*Limitations and Future Work*

These two studies are the first dietary clinical investigations of an ulvan, and specifically of SXRG84. They indicated a highly consistent anti-inflammatory effect. However, the lipid effects were not replicated in Study 2, possibly because of differences in the study population.

The use of the crossover designs—as in Study 2—in randomized clinical trials are popular because they can reduce bias from confounding variables and allow participants to act as their own control. Crossover trials are appropriate for use where treatment effects are short-lived in chronic conditions [46]. However, crossover trials also have complications, including the possibility of "order effects". In Study 2, there were no consistent effects found for either of the treatment groups or the placebo groups when examining the significant outcomes. As such, we analyzed the data as four groups instead of collapsing them into two treatment groups. The need to analyze four separate groups does undermine the strengths of running a crossover trial. The true baseline at week zero was used as a covariate in all of the metabolic and inflammatory outcome measures in an attempt to reduce any carryover effects. Study 2 did not include a washout period, with the belief that six weeks on the placebo post SXRG84 treatment would not result in any carried over benefits. Furthermore, any improvements seen after six weeks on the SXRG84 treatment should have returned to baseline after the placebo in the second arm. Indeed, the gut microbiota can revert back to their initial state within 48 h of ceasing specific diets [47]. Figure 6 suggests that the gut microbiota shift that was observed in the placebo group following the SXRG84 was minimal, and many shifts observed after six weeks on the SXRG84 reverted back in the subsequent placebo period. For example, Bacteroides decreased following the SXRG84 treatment and then increased again following the placebo. Regardless, not including a washout period can make identifying treatment effects difficult. Future crossover trials examining the gut microbiome should consider a washout period.

Lastly, the multiple comparisons used in this study makes the likelihood of type I errors more common and need to be considered when assessing the results. We did not adjust for multiple comparisons, as this can have the opposite effect and make type II errors more common [48], therefore making it difficult to determine whether any effect exists. Due to the novel nature of this work, we would prefer to present unadjusted p-values to identify potential treatment effects.

In conclusion, Study 1 showed that the dietary inclusion of SXRG84 had a beneficial effect on a number of lipid and inflammation markers and showed a relationship broadly consistent to gut flora shifts in overweight and obese humans. The results of Study 2 failed to confirm the reduction in non-HDL cholesterol from Study 1 but did confirm the anti-inflammatory potential of SXRG84 in overweight adults across a range of specific cytokine markers. These anti-inflammatory effects may exert benefits to the host as inflammatory cytokines are interrelated with metabolic and cardiovascular diseases. Three genera (*Fusicatenibacter*, *Parabacteroides*, and *Clostridium)*, which consistently responded whilst on the SXRG84 treatment, were identified, supporting the prebiotic potential of this extract.

Importantly, there were no changes in blood counts or other markers that indicated a compromise in health. The potential for the supplements as a preventative or additional therapy is apparent but requires further investigation. It is evident that effects will also be highly specific to an individual considering their baseline metabolic state as well as gut flora composition.
