Investigating the effect of herbal antioxidants on the process of colon cancer treatment: A systematic review

Abstract This study is asystematic review conducted in July, 2017. Related studies were searched in scientific databases of Biomed central, PubMed, Sciencedirect, Science of the Web, and Scopus according to the search method of Cochrane with related keywords since the creation of the desired databases. Among the studies, 4 studies did not achieve any of the outcomes, two studies had achieved all the desired outcomes and one study achieved some of the outcomes. Although, the use of antioxidants has reduced the risk of cancer or improved the patient’s condition in all the studies, these effects are not statistically significant generally.


Background
Colorectal cancer is the most common cancer of the digestive tract and the fourth cause of cancer death in the world (Iacopetta, 2002). This cancer is divided into two types: hereditary and nonhereditary. Almost 80% of cancer cases are non-hereditary (Naccarati etal., 2007). Annually, more than 1,200,000 people are diagnosed with this cancer in the world and over 600,000 people die, and this number is increasing (Ferlay etal., 2015). This cancer is ahighly heterogeneous disease caused by genetic and environmental factors and it progresses through the gradual accumulation of genetic and environmental changes leading to the transformation of normal mucosal cells into cancer cells (Binefa etal., 2014

PUBLIC INTEREST STATEMENT
Colorectal cancer is the most common cancer of the digestive tract and the fourth cause of cancer death in the world. More than 70% of cancer cases are related to the lifestyle and diet of individuals. Diet can affect the amount of oxidative stress in individuals and change it. Herbal antioxidants can inhibit oxidative stress by exchanging electrons with free radicals. The authors have put together the data from different articles and made asystemic review to show the effect of herbal antioxidants on the process of colon cancer treatment. They have compared different studies and concluded that although the antioxidants have reduced the risk of cancer or improved patients condition but these affects are not statistically significant in the long-term usage.
environmental changes such as infection, diet and lifestyle which could be involved in relevant cancers (Farrington etal., 2005;Gagnière etal., 2016).
More than 70% of cancer cases are related to the lifestyle and diet of the individuals (Binefa etal., 2014). Diet can affect the amount of oxidative stress in the individuals and change it. Oxidative stress is caused by the imbalance between the production of free radicals and active oxygen species (including superoxide anion, hydroxyl free radical, hydrogen peroxide, etc). There are also part of the antioxidant defense system in animals and humans (Chompoo etal., 2012). In aerobic biological systems, defense mechanisms have been designed to deal with these free radicals and active oxygen species, in order to neutralize or minimize the harmful effects of these invading factors. Some components of the defense system include enzymes (such as superoxide dismutase, glutathione peroxidase, catalase, etc) that are synthesized in the body, but some other components of this system such as vitamin E, beta-carotene and so on must be supplied through diet. Oxidative stress causes destructive effects on macromolecules, including DNA, proteins, and lipids. Herbal antioxidants which are compounds with conjugated double bonds can inhibit oxidative stress by exchanging electrons with free radicals. Some of these compounds are scavengers of free radicals and some repair the damage caused by free radicals (Panchawat etal., 2010;Rice-Evans etal., 1996). Antioxidants collectors of active oxygen species protect the cells from oxidative stress that initiate and direct carcinogenicity through the induction of gene mutation, DNA damage, genomic instability, cell differentiation, and inflammation. Carotenoids (such as beta carotene, precursor of vitamin A, and lycopene), vitamins Cand Ehave antioxidant and anti-inflammatory properties. Therefore, these vitamins and minerals are introduced as antioxidant food that can affect cancer cells and control or treat them (Song etal., 2015). Therefore, interventions effective on oxidative stress can probably help to prevent and treat patients. Therefore, to investigate the effects of interventions, this study was conducted with the aim of systematically reviewing previous studies and the effects of interventions made to identify the effect of herbal antioxidants on the process of colon cancer treatment.

Materials and methods
This study is asystematic review conducted to identify the randomized clinical trials interventions in investigating the effect of herbal antioxidants on the prevention and treatment of colorectal cancer. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) checklist was used to clear how the study was conducted (Moher & Shamseer etal., 2015). The electronic search of related databases was done in July2017.
Studies were selected using the Cochrane search strategy from the scientific databases of Web of Science (Topic (Title, Abstract, Keywords); 1983 to present), PubMed & BioMed Central (Title/ Abstract; 1836 to present), Scopus (Article, Abstract, Keywords; 1836 to present), and Science Direct (Title, Abstract, Keywords; 1823 to present). The desired search strategy was developed by an expert in the field of systematic studies, using the study conducted by Myung etal. (2009). Therefore, awide range of terminology and vocabulary related to the purpose of the study were considered for the search strategy. Also, only clinical trial studies were used since the creation of databases and searching the scientific databases was limited to English language. The following keywords were used for the combined search strategy: "Colorectal Neoplasm" OR "Colorectal Neoplasms" OR "Colorectal cancer" AND Prevent* OR prevention OR diagnosis* OR screening OR therapy* OR therapeutics AND garcinia* OR GIE OR "garciniaindica" OR antioxidant OR "herbal antioxidant" OR "antioxidant" OR "herbal antioxidant" OR "anti-oxidant" OR carotenoid* OR flavonoid OR polyphenelie* OR phenylpropanoid OR anthocyanin OR tannin OR lignin OR "vitamin C" OR "vitamin C" OR "vitamin-C" OR "scorbic acid" OR "vitamin E" OR "tocopherol" OR "catechins" OR polyphenole* OR "phenolic acid" OR vitamin* OR phenol* OR "nonflavonoid phenolic" OR "organic antioxidant" AND effect* OR control* OR evaluation* OR program* AND "randomized controlled trial" OR "cluster-randomized controlled trial" OR "cluster-randomized controlled trials" OR "controlled clinical trial" OR "randomized controlled trials" OR "clinical trial" OR human NOT animal.
Database search was done by only one of the researchers. First, all articles identified from different sources were collected by one of the researchers using the Endnote software. After collecting these articles, repetitive articles were eliminated and this was done by two researchers who independently reviewed all the articles obtained and excluded the articles not related to the subject using the inclusion criteria. Then, the abstracts of the remaining articles were studied independently by the two researchers. Finally, the full texts of the selected articles were investigated by two researchers until the articles that completely matched the inclusion criteria were identified. After extracting all the desired articles based on the objectives of the study, with amore detailed investigation in the next step, articles which cited these remaining studies were found and all sources of the desired articles were investigated. Articles that are related to the purpose of the study were extracted and these articles were carefully investigated independently by two researchers in terms of relevance to the inclusion criteria. In all the stages, the cases of disagreements were resolved through dialogue and eventually by the third researcher's opinion. Ethics and trust were observed in using the sources in all the research stages.
Related studies were selected using systematic search approach from scientific articles, scientific and government reports. The criterion (PICO) which includes four components of Population, Intervention, Comparator and Outcome was used for articles selection (Schardt etal., 2007).
Population: The studies were investigated to systematically review which of them had been conducted on both genders and without age and race/ethnic restrictions on the effect of herbal antioxidants on patients with at least one polyp in the rectum or colon cancer.
Types of intervention: Clinical trials and randomized controlled trials (RCT) on the effect of any herbal antioxidants on the process of colon cancer treatment were included in the study. Descriptive, qualitative, review, structured review, meta-analysis, and semi-experimental studies (before and after controlled and before and after without control) on the effect of herbal antioxidants on the process of colon cancer treatment were excluded. Also, studies investigating the effect of herbal antioxidants on other cancers were excluded. The interventions included providing dietary and pharmaceutical supplements and changes in the diet of the subjects studied.
Comparator/Control: Considering studies where the control group participated in the study in any way.
Types of Outcome: These studies were evaluated based on the results obtained from observation, self-reporting, and biological monitoring.
The results were reported through apositive increase in the effect of herbal antioxidants, significant statistical changes in reducing the risk of polyps or colon cancer, making difference in cancer cells, reducing the number of polyps and so on. Statistical significance was considered at 0.05 for all the study results. Data were extracted from the desired articles by two researchers from September to October, 2017. The output data in the WORD software were author name/year of publication/country, participants' characteristics, interventions, outcomes, and studies' quality notes (Table 1).
In order to determine the quality of the desired articles, two of the trained researchers investigated the studies individually using the tools (EPHPP) related to the Institute of National Collaborating Center for Methods and Tools (National Collaborating Centre for Methods and Tools, 2010). This checklist divided the studies into three groups of powerful, moderate, and weak according to the six bias components in the selection of samples, type of study, confounding, data collection methods, and loss. An instruction was used to investigate the tool and tocopherols.
-C: The volunteers were given supplements of aplacebo.
-Follow-up: The average length of treatment was 10 ± 2 and 16 ± 3 (days ±SEM) for the lycopene and placebo groups, respectively, with arange of 2-65 days.
-In the placebo-treated group, there was asmall nonsignificant increase in lycopene plasma levels.
-The plasma concentration of insulin-like growth factor-I decreased significantly by about 25% after tomato lycopene extract supplementation as compared with the placebotreated group (P < 0.05).
-No significant change was observed in insulin-like growth factor-I-binding protein-3 or insulin-like growth factor-II, whereas the insulin-like growth factor-I/insulin-like growth factor-I binding protein-3 molar ratio decreased significantly (P < 0.05). control disagreements between the researchers were resolved through discussion. Also, the internal reliability for using tools was calculated using Kappa coefficient.

Results
The initial search led to the finding of 3113 studies after eliminating repetitive studies. Atotal of 800 studies were excluded because they were not in line with the study objectives and inclusion criteria. Then, 515 abstracts of these studies were carefully investigated independently by two researchers. Atotal of 360 abstracts of the studies were excluded because they were not in line with the study objectives and inclusion criteria. 155 texts of the studies were carefully investigated independently by two researchers. From these 155 texts of the studies, 25 texts were from studies which are common in terms of the desired factor and response and 7 studies were completely common in terms of the type of study, the desired factor, and the response (colon cancer). 18 texts of the papers were excluded because they are semi-experimental studies, systematic reviews and studies of merely intervention on cancer cells. Finally, seven randomized trial studies were identified for evaluation in this study (Figure 1).
In evaluating papers based on race/ethnicity of participants, most studies used local communities to evaluate the antioxidant intervention (Albanes etal., 2000;Cascinu etal., 2000;McKeown-Eyssen etal., 1988;Mobarhan etal., 1994;Walfisch etal., 2007, but, in two studies enrolled participants with different races (white, black, Hispanic, etc.) (Emmons etal., 2005;Wactawski-Wende etal., 2006). The effect of the intervention in studies with local populations of Italy, Finland and Canada was not significant, but in the US study, differences were observed in the response of the intervention and control groups without race effect. In the study of the Israeli population despite some appropriate results, however, the effect of the intervention was insignificant.
Papers assessed according to amount of antioxidant and results indicated that participates in four studies with respectively twice daily 500mg calcium with 200 IU of vitamin D, daily 30,000 IU of axerophtol palmitate (vitamin A) plus 1 g ascorbic acid (vitamin C) plus 70mg of dl-alphatocopherol acetate (vitamin E) and 2 g natural calcium, daily α-tocopherol and β-carotene, and 400mg each of vitamins Cand Eevery 4 months did not achieve any of the desired outcomes (Albanes etal., 2000;Cascinu etal., 2000;McKeown-Eyssen etal., 1988;Wactawski-Wende etal., 2006). In two studies, that the interventions were grounded in the Social Cognitive Theory for 4 weeks and 90 capsules each containing 30mg β-carotene (BC) for 90 days, there were asignificant difference between the groups (Emmons etal., 2005;Mobarhan etal., 1994). In one study, the volunteers were given supplements of anatural source of lycopene some results are responsible between groups og intervention (Walfisch etal., 2007).
In terms of investigating the quality of studies and bias, none of the studies had poor or moderate quality, but all were powerful in terms of investigating quality (Albanes etal., 2000;Cascinu etal., 2000;Emmons etal., 2005;McKeown-Eyssen etal., 1988;Mobarhan etal., 1994;Wactawski-Wende etal., 2006;Walfisch etal., 2007). Kappa coefficient was calculated to investigate the internal consistency between the two evaluators of the EPHPP components. Kappa coefficient was in good agreement for withdrawals and selection bias (k = 0.60 to 0.80) and was in very good agreement for others (k = 0.80 to 1.00) ( Table 2).

Discussion
This systematic review investigated randomized clinical trials interventions which had investigated the effect of herbal antioxidants on the prevention and treatment of colorectal cancer. After reviewing the studies by the researchers, finally seven studies were identified for evaluation in this study. The seven studies were randomized trials. The sample size of most of the studies was small and the duration of follow up, regarding the impact of interventions, was mainly long. The results of these studies were mainly determined using two methods, such as laboratory measurements and self-reporting in two studies and self-reporting in one study. The intervention approach used in all the studies was by using capsules or pills containing herbal antioxidants or combining drug with herbal diet. From the studies investigated, four studies did not achieve any of the desired outcomes, two achieved all the desired outcomes, and one achieved some of the outcomes. All the studies were powerful in terms of the quality of the investigation.
According to the systematic review conducted by Bjelakovic etal. (Bjelakovic etal., 2008) to investigate the effects of antioxidants used in preventing gastrointestinal cancers, finally, 20 studies were achieved and all of them were randomized trials. Also in another study, the results showed that all the studies that investigated antioxidants to prevent gastrointestinal cancer were studies that used randomized trials (Bjelakovic etal., 2004). Therefore, the results of the two studies mentioned were consistent with the results of this study. From the studies investigated, most of them were published after 2000. In the meta-analysis study of Liu etal., which was conducted to evaluate the relationship between the consumption of multivitamins and the incidence of colon cancer, most of these studies were published after 2000, and as such were consistent with the results of this study (Liu etal., 2015). In this work, most of the studies reviewed were conducted in developed countries such as the United States, and this is clear from the previously reviewed studies on the effect of antioxidants consumption in preventing and treating colon cancer (Bjelakovic etal., 2008(Bjelakovic etal., , 2004Liu etal., 2015). This can be considered probably due to the access of these populations to asufficient and effective level of vitamins and herbal antioxidants and paying attention to the health problems of this population is also an important research priority in these countries. But the situation in developing countries is that this population is given less consideration by the researchers, because of the health problems in other groups.
The sample size of most of the studies was small and the number of participants in four out of seven studies was less than 200 individuals. Also, the duration of follow up, the impact of interventions was mainly long. In other studies, similar results were obtained (Bjelakovic etal., 2008(Bjelakovic etal., , 2004. The reason for this can be the nature of design and follow up in the randomized trial studies which dealt with the real patients and required long-term follow-up in order to achieve minimum effectiveness in intervention.
Outcomes in the studies were determined using laboratory measurements. The measurement of outcomes was based on self-reporting in astudy, and two methods of self-reporting and laboratory measurements were used to investigate the outcomes in two studies. From the papers investigated, four studies did not achieve any of the desired outcomes, thus, no statistically significant difference was observed between the control and intervention groups. Two studies achieved all the desired outcomes, therefore, there was asignificant difference between the groups. But only one study achieved some of the outcomes. Similar review studies reported similar results (Bjelakovic etal., 2008(Bjelakovic etal., , 2004. The different effects of antioxidants in the groups with different risks can be considered as the reason why antioxidants are reported effective in some populations and ineffective in some others.
In all studies, effect of intervention between groups assessed by controlling on some confounder such as the age, race, and amount of antioxidant.
The intervention approach used in all the studies was by using capsules or pills containing herbal antioxidants or combining drug with herbal diet. The reason for the greater use of pills and capsules in interventions can be considered by matching and controlling the amount of intervention in the individuals as compared to the dietary intervention. Also, in dietary interventions, there is greater probability that the subjects of the study will not observe the study protocol. Only few studies had used other approaches and it can be argued that these interventions were not well received due to the cost, time-consumed, and long-term outcomes. In similar studies, the intervention approach was often similar to the approach of this study (Bjelakovic etal., 2008(Bjelakovic etal., , 2004Liu etal., 2015).
The findings of the studies were determined by laboratory measurements. The measurement of the outcomes was done based on self-reporting only in astudy and two methods of self-reporting and laboratory measurements were used to investigate the outcomes in two studies. Self-reporting requires less time, cost, and facilities than the laboratory methods. But the accuracy of the conclusions in the laboratory method is more than the self-reporting methods. Also, most studies that have used the laboratory methods had been conducted in developed countries and the reason can be the high cost of these experiments and the need for expert force.
The strengths of this study can be considered by looking at aparticular outcome, colorectal cancer, instead of considering several outcomes. Also, more relevant databases were investigated using comprehensive search strategy than other studies. The weaknesses of this study were considered as investigating only English-language studies and the lack of use of other language studies.

Conclusions
There are significant differences in the study for using self-expression method to investigate the outcomes, but no significant difference was observed in most of the studies that have used the laboratory method. Although consuming antioxidants has reduced the risk of cancer or improved the patient's condition in all the studies, but these effects are not significant in most studies. Due to the accuracy of the laboratory methods as compared to the self-expression method, it can be concluded that consuming antioxidants reduces the risk of colon cancer among healthy people and increases the probability of improvement in the community of patients with colon cancer, but there is no statistically significant difference between these interventions and the control group (without treatment) in the long term.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.