A phase I single-blind clinical trial to evaluate the safety of oil palm phenolics (OPP) supplementation in healthy volunteers

Plant phenolics are being increasingly consumed globally with limited scientific and clinical evidence pertaining to safety and efficacy. The oil palm fruit contains a cocktail of phenolics, and palm oil production results in high volumes of aqueous by-products enriched in phenolics and bioactives. Several lines of evidence from in vitro and in vivo animal studies confirmed that the aqueous extract enriched in phenolics and other bioactives collectively known as oil palm phenolics (OPP) is safe and has potent bioactivity. A phase one clinical trial was conducted to evaluate the safety and effects of OPP in healthy volunteers. In this single-blind trial, 25 healthy human volunteers were supplemented with 450 mg gallic acid equivalent (GAE)/day of OPP or control treatments for a 60-day period. Fasting blood and urine samples were collected at days 1, 30 and 60. Medical examination was performed during these trial interventions. All clinical biochemistry profiles observed throughout the control and OPP treatment period were in the normal range with no major adverse effect (AE) or serious adverse effect (SAE) observed. Additionally, OPP supplementation resulted in improvement of total cholesterol and LDL-C levels, compared to the control treatment. The outcomes support our previous observations that OPP is safe and may have a protective role in reducing cholesterol levels.

the mice supplemented with OPP as a drink were significantly lower compared to controls 11 . Microarray profiling of BALB/c mice supplemented with OPP showed that several genes related to hepatic lipid catabolism were up-regulated, indicating suppression of liver fat and visceral fat accumulation in the body. Additionally, genes involved in cholesterol biosynthesis were down-regulated, suggesting a possible role in preventing hypercholesterolemia 12 . OPP also attenuated atherosclerosis 13,14 as well as responded positively towards the distal colonic contractility and motility 15 in animal models.
Long-term intake of OPP protected healthy, young Nile rats against diabetes onset, as measured by glucose, blood lipids, and weights of livers and kidneys 11 . These outcomes were probably due partly at least to the phenolic compounds in OPP and their protection of ß-cells in the pancreatic islets against oxidative stress, thereby maintaining the integrity and ability of ß-cells to produce insulin 16 . Recently, the anti-diabetic potential by OPP supplementation has been postulated to be a result of enhanced insulin sensitivity and reduced glucose absorption/output and not an increase in insulin secretion, as indicated by down-regulation of insulin signaling genes 17 .
The wide spectrum of bioactivities demonstrated by OPP in in vitro and in vivo systems suggests its potential application for a range of chronic diseases. Currently however, there is still no clinical data on the effects of OPP in humans in terms of safety and efficacy. Therefore, we conducted a phase one clinical trial to evaluate the physiological effects of OPP in healthy human subjects. The primary objective of this trial was to evaluate if there were any Adverse Effects (AE), and/or Serious Adverse Effects (SAE) following OPP administration (up to 60 days supplementation). Secondary objectives were to evaluate physiological effects which resulted from OPP supplementation. Data from this trial would be important for regulatory requirements and developing a safety profile of OPP besides understanding the physiological roles of OPP in humans under normal conditions.

Methods
Volunteers. Twenty-five healthy volunteers, consisting of 11 males and 14 females were recruited from the Malaysian Palm Oil Board (MPOB). Volunteers were thoroughly briefed on the objectives, design, and trial protocol before they signed the informed consent forms. All volunteers were normolipaemic, nonsmokers, and did not show any clinical symptoms associated with cardiovascular disease. They were examined for their health status and medical history by a medical officer before participation. Through the administration of a questionnaire and dietary interview, we established that none of the volunteers consumed any vitamin or herbal supplements, or were taking any prescribed medication. Female volunteers were not pregnant, lactating, or taking contraceptives at the time of recruitment. The recruitment was completed with the following baseline characteristics: Means ± SDs: Age, 29.24 ± 4.31 y; body mass index, 22.41 ± 3.99 kg/m 2 ( Study protocol. The trial was a single-blind trial, with comparison to control treatment. The study design enabled each subject to serve as his/her own control. On day one of the trial, volunteers attended a clinic for baseline measurement of plasma clinical profiles, after an overnight fast of at least 10 hours. Fasting blood (20 mL) and urine samples were taken. Volunteers were assigned into two intervention groups, where one group (treatment group) was given 150 mL of OPP (containing 1500 mg/L GAE OPP), twice per day. The total amount of phenolic compounds supplemented was 450 mg GAE/day. The composition of OPP was analyzed according to the method described previously 2 . Concentrations of major phenolic compounds are tabulated in Table 2. The other group (control group), was given 150 mL of control drinks (drinking water), twice per day. Both treatment and control drinks were made available in 250 mL amber glass bottles and kept refrigerated until distribution to the volunteers. In order to ensure compliance, all volunteers consumed both drinks in front of the investigator. All drinks consumption record was verified by the investigator and documented accordingly. The trial design allowed the investigators to detect early physiological outcomes from OPP intake. The trial period for OPP and control treatments was 60 days each where a four-week (28 days) wash-out period was allowed between both treatments. Volunteers were allowed to maintain their habitual diets and life styles during this period. Fasting blood (20 mL) and urine were also sampled at day 30 and 60. In every clinic visit, volunteers were thoroughly examined for their health status by the medical officer. The measurement of body weight (Table 3) and blood pressure ( Blood samples were drawn from the volunteers by an experienced and well-trained phlebotomist using syringes. A 20 mL blood sample was then transferred into blood collection tubes either with or without ethylenediamine tetra acetic acid (EDTA). The EDTA treated blood samples were centrifuged at 3000 × g for 20 minutes at 7° C to obtain plasma samples. Urine samples were also taken prior to the blood sampling. Samples were immediately analyzed by Cobas 8000-Integrated chemistry and immunoassay Electrochemiluminescence-ECLIA platform (Roche, Indianapolis USA), Immulite 2000 XPi-Chemiluminescence immunoassay platform (Siemens, Erlargen, Germany), Urisys 2400-Fully automated Urine FEME analyzer (Roche, Indianapolis, USA) and Cell-Dyn Ruby-Fully automated 5-part hematology analyser (Abbott, Illinois, USA) for clinical biochemistry profiles (as tabulated in Table 5 to Table 11). A comprehensive clinical trial protocol of the study is provided in the Supplementary File. Statistical analysis. Wilcoxon-Signed Test was performed to compare significance of differences between parameters of interest before (baseline, day 1) and after treatment (days 30, and 60) for each treatment. Effects of treatment on parameters of interest were analyzed for their time × treatment interaction, using Two-factor repeated measures analysis of variance (ANOVA) with an interaction term to detect whether there was a significant difference of plasma and urine profiles between OPP and control treatments. If there was any significant time × treatment interaction, the value at the specific day of treatment was extensively compared using Wilcoxon-Signed Rank Test. To increase the stringency of the analysis, Bonferroni correction for multiple testing was applied. Statistical analysis was performed using Statistical Package for Social Sciences (SPSS ® ) for    Table 3. Body weight changes in volunteers (mean ± SD).

Results
All 25 volunteers completed the trial. Table 1 shows the demographic characteristics of the volunteers. The trial was completed with the following baseline characteristics: Means ± SDs: Age, 29.24 ± 4.31 y; body mass index, 22.41 ± 3.99 kg/m 2 . There was no change in body weight of all volunteers (Table 3). During the trial, there was no dropout and none of the volunteers demonstrated any adverse effects (AE) or serious adverse effects (SAE). Sample size or number of volunteers was determined based on calculation of Effect Size (ES). Data from other human studies investigating the antioxidant effects of phenolic compounds on several plasma biochemistry profiles 18-21 were pooled and calculated for their ES. Estimation of sample size was conducted by introducing the ES value into the Altman Normogram. Calculations and statistical analysis for this part were carried out according to previous statistical methods 22,23 . This analysis indicated that the minimum number of volunteers required for this study was 23 per group. Hypothetically, this number would allow investigators to observe any difference between control and treatment groups.
All clinical biochemistry parameters including blood pressure (Table 4), liver function (Table 5), renal function (Table 6), haematology (Table 7), lipid profiles (Table 8), glucose and insulin levels (Table 9), hsCRP, whole blood HbA1C, urine albumin, ACR profile (Table 10) and electrolytes (Table 11) were in the normal range according to standard clinical safety references. There were no significant differences in any of these clinical biochemistry profiles between OPP and control treatments during the whole intervention period.
The OPP dose used in the trial resulted in lower plasma TC ( Fig. 1 and Table 8). After 60 days of OPP supplementation, plasma TC level in the OPP treatment group was significantly lower compared to the control treatment (p = 0.025) based on the Wilcoxon signed rank test. Plasma LDL-C was also significantly lower when volunteers were supplemented with OPP (p = 0.04), (Fig. 2 and Table 8). As several dependent and independent statistical tests were performed in this study, in order to rule out false positive results and increase the stringency of the test, Bonferroni correction was applied. Following the Bonferroni correction, the decrease in TC and LDL-C was deemed to be not significant. There was no significant difference of HDL-C (Fig. 3), TAG and TC/HDL ratio (

Discussion
Dosage of the OPP was determined based on the safety and functionality of a similar intake of GAE from pomegranate juice. Previous investigation demonstrated that pomegranate dietary supplement is safe when ingested by healthy human volunteers in amounts up to 1420 mg/day, providing a total of 870 mg GAE/day 21 . No adverse effects or changes in haematology, serum chemistry, or urinalyses related to the pomegranate dietary supplement consumption were observed during the trial period. In the current trial, volunteers were given a total of 300 mL of OPP with a concentration of 1500 mg/L GAE (150 mL of OPP in the morning session and another 150 mL of OPP in the afternoon session). This provided a total of 450 mg GAE/day to each volunteer for 60 days.
The dose was selected based on translation of the effective dose used in our previous pre-clinical trials on animal models supplemented with OPP where positive effects on clinical biochemistry profiles were demonstra ted 2,11,12,14,16 . Most clinical trials have attempted to deliver large quantities of individual polyphenols to overcome the issues of rapid polyphenol clearance in the human circulation 24 . Supplementation of phenolics in high doses appears to be the only reliable method to achieve measurable concentrations approaching those seen in in vitro and in vivo animal studies, thus enabling investigators to observe any beneficial effects that may occur due to the supplementation.
In most clinical trials investigating the physiological effects of phenolic substances, safety aspects are rarely documented. It is further suggested that detailed safety profiles should be included when findings are reported and published 24 . Following OPP supplementation in our trial, no AE/SAE or changes in haematology, serum and plasma chemistry, as well as urinalyses were observed. Moreover, none of the volunteers showed any allergic reactions towards OPP supplementation. All clinical biochemistry liver profiles were in the normal range according to the standard clinical safety references, indicating no toxic effects from OPP supplementation. In our previous investigations on animals, no toxicity effects from OPP supplementations (up to 2400 mg GAE/L dose) were found, as far as the histology, haematology and clinical biochemistry profiles were concerned 2,11,12,14,16 . Therefore, it can be concluded that OPP at the dose tested in the current trial is generally safe for human consumption.
Introduction of the control treatment arm into the trial was necessary as previously recommended 25,26 . Furthermore, a much longer wash out period (four-weeks) was implemented in the current trial, to ensure the duration was sufficient to avoid any possible effects of the repeated ingestions 25 . Although it has been reported that a one-week wash out period is adequate 27,28 given that most polyphenols are rapidly metabolized and reach maximum concentration within hours 20 , this does not take into account the fact that phenolic extracts including pomegranate supplement and OPP contain other compounds such as fibre, sugars, organic acids and vitamins besides phenolics. A longer washout period is thus advantageous.
Most of the human trials investigating the effects of phenolics and polyphenols were focused on evaluating beneficial effects. Safety issues such as side effects, acute or chronic/long-term effects have rarely been reported in the literature 24 . A primary aim of this trial however, was to ascertain the safety of consuming OPP.
Among all results, we carefully evaluated the response of plasma lipid profiles (TC, LDL-C and HDL-C), glucose and insulin levels towards OPP supplementation, given that OPP had demonstrated positive effects in lowering lipid and diabetic profiles in our previous studies using animal models 11 . The OPP dose utilized in the current trial reduced plasma TC and LDL-C compared to the control treatment. Although the reduction was significant based on the Wilcoxon signed rank test, the application of the Bonferroni correction suggested that the results were not significant. Previous pre-clinical studies demonstrated the ability of OPP in lowering plasma    12,14 . Further clinical tests to specifically confirm the effect of OPP on lipid profiles in humans with raised TC may be necessary. Blood lipid profiles (TAG, TC, LDL-C, HDL-C) have been routinely used as indicators in evaluating the physiological potential of phenolic supplementations in human trials. In most cases, no improvement has been clearly demonstrated in healthy subjects when phenolic substances were supplemented 29 . A previous human intervention trial also found no significant differences in lipid profiles namely TAG, TC, LDL-C, and HDL-C between treatment and control groups following red wine consumption for two weeks 30 . However, the phenolic content used in that trial was much lower (62 mg epicatechin equivalent/day) compared to our trial. Even larger doses such as 852 mg GAE/day of total phenolic intake of cranberry juice for the same period did not significantly alter plasma lipid profiles in healthy population 31 .
Our findings are in agreement with a previous trial where 49 males and 11 females were supplemented with 544 mg/day polyphenol extract powder (dissolved as drinks) for two months 32 . There were no significant differences in serum TAG, TC, HDL-C and LDL-C after consumption of the polyphenols in two different intervention groups. In another human trial, supplementation of 710 mg/day or 1420 mg/day of pomegranate polyphenol extract in the form of capsules for four weeks in 64 healthy subjects also did not significantly improve plasma OPP treatment (n = 25) Control treatment (n = 25)   Table 8. Plasma lipid profiles (mean ± SD). 1 Although time × treatment interaction was significant, p = 0.001 (2-factor repeated measures ANOVA) and TC was significantly lower for OPP compared to control treatment (at day 60), based on Wilcoxon signed rank test (p = 0.025), no significant difference of total cholesterol level was found between OPP and control treatments at day 1, 30 and 60 following Bonferroni correction. 2 Baseline HDL-C level was significantly different between treatments, p = 0.038 (Wilcoxon signed rank test).
The value was further used as covariate in repeated measures. Although the time × treatment analysis was significant, p = 0.011 (2-factor repeated measures ANOVA), no significant difference of HDL-C level was found between treatments at day 30 and 60 based on Wilcoxon signed rank test. Similarly, no significant difference of HDL-C level was found between treatments at day 30 and 60 after Bonferroni correction. 3 Although the time × treatment interaction was significant, p = 0.018 (2-factor repeated measures ANOVA) and LDL-C was significantly lower for OPP compared to control treatment (at day 60) based on Wilcoxon signed rank test (p = 0.04), no significant difference of LDL-C level was found between OPP and control treatments at day 1, 30 and 60 following Bonferroni correction. TAG, TC, HDL-C and LDL-C levels 21 . Supplementation with 120 mg/day of isoflavones through consumption of soy preparation powder for 12 weeks significantly reduced TC, LDL-C and cholesterol/HDL ratio but not HDL-C levels. These effects were however demonstrated in diabetic volunteers 27 . In a later study which applied the same recruitment period and the same type of isoflavone preparation but at a slightly higher dose (132 mg/day), TAG, TC, LDL-C and HDL-C levels were not significantly changed as well 33 . Theoretically as an antioxidant, OPP may provide positive effects in improving the lipid profiles in volunteers with raised TC levels. However, in the current trial, no significant differences in TAG and HDL-C levels between OPP and control treatments were observed. It is important to note that the volunteers were healthy and their lipid profiles were already in the normal range and hence it is not surprising that no significant changes were observed following OPP supplementation.  Table 9. Endocrinology profiles (mean ± SD).

OPP treatment (n = 25) Control treatment (n = 25)
Serum hsCRP (mg/L)  The effect of phenolic supplementations on lipid profiles in humans is rather unique. Several contradictory findings from the literature suggest that recruitment periods, doses and forms of phenolic supplementations, as well as health condition of recruited volunteers affect the results of the supplementations. A short supplementation period of 150 mg/day of polyphenols rich pine bark extract in capsule form for six weeks significantly improved LDL-C and HDL-C profiles in 25 healthy subjects 34 . In comparison to the placebo treatment, supplementation with the same type of polyphenols 34 but at slightly lower dose for a much longer period (125 mg/day) in diabetics with mild hypertension resulted in significantly lower LDL-C levels after 8 and 12 weeks 35 . A significant LDL-C lowering effect was also demonstrated in diabetic volunteers even after four weeks of supplementation with 321 mg/day of flavanol-rich drinks 36 . However, a two-week supplementation period of 800 mg/day polyphenol capsules did not significantly alter blood lipid profiles in 35 healthy male subjects 28 . Even high dose of phenolic substance at longer period did not improve lipid profiles, as previously demonstrated in obese subjects who received 800 mg/day of catechin supplement capsules for eight weeks 37 .
In the current study, following 60 days of OPP supplementation, plasma glucose levels declined gradually from their respective baseline values. However, this observation did not reflect any of the glucose-reducing effects as previously suggested in our animal studies 11,16 , since a similar trend was even demonstrated in the controls. Nevertheless, no significant changes in glucose and insulin levels were observed after 60 days of OPP supplementation. Furthermore, there was no significant difference in plasma glucose at day 30 or 60 between both OPP and control treatments. However, this is not surprising as this trial was with healthy volunteers and reduction in glucose levels was not anticipated. OPP may have elicited a different effect in diabetic or pre-diabetic volunteers.
Most of the glucose lowering effects by polyphenols in healthy populations were demonstrated in the postprandial trial, where subjects were challenged with the oral glucose tolerance test (OGTT) coupled with polyphenol treatments 38,39 . No significant effects were however demonstrated during long-term supplementation trials in healthy populations 19,21,30,37,40 and even in pathological populations 18,27,33,36,40,41 . Similar to other antioxidants, the effects of polyphenol treatments are likely to be influenced by doses, durations and delivery forms of supplementations, as well as the health status of the studied populations. Higher polyphenol treatments such as ≈1500 mg/day for two weeks Figure 1. Total cholesterol level in plasma following OPP and control treatments for 60 days. # Time × treatment interaction was significant, p = 0.001 (2-factor repeated measures ANOVA). Based on Wilcoxon signed rank test, TC was significantly lower for OPP treatment compared to control treatment (at day 60), p = 0.025. However, following Bonferroni correction, no significant difference of total cholesterol level was found between OPP and control treatments at day 1, 30 and 60. Figure 2. LDL-C level following OPP and control treatments for 60 days. # Time × treatment interaction was significant, p = 0.018 (2-factor repeated measures ANOVA). Based on Wilcoxon signed rank test, LDL-C was significantly lower after OPP compared to control treatment (at day 60), p = 0.04. However, following Bonferroni correction, no significant difference of LDL-C level was found between OPP and control treatments at day 1, 30 and 60.
Scientific RePORTS | (2018) 8:8217 | DOI:10.1038/s41598-018-26384-7 significantly reduced plasma glucose level in type 2 diabetic volunteers 42 . However, the actual effects of the polyphenol treatment in the study were not clearly established since the diabetic volunteers were also prescribed with the oral blood glucose-lowering drugs, although the comparison was made with the control treatment 42 . Furthermore, the relationship between blood glucose and insulin levels in diabetic volunteers has not yet been conclusive since some polyphenol treatments resulted in significantly improved insulin profiles but not glucose levels 19 .
Although there are no clear guidelines in determining the upper-tolerable intake of phenolic supplementations, future human trials should establish the effective minimum dose of phenolic intake that could significantly improve desired clinical profiles. Our phase one clinical trial demonstrated the safety and physiological effects of OPP, as supported by the current available data of clinical biochemistry profiles. Supplementation of 450 mg GAE/day OPP is found to be safe for human consumption. Although the results were not significant following Bonferroni correction, the observed decrease in TC and LDL-C suggests that at appropriate doses, OPP may provide several beneficial effects in improving lipid profiles in humans and the current observations justify the need to continue with dose-exploration trials of OPP in future. Figure 3. HDL-C level following OPP and control treatments for 60 days. *Baseline HDL-C level was significantly different between treatments, p = 0.038 (Wilcoxon's signed-rank test). Value was further used as covariates in repeated measures. Although the time × treatment analysis was significant, p = 0.011 (2-factor repeated measures ANOVA), no significant difference of HDL-C level was found between treatments at day 30 and 60 (Wilcoxon signed rank test). Similarly, no significant difference of HDL-C level was found between treatments at day 30 and 60 after Bonferroni correction.