This is a free-living study that showed that the regular addition of two servings of soluble cocoa products, containing 15.1 and 22.0% (w/w) of NSP, respectively, to a Spanish diet is an efficacious alternative to increase fibre intake to recommended levels and the product richer in NSP (22.0%, w/w) also promotes healthier bowel habits. The servings used in this study correspond to the quantity of the product that can reasonably be expected to be consumed. The consumption rate of two cocoa beverages per day reproduces real conditions in the Spanish population and may be considered moderate [, http://www.cacaoychocolate.com/consumoen.html]. The only dietary modification introduced in this study, apart from consuming the cocoa products, was the restriction of certain fruits and vegetables. However taking into account the restrictions, the intake of DF remained within the range estimated in the Spanish population (16.3 - 18.4 g/day)  and was similar to that in European countries (18.5 g/day) . Up to date, there is no gold standard for fibre intake assessment being dietary food records a well accepted alternative, particularly if the influence of smoking, alcohol intake, gender and education is considered . With this regard, volunteers who participated in this study were non-smokers, consumed very low amounts of alcohol and had a similar, medium-high, education level. In the non-cocoa stage, the intake of fibre (Table 2) was near to the lower limit of the various daily fibre recommendations, which range from 15 to 35 g . The regular consumption of either cocoa product improved the intake of DF with respect to the non-cocoa stage, particularly that corresponding to cocoa product B. The higher intake of fibre associated to consuming cocoa product B compared to A is due both to their unlike NSP content (22.00% vs. 15.09%, respectively) and to the different net quantity consumed daily of each cocoa product (30 g of cocoa product B vs. 15 g of A) which provided 6.6 and 2.26 g/d of total NSP.
According to the bowel function and symptom questionnaires, both cocoa diets were well tolerated and volunteers did not describe adverse gastrointestinal symptoms such as abdominal pain or bloating compared to the non-cocoa period. The only negative effect described was a higher perception of flatulence when the product richer in NSP was consumed. This outcome was not completely supported with the perception of passing gas during the day, which was slightly higher in the cocoa product B stage without reaching a statistical difference compared to other stages. This symptom has been widely described as a side-effect related to fibre intake [13, 14]. Excepting this result, there was a predominant positive gastrointestinal tolerance to the fibre-rich cocoa periods which may be attributed to the adaptation hypothesis, according to which a continuous intake of DF relieves negative gastrointestinal side-effects based on the equilibrium that is reached between the gas-producing and gas utilizing bacteria . Other interesting outcomes associated with the consumption of product B, was the lower perception of producing less hard stools with lower stool consistency without leading to the feeling of having diarrhoea. In agreement, Castillejo et al.  described that a significantly higher number of parents reported a subjective improvement in stool consistency in their chronically constipated children who consumed a supplement of cocoa husk.
Four positive significant changes on bowel function over time were described associated to consuming the cocoa product richer in fibre (6.6 g/d of total NSP): a reduced perception of being constipated, a higher record of having a bowel movement per day, supported by an objective higher number of daily bowel movements, and a lower time to have the bowel movement. The higher frequency of daily bowel movements is in agreement with other studies. Jenkins et al.  obtained a significantly higher frequency of movements per day when a diet containing a cocoa-bran cereal (total DF 25 g/day) was consumed for 2 weeks versus a low-fibre cereal (total dietary fibre 5.6 g/day). Similarly, Dahl et al.  described that the addition of a moderate amount of finely processed pea hull fibre to usual food (doses ranging 20-26 g/day) helped to normalize and improve bowel function in elderly subjects resulting in a significant increase in bowel movements among other improvements of clinical symptoms of constipation. Similar effects where observed with fibre supplementation (the doses used were "age in years plus 5 g/day") in chronically constipated children .
In addition to changes in bowel function based on the subject's own reports recorded in the subjective questionnaires and the diary, an objective measurement of intestinal transit time was carried out which evaluates overall colonic function. The reported average transit time usually takes 5 days . On one hand, we were aware that using a faecal dye to estimate intestinal transit time presents the limitation that transit time is influenced by the level of either food or fibre in the diet . In order to minimize the possible confounding effect that could be produced by other food items consumed simultaneously with the dye, the dye capsule was administered with a standardised fibre-free breakfast, formed by two slices of white bread and semi-skimmed milk which was administered in each of the stages of the study. Additionally, the dye was consumed on one of the days in which the 72-h food record was filled out, so volunteers' food intake was recorded. In this sense, no differences in DF consumption were observed between the three days in each stage (data not shown). On the other hand, the advantages of using brilliant blue are that it is non-invasive, compared to other dyes its differentiation is quite unequivocal, and it does not require stool collection . Our results are in agreement with the appearance rate of 24 to 36 hours and the disappearance rate between 48 and 72 hours previously described in healthy individuals . The dye appearance time was about 1.5 hours shorter when cocoa product B was consumed compared to the non-cocoa stage, while the disappearance time was 4 hours shorter. However these differences did not reach the level of statistical significance. These effects may be attributed to the higher content of water insoluble forms of NSP in cocoa product B (B: 20.3% versus A: 12.0%) which are known to promote colonic mucosal health through inducing changes such as increasing faecal bulking or reducing transit time [12, 34]. Therefore, the positive self-perceived symptoms previously described associated to consuming cocoa product B are supported with objective shorter intestinal transit times, although not statistically lower.
Surprisingly, product A showed higher dye appearance and disappearance times compared to the non-cocoa stage. Regarding these results, it should be noted that very high intraindividual variance was obtained in the cocoa product A stage, with abnormally high transit times in some cases, which is in agreement with the elevated variation in intestinal transit time described in other studies . These results may be attributed to other life-style factors that were not possible to control in this free-living study. Walking, running, and strength training have been described to reduce gastrointestinal transit time, [35–37] although the effectiveness of physical activity in the management of constipation remains controversial. Recently, a study performed in normally active subjects, between 24-77 y, failed to prove this relationship . In contrast, in a group of subjects over 40 years of age, dietary factors have little influence on defecation habits because of the stability of large bowel function, and exercise, including walking, suggested to be more useful in preventing constipation . In this study, the physical activity values are obtained based on the assumption that 1 MET is approximately 1 kcal/min may have led to an overestimation of energy expenditure although this fact is not particularly relevant as the same factor was used along the study and physical activity was not different.
As cocoa products are relatively high energy food, in fact the energy intake was significantly lower (p = 0.045) at the non-cocoa stage, their addition to a diet could induce weight gain. However, the consumption of the cocoa products did not lead to an increase in body weight, without disregarding the intervention stages were 4 weeks long. These results are in agreement with previous studies in humans , in which the energy intake was higher in dark chocolate consumers than in nonconsumers, although BMI in the former group was lower. The association between dark chocolate consumption and lower BMI has been attributed to the physiologic activity of polyphenols, which have also been described with tea polyphenols in humans . The mechanisms involved in the antiobesity effects of cocoa have been studied in rats and it seems that cocoa modulates lipid metabolism, especially by decreasing fatty acid synthesis and transport systems, and enhances part of the thermogenesis mechanism in liver and white adipose tissue .
The present study presents certain methodological limitations: the sample size was calculated according to total cholesterol, therefore it is not possible to know if the study was adequately powered, especially after having observed that bowel habits show high variability. Other components of the cocoa products and diet may have influenced the outcome on bowel habits. Faecal dyes cannot be quantitated in faeces.
Strengths of the study are that it is the only study that has used multiple assessments to investigate the effect of cocoa bran on bowel habits in healthy adults. To our knowledge, there are only two other human studies that focus on the same issue, one having been carried out in constipated children and thus the questionnaires were answered by their parents , and the other uses only one tool, the symptom diary, to evaluate the effects of cocoa bran on bowel habits . Moreover, it has been carried out in 44 subjects which is a relatively high number compared to similar studies [1, 9].