WHI 2006.

Study characteristics
Methods	RCT Women's Health Initiative (WHI) Summary risk of bias: low
Participants	Postmenopausal women aged 50 ‐ 79 with or without CVD at baseline (USA) CVD risk: low in those without CVD at baseline, high in those with CVD Control without CVD at baseline: randomised 29,294, analysed 29,294 Intervention without CVD at baseline: randomised 19,541, analysed 19,541 Control with CVD at baseline: randomised 1369, analysed 1369 Intervention with CVD at baseline: randomised 908, analysed 908 Mean years in trial: control 8.1, intervention 8.1 % male: 0 Age: mean (both with and without CVD at baseline) int 62.3 (SD 6.9), control 62.3 (SD 6.9) Ethnicity (women both with and without CVD at baseline): white 82%, black 11%, Asian or Pacific Islander 2%, unknown 1%, American Indian or Alaskan native < 1%. No statistically significant effects of the intervention on CHD events was seen for any ethnic subgroup. Statins use allowed? Yes % taking statins: 12% of women recruited were on lipid‐lowering medication (these were a mixture of participants with and without CVD at baseline).
Interventions	Reduced fat vs usual diet Control: diet‐related education materials Intervention: low‐fat diet (20%E from fat), reduce saturated fat to 7%E with increased fruit and vegetables Control methods: given copy of 'Dietary Guidelines for Americans' Intervention methods: 18 group sessions with trained and certified nutritionists in the 1st year, quarterly maintenance sessions thereafter, focusing on diet and behaviour modification Intervention delivered face‐to‐face in a group by nutritionists Intake data all relate to the full WHI cohort (not divided by whether participants have CVD at baseline or not) Total fat intake, %E (at 6 years): int 28.8 (SD 8.4), cont 37.0 (SD 7.3) (mean difference ‐8.20, 95% CI ‐8.34 to ‐8.06) significant reduction Saturated fat intake, %E (at 6 years): int 9.5 (SD3.2), cont 12.4 (SD3.1) (mean difference ‐2.90, 95% CI ‐2.96 to ‐2.84 for full WHI population) significant reduction PUFA intake, %E (at 6 years)§: int 6.3 (SD?), cont 7.6 (SD?) (mean difference ‐1.30, 95% CI ‐1.72 to ‐0.88 assuming missing SDs were 5) significant reduction PUFA n‐3 intake: not reported PUFA n‐6 intake: not reported MUFA intake, %E (at 6 years)§: int 11.1 (SD?), cont 14.3 (SD?) (mean difference ‐3.20, 95% CI ‐3.62 to ‐2.78 assuming unclear SDs were 5) significant reduction CHO intake, %E (at 6 years)§: int 53.9 (SD?), cont 46.3 (SD?) (mean difference 7.60, 95% CI 5.91 to 9.29 assuming SDs of 20) significant increase Protein intake, %E (at 6 years)§: int 17.7 (SD?), cont 17.0 (SD?) (mean difference 0.70, 95% CI 0.28 to 1.12 assuming SDs of 5) significant increase Trans fat intake, %E (at 6 years)§: int 1.8 (SD?), cont 2.4 (SD?) (mean difference unclear, no SDs assumed) Replacement for saturated fat: mainly carbohydrate, some protein (based on dietary achievement) Style: dietary advice Setting: community §Amongst the 881 intervention and 1373 control participants with blood samples at baseline, with or without CVD at baseline (Howard 2010)
Outcomes	Stated trial outcomes: breast cancer, mortality, other cancers, cardiovascular events, diabetes Data available on total mortality? yes* Cardiovascular mortality? yes Events available for combined cardiovascular events: CHD, stroke, heart failure, angina, peripheral vascular disease, revascularisation, pulmonary embolism, DVT Secondary outcomes: cancer deaths*, cancer diagnoses*, stroke, non‐fatal MI, diabetes diagnosis* Tertiary outcomes: weight, BMI, total, LDL and HDL cholesterol, TGs, systolic and diastolic BP (Lp(a) and HOMA reported as geometric means) * these are only available for the whole cohort, not split between low and high CVD risk groups
Notes	Study duration over 8 years Study aim was to reduce total fat to 20%E, reduce saturated fat to 7%E and increase fruit and vegetable intake (Patterson 2003), and saturated fat intake in the intervention group was significantly lower than in control SFA reduction aimed and achieved Total serum cholesterol, difference between intervention and control, mmol/L: ‐0.09 (95% CI ‐0.15 to ‐0.02), statistically significant reduction Trial dates: Recruitment was between 1993 and 1998 Funding: National Heart, Lung and Blood Institute of the National Institutes of Health Declarations of Interest of primary researchers: Declarations varied from paper to paper, but this is a typical one from Beresford 2006 "Dr Black has received research grants from Pfizer and AstraZeneca, was on the speakers bureaus for Pfizer, Novartis, Sanofi‐Aventis, Bristol‐Meyers Squibb, Searle, Pharmacia, and Boehringer and served as a consultant of on an advisory board for Myogen, Merck Sharp and Dohme, Novartis, Mylan‐Bertek, Pfizer, Bristol‐Meyers Squibb, and Sanofi‐Aventis. Dr Howard has served on the advisory boards of Merck, Schering Plough, and the Egg Nutrition Council, has received research support from Merck and Pfizer, and has consulted for General Millls. Dr Assaf is an employee of Pfizer. No other disclosures were reported."
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated permuted block algorithm stratified by clinical centre and age
Allocation concealment (selection bias)	Low risk	Allocations developed by the WHI Clinical Coordinating Center
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants aware of allocation
Blinding of outcome assessment (detection bias) CVD outcomes	Low risk	Trained clinic staff, who were responsible for anthropometric assessments and administration of FFQs, were blinded to treatment assignments to the extent practical. The dietary intervention staff did not conduct clinical assessments, and clinic staff were not permitted to participate in any intervention activities; participants were instructed not to discuss nutrition activities with clinic staff.
Blinding of outcome assessment (detection bias) All‐cause mortality	Low risk	Blinding not relevant for mortality assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT analysis
Selective reporting (reporting bias)	Low risk	Trials register 1999, study completion 2005, but outcomes not stated in trials register. However, outcomes were well published; trialists were asked for data.
Free of systematic difference in care?	High risk	Intervention participants received 18 group sessions with behavioural modification plus quarterly maintenance sessions thereafter; control groups received a leaflet. See control and intervention methods in the Interventions section of the table of Characteristics of included studies
Stated aim to reduce SFA	Low risk	Aim to reduce SFA stated
Achieved SFA reduction	Low risk	SFA reduction achieved
Achieved TC reduction	Low risk	Statistically significant TC fall
Other bias	Low risk	None noted