Weight reduction and blood pressure: averages, not a per-kilo promise
A meta-analysis of 25 randomized trials quantified pressure changes while exposing the limits of a simple weight-loss rule.

Illustrative food image, not a photograph of the study intervention.
What the randomized evidence combined
Neter and colleagues pooled 25 randomized controlled trials published between 1966 and 2002, comprising 34 strata and 4,874 participants. Weight reduction was achieved through energy restriction, increased physical activity, or both. This was therefore not a trial of one named diet, a single food, or an exclusively plant-based eating pattern. The researchers used a random-effects model to account for differences among trials. That design supports examining the average blood-pressure effect of the weight-reduction interventions, while leaving an important attribution limit: the pooled result does not isolate one menu or ingredient as the cause of the observed changes.
Neter JE, Stam BE, Kok FJ, Grobbee DE et al.. Hypertension (Dallas, Tex. : 1979), 2003. PMID 12975389 · DOI
The pooled pressure change
Across the trials, net weight reduction averaged 5.1 kg, with a 95% confidence interval of 4.25 to 6.03 kg lost. The pooled systolic blood-pressure reduction was 4.44 mm Hg, with a confidence interval from 2.95 to 5.93 mm Hg lower. Diastolic pressure fell by 3.57 mm Hg, with a confidence interval from 2.25 to 4.88 mm Hg lower. These estimates describe differences summarized across trial populations, not a required weight target for an individual. They also concern blood-pressure measurements: the abstract does not report a pooled reduction in strokes, heart attacks or deaths from these interventions.
Neter JE, Stam BE, Kok FJ, Grobbee DE et al.. Hypertension (Dallas, Tex. : 1979), 2003. PMID 12975389 · DOI
Why the per-kilogram figure is not a calculator
Expressed per kilogram of weight loss, the estimated reductions were 1.05 mm Hg systolic and 0.92 mm Hg diastolic. The respective 95% confidence intervals were 0.66 to 1.43 and 0.55 to 1.28 mm Hg lower. These figures are a convenient description of the pooled evidence, not a personalized conversion formula. The trials differed, and the interventions included changes in energy intake, activity, or both. Multiplying a planned weight change by one of these numbers would imply individual predictability that a population summary does not provide. The useful message is the direction and approximate scale of the average association within these randomized interventions.
Neter JE, Stam BE, Kok FJ, Grobbee DE et al.. Hypertension (Dallas, Tex. : 1979), 2003. PMID 12975389 · DOI
Subgroups reinforce the need for context
Trial populations losing more than 5 kg on average had larger blood-pressure reductions than populations losing less. Systolic reductions were 6.63 versus 2.70 mm Hg, and diastolic reductions were 5.12 versus 2.01 mm Hg. The diastolic effect was also larger in populations taking antihypertensive medication than in untreated populations: 5.31 versus 2.91 mm Hg lower. These are comparisons between population subgroups, not instructions to change medication or pursue a specified weight loss. The review supports weight-reduction interventions as relevant to blood-pressure management, but it does not establish that the same numerical response will occur in every person or with every eating pattern.
Neter JE, Stam BE, Kok FJ, Grobbee DE et al.. Hypertension (Dallas, Tex. : 1979), 2003. PMID 12975389 · DOI