Impact of Abdominal Obesity on Incidence of Adverse Metabolic Effects Associated With Antihypertensive Medications
We assessed adverse metabolic effects of atenolol and hydrochlorothiazide among hypertensive patients with and without abdominal obesity using data from a randomized, open-label study of hypertensive patients without evidence of cardiovascular disease or diabetes mellitus. Intervention included randomization to 25 mg of hydrochlorothiazide or 100 mg of atenolol monotherapy followed by their combination. Fasting glucose, insulin, triglycerides, high-density lipoprotein cholesterol, and uric acid levels were measured at baseline and after monotherapy and combination therapy. Outcomes included new occurrence of and predictors for new cases of glucose ≥100 mg/dL (impaired fasting glucose), triglyceride ≥150 mg/dL, high-density lipoprotein ≤40 mg/dL for men or ≤50 mg/dL for women, or new-onset diabetes mellitus according to the presence or absence of abdominal obesity. Abdominal obesity was present in 167 (58%) of 395 patients. Regardless of strategy, in those with abdominal obesity, 20% had impaired fasting glucose at baseline compared with 40% at the end of study (P<0.0001). Proportion with triglycerides ≥150 mg/dL increased from 33% at baseline to 46% at the end of study (P<0.01). New-onset diabetes mellitus occurred in 13 patients (6%) with and in 4 patients (2%) without abdominal obesity. Baseline levels of glucose, triglyceride, and high-density lipoprotein predicted adverse outcomes, and predictors for new-onset diabetes mellitus after monotherapy in those with abdominal obesity included hydrochlorothiazide strategy (odds ratio: 46.91 [95% CI: 2.55 to 862.40]), female sex (odds ratio: 31.37 [95% CI: 2.10 to 468.99]), and uric acid (odds ratio: 3.19 [95% CI: 1.35 to 7.52]). Development of adverse metabolic effect, including new-onset diabetes mellitus associated with short-term exposure to hydrochlorothiazide and atenolol was more common in those with abdominal obesity.
- abdominal obesity
- metabolic syndrome
- new-onset diabetes mellitus
It is estimated that >72 million US adults are obese, affecting 33% of men and 35% of women.1 This epidemic is associated with increased mortality,2 primarily via metabolic and cardiovascular (CV) complications. Abdominal fat accumulation increases CV disease risk independent of overall adiposity.3 The presence of abdominal obesity provides additional predictive information for the development of CV morbidity and mortality compared with increased body mass index alone.4 Hypertension requiring treatment is highly prevalent in those with obesity and abdominal obesity, regardless of sex or ethnicity, and is poorly controlled.5–7 Some antihypertensives are associated with adverse metabolic effects (AMEs), including hyperglycemia, hypertriglyceridemia, and hyperuricemia.8 The predisposing factors for these AMEs are unknown, but preexisting abdominal obesity may contribute. Although there are many studies in older, high-risk populations indicating that thiazide diuretics and β-blockers increase the incidence of new-onset diabetes mellitus compared with other antihypertensive regimens,9 our knowledge is incomplete with regard to the development of AMEs in those with a lower CV risk profile and the contribution of abdominal obesity to this risk.
Accordingly, we investigated early AMEs and clinical characteristics predictive of early AMEs among those with and without abdominal obesity in the Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR) Study. We hypothesized that antihypertensive-induced AMEs develop preferentially in those with abdominal obesity.
This study includes analysis of the AME data from a contemporary population of low-risk hypertension subjects enrolled in the PEAR Study. PEAR is an ongoing, prospective, randomized, parallel group titration study undertaken primarily to evaluate the pharmacogenomic determinants of the antihypertensive and adverse metabolic responses to hydrochlorothiazide (HCTZ) and atenolol in hypertensive patients without a history of heart disease or diabetes mellitus. Details regarding study design and enrollment criteria have been published previously.10
Men and women with mild-to-moderate essential hypertension within age 17 to 65 years are being recruited from primary care populations at the University of Florida, Emory University, and the Mayo Clinic. Enrolled subjects had newly diagnosed, untreated or treated hypertension. Those in whom hypertension was treated had antihypertensives discontinued with a minimum washout of 18 days.
The study has been conducted in accordance with the provisions of the Declaration of Helsinki and was approved by the institutional review boards at each institution. Written, informed consent was provided by each patient before participation. Those meeting the blood pressure (BP) inclusion criteria were randomly assigned at each study site to receive HCTZ or atenolol. HCTZ and atenolol were initiated at 12.5 mg and 50.0 mg daily, respectively, and titrated to 25.0 mg and 100.0 mg daily on the basis of BP. After ≥6 weeks of 25 mg of HCTZ or 100 mg of atenolol or maximum tolerated dose, response to monotherapy was assessed. Then, those with BP remaining >120/70 mm Hg had the second drug added and titrated to the maximum dose. Response to combination therapy was assessed after 6 weeks on both drugs at the maximum tolerated dose.
Height and weight were measured to the nearest 0.1 cm and 0.1 kg, respectively. Waist circumference was measured to the nearest 0.5 cm by placing a tape measure snugly around the abdomen at the level of the umbilicus just above the uppermost lateral border of the right iliac crest and at the normal minimal respiration with the patient in standing position and with his/her hands by the side. Patients were categorized as having abdominal obesity if their waist measured ≥35 in for women or ≥40 in for men.11
At baseline, fasting blood samples were collected for glucose, insulin, potassium, magnesium, uric acid, and a lipid profile. Blood samples were collected again, after a 12-hour fast, at the completion of the monotherapy and combination therapy phases of the study. Insulin sensitivity using homeostasis model assessment-insulin resistance (HOMA-IR) was calculated at each time point according to a validated formula.12 Estimated glomerular filtration rate was calculated using the Modification of Diet in Renal Disease equation.13
Serum potassium, magnesium, glucose, triglyceride, high-density lipoprotein (HDL) cholesterol, and uric acid concentrations were measured on an Hitachi 911 Chemistry Analyzer (Roche Diagnostics) at the central laboratory at the Mayo Clinic. Potassium and magnesium concentrations were determined by an ion selective electrode, and glucose, triglycerides, HDL cholesterol, and uric acid concentrations were determined spectrophotometrically by automated enzymatic assays. Low-density lipoprotein (LDL) cholesterol was computed. Plasma insulin was measured using the Access Ultrasensitive Insulin immunoassay system (Beckman Instruments). All of the samples were tested in duplicate, and data reported are the means of the duplicate samples.
Within abdominal obesity categories, we compared the percentage of patients with glucose <100 and ≥100 mg/dL, triglycerides <150 and ≥150 mg/dL, and HDL <40 for male and <50 for female (hereafter termed “low HDL”) or ≥40 for male and ≥50 for female at baseline and after monotherapy and combination therapy in the 2 treatment strategies, respectively. These categories were picked on the basis of their inclusion in the definition of metabolic syndrome (MetSyn).11 In addition, we compared the percentage of patients with incident diabetes mellitus (fasting glucose: ≥126 mg/dL) after monotherapy and combination therapy within waist groups and treatment strategies. Lastly, we assessed the clinical characteristics that were predictive for development of these outcomes after monotherapy in those in whom the outcomes were absent at baseline.
Nonparametric tests were used when data were nonnormally distributed. Baseline characteristics were compared between patients with and without abdominal obesity using χ2 and Wilcoxon rank-sum tests. McNemar test was used to compare counts of dichotomous traits of MetSyn criteria in the abdominal obesity groups before and after treatment. Continuous variables repeatedly measured over time were compared using the Friedman test.
We used a multivariable logistic regression analysis to determine factors associated with having a glucose level ≥100 mg/dL after monotherapy among those with a level <100 mg/dL at baseline. Patient baseline clinical characteristics and laboratory variables were entered in a backward logistic regression model. Variables with a P value of <0.05 were retained in the model. The same logistic regression analysis was performed for triglycerides ≥150 mg/dL, low HDL, and new-onset diabetes mellitus. The baseline variables entered in the model were as follows: age, sex, race, treatment strategy, waist status, pulse, home BP, glucose, insulin, triglyceride, uric acid, HDL, LDL, potassium, estimated glomerular filtration rate, and current smoking status. Because potassium during treatment with HCTZ has been associated with dysglycemia, on-treatment change in potassium was also entered in the model.
All of the statistical analyses were performed using SAS 9.1 (SAS Institute). P value <0.05 was considered significant for a single test. For multiple comparisons we used the Bonferroni adjustment, and the critical P value was 0.017 after adjustment.
From a total of 418 patients who had completed the PEAR Study at the time of this analysis, we excluded those with missing waist circumference measurements (n=13) and those with nonfasting laboratory measurements (n=10). Baseline characteristics of the 395 patients included in this analysis are summarized in Table 1, according to the presence or absence of abdominal obesity. Among these middle-aged hypertensive patients, 58% had abdominal obesity. At baseline, compared with those without abdominal obesity, those with abdominal obesity had significantly higher fasting glucose, insulin, HOMA-IR, and triglycerides and lower HDL. They were also significantly more likely to have MetSyn compared with those without abdominal obesity (odds ratio: 10.1; 95% CI: 6.13 to 16.57). Less than one quarter of all patients had impaired fasting glucose. Mean duration of antihypertensive washout was 29±16 days, and there was no difference in any of the metabolic parameters of interest at baseline comparing those who had received previous treatment with a β-blocker and/or thiazides diuretic and those who had received other classes of antihypertensive agents. Mean duration of monotherapy and combination therapy in each of the treatment strategies was 9 weeks, resulting in a mean total follow-up of 18 weeks for all of the patients. In the atenolol strategy, 86% of patients were taking 100 mg of atenolol, and 78% of patients were taking 25 mg of HCTZ at the end of combination therapy. In the HCTZ strategy, 98% of patients were taking 25 mg of HCTZ, and 68% of patients were taking 100 mg of atenolol at the end of combination therapy. There were no significant differences comparing the percentage on maximum dose between those with and without abdominal obesity.
Table 2 summarizes mean biochemical parameters, BP parameters, and weight according to treatment strategy in those without and with abdominal obesity. In those randomized to the atenolol strategy (top half of Table 2), glucose, triglycerides, and uric acid were significantly increased during follow-up irrespective of the presence or absence of abdominal obesity at baseline. In those randomized to the HCTZ strategy (bottom half of Table 2), uric acid was significantly increased during follow-up irrespective of the presence or absence of abdominal obesity at baseline. However, in the HCTZ strategy, those with abdominal obesity at baseline also exhibited significantly increased glucose, insulin, HOMA-IR, and triglyceride levels during follow-up. Insulin and HOMA-IR were not significantly affected by monotherapy or combination therapy in either treatment strategy in the group without abdominal obesity, nor were they affected in those with abdominal obesity treatment with the atenolol strategy.
Regardless of abdominal obesity status, potassium was significantly decreased in both treatment strategies, related to the initiation or add-on of HCTZ, although mean potassium did not fall below 4.0 meq/L. There was not a difference in the potassium decrease in those with and without abdominal obesity after HCTZ monotherapy (P=0.96) or add-on therapy (P=0.08). Systolic and diastolic BPs and pulse were significantly and similarly decreased by both treatment strategies, whereas weight was not significantly affected by either strategy.
In subjects without abdominal obesity (Figure, left), the majority had glucose <100 mg/dL, high HDL, and triglycerides <150 mg/dL at baseline. Although neither treatment strategy significantly altered the proportion of patients with glucose levels ≥100 mg/dL, drug treatment significantly increased the proportion with low HDL and triglyceride levels ≥150 mg/dL.
In subjects with abdominal obesity (Figure, right), both treatment strategies resulted in significantly increased proportions with glucose levels ≥100 mg/dL, approximately doubling after combination therapy. The proportion with triglyceride levels ≥150 mg/dL increased by 30% to 50% after combination therapy. Order of initiation (atenolol versus HCTZ) did not impact the glucose or triglyceride outcomes within either abdominal obesity group.
A total of 17 patients developed new-onset diabetes mellitus during the follow-up period. In the abdominally obese patients, new-onset diabetes mellitus occurred in 13 (6%) of 224 patients, 11 patients in the HCTZ strategy and 2 patients in the atenolol strategy (P=0.0189). In those without abdominal obesity, 4 (2%) of 164 patients developed new-onset diabetes mellitus, 2 patients in the HCTZ strategy and 2 patients in the atenolol strategy (P=1.00). Of the 17 patients who developed new-onset diabetes mellitus, 13 (76%) met the criteria for MetSyn at baseline, and mean baseline fasting glucose was 101±11.8 mg/dL.
Table 3 summarizes baseline predictors of developing glucose ≥100 mg/dL, triglycerides ≥150 mg/dL, low HDL, or new-onset diabetes mellitus after exposure to atenolol or HCTZ monotherapy in those in whom the condition was not present at baseline.
Multiple randomized, controlled studies in patients with or at increased risk for CV disease have associated long-term use of thiazide diuretics and/or β-blockers with new-onset diabetes mellitus compared with other antihypertensive medications.9 Our findings in a contemporary sample of hypertensives without CV disease or diabetes mellitus demonstrate that the AMEs of atenolol and HCTZ begin within 9 weeks of initiation and are most pronounced in patients with abdominal obesity with longer duration of exposure to HCTZ. Importantly, in patients with abdominal obesity, we observed a significant increase in the proportion with adverse metabolic phenotypes, including impaired fasting glucose, increased triglycerides, and low HDL, which are known to increase the risk of developing diabetes mellitus and long-term CV adverse outcomes.
In PEAR participants with hypertension and abdominal obesity, we observed significantly higher baseline values for glucose, insulin, and triglycerides; lower HDL values; and a prevalence of MetSyn compared with those without abdominal obesity, indicating a population at increased risk for developing diabetes mellitus.14,15 After 9 to 18 weeks of exposure to 25 mg of HCTZ either alone or in combination with atenolol, patients with abdominal obesity had a significant increase in glucose and a significant increase in the proportion with glucose ≥100 mg/dL or impaired fasting glucose. Although the notion that treatment with thiazide diuretics increases glucose and worsens glucose tolerance is not new,16 it is considered by some to be an “innocent” adverse effect that takes many years to develop and may not be associated with adverse CV outcomes.17,18 Others have shown a significant association between antihypertensive associated incident diabetes mellitus and CV outcomes, including stroke, myocardial infarction, and death.19,20 Importantly, in patients who have both abdominal obesity and hypertension, a glucose level ≥100 mg/dL results in a diagnosis of MetSyn,11 resulting in a 3- to 5-fold increase in the risk for diabetes mellitus.15 The risk for long-term CV morbidity and mortality is increased 2- to 4-fold and increases proportionately with an increasing number of MetSyn components,21 suggesting the importance of preventing or aggressively treating each of the MetSyn criteria. Among our patients, who were exposed to short-term antihypertensive therapy, the majority who developed diabetes mellitus also had MetSyn and impaired fasting glucose at baseline.
The hyperglycemic effects of thiazide diuretics have been associated with diuretic-induced hypokalemia22,23; however, we did not confirm this association in PEAR participants.24 In a cross-sectional study, serum potassium was found to be independently associated with plasma glucose abnormalities in abdominally obese hypertensives treated with thiazide diuretics.25 The data presented here did not confirm this finding, because neither baseline potassium nor change in potassium during treatment predicted the development of either impaired fasting glucose or new-onset diabetes mellitus.
Visceral fat may influence metabolism and promote insulin resistance via the liver through the portal circulation, and, recently, treatment with HCTZ for 12 weeks was associated with liver fat accumulation and fat redistribution from the subcutaneous to the visceral space in patients with abdominal obesity.26 This fat redistribution was associated with aggravated insulin resistance and low-grade inflammation. This mechanism may partially explain the insulin resistance that we observed in our abdominally obese subjects assigned to the HCTZ strategy but not in those assigned to the atenolol strategy.
Although the number of cases of new-onset diabetes mellitus that we report here is small, when extrapolated to the large population of abdominally obese individuals with hypertension who are likely to be exposed to HCTZ and/or atenolol, the impact is significant. Baseline predictors of new-onset diabetes mellitus during treatment with antihypertensives in long-term observational and prospective treatment trials consistently include age, Hispanic ethnicity, BMI, waist circumference, glucose (fasting and nonfasting), HDL, female sex, and treatment with a thiazide diuretic and/or β-blocker.27 In the PEAR Study, baseline predictors of new-onset diabetes mellitus after the short-term use of antihypertensives are remarkably consistent with this list. Importantly, we add an additional predictive factor, uric acid, which was associated with a >3-fold excess risk of new-onset diabetes mellitus. Baseline uric acid was also a significant predictor of elevated triglycerides posttreatment in the PEAR Study. In addition, during follow-up, we observed a significant increase in uric acid without regard to abdominal obesity status, particularly associated with the addition of HCTZ. This uric acid elevation may have serious long-term consequences in terms of increased risk of CV events, as well as offsetting benefits of BP lowering.28
There are some limitations of this study worthy of mention. We only measured fasting glucose and not glucose tolerance on the basis of the outcome of an oral glucose tolerance test. As a result, we likely underdetected new-onset diabetes mellitus cases that might have been diagnosed on the basis of impaired glucose tolerance, which has been associated with short-term use of HCTZ.29 This may also have contributed to the wide CIs observed in Table 3 and, thus, should be interpreted with some caution. We are unable to adequately assess the impact of dose of HCTZ or atenolol on development of AMEs, because the majority of all of the patients in both treatment strategies required the highest protocol-specified dose (25 mg of HCTZ and 100 mg of atenolol). It has been suggested that the AMEs associated with HCTZ are dose related and are minimized or prevented with lower doses than those used in this study; however, PEAR data suggest that few patients achieve their BP goal with 12.5 mg of HCTZ daily.30 In addition, lower doses of thiazides have been associated with neutral or negative long-term morbidity and mortality outcomes and, thus, may not be optimal long-term treatments. Similarly, because of the lack of a control (untreated) group, we are unable to detect temporal changes that might have occurred even without treatment. Lastly, we only have a single fasting blood glucose level of ≥126 mg/dL in patients classified as having new-onset diabetes mellitus.
In conclusion, we observed AMEs, including new-onset diabetes mellitus, in a contemporary sample of hypertensive study participants after only short-term treatment with HCTZ and atenolol. These adverse effects were more prominent in those with abdominal obesity, despite normal potassium levels. Our findings reinforce guidelines and recommendations that indicate thiazide diuretics and β-blockers be used with caution in patients with or at risk for developing impaired fasting glucose or MetSyn to prevent the development of diabetes mellitus and the associated long-term adverse consequences.31,32
New data from the Centers for Disease Control and Prevention indicate that, in the United States, the prevalence of obesity in whites ranges from 9% to 30%. Compared with whites, blacks had 51% higher and Hispanics had 21% higher prevalence of obesity.33 Given that ≈50% of obese individuals also have hypertension,5–7 understanding the impact the AMEs on thiazide diuretics and β-blockers in this growing population is important. Our data add significantly to the growing body of literature related to the metabolic effects of antihypertensives by demonstrating that the AMEs associated with both thiazide diuretics and β-blockers occur very early in therapy in those with abdominal obesity, and, in our population of hypertensives with abdominal obesity, HCTZ was strongly associated with new-onset diabetes mellitus after just 9 to 18 weeks of exposure. Because treatment for hypertension usually requires lifelong therapy, and the likelihood of developing AMEs increases with increasing exposure duration, particularly to HCTZ, prescribers should consider not only the BP-lowering properties, but also the AMEs, which could include diabetes mellitus and it is associated CV outcomes, when developing a hypertension treatment regimen for patients with abdominal obesity and/or MetSyn.
We acknowledge and thank the study participants, support staff, and study physicians: Drs George Baramidze, Carmen Bray, Kendall Campbell, Robert Whit Curry, Frederic Rabari-Oskoui, Dan Rubin, and Siegfried Schmidt, for their valuable contributions.
Sources of Funding
This work is supported by a grant from the National Institutes of Health, grant U01 GM074492, funded as part of the Pharmacogenetics Research Network. In addition, this work is supported by the following grants from the National Institutes of Health, National Center for Research Resources: M01 RR00082 to the University of Florida, UL1 RR025008 and M01 RR00039 to Emory University, and UL1 RR024150 to Mayo Clinic, as well as the National Heart Lung and Blood Institute grants K23HL08655 (to R.M.C.-D.) and K23HL091120 (to A.L.B.).
R.M.C.-D. and S.W. had full access to all of the data used in this study and take full responsibility for the integrity of the data and the accuracy of the data analysis.
I.Z. is an employee of the US Food and Drug Administration, and the views expressed in this article do not necessarily reflect the official policy of the Food and Drug Administration. No official endorsement by the Food and Drug Administration is intended or should be inferred.
PEAR has been registered at www.clinicaltrials.gov (identifier NCT00246519).
- Received July 22, 2009.
- Revision received August 9, 2009.
- Accepted October 19, 2009.
Colin Bell A, Adair LS, Popkin BM. Ethnic differences in the association between body mass index and hypertension. Am J Epidemiol. 2002; 155: 346–353.
Bramlage P, Pittrow D, Wittchen HU, Kirch W, Boehler S, Lehnert H, Hoefler M, Unger T, Sharma AM. Hypertension in overweight and obese primary care patients is highly prevalent and poorly controlled. Am J Hypertens. 2004; 17: 904–910.
Lithell HO. Effect of antihypertensive drugs on insulin, glucose, and lipid metabolism. Diabetes Care. 1991; 14: 203–209.
Johnson JA, Boerwinkle E, Zineh I, Chapman AB, Bailey K, Cooper-Dehoff RM, Gums J, Curry RW, Gong Y, Beitelshees AL, Schwartz G, Turner ST. Pharmacogenomics of antihypertensive drugs: rationale and design of the Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR) Study. Am Heart J. 2009; 157: 442–449.
Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002; 106: 3143–3421.
Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004; 27: 1487–1495.
Ford ES, Li C, Sattar N. Metabolic syndrome and incident diabetes: current state of the evidence. Diabetes Care. 2008; 31: 1898–1904.
Alderman MH. New onset diabetes during antihypertensive therapy. Am J Hypertens. 2008; 21: 493–499.
Barzilay JI, Davis BR, Cutler JA, Pressel SL, Whelton PK, Basile J, Margolis KL, Ong ST, Sadler LS, Summerson J. Fasting glucose levels and incident diabetes mellitus in older nondiabetic adults randomized to receive 3 different classes of antihypertensive treatment: a report from the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Arch Intern Med. 2006; 166: 2191–2201.
Dunder K, Lind L, Zethelius B, Berglund L, Lithell H. Increase in blood glucose concentration during antihypertensive treatment as a predictor of myocardial infarction: population based cohort study. BMJ. 2003; 326: 681–685.
Butler J, Rodondi N, Zhu Y, Figaro K, Fazio S, Vaughan DE, Satterfield S, Newman AB, Goodpaster B, Bauer DC, Holvoet P, Harris TB, de Rekeneire N, Rubin S, Ding J, Kritchevsky SB. Metabolic syndrome and the risk of cardiovascular disease in older adults. J Am Coll Cardiol. 2006; 47: 1595–1602.
Zillich AJ, Garg J, Basu S, Bakris GL, Carter BL. Thiazide diuretics, potassium, and the development of diabetes: a quantitative review. Hypertension. 2006; 48: 219–224.
Agarwal R. Hypertension, hypokalemia, and thiazide-induced diabetes: a 3-way connection. Hypertension. 2008; 52: 1012–1013.
Smith SM, Anderson SD, Wen S, Gong Y, Turner ST, Cooper-Dehoff RM, Schwartz GL, Bailey K, Chapman AB, Hall KL, Feng H, Boerwinkle E, Johnson JA, Gums JG. Thiazide-Induced hyperglycemia is not related to hypokalemia: results from the Pharmacogenomic Evaluation of Antihypertensive Response (PEAR) Study. Pharmacotherapy. 2009; 29: 1157–1165.
Eriksson JW, Jansson PA, Carlberg B, Hagg A, Kurland L, Svensson MK, Ahlstrom H, Strom C, Lonn L, Ojbrandt K, Johansson L, Lind L. Hydrochlorothiazide, but not Candesartan, aggravates insulin resistance and causes visceral and hepatic fat accumulation: the Mechanisms for the Diabetes Preventing Effect of Candesartan (MEDICA) Study. Hypertension. 2008; 52: 1030–1037.
Bakris G, Stockert J, Molitch M, Zhou Q, Champion A, Bacher P, Sowers J. Risk factor assessment for new onset diabetes: literature review. Diabetes Obes Metab. 2009: 177–187.
Johnson JA, Gong Y, Bailey KR, Cooper-Dehoff RM, Chapman AB, Turner ST, Schwartz GL, Campbell K, Schmidt S, Beitelshees AL, Boerwinkle E, Gums JG. Hydrochlorothiazide and atenolol combination antihypertensive therapy: effects of drug initiation order. Clin Pharmacol Ther. 2009; 86: 533–539.
Torre JJ, Bloomgarden ZT, Dickey RA, Hogan MJ, Janick JJ, Jyothinagaram SG, Siragy HM; for the AACE Hypertension Task Force. American Association of Clinical Endocrinologists medical guidelines for clinical practice for the diagnosis and treatment of hypertension. Endocr Pract. 2006; 12: 193–222.