acarbose

• Molecular Formula: C25H43NO18
• Molecular Weight: 645.6 g/mol
• SMILES: CC1C(C(C(C(O1)OC2C(OC(C(C2O)O)OC3C(OC(C(C3O)O)O)CO)CO)O)O)NC4C=C(C(C(C4O)O)O)CO
• InChI: InChI=1S/C25H43NO18/c1-6-11(26-8-2-7(3-27)12(30)15(33)13(8)31)14(32)19(37)24(40-6)43-22-10(5-29)42-25(20(38)17(22)35)44-21-9(4-28)41-23(39)18(36)16(21)34/h2,6,8-39H,3-5H2,1H3/t6-,8+,9-,10-,11-,12-,13+,14+,15+,16-,17-,18-,19-,20-,21-,22-,23-,24-,25-/m1/s1
• InChIKey: XUFXOAAUWZOOIT-SXARVLRPSA-N
• Solubility: In water, 1.0X10+6 mg/L /miscible/ at 25 °C (est)

Standard α-glucosidase inhibitors lack pancreatic α-amylase inhibition and cardiovascular endpoint data, limiting their utility in metabolic research. Acarbose resolves this gap. - **Mechanistic differentiation**: Inhibits both intestinal α-glucosidases and pancreatic α-amylase; competitive and reversible binding. - **Quantitative outcomes**: STOP-NIDDM trial data show 49% RRR in major CV events (HR 0.51; P=0.03) and 91% RRR in MI (HR 0.09; P=0.02) for IGT patients. - **Supply**: Commercial-grade acarbose for oral dosing studies, monotherapy or combination research.

Acarbose Technical Baseline for Research & Procurement

Acarbose is a pseudo-tetrasaccharide α-glucosidase inhibitor (AGI) that competitively and reversibly binds to α-glucosidase hydrolase enzymes in the small intestinal brush border, delaying the hydrolysis of complex carbohydrates and thereby attenuating postprandial glycemic excursions [1]. Unlike sulfonylureas or insulin secretagogues, acarbose does not stimulate insulin release, conferring a low intrinsic hypoglycemia risk [1]. As the most extensively studied AGI globally, acarbose is approved for the management of type 2 diabetes mellitus (T2DM) as monotherapy or in combination with other antihyperglycemic agents, and it has demonstrated unique pleiotropic effects on cardiovascular risk parameters and body weight regulation that distinguish it from other agents within and beyond its class [2].

Mechanism

Competitive α-glucosidase and pancreatic α-amylase inhibitor for postprandial glucose modulation studies.

Model Context

Supports carbohydrate digestion and glycemic excursion models; non-insulin secretagogue mechanism.

Differentiation

Dual enzyme inhibition profile distinguishes it from miglitol and voglibose in class-comparison studies.

• [1] Derosa G, Maffioli P. α-Glucosidase inhibitors and their use in clinical practice. Arch Med Sci. 2012 Nov 9;8(5):899-906. doi: 10.5114/aoms.2012.31621. View Source
• [2] Van de Laar FA, Lucassen PL, Akkermans RP, Van de Lisdonk EH, Rutten GE, Van Weel C. α-Glucosidase inhibitors for patients with type 2 diabetes: results from a Cochrane systematic review and meta-analysis. Diabetes Care. 2005 Jan;28(1):154-63. doi: 10.2337/diacare.28.1.154. View Source

Acarbose Differentiation Drivers

Although acarbose, miglitol, and voglibose are all classified as α-glucosidase inhibitors, they exhibit clinically meaningful differences in enzyme inhibition profiles, glycemic efficacy, adverse effect burden, and pleiotropic outcomes that preclude simple therapeutic interchange [1]. Acarbose uniquely inhibits pancreatic α-amylase in addition to intestinal α-glucosidases, a mechanistic distinction from miglitol and voglibose that influences both postprandial glucose dynamics and gastrointestinal tolerability [1]. Furthermore, acarbose is the only AGI with robust, long-term cardiovascular outcome data from landmark trials such as STOP-NIDDM and ACE, providing a differentiated evidence base that informs guideline positioning and formulary decisions [2]. Consequently, procurement and research selection of acarbose versus alternative AGIs or other glucose-lowering agents must be guided by precise, quantifiable comparator data to align with specific therapeutic, metabolic, or investigational objectives.

Enzyme Profile

Acarbose inhibits pancreatic α-amylase; miglitol and voglibose do not. Postprandial glucose dynamics may differ significantly.

Endpoint Evidence

Only acarbose has long-term cardiovascular endpoint data from IGT studies (e.g., STOP-NIDDM); other AGIs lack comparable evidence.

Tolerability Context

GI tolerability profiles vary; acarbose shows higher GI event rates than voglibose in reported analyses, which may affect model endpoints.

• [1] Derosa G, Maffioli P. α-Glucosidase inhibitors and their use in clinical practice. Arch Med Sci. 2012 Nov 9;8(5):899-906. doi: 10.5114/aoms.2012.31621. View Source
• [2] Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M; STOP-NIDDM Trial Research Group. Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial. JAMA. 2003 Jul 23;290(4):486-94. doi: 10.1001/jama.290.4.486. View Source

Acarbose Quantitative Evidence for Decision-Making

Non-Inferior Glycemic Efficacy with Superior Weight Loss

In a 48-week, randomized, open-label, non-inferiority trial (MARCH study) comparing acarbose to metformin as initial therapy in newly diagnosed T2DM patients (baseline HbA1c ~7.5%), acarbose demonstrated statistically non-inferior HbA1c reduction at both 24 and 48 weeks. At week 48, acarbose reduced HbA1c by -1.11% compared to metformin's -1.12%, with a between-group difference of only 0.01% (95% CI -0.12 to 0.14; p=0.8999) [1]. Critically, acarbose achieved this equivalent glycemic control while producing significantly greater weight loss than metformin, a key differentiating factor for overweight or obese T2DM patient populations [1].

HbA1c vs. Metformin

Head-to-head

-1.11% (acarbose) vs. -1.12% (metformin); difference 0.01% (95% CI -0.12 to 0.14; p=0.8999)

Reported non-inferior glycemic endpoint context; weight reduction higher with acarbose.

48-week RCT, newly diagnosed T2DM model; baseline HbA1c ~7.5%.

Evidence Dimension	Change in HbA1c from baseline at 48 weeks
Target Compound Data	-1.11% (acarbose)
Comparator Or Baseline	-1.12% (metformin)
Quantified Difference	0.01% (95% CI -0.12 to 0.14); p=0.8999
Conditions	48-week, randomized, open-label, non-inferiority trial in newly diagnosed T2DM patients; baseline HbA1c ~7.5%; n=784 modified ITT

Why This Matters

This evidence establishes acarbose as a viable first-line alternative to metformin, particularly when metformin is contraindicated or not tolerated, while offering a weight-loss advantage that metformin does not provide, thereby supporting distinct clinical and formulary positioning.

• [1] Yang W, Liu J, Shan Z, Tian H, Zhou Z, Ji Q, Weng J, Jia W, Lu J, Liu J, Xu Y, Yang Z, Chen W. Acarbose compared with metformin as initial therapy in patients with newly diagnosed type 2 diabetes: an open-label, non-inferiority randomised trial. Lancet Diabetes Endocrinol. 2014 Jan;2(1):46-55. doi: 10.1016/S2213-8587(13)70021-4. View Source

Weight Loss Advantage over DPP-4 Inhibitors

A systematic review and network meta-analysis of randomized controlled trials, incorporating data from 11,877 participants, compared acarbose (at optimal dose of 300 mg/day) with dipeptidyl peptidase-4 inhibitors (DPP-4is) at their recommended doses [1]. The analysis demonstrated that acarbose and DPP-4is have similar glucose-lowering effects on HbA1c (mean difference 0.05%; 95% credible interval -0.25 to 0.33) and 2-h postprandial glucose (MD 0.96 mmol/L; 95% CrI -0.56 to 2.54) [1]. However, acarbose was superior to DPP-4is regarding weight loss, with a mean difference of -1.23 kg (95% CrI -2.08 to -0.33) [1].

Weight vs. DPP-4i

Head-to-head

Acarbose -1.23 kg (95% CrI -2.08 to -0.33) vs. DPP-4 inhibitors; HbA1c difference 0.05% (95% CrI -0.25 to 0.33)

Reported higher weight reduction; glycemic control similar. Supports obesity-model endpoint differentiation.

Network meta-analysis, 11,877 participants; acarbose 300 mg/day.

Evidence Dimension	Change in body weight from baseline
Target Compound Data	Acarbose group weight change (relative to DPP-4i)
Comparator Or Baseline	DPP-4 inhibitors (various agents at recommended doses)
Quantified Difference	-1.23 kg (95% CrI -2.08 to -0.33)
Conditions	Network meta-analysis of 48 monotherapy studies; 11,877 participants; acarbose 300 mg/day vs. DPP-4is at recommended doses

Why This Matters

For overweight or obese T2DM patients, the superior weight-loss effect of acarbose compared to DPP-4 inhibitors provides a clinically meaningful advantage, guiding therapeutic selection toward agents that address both glycemic control and adiposity.

• [1] Zhang F, Xu S, Tang L, Pan X, Tong N. Acarbose With Comparable Glucose-Lowering but Superior Weight-Loss Efficacy to Dipeptidyl Peptidase-4 Inhibitors: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials. Front Endocrinol (Lausanne). 2020 Jun 5;11:288. doi: 10.3389/fendo.2020.00288. View Source

Superior Evening Postprandial Glucose Control vs. Voglibose

In a 24-week, randomized, active-controlled, multicenter study comparing acarbose (300 mg/day, n=59) to voglibose (0.9 mg/day, n=62) as add-on therapy to basal insulin, both agents reduced HbA1c similarly by approximately 0.7% from baseline (acarbose: 8.43% ± 0.71% to 7.71% ± 0.93%; voglibose: 8.38% ± 0.73% to 7.68% ± 0.94%) [1]. However, a key point of differentiation emerged in postprandial glucose control: the 1-hour post-dinner glucose level decreased significantly more in the acarbose group (from 233.54 ± 69.38 mg/dL to 176.80 ± 46.63 mg/dL) compared with the voglibose group (from 224.18 ± 70.07 mg/dL to 193.01 ± 55.39 mg/dL) [1].

Postprandial vs. Voglibose

Head-to-head

1-h post-dinner glucose: acarbose 176.80 ± 46.63 mg/dL vs. voglibose 193.01 ± 55.39 mg/dL (25.57 mg/dL greater reduction)

Reported evening postprandial glucose control advantage in insulin-treated models.

24-week RCT, add-on to basal insulin; HbA1c reduction similar (~0.7%).

Evidence Dimension	1-hour post-dinner blood glucose level at week 24
Target Compound Data	Acarbose: 176.80 ± 46.63 mg/dL (change: -56.74 mg/dL)
Comparator Or Baseline	Voglibose: 193.01 ± 55.39 mg/dL (change: -31.17 mg/dL)
Quantified Difference	Acarbose achieved a 25.57 mg/dL greater reduction in post-dinner glucose
Conditions	24-week RCT; add-on to basal insulin ± metformin/sulfonylurea; acarbose 300 mg/day vs. voglibose 0.9 mg/day

Why This Matters

This demonstrates that while overall HbA1c efficacy is equivalent, acarbose provides superior control of evening postprandial hyperglycemia—a critical factor in reducing glycemic variability and potentially mitigating nocturnal glucose excursions in insulin-treated patients.

• [1] Lee MY, Choi DS, Lee MK, Lee HW, Park TS, Kim DM, Chung CH, Kim DK, Kim IJ, Jang HC, Park YS, Kwon HS, Lee SH, Shin HK. Comparison of Acarbose and Voglibose in Diabetes Patients Who Are Inadequately Controlled with Basal Insulin Treatment: Randomized, Parallel, Open-Label, Active-Controlled Study. J Korean Med Sci. 2014 Jan;29(1):90-7. doi: 10.3346/jkms.2014.29.1.90. View Source

Efficacy and GI Tolerability Compared to Miglitol

A Cochrane systematic review and meta-analysis of 41 randomized controlled trials (30 acarbose, 7 miglitol, 1 voglibose, 3 combined) compared the efficacy of AGIs as monotherapy [1]. Compared with placebo, acarbose reduced glycated hemoglobin (GHb) by -0.77%, while miglitol reduced GHb by -0.68% [1]. However, a single-blind, randomized, parallel-group comparative trial of 90 T2DM patients found that the adverse effect profile was better with voglibose (6.66%) than miglitol (16.66%) and acarbose (33.33%), indicating that acarbose has the highest rate of gastrointestinal adverse events among the three AGIs [2].

Tolerability vs. Miglitol

Cross-study

GHb reduction vs. placebo: acarbose -0.77%, miglitol -0.68%. GI AE rate: acarbose 33.33%, miglitol 16.66%, voglibose 6.66%.

Reported glycemic efficacy numerically higher; GI tolerability endpoints differ markedly among AGIs.

Cochrane meta-analysis (41 RCTs) and 90-patient comparative study.

Evidence Dimension	Reduction in glycated hemoglobin (GHb) vs. placebo; Adverse event incidence
Target Compound Data	Acarbose: -0.77% GHb reduction vs. placebo; 33.33% adverse event rate
Comparator Or Baseline	Miglitol: -0.68% GHb reduction vs. placebo; Voglibose: 6.66% adverse event rate
Quantified Difference	Acarbose GHb reduction: 0.09% greater than miglitol; Acarbose adverse event rate: 16.67% higher than miglitol, 26.67% higher than voglibose
Conditions	Cochrane meta-analysis of 41 RCTs; 90-patient single-blind RCT for adverse events

Why This Matters

While acarbose demonstrates numerically greater glycemic efficacy than miglitol in placebo-controlled analyses, its higher adverse event burden necessitates careful patient selection and counseling regarding gastrointestinal tolerability, which may influence procurement decisions based on target population characteristics.

• [1] Van de Laar FA, Lucassen PL, Akkermans RP, Van de Lisdonk EH, Rutten GE, Van Weel C. α-Glucosidase inhibitors for patients with type 2 diabetes: results from a Cochrane systematic review and meta-analysis. Diabetes Care. 2005 Jan;28(1):154-63. doi: 10.2337/diacare.28.1.154. View Source
• [2] Ismail TSES, Deshmukh SA. Comparative study of effect of alpha glucosidase inhibitors-miglitol, acarbose and voglibose on postprandial hyperglycemia and glycosylated hemoglobin in type-2 diabetes mellitus. International Journal of Pharma and Bio Sciences. 2012;3(3). View Source

Cardiovascular Risk Reduction in the STOP-NIDDM Trial

The STOP-NIDDM trial, a randomized, double-blind, placebo-controlled study in 1,368 patients with impaired glucose tolerance (IGT), evaluated the effect of acarbose (100 mg three times daily) on cardiovascular outcomes over a mean follow-up of 3.3 years [1]. Acarbose treatment was associated with a 49% relative risk reduction in the development of major cardiovascular events (hazard ratio [HR], 0.51; 95% CI, 0.28-0.95; P = 0.03) and a 2.5% absolute risk reduction [1]. Specifically, acarbose reduced the risk of myocardial infarction by 91% (HR, 0.09; 95% CI, 0.01-0.72; P = 0.02) [1]. Additionally, acarbose was associated with a 34% relative risk reduction in new-onset hypertension (HR, 0.66; 95% CI, 0.49-0.89; P = 0.006) and a 5.3% absolute risk reduction [1].

CV Risk in IGT

Class-level inference

49% RRR major CV events (HR 0.51; 95% CI 0.28-0.95; P=0.03); 91% RRR MI; 34% RRR new hypertension

Reported cardiovascular endpoint reduction in IGT model; unique among AGIs.

STOP-NIDDM; 1,368 patients, 3.3-year follow-up.

Evidence Dimension	Relative risk reduction in major cardiovascular events
Target Compound Data	Acarbose: 49% RRR (HR 0.51; 95% CI 0.28-0.95; P=0.03)
Comparator Or Baseline	Placebo (no other AGI has comparable CV outcome trial data)
Quantified Difference	49% RRR; 2.5% ARR; 91% RRR in MI; 34% RRR in hypertension
Conditions	STOP-NIDDM trial: 1,368 IGT patients; acarbose 100 mg TID; mean follow-up 3.3 years

Why This Matters

Acarbose is the only α-glucosidase inhibitor with robust, long-term cardiovascular outcome data demonstrating significant reductions in cardiovascular events and hypertension, providing a unique evidence-based rationale for its selection in patients with IGT or high cardiovascular risk, a distinction not shared by miglitol or voglibose.

• [1] Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M; STOP-NIDDM Trial Research Group. Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial. JAMA. 2003 Jul 23;290(4):486-94. doi: 10.1001/jama.290.4.486. View Source

Safety Profile Comparison with SGLT2 Inhibitors

A descriptive analysis of the WHO VigiAccess® database (1983-2024) encompassing 9,410 adverse drug reaction (ADR) reports characterized the safety profiles of α-glucosidase inhibitors [1]. Gastrointestinal disorders were the most frequently reported system organ class for all synthetic AGIs, with acarbose reporting the highest proportion at 43.10%, compared to voglibose at 39.02% and miglitol at 40.81% [1]. For acarbose specifically, diarrhoea (report rate 7.51%, ROR = 6.53, 95% CI 6.07-7.02) and abdominal distension (report rate 7.45%, ROR = 56.69, 95% CI 52.68-61.00) showed the highest reporting rates [1]. In contrast, a separate network meta-analysis of glucose-lowering medications in older people reported that SGLT2 inhibitors prevent hospitalization for heart failure (RR, 0.66; 95% CI, 0.57 to 0.77) and renal composite outcomes (RR, 0.69; 95% CI, 0.53 to 0.89) [2], benefits not established for acarbose.

Safety vs. SGLT2i

Cross-study

Acarbose GI ADR 43.10% (diarrhoea ROR 6.53, abdominal distension ROR 56.69). SGLT2i: HF hospitalization RR 0.66, renal composite RR 0.69.

Reported GI tolerability burden higher; cardio-renal protective endpoints not established for acarbose.

WHO VigiAccess® and network meta-analysis.

Evidence Dimension	Proportion of gastrointestinal disorder ADR reports; Cardio-renal protective effects
Target Compound Data	Acarbose: 43.10% GI disorder reports; diarrhoea ROR 6.53; abdominal distension ROR 56.69
Comparator Or Baseline	SGLT2 inhibitors: RR 0.66 for heart failure hospitalization; RR 0.69 for renal composite outcome
Quantified Difference	Acarbose has 43.10% GI ADR burden but lacks demonstrated cardio-renal protective effects of SGLT2 inhibitors
Conditions	WHO VigiAccess® real-world pharmacovigilance analysis; Network meta-analysis of glucose-lowering medications

Why This Matters

This evidence delineates the distinct safety and efficacy profiles of acarbose versus SGLT2 inhibitors, guiding procurement and therapeutic selection: acarbose is appropriate for postprandial glucose control with attention to GI tolerability, while SGLT2 inhibitors are preferred when cardio-renal protection is a primary objective.

• [1] Du C, et al. Characteristic analysis of adverse reactions to natural and synthetic alpha-glucosidase inhibitors: a descriptive analysis using World Health Organisation-VigiAccess. J Chromatogr B Analyt Technol Biomed Life Sci. 2025 Dec 15:1267:124800. doi: 10.1016/j.jchromb.2025.124800. View Source
• [2] Evaluation of glucose-lowering medications in older people: a comprehensive systematic review and network meta-analysis of randomized controlled trials. 2024. View Source

Acarbose Research and Application Scenarios

First-Line Monotherapy When Metformin Is Contraindicated

Based on the MARCH trial demonstrating non-inferior HbA1c reduction (-1.11% vs. -1.12% at 48 weeks; difference 0.01%, 95% CI -0.12 to 0.14) and superior weight loss compared to metformin [1], acarbose should be prioritized as initial monotherapy in patients with newly diagnosed T2DM who cannot take metformin due to gastrointestinal intolerance, renal impairment, or other contraindications. This evidence supports acarbose's inclusion in first-line treatment algorithms and formulary listings as a metformin alternative.

Add-On to Basal Insulin for Postprandial Control

In T2DM patients inadequately controlled on basal insulin (with or without metformin/sulfonylurea), acarbose 300 mg/day provides a clinically meaningful advantage over voglibose 0.9 mg/day in reducing 1-hour post-dinner glucose levels (176.80 ± 46.63 mg/dL vs. 193.01 ± 55.39 mg/dL; acarbose achieved a 25.57 mg/dL greater reduction) [2]. This superior evening postprandial control makes acarbose the preferred AGI for add-on therapy in insulin-treated patients, particularly those with significant nocturnal hyperglycemia or high glycemic variability.

Weight-Loss Preferred Therapy for Overweight/Obese T2DM

Network meta-analysis data demonstrate that acarbose, at its optimal dose of 300 mg/day, provides a significant weight-loss advantage of -1.23 kg (95% CrI -2.08 to -0.33) compared to DPP-4 inhibitors, while achieving comparable glycemic control (HbA1c difference 0.05%; 95% CrI -0.25 to 0.33) [3]. Consequently, acarbose is a superior therapeutic choice for overweight or obese T2DM patients when weight gain associated with sulfonylureas, thiazolidinediones, or insulin is undesirable, or when the weight-neutral profile of DPP-4 inhibitors is insufficient.

Cardiovascular Risk Reduction in Impaired Glucose Tolerance

The STOP-NIDDM trial provides unique, class-differentiating evidence that acarbose (100 mg TID) reduces the relative risk of major cardiovascular events by 49% (HR 0.51; 95% CI 0.28-0.95; P=0.03), myocardial infarction by 91% (HR 0.09; 95% CI 0.01-0.72; P=0.02), and new-onset hypertension by 34% (HR 0.66; 95% CI 0.49-0.89; P=0.006) in patients with IGT [4]. Acarbose should be strongly considered for procurement and clinical use in prediabetic populations or patients with IGT and elevated cardiovascular risk, as no other AGI has demonstrated comparable cardiovascular outcome benefits.

Application

Selection Property

Validation Focus

α-Glucosidase inhibition models (metformin comparison)

Non-insulin secretagogue; weight endpoint differentiation

HbA1c and body weight change endpoints in T2DM models

Postprandial glucose excursion studies with basal insulin

Evening postprandial glucose control advantage over voglibose

1-h post-dinner glucose and glycemic variability endpoints

Obesity/diabetes weight-management research

Weight-loss advantage vs. DPP-4 inhibitors; comparable glycemic control

Body weight change and HbA1c in overweight/obese models

Impaired glucose tolerance and cardiovascular endpoint studies

Unique CV outcome data (STOP-NIDDM); class-differentiating evidence

Major CV events, MI, and hypertension incidence in IGT models

• [1] Yang W, Liu J, Shan Z, Tian H, Zhou Z, Ji Q, Weng J, Jia W, Lu J, Liu J, Xu Y, Yang Z, Chen W. Acarbose compared with metformin as initial therapy in patients with newly diagnosed type 2 diabetes: an open-label, non-inferiority randomised trial. Lancet Diabetes Endocrinol. 2014 Jan;2(1):46-55. doi: 10.1016/S2213-8587(13)70021-4. View Source
• [2] Lee MY, Choi DS, Lee MK, Lee HW, Park TS, Kim DM, Chung CH, Kim DK, Kim IJ, Jang HC, Park YS, Kwon HS, Lee SH, Shin HK. Comparison of Acarbose and Voglibose in Diabetes Patients Who Are Inadequately Controlled with Basal Insulin Treatment: Randomized, Parallel, Open-Label, Active-Controlled Study. J Korean Med Sci. 2014 Jan;29(1):90-7. doi: 10.3346/jkms.2014.29.1.90. View Source
• [3] Zhang F, Xu S, Tang L, Pan X, Tong N. Acarbose With Comparable Glucose-Lowering but Superior Weight-Loss Efficacy to Dipeptidyl Peptidase-4 Inhibitors: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials. Front Endocrinol (Lausanne). 2020 Jun 5;11:288. doi: 10.3389/fendo.2020.00288. View Source
• [4] Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M; STOP-NIDDM Trial Research Group. Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial. JAMA. 2003 Jul 23;290(4):486-94. doi: 10.1001/jama.290.4.486. View Source