Original Article Elizabeth Kos, MD 1 ; Mary Jo Liszek, MD 2 ; Mary Ann Emanuele, MD 3 ; Ramon Durazo-Arvizu, PhD 4 ; Pauline Camacho, MD, FACE 3 ABSTRACT Objective: To determine the effect of metformin on 25-hydroxyvitamin D [25(OH)D] and vitamin levels in patients with type 2 diabetes mellitus. Methods: We performed a retrospective review of medical records of patients treated between 2003 and 2009 at Loyola University Medical Center, Maywood, Illinois, in both ambulatory primary care and endocrinology clinics. The study cohort consisted of 706 patients with type 2 diabetes mellitus who were 20 to 93 years old (mean age, 63 ± 13) and had a mean body mass index of 33.1 kg/m 2. Of these patients, 42% were treated with metformin, and 34% had been diagnosed with osteoporosis or osteopenia. Results: Patients taking metformin had statistically significant lower vitamin levels than those not receiving metformin (P<.0001; 95% confidence interval [CI] = -220 to -84 pg/ml). No statistically significant difference was found between users and nonusers of metformin in regard to 25(OH)D levels when adjusted for variables (P =.297; 95% CI for mean difference = -0.7 to 2.2 ng/ml). Metformin use did not adversely affect successful treatment of vitamin D deficiency in this patient population as a whole, nor did it affect the subgroup with osteoporosis Submitted for publication December 10, 2010 Accepted for publication August 11, 2011 From the 1 Department of General Internal Medicine, 2 Department of Internal Medicine and Pediatrics, 3 Department of Endocrinology and Metabolism, and 4 Department of Epidemiology, Loyola University Medical Center, Maywood, Illinois. Address correspondence to Dr. Elizabeth Kos, Loyola University Medical Center, 2160 South 1st Avenue, Room 7611A, Maywood, IL 60153. E-mail: ekos@lumc.edu. Published as a Rapid Electronic Article in Press at http://www.endocrine practice.org on September 22, 2011. DOI:10.4158/EP11009.OR To purchase reprints of this article, please visit: www.aace.com/reprints. Copyright 2012 AACE. (P =.956). The patients with osteoporosis had statistically significant lower baseline 25(OH)D levels in comparison with those without osteoporosis, when adjustments were made for all variables (P =.003; 95% CI = 0.7 to 3.5 ng/ ml). Conclusion: This study confirms the higher prevalence of vitamin deficiency in metformin-treated patients with type 2 diabetes than in those not treated with metformin. This study also suggests that vitamin D deficiency is not a clinical concern among metformin-treated patients with type 2 diabetes and that metformin does not negatively affect treatment of vitamin D deficiency in these patients. (Endocr Pract. 2012;18:179-184) Abbreviations: BMI = body mass index; CI = confidence interval; 25(OH)D = 25-hydroxyvitamin D INTRODUCTION Type 2 diabetes mellitus currently affects 8% of the population of the United States and 11% of adults beyond the age of 20 years. This disease has substantial associated morbidity and mortality, with the potential to affect adversely the cardiovascular, renal, and neurologic systems of each person with the disease (1). Treatment of type 2 diabetes includes many options such as lifestyle changes, a variety of medications, and insulin. One medication that has proved effective in treatment of type 2 diabetes and is considered the first choice for oral management, as recommended by the American Diabetes Association, is a drug in the biguanide class, metformin. Metformin acts through multiple mechanisms including decreasing hepatic glucose output, increasing insulinmediated glucose use in peripheral tissues, and increasing intestinal glucose utilization. Most side effects of metformin are mild and can include gastrointestinal distress, soft stools, and diarrhea (2). It is well documented that these gastrointestinal side effects can lead to malabsorption of ENDOCRINE PRACTICE Vol 18 No. 2 March/April 2012 179
180 Metformin and Vitamin D and Levels, Endocr Pract. 2012;18(No. 2) vitamin in a dose- and time-dependent manner (3,4). The effect of metformin on vitamin levels in patients with diabetes has been studied for decades; on average, 10% to 30% of patients show a malabsorptive deficiency of vitamin (5,6). The mechanism has not been fully elucidated, although it has been speculated to be bacterial overgrowth attributable to diabetes, alteration of small bowel motility, alteration of bacterial flora, or even the effect of calcium on cell membranes (7). Long-term treatment with metformin has been shown to have a progressive effect, with increasingly significant declines over time. Although the clinical significance of short-term treatment is variable, it has been recommended that those receiving long-term treatment undergo screening (8). Vitamin D has a small but important role in diabetes. Investigators have suggested that an association exists between decreased vitamin D and calcium status and the risk of type 2 diabetes as well as between vitamin D deficiency and impaired glucose-mediated insulin release (9,10). Vitamin D is produced endogenously when ultraviolet rays initiate synthesis in the skin, but vitamin D is also absorbed in the intestine from various natural and fortified food sources (11). If gastrointestinal symptoms from metformin use can create an environment conducive to the malabsorption of vitamin, it is probable that vitamin D could also be malabsorbed. This could adversely affect treatment efficacy as well as expose certain patient populations to the risk of symptoms of vitamin D deficiency. Currently, no published studies have examined the effect of metformin on vitamin D levels in a population of patients with diabetes. This study focuses on the prevalence of deficiencies of both vitamin D and vitamin in a population of patients with type 2 diabetes using the medication metformin as primary treatment. OBJECTIVE The primary objective of this study was to determine the effect of metformin on vitamin D and levels in patients with type 2 diabetes mellitus whose primary treatment was metformin. The secondary objective was to ascertain whether use of metformin affected treatment of vitamin D deficiency. RESEARCH DESIGN AND METHODS This study was approved by the Institutional Review Board of Loyola University Medical Center, Maywood, Illinois. Medical records and laboratory data of patients with a preexisting diagnosis of type 2 diabetes mellitus seen at 1 of the 3 Loyola University Medical Center general medicine or subspecialty endocrinology clinics from 2003 to 2009 were retrospectively reviewed. These were consecutive patients seen in these clinics for general follow-up appointments, and the sample size was determined by the number of these patients who fulfilled the inclusion criteria during the specified years of evaluation. Patients were included in the study if they had a preexisting diagnosis of type 2 diabetes and had multiple documented levels of either 25-hydroxyvitamin D [25(OH)D] or vitamin for review. Exclusion criteria included patients with newly diagnosed type 2 diabetes mellitus or patients newly started on metformin as primary or supplemental therapy for their diabetes. There were no exclusion criteria related to preexisting malabsorptive gastrointestinal disease or medications known to cause malabsorption of vitamin D or vitamin. Baseline demographic information of these patients was collected from the electronic medical records and included age, sex, body mass index (BMI), race, multivitamin or vitamin supplementation, season during which 25(OH)D levels were determined, and preexisting diagnosis of osteoporosis or osteopenia. Levels of the following laboratory data were also collected: 25(OH)D (in ng/ml) for up to 8 visits over 4 years and vitamin (in pg/ml) for up to 4 visits over 4 years. No values past the eighth for vitamin D or the fourth for vitamin were included because the number of patients with these values declined substantially after these visits. More documented values existed for 25(OH)D, inasmuch as patients are regularly screened for vitamin D deficiency in both the general medicine and the endocrinology subspecialty clinics. Levels of 25(OH)D were determined with use of the DiaSorin Liaison chemiluminescent immunoassay (DiaSorin, Saluggia, Italy) with a reference range of 30 to 100 ng/ ml. Other measures of vitamin such as methylmalonic acid were not used for this study, primarily because of lack of significant numbers of patients with multiple determinations for analysis. Likewise, duration of metformin therapy could not be assessed because length of treatment was not recorded for the majority of patients in the electronic medical records. Thus, because it was not pertinent to the primary objective and few patients had accurate treatment duration documented, it was not included in the analysis. This retrospective review revealed that the patients deficient in 25(OH)D were treated in the traditional manner with 50,000 IU of vitamin D once weekly for months, with a reassessment of 25(OH)D levels at 3 months. There was no standardization for timing of replacement therapy. All statistical analyses were performed with use of Stata version 11.0 software for Windows (StataCorp, College Station, Texas), 2-tailed tests, and P<.05 as the threshold for statistical significance. Mixed-effects regression models were used to investigate the association between vitamin D (or vitamin ) and use of metformin, with adjustments for confounding variables (that is, age, osteoporosis, and BMI). These methods allow the use of all available data when some values are missing at random and also account for the within-person correlation of the
Metformin and Vitamin D and Levels, Endocr Pract. 2012;18(No. 2) 181 measurements. Statistical power was noted to be greater than 80%. RESULTS Demographic baseline characteristics of study patients with multiple 25(OH)D and multiple vitamin measurements are reported in Tables 1 and 2, respectively. A total of 706 patients with type 2 diabetes mellitus, ranging in age from 20 to 93 years, were identified for this study. Overall, the mean age was 63 ± 13 years (standard deviation), and the mean BMI was 33.1 kg/m 2. Of these patients, 42% were treated with metformin (mean daily dose, 1.5 g), and 34% had been diagnosed with osteoporosis or osteopenia. Of the 706 patients, 30.8% described their race as nonwhite (17.7% of patients self-reported race as African American, 69.2% as Caucasian, 8.6% as Hispanic, and 4.5% as other/prefer not to mention). A total of 37.6% of patients in this investigation reported taking a multivitamin. Only 6.6% of patients were recorded as taking either an orally administered or an intramuscularly administered vitamin replacement. Of the 706 patients, 4 (0.6%) did have known malabsorptive disease (3 patients diagnosed with celiac disease and 1 patient with Crohn disease). Of these 706 patients, 698 had multiple vitamin D determinations and were included in the vitamin D analysis (Table 1). Blood specimens were collected during all seasons, with a slight preponderance of winter months (22.0% in the spring months, 21.5% in the summer months, 24.7% in the fall months, and 31.8% in the winter months). Of the 706 patients, 347 had multiple vitamin measurements and were therefore included in the vitamin analysis (Table 2). Levels of 25(OH)D in those patients treated with metformin were not significantly different from those not treated with metformin (P =.297). The adjusted mean difference between the groups was estimated by means of the regression model as 0.7 ng/ml (95% confidence interval [CI] = -0.7 to 2.2) when adjustments were made for age, osteoporosis, and BMI. The presence of osteoporosis was significantly associated with lower 25(OH)D levels regardless of metformin use (P =.002; mean difference = 2.36 ng/ml; 95% CI = 0.8 to 3.9) even after adjustment for the study population s baseline mean 25(OH)D level (P =.003; mean difference = 2.09 ng/ml; 95% CI = 0.7 to 3.5). Metformin therapy had no effect on normalization of 25(OH)D levels over time in those patients treated with vitamin D replacement (P =.956 for interaction between time and metformin use) (Fig. 1). Vitamin levels were significantly lower in those treated with metformin than in those not treated with metformin (P<.0001). The regression-adjusted mean vitamin difference was -152.2 pg/ml (95% CI for mean difference = -220 to -84) when adjustments were made for age, BMI, and prior diagnosis of osteoporosis. These results are demonstrated in Figure 2. Osteoporosis and age did not have a significant effect on vitamin levels (osteoporosis: P =.224; mean difference = 42.6 pg/ml; 95% Factor Table 1 Baseline Characteristics of 698 Patients With Multiple 25-Hydroxyvitamin D Levels Value Age (y) a 63 ± 12 Body mass index (kg/m 2 ) a 33 ± 8 Male (%) 29.20 Metformin No metformin P use (n = 403) use (n = 295) value Baseline vitamin D (ng/ml) 22.9 ± 11.1 22.3 ± 11.5.297 Age (y) 61.9 ± 11.6 64.7 ± 13.7.096 Body mass index (kg/m 2 ) 34.7 ± 8.6 32.0 ± 8.081 With Without osteoporosis osteoporosis P (n = 241) (n = 457) value Baseline vitamin D (ng/ml) 21.8 ± 11.5 23.98 ± 11.003 Age (y) 70.21 ± 9.9 59.9 ± 12.9 Body mass index (kg/m 2 ) 31.64 ± 7.9 33.96 ± 8.5 a Values are shown as mean ± standard deviation. All other values (except sex) are shown as mean ± standard error.
182 Metformin and Vitamin D and Levels, Endocr Pract. 2012;18(No. 2) Factor Table 2 Baseline Characteristics of 347 Patients With Multiple Vitamin Levels Value Age (y) a 65 ± 13 Body mass index (kg/m 2 ) a 32.8 ± 8 Male (%) 30.30 Metformin No metformin P use (n = 142) use (n = 205) value Baseline vitamin (pg/ml) 496 ± 282 637 ± 352.0001 Age (y) 63.5 ± 12 66.5 ± 13.6.069 Body mass index (kg/m 2 ) 34.6 ± 8 31.7 ± 8.048 a Values are shown as mean ± standard deviation. All other values (except sex) are shown as mean ± standard error. CI = -26 to 111; age: P =.069; rate of change by year of age = -2.4 pg/ml; 95% CI = -4.9 to 0.2), whereas BMI appeared to have a statistically significant inverse association with vitamin levels (P =.048; rate of change by BMI unit = -3.96 pg/ml; 95% CI = -7.9 to -0.02). After adjustments were made for baseline levels of vitamin, BMI no longer had a statistically significant effect (P =.413; rate of change per BMI unit = -1.15 pg/ml; 95% CI = -3.9 to 1.6). Age also appeared statistically significant after baseline levels of vitamin were taken into account (P =.022; rate of change per year of age = -2.08 pg/ml; 95% CI = -3.9 to -0.3). DISCUSSION This study attempted to determine the effect of metformin on 25(OH)D levels in patients with type 2 diabetes mellitus as well as to validate previous studies regarding Fig. 1. Change in 25-hydroxyvitamin D levels (ng/ml) (vertical axis) over time (horizontal axis). *Average time between visits was 6 months. No significance was seen (P =.297) when adjustments were made for age, sex, and body mass index. Metformin did not affect normalization of 25-hydroxyvitamin D levels in vitamin D-deficient patients treated with replacement therapy (P =.956).
Metformin and Vitamin D and Levels, Endocr Pract. 2012;18(No. 2) 183 Fig. 2. Change in vitamin levels (pg/ml) (vertical axis) over time (horizontal axis). *Average time between visits was 1 year. Vitamin levels in those patients treated with metformin were significantly lower than in those not receiving metformin (P<.0001) by 152 ± 34 pg/ml (standard deviation) when adjustments were made for age, BMI, and prior diagnosis of osteoporosis. Body mass index appeared to have a statistically significant effect on vitamin levels (P =.048) but was not clinically significant with a decrease of only 3 ± 2 pg/ml. metformin-related vitamin deficiency. In regard to vitamin levels, this study clearly shows a decrease associated with metformin therapy, and this decrease could not be explained by any other variables in this analysis. This result was expected because it is well documented in the literature that metformin does cause a decrease in vitamin levels in a time- and dose-dependent manner. In contrast, when we evaluated the effect of metformin on 25(OH)D levels, this same decrease was not observed. In fact, the only variable related to decreasing 25(OH)D levels was a prior diagnosis of osteoporosis. The absence of an effect of metformin on 25(OH)D levels could support a hypothesis that the mechanism of action causing the metformin-related vitamin deficiency is not entirely malabsorption. Our study design had intrinsic limitations most notably, the data were collected retrospectively. We could not account for certain variables, including but not limited to exposure of patients to sunlight. In addition, any vitamin D deficiency detected during the study period we examined would have been treated appropriately with orally administered vitamin D supplementation, as discussed in the Research Design and Methods section. This allowed us to conclude that treatment of vitamin D deficiency and time to normalization of 25(OH)D levels were not negatively affected by metformin use. Unfortunately, any lowering effect that metformin may have had on 25(OH)D levels may have been masked by such repletion treatment. This factor could have skewed the results toward no significance between treatment and control groups. A prospective randomized trial with control for the aforementioned variables would need to be conducted to characterize the full effect of metformin on 25(OH)D levels in patients with type 2 diabetes. CONCLUSION In summary, this study confirms the higher prevalence of vitamin deficiency in metformin-treated patients with type 2 diabetes than in those not receiving metformin. It also suggests that vitamin D deficiency attributable to metformin therapy is not a clinical concern among this same patient population. In addition, metformin does not appear to have a negative effect on treatment of vitamin D deficiency in these patients. This finding has important implications for many elderly patients treated with metformin who have type 2 diabetes mellitus in conjunction with osteoporosis and vitamin D deficiency, inasmuch as metformin would not interfere with treatment of their comorbidities.
184 Metformin and Vitamin D and Levels, Endocr Pract. 2012;18(No. 2) ACKNOWLEDGMENT This research was published in abstract form in the meeting syllabus and presented in poster form at the 19th Annual Meeting and Clinical Congress of the American Association of Clinical Endocrinologists; April 21-25, 2010; Boston, Massachusetts. AUTHOR CONTRIBUTIONS Dr. Elizabeth Kos researched the data, contributed to the Discussion, wrote the manuscript, and reviewed and edited the manuscript. Dr. Mary Jo Liszek researched the data and contributed to the Discussion. Dr. Mary Ann Emanuele contributed to the Discussion and reviewed and edited the manuscript. Dr. Ramon Durazo-Arvizu contributed to the Discussion and reviewed and edited the manuscript. Dr. Pauline Camacho researched the data, contributed to the Discussion, and reviewed and edited the manuscript. DISCLOSURE The authors have no multiplicity of interest to disclose. REFERENCES 1. National Diabetes Information Clearinghouse (NDIC). National Diabetes Statistics, 2011. NIH publication no. 11-3892, February 2011. http://diabetes.niddk.nih.gov/dm/ pubs/statistics/. Accessed for verification October 5, 2011. 2. Nathan DM, Buse JB, Davidson MB, et al. Management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy; a consensus statement from the American Diabetes Association and the European Association for the Study of Diabetes [published correction appears in Diabetes Care. 2006;29:2816-2818]. Diabetes Care. 2006;29:1963-1972. 3. Ting RZ, Szeto CC, Chan MH, Ma KK, Chow KM. Risk factors of vitamin B12 deficiency in patients receiving metformin. Arch Intern Med. 2006;166:1975-1979. 4. Tomkin GH, Hadden DR, Weaver JA, Montgomery DA. Vitamin B12 status of patients on long-term metformin therapy. Br Med J. 1971;2:685-687. 5. Adams JF, Clark JS, Ireland JT, Kesson CM, Watson WS. Malabsorption of vitamin B12 and intrinsic factor secretion during biguanide therapy. Diabetologia. 1983;24: 16-18. 6. Bauman WA, Shaw S, Jayatilieke E, Spungen AM, Herbert V. Increased intake of calcium reverses vitamin B12 malabsorption induced by metformin. Diabetes Care. 2000;23:1227-1231. 7. Wulffelé MG, Kooy A, Lehert P, et al. Effects of shortterm treatment with metformin on serum concentrations of homocysteine, folate and vitamin B12 in type 2 diabetes mellitus: a randomized, placebo-controlled trial. J Intern Med. 2003;254:455-463. 8. De Jager J, Kooy A, Lehert P, et al. Long term treatment with metformin in patients with type 2 diabetes and risk of vitamin B-12 deficiency: randomised placebo controlled trial. BMJ. 2010;340:c2181. doi:10.1136/bmj.c2181. 9. Pittas AG, Lau J, Hu FB, Dawson-Hughes B. The role of vitamin D and calcium in type 2 diabetes: a systemic review and meta-analysis. J Clin Endocrinol Metab. 2007; 92:2017-2029. 10. Chiu KC, Chu A, Go VL, Saad MF. Hypovitaminosis D is associated with insulin resistance and beta cell dysfunction. Am J Clin Nutr. 2004;79:820-825. 11. National Institutes of Health (NIH) Office of Dietary Supplements. Dietary supplement fact sheet: vitamin D. http://ods.od.nih.gov/factsheets/vitamind_pf.asp. Accessed for verification October 5, 2011.