High Prevalence of Osteoporosis in Patients With Chronic Pancreatitis: A Systematic Review and Meta-analysis

Transcription

1 Clinical Gastroenterology and Hepatology 2014;12: High Prevalence of Osteoporosis in Patients With Chronic Pancreatitis: A Systematic Review and Meta-analysis Sinead N. Duggan,* Niamh D. Smyth, Anne Murphy, David MacNaughton, jj Stephen J. D. O Keefe, and Kevin C. Conlon* *Centre for Pancreatico-Biliary Diseases, Professorial Surgical Unit, Trinity College Dublin, Tallaght Hospital, Dublin, Ireland; Department of Nutrition & Dietetics, and Tallaght Hospital Library, Tallaght Hospital, Dublin, Ireland; jj Hamilton Library, Trinity College Dublin, Dublin, Ireland; and Small Intestinal Rehabilitation & Transplant Center, Clinical Nutrition Service, Pittsburgh University Hospital, Pittsburgh, Pennsylvania BACKGROUND & AIMS: Patients with chronic pancreatitis may be at high risk for osteoporosis and osteopenia. We performed a systematic review and meta-analysis to determine the prevalence of osteoporosis and osteopenia in patients with chronic pancreatitis. METHODS: Articles were identified from MEDLINE, EMBASE, and SCOPUS databases (through October 2012) and a manual search of the literature. The primary outcome measure was bone density, measured by dual-energy X-ray absorptiometry (T-score or Z-score). When available, data on the prevalence of osteopenia, bone mineral density, and bone mineral content also were recorded. RESULTS: CONCLUSIONS: Ten studies including 513 patients were eligible for inclusion. Based on a random-effects model, the pooled prevalence rate for osteoporosis among patients with chronic pancreatitis was 23.4% (95% confidence interval, ). The pooled prevalence for osteopenia was 39.8% (95% confidence interval, ). The pooled prevalence rate for either osteoporosis or osteopenia was 65% (95% confidence interval, ). Based on meta-analysis, almost 1 of 4 patients with chronic pancreatitis have osteoporosis, and almost two-thirds of patients have either osteoporosis or osteopenia. Osteoporosis and osteopenia are underappreciated sources of morbidity in patients with chronic pancreatitis. Bone health management guidelines are urgently required in patients with chronic pancreatitis. Keywords: Bone Disease, Metabolic; Demineralization; Risk Factor. Osteoporosis is characterized by structural deterioration of bone tissue and low bone mass, leading to bone fragility and increased risk of fracture. Osteoporosis is a major public health problem because of its potentially severe consequences for both patients and the health care system. 1 Chronic pancreatitis is a progressive inflammatory condition resulting in exocrine and endocrine dysfunction. Exocrine dysfunction leads to reduced production of pancreatic digestive enzymes, and the resultant maldigestion and malabsorption of ingested nutrients leads to malnutrition and nutrient deficiency. Osteoporosis has been described in chronic pancreatitis, with 21% of chronic pancreatitis patients having osteoporosis in the earliest study published in Since then, a number of studies have shown a varied prevalence of osteoporosis from 5% to 39%. 3 5 From these studies and others, it is reasonable to assume that patients with chronic pancreatitis may be at risk for osteoporosis and osteopenia (termed osteopathy). However, to date, no international consensus guidelines have recommended the assessment or monitoring of bone health in chronic pancreatitis. To address this uncertainty, we conducted a systematic review of the literature and a meta-analysis of the data to estimate the prevalence of osteoporosis and osteopenia in chronic pancreatitis. Methods Criteria for Consideration of Studies Observational studies that reported data on the prevalence of osteoporosis in patients with chronic Abbreviations used in this paper: BMC, bone mineral content; BMD, bone mineral density; BMI, body mass index; DXA, dual-energy X-ray absorptiometry by the AGA Institute /$

2 220 Duggan et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2 pancreatitis were included. The search was not limited by sex, geographic location, or publication status. Studies that were limited solely to pediatric patients (age, <18 y) were excluded. The primary outcome measure of interest was the prevalence of osteoporosis or osteopenia based on bone density measured by dual-energy X-ray absorptiometry (DXA) (T-scores or Z-scores), and, where available, bone mineral density (BMD) (g/cm 2 ) and bone mineral content (BMC) (g/cm). Literature Search The following bibliographic databases were searched for studies on chronic pancreatitis and osteoporosis: Ovid MEDLINE (1946 to October 31, 2012), Elsevier EMBASE (1980 to October 31, 2012), and SciVerse SCOPUS (1966 to October 31, 2012). No date or language restrictions were used, but searches were limited to human beings. Searches were conducted during October of 2012, and search updates were set-up to send updates to the authors automatically from MEDLINE, EMBASE, and SCOPUS. The last update considered for inclusion was sent at the end of October The search strategy was developed for Ovid Medline and translated for use on EMBASE and SCOPUS. We searched for articles with combinations of subject headings and key words relating to Pancreatitis, Chronic or exocrine Pancreatic Insufficiency and Bone Density or Absorptiometry, Photon or Bone Diseases, Metabolic. Further searches were performed by scanning the reference lists of the primary and review articles to identify studies not found by the electronic search. We also conducted searches of conference proceedings. Last, we searched the Cochrane central register of controlled trials (The Cochrane Library) but did not find any randomized studies that provided data for our analysis. Both independent reviewers (S.N.D. and N.D.S.) were supported in developing search terms by medical librarians (D.M. and A.M.) in separate institutions. Study Selection Criteria All citations identified by literature search were screened independently by 2 reviewers (S.N.D. and N.D.S.) using article titles and abstracts. The full texts of potentially relevant articles were sought and the selection criteria were applied. Conference proceedings were considered for inclusion if they contained adequate relevant information for review. Reviewers were not blinded to author names or institutions. Studies were selected according to predefined criteria. For both conference proceedings and full-text articles in which there were missing data or a requirement for clarity, the study authors were contacted by . In the case of non English language articles, translations were performed by Corporate Translation Services in Dublin, Ireland (ISO 9001:2008 certified). Assessment of the Quality of Individual Studies The quality of the studies was evaluated independently by the 2 reviewers using the Newcastle Ottawa Scale, which uses a star rating system to judge the quality of observational studies. 6 This scale awards a maximum of 9 stars to each study: up to 4 stars for selection of participants, 2 stars for comparability of participants on the basis of the design or analysis, and 3 stars for ascertainment of exposure. We assigned scores of 0 to 3, 4 to 6, and 7 to 9 for low-, moderate-, and high-quality studies, respectively. Data Extraction and Statistical Analyses Data extraction was performed independently (by S.N.D. and N.D.S.) using a predefined data extraction form. Osteoporosis and osteopenia rates (osteopathy rates) were recorded. BMD and BMC were recorded where available for patients and controls. Additional data sought included the following: study design, sex, age, etiology, DXA scanner used, sites assessed, exocrine function, disease severity, and body mass index (BMI). Studies that reported the rate of osteoporosis were deemed eligible for meta-analysis. If studies also reported the rate of osteopenia, the overall prevalence of osteopathy also was calculated by meta-analysis. Data were meta-analyzed using Comprehensive Meta-Analysis software (version ; Englewood, NJ) and Forest plots were constructed using Excel (Microsoft Office 2007; Microsoft, Redmond, WA) as described by Neyeloff et al. 7 Data were presented as prevalence rates (with percentages) and corresponding 95% confidence intervals. A pooled estimation was computed using a random-effects model to provide a more conservative estimate of the prevalence, allowing for variations between studies. Statistical heterogeneity between studies was calculated as I 2 (values ranged between 0% and 100%, with values closer to 0% indicating less heterogeneity). The I 2 quantity describes the percentage of total variation across studies that is caused by heterogeneity rather than chance. Details on reported statistical associations with age, sex, smoking, exocrine function, BMI, disease duration, duration of symptoms, smoking, vitamin D level, diabetes, and bone biochemistry within individual studies were recorded and described qualitatively. The Metaanalysis of Observational Studies in Epidemiology guidelines 8 were adhered to where appropriate. P values less than.05 were considered statistically significant. Results Search Results The search and selection process is summarized in Figure 1. Manual searches of the reference sections of relevant articles and reviews did not provide any further

3 February 2014 Osteoporosis in Chronic Pancreatitis 221 BMD or BMC were reported by 3 studies only. 2,4,9 In the majority of studies, osteoporosis and osteopenia were defined according to T-scores (a T-score between 1.0 and 2.5 standard deviations was defined as osteopenia; a T-score below 2.5 standard deviations was defined as osteoporosis). T-scores compare bone density values with those of young adults (peak bone mass). Two studies 5,12 defined osteoporosis and osteopenia according to Z- scores (Z-score < 1 was defined as osteopenia, Z-score < 2 was defined as osteoporosis). Meta-analysis of Osteoporosis Prevalence Figure 1. Flow diagram for the assessment of studies identified in the systematic review. articles. Hand-searching through conference proceedings yielded 2 additional studies. Summary of Studies Included Eleven studies met the inclusion criteria and were included in the review (Table 1). 2 5,9 15 Of these, 8 were full articles and 3 were abstracts of conference proceedings. Of the 11 studies, 8 were conducted in Europe, 2 in India, and 1 in South America. More than half (6 of 11) were published during 2011 and One study 10 was in Russian and was available in abstract form only. After corresponding with the author, the full-text version was obtained and translated into English. One study reported on both adults and children, however, in accordance with the predefined criteria, only data for adults were included. Three studies were limited to men. 2,9,11 The mean ages of patients ranged from 46.6 to 60 years, with the notable exception of the studies from India, both of which included considerably younger patients (mean age, 31 and 36 y). 5,9 The underlying etiology for chronic pancreatitis was mixed in all studies with the exception of one, 5 which included only patients with tropical chronic pancreatitis (study from India). There was also variation in the types of DXA scanners used (Table 2). All studies reported bone density at the lumbar spine, but varied in reporting data regarding hip bone density (such as total hip, left hip, right hip, and Ward s triangle). Details of In total, 555 patients with chronic pancreatitis underwent bone density assessment. One study 11 did not report absolute osteopathy prevalence rates and the data were presented as the percentage of normal. This study was excluded from the meta-analysis but used for qualitative assessments of factors associated with osteoporosis in chronic pancreatitis. Figure 2 illustrates the prevalence of osteoporosis in 10 eligible studies that included 513 patients with chronic pancreatitis. Based on the randomeffects model, the pooled prevalence rate of osteoporosis was 24.3% (95% confidence interval, ). Assessment of heterogeneity gave an I 2 value of 73.6%, indicating considerable heterogeneity. The pooled prevalence rate for osteopenia among patients was 39.8% (95% confidence interval, 29.1% 51.6%) (Forest plot not shown). Figure 3 summarizes the prevalence of overall osteopathy (osteoporosis and osteopenia). A random-effects model yielded a pooled osteopathy prevalence of 65% (95% confidence interval, ). We estimated the pooled prevalence of osteoporosis excluding the studies from India, which included mostly tropical chronic pancreatitis patients associated with severe malnutrition (Forest plots not shown). When the 2 studies from India 5,9 were removed, the pooled prevalence for osteoporosis was slightly lower at 20.6% (95% confidence interval, ); however, the I 2 value for statistical heterogeneity was still high at 76.6%. Regarding controls, of 2 studies with usable data, the osteoporosis rate for controls ranged from 8.6% to 10.2%, whereas the osteopenia rate for controls from one study was 33.9%. Study Quality On the basis of the Newcastle Ottawa Scale, 4 studies, 4,5,9,11 all of which were controlled, received 7 stars or more (of a possible 9 stars) and were classified as high-quality studies. The other studies, all of which were uncontrolled, received 3 stars or less and were classified as lower-quality studies. The conference proceedings were not amenable to quality assessment. There was 100% agreement between the 2 independent reviewers regarding the quality scoring.

4 222 Duggan et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2 Table 1. Description of 11 Studies Included in the Systematic Review Study Publication type/study design Subjects, n Sex, M/F Age, y (SD) Patients Controls Patients Controls Patients Controls Etiology Metaanalysis JA/cross-sectional 73-56/ (13.2) - 89% idiopathic (6.4) No Duggan et al, JA/case-control /13 43/ (12.5) (11) 38.7% alcohol Yes 2012, Ireland 4 Sudeep et al, JA/case-control /0 35/ (9) 38.6 (5.2) 65% TCP Yes 2011, India 9 35% idiopathic Joshi et al, 2011, India 5 JA/case-control /34 a 50/50 31 (10) 32.6 (9.6) 100% TCP Yes Drozdov et al, JA/cross-sectional /53-51 (10.2) - 52% alcohol Yes 2010, Russia 10 48% biliary Dujshikova et al, Yes 2008, Czech Republic 3 11% alcohol Mann et al, JA/case control /0 20/ (13.5) 48.9 Not specified 2003, Germany 11 Haaber et al, JA/cross-sectional 58-32/26-55 (11) b % alcohol Yes 2000, Denmark 12 Moran et al, JA/cross-sectional 14-14/ % alcohol Yes 1997, Argentina 2 29% idiopathic Sikkens et al, 2012, The Netherlands 13 CPr/cross-sectional 40-23/17-52 (11) - 50% alcohol 30% idiopathic 20% other Yes Skipworth et al, 2012, United Kingdom 14 CPr/retrospective review 64-44/ % alcohol 25% idiopathic 11% biliary 12% other Gubergrits et al, Cpr/cross-sectional 24 - Not specified Not specified Not specified Yes 2011, Ukraine 15 Yes CPr, conference proceedings; JA, journal article; TCP, tropical chronic pancreatitis. a Represents total number recruited, sex breakdown for those undergoing DXA unknown. b Values were available only for the pancreatic enzyme insufficiency group. Factors Affecting Bone Mineral Density in Chronic Pancreatitis: Qualitative Results Subgroup analyses were reported in all studies, and are summarized in Table 3 according to the presence of statistical association or no association. Age at Assessment Counterintuitively, the 4 articles with the highest prevalence of osteoporosis (>29%) had generally younger patients (mean age, 31, 36, 48, 51 y) than those with a lower osteoporosis prevalence (mean age, 52, 53, 55, 56, 60 y), with one exception in the lower-incidence group 3 (mean age, 46.6 y). The 2 studies from India 5,9 in particular included patients in their early to mid-30s on average. There was, however, no observed age pattern when looking at the prevalence of overall osteopathy (osteoporosis and osteopenia). Sex Three studies examined men only, whereas the remaining studies included both men and women. Only 3 studies reported testing for an effect of sex, of which 2 studies found no difference in regard to BMD. 4,5 Only one study reported lower BMD in women compared with men. 10 The remaining studies did not report any effect of sex. Body Mass Index Patients from the 2 studies from India 5,9 had a notably lower BMI than patients from other studies (predominantly tropical chronic pancreatitis etiology), and had a high prevalence of osteoporosis. Both studies noted positive statistical associations between BMI and BMD. Subjects in all other studies had a mean BMI in the normal range, with the exception of the Irish study, 4 in which subjects had a mean BMI that was marginally in the overweight range. The Irish 4 and UK 14 studies reported a statistical association with BMI (decreased BMI was associated with a decreased BMD). Etiology All but one study 5 had a mixed etiology (only tropical chronic pancreatitis patients). Because of the mixed nature of the study groups, there were no apparent associations with BMD. Exocrine Function Most studies reported pancreatic enzyme insufficiency of varying degrees (one study 3 did not report on exocrine dysfunction). To assess pancreatic exocrine function 6 studies measured fecal elastase-1, 4,5,10,11,13,15

5 Table 2. Prevalence of Osteoporosis and Osteopenia Among Chronic Pancreatitis Patients and Controls Study DXA details Sites Duggan et al, Sudeep et al, Lunar Prodigy Advance (GE Medical Systems, Belgium) QDR 4500 (Hologic, Bedford, MA) Joshi et al, ,a QDR 4500 (Hologic) Left hip Distal radius LV Drozdov et al, Lunar DPX Hip (GE Medical Systems) LV Dujshikova et al, ,c Lunar (GE Medical Systems) Mann et al, ,d Lunar Radiation (Madison, WI) Haaber et al, ,a Norland XR 36 (Norland Instruments, Fort Atkinson, WI) CP patients Osteoporosis, n(%) Osteopenia, n (%) BMD (SD) BMC (SD) Osteoporosis, n(%) Controls Osteopenia, n(%) BMD (SD) BMC (SD) Total hip Right FN LV L1:L4 18 (34) 21 (39.6) (10.2) 20 (33.9) LV 9 (29) (0.135) 51.9 (10.3) 3 (8.6) (0.098) (7.62) Whole proximal femur LV L1:L4 Ward s triangle in FN LV L2:L4 Total body LV L2:L4 22 (39) 14 (25) - - b b (39) 18 (18) (5) 19 (26) DEXA Ward 92.2 (5.2) normal DEXA Ward 97.1 (3.1) normal (22.4) 36 (62) (21.4) 10 (71.4) 1.07 (0.05) Moran et al, Lunar DPX (Madison, WI) FN LV L2:L4 e Sikkens et al, Lunar prodigy Left femur 4 (10) 18 (45) (GE Medical Systems) LV L2:L4 Skipworth et al, Not specified FN 12 (19) 27 (42) LV L1:L4 Gubergrits et al, Explorer QDR (Hologic, Bedford, MA) Not specified 5 (20.8) 12 (50) CP, chronic pancreatitis; FN, femoral neck; LV, lumbar vertebrae. a Data are presented as Z-scores. b Control values are not discernible from text. c Further 6 patients (8%) classified with osteomalacia. d BMD values are presented as a percentage of normal. e Data are presented for lumbar spine. February 2014 Osteoporosis in Chronic Pancreatitis 223

6 224 Duggan et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2 Figure 2. Forest plot of the pooled prevalence of osteoporosis in chronic pancreatitis using a random-effects model. 3 studies measured stool fecal fat, 2,9,12 and the remaining study did not provide details on the measurement method used. 14 Two studies 2,14 reported that all subjects had measurable exocrine impairment or clinical steatorrhea. Five studies 9 11,13,15 reported statistical associations between lower fecal elastase-1 levels (or high fecal fat) and reduced BMD, although one study reported the association only with BMC. 9 Three studies found no association. However, of those 3 studies, 2 studies 4,12 reported pancreatic enzyme insufficiency in less than half the group, leading to smaller subgroup numbers, while the third study 2 had a very low sample size (n ¼ 14) and therefore was likely to be underpowered to detect subgroup differences if they existed. Disease Severity Only 4 studies 3,4,10,11 provided details regarding chronic pancreatitis severity, and, of these, only 2 used the recognized Cambridge criteria to classify severity. 4,11 Although one study found an association with disease severity 11 (lower BMD in higher-severity grades), the other study did not. 4 Vitamin D Seven studies 2 5,9,11,12 considered associations between low serum vitamin D level and low BMD, 4 of which found no association. Two studies 4,5 found that controls also had low serum vitamin D level, with no statistically significant differences between controls and patients. Conversely, 2 groups 8,11 reported differences between control and patient serum vitamin D levels. Smoking Three studies, 4,13,14 all of which were published in 2012, considered the relationship between smoking and osteopathy rates. Although 2 of these found no association, 1 study 4 reported a statistically significant relationship between heavy smoking and reduced BMD. Figure 3. Forest plot of the pooled prevalence of overall osteopathy (osteoporosis and osteopenia) in chronic pancreatitis using a random-effects model.

7 Table 3. Details of Disease Severity and Statistically Significant Associations Between BMD and Specific Factors Study Setting Exocrine function Disease severity BMI mean (SD) (CP) Duggan et al, OPD, acute hospital 30.4% PEI Cambridge classification 27.4% severe Statistical association 25.6 (5) Increasing age, heavy smoking history, lower BMI Subgroup analysis: association with low BMD Sudeep et al, OPD, acute hospital 69% steatorrhea Not specified (2.86) Duration of symptoms, Vitamin D lower BMI Joshi et al, OPD, acute hospital 46% steatorrhea Not specified 19 (3.1) Lower BMI Vitamin D Drozdov et al, Not specified 59% PEI 44% complicated CP 7%, underweight; 61%, normal weight; 26%, overweight; 6%, obese a Lower fecal elastase-1, duration of the disease No association Disease severity, fecal elastase-1, vitamin D, sex, education level Dujshikova et al, Acute hospital Not specified 27.4% severe Not specified Lower vitamin D - Mann et al, Acute hospital Mean fecal elastase, Cambridge Not specified Disease severity, lower PTH, OC, CICP, BAP (in-patients) (SD, 75.7) classification 20% severe fecal elastase-1, vitamin D Haaber et al, OPD, acute hospital 44.8% PEI, steatorrhea Not specified 22 (3) b - Disease duration, vitamin D, PTH, calcium, PEI, steatorrhea Moran et al, Acute hospital All subjects steatorrhea Not specified 22.6 (3.2) - Alcohol, fecal fat, secretin test, BMI, age, vitamin D, PTH, serum calcium Sikkens et al, OPD, acute hospital 70% PEI Not specified 24 (5) Lower fecal elastase-1 Smoking, alcohol Skipworth et al, Tertiary center All subjects PEI Not specified 23.3 Lower BMI, diabetes Etiology, smoking, bone biochemistry, disease duration Gubergrits et al, Not specified Not specified Not specified Not specified Lower fecal elastase-1 - BAP, bone-specific alkaline phosphatase; CICP, carboxy-terminal propeptide of type I procollagen; CP, chronic pancreatitis; OC, osteocalcin; OP, out-patients department; PEI, pancreatic exocrine insufficiency; PTH, parathyroid hormone; SD, standard deviation. a Classified as BMI greater than 28 kg/m 2. b Values were available only for the PEI group. - February 2014 Osteoporosis in Chronic Pancreatitis 225

8 226 Duggan et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2 Duration of Disease and Symptoms One Indian study 9 found a statistical association between duration of symptoms and low BMD, and another study 10 reported a link between the duration of disease and low BMD. Two other studies that considered the relationship between disease duration and BMD found no such relationship. 12,14 Discussion In this systematic review on the prevalence of osteoporosis and osteopenia in chronic pancreatitis, 11 studies were reviewed. 2 5,9 15 We showed a high prevalence of osteoporosis in chronic pancreatitis, and also a high prevalence of overall osteopathy. By using the available data, we calculated that the prevalence of osteoporosis in chronic pancreatitis may be as high as 1 patient in 4. Similarly, by calculating a pooled prevalence, as many as two thirds of chronic pancreatitis patients may have osteopathy (either osteopenia or osteoporosis). Because of the heterogeneous data and the small sample size in individual studies, making definitive or broad statements on subgroup associations is not warranted. Nevertheless, certain patterns were evident from the studies included. Pancreatic enzyme insufficiency was associated with a lower bone density in 5 of 9 studies in which it was measured. It is reasonable to assume that a reduction in enzyme production affecting nutrient absorption would affect BMD. Pancreatic enzyme replacement therapy is the mainstay of treatment for pancreatic enzyme insufficiency, 16 and this finding supports the necessity of timely and adequate prescription of pancreatic enzyme replacement therapy, although there are no published data on the effect of pancreatic enzyme replacement therapy on bone density. Low serum vitamin D levels are probably contributory to low BMD in chronic pancreatitis owing to malabsorption and diminished sunshine exposure in chronic illness. However, the available data failed to show direct associations between serum vitamin D level and low BMD. This may be owing to seasonal variation in vitamin D levels. These data suggest that vitamin D deficiency is not the main driver of bone demineralization. Two of the high-quality studies included found that vitamin D levels in the affected populations, although low, were not different from controls. In certain etiologies, malnutrition (as evidenced by low BMI) was a strongly associated factor. BMI is a recognized indicator of nutritional status and therefore this association may show that undernutrition in chronic pancreatitis is an important contributor to loss of BMD. Because female sex (postmenopausal) is a known risk factor for low BMD, failure to identify a relationship between sex and low bone density is surprising and may suggest that factors relating to the chronic pancreatitis disease process itself are the drivers of bone demineralization. Similarly, although one would expect older patients to have lower BMD owing to normal age-related bone demineralization, there was no apparent association between age and BMD. Therefore, although it could be argued that the complications of chronic pancreatitis might cause death before the economic burden of osteoporosis is realized, the studies in this review showed that osteoporosis is prevalent in a relatively young patient group. It is probable that increasing age and, for women, menopause, will have an additive effect on osteopathy. It is likely that the pathogenesis of low BMD is multifactorial and is driven by a mix of co-existing factors. In celiac disease and inflammatory bowel disease, normal regulatory control of bone metabolism may be affected adversely by proinflammatory cytokines released from inflamed intestines as well as from mature T cells. 17 The effect of chronic inflammation on bone metabolism in chronic pancreatitis has not been characterized. We are aware of only one study that has investigated bone histomorphometry in chronic pancreatitis. 18 In a study from South Africa with 13 African male (alcoholinduced) chronic pancreatitis patients and 37 controls, Schnitzler et al 18 obtained bone biopsy specimens to examine cortical and trabecular iliac crest bone by histomorphometry for microarchitectural and bone turnover abnormalities. They showed loss of cortical thickness and trabecular bone volume with microarchitectural deterioration in chronic pancreatitis patients compared with controls. In addition, 25-OH-D (vitamin D) levels were lower in chronic pancreatitis patients than in controls. They reported that the histomorphometric data were consistent with alcohol bone disease. There is also a gap in the literature regarding the prevalence of fractures in patients with chronic pancreatitis, which is ultimately the important clinical end point of osteoporosis. In one noteworthy study, Tignor et al 19 conducted a retrospective cohort study that examined fracture prevalence in 3192 chronic pancreatitis patients. They estimated the prevalence of fractures in chronic pancreatitis compared with high-risk gastrointestinal diseases and reported that the prevalence of fracture was 4.8% in chronic pancreatitis, compared with 1.1% in controls, 3.0% in Crohn s disease, and 5.0% in celiac disease. They concluded that there is an increased risk for fragility (low-trauma) fracture in chronic pancreatitis patients, which was at least comparable with the other high-risk groups. The systematic review and meta-analysis had several limitations. The clinical and methodologic diversity of the included studies means that any conclusions, particularly of meta-analyses, must be interpreted with caution. Although we conducted an overall meta-analysis of the prevalence, the data did not support a meta-analysis of the various subgroups. There is a possibility of caseselection bias in individual studies, which may inflate the true prevalence. Similarly, publication bias may

9 February 2014 Osteoporosis in Chronic Pancreatitis 227 preclude the publication of negative studies, which may overestimate the true prevalence further. Studies differed in terms of patient type, disease severity, and assessment methods. Individual studies had a small sample size (n < 100; indeed, one study had only 14 patients), leading to wide confidence intervals. The reporting quality of included studies varied somewhat, with only 4 studies achieving high-quality scores, and 4 studies being classified as low quality. One of the studies was authored by one of the reviewers (S.N.D.), which may have introduced bias in the study assessment. Notably, only 4 of 11 studies included the recruitment of a comparative control arm. Finally, combining studies from vastly different countries is also a limitation. For example, although 8 of the studies originated in Europe, 2 were from India and 1 was from South America. However, in the case of the studies from India, although the patients were younger than those in the European studies, the rates of osteoporosis and overall osteopathy were not vastly different and these studies were not considered outliers. Perhaps more importantly, most or all of the patients in the 2 studies from India had tropical chronic pancreatitis, which is associated with malnutrition. Therefore, bone demineralization for this group may be partly secondary to undernutrition. However, when these studies were excluded from the metaanalysis, the results were not greatly altered. Despite the limitations, the systematic review was performed within the constraints of rigorous methodology. The British Society of Gastroenterology published guidelines for osteoporosis in inflammatory bowel disease and celiac disease in In its systematic review, it reported 18 cross-sectional studies of bone density in Crohn s disease, reporting an osteoporosis prevalence of between 0% and 58%. Although acknowledging the limitations of a pooled analysis, it reported that T-scores were within the osteoporotic or osteopenic range in more than half of patients. In our study, the comparable figure was 65%, a finding that is even more striking when you consider that patients with chronic pancreatitis do not have the added risk factor of cumulative corticosteroid doses that are associated with reduced BMD in Crohn s disease. 1 The American Gastroenterological Association, as part of its recommendations on osteoporosis in gastrointestinal disease, recommended that patients with one or more known osteoporosis risk factors should undergo initial screening with DXA. 20 These recommendations were for inflammatory bowel disease, celiac disease, and postgastrectomy patients, but did not cover chronic pancreatitis, likely owing to a lack of studies at the time. However, based on the data from this systematic review, we suggest that patients with chronic pancreatitis should be afforded the same screening protocol. In summary, our systematic review found limited data on the prevalence of osteoporosis in chronic pancreatitis. We suggest that high-quality prevalence studies that clearly define the study population and outcomes, and include larger sample sizes and a wellmatched control arm, should be performed. We also suggest that studies focusing on bone metabolism are warranted to fully understand the bone remodeling process in chronic pancreatitis. Nevertheless, the available data show a notably high prevalence of osteoporosis and overall osteopathy, which merits the development of bone health guidelines for the management of this patient group. Our findings have direct implications for clinical practice. We recommend that bone-health assessment should form an integral part of the medical/nutritional work-up of patients with chronic pancreatitis to prevent and treat osteoporosis and, ultimately, to prevent fracture and its associated morbidity. References 1. Lewis NR, Scott BB. Guidelines for osteoporosis in inflammatory bowel disease and coeliac disease. Written and approved by the British Society of Gastroenterology Available at: Moran CE, Sosa EG, Martinez SM, et al. Bone mineral density in patients with pancreatic insufficiency and steatorrhea. Am J Gastroenterol 1997;92: Dujsikova H, Dite P, Tomandl J, et al. Occurrence of metabolic osteopathy in patients with chronic pancreatitis. Pancreatology 2008;8: Duggan SN, O Sullivan M, Hamilton S, et al. Patients with chronic pancreatitis are at increased risk for osteoporosis. Pancreas 2012;41: Joshi A, Reddy SV, Bhatia V, et al. High prevalence of low bone mineral density in patients with tropical calcific pancreatitis. Pancreas 2011;40: Wells G, Shea B, O Connell D. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analyses. In: Proceedings of the 3rd Symposium on Systematic Reviews Beyond the Basics: improving quality and impact. Oxford, England: Available at: programs/clinical_epidemiology/oxford.htm. 7. Neyeloff JL, Fuchs SC, Moreira LB. Meta-analyses and Forest plots using a microsoft excel spreadsheet: step-by-step guide focusing on descriptive data analysis. BMC Res Notes 2012;5: Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology. A proposal for reporting. JAMA 2000;283: Sudeep K, Chacko A, Thomas N, et al. Predictors of osteodystrophy in patients with chronic nonalcoholic pancreatitis with or without diabetes. Endocr Pract 2011;17: Drozdov VN, Chernyshova IV, Vinokurova LV, et al. [Role of exocrine pancreatic insufficiency in reducing of the bone mineral density in patients with chronic pancreatitis]. Eksp Klin Gastroenterol 2010;8: Mann ST, Stracke H, Lange U, et al. Alterations of bone mineral density and bone metabolism in patients with various grades of chronic pancreatitis. Metabolism 2003;52: Haaber AB, Rosenfalck AM, Hansen B, et al. Bone mineral metabolism, bone mineral density, and body composition in patients with chronic pancreatitis and pancreatic exocrine insufficiency. Int J Pancreatol 2000;27: Sikkens ECM, Cahen DL, Kuipers EJ, et al. Prosepctive evaluation of the prevalence of fat-soluble vitamin deficiencies and decreased bone mineral density in chronic pancreatitis. Gastroenterology 2012;142:s461 s462.

10 228 Duggan et al Clinical Gastroenterology and Hepatology Vol. 12, No Skipworth JRA, Chapman MH, Johnson G, et al. Bone mineral density in patients with chronic pancreatitis. Pancreatology 2012;12:e Gubergrits NB, Kutyrkina IL. Mineral density of bones in chronic pancreatitis. Pancreatology 2011;11(Suppl 1): Toouli J, Biankin AV, Oliver MR, et al. Management of pancreatic exocrine insufficiency: Australasian Pancreatic Club recommendations. Med J Aust 2010;193: Cashman KD. Altered bone metabolism in inflammatory disease: role for nutrition. Proc Nutr Soc 2008;67: Schnitzler CM, Mesquita JM, Shires R. Cortical and trabecular bone microarchitecture and turnover in alcohol-induced chronic pancreatitis: a histomorphometric study. J Bone Miner Metab 2010;28: Tignor AS, Wu BU, Whitlock TL, et al. High prevalence of lowtrauma fracture in chronic pancreatitis. Am J Gastroenterol 2010;105: American Gastroenterological Association medical position statement: guidelines on osteoporosis in gastrointestinal diseases. Gastroenterology 2003;124: Reprint requests Address requests for reprints to: Sinead N. Duggan, PhD, 1.29 Department of Surgery, Trinity Centre for Health Sciences, Trinity College Dublin, Tallaght Hospital, Dublin 24, Ireland. siduggan@tcd.ie; fax: (00) (353) Acknowledgments The authors wish to acknowledge Tim Grant and Ricardo Segurado (Medical Statisticians from the Centre for Support and Training in Analysis and Research, University College Dublin) for statistical assistance and advice. The authors also wish to acknowledge contributions to discussions by members of The Centre for Pancreatic-Biliary Disease, and the Department of Clinical Nutrition & Dietetics, Tallaght Hospital, Dublin, Ireland. The authors are grateful to the following researchers for providing additional information on individual studies: E. Sikkens, M. Bruno, E. Bhatia, V. Reddy, J. Skipworth, S. Pereira, S. Duggan, and V. Drozdov. The authors also thank Corporate Translation Services, Dublin, Ireland (ISO 9001:2008 certified) for translation of Russian manuscripts. Conflicts of interest The authors disclose no conflicts. Funding This study was supported by an unrestricted grant from the Health Research Board, Ireland, by means of a Health Professional s Fellowship (HPF 2009/046 to S.N.D.).