# Measuring bioelectric impedance outputs following coffee consumption in healthy adults using an 8-electrode segmental BIA device

## TL;DR

This study found that drinking coffee before a BIA test does not significantly affect the results, suggesting that coffee consumption may not need to be restricted before such measurements.

## Contribution

The study challenges the need to restrict coffee consumption before BIA measurements by showing no significant effect of caffeine on BIA outputs.

## Key findings

- Caffeine consumption did not significantly affect BIA outputs.
- Changes in impedance, resistance, and reactance were detected but not linked to caffeine.
- Effect sizes were small, indicating minimal practical significance of coffee consumption on BIA measurements.

## Abstract

Bioelectric impedance analysis (BIA) is increasingly used to measure body composition in athletic, clinical and research settings. The reliability of BIA measurements relies on following procedures carefully. However, some practices for ensuring reliable measures may be unnecessarily restrictive. Previous research using BIA outputs as study outcome measures, has required participants to avoid coffee and caffeine-containing foods and beverages prior to measurements. The aim of this study was to determine whether BIA outputs are affected by coffee consumption at different caffeine concentrations.

This study used a blinded, randomized, cross-over trial design. Participants (n = 13) received one of three treatments per visit: 200 mL hot water (80°C) mixed with 5 g instant coffee, 2.5 g instant coffee with 2.5 g decaffeinated coffee, or 5 g decaffeinated instant coffee. Body composition and fluid parameters were measured over 50 minutes using a Seca mBCA 515 device.

The treatment predictor (p > 0.05) and sex-time-treatment interaction for all outcomes was found to be non-significant (p > 0.05). The time predictor was statistically significant (p < 0.05) for impedance, resistance and reactance but not for phase angle ϕ50 (p = 0.731), ϕ5 (p = 0.059) or urine osmolality (p = 0.066). The sex predictor was statistically significant for Z50 (p = 0.001), Z5 (p = 0.002), R50 (p = 0.001), R5 (p = 0.002), ϕ50 (p = 0.01), ϕ5 (p = 0.049), fat mass (%) (p = 0.016) and fat free mass (%) (p = 0.016). The effect size for this predictor was η2G < 0.336. A significant sex-time interaction was found for Z50 (p = 0.025) with a small effect size (η2G < 0.01). Effect sizes for the treatment predictor and time-treatment interaction were found to be small (η2G < 0.01). Effect size for the time predictor was small (η2G > 0.01).

Changes in impedance, resistance and reactance were detected over the course of the experiment, and these changes were greater than could be explained by the technical error of measurement. However, the amount of caffeine in coffee did not affect BIA outputs. Effect sizes were small, suggesting little practical significance of drinking coffee before taking BIA measurements. Therefore, coaches, athletes, researchers and clinicians may be able to obtain reliable BIA measurements even when coffee has been consumed up to 50 minutes prior to measurements, however, fluid consumption and being adequately hydrated should still be considered.