To determine if blood pressure (BP) response to fiber intake is related to type of fiber and subject characteristics such as age, sex, and hypertensive status.

• Human subjects
• English language literature

• Lack of appropriate BP data
• Intentional co-intervention from which the effect of fiber supplementation could not be separated
• Lack of concurrent placebo group

Recruitment

MEDLINE search with the following logic: Fiber or fibre AND blood pressure or hypertension AND trial or intervention or random or study.  Search limited to studies in human subjects and English-language literature. Original publication only.

Design:  Meta-Analysis

Blinding Used (if appllicable):  not applicable

Intervention (if applicable):

Data abstraction:

• Data collected on:
  • Differences in systolic (SBP) and diastolic (DBP) blood pressure
    • For 2 trials this was estimated from a graph in the article
  • Trial design
  • Duration of the trial
  • Type of fiber and dose
  • Whether BP was a primary or secondary outcome
  • Sample size
  • Characteristics of the sample population
  • Mean age
  • Sex
  • Baseline BP
  • Baseline body weight
  • Baseline body mass index (BMI)
  • Change in body weight during intervention
  • Use of anti-hypertensive drugs
  • Baseline fiber intake
  • Use of anti-hypertensive drugs
  • Baseline (habitual) fiber intake
•  BP measurements abstracted by priority:
• • Sitting position
  • Supine if sitting not available, then
  • Awake ambulatory if previous listed not available.
• Populations were considered hypertensive if >50% of subject were on anti-hypertensive medications or baseline blood pressure > 140/90 mmHg

Statistical Analysis

• Internal validity guaranteed by random assignment
• Quality of studies were quantified by scoring blinding by the type of treatment
• Variation (within and between study)- random effects model
• For population description, averages calculated for age, proportion of men, SBP, DBP, body weight, BMI and fiber intake in control and intervention groups.
• BP change
  • Parallel trials: net change = change in control group – change in intervention group
  • Crossover trials: net change = final control period blood pressure-final intervention period blood pressure
  • Standard errors (SE) obtained or derived from confidence intervals (CI) P values, or individual SEs for blood pressure changes
• Q statistics: to test homogeneity of effect size
• BP estimates – 95% CI; 2 sided P values < 0.05 were statistically significant.
• Subgroups evaluation stratified meta-analyses for between group effects of fiber supplementation.  Repeated analysis using a multivariate model.
  • Mean age                               <= 40 years        vs.   > 40 years
  • Hypertensive status                 no                     vs.   yes
  • Sex                                         < 50% male        vs.   >= 50% male
  • BMI                                         <= 28 kg/m²       vs.   > 28 kg/m²
  • Type of fiber                           soluble                vs.   insoluble           vs.   mixed
• Associations of fiber dose & intervention duration - linear regression analysis with intercepts & ß regression coefficients with 2-sided P values
• Publication bias – examined by funnel graph and weighted against net changes in BP.  Nonparametric “trim and fill” method used to adjust for bias.

Timing of Measurements:  not applicable

Dependent Variables

• systolic blood pressure
• diastolic blood pressure

Independent Variables

• fiber intake

Control Variables

• Fiber dose
• Duration of intervention
• Type of intervention
• Change in body weight 
• Age
• Sex
• Hypertensive status

Initial N:  72 trials met inclusion criteria.  48 trials were then excluded based on exclusion criteria.

Attrition (Final N):  24 trials with 25 relevant strata remained, representing 1404 subjects.  1 trial was missing data for diastolic blood pressure.  Men only trials-3; Women only trials-6; remainder included both.

Age:  Mean age:42 years (median 40 years; range, 23-63 years)

Ethnicity: not mentioned

Other relevant demographics:

Anthropometrics:

Location:  Worldwide studies

Other Findings

• Sample size range: 12 to 201 subjects
• Mean trial duration: 9.0 weeks (median: 8 weeks; range 2-24 weeks)
• Soluble fiber in 11 trials; insoluble fiber in 7 trials; mixture in remaining trials.
  • Diet intervention in 4 trials
  • Fiber dose range 3.5 to 42.6 g/day.
• Unweighted averages
  • SBP change = -1.54 mmHg (95% CI, -3.48 to 0.39)
  • DBP change = -1.34 mmHg (95% CI, -2.41 to –0.27)
• Significant heterogeneity among trials
  • Systolic X²₂₄  = 24.7, P = 0.002
  • Diastolic X²₂₃ = 37.1, P = 0.03
• Weighted overall effect of fiber supplementation
  • SBP: -1.13 mmHg (95% CI, -2.49 to 0.23)
  • DBP: -1.26 mmHg (95% CI, -2.04 to –0.48)

• Changes in BP less pronounced with increasing fiber dose
  • SBP: -1.55 + 0.032 mmHg / g fiber, P = .60
  • DBP: -1.89 + 0.047 mmHg / g fiber, P = .16
• No relation of BP to duration of intervention
  • SBP: -0.77 – 0.046 mmHg / wk, P = .73
  • DBP: -1.15 – 0.013 mmHg / wk, P = .88
• Type of intervention
  • Soluble fiber
    • SBP: -1.32 mmHg (95% CI, -3.19 to 0.55)
    • DBP: -0.82 mmHg (95% CI, -1.83 to 0.18)
  • Insoluble fiber
    • SBP: -0.23 mmHg (95% CI, -2.88 to 2.42)
    • DBP: -0.57 mmHg (95% CI, -1.86 to 0.72)
  • Mixture of soluble & insoluble fiber
    • SBP: -1.74 mmHg (95% CI, -4.49 to 1.02)
    • DBP: -2.22 mmHg (95% CI, -3.40 to –1.03)
• Change in body weight appeared to not be significantly significant.
• Significantly larger reduction of SBP in older vs. younger groups (P = .04)
• Small trials with large SBP reductions may be overrepresented.

The present meta-analysis shows that dietary fiber has a small BP-lowering effect.  Increasing the intake of fiber in Western populations, where intake is far below recommended levels, may contribute to the prevention of hypertension.

Authors noted the following limitations:

• Not able to completely rule out BP effects from magnesium and potassium in the supplemented fiber.
• Small trials with large SBP reductions may be overrepresented.  However, these trials did have small weights in the meta-analysis.
• Data in this study shows some support for more effect with soluble over insoluble fiber.