Americans' Use of Dietary Supplements That Are Potentially Harmful in CKD

Vanessa Grubbs, MD, MPH; Laura C. Plantinga, ScM; Delphine S. Tuot, MD, MAS; Elizabeth Hedgeman, MS, MPH; Rajiv Saran, MD, MS; Sharon Saydah, PhD; Deborah Rolka, MS; Neil R. Powe, MD, MPH, MBA


Am J Kidney Dis. 2013;61(5):739-747. 

In This Article


Study Population

The study population was drawn from the National Health and Nutrition Examination Survey (NHANES).[10] NHANES is a well-established representative survey of noninstitutionalized civilian residents in the United States conducted by the National Center for Health Statistics of the US Centers for Disease Control and Prevention. It consists of a standardized in-home interview followed by physical examination and blood and urine collection at a mobile examination center. All participants provide written informed consent. The protocol was approved by the National Center for Health Statistics Research Ethics Review Board. We included 21,169 nonpregnant adult participants from NHANES 1999–2008 who met our study criteria. From a denominator of 24,693 adults 20 years and older, we excluded 9 with missing dietary supplement use data, an additional 2,262 with missing kidney function data, and finally 1,253 more with estimated glomerular filtration rate (eGFR) <15 mL/min/1.73 m2. We excluded those with very low eGFR because our goal was to focus on individuals who would benefit most from identifying behaviors that may predispose to nephrotoxicity or CKD progression. After these exclusions, there were no pregnant participants remaining.


Serum creatinine was measured by the modified kinetic method of Jaffé using different analyzers in different survey years. Creatinine levels were calibrated as specified in NHANES documentation.[11,12] Random spot urine samples were obtained, urine albumin was measured using solid-phase fluorescence immunoassay, and urine creatinine was measured using the modified Jaffé kinetic method in the same laboratory[11] on frozen samples.


Participants who responded "yes" to the question "Have you used or taken any vitamins, minerals, or other dietary supplements in the past month?" were asked to provide bottles for the individual supplements they took. Each provided supplement was classified as either potentially harmful in the setting of CKD or "other." A supplement was considered potentially harmful if it contained at least one of 37 distinct herbs identified from a literature review and expert opinion by the Council on Renal Nutrition for the National Kidney Foundation (NKF).[13] We reviewed supplement ingredients using the variable dsdingr (ingredient name) in the Dietary Supplement Database-File 4 to determine the presence of any of the 37 potentially toxic herbs. For ingredients noted as "proprietary blends," we located the actual product label to identify ingredients. In the event the product label could not be found, the variable dsdbcnam (blend component name, also in File 4) was queried. Participants were classified as taking any potentially harmful supplement, taking only "other" supplements, or taking no supplements.

We defined CKD status as no CKD; at risk only, by the presence of strong CKD risk factors (including diabetes, hypertension, or cardiovascular disease); stages 1/2 CKD, as albuminuria only (urine albumin-creatinine ratio ≥30 mg/g) with eGFR ≥60 mL/min/1.73 m2; or stages 3/4 CKD, as eGFR of 15–59 mL/min/1.73 m2, regardless of albuminuria. eGFR was calculated using the CKD Epidemiology Collaboration (CKD-EPI) creatinine equation: GFR = 141 × min(SCr/κ,1)α × max(SCr/κ,1)−1.209 × 0.993age × 1.018 [if female] × 1.159 [if black], where SCr is serum creatinine (mg/dL), κ is 0.7 for females and 0.9 for males, α is −0.329 for females and −0.411 for males, min indicates the minimum of SCr/κ or 1, and max indicates the maximum of SCr/κ or 1.[14]

We defined diabetes by participant self-report. Hypertension was defined by self-report or an average of second and third blood pressure readings ≥140 mm Hg systolic or ≥90 mm Hg diastolic. Cardiovascular disease included self-report of coronary artery disease, stroke, heart attack, congestive heart failure, or angina.

We categorized age into 3 groups (20–44, 45–64, and ≥65 years) and used NHANES categories of self-reported race/ethnicity as non-Hispanic white, non-Hispanic black, Mexican American, or other. Educational attainment was categorized as more than high school, high school or high school equivalent, and less than high school. We categorized income using the US Census Bureau's poverty index ratio (the ratio of family income to federal poverty level, where ≤1.00 is considered below the poverty level) into 3 groups (poverty index ratio ≤1.00, >1-<3, or ≥3). We included self-reported arthritis and cancer as comorbid conditions ascertained uniformly in our study population that may prompt individuals to take dietary supplements as a means of prevention or treatment.[15]

Tobacco use (no/past vs ongoing) was defined by significant lifetime use of cigarettes (≥100), snuff (≥20 times), and/or chewing tobacco (≥20 times). Current alcohol use was categorized as none/moderate versus heavy (>7 drinks per week for women or >14 drinks per week for men). We defined health care utilization as the number of health care visits within the last 12 months as a continuous variable to examine the effect of encounters with health care providers on supplement use.

Statistical Analysis

We calculated the proportion of all reported supplements containing at least one NKF-identified herb and the proportion of potentially harmful supplements containing each specific NKF-identified herb. We used ordinal logistic regression to test whether supplement use varied by survey year. For US adults 20 years or older, we estimated the prevalent use of potentially harmful or only other supplements overall and by CKD status within groups defined by demographic characteristics, comorbid conditions, health-related behaviors, and health care visits. For the subpopulation of US adults who took any supplement, we used χ2 analysis to test whether potentially harmful supplement use varied by CKD status within each of these groups. We estimated the frequency and duration of any potentially harmful supplement use and used χ2 analysis to test whether these estimates were associated with CKD status. Finally, we used multivariable logistic regression to assess the presence, direction, strength, and independence of the association between taking a potentially harmful supplement and CKD status in those taking any supplement. We added covariates to the model sequentially to examine their incremental effects on the likelihood of taking a potentially harmful supplement. We performed sensitivity analyses with CKD defined by GFR estimated according to the isotope-dilution mass spectrometry−traceable 4-variable Modification of Diet in Renal Disease (MDRD) Study equation.[16] All analyses used recommended sampling weights[11] and were performed using the "svy" commands in Stata, version 12.0 (StataCorp LP) to account for the study design.