Letters RESEARCH LETTER Presence of Banned Drugs in Dietary Supplements Following FDA Recalls The US Food and Drug Administration (FDA) initiates class I drug recalls when products have the reasonable possibility of causing serious adverse health consequences or death.1 Recently, the FDA has used class I drug recalls in an effort to remove dietary supplements adulterated with pharmaceutical ingredients from US markets. Approximately half of all FDA class I drug recalls since 2004 have involved dietary supplements adulterated with banned pharmaceutical ingredients.2,3 Prior research has found that even after FDA recalls, dietary supplements remain available on store shelves.4 However, it is not known if the supplements on sale after FDA recalls are free of the adulterants. In the present study, dietary supplements purchased at least 6 months after FDA recalls were analyzed to determine if banned drugs were still present. Methods | Dietary supplements were analyzed if they met the following inclusion criteria: (1) recalled due to adulteration with pharmaceutical ingredients between January 1, 2009, and December 31, 20123,5; (2) available for purchase in July or August 2013 directly from websites of supplement manufacturers or retailers (as opposed to general e-commerce sites such as Amazon.com, eBay Inc, or Alibaba Group); and (3) the supplement name, manufacturer, and distributor listed on the purchased supplement was identical to the information provided in the FDA recall. Dietary supplements were analyzed by Flora Research Laboratories (J.N-K.). Samples were labeled with the marketing claim on the supplement label (eg, weight loss, sexual enhancement, or sports enhancement), but did not include the supplement name, manufacturer, and prior FDA findings. Analyses were performed using the same methods that the FDA’s field laboratories use to screen for clandestine adulteration. In short, dietary supplements were analyzed using either gas chromatography mass spectrometry or liquid chromatography tandem mass spectrometry in data-triggered mode. Adulterants, except for anabolic steroids, were confirmed against a standard using retention time, mass spectrum, and UV spectrum. Results | The FDA recalled 274 dietary supplements between January 1, 2009, and December 31, 2012. Twenty-seven of the 274 recalled supplements (9.9%) met our inclusion criteria and were analyzed. Supplements were purchased a mean (SD) of 34.3 (11.5) months after the FDA recall (range, 8-52 months). Seventy-four percent of supplements (20/27) were produced by US manufacturers. One or more pharmaceutical adulterant was identified in 66.7% of recalled supplements still available for purchase (18/27; Table). Supplements remained adulterated in 85% (11/ 13) of those for sports enhancement, 67% (6/9) for weight loss, and 20% (1/5) for sexual enhancement. Of the subset of supplements produced by US manufacturers, 65% (13/20) remained adulterated with banned ingredients. Sixty-three percent of analyzed supplements (17/27) contained the same adulterant identified by the FDA. Six of the 27 (22.2%) supplements contained 1 or more additional banned ingredients not identified by the FDA (Table). Some supplements contained both the previously identified adulterant as well as additional pharmaceutical ingredients. Banned substances identified in recalled supplements included sibutramine, sibutramine analogs, sildenafil, fluoxetine, phenolphthalein, aromatase inhibitor, and various anabolic steroids. One novel adulterant, benzyl sibutramine, was first described as recently as 2013.6 Discussion | To our knowledge, this is the first study to determine if adulterants remain in supplements sold after FDA recalls. We found that 66.7% of recalled supplements still available for purchase at least 6 months after FDA recalls remained adulterated with banned ingredients. Our study has several limitations. First, we limited testing to common adulterants expected based on marketing claims (eg, weight loss supplements were tested for adulterants commonly found in weight loss products). Second, our analyses may have failed to detect recently introduced drug analogs. Third, although every effort was made to purchase recently manufactured supplements, it is not known if all supplements were manufactured after the FDA recall. Action by the FDA has not been completely effective in eliminating all potentially dangerous adulterated supplements from the US marketplace. More aggressive enforcement of the law, changes to the law to increase the FDA’s enforcement powers, or both will be required if sales of these products are to be prevented in the future. Pieter A. Cohen, MD Gregory Maller, BS Renan DeSouza James Neal-Kababick, BS Author Affiliations: Harvard Medical School, Boston, Massachusetts (Cohen); Cornell University, Ithaca, New York (Maller, DeSouza); Flora Research Laboratories, Grants Pass, Oregon (Neal-Kababick). Mr Maller is now with a venture capital firm. Corresponding Author: Pieter Cohen, MD, 236 Highland Ave, Somerville, MA 02143 (pcohen@challiance.org). Author Contributions: Dr Cohen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Cohen. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: Cohen. jama.com JAMA October 22/29, 2014 Volume 312, Number 16 1691 Copyright 2014 American Medical Association. All rights reserved. Letters Table. Pharmaceutical Adulterants Identified in Recalled Dietary Supplements Purchased at Least 8 Months After US Food and Drug Administration (FDA) Recalls Recalled Supplement 2a, 17a Methastadrol 4-ad Açai-Man Mangosteen Botanical Slimming 100% Natural Softgel E-pol: Inslinsified Everlax capsules EverSlim Finaflex 550-XD Forged Extreme Mass Joyful Slim M-Drol Magic Power Coffee Massdrol Mince Belle Novedex XT On Cycle II Hardcore P-Plex Pandora: Sexual Enhancer for Women RockHard Weekend Testra-flex Slim-30 Slim Forte Slimming Capsule Slim Xtreme Herbal Slimming Capsule Stamina-RX Trenadrol X-TREN Zi Xiu Tang Bee Pollen Capsule Date of FDA Recalla,b November 3, 2009 September 16, 2010 February 3, 2012 September 2, 2011 Date Purchased August 2013 July 2013 July 2013 August 2013 Expiration Date on Purchased Supplement Adulterant Found by FDAa,c January 2014 Steroid or steroid-like compound or analog March 2015 Aromatase inhibitor Not available Tadalafil Not available Sibutramine Adulterant Found After FDA Recallc Anabolic steroid None identified None identified Sibutramine November 3, 2009 August 2013 February 3 and 22, 2012 July 2013 February 3, 2012 July 2013 November 3, 2009 August 2013 November 3, 2009 August 2013 July 22, 2010 July 2013 November 3, 2009 August 2013 June 25, 2010 August 2013 November 3, 2009 August 2013 February 3, 2012 July 2013 January 15 and October 7, 2010 July 2013 November 3, 2009 August 2013 November 3, 2009, and January 15, 2010 December 23, 2010 August 2013 July 2013 Not available Not available January 2018 June 2014 November 2011 December 2013 Not available February 2015 Not available Not available July 2013 Not available October 2012 October 2013 Steroid or steroid-like compound or analog Two anabolic steroids Sibutramine None identified Sibutramine Steroid or steroid-like compound or analog Steroid or steroid-like compound or analog Desmethylsibutramine (an analog of sibutramine) Fluoxetine, sibutramine Two anabolic steroids Anabolic steroid None identified Steroid or steroid-like compound or analog Anabolic steroid Hydroxythiohomosildenafil (an analog of sildenafil) Steroid or steroid-like compound or analog Sibutramine Aromatase inhibitor and steroid or steroid-like compound or analog Sildenafil Anabolic steroid Fluoxetine, N-didesmethyl sibutramine (an analog of sibutramine) Aromatase inhibitor and an anabolic steroid Steroid or steroid-like compound or analog Two anabolic steroids Steroid or steroid-like compound or analog Analog of sildenafil Anabolic steroid impurities None identified November 9, 2009, and December 22, 2010 January 15, 2010 July 2013 July 2013 July 16 and August 18, 2010 July 27, 2011 July 2013 July 2013 May 11, 2011 July 2013 June 15, 2009 August 2013 November 3, 2009 August 2013 November 3, 2009, and January 15, 2010 October 24, 2012 August 2013 July 2013 March 2014 May 2014 December 2015 April 2018 Sulfoaildenafil (an analog of sildenafil) Steroid or steroid-like compound or analog Desmethyl sibutramine (an analog of sibutramine) Sibutramine January 2015 Sibutramine September 2014 Not available September 2014 July 2015 Benzamidenafil (an analog of sildenafil) Steroid or steroid-like compound or analog Steroid or steroid-like compound or analog Sibutramine None identified Anabolic steroid None identified Sibutramine, phenolphthalein Sibutramine, phenolphthalein, benzyl sibutramine (an analog of sibutramine) None identified Anabolic steroid None identified Sibutramine, phenolphthalein a Information available at http://www.fda.gov/ForConsumers/ProtectYourself/HealthFraud/ucm255499.htm. b May have been recalled more than once. c May have included more than 1 adulterant. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Maller, DeSouza. Obtained funding: Cohen. Administrative, technical, or material support: Maller, Neal-Kababick. Study supervision: Cohen. Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Mr Maller reported receiving personal fees from NY State Senator Jeffrey Klein and SSP Nutrition outside of the submitted work. Mr Neal-Kababick reported being the vice chair of the US Pharmacopeia expert panel on Adulteration of Dietary 1692 JAMA October 22/29, 2014 Volume 312, Number 16 Copyright 2014 American Medical Association. All rights reserved. jama.com Letters Supplements with Drugs and Drug Analogues; reported being a co-owner of Flora Research Laboratories (some of the clients are dietary supplement manufacturers); and reported serving as an expert witness in cases involving the investigation of quality issues in the production of dietary supplements. No other disclosures were reported. Funding/Support: This research was supported in part by a grant from Consumer Union. Role of the Funder/Sponsor: Consumer Union had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Additional Contributions: We acknowledge Mercy Imahiyerobo, MPH, JD (Harvard School of Public Health), for her assistance in identifying recalled supplements; she was not compensated for her contribution. Ms Imahiyerobo is now with a private law firm. 1. US Food and Drug Administration. Safety: recalls, market withdrawals, & safety alerts. http://www.fda.gov/Safety/Recalls/ucm165546.htm. Accessibility verified September 29, 2014. 2. Harel Z, Harel S, Wald R, Mamdani M, Bell CM. The frequency and characteristics of dietary supplement recalls in the United States. JAMA Intern Med. 2013;173(10):926-928. 3. US Food and Drug Administration. Tainted_supplements_CDER. http://www.accessdata.fda.gov/scripts/sda/sdNavigation.cfm?filter =&sortColumn=1d&sd=tainted_supplements_cder&page=1. Accessibility verified September 29, 2014. 4. Cohen PA, Benner C, McCormick D. Use of a pharmaceutically adulterated dietary supplement, Pai You Guo, among Brazilian-born women in the United States. J Gen Intern Med. 2012;27(1):51-56. 5. US Food and Drug Administration. Recalls—health fraud. http://www.fda.gov /ForConsumers/ProtectYourself/HealthFraud/ucm255499.htm. Accessibility verified September 29, 2014. 6. Mans DJ, Gucinski AC, Dunn JD, et al. Rapid screening and structural elucidation of a novel sibutramine analogue in a weight loss supplement: 11-desisobutyl-11-benzylsibutramine. J Pharma Biomed Anal. 2013;83:122-128. COMMENT & RESPONSE Insulin vs Sulfonylureas for Second-Line Diabetes Treatment To the Editor Dr Roumie and colleagues1 reported that compared with sulfonylureas the addition of insulin to metformin to improve glycemic control was associated with an increased risk of a composite of nonfatal cardiovascular outcomes and all-cause mortality in patients with diabetes mellitus. Although the results presented require verification in carefully designed clinical trials, we have some concerns about the current analyses. Roumie et al1 stated that subgroup analyses stratifying by age were consistent with the primary analysis. However, eFigure 3 showed that the hazard ratio (HR) for cardiovascular events or death was 1.14 (95% CI, 0.82-1.59) for patients younger than 65 years, whereas for those aged 65 years or older, the HR was 1.51 (95% CI, 1.19-1.92). For the composite death outcome, the HR was 1.28 (95% CI, 0.85-1.97) for patients younger than 65 years, whereas for those aged 65 years or older, the HR was 1.66 (95% CI, 1.29-2.14). Contrary to the conclusion made by the authors, the results suggest that insulin added to metformin is safe among patients younger than 65 years but appears to be hazardous in older patients with diabetes. Part of the explanation may be differences in types of insulin used and hypoglycemia by age group. Of the patients in the metformin plus insulin group, 47% used long-acting agents only; 22%, both long- and short-acting agents; 17%, premixed insulin; and 11%, short-acting agents only. Although premixed insulin use in older adults can increase the risk of hypoglycemia compared with long-acting analogs,2 even the choice of regular or analog insulin can influence the frequency of hypoglycemic episodes when multiple injections are used.3 Thus, the type of insulin chosen for older patients with diabetes should be taken into consideration in such database analyses. In addition, below a hemoglobin A1c level of 8%, the risk of treatment-induced hypoglycemia increases.4 Because mean achieved hemoglobin A1c levels were 7% among sulfonylurea users and 6.9% among insulin users in this study, missing hypoglycemia data are of concern. Even though Roumie et al1 used an extensive database, they were unable to obtain data on hypoglycemia frequency, especially the number of severe events requiring hospital admission. Such information is essential for appropriate interpretation of the study. Although sulfonylurea drugs can also increase the risk hypoglycemia of among older patients, comparative data are still required. Ilker Tasci, MD Umut Safer, MD Author Affiliations: Department of Internal Medicine, Gulhane School of Medicine, Ankara, Turkey. Corresponding Author: Ilker Tasci, MD, Department of Internal Medicine, Gulhane School of Medicine, GATA Ic Hastaliklari BD, Etlik 06018 Ankara, Turkey (itasci@gata.edu.tr). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Roumie CL, Greevy RA, Grijalva CG, et al. Association between intensification of metformin treatment with insulin vs sulfonylureas and cardiovascular events and all-cause mortality among patients with diabetes. JAMA. 2014;311(22): 2288-2296. 2. Janka HU, Plewe G, Busch K. Combination of oral antidiabetic agents with basal insulin versus premixed insulin alone in randomized elderly patients with type 2 diabetes mellitus. J Am Geriatr Soc. 2007;55(2):182-188. 3. Velussi M. Lispro insulin treatment in comparison with regular human insulin in type 2 diabetic patients living in nursing homes. Diabetes Nutr Metab. 2002; 15(2):96-100. 4. Bramlage P, Gitt AK, Binz C, Krekler M, Deeg E, Tschöpe D. Oral antidiabetic treatment in type-2 diabetes in the elderly: balancing the need for glucose control and the risk of hypoglycemia. Cardiovasc Diabetol. 2012;11:122. In Reply We agree with Drs Tasci and Safer that the risk of metformin and insulin compared with metformin plus sulfonylurea as a second-line diabetes treatment after failure of metformin monotherapy may differ in certain populations. They point to eFigure 3 that shows a statistically significant increased risk in persons aged 65 years or older and no statistically significant increase in younger persons. However, the confidence intervals for the HRs in the 2 age groups have considerable overlap and a formal test for interaction between metformin plus insulin and age is not significant (P = .22). Thus, we have not proven significant differences in risk by age. The numerical differences observed could represent real differences or simply chance variation in the estimated associations. jama.com JAMA October 22/29, 2014 Volume 312, Number 16 1693 Copyright 2014 American Medical Association. All rights reserved.