Meta-analysis or Systematic Review

M - Click here for explanation of classification scheme.

NEUTRAL: See Quality Criteria Checklist below.

• To determine the effect of calcium supplementation on blood pressure (BP) 
• To examine calcium and blood pressure response relative to subject characteristics such as age, gender and calcium intake.

• Randomized controlled trials in humans
• Published between 1966 and June 2003
• Any language
• Normotensive or hypertensive persons
• Non-pregnant females
• Co-intervention with vitamin D acceptable.

• Duplicate publication
• Insufficient BP data
• Co-intervention except vitamin D
• Study population with renal disease or hyperparathyroidism
• No control group
• Study duration less than two weeks.

Recruitment

• MEDLINE database searched with the following MESH terms: (calcium NOT calcium antagonis* NOT calcium channel NOT calcium entry NOT calcium blocker) AND (blood pressure OR hypertension)
• Retrieved articles, meta-analyses and systematic reviews searched manually for additional eligible studies.

Design 

Meta-analysis.

Intervention

• Data abstraction completed by two authors; a third author helped reach consensus if necessary
• Studies were scored based on blinding: open, single-blind or double-blind
• The following were extracted from each study:
  • Changes in systolic blood pressure (SBP) and diastolic blood pressure (DBP) as well as variance measures
  • Dose and source of calcium
  • Characteristics of trial population:
    • Mean age
    • Gender
    • Race
    • Initial SBP and DBP
    • Habitual or initial calcium intake.
• In trials with different doses or formulations of calcium compared to a single control group, average BP changes in the intervention group were compared to the control group
• In trials with multiple BP measurements over time, the longest follow-up period while still receiving randomized treatment was used
• If raw data is not available from the article, BP adjusted for confounders was used
• For identical studies with multiple publications, all available information was collected and the most recent article was cited
• Data for mean age, gender and calcium intake were input, obtained from authors or derived from meta-analysis or review if not available.

Statistical Analysis

•  Variance:
  • If not available, determined from confidence intervals, P-values or individual variance measures between intervention and control groups or periods
  • Pooled for parallel trials if no measures available for paired differences
  • If measures are not available at the end of follow-up, variance was assumed to be similar to baseline values.
• Homogeneity of effect size was assessed with Q statistics
• Random effects model used to assess BP:
  • BP effect was weighted by the reciprocal of its variance (1/SE²)
  • P≤0.05 was considered significant (two-sided) with 95% CI reported.
• In sub-group analyses, BP response and median calcium dose were compared to trial population characteristics such as median age, gender, initial BP and initial or habitual calcium intake
• Additionally, calcium dose vs. BP was adjusted for above confounders (except when used as a stratification variable) using a multivariate random effects model
• Publication bias was examined visually by plotting weighted blood pressure effect (1/SE²) against net changes in SBP. To adjust for publication bias, the non-parametric "trim and fill" method was used.

Dependent Variables

• Primary:
  • Change in BP:
    • Parallel trials: Change in baseline BP in intervention group subtracted from change in baseline BP in control group
    • Crossover trials: If baseline BP was similar for intervention and control periods, the change in BP was determined by subtracting the final BP during the control period from the final BP during the intervention period. Otherwise, change in BP was determined as it was for parallel trials.
• Secondary:
  • Change in BP vs trial population characteristics:
    • Median age: Less than 45 years vs. 45 years or more
    • Initial BP: Less than 140/90mm Hg vs. 140/90mm Hg or more
    • Gender distribution: 50% or less, or more than 50% males
    • Median initial calcium intake: Less than 800mg vs.800mg or more per day.

Independent Variables

Calcium dose was obtained from dietary sources or supplements.
 

Initial N

N=40 (2,492 subjects):

• Crossover design: N=18 studies (572 subjects)
• Parallel design: N=22 studies (1,920 subjects)
• Blinding:
  • Double blind: N=32 studies
  • Open: N=8 studies.
• Trial sample size ranged from seven to 452 subjects
• Trials with men only: N=5
• Trials with women only: N=4
• Trials with both men and women with varying percentages: N=30
• Median duration of intervention: An average of 9.5 weeks with a range of three weeks to 208 weeks.

Age

Ages of 43.7±14.3 years (mean±SD), with a range of 11 years to 77 years.

Ethnicity

• Trials with predominately White participants: N=31
• Trials with White and African American participants: N=1
• Trials with predominately African American participants: N=4
• Trials with predominantly Asian participants: N=4.

Other Relevant Demographics

• Blood pressure:
  • Participants with hypertension: N=2,258
  • Participants with normotension: N=234
  • Trials with hypertensive-only participants: N=36
  • Trials including both hypertensive and normotensive participants: N=4.
• Calcium supplementation:
  • Dose:
    • Mean: A total of 1,200mg per day
    • Median: A total of 1,055mg per day
    • Range: A total of 355mg to 2,000mg per day.
  • Source of calcium:
    • Dietary sources only: N=4 studies
      • Dairy products: N=3 studies
      • Calcium citrate malate enriched juice: N=1 studies.
    • Dietary sources plus supplement (calcium carbonate): N=1 studies
    • Supplements only: N=35
      • Calcium gluconate: N=11 studies
      • Calcium carbonate: N=16 studies
      • Calcium citrate: N=2 studies
      • Calcium carbonate + calcium citrate: N=1 studies
      • Calcium carbonate + vitamin D: N=1 studies
      • Calcium gluconate + calcium yeast: N=1 studies
      • AA calcium: N=1 studies
      • Calcium supplement (not defined): N=1 studies
      • Source of calcium not provided: N=1 studies.

Location

The Netherlands:

• Wageningen University, Wageningen
• Erasmus Medical Center, Rotterdam
• Maastrict University, Maastrict
• University Medical Center Utrecht, Utrecht.

•  Overall effect of calcium supplementation (mean dose of 1,200mg daily) on BP:
  • SBP, mm Hg: –1.86 (95% CI: –2.91 to –0.81)
  • DBP, mm Hg: –0.99 (95% CI: –1.61 to –0.37).
• Sub-group analyses of calcium supplementation on BP:
  • BP responses to calcium supplementation did not differ significantly for the following in unadjusted or adjusted models:
    • Age: Less than 45 years vs. more than 45 years
    • Male gender distribution: 50% or less vs. more than 50%
    • Initial BP: Less than 140/90mm Hg vs.140/90mm Hg or more
    • Initial calcium intake: 800mg or less vs. more than 800mg per day
    • Calcium dose: 1,000mg or less vs. more than 1,000mg per day.
• Asians with a usual calcium intake of 400mg to 550mg per day (four trials with 154 participants) experienced a greater blood pressure response to calcium supplementation compared to Blacks and Whites in both unadjusted and adjusted models
  • Asians: –9.66/–4.62mm Hg (SBP/DBP, adjusted)
  • Blacks: –1.54/–0.32mm Hg (SBP/DBP, adjusted)
  • Whites: –1.21/–0.78mm Hg (SBP/DBP, adjusted).

Other Findings

• Significant heterogeneity for effect size (Q statistic) for both SBP and DBP (P<0.0001)
• Visual examination of funnel plot suggested that small trials with larger decreases in SBP were possibly over-represented
• Results of non-parametric fit and trim model suggested two trials may have been missing.

Calcium intake of approximately 1,000mg daily, from dietary sources or supplements, had a positive effect on blood pressure response, decreasing SBP by 2mm Hg and DBP by 1mm Hg. Populations with a habitually low calcium intake may experience more improvement in BP than populations with an adequate intake of calcium. Age, gender and the presence of hypertension did not alter the BP response to calcium supplementation.

University/Hospital:

Wageningen University; Erasmus Medical Center, Rotterdam; Maastrict University; University Medical Center Utrecht; The Netherlands

• Initial or habitual calcium intake was not reported for 16 trials. Mean intake was imputed from the population from which the subjects were recruited.
• In the methods, the authors did not discuss whether any studies were identified in the the manual search or whether any non-English or unpublished trials were identified
• The authors attempted to assess quality of studies by assessing whether a trial was blinded
• While the authors assessed homogeneity of the effect size among the trials, they did not address potential factors accounting for homogeneity in their discussion
• Total number of participants was provided but would have liked to have seen a more detailed breakdown by gender and ethnicity
• The outcomes reflect only those who completed the trials.