In vitro studies

The focus on single plant components in the comparison of crops from organic and conventional production, as discussed further below, disregards the fact that compounds in food do not exist and act separately, but in their natural context [49]. In vitro studies of effects of entire foods in biological systems such as cell lines can therefore potentially point at effects that cannot be predicted from chemical analyses of foods, although a limitation is that most cells in humans are not in direct contact with food or food extracts.

Two studies have investigated the effect of organic and conventional crop cultivation on cancer cell lines, both using crops produced under well-documented agricultural practices and with several agricultural and biological replicates. In the first study extracts from organically grown strawberries exhibited stronger antiproliferative activity against one colon and one breast cancer cell line, compared to the conventionally produced strawberries [50]. In the second study [51] the extracts of organic naturally fermented beetroot juices induced lower levels of early apoptosis and higher levels of late apoptosis and necrosis in a gastric cancer cell line, compared to the conventional extracts. Both studies thus demonstrated notable differences in the biological activity of organic vs. conventionally produced crop extracts in vitro, which should inspire further research. However, neither of these studies allows for the distinction of a selective antiproliferative effect on cancer cells, and general cell toxicity. Therefore it cannot be determined which of the organic or conventional food extracts, if any, had the preferable biological activity in terms of human health.

Animal studies of health effects

Considering the difficulties of performing long-term dietary intervention studies in humans, animal studies offer some potential of studying long-term health effects of foods in vivo. However, extrapolation of the results from animal studies to humans is not straight-forward. Studies in this field started almost 100 years ago. A review of a large number of studies [52] concluded that positive effects of organic feed on animal health are possible, but further research is necessary to confirm these findings. Here we focus on the main health aspects.

In one of the best-designed animal studies, the second generation chickens receiving the conventionally grown feed demonstrated a faster growth rate. However, after an immune challenge, chickens receiving organic feed recovered more quickly [53]. This resistance to the challenge has been interpreted as a sign of better health [54, 55].

In one carefully conducted crop production experiment, followed by a rat feeding trial, the production system had an apparent effect on plasma-IgG concentrations but not on other markers of nutritional or immune status [56]. A two-generational rat study based on feed grown in a factorial design (fertilisation x plant protection) of organic and conventional practices revealed that the production system had an effect on several physiological, endocrine and immune parameters in the offspring [57]. Most of the effects identified were related to the fertilisation regimen. None of these studies found that any of the feed production systems was more supportive of animal health.

Several other studies, mostly in rats, have reported some effect of the feed production system on immune system parameters [57–60]. However, the direct relevance of these findings for human health is uncertain.

Collectively, in vitro and animal studies have demonstrated that the crop production system does have an impact on certain aspects of cell life, the immune system, and overall growth and development. However, the direct relevance of these findings for human health is unclear. On the other hand, these studies may provide plausibility to potential effects of conventional and organic foods on human health. Still, most of the outcomes observed in animal studies have not been examined in humans so far.

Plant protection in organic and conventional agriculture

Most of the pesticides approved for organic agriculture are of comparatively low toxicological concern for consumers because they are not associated with any identified toxicity (e.g. spearmint oil, quartz sand), because they are part of a normal diet or constitute human nutrients (e.g. iron, potassium bicarbonate, rapeseed oil) or because they are approved for use in insect traps only and therefore have a negligible risk of entering the food chain (i.e. the synthetic pyrethroids lambda-cyhalothrin and deltamethrin, and pheromones). Two notable exceptions are the pyrethrins and copper. Pyrethrins, a plant extract from Chrysanthemum cinerariaefolium, share the same mechanism of action as the synthetic pyrethroid insecticides, but are less stable. Copper is an essential nutrient for plants, animals and humans, although toxic at high intakes and of ecotoxicological concern due to toxicity to aquatic organisms.

Plant protection practices developed in and for organic agriculture may be of benefit to the entire agricultural system [67–70]. This is of specific value for the transition towards sustainable use of pesticides in the EU, which has a strong emphasis on non-chemical plant protection measures including prevention and biological agents [63, 64]. Further, steam treatment of cereal seeds for the prevention of fungal diseases (http://thermoseed.se/) has been developed driven by the needs of organic agriculture as an alternative to chemical seed treatments [71, 72]. These methods are now also being marketed for conventional agriculture, specifically for integrated pest management (IPM) [73].

Pesticide use – Exposure of consumers and producers

One main advantage of organic food production is the restricted use of synthetic pesticides [5, 6], which leads to low residue levels in foods and thus lower pesticide exposure for consumers. It also reduces the occupational exposure of farm workers to pesticides and drift exposures of rural populations. On average over the last three available years, EFSA reports pesticide residues below Maximum Residue Levels (MRL) in 43.7% of all and 13.8% of organic food samples. MRLs reflect the approved use of a pesticide rather than the toxicological relevance of the residue. There are no separate MRLs for organic products. A total of 2.8% of all and 0.9% of organic samples exceeded the MRL, which may be due to high residue levels or due to low levels but unapproved use of a particular pesticide on a particular crop [74–76]. Of higher toxicological relevance are risk assessments, i.e. expected exposure in relation to toxicological reference values. On average 1.5% of the samples were calculated to exceed the acute reference dose (ARfD) for any of the considered dietary scenarios, with the organophosphate chlorpyrifos accounting for approximately half of these cases and azole fungicides (imazalil, prochloraz, and thiabendazole) for approximately 15%. None (0%) of the organic samples exceeded the ARfD [74]. Residues of more than one pesticide were found in approximately 25% of the samples but calculations of cumulative risks were not included in the reports [74–76].

The only cumulative chronic risk assessment comparing organic and conventional products known to us has been performed in Sweden. Using the hazard index (HI) method [77], adults consuming 500 g of fruit, vegetables and berries per day in average proportions had a calculated HI of 0.15, 0.021 and 0.0003, under the assumption of imported conventional, domestic conventional, and organic products, respectively [78]. This indicates an at least 70 times lower exposure weighted by toxicity for a diet based on organic foods. There are several routes by which pesticides not approved for use in organic agriculture may contaminate organic products, including spray drift or volatilisation from neighbouring fields, fraudulent use, contamination during transport and storage in vessels or storages where previously conventional products have been contained, and mislabelling by intention or mistake. Overall, however, current systems for the certification and control of organic products ensure a low level of pesticide contamination as indicated by chronic and acute risks above, although they still can be improved [79].

The general population’s exposure to several pesticides can be measured by analysing blood and urine samples, as is routinely done in the US [80] although not yet in Europe. However, a few scattered European studies from France [81–83], Germany [84], the Netherlands [85], Spain [86], Belgium [87], Poland [88] and Denmark [89] have shown that EU citizens are commonly exposed to organophosphate and pyrethroid insecticides. A general observation has been higher urinary concentrations of pesticide metabolites in children compared to adults, most likely reflecting children’s higher food intake in relation to body weight and maybe also more exposure-prone behaviours. The urinary concentrations of generic metabolites of organophosphates (dialkyl phosphates, DAPs) and pyrethroids (3-phenoxybenzoic acid, 3-PBA) found in most of the European studies were similar to or higher than in the US studies. Although urinary metabolite concentration might overestimate the exposure to the parent compounds, due to ingestion of preformed metabolites in food items, several studies have reported associations between urinary metabolite concentrations and neurobehavioral deficits as described below. Besides, the metabolites are not always less toxic than the parent compounds [90].

For the general population, pesticide residues in food constitute the main source of exposure for the general population. This has been illustrated in intervention studies where the urinary excretion of pesticides was markedly reduced after 1 week of limiting consumption to organic food [91–93]. Similar conclusions emerged from studies investigating associations between urinary concentrations of pesticides and questionnaire information on food intake, frequency of different foodstuffs and organic food choices. Thus a high intake of fruit and vegetables is positively correlated with pesticide excretion [94], and frequent consumption of organic produce is associated with lower urinary pesticide concentration [95].

Pesticide exposure and health effects

The regulatory risk assessment of pesticides currently practised in the EU is comprehensive, as a large number of toxicological effects are addressed in animal and other experimental studies. Nonetheless, there are concerns that this risk assessment is inadequate at addressing mixed exposures, specifically for carcinogenic effects [96] as well as endocrine-disrupting effects [97, 98] and neurotoxicity [99]. Furthermore, there are concerns that test protocols lag behind independent science [100], studies from independent science are not fully considered [101] and data gaps are accepted too readily [102]. These concerns primarily relate to effects of chronic exposure and to chronic effects of acute exposure, which are generally more difficult to discover than acute effects. Most studies rely on urinary excretion of pesticide metabolites and a common assumption is that the subjects were exposed to the parent chemicals, rather than the metabolites.

The overall health benefits of high fruit and vegetable consumption are well documented [31, 35]. However, as recently indicated for effects on semen quality [103], these benefits might be compromised by the adverse effects of pesticide residues. When benefits are offset by a contaminant, a situation of inverse confounding occurs, which may be very difficult to adjust for [104]. The potential negative effects of dietary pesticide residues on consumer health should of course not be used as an argument for reducing fruit and vegetable consumption. Neither should nutrient contents be used to justify exposures to pesticides. Exposures related to the production of conventional crops (i.e. occupational or drift exposure from spraying) have been related to an increased risk of some diseases including Parkinson’s disease [105–107], type 2 diabetes [108, 109] and certain types of cancers including non-Hodgkin lymphoma [110] and childhood leukaemia or lymphomas, e.g. after occupational exposure during pregnancy [105, 111] or residential use of pesticides during pregnancy [105, 112] or childhood [113]. To which extent these findings also relate to exposures from pesticide residues in food is unclear. However, foetal life and early childhood are especially vulnerable periods for exposure to neurotoxicants and endocrine disruptors. Even brief occupational exposure during the first weeks of pregnancy, before women know they are pregnant, have been related to adverse long-lasting effects on their children’s growth, brain functions and sexual development, in a Danish study on greenhouse worker’s children [114–118].

In order to assess the potential health risk for consumers associated with exposure to dietary pesticides, reliance on epidemiological studies of sensitive health outcomes and their links to exposure measures is needed. Such studies are complicated both by difficult exposure assessment and the necessary long-term follow-up. The main focus so far has been on cognitive deficits in children in relation to their mother’s exposure level to organophosphate insecticides during pregnancy. This line of research is highly appropriate given the known neurotoxicity of many pesticides in laboratory animal models [99] and the substantial vulnerability of the human brain during early development [119].

Most of the human studies have been carried out in the US and have focused on assessing brain functions in children in relation to prenatal organophosphate exposure. In a longitudinal birth cohort study among farmworkers in California (the CHAMACOS cohort), maternal urinary concentrations of organophosphate metabolites in pregnancy were associated with abnormal reflexes in neonates [120], adverse mental development at 2 years of age [121], attention problems at three and a half and 5 years [122], and poorer intellectual development at 7 years [123]. In accordance with this, a birth cohort study from New York reported impaired cognitive development at ages 12 and 24 months and 6 – 9 years related to maternal urine concentrations of organophosphates in pregnancy [124]. In another New York inner-city birth cohort, the concentration of the organophosphate chlorpyrifos in umbilical cord blood was associated with delayed psychomotor and mental development in children in the first 7 years of life [125], poorer working memory and full-scale IQ at 7 years of age [126], structural changes, including decreased cortical thickness, in the brain of the children at school age [127], and mild to moderate tremor in the arms at 11 years of age [128]. Based on these and similar studies, chlorpyrifos has recently been categorised as a human developmental neurotoxicant [129]. Recent reviews of neurodevelopmental effects of organophosphate insecticides in humans conclude that exposure during pregnancy – at levels commonly found in the general population – likely have negative effects on children’s neurodevelopment [130–132]. In agreement with this conclusion, organophosphate pesticides considered to cause endocrine disruption contribute the largest annual health cost within the EU due to human exposures to such compounds, and these costs are primarily due to neurodevelopmental toxicity, as discussed below.

Since growth and functional development of the human brain continues during childhood, the postnatal period is also assumed to be vulnerable to neurotoxic exposures [119]. Accordingly, five-year-old children from the CHAMACOS cohort had higher risk scores for development of attention deficit hyperactive disorder (ADHD) if their urine concentration of organophosphate metabolites was elevated [122]. Based on cross-sectional data from the NHANES data base, the risk of developing ADHD increases by 55% for a ten-fold increase in the urinary concentration of organophosphate metabolites in children aged 8 to 15 years [133]. Also based on the NHANES data, children with detectable concentrations of pyrethroids in their urine are twice as likely to have ADHD compared with those below the detection limit [134]. In addition, associations between urinary concentrations of pyrethroid metabolites in children and parent-reported learning disabilities, ADHD or other behavioural problems in the children have recently been reported in studies from the US and Canada [135, 136].

So far only few prospective studies from the EU addressing associations between urinary levels of pesticides and neurodevelopment in children from the general population have been published. Three studies are based on the PELAGIE cohort in France and present results for organophosphates and pyrethroids respectively [81, 82, 137]. While no adverse effects on cognitive function in six-year-old children were related to maternal urine concentrations of organophosphates during pregnancy, the concentration of pyrethroid metabolites was associated with internalising difficulties in the children at 6 years of age. Also, the children’s own urinary concentrations of pyrethroid metabolites were related to decrements in verbal and memory functions and externalising difficulties and abnormal social behaviour. While this sole European study did not corroborate US birth cohort studies results showing that exposure during pregnancy to organophosphate insecticides at levels found in the general population may harm brain development in the foetus, the exposure levels measured in the PELAGIE cohort were considerably lower for both organophosphates and pyrethroids than those measured in other European studies as well as in studies from the US and Canada. For example, the median urine concentration of organophosphate metabolites in pregnant women in the PELAGIE cohort was 2 – 6 times lower than for pregnant women in other studies [85, 122, 138] and the concentration of the common pyrethroid metabolite 3-PBA was only detectable in urine samples from 30% of the women compared to 80–90% in other studies [88, 139]. Thus, to supplement the French study and the previously mentioned Danish study of greenhouse worker’s children, additional studies that include more representative exposure levels for EU citizens are desirable.

Although exposure levels found in European countries are generally similar to or slightly higher than concentrations found in the US studies, the risk of adverse effects on neurodevelopment in European populations needs to be further characterised. The organophosphate insecticides contributing to the exposure might differ between the US and the EU, also in regard to oral and respiratory intakes. According to the European Food Safety Agency (EFSA), of all the organophosphate insecticides, chlorpyrifos most often exceeds the toxicological reference value (ARfD) [74]. A recent report utilised US data on adverse effects on children’s IQ levels at school age to calculate the approximate costs of organophosphate exposure in the EU. The total number of IQ points lost due to these pesticides was estimated to be 13 million per year, representing a value of about € 125 billion [140], i.e. about 1% of the EU’s gross domestic product. Although there is some uncertainty associated with this calculation, it most likely represents an underestimation, as it focused only on one group of pesticides.

Unfortunately, epidemiological evidence linking pesticide exposure and human health effects is rarely regarded as sufficiently reliable to take into account in the risk assessment conducted by regulatory agencies. For example, the conclusion from the epidemiological studies on chlorpyrifos is that an association of prenatal chlorpyrifos exposure and adverse neurodevelopmental outcomes is likely, but that other neurotoxic agents cannot be ruled out, and that animal studies show adverse effects only at 1000-fold higher exposures [141]. A recent decrease of the maximum residue limit for chlorpyrifos in several crops [142, 143] was based on animal studies only [144], but the limits for the sister compound, chlorpyrifos-methyl were unchanged. This case highlights a major limitation to current approaches to protecting the general population against a broad variety of pesticides.

Overall crop composition

Metabolomics [148–152], proteomics [153, 154] and transcriptomics [155, 156] studies in controlled field trials provide evidence that the production system has an overall influence on crop development, although there is no direct relevance of these studies for human health. Furthermore, the generally lower crop yield in organic systems [146] as such indicates an effect of management strategy on plant development.

Several systematic reviews and meta-analyses [25, 157–159] with different scopes, inclusion criteria and statistical methods have summarised several hundred original studies reporting some aspect of plant chemical composition in relation to conventional and organic production, in search of overall trends across crops, varieties, soils, climates, production years etc. While the overall conclusions of these systematic reviews look contradictory at first sight, there is agreement between them in most of the detailed findings:

Nitrogen and phosphorus

Existing systematic reviews have consistently found lower total nitrogen (7% [157], 10% [159]) and higher phosphorus (standardised mean difference (SMD) 0.82 [25], 8% [157]) in organic compared to conventional crops. These findings lack direct relevance for human health. However, considering the differences in fertilisation strategies discussed above, and the fundamental importance of N, P [160–162], and the N:P ratio [163] for plant development, this may lend some plausibility to other observed effects of the production system on crop composition.

Vitamins

Systematic reviews generally agree that the concentration of macronutrients, vitamins, and minerals in crops is either not at all or only slightly affected by the production system. For example, ascorbic acid (vitamin C) has received most attention in this context. Meta-analyses report only small effect sizes of the organic production system on vitamin C content [25, 158, 159].

Polyphenols

(Poly)phenolic compounds are not essential nutrients for humans but may play a role in preventing several non-communicable diseases, including cardiovascular disease, neurodegeneration and cancer [164]. The detailed mechanisms are complex and not fully understood [164]. Several environmental and agronomic practices affect the phenolic composition of the crop, including light, temperature, availability of plant nutrients and water management [165]. Under conditions of high nitrogen availability, many plant tissues show a decreased content of phenolic compounds, although there are examples of an opposite relationship [165].

Meta-analyses report modest effect sizes of the production system on total phenolics content, e.g. an increase of 14 – 26% [25, 158, 159]. For some narrower groups of phenolic compounds, larger relative concentration differences (in percent) between organic and conventional crops have been reported [159]. However, such findings represent unweighted averages typically from small and few studies, and are therefore less reliable.

Collectively the published meta-analyses indicate a modestly higher content of phenolic compounds in organic food, but the evidence available does not constitute a sufficient basis for drawing conclusions on positive effects of organic compared to conventional plant products in regard to human health.

Cadmium and other toxic metals

Cadmium (Cd) is toxic to the kidneys, can demineralise bones and is carcinogenic [166]. Cd is present naturally in soils, and is also added to soils by P fertilisers and atmospheric deposition. Several factors, including soil structure and soil chemistry, humus content and pH, affect the plant availability of Cd [167]. The application of Cd-containing fertilisers increases Cd concentrations in the crops [167, 168]. Low soil organic matter generally increases the availability of Cd for crops [169], and organically managed farms tend to have higher soil organic matter than conventionally managed farms [11].

The source of Cd in mineral fertilisers is the raw material phosphate rock. The European average Cd content in mineral fertilisers is reported as 68 mg Cd/kg P [170] or 83 mg Cd/kg P [171]. The content of Cd in farmyard manure is variable but apparently in many cases lower: Various types of animal manure in a German collection averaged between 14 and 37 mg Cd/kg P [172].

Smith-Spangler et al. [25] found no significant difference in the Cd content of organic and conventional crops (SMD = −0.14, 95% CI -0.74 – 0.46) in their meta-analysis, while Barański et al. [159] report significantly 48% higher Cd concentration in conventional compared to organic crops (SMD = -1.45, 95% CI -2.52 to −0.39) in another meta-analysis largely based on the same underlying original studies, albeit with different inclusion criteria. We contacted the authors of these meta-analyses in order to understand this discrepancy. An updated version of the Barański meta-analysis, in which some inconsistencies have been addressed and which has been provided by the original authors [173], shows a significant 30% (SMD = −0.56, 95% CI -1.08 to −0.04) elevations of Cd contents in conventional compared to organic crops; in subgroup analysis, this difference is restricted to cereal crops. No updated meta-analysis was available for Smith-Spangler’s analysis [25]; apparently, two large well-designed studies with tendencies towards a lower Cd content in organic crops were not considered [174, 175] although they appear to fulfil the inclusion criteria. Also, a correction for multiple testing has been imposed, which may be overly conservative, given the prior knowledge that mineral fertilisers constitute an important source of Cd to soils and crops. It is unclear how these points would affect the results of Smith-Spangler’s meta-analysis.

There are short-term and long-term effects of Cd influx from fertilisers on the Cd content of crops [167] but no long-term study comparing Cd content in organic and conventional crops is available. In absence of such direct evidence, two long-term experiments indicate a higher slope in Cd concentration over time for minerally fertilised compared to organically fertilised cereal crops [176, 177], after over 100 years of growing.

A lower Cd content of organic crops is therefore plausible due to a lower Cd content in the fertilisers used in organic farming, and potentially due to higher soil organic matter in organic farmland. The general population’s Cd exposure is close to, and in some cases above, the tolerable intake and therefore their exposure to Cd should be reduced. For non-smokers, food is the primary source of exposure, with cereals and vegetables being the most important contributors [168].

For other toxic metals including lead, mercury and arsenic, no differences in concentration in organic and conventional crops have been reported [25, 159]. Uranium (U) is also present as a contaminant of concern in mineral P fertilisers [178], but less so in organic fertilisers [179], and consequently manure-based cropping systems have a lower U load than mineral-fertilised systems at equal P load [179]. Uranium appears to accumulate in mineral-fertilised soils [180], and agricultural activity may increase the U content of surface and groundwater [181, 182]. However, no evidence was found comparing uranium contents of organic and conventional products.

Fungal toxins

Regarding fungal toxins in crops, one meta-analysis has reported a lower contamination of organic compared to conventional cereal crops with deoxynivalenol (DON), produced by certain fusarium species [25]. Although not fully understood, fungicide applications may alter fungal communities on cereal leaves, potentially weakening disease-suppressive species [183, 184]. Also, crop rotations including non-cereal crops may contribute to lower infestation with fusarium [185], while N availability is positively associated with cereal DON content [186]. These factors give plausibility to the observed lower DON contamination in organic cereals. In the EU, the mean chronic exposure of toddlers, infants and children to DON is above the tolerable daily intake (TDI), with grains and grain-based products being the main contributors to total exposure. The TDI is based on decreased body weight gain observed in mice [187]. The production system does not have any observed effect on the concentration of ochratoxin A (OTA), another fungal toxin of importance in cereal production [25].

Fatty acids

Much of the focus of existing research on compositional differences of organic and conventional animal-based foods is on the fatty acid composition, with a major interest in omega-3 FAs due to their importance for human health. Some studies also address the content of minerals and vitamins.

The FA composition of the feed is a strong determinant of the fatty acid composition of the milk, egg or meat [190, 191]. Grass and red clover, typical roughage feeds, contain between 30% and 50% omega-3 FA of total FA, while the concentrate feeds cereals, soy, corn, and palm kernel cake all contain below 10% omega-3 FA of total FA [190]. Like humans, farm animals turn a small part of dietary alpha-linolenic acid into long-chain omega-3 fatty acids with the help of elongase and desaturase enzymes.

For cow’s milk, a recent meta-analysis reports conclusively an approximately 50% higher content of total omega-3 fatty acids (as percent of total fatty acids) in organic compared to conventional milk [192], generally confirming earlier reviews [25, 189]. Also, the content of ruminant FAs (a group of natural trans FAs produced in the cow’s rumen) is higher in organic milk. The content of saturated fatty acids, mono-unsaturated fatty acids and omega-6 PUFA was similar in organic and conventional milk [192].

A considerable statistical heterogeneity in these findings is reported. Individual differences described above are based on results from between 11 and 19 included studies. The observed differences are plausible, because they are directly linked to differences in feeding regimens. It should also be noted that several other factors influence the fatty acid composition in milk [193]. Specifically, the season (indoor vs. outdoor) has an impact on the feeding regime [188] and therefore on the omega-3 content of milk. However, the content of omega-3 fatty acids is higher in organic milk during both the outdoor and indoor seasons [189].

For eggs, it is likewise well described that the FA composition of the feed [190] and consequently the access to pasture [194, 195] such as in organic systems, is a strong determinant of the fatty acid composition of the egg. However, only few studies have compared the FA composition in organic and conventional eggs [196] and a systematic review is not available. A higher omega-3 content of organic eggs is plausible but has not been documented.

A total of 67 original studies report compositional aspects of meat (mainly beef, chicken, lamb, and pork) from organic and conventional husbandry and were recently summarised in a meta-analysis [197]. Based on 23 and 21 studies respectively, the content of total PUFA and omega-3 PUFA was found to be significantly higher (23 and 47%, respectively) in organic compared to conventional meats. Weighted by average consumption in Europe, choosing organic instead of conventional meat, while maintaining a constant consumption, increased the intake of PUFA and omega-3 FA from meat by 17 and 22%, respectively [198]. These findings are plausible, especially in the case of omega-3 PUFA, considering the known differences in feeding regimens in organic and conventional production. However, few studies were available for each analysis, leaving many analyses with high uncertainty and poor statistical power. Furthermore, fatty acid metabolism differs between ruminants and monogastric animals [190]. Also, the actual differences in feeding regimens between conventionally and organically raised animals may differ by species, and by country. The variation between studies and between species was large, and the overall reliability of these results is therefore lower compared to milk above. This meta-analysis therefore indicates a plausible increase in omega-3 contents in organic meats, but more well-designed studies are needed to confirm this effect [197].

Dairy products account for 4–5% of the total PUFA intake in most European populations, while meat and meat products contribute another 7–23% [199]. The contribution of milk fat to omega-3 PUFA intake (approximated as intake of α-linolenic acid) has been estimated at 5–16% [200, 201], while meat contributes with 12–17% [201, 202]. The effect of exchanging organic for conventional dairy products on omega-3 PUFA intake while maintaining a constant consumption has not been examined rigorously. From the intake and composition data presented here, it can be estimated that choosing organic products would increase the average dietary omega-3 PUFA intake by 2.5–8% (dairy) and by a less certain 2.5–4% (meat). A recent preliminary estimate based on FAO food supply data resulted in similar numbers [198]. For certain population groups and fatty acids, these numbers could be higher, and an increased omega-3 PUFA consumption is generally desirable, as some subpopulations have a lower-than-recommended intake of omega-3 PUFA [203]. However, overall, the effect of the animal production system on omega-3 PUFA intake is minor, and no specific health benefits can be derived. Furthermore, other dietary omega-3 PUFA sources, specifically certain plant oils and fish, are available that carry additional benefits [204–206]. The existence of specific health benefits of ruminant trans fatty acids (as opposed to industrial trans fatty acids) is indicated by some studies [207] but not strongly supported [208]. Taking into account the actually consumed amounts of ruminant trans fatty acids, this is likely lacking public health relevance [208].

Trace elements and vitamins

A recent meta-analysis points to a significantly higher content of iodine (74%) and selenium (21%) in conventional milk and of iron (20%) and tocopherol (13%) in organic milk based on six, four, eight and nine studies respectively [192]. Iodine deficiency during pregnancy and infancy leads to impairment of brain development in the offspring, while excess iodine intake is associated with similar effects, and the window of optimal iodine intake is relatively narrow [209]. Overall, iodine intake in Europe is low and mild deficiency is prevalent [210]. The preferred way of correcting deficiency is salt iodisation [210, 211], because salt is consumed almost universally and with little seasonal variation [212].

Feed iodine supplementation is not linked by regulation to the production system in the EU, as iodine is listed as approved feed additive, and the maximum amount of supplementation is the same for all milk production. Optimum dairy cow supplementation should be seen in relation to other national strategies for human iodine intake. This should also take into account human subpopulations with low or no intake of dairy products.

For tocopherol, selenium and iron, a higher content is generally desirable, and in the case of selenium milk is an important source. However, the concentration differences between organic and conventional milk are modest and based on a few studies only.