CELIAC DISEASE IS A common cause of intestinal malabsorption. Bone Loss in Celiac Disease Is Related to Secondary Hyperparathyroidism*

Transcription

1 JOURNAL OF BONE AND MINERAL RESEARCH Volume 14, Number 4, 1999 Blackwell Science, Inc American Society for Bone and Mineral Research Bone Loss in Celiac Disease Is Related to Secondary Hyperparathyroidism* PETER L. SELBY, 1 MICHAEL DAVIES, 1 JUDITH E. ADAMS, 2 and E. BARBARA MAWER 1 ABSTRACT Celiac disease is a major cause of intestinal malabsorption. Previous studies have demonstrated that celiac disease is associated with significant osteoporotic bone loss. These studies have suggested that successful treatment of the malabsorption is associated with amelioration of the bone loss. Such studies have failed to examine bone mass at peripheral skeletal sites which is more likely to be responsive to changes in parathyroid hormone (PTH) in response to calcium malabsorption. We have examined bone density in the lumbar spine, femoral neck, and distal forearm in 35 patients with celiac disease who had been established on gluten-free diet. In addition, the concentrations of PTH and 1,25-dihydroxyvitamin D (1,25(OH) 2 D) were measured. Bone density was below that expected for the subject s age and gender at all sites. This was most marked in the distal forearm where the bone density was 1.40 SD below expected (p < ). In the forearm, there was a negative relationship between bone density and PTH concentration (r = 0.49, p = 0.009). In the forearm and lumbar spine, there was a negative relationship between 1,25(OH) 2 D concentration and bone density. Bone mass was not related to the concentration of 25-hydroxyvitamin D at any of the skeletal sites measured. Bone density is reduced in the peripheral skeleton in celiac disease and this deficit persists despite treatment with apparent normalization at axial skeletal sites. This reduction in bone mass is related to the presence of secondary hyperparathyroidism which should be sought in all patients with treated celiac disease. (J Bone Miner Res 1999;14: ) INTRODUCTION CELIAC DISEASE IS A common cause of intestinal malabsorption affecting up to 1 in 2000 people. Since it affects the small intestine, which is the major site for calcium and vitamin D absorption, it is not surprising that disturbances of calcium and bone metabolism have long been recognized in patients with celiac disease. Although osteomalacia has often been thought of as the principal skeletal consequence of celiac disease, (1 3) the relatively recent advent of techniques for the measurement of bone mineral density (BMD) has led to the recognition that osteoporosis *This work was presented in part to the 1995 meeting of the Adult Working Group of the American Society of Bone and Mineral Research. is also common in vitamin D replete patients with celiac disease. (4 10) Indeed there is evidence to suggest that the prevalence of asymptomatic celiac disease is markedly increased in patients with osteoporosis. (11) The majority of such studies have been undertaken in subjects with recently diagnosed celiac disease, in whom a marked reduction of bone in the lumbar spine has been observed. More recently, several workers have reported that this deficit is largely attenuated by successful treatment of the celiac disease with a gluten-free diet. (9,10,12) Although some of the earlier studies have examined forearm bone mass, (4,5) little attention has been paid to bone loss at other skeletal sites, particularly in the appendicular skeleton. In celiac disease, malabsorption of calcium leads to secondary hyperparathyroidism. It is well recognized that excess parathyroid hormone (PTH) can be associated with 1 University of Manchester Bone Disease Research Centre, Department of Medicine, Manchester Royal Infirmary, Manchester, United Kingdom. 2 Department of Diagnostic Radiology, Stopford Building, Manchester, United Kingdom. 652

2 OSTEOPOROSIS IN CELIAC DISEASE 653 bone loss from the appendicular skeleton while axial bone mass is relatively well preserved. (13,14) By concentrating on the BMD of the lumbar spine, the previous workers might have therefore overlooked significant bone loss from other potentially important fracture sites. To determine the effect of celiac disease on the appendicular skeleton, we have examined bone mass at a variety of skeletal sites in a group of patients with treated celiac disease. We have also related these BMD measurements to the presence of abnormalities of calcium homeostasis as indicated by changes in the PTH vitamin D axis. Patients MATERIALS AND METHODS Thirty-five patients (26 female) with celiac disease previously diagnosed by intestinal biopsy and subsequently treated with a gluten-free diet were studied. Each patient had blood taken after an overnight fast for measurement of the serum concentrations of the calcium regulating hormones, PTH, and 1,25-dihydroxyvitamin D (1,25(OH) 2 D), together with 25-hydroxyvitamin D (25(OH)D), the main storage form of vitamin D. In addition, they attended the University of Manchester Department of Diagnostic Radiology, where BMD was measured at a variety of skeletal sites. Since these measurements were all undertaken as part of the normal clinical care of the patients, ethical approval was not necessary. Biochemical measurements Blood was allowed to clot immediately after venepuncture when it was immediately separated and the serum stored at 20 C until analysis. PTH was measured by an immunoradiometric assay, which detects intact PTH(1 84) molecules (Nichols Institute, Saffron Walden, U.K.). The normal range is pg/ml, and the inter- and intra-assay coefficients of variation (CVs) are 5.6% and 3.4%, respectively. 1,25(OH) 2 D was measured by radioimmunoassay following separation of the vitamin D metabolites by high performance liquid chromatography. (15) The reference range is pmol/l and the inter-and intra-assay CVs are 10.7% and 7.8%, respectively. 25(OH)D was measured by a competitive protein binding assay. (16) The normal range is nmol/1, and the inter- and intra-assay CVs are 8.7% and 6.5%, respectively. BMD measurements BMD was measured in the distal forearm by single photon absorptiometry using a Nuclear Data ND 1100 scanner (John Rouse Instruments, High Wycombe, U.K.). The proximal measurement site was chosen. This is almost exclusively cortical bone and the measurement precision (CV) at that site is 1%. Bone mineral content (g/cm) and BMD (g/cm 2 ) were measured. The BMD of the lumbar spine (L2 L4) and the right femoral neck were measured by dual-energy X-ray absorptiometry (DXA) using a Lunar DPX-L scanner (Lunar Corporation, Madison, WI, U.S.A.). Measurement precision is better than 1% for the spine and 2.5% for the femoral neck. BMD is measured in grams per square centimeter. Integral bone is measured with cortical/trabecular ratios of 50:50 in the lumbar spine and 60:40 in the femoral neck. (17) BMD in the lumbar spine was also measured by quantitative computed tomography (QCT) using a GE 9800 scanner (General Electric, Milwaukee, WI, U.S.A.). This provides a measure of spinal trabecular bone with a precision of 1%. In each patient a 10-mm slice was performed through the midplane of four vertebrae, T12 L3. A calibration phantom of dipotassium hydrogen phosphate was used and BMD was expressed in milligrams per milliliter as a mean value for the four vertebrae measured. The BMD at each site was expressed as a standard normal deviate (Z score) defined as: Z score (measured value age- and gender-matched mean value)/(age- and gender-matched SD) In each case, the manufacturer s normal range and SD was used to calculate the Z score. This technique eliminates the marked influence of both age and gender upon BMD and allows the effects of the disease to be examined directly. Statistical analysis Statistical analysis was undertaken using SPSS v7.0 (SPSS Inc., Chicago, IL, U.S.A.). With the exception of PTH, which was log normal, none of the other variables deviated significantly from a normal distribution. All statistical tests were accordingly carried out following logarithmic transformation of PTH. Standard parametric tests were used to examine the deviation of each of the BMD variables from zero and also to look at the association between the variables. Clinical state RESULTS The mean age of patients was 50.6 ± 13.4 (SD) years. All the patients had been diagnosed for at least 1 year. The mean duration of diagnosis was 9.9 ± 9.3 (SD) years. All patients claimed to be compliant with their gluten-free diet. Response to the diet had not been assessed by repeat biopsy, but in the majority of cases there had been a clinical response to treatment. Of the 26 women studied, 17 were postmenopausal. Six of these were receiving hormone replacement therapy at the time of study and a further three had received such therapy in the past. Information on vitamin D supplementation was available in 29 subjects; 9 patients were taking such supplements in a dose of g ( IU/day), and 1 patient had received such therapy in the past, but the majority had never received vitamin D supplementation. One patient had evidence of an irondeficient anemia but no other nutritional deficiencies were manifest. Metabolic state None of the subjects was vitamin D deficient (Fig. 1). Plasma calcium (corrected for albumin concentration) was

3 654 SELBY ET AL. FIG. 1. Serum concentrations of the calcium-regulating hormones in patients treated for celiac disease. 25D, 25- hydroxyvitamin D; 1,25D, 1,25-dihydroxyvitamin D; PTH, parathyroid hormone. For each hormone the shaded box represents the normal reference range. normal at 2.21 mmol/l (95% confidence interval [CI] ) as was the plasma phosphate at 1.02 mmol/l (95% CI ). Renal function was normal as was plasma alkaline phosphatase activity. The mean 25(OH)D was 52.7 nmol/1 (95% CI ). However, there was evidence of secondary hyperparathyroidism in 28% of subjects (Fig. 1) with a geometric mean PTH of 37 pg/ml (95% CI 29 48). This was accompanied by elevated serum concentrations of 1,25(OH) 2 D, mean pmol/1 (95% CI ) with 44% of values being above the reference range (Fig. 1). The concentration of vitamin D was not significantly different in those recorded as taking vitamin D supplements. There was the expected negative relationship between serum PTH and 25(OH)D concentrations (r 0.662, p < 0.001) (Fig. 2) and a positive relationship between the serum concentrations of PTH and 1,25(OH) 2 D(r 0.488, p 0.005) (Fig. 2). There was no relationship between the serum concentrations of any of the calcium-regulating hormones and the age of the patient or the duration of disease. Bone mineral density The results of the BMD measurements at the various sites are given in Fig. 3. Spinal BMD measured with DXA was not significantly different from that expected for each subject s age and gender, mean Z score 0.44 (SE 0.29, p vs. zero 0.15). The same was also true at the femoral neck, mean Z score 0.27 (SE 0.21, p 0.22). When spinal trabecular BMD was estimated using QCT, the results were almost identical to those obtained by DXA; the mean Z score being 0.35 (SE 0.27, p 0.21). In the forearm, the deficit was more pronounced and reached statistical significance with a mean Z score of 1.40 (SE 0.22, p < ). BMD measured at each skeletal site was correlated with that measured elsewhere in the skeleton. The correlation coefficients ranged between 0.40 and There was a negative relationship between the BMD measured at the forearm and serum PTH concentration (Fig. 4; r 0.48, p 0.009). There was no relationship between serum PTH FIG. 2. Inter-relationship between calcium-regulating hormones in patients with treated celiac disease. and BMD at any of the other sites. There was a similar negative correlation between BMD in the forearm and the serum concentration of 1,25(OH) 2 D (Fig. 4; r 0.50, p 0.007). In addition, there was a weaker negative relationship between the serum concentration of 1,25(OH) 2 D and BMD as measured in the lumbar spine by either DXA (r 0.43, p 0.018) or QCT (r 0.46 n 25, p 0.018). Adjusting for the relationship between PTH and 1,25(OH) 2 D concentrations by taking partial correlation coefficients reduced the strength of these relationships. In the proximal forearm, the partial correlation between bone mass and 1,25(OH) 2 D allowing for PTH was 0.34 (p 0.081) and in the spine by DXA it was 0.36 (0.054). There was no relationship between BMD at any of the sites measured and the serum concentrations of 25(OH)D (Fig. 4).

4 OSTEOPOROSIS IN CELIAC DISEASE 655 FIG. 3. BMD (mean standard error) at the various sites measured in patients treated for celiac disease. ***p < Equally there was no significant difference in bone mass between those receiving vitamin D supplementation and those not. DISCUSSION FIG. 4. Relationship between calcium-regulating hormones and BMD in patients with treated celiac disease. Several workers have reported that BMD is reduced in patients with celiac disease. (4 10) For the most part, these studies have examined BMD in subjects in whom the disease is newly diagnosed. More recently, some studies have suggested that following the institution of gluten-free diet there is an improvement in BMD. (9,10,12) Such studies have concentrated on BMD at the spine and femur and have not reported changes in the appendicular skeleton. In as much as the BMD in the femoral neck and spine, when measured by DXA, was not significantly reduced in our patients who had been on a long-term gluten-free diet the results reported here would confirm these observations. However, our results, together with those of other workers, do need to be interpreted with some degree of caution. In particular the changes in BMD observed in the spine and femur are relatively small and so there is a substantial risk of a type 2 statistical error. Indeed nearly 120 subjects would be necessary to have 80% power to detect changes of the magnitude detected in our patients. Given the known relationship between bone mass and fracture risk, even such a modest loss of BMD might be expected to lead to a 35% increase in osteoporotic fracture rates (18) and so could be of major clinical importance in patients with celiac disease. Two relatively small studies of bone mass in celiac disease have demonstrated a deficit in bone mass at the forearm but offer conflicting evidence as to the effect of a gluten-free diet on this. (4,5) The more recent studies have failed to examine BMD in the appendicular skeleton. Since it is known that hyperparathyroidism leads to preferential bone loss from this part of the skeleton, it is no surprise that bone loss is evident at this site, even in subjects in whom bone is ostensibly well preserved centrally. This may indicate a propensity for patients with celiac disease, even after treatment, to be at an even greater increased risk of fracture, especially in peripheral sites where cortical bone predominates. The magnitude of the BMD deficit at the forearm suggests that the risk of fractures of the spine, hip, or forearm might still be doubled in patients with treated celiac disease. Even though causation cannot be assumed from correlation, the strong inverse relationship between BMD and plasma PTH suggests that the loss of bone may be related to the presence of secondary hyperparathyroidism. This hypothesis would be amenable to testing by examining whether the BMD of patients with celiac disease improves when the secondary hyperparathyroidism is suppressed. To some extent it might be argued that this has already been demonstrated by the study of McFarlane et al. (9) ; however, that study did not examine BMD changes in sites such as the distal radius which are most susceptible to PTH. Hence it might well have been relatively inefficient at detecting such changes.

5 656 Although the relationship between forearm BMD and PTH was weaker than that between BMD and 1,25(OH) 2 D we believe that this latter relationship may be of less physiological significance than the former. PTH concentration is one of the major determinants of 1,25(OH) 2 D concentration. This is illustrated in our own results by the correlation between the concentrations of these two hormones (Fig. 2). It is therefore important to allow for the effect of PTH on 1,25(OH) 2 D concentration before comparing the effects of these two hormones on bone mass. If allowance is made for this by calculating the partial correlation between 1,25(OH) 2 D and proximal radial bone density allowing for the effect of PTH, there is a reduction in the observed correlation coefficient from 0.50 to This residual relationship, although no longer statistically significant, raises the possibility that we may be seeing evidence of the intrinsic bone resorbing activity of 1,25(OH) 2 D (19) over and above that of PTH. This is of a similar magnitude to the effect of 1,25(OH) 2 D elsewhere in the skeleton and accounts for 12% of the variation in bone mass at this site compared with 23% of the variation in forearm bone mass which is attributable to changes in PTH. The fact that the majority of our patients had normal vitamin D nutrition as measured by serum 25(OH)D concentrations excludes gross vitamin D deficiency as the cause of secondary hyperparathyroidism. However, in the U.K. the normal range for vitamin D is such that many subjects have levels that we have previously defined as vitamin D insufficient (20) and hence capable of increasing PTH levels. Nonetheless, we do not believe that these patients with celiac disease differed in this respect from the general U.K. population since, provided that patients receive adequate solar exposure, vitamin D supply is not a problem even when there is steatorrhea. (21) It is well described that patients with celiac disease have increased endogenous fecal calcium losses (3) and it is likely that the major stimulus to secondary hyperparathyroidism is calcium deficiency consequent upon this. (22) Indeed the secondary hyperparathyroidism seen following gastrointestinal calcium loss can lead to increased catabolism of vitamin D and might contribute to vitamin D deficiency. (23) While vitamin D deficiency was not a problem in these patients the fact that this mechanism is operative in celiac disease can be seen from the strong negative relationship between the serum concentrations of PTH and 25(OH)D (Fig. 2). It follows that the appropriate means of alleviating this calcium deficiency is by the administration of calcium supplements to promote increased calcium absorption and hence bring about a neutral calcium balance. (23) It is impossible to determine what the optimal calcium intake would be from our data but the balance studies of Melvin et al. suggest that in untreated subjects around 2 g of calcium intake per day is necessary to ensure a neutral balance. (3) This is 0.5 g over that required to maintain neutral balance in patients with osteoporosis or normal postmenopausal women. (24) The relationship between bone loss and PTH offers a potential means of monitoring the treatment of celiac disease and identifying those patients at particular risk of osteoporosis. Although facilities for the measurement of BMD are not uniformly available, it is always possible to obtain blood for the estimation of PTH. We would recommend that patients with celiac disease who have raised levels of PTH should receive adequate doses of calcium supplements to ensure that PTH is returned to the normal range even if the patient is believed to be complying with a gluten-free diet. REFERENCES SELBY ET AL. 1. Juergens JL, Scholz DA, Wollager EE 1956 Severe osteomalacia associated with occult steatorrhoea due to non tropical sprue: Report of 5 cases. Arch Intern Med 98: Hajjar ET, Vincenti F, Salti CS 1979 Gluten induced enteropathy: Osteomalacia as a principal manifestation. Arch Intern Med 131: Melvin KEW, Hepner GW, Bordier P, Neale G, Joplin GF 1970 Calcium metabolism and bone pathology in adult coeliac disease. Q J Med 39: Caraceni MP, Molteni N, Bardella MT, Ortolani S, Nogara A, Bianchi PA 1988 Bone and mineral metabolism in adult celiac disease. Am J Gastroenterol 83: Bode S, Hassager C, Gudmand-Hoyer E, Christiansen C 1991 Body composition and calcium metabolism in adult treated coeliac disease. Gut 32: Mazure R, Vazquez H, Gonzalez D, Mautalen C, Pedreira S, Boerr L, Bai JC 1994 Bone mineral affection in asymptomatic adult patients with celiac disease. Am J Gastroenterol 89: Valdimarsson T, Toss G, Ross I, Lofman O, Strom M 1994 Bone mineral density in coeliac disease. Scand J Gastroenterol 29: Pistorius LR, Sweidan WH, Purdie DW, Steel SA, Howey S, Bennett JR, Sutton DR 1995 Coeliac disease and bone mineral density in adult female patients. Gut 37: McFarlane XA, Bhalla AK, Robertson DAF 1996 Effect of a gluten-free diet on osteopenia in adults with newly-diagnosed celiac-disease. Gut 39: Corazza GR, Di Sario A, Cecchetti L, Tarozzi C, Corrao G, Bernandi M, Gasbarrini G 1995 Bone mass and metabolism in patients with celiac disease. Gastroenterology 109: Lindh E, Ljunghall S, Larsson K, Lavo B 1992 Screening for antibodies against gliadin in patients with osteoporosis. J Intern Med 231: Valdimarsson T, Lofman O, Toss G, Strom M 1996 Reversal of osteopenia with diet in adult celiac-disease. Gut 38: Whitehouse RW, Adams JE, Davies M, Adams PH 1993 Factors influencing bone mass in primary hyperparathyroidism. Calcif Tissue Int 52:178 (abstract). 14. Wishart J, Horowitz M, Need A, Nordin BE 1990 Relationship between forearm and vertebral mineral density in postmenopausal women with primary hyperparathyroidism. Arch Intern Med 150: Mawer EB, Berry JL, Cundall JP, Still PE, White A 1990 A sensitive radioimmunoassay using a monoclonal antibody that is equipotent for ercalcitriol and calcitriol (1,25-dihydroxyvitamin D 2 and D 3 ). Clin Chim Acta 190: Selby PL, Davies M, Marks JS, Mawer EB 1995 Vitamin D intoxication causes hypercalcaemia by increased bone resorption which responds to pamidronate. Clin Endocrinol 43: Faulkner KG, Gluer CC, Majumdar S, Lang P, Engelke K, Genant HK 1991 Noninvasive measurements of bone mass, structure and strength: Current methods and experimental techniques. Am J Radiol 157: Marshall D, Johnell O, Wedel H 1996 Meta-analysis of how well measures of bone mineral density predicts occurrence of osteoporotic fracture. BMJ 312: Raisz LG, Trummel CC, Holick MF, DeLuca HF ,25-

6 OSTEOPOROSIS IN CELIAC DISEASE 657 dihydroxycholecalciferol: A potent stimulator of bone resorption in tissue culture. Science 175: Peacock M, Selby PL, Francis RM, Brown WB, Hordon L 1985 Vitamin D deficiency, insufficiency, sufficiency and intoxication: What do they mean? In: Norman A, Schaefer K, Grigoleit MG, Herrath D (eds.) Vitamin D: A Chemical, Biochemical and Clinical Update. Walter de Gruyter, Berlin, Germany, pp Mawer EB 1997 Vitamin D deficiency in patients with intestinal malabsorption. Nutrition 13: Mawer EB, Davies M 1997 Bone diseases associated with gastrointestinal and hepatobiliary disease. In: Feldman D, Glorieux FH, Pike JW (eds.) Vitamin D. Academic Press, San Diego, CA, U.S.A., pp Davies M, Heys SE, Selby PL, Berry JL, Mawer EB 1997 Increased catabolism of 25-hydroxyvitamin D in patients with partial gastrectomy and elevated 1,25-dihydroxyvitamin D levels: Implications for metabolic bone disease. J Clin Endocrinol Metab 82: Selby PL 1994 Calcium requirement A reappraisal of the methods used in its determination and their application to patients with osteoporosis. Am J Clin Nutr 60: Address reprint requests to: Dr. P.L. Selby University Department of Medicine Manchester Royal Infirmary Oxford Road Manchester M13 9WL, U.K. Received in original form December 23, 1998; in revised form September 16, 1998; accepted December 15, 1998.