Dark chocolate or tomato extract for prehypertension: a randomised controlled trial

The trial was conducted between June and December 2007 in Adelaide, South Australia. Recruitment started in February 2007 by advertising in local newspapers, the University websites, flyers around University noticeboards, a public hospital's health promotion unit, and items on local radio and evening television news programs. We received about 200 enquiries regarding the study, with 87 individuals filling in a screening questionnaire assessing eligibility criteria including recent blood pressure readings and availability for the 6-month study. Inclusion criteria were: 1. Two of either four systolic or diastolic blood pressure readings in the prehypertensive range (SBP 120–139 or DBP 80–89 mm Hg), as measured by our research nurse; 2. Not taking any antihypertensive prescription medication. Exclusion criteria were: 1. Diabetes mellitus; 2. Allergy to or intolerance of tomato or chocolate products. Forty-eight individuals did not meet inclusion criteria (n = 31) or decided not to participate (n = 17) including some who did not want to take the chance of not being randomised to the chocolate group. The remaining 39 healthy volunteers, aged between 22 and 73 years, provided written consent and were allocated to one of three treatment arms (Figure 1).

The study was approved by the Human Research Ethics Committee of The University of Adelaide.

Our trial consisted of two phases: a three group parallel randomised controlled trial (phase 1), followed by a crossover of the two active treatment arms (phase 2). The duration of each study arm was 12 weeks, consisting of an 8-week intervention period followed by a 4-week washout period. Participants were randomly allocated to the chocolate (group 1), tomato extract (group 2), or placebo control (group 3) group by permuted block randomisation using the SAS 9.1 software package [29]. Participants remained in these three groups for 12 weeks (phase 1), followed by the two active treatment groups (group 1 and 2) being crossed over to the other active treatment group for an additional 12 weeks (phase 2). The control group (group 3) ceased participation after phase 1 (at 12 weeks) (Figure 1).

Over an 8-week period, group 1 participants were allocated a 50 g daily dose (half of a 100 g block) of commercially available dark chocolate with 70% cocoa, containing 750 mg of polyphenols (Haighs Chocolates, Adelaide, South Australia [30]); group 2 was allocated one commercially available tomato extract capsule per day containing 15 mg lycopene (Lyc-o-Mato^® Tomato Lycopene Complex capsules, LycoRed Ltd, Beer Sheva, Israel, equivalent to four to five medium size tomatoes) [31]; and group 3 was allocated one placebo capsule per day (matched to tomato extract capsules by colour, odour and size, and containing mainly soy oil). Polyphenol content of the trial chocolate was analysed by the Australian Government National Measurement Institute using the colorimetric Folin-Ciocalteu assay. The trial chocolate contained 15 mg/g of total polyphenols, equivalent to 395 umol/g ORAC (oxygen radical absorbance capacity).

To conceal allocation from investigators, trained staff not involved in trial design and analysis handed out intervention packs to participants. Blinding of participants to chocolate was impractical, however blinding of participants in the capsule groups was achieved by identical packaging of active tomato extract and placebo capsules.

Blood pressure, weight and abdominal circumference were measured every 4 weeks by a trained research nurse, and questionnaires assessing physical activity, smoking status, alcohol intake, general health, mood and energy levels, habitual chocolate and tomato product intake, ease and acceptability of treatment, were administered at baseline and after each intervention period (at 0, 8 and 20 weeks). The questionnaire consisted of 5-point Likert-scales, and closed and open ended questions, and can be obtained from the corresponding author.

Habitual lycopene intake at baseline was estimated using summary values of lycopene content in tomato products based on five databases by Porrini M and Riso P [32]. Participants were counselled to maintain their usual diet and physical activity during the trial. Compliance and any additional chocolate or tomato product intake was assessed using participants' daily diary entries. A follow-up questionnaire was administered 3 months after the trial to assess chocolate and tomato consumption habits.

Blood pressure was assessed by a trained research nurse using a digital calibrated sphygmomanometer with appropriately sized cuffs (Omron HEM-907, JA Davey Pty Ltd). Participants rested for 5 minutes before four blood pressure readings were taken in intervals of 3 minutes in alternate upper arms in a seated position. Median SBP and DBP at one time point were used for analysis to minimise the effect of blood pressure variability.

On a population level a reduction of 4 to 5 mm Hg in SBP and 2 to 3 mm Hg in DBP has been estimated to reduce the risk of cardiovascular morbidity and mortality by up to 20 per cent [33]. We calculated that a sample size of 60, with 20 individuals in each group, would allow detection of this difference at a power of 0.75 with 95% confidence assuming a correlation of 0.5 among repeated measures.

As random number generation was based on a sample size of 60, which we were unable to match in practice, numbers allocated to each of the three groups were slightly disproportionate. Smaller group sizes between n = 10 and 15 in study 1 (three group parallel RCT) reduced the power to 0.5 for detecting a difference of 4.5 mm SBP and 2.5 mm DBP, while reductions of 5–6 mm SBP and 3–4 mm DBP would retain the power at 0.8. However, larger group sizes in the crossover analysis (chocolate groups n = 26, capsule groups n = 36) assessing acceptability of the two treatment forms led to a reasonable effect size > 0.9.

Analyses were performed using SPSS version 15.0 and SAS version 9.1 [29, 34]. Between-group differences at one time point were assessed by Kruskal-Wallis test (continuous variables) and Fisher's-Exact test (binominal variables). A linear mixed model was used to assess within-group and between-group differences of blood pressure over time. Multiple comparisons were adjusted by Holm's Stepdown Bonferroni procedure. P-values < 0.05 were considered statistically significant. Blood pressure changes over time were analysed separately for phase 1 (parallel three group RCT) and phase 2 (two group crossover). Adequate washout between studies was assumed. Acceptability of two treatment forms (chocolate or capsule) was analysed by student t-test.

We collected baseline characteristics of all 39 randomised prehypertensive individuals. However, three participants withdrew before the second time point for data collection scheduled at week 4. All comparisons were subsequently based on the thirty-six remaining participants. At baseline, groups did not differ in anthropometric measures, habitual chocolate or tomato product intake, physical activity, alcohol intake, general health, energy and mood levels. None of the participants smoked. A significantly higher proportion of participants in the chocolate and tomato extract groups took daily fish oil and/or garlic supplements than in the control group (Table 1). As there is evidence to suggest that fish oil or garlic supplementation may have a blood pressure reducing effect in hypertensive individuals [35, 36], dietary supplementation was subsequently included as a binominal covariate in the analyses of blood pressure. While median SBP and DBP at baseline did not differ statistically between groups, observed differences were clinically relevant, and were therefore included as covariates in the linear mixed model analyses of blood pressure changes over time.

As mentioned above three individuals withdrew after baseline assessment but before the 4-week assessment. In addition, one individual withdrew after crossover to study 2 (after 12 wks but before the 16 wk assessment). Two individuals withdrew when allocated to a chocolate group (one in phase 1, one in phase 2), both indicating that they found the dark chocolate unpalatable (Figure 1). One individual allocated to the tomato extract capsules reported gastrointestinal upset and one individual allocated to the control group withdrew due to illness apparently unrelated to our trial. In total, 36 participants completed phase 1, twenty-five completed phase 2, and attended all monitoring sessions, respectively. Self-reported diary entries indicated excellent compliance with the study protocol and consumption of provided chocolate or capsules. Additional chocolate or tomato product consumption during the intervention and washout periods was negligible (data not shown).

Weight, body mass index (BMI), abdominal circumference, physical activity, general health, energy and mood levels were assessed at three time points at 0, 8 and 20 weeks (baseline, end of intervention period in phase 1, and end of intervention period in phase 2). No significant changes over time within the three groups were observed (Table 3). Differences in characteristics between groups remained insignificant (data not shown). Not all participants agreed to have their abdominal circumference measured at all time points, therefore mean values are not representative of all participants and changes over time need to be interpreted cautiously.

At the end of the intervention periods of phases 1 and 2 we assessed acceptability and willingness to continue with treatment long-term with daily consumption of chocolate or a capsule. The two chocolate groups of phase 1 and crossover phase 2 were analysed together, as well as all capsule groups (tomato extract and placebo groups of phase 1 and tomato extract group after crossover in phase 2). Differences in acceptability between the two treatment forms were highly significant (p < 0.0001, effect size 0.9): half of the participants allocated to the chocolate treatment found it hard to eat 50 g of dark chocolate every day and 20% considered it an unacceptable long-term treatment option, whereas all participants found it easy and acceptable to take a capsule each day for blood pressure (Table 4). Unacceptable taste, concerns about fat content, and side effects such as constipation and headache were reported reasons for unacceptability of chocolate as a regular treatment option. The majority of participants stated that they would be willing to continue treatment if it was effective (Table 4). Follow-up of participants three months after the trial revealed that tomato product consumption did not change significantly, although 80% had intended to add more tomatoes to their diet. However, dark chocolate consumption increased after the trial in all groups (Table 4).