Journal of Internal Medicine 1998; 243: 123–126

© 1998 Blackwell Science Ltd 123

Bone mineral density in patients with Crohn’s disease duringlong-term treatment with azathioprine

C.-H. FLORÉN a , B. AHRÉN b , M. BENGTSSON a , J. BARTOSIK a & K. OBRANT c

From the Departments of Internal Medicinea and Orthopedicsc, and the Wallenberg Laboratoryb, Malmö University Hospital, Lund University, Malmö,Sweden

Abstract. Florén C-H, Ahrén B, Bengtsson M,Bartosik J, Obrant K (Lund University, Malmö,Sweden). Bone mineral density in patients withCrohn’s disease during long-term treatment withazathioprine. J Intern Med 1998; 243: 123–26.

Objectives. To ascertain whether patients withCrohn’s disease treated with azathioprine main-tained bone mineral mass better than patients treat-ed with steroids alone.Design. Retrospective study.Setting. University Hospital of Malmö, Sweden.Subjects. A total of 59 patients with ileocolonic, ileo-caecal or colonic Crohn’s disease.Methods. Bone mass was assessed by dual photon X-

ray absorptiometry at the level of L2 – L4.Results. Patients treated with a high lifetime dose ofsteroids (. 5 g prednisolone) had significantly(P 5 0.011) lower Z-score of L2–L4 (20.87 6 1.11;11 SD) than steroid-treated patients, who hadreceived a low dose of prednisolone (, 5 g) (0.08 61.16 SD). Azathioprine did not negatively influencethe steroid effect on bone mineral density.Conclusions. Azathioprine does not seem to affectbone mineral density by itself. However, by beingsteroid-saving, it seems to conserve bone mineralmass in patients with Crohn’s disease.

Keywords: azathioprine, bone mineral density, corti-costeroids, Crohn’s disease, prednisolone.

Introduction

In patients with Crohn’s disease, one of the corner-stones of treatment is corticosteroids, which are usedto treat exacerbations of disease activity, and arethen tapered off when disease activity diminishes.Several studies have convincingly demonstrated theclinical value of this treatment [1–3]. Steroids are,though, afflicted with a variety of adverse effects,including negative effects on bone mineral mass [4].This effect is of importance in Crohn’s disease, sincethese patients usually already have a low bone min-eral mass when diagnosed [5], and steroids cantherefore exacerberate an already negative balance ofbone mineral turnover [6]. Therefore, to diminishthis negative effect, a low dose of steroids is desirable.This may be accomplished by addition of substanceswithout negative effects on bone mineral mass,which reduce the requirement for steroids.

In patients with Crohn’s disease, a treatmentschedule of sparing steroids by combining steroids

with immunosuppressants like azathioprine hasemerged [2]. Placebo-controlled studies have shownthat azathioprine together with steroids inducesremission significantly more than steroids alone, andalso that azathioprine maintains remission inpatients with Crohn’s disease significantly betterthan placebo [7]. However, whether azathioprineaffects bone mineral mass in steroid-treated patientswith Crohn’s disease is not known. In order to evalu-ate whether azathioprine is beneficial to bone miner-al mass, we invited patients with ileocolonic,ileocaecal or colonic Crohn’s disease to undergo bonedensitometric measurements. The specific questionposed was whether patients treated with azathio-prine maintained bone mineral mass better thanpatients treated with steroids alone.

Materials and methods

At the Division of Gastroenterology and Hepatology,Department of Internal Medicine at the Malmö

C.-H. FLORÉN et al.124

© 1998 Blackwell Science Ltd Journal of Internal Medicine 243: 123–126

University Hospital, Malmö, all patients, less than70 years of age, with Crohn’s disease, localized to thedistal 3 dm of ileum together with involvement incaecum or in colon, or in colon alone, were invited tohave their bone mineral density measured. A total of59 patients accepted. The 31 females and 28 maleshad a mean disease history since diagnosis of10 years (range 1–30 years) and the mean age was38 years (range 20–69 years). In order to avoidpatients with absorption defects of calcium and vita-min D, patients with small bowel disease or proximalbowel disease were not included in the study. Also,the therapeutic tradition in Malmö has primarilybeen to treat Crohn’s disease patients with colonicinvolvement with azathioprine [8]. Of these 59patients, 24 had ileocolonic disease, 17 ileocaecaldisease and 18 colonic disease. The diagnosis and dis-ease extent in all patients was confirmed by histologyand endoscopical and radiological examinationsaccording to established criteria [3]. All the medicalrecords from these patients were reviewed and totallifetime doses of steroids and azathioprine were cal-culated. A total of 21 patients had not been treatedwith steroids or azathioprine, and formed the controlgroup. Of these, 15 had been treated surgically (onehad a colectomy, two had right-sided hemicolec-tomies and 12 had ileocaecal resections). The otherpatients were treated either with steroids alone(n 5 18) or with a combination of azathioprine and

steroids (n 5 20). The current dose of steroids in the38 steroid-treated patients was nil, except in sixpatients, who were treated with doses from 2.5 to20 mg (mean 12.5 mg).

The bone mineral density was measured by dualphoton X-ray absorptiometry (Lunar DPX-L,Madison, USA). A preset algorithm automaticallycalculated the bone mineral density of the lumbarspine as the mean of three different vertebrae, L2–L4,as well as the femoral neck. The precision of themethod has been shown to be 0.5% for the lumbarspine and 1.6% for the hip (neck) [9].

Since the probands represented different ages,comparison was made with Z-scores from normativedata provided by the manufacturer of the equipment[10]. This reference material does not differ fromSwedish reference material from Malmö [9]. Therewas a significant correlation between the Z-scores ofL2–L4 and the hip (r 5 0.67; P , 0.001, n 5 59)(Fig. 1), and because of this henceforth only the Z-score of L2–L4 is presented. Moreover, the correlationbetween Z-score in the hip and Z-score in the lumbarregion is not dependent on age (, 30 years; n 5 14;r 5 0.58 compared to . 60 years; n 5 6; r 5 0.78).No difference existed between the males (r 5 0.73)and females (r 5 0.58). In the patient material onlysix patients were older than 60 years and none hadvertebral deformities.

Statistics

Mean 6 SEM is demonstrated unless otherwisestated. Statistical comparisons between groups wereperformed with Mann–Whitney U-test, withKruskal–Wallis test for multiple comparisons.

Results

Of the 59 patients investigated with disease localizedto the distal ileum and colon or colon alone, 21 hadnot been treated with steroids and azathioprine. TheZ-score of L2–L4 in these patients was 20.30 6 0.20,which was not significantly different from patientstreated with steroids and/or azathioprine 0.08 6

0.21; P 5 0.20; n 5 38). However, patients treatedwith a high lifetime dose of steroids (. 5 g pred-nisolone) had significantly lower Z-scores thansteroid-treated patients who had received a low doseof prednisolone during their lifetime (, 5 g pred-nisolone; P 5 0.011, Table 1). As in this patient

Fig. 1 Correlation between the Z-score of the hip and the Z-scoreof L2–L4 in 59 patients with Crohn’s disease.

BONE MINERAL DENSITY IN CROHN’S DISEASE 125

© 1998 Blackwell Science Ltd Journal of Internal Medicine 243: 123–126

material there is a significant correlation (Fig. 1)between Z-score in the hip and Z-score in the lumbarregion, L2–L4, the conclusion that a high lifetime doseof steroids reduces bone mineral density is true notonly for the lumbar spine but also for the hip.

Because the therapeutical tradition is to combinesteroids with azathioprine, and then taper offsteroids, there was a good correlation between life-time dose of steroids and lifetime dose of azathio-prine (r 5 0.53; P , 0.001; n 5 38, Fig. 2), except inthe case of one patient (treated elsewhere and notincluded in Fig. 2) with a high lifetime dose ofsteroids (50 g prednisolone) not combined with aza-thioprine. In patients treated with a therapeutic doseof azathioprine (defined as treatment of 12 g in a life-time dose) together with steroids, the Z-score in the

lumbar region was 20.42 2 0.27 (n 5 24) vs.20.20 2 0.27 in patients treated with steroids aloneor steroids in combination with a nontherapeuticdose of azathioprine (12 g in a lifetime dose) (NSn 5 14). The minimal requirement for a therapeuticdose of azathioprine, 12 g, was chosen, as it takes3–4 months of azathioprine treatment before aneffect can be seen [7]. To evaluate the possible influ-ence of azathioprine on bone mineral density, wesubdivided the patients into four different groups,depending upon lifetime doses of steroids and aza-thioprine (Table 2). The patients were subdivided intohigh and low lifetime dose of steroids (. 5and , 5 g prednisolone, respectively) and treatmentwith at least 50 g of azathioprine as lifetime dose(Aza 1) or no Aza (defined as no treatment or treat-ment less than 50 g). Thus a higher dose than thetherapeutic dose (. 12 g) was chosen, in order tofurther test the effect of azathioprine on bone miner-al density. As can be seen in Table 2, azathioprine didnot negatively influence the steroid effect on bonemineral density, i.e. the steroid dose itself seems to bethe main parameter affecting bone mineral density inpatients with Crohn’s disease.

Discussion

One of the main drawbacks of steroid treatment isthe metabolic adverse effect on bone metabolism,causing severe osteoporosis [4]. In order to avoid thisproblem, other therapeutic modalities, such as

Fig. 2 Correlation between lifetime dose of prednisolone andlifetime dose of azathioprine in 38 patients with Crohn’s disease.

Table 1 Z-score (L2–L4) as a function of lifetime dose of steroids in59 subjects with Crohn’s disease, divided into groups according todose of steroid

Z-score (L2–L4)

n mean SD SEM

Low dose prednisolone (,5 g) 23 20.08 1.16 0.24High dose prednisolone (.5 g) 15 20.87a 1.11 0.29Control group (no steroid/

no azathioprine treatment) 21 20.30 1.22 0.20

aIndicates a probability level of random difference versus the low-dose prednisolone group of P 5 0.011.

Table 2 Z-score (L2–L4) in groups of patients with high or lowlifetime doses of azathioprine and prednisolone

Z-score (L2–L4)

Groups mean SD SEM

1 Aza 1

Low steroids 20.33 1.09 0.41n 5 7

2 Aza 1

High steroids 20.89 1.27 0.38n 5 11

3 Aza-Low steroids 20.04 1.20 0.30n 5 16

4 Aza-High steroids 20.82 0.58 0.29n 5 4

aAza1, .50 g azathioprine; Aza-, ,50 g azathioprine; Highsteroids, .5 g prednisolone; Low steroids, ,5 g prednisolone.

C.-H. FLORÉN et al.126

© 1998 Blackwell Science Ltd Journal of Internal Medicine 243: 123–126

immunosuppressants, have been used to treatpatients with Crohn’s disease. A recent meta-analysishas shown, by using existing placebo- controlledstudies, that azathioprine maintains remission inpatients with Crohn’s disease and that it also, togeth-er with steroids, induces remission significantly bet-ter than placebo [7]. A drawback to azathioprinetreatment is, however, its adverse effects – for exam-ple, neutropenia – and that it takes 3–4 months oftreatment before an effect can be seen [7, 11]. In thepresent cross-sectional, retrospective study ofpatients with Crohn’s disease, mainly localized to thecolon, we evaluated whether addition of azathio-prine affects the adverse effects of steroid on bonemineral density. The results, with their inherentselection bias, show that the dominating effect onbone mineral density is the use of steroids, and thatazathioprine by itself has only little effect on bonemineral density. Also earlier experimental data in therat has shown that azathioprine by itself caused noloss of bone volume and that it does not altercyclosporin A-induced osteopenia [12]. Hence, it isunlikely that azathioprine affects the steroid-inducedadverse effect on bone remodelling. However, sinceazathioprine is used to diminish inflammatory activi-ty, the drug is steroid-sparing, meaning that a patienttreated with azathioprine can reduce the dose ofsteroids. Our finding that azathioprine seems, byitself, to only marginally affect bone mineral density,therefore means that, in the clinical setting, it doesconserve bone mineral mass by its steroid-sparingeffect.

References1 Summers RW, Switz DM, Sessions JT, Becletel JM, Bert WR,

Kern F, Singleton JW. National cooperative Crohn’s disease

study; results of drug treatment. Gastroenterology 1979; 77:847–69.

2 Ewe K, Press AG, Sirge CC, Stufter M, Ueberschaler B, HommelG, Meyer zum Büschenfelde K-H. Azathioprine combined withprednisolone or monotherapy with prednisolone in activeCrohn’s disease. Gastroenterology 1993; 105: 362–72.

3 Podolsky DK. Inflammatory bowel disease. N Engl J Med 1991;325: 928–37; 1008–16.

4 Hahn TJ, Boisseau VC, Avioli LV. Effect of chronic corticos-teroid administration on diaphyseal and metaphyseal bonemass. J Clin Endocrin Metab 1974; 39: 274–87.

5 Gosh S, Cowen S, Hanna WJ, Ferguson A. Low bone mineraldensity in Crohn’s disease but not in ulcerative colitis, at diag-nosis. Gastroenterology 1994; 107: 1031–9.

6 Silvennoinen JA, Karttunen TJ, Niemelä SE, Manelius JJ,Lehtola JK. A controlled study of bone mineral density inpatients with inflammatory bowel disease. Gut 1995; 37:71–6.

7 Pearson DC, May GR, Fick GH, Sutherland LR. Azathioprineand 6-merkapto-purine in Crohn’s disease. A meta-analysis.Ann Intern Med 1995; 122: 132–42.

8 Nyman M, Hansson I, Eriksson S. Long-term immunosuppres-sive treatment in Crohn’s disease. Scand J Gastroent 1987; 20:1197–203.

9 Karlsson M, Gärdsell P, Johnell O, Nilsson B, Åkesson K,Obrant K. Bone mineral normative data in Malmö, Sweden.Comparison with reference data and hip fracture incidence inother ethnic groups. Acta Orthop Scand 1993; 64 (2):168–72.

10 Lunar News. Madison, USA: Lunar Corporation, 1991 (a, Jan;b, March).

11 Connell WR, Kamm MA, Ritchie JK, Lennard-Jones JE. Bonemarrow toxicity by azathioprine in inflammatory bowel dis-ease: 27 years of experience. Gut 1993; 34:1081–1085.

12 Bryer HP, Isserow JA, Armstrong EC, Mann GN, Rucinski B.Buchinsky FJ et al. Azathioprine alone is bone sparing anddoes not alter cyclosporine A-induced osteopenia. J Bone MinerRes 1995; 10: 132–8.

Received 23 January 1997; accepted 11 July 1997.

Correspondence: C.-H. Florén MD PhD, Division of Gastroenterologyand Hepatology, Department of Internal Medicine, MalmöUniversity Hospital, Lund University, S-205 02 Malmö, Sweden(fax: +40 33 62 08).