The cancer-preventive effects of fruit and vegetables have been intensively studied for over 30 years. According to early epidemiological research, there is an inverse association between the intake of fruit and vegetables and the risk of developing cancers [1
]; however, recent epidemiological studies indicate that for common cancers such as breast, colorectal, lung and prostate cancer, small or no association between total fruit and vegetable consumption and cancer risk exists [4
]. In early studies, the cancer-preventive effects of fruit and vegetables were mainly ascribed to their content of minerals, fibers and antioxidants. However, fruit and vegetables have very varied composition of nutrients and other constituents, and therefore it is still possible that there are cancer-preventive effects to be identified for the intake of individual fruit and vegetables. In particular, vegetables contain a wide variety of compounds with many interesting bioactivities that are unrelated to antioxidant effects. Some of these bioactive constituents may contribute to the potential cancer-preventive effects of vegetables and thus may help to obtain a deeper understanding of their health promoting effects in general [6
]. This is, for example, the case if we look at apiaceous vegetables such as carrots that are consumed worldwide and in particular in North America and in European countries.
Carrots are rich in carotenoid antioxidants such as β-and α-carotene and epidemiological studies have shown that a high content of β-carotene in the blood is correlated with a low incidence of cancers and other diseases [9
]. In most European countries and North America, more than 50% of the β-carotene intake is provided by carrots, although in these regions of the world, carrot consumption is better correlated with the intake of α-carotene [12
]. Several studies have also found stronger negative correlations of developing cancer and in particular lung cancer with intake of α-carotene rather than β-carotene [13
]. Thus, it is widely accepted that carrots play a central role as a protecting vegetable against development of cancer, which is supported by recent meta-analysis studies on carrot consumption in relation to the development of breast, gastric, lung and prostate cancer [15
]. The cancer-preventive effect of this vegetable has mainly been explained by its high contents of carotenoids; however, intervention studies have shown that supplementation with carotenoids does not protect against development of this disease [10
]. Hence, α-carotene and β-carotene may be biomarkers for the intake of other bioactive constituents in carrots with cancer-preventive effects. Such potential anticancer compounds are indeed present in carrots and includes mainly phenylpropanoids [20
] and polyacetylenic oxylipins [21
] of which the latter type of bioactive constituents are the most studied both in vitro and in vivo.
The major polyacetylenic oxylipins in carrots are falcarinol (FaOH) and falcarindiol (FaDOH) and carrots are the major dietary source of this type of bioactive compounds, although they are also present in other apiaceous vegetables such as celery, celeriac, fennel, and parsley [28
]. FaOH and FaDOH have received considerable attention in recent years due to their cytotoxic and anti-inflammatory activities in vitro [21
] and, recently, the anti-neoplastic effects of these polyacetylenic oxylipins have been demonstrated in cancer-primed rat models for colorectal cancer (CRC) [21
CRC is the third cause of cancer-related death in developed countries and is probably associated with a modern lifestyle typified by limited physical activity, alcohol consumption and dietary changes [35
]. CRC is a metastatic type of cancer that is developed in a multistep process, from normal epithelial cells via inflammation to aberrant crypt foci and progressive adenoma stages, to carcinomas [37
]. To reduce the incidence and consequences of CRC, effective prevention and treatment strategies need to be identified. Due to the long precancerous stage of this disease, dietary intervention may exert favorable effects on polyp formation and/or inhibition of adenomas transformation to CRC. Recent findings indicate that long-term consumption of a diet rich in vegetables may prevent the development of CRC [39
]. Chronic inflammation appears to play key role in the development of CRC. Cyclooxygenase 2 (COX-2) levels are low in normal tissue but are rapidly induced as an early response to growth factors, cytokines and tumor promoters associated with inflammation, abnormal proliferation, angiogenesis, invasion, and metastasis, and the existence of an association between CRC and COX-2 overexpression appears to be established [41
]. Studies on the anti-neoplastic effects of FaOH and FaDOH in tumors of cancer-primed rats for CRC have demonstrated that these polyacetylenic oxylipins inhibit pro-inflammatory and transcription factor biomarkers for inflammation and cancer such as COX-2, interleukin 6 (IL-6), tumor necrosis factor alfa (TNF-α) and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) [21
]. Hence, carrots contain bioactive compounds that target CRC development and thus carrot intake could be one of the most important dietary measures for the prevention of CRC. Polyacetylenic oxylipins are, however, sensitive to heat, light and oxidation, and previous investigations of blanched or cooked carrots have shown that the content of these bioactive compounds may be reduced up to 70% by thermal processing [44
]. Furthermore, carrot juice represents another source of polyacetylenic oxylipins, but the content of these bioactive compounds is also reduced in carrot juice compared to raw carrots due to their low water solubility and possible thermal treatment as well as pH of the carrot juice [46
]. On the other hand, the availability of these compounds may be increased in processed vegetables due to alteration of the structure and digestibility of the food thus increasing their bioavailability, which to some extent can offset the cancer-preventive effect of raw versus processed carrots. This is consistent with the fact that no major differences in the cancer-preventive effects of raw versus cooked vegetables, including carrots, have been observed [47
However, more specific studies of the effect in a human population are still missing in order to conclude whether carrot intake reduces risk of CRC and what quantity of carrot intake is sufficient. The aim of this study was to investigate the risk of being diagnosed with CRC as predicted by intake of carrots in a large Danish study population with a long follow-up.
Out of 160,725 invited individuals, 57,053 participated in the Diet, Cancer and Health study. In total, 585 individuals were excluded due to previous CRC diagnosis and 593 individuals were excluded due to missing information for covariates. Complete information was obtained for 55,875 individuals who were eligible for analysis. In total, 1889 (3.38%) were diagnosed with CRC during follow-up (Figure 1
). Follow-up time varied from 3 to 8438 days, with a mean of 6845 days.
Investigation of the 11 groups of self-reported daily intake of raw carrot and register-based incidence of CRC showed a pattern (adjusted for age and gender), in which groups reporting an intake less than 32 g/day had an insignificant decrease in risk between 0.88 and 0.91 (hazard ratio (HR)). Groups above 32 g/day had significantly decreased risk between 0.68 and 0.81 (HR), although insignificant in subgroups sized below 1300.
At baseline, 7916 individuals reported never eating raw carrots during the previous 12 months. In total, 31,545 had a calculated consumption of less than 32 g/day and 16,414 individuals had a calculated consumption higher than 32 g/day. The incidence of CRC was 3.9% for those eating no raw carrots, 3.5% for those eating less than 32 g/day and 2.9% for those eating more than 32 g/day. X2
-tests showed significant differences in the incidence of CRC according to raw carrot intake, gender, age groups, NSAID intake, BMI, and alcohol intake. No significant differences were seen according to previous cerebral or coronary artery thrombosis, MET score, all other vegetable intake or intake of other root vegetables (Table 1
3.1. Comparison of High, Low and No Intake of Carrots
The cumulative incidence proportion of CRC was higher with a decreasing intake of raw carrots. The cumulative incidence proportions increased throughout follow-up for all three exposure groups (Figure 2
The univariate cox proportional hazard regression model (model I) showed significant decreases in risk of CRC in those with an intake less than 32 g/day (HR 0.82 CI95% 0.72; 0.93) and in those with an intake over 32 g/day (HR 0.66 CI95% 0.57; 0.76), compared to those with no intake of raw carrots. After adjusting for gender and age (model II) the risk differences were decreased and a significant difference were only seen in subgroup eating more than 32 g/day (HR 0.79 CI95% 0.68; 0.91). The subsequent adjustments (model III through V) had only minimal effect on the risk estimates for raw carrot intake. This resulted in a significant difference in risk for subgroup eating more than 32 g/day (HR 0.83 CI95% 0.71; 0.98), and an insignificant difference in risk for subgroup eating less than 32 g/day (HR 0.93 CI95% 0.82; 1.06), compared to subgroup eating no raw carrot, after full model adjustments (Figure 3
3.2. Sensitivity Analyses
Sensitivity analyses showed that carrot juice, prepared/cooked carrots or total carrot intakes were not statistical significantly associated with incidence of CRC. Analyses of possible interactions between raw carrot intake and each of the covariates from the model V did not identify any significant interactions. Sensitivity analysis conducted using age as time resulted in similar HRs for raw carrot intake >32 g/day as did the main analyses. Risks of CRC were statistically significant at 0.71, 0.80, 0.80, 0.85 and 0.84 respectively in model I through V for those with an intake >32 g/day compared to those eating no raw carrot. Examination of Schoenfeld residuals verified the proportional hazards assumption.
In this prospective cohort study, a population of 57,053 Danes was used to explore the association between dietary carrot intake and risk of CRC. The results showed that high carrot intake corresponding to >32 g raw carrot per day was associated with a 17% decreased risk of CRC, whereas an insignificant difference in risk of CRC was observed for those eating less than 32 g raw carrot per day, compared to those eating no raw carrot. To our knowledge, this is the first study evaluating the direct relationship between carrot intake and the incidence of CRC, although in a previous study using data from two case-control studies that included 1225 cases of CRC, it was shown that raw vegetables and in particular raw carrots caused risk reductions of 20% for CRC [60
]. A preventive role of carrot in the development of CRC is plausible because carrots is the main dietary source of the bioactive polyacetylenic oxylipins FaOH and FaDOH, which have demonstrated anti-proliferative effects in cancer cells [25
] and anti-neoplastic effects in CRC-primed rats in a dose-dependent relationship [21
]. The preventive role of FaOH and FaDOH in relation to CRC is probably linked to their alkylating properties leading to covalent alkylation and inhibition of pro-inflammatory markers, enzymes, and inflammatory transcription factors that play an important role in cancer development and in particular CRC [21
]. According to the preclinical rat studies, a cancer-preventive dose of polyacetylenes can be obtained, at an intake above 30 g of raw carrots per day, which furthermore substantiates the cancer-preventive effects of carrots as demonstrated in this cohort study.
Based on the numerous in vitro studies on the anti-proliferative effects of FaOH and FaDOH on cancer cells, it is clear that FaOH is more cytotoxic than FaDOH [21
]. In addition, it has been shown that FaOH inhibits in vitro, the growth of the human epithelial colorectal adeocarcinoma cell line Caco-2, and that this effect is synergistically enhanced when combined with FaDOH [25
]. In the recent studies on the anti-neoplastic effects of FaOH and FaDOH in CRC-primed rats the ratio of these polyacetylenes were 1:1 in the tested rat diets but other ratios may have resulted in even higher or lower CRC-preventive effects [21
]. Thus, not only the concentration of FaOH and FaDOH in carrots may be important for the cancer-preventive effects of carrots but also the ratio between these polyacetylenes.
Factors that have an influence on the content of polyacetylenic oxylipins in consumed carrots, includes besides processing as mentioned in the introduction also storage and location of cultivation and not at least genotype. The cropping system on the other hand, i.e., organic versus non-organic cultivation does not affect the content of these bioactive constituents [45
]. Consequently, there is a potential to optimize the content of polyacetylenic oxylipins in carrots and thus to increase the possible cancer-preventive effect of carrots and carrot products. There are considerable differences in the contents and ratio of FaOH and FaDOH between different carrot cultivars available in Denmark and on the marked in general. This complicate the picture in relation to the preventive effect of carrot intake on CRC but also show that there is a huge potential of optimizing the intake of FaOH and FaDOH in selecting carrot cultivars for human consumption with an optimal composition profile of these polyacetylenic oxylipins. Therefore, choosing the right carrot cultivar could increase the preventive effect on CRC. Bitterness is considered as an undesirable taste of carrots, which is primarily due to the content of FaDOH and di-caffeic acids that are mainly present in the peel and outer-layers of carrots [62
]. Therefore, raw carrots are usually peeled before intake. Peeling not only changes the composition profile of FaOH and FaDOH of raw peeled carrots but also has an effect on the intake of these polyacetylenes because their concentration are higher in the peel per g fresh weight compared to the rest of the carrot, and this is particular the case for FaDOH. Thus, an optimal CRC-preventive effect of carrots not only depends on the cultivar but also how the carrots are processed before consumption.
Other apiaceous vegetables, such as celery, celeriac, fennel, parsley, and parsnip, do also contain FaOH and FaDOH and for some of these vegetables the concentrations of FaOH and FaDOH in the edible parts of these vegetables are even higher than in carrots [28
]. However, these vegetables seems not to contribute significantly to the overall intake of polyacetylenes mainly because they are consumed in much lower quantities compared to carrots and because the root vegetables such as celeriac and parsnip are often subjected to thermal processing resulting in considerably losses of polyacetylenes as is the case for carrots. Intake of all other vegetables, including apiaceous vegetables, were found not to have any significant impact on risk of CRC because no other root vegetables nor all other vegetables were associated with incidence of CRC as demonstrated in model V (Figure 3
). Hence, previous studies reporting lower cancer incidence with higher vegetable intake, may partially be due to the intake of carrots.
The study is based on self-reported 1 year recall of different food intakes and this is usually correlated with some bias. First of all, the self-reported recall will result in some bias as participants will have difficulties in remembering food intake as well as over-estimation of healthy food components [63
]. Furthermore, carrot eaters may have a healthier behavior in general, although the adjustments for MET, other vegetable intake, smoking and alcohol intake did not affect the hazard ratios notably. If participants’ self-reported raw carrot intake is either randomly over- and underestimating actual intake, or systematically overestimating, the decrease in risk of CRC correlated with raw carrot intake is probably underestimated. Further, the true decrease in risk of developing CRC would probably be even greater if the type and handling of carrots were controlled.
In the food frequency questionnaire, the participants reported intake of raw carrots per week and in our study, effect on incidence of CRC was seen in intake as low as 2–4 carrots corresponding to >32 g raw carrot each day, i.e., half of a small-sized raw carrot each day. In the present study, the carrot cultivars were not reported. However, if we look at the contents of polyacetylenes in an average carrot, this will not give 24 h protection of the epithelial cells and so the effect of the carrots seems to be long term. The halftime of FaOH and FaDOH has been shown to be approximately 5–6 h in the blood circulation after intake of carrot juice [30
Prepared/cooked carrots, carrot juice and total carrot intake were not statistical significantly associated with incidence of CRC as expected even though this may increase the bioavailability of bioactive polyacetylenic oxylipins. This is probably due to a great loss of FaOH and FaDOH in preparation or preserving methods as described in the introduction.
Participants differed from non-participants, as the non-participants as a whole had a lower socio-economic status than participants [48
]. This would mean that the results may not be transferrable to the general population, although the possibility of polyacetylenic oxylipins affecting socio-economic strata in different ways is considered unlikely. This is also supported by the fact that no interactions were found between raw carrot intake and any covariate, even though health-affecting behaviors are often associated with socio-economic status.