Systematic review of the prevalence and development of osteoporosis or low bone mineral density and its risk factors in patients with inflammatory bowel disease

Table 1 Study characteristics

Ref.	Publication year	Country of study	Type of study	DXA scan1	No. of patients	Average age of patients	Cohort	Control group	Follow- up	Comments
Andreassen et al[1]	1998	Denmark	Cross-sectional study inviting all IBD patients from a well-defined area	Yes	115	37 (16-75); median (range)	CD only	No	No
Andreassen et al[2]	1999	Denmark	Cross-sectional case-control study inviting all IBD patients from a well-defined area	Yes	113	37 (16-75); median (range)	CD only	Yes; n = 113	No	Same cohort used as in Andreassen et al[14] (1998)
Bernstein et al[3]	20032	United States	Cross-sectional data extracted from population-based Manitoba IBD research registry	Yes	70; UC: n = 12; CD: n = 58	33.0 (7.4); mean (SD)	UC and CD	No	No	Includes only premenopausal women
Bernstein et al[4]	20033	United States	Cross-sectional data extracted from population-based Manitoba IBD research registry	Yes	66 (DXA results: n = 70); UC: n = 11; CD: n = 55	33.3 (18-44); mean (range)	UC and CD	No	No	Includes only premenopausal women. Same cohort used as in Bernstein (2002)
Haugeberg et al[5]	2001	Norway	Cross-sectional data from a population-based study. Case control study	Yes	55	38.5 (12.7); mean (SD)	CD only	Yes; n = 52	No
Jahnsen et al[6]	1997	Norway	Cross-sectional case control study	Yes	60	36 (21-75); median (range)	CD only	Yes; n = 60	No	Includes a cohort of UC patients that is not population-based which was therefore not included
Jahnsen et al[7]	2004	Norway	Follow-up study	Yes	60	36 (21-75); median (range)	CD only	No	Yes	Includes a cohort of UC patients that is not population-based which was therefore not included
									2 yr	Same cohort used as in Jahnsen (1997)
Leslie et al[8]	2008	Canada	Follow-up study with cohort extracted from population-based Manitoba IBD research registry	Yes	101; UC: n = 45; CD: n = 56	46.9 (15.5); mean (SD)	UC and CD	No	Yes; 2.3 ± 0.3 yr
Leslie et al[9]	2009	Canada	Follow-up study with cohort extracted from population-based Manitoba IBD research registry	Yes	101 UC: n = 45 CD: n = 56	47 (15); mean (SD)	UC and CD	No	Yes; 2.3 ± 0.3 yr	Same cohort used as in Leslie et al[19] (2008)
Schoon et al[10]	2000	The Netherlands	Cross-sectional cohort	Yes	119	42 (14); mean (SD)	CD only	No	No
Targownik et al[11]	2012	Canada	Follow-up study with data extracted from population-based Manitoba IBD research registry	Yes	86; UC: n = 32; CD: n = 50; Unclass: n = 4	46.7 (14.9); mean (SD); 46 (35-57) median (IQR)	UC and CD	No	Yes; 4.3 ± 0.3 yr	Same cohort used as in Leslie et al[19] (2008)
Tsai et al[12]	2015	Taiwan	Follow-up case control study with data extracted from population-based registry	No	3141; UC: n = 1489; CD: n = 1652	46.7 (35.6-61.0); median (IQR)	UC and CD	Yes; n = 12564	Yes; 6.49 ± 3.09 yr	Diagnosis of osteoporosis based on ICD-10 codes

¹BMD scores derived from DXA scan.

²published in May.

³published in November. CD: Crohn’s disease; UC: Ulcerative colitis; IQR: interquartile range; Unclass.: unclassified; IBD: inflammatory bowel diseases.

Table 2 Quality assessment according to the Newcastle–Ottawa Scale

	Schoon et al[13] (2000)	Jahnsen et al[12] (1997)	Jahnsen et al[16] (2004)	Tsai et al[18] (2015)	Targownik et al[3] (2012)	Leslie et al[20] (2009)	Leslie et al[19] (2008)	Andreassen et al[14] (1998)	Andreassen et al[15] (1999)	Bernstein et al[21] (2003, May)	Bernstein et al[22] (2003, November)	Haugeberg et al[17] (2001)
Selection	**	***	**	****	**	**	**	**	***	**	**	***
Comparability		**		**					**			**
Outcome	*	*	**	**	***	**	**	*	*	*	*	*
Total number of stars allocated	3	6	4	7	5	4	4	3	6	3	3	6

Table 3 Overview of most relevant risk factors for low bone mineral density or osteoporosis

Risk factors for reduced BMD	CD	CD + UC	Comments
General risk factors
Gender[5,6,9,10]	+/-	+	Female gender was found to be significantly correlated by Leslie et al[20] (2009) investigating both CD and UC patients. In CD studies, Haugeberg et al[5] found female gender to be a predictive factor for osteoporosis. Jahnsen et al[6] found men to have lower Z-scores than women, whereas Schoon et al[10] found no significant association.
Age[2,3,5,9]	+	+/-	Age was significantly associated in the CD studies. However, Haugeberg et al[5] found patients with reduced BMD to be significantly younger than those without reduced BMD.
Weight[2,3,5,9]	+, -1	+	Low weight was found to be a risk factor for low BMD in both CD + UC cohorts. In CD cohorts, Andreassen et al[15] (1999) found a significant positive correlation only in males. Haugeberg et al[17] found a positive correlation between weight and BMD for both genders.
BMI[2,5,6,9]	+/-	+	Leslie et al[20] (2009), the only study investigating BMI in CD + UC, found a positive correlation between BMI and BMD. Haugeberg et al[17] found a significant association for CD patients in a bivariate analysis, but not in a multiple linear regression analysis.
Steroid treatment[2,3,5,6,9]	+/-2	-	Multiple risk factors related to steroid usage were investigated. No correlation was found in CD + UC. However, most CD studies did find a correlation.
Height[3,5,9]	+/-	+/-
Smoking[3,5,6]	-	-
Vitamin D supplement[3-5]	-	-
Calcium supplement[3-5]	-	-
Serum 25(OH)D[1,5,8]	+/-	+/-
Serum calcium[1,5,8]	-	-
Serum parathyroid hormone[1,5,8]	+/-	+
Disease-specific risk factors
UC diagnosis[3,9]	Not relevant	-
CD diagnosis[3,6,9]	Not relevant	+/-
Disease location[1,3,5]	-	-
Disease duration[2,3,5,6]	+3,-	-
Surgery[2,3,5,6]	+/-	-

¹Only in females.

²Only in males.

³Postmenopausal females. A plus sign means that a significant association was found and a minus sign means that no association was found. If studies found different results, both signs are present. CD: Crohn’s disease; UC: Ulcerative colitis; BMI: Body mass index; BMD: Bone mineral density; +: Positive association; -: No association; +/-: Significant association and no association were found, depending on the study and/or statistical analysis carried out.

Table 4 Overview of the most relevant risk factors for change in bone mineral density over time

Risk factors for change in BMD	CD	CD + UC	Comments
General risk factors
Gender[8,9,11,12]	No data	+/-	No difference was found between genders in one study cohort[8,9,12], whilst another cohort[11] found a greater incidence of osteoporosis in women than in men.
Age[8,9,11,12]	No data	+/-
Weight[9,11]	No data	+
BMI[7,9,11]	+	+
Steroid treatment[7-9,11]	-	+/-
Smoking[7]	-	No data
Serum 25-OH D[7,8,11]	+	+/-
Disease-specific risk factors
Diagnosis[9,11,12]	Not relevant	+1, -	One[13] out of three studies found CD to be associated with an increased risk of osteoporosis. The others found no associations.
Disease location[7]	-	No data
Disease activity[11]	No data	-

¹Only in Crohn’s disease. CD: Crohn’s disease; UC: Ulcerative colitis; +: Positive association; -: No association; +/-: Significant association and no association were found, depending on the study and/or statistical analysis carried out. A plus sign means that a significant association was found and a minus sign means that no association was found. If studies found different results, both signs are present. Follow-up time for the respective studies was as follows: Jahnsen et al[16]: 2 yr; Leslie et al[19]: 2.3 ± 0.3 yr; Leslie et al[20]: 2.3 ± 0.3 yr; Targownik et al[3]: 4.3 ± 0.3 yr; Tsai et al[18]: 6.49 ± 3.09 yr.