Table 2

Stratified meta-analyses of the effects of fenugreek on fasting blood glucose and 2 hour postload glucose according to trial and participant characteristics
Characteristics Fasting blood glucose 2 hr-postload glucose
n Pooled estimate I2 (95% CI) Phetr PEM n Pooled estimate I2 (95% CI) Phetr PEM
Overall 11 -0.96 (-1.52, -0.40) 80 (65, 89) <0.001 8 -2.19 (-3.19, -1.19) 71 (40, 86) 0.001
Study design
  Parallel 5 -0.78 (-1.93, 0.37) 86 (69, 94) <0.001 0.48 4 -1.71 (-2.73, -0.70) 59 (0, 86) 0.07 0.33
  Crossover 6 -1.20 (-2.03, -0.38) 77 (49, 90) 0.01 4 -3.32 (-5.90, -0.75) 82 (54, 93) 0.001
Region
  India 7 -1.43 (-2.26, -0.60) 48 (0, 78) 0.07 0.21 6 -2.60 (-4.06, -1.13) 73 (39, 88) 0.002 0.50
  Others 4 -0.55 (-1.25, 0.16) 89 (75, 95) <0.001 2 -1.59 (-3.08, -0.10) 74 (0, 94)
Study precisiona
  Below median 5 -0.54 (-1.15, 0.07) 86 (69, 94) <0.001 0.10 4 -1.55 (-2.52, -0.57) 67 (2, 89) 0.03 0.11
  Above median 6 -1.68 (-2.62, -0.74) 43 (0, 77) 0.12 4 -3.42 (-5.56, -1.28) 67 (4, 89) 0.03
Study populationb
  Healthy 2 0.05 (-0.51, 0.61) 82 (23, 96) 0.02 - - - -
  T1DM 1 -4.20 (-6.73, -1.67) - - 0.01 1 -3.20 (-6.92, 0.52) - - 0.79
  T2DM 8 -1.21 (-1.69, -0.73) 26 (0, 67) 0.22 7 -2.14 (-3.19, -1.09) 74 (45, 88) 0.001
Randomization
  Yes 9 -0.94 (-1.56, -0.32) 83 (69, 91) <0.001 0.83 6 -2.54 (-3.87, -1.21) 73 (37, 88) 0.003 0.49
  Unknown 2 -1.11 (-2.10, -0.12) 0 (0, 100) 0.49 2 -1.58 (-3.61, 0.45) 81 (17, 95) 0.02
Blinding status
  Double-blinded 4 -0.44 (-1.22, 0.34) 88 (72, 95) <0.001 0.15 2 -1.95 (-3.02, -0.88) 16 (0, 55) 0.28 0.74
  Unknown 7 -1.37 (-2.08, -0.67) 47 (0, 78) 0.08 6 -2.43 (-3.82, -1.04) 78 (50, 90) 0.000
Daily dose (grams)b
  <5 3 0.02 (-0.46, 0.50) 63 (0, 89) 0.07 1 -1.34 (-2.83, 0.14) - -
  5-10 4 -1.27 (-1.80, -0.74) 21 (0, 88) 0.28 0.01 4 -1.55 (-2.52, -0.57) 67 (2, 89) 0.04 0.02
  >10 4 -2.07 (-3.58, -0.55) 68 (9, 89) 0.02 3 -4.42 (-5.96, -2.89) 8 (0, 35) 0.58
Preparationb
  Powdered seed 5 -1.12 (-1.63, -0.61) 29 (0, 72) 0.23 4 -1.55 (-2.52, -0.57) 67 (2, 89) 0.03
  Hydro-alcoholic extract 3 0.02 (-0.46, 0.50) 63 (0, 89) 0.07 0.002 1 -1.34 (-2.83, 0.14) - - 0.05
  Debitterized seed powder 3 -2.68 (-3.82, -1.54) 0 (0, 90) 0.42 3 -4.42 (-5.96, -2.89) 0 (0, 90) 0.58
Study duration (days)
  <30 5 -1.48 (-2.64, -0.32) 80 (52, 91) 0.001 0.39 3 -4.42 (-5.96, -2.89) 8 (0, 35) 0.58 0.02
  > = 30 6 -0.78 (-1.68, 0.11) 83 (65, 92) <0.001 5 -1.49 (-2.28, -0.70) 55 (0, 84) 0.06
Mean Age (years)
  <43.1 4 -0.73 (-1.61, 0.16) 86 (67, 94) <0.001 2 -4.14 (-5.80, -2.47) 0 (0, 100) 0.58
  > = 43.1 5 -1.11 (-1.76, -0.47) 45 (0, 80) 0.12 0.60 4 -1.90 (-3.19, -0.61) 67 (4, 89) 0.03 0.21
Mean BMI (kg/m2)
  <25 3 -1.62 (-3.21, -0.02) 93 (82, 97) <0.001 0.29 2 -2.52 (-3.75, -1.30) 0 (0, 100) 0.71 0.17
  > = 25 3 -0.19 (-1.14, 0.76) 74 (12, 92) 0.02 2 -1.01 (-1.72, -0.30) 0 (0, 100) 0.62

Abbreviations: BMI body mass index, CI confidence interval, Phetr p value for heterogeneity, PEM p value for effect modification.

aThe cutoffs are based on the median of standard error of the effect size for fasting blood glucose and 2hr glucose respectively; For mean age and study duration, the cutoffs are based on the median values of 11 data points.

bPEM-value for the overall F-test for a common mean amongst the three groups (i.e., study population, and fenugreek drug preparation format), the overall p-value for daily dose was obtained by modeling this as a continuous variable in meta-regression analysis.

Neelakantan et al.

Neelakantan et al. Nutrition Journal 2014 13:7   doi:10.1186/1475-2891-13-7

Open Data