r u e l o v e ' s criteria. N o n e of the patients with ileoanal anastomosis had biological inflammatory syndrome. Three of them had pouchitis, with endoscopic lesions in the reservoir (11).

1397

PIGOT ET AL TABLE 1. CHARACTERISTICS OF 61 PATIENTS WITH INFLAMMATORY BOWEL DISEASE INCLUDED IN STUDY

Patients Characteristics

Crohn' s disease

Ulcerative colitis

Number Sex (F/M) Age (years) Weight(kg) BMI (kg/m 2) Digestive resection Small bowel ( > 15 cm) Colon ( > 50 era) Disease duration (mo) Steroid dose Total lifetime (mg) Daily (mg/day) Calorie intake (kcal/day) Calcium intake (mg/day) Fiber intake (g/day)

27t 17/10 34 --13 59 11 21.4 3.2

21 11/10 36 -- 10 61 9 11 21.9 • 3.3

5 0 63 11029 7.6 1893 690 15

•

75

• 15291 • 8.3 • 625 • 377 • 5

0 1 • 100

75 2548 4.0 2292 694 15

- 3962 6.2 - 608 • 392 • 6

lleoanal anastomosis

13 7/6 38 --63 22.3 •

P value*

NS 10 12 2.6

31

3 13 • 16~t

5223 4.4 2164 618 16

• 7014 • 7.1 • 692 • 280 • 9

0.023w NS NS (0.06)w NS NS

tValues given are either the number of patients with the characteristic or mean - standard deviation. *Statistical significance determined by the Student's unpaired test between each of the three groups, and chi-square test for comparison of incidences. A P value less than 0.05 was considered significant. ~tElapsed time between coloproctectomy and the present visit. w Crohn's disease and ulcerative colitis groups.

Eighteen (29%) patients had never received oral steroid therapy. Thirteen of these patients were treated with sulfasalasine, and five with 5-aminosalicylic acid (Pentasa). After coloproctectomy no patient received steroids longer than two months. Steroid enemas had been used in nine patients (four with ulcerative colitis, and five with ileoanal anastomosis); in three cases it had not been associated with systemic steroid therapy. Three patients with Crohn's disease and one with ulcerative colitis were receiving azathioprine (1.5-1.75 mg/kg body weight). All patients had a normal clinical nutritional status. Body mass index was 21.79 --4- 3.09 kg/m2 (range: 16.84-29.41) in the 61 patients. No patients complained of bone pain. Physical activity was stage III in two cases, II in six cases, and normal in the other patients. Of the 35 women, four had irregular menses and four were menopausal, none received hormone replacement therapy. Three patients with Crohn's disease and one with ileoanal anastomosis had serum calcium levels below the normal range, including one patient with elevated alkaline phosphatase (with histologically proven sclerosing cholangitis). Their Z scores at lumbar spine level were +0.61, +0.05, -1.50, and -3.00, respectively. Dietary Inquiry. Mean total caloric intake was 2080 --- 628 (1006-3529) kcal/day, which repre-

1398

sented a caloric intake of 1.48 --- 0.40 (0.74-2.21) times their basal metabolic expenditure. Protein, fat, and carbohydrate represented 16%, 36%, and 48% of the total caloric intake, respectively. Calcium intake was 675 --- 357 mg/day and showed a significant positive correlation with total calorie intake (r = 0.57, P = 0.0001). Mean fiber intake was 15 - 6 g/day. None of the patients had been taking vitamin D or a calcium supplement. Bone Examination (Figure 1). Three patients had vertebral crush fractures on x-ray examination (a 32-year-old woman with an ileoanal anastomosis and normal menses, a 45-year-old man, and a 70year-old woman with Crohn's disease). L u m b a r Z scores 2"

-o-

Fig 1. Lumbar bone mineral density of 61 patients with inflammatory bowel disease expressed as standard deviations from normal values corrected for sex and age (Z scores). Digestive Diseases and Sciences, Vol. 37, No. 9 (September 1992)

BONE MINERAL DENSITY IN INFLAMMATORY BOWEL DISEASE TABLE 2. BONE MINERAL DENSITY EXPRESSEDAS Z SCORES (NUMBER OF STANDARD DEVIATIONS FROM NORMAL VALUES CORRECTED FOR SEX AND AGE) OF 61 PATIENTS WITH INFLAMMATORYBOWEL DISEASE Patients

Number of patients Z scores Lumbar spine Femoral neck Ward's triangle Total femur Number of patients (%) with: Z score < 1 SD~: Z score < 2 SD$

Crohn's disease

Ulcerative colitis

lleoanal anastomosis

27

21

13

-1.11 -0.83 -0.66 -0.88

-+ -+ -+ -+

1.37t 1.66 1.42 1.25

-0.93 -0.70 -0.74 -0.70

16 (59) 8 (30)

-+ 1.21 +- 1.28 -+ 1.26 -+ 0.94

-1.53 -1.17 -1.34 -1.03

12 (57) 6 (29)

-+ -+ -+ +-

P value*

1.73 1.24 1.27 1.31

NS NS NS NS

8 (61) 3 (23)

NS NS

tValues given are mean -+ SD. *Statistical significance determined by the Student's unpaired test between each of the three groups, and the chi-square test for comparison of incidences. A P value less than 0.05 was considered significant. SAt lumbar and/or total femoral site.

Results of DEXA bone examination are reported in Table 2 for each group. Mean lumbar spine Z score in our population was -1.14 _+ 1.40 (-5.10, 1.83); it was -1.26 -+ 1.38 for males, and -1.05 --1.43 for females. At the femoral neck level, mean Z score was -0.86 --- 1.44 (-4.03; 3.25), -0.90 --1.46, -0.84 - 1.46 for all patients, males, and females, respectively. A significant positive correlation was found between lumbar and femoral Z scores (r = 0.72, P < 0.0001).

Patients who never received steroid therapy were compared to the others (Table 3); patients with bone mineral density more than 2 SD below normal values are reported in Table 4. These data show that patients with Crohn's disease received steroid therapy more often than others. No patient without, but 33% of those with, previous steroid therapy had a Z score more than 2 SD below the normal value corrected for sex and age. The Z scores and BMD were not different between patients who were on

TABLE 3. COMPARISONS BETWEEN PATIENTS WITH AND WITHOUT PREVIOUS STEROID THERAPY IN 61 PATIENTS WITH INFLAMMATORYBOWEL DISEASE With steroid Number Sex ratio (M/F) Age (years) Weight (kg) BMI (kg/m2) Diagnosis Crohn's disease Ulcerative colitis Ileoanai anastomosis Steroid dose (mg): Total Daily Z score~t Lumbar spine Total femur Femoral neck Ward's triangle Number (%) with a Z score