Knowledge update on macroalgae food and feed safety

In 2016, a technical report on potential risks posed by macroalgae for application as feed and food, from a Norwegian perspective was published by the Institute of Marine Research (Duinker et al., 2016) . The Norwegian Food Safety Authority (NFSA) has now requested the Institute of Marine Research (IMR) for an updated knowledge status on macroalgae, with special emphasis on data on iodine and metals and new data that could change the conclusions from the 2016 report. Other data on macroalgal food and feed safety generated by the IMR since 2016, that might fill some of the knowledge gaps pointed out in the 2016 report, are also requested.

One of the main conclusions from the 2016 report was that macroalgae, in particular brown algae, may contain elevated levels of inorganic arsenic, total arsenic and cadmium. Further, the amount of data on Norwegian seaweed was too low, allowing only general conclusions. In 2018, EFSA made a call for data in the commission recommendation (EU) 2018/464 on the monitoring of metals and iodine in seaweed, halophytes and products based on seaweed, including both local and imported products.

Since 2016 the amount of chemical data generated by the IMR on Norwegian as well as imported macroalgae has increased substantially, and a monitoring programme was started according to EFSAs requirements. The main updated conclusion is that it now can be discriminated between individual species that have high levels of these components and others that are within the normal range. The 2016 report also revealed a lack of information on seasonal and geographical variation, which is still lacking except for a few species as discussed below. Some work has also been done on the presence of bacteria on kelp and will be summarised here. Studies on bioavailability of metals from kelp have been performed together with studies on the effect of cooking on iodine content, and ongoing student work on iodine is also described. Work on the use of macroalgae in fish feed is summarised as well as analyses of minerals. Finally, the process of submitting occurrence data to EFSA is described and discussed.

Samples were retrieved by field work by IMR staff, kelp growers and wild harvesters during the period 2014-2019. Samples were freeze dried and homogenised before further analysis. The analyses are ISO accredited for most elements and otherwise analysed in our accredited labs, hence fulfilling the request in EFSA’s call for data that the analyses should be carried out in accordance with Annex III to Regulation (EC) No 882/2004. Samples of arame, kombu, hijiki, wakame and nori are imported from production in Asia.

An overview of the species collected is given in Table 1.

Determination of metals (including Cd, Hg, Pb, Fe, Zn and Se) by ICPMS (IMR method 197)

Two parallels were weighed from each sample. The metals were determined by inductively coupled plasma-mass spectrometry (ICP-MS) after decomposing in microwave oven as described by Julshamn et al. (2007). The method is accredited for cadmium (Cd), mercury (Hg), lead (Pb), zinc (Zn) and selenium (Se).

Determination of macro minerals (Na, Mg, K, Ca and P) by ICPMS (IMR method 382)

Two parallels were weighed from each sample. The concentrations of the macro minerals (Na, Mg, K, Ca and P) were determined by inductively coupled plasma-mass spectrometry (ICP-MS), after acid wet digestion in a microwave oven. The concentrations were determined using an external calibration (standard curve) and the method is accredited according to ISO 17025.

Determination of inorganic arsenic by HPLC-ICPMS (IMR method 261)

The sample was added 10 ml 0.07 mol/l HNO ₃ in 3 % H ₂ O ₂ and extracted in microwave oven for 20 minutes at 90 °C. The sample was then cooled, centrifuged and filtrated prior to analysis. Inorganic arsenic was selectively separated from other arsenic compounds by anion exchange HPLC and determined as As5+ by ICP-MS. The method is accredited according to iso-17025.

Determination of iodine

Samples were added 1 ml tetrametylammonium hydroxyide (TMAH) and 5 ml deionized water before extraction at 90 ° C ± 3 ° C for 3h. The samples were then diluted and centrifuged. Prior to quantification, the samples were filtered through a 0.45 µ m single use syringe and disposal filter. Tellurium which was used as an internal standard in order to correct for instrument drift. Iodine concentration in the samples was determined by inductively coupled plasma-mass spectrometry (ICP-MS).

3.1 - Occurrence of inorganic arsenic, iodine, cadmium, lead and mercury in macroalgae

Both fresh and dried macroalgae have been sampled and analysed. Dried products show higher concentrations than fresh, since dry matter percentage typically range from 10 to 30 % (see appendix) resulting in 3-10 times higher concentrations of metals after drying. As an example, median iodine concentration in 141 samples of fresh sugar kelp (Saccharina latissima) was 410 mg/kg wet weight while 16 samples of dried products had median concentration of 3650 mg/kg dry weight. For comparison purposes, all concentrations are converted to dry weight basis in the following review. 27 species have been sampled and analysed, and 14 of these have five or more samples.

The tables include both mean and median values, since in many instances the mean values are affected by a few atypically high values, and median values are more representable for the typical values. The 25 % quartile range is presented in addition to the minimum and maximum values for the same reason, and the upper 75 % quartile value is more suited for comparison of typically high values, not affected by a few extreme values.

At present there are no maximum levels (MLs) for minerals in seaweed as food in the EU, except for mercury, and a call for data for the period 2018-2021 has been made to generate a basis for evaluating establishment of MLs. As stated in the Commission Recommendation (EU) 2018/464 on the monitoring of metals and iodine in seaweed, halophytes and products based on seaweed: “For arsenic, cadmium and lead, maximum levels (MLs) for various foodstuffs are established under Commission Regulation (EC) No 1881/2006. However, currently no MLs are established for these substances in seaweed and halophytes, except for the MLs established under this Regulation for food supplements consisting exclusively or mainly of seaweed or products derived from seaweed”. Concentrations in fresh seaweed have been converted to dry weight concentrations for comparisons with dry seaweed products. Dry weight percentages are presented in the appendix.

3.1.1 - Inorganic arsenic

Content of inorganic arsenic show a particularly high variation (Table 1). Among the types of seaweed that have been sourced in Norwegian waters from mainly wild harvest, but also from cultivation, the oar weed (Laminaria digitata) shows the highest concentrations. The variation spans from the lowest concentrations to the highest concentrations in Table 1, and the highest concentrations are at the level of the Asian produced hijiki ( Sargassum fusiforme ), which several European countries (including Norway), USA and Canada have issued warnings for due to the high level of inorganic arsenic and increased risk of cancer. 50 % of the samples of oar weed have concentrations of inorganic arsenic above 24 mg/kg dry weight, and the three highest samples have concentrations between 63 and 79 mg/kg dry weight. Also cultivated oar weed showed very high concentrations. In a small study, inorganic arsenic was seen to increase substantially in oar weed between March and June (Duinker, 2014) , but samples submitted by the industry in 2018 showed an opposite pattern. The concentration in one sample of Hijiki was 59 mg/kg dry weight, and levels from literature typically vary between 50 and 90 mg/kg dry weight (Rose et al., 2007; Yamashita, 2014) . A related species to hijiki, the invasive wireweed ( Sargassum muticum ) also showed high concentration of inorganic arsenic in four samples with concentrations between 48 and 68 mg/kg dry weight. A native species, Halidrys siliquosa (sea oak, family Sargassaceae also including the Sargassum species), is also related to Hijiki and had concentrations between 2.4 and 42 mg/kg dry weight. These two latter species are not commonly used at the moment. On the other side, a species that is closely related to oar weed with quite similar morphology, Laminaria hyperborea (tangle), showed concentrations in the lower end of all seaweed species between 0.03 and 0.04 mg/kg dry weight in four samples analysed. Several companies report a shift from L. digitata to L. hyperborea in their products. More samples are needed of both the Sargassum related species and L. hyperborea to confirm these findings.

The species with high concentrations of iAs range in median concentrations from 22 to 59 mg/kg dry weight, and there is then a distinct gap down to the other species where median concentrations are 100 times lower from 0.23 and below. 

3.1.2 - Cadmium

Regarding cadmium, the levels are in accordance with the 2016-report (Duinker et al., 2016), but the available information of specific species is improved. There is no group of species that show drastically elevated concentrations like for inorganic arsenic, and the concentration ranges overlap (Table 2). Red and brown algae have the highest concentrations, while green algae are very low in cadmium. Sugar kelp has the highest concentrations for the northern localities as discussed below (“Geographic variation of cadmium in farmed sugar kelp”). 

3.1.3 - Iodine

The concentrations of iodine are in accordance with the 2016 report (Duinker et al., 2016), but with more detailed information at the species level. The kelp species in the Saccharina and Laminaria families, both in Norway and Asia, have the highest levels with typical concentrations between 3 000 and 4 000 mg/kg dry weight and maximum concentrations around 10 000 mg/kg dry weight (Table 3), which is higher than any other food group. Winged kelp and the related Wakame from Asia, however, have a concentration range one tenth or less of the Laminaria and Saccharina species, which is part of the reason that winged kelp is getting more popular for cultivation in Norway these days. In general, the brown perennial species are intermediate, and the green and red algae are relatively low in iodine concentrations. One exception here is Vertebrata lanosa (wrack siphon weed) which grows as a symbiont on Ascophyllum and has ten times higher iodine concentrations compared to the other red algae.

For adults, recommended daily intake of iodine is 150 µg and maximum daily intake is 600 µg. Nori sheets contained about 60 µg per sheet, and between 2 and 10 sheets should cover the range 150-600 µg iodine. 1 ml (1/5 teaspoon) of dried kelp flakes weighing approximately 100 mg correspond to approximately 400 µg of iodine, given a typical concentration of 4 000 mg/kg dry weight. Similar calculations could make producers able to advice intake to the consumers.

3.1.4 - Lead

There are generally low levels of lead in seaweed (Table 5), although large variation is seen within each species with a few relatively high values 5-10 times higher than the 75% quartile for some of the species. Such high values should be followed up in future studies. 

3.1.5 - Mercury

The levels of mercury are generally low. Most mercury concentrations were below the limit of quantification (loq), and only 100 of 406 samples were above loq. Only five of the species had more than 50% of mercury concentrations above loq so that mean concentrations could be presented (Table 6). However, this does not reflect higher concentrations of mercury as loq was quite variable. Samples with high total mineral content were diluted more than other samples, resulting in higher loq, and some species had hence more concentrations above loq but lower concentrations than other species.

The highest max concentrations were found in sugar kelp, oar weed and wrack siphon weed with 0.08, 0.07 and 0.07 mg/kg dry weight, respectively.

3.1.6 - Commercially cultivated species: Sugar kelp and winged kelp

The higher number of data in this update also allows more precise consideration of sugar kelp and winged kelp, which are the most common species for both cultivation and wild harvest in Norway at the moment. Neither of these species are among the high concentration species regarding inorganic arsenic. Sugar kelp is number seven and winged kelp number 12 in the list based on decreasing median concentrations of inorganic arsenic (Table 2). For the winged kelp one extreme value of inorganic arsenic was seen with more than ten times higher concentration than the 75 % percentile. This could be associated with epiphyte algae as suggested below (3.2), and this should be followed up with further studies. Winged kelp and sugar kelp have intermediate cadmium concentrations present as number six and eight on the list with decreasing median values (Table 3). Regarding iodine, sugar kelp is among the species with the highest concentrations and winged kelp intermediate (Table 4). The higher concentrations of cadmium in sugar kelp are found in more northern areas as discussed below (“Geographic variation of cadmium in farmed sugar kelp”).

3.2 - Factors affecting the levels of metals in kelp

As discussed below (“Geographic variation of cadmium in farmed sugar kelp”), cadmium concentrations in farmed sugar kelp showed a clear increase from south to north in Norway. The same section also discusses how iodine is reduced during cooking and lack of seasonal variation for farmed kelp.

During a master thesis study of wild and cultivated sugar kelp and winged kelp (Kleppe, 2016), cadmium decreased with increasing size of the plants, which suggests that fast growing individuals have lower concentrations of cadmium than slow growing individuals. Cadmium concentrations also decreased from the stipes and growth zone towards the tip. Iodine showed a similar trend within the plants, but no clear correlation with size. Inorganic arsenic showed no consistent variation but was usually highest in the mid-section of the blades of both species. Fouling was found to increase metal concentrations with higher cadmium concentrations in areas of sugar kelp covered with the bryozoan Membranipora membranacea , while an unknown filamentous alga increased concentrations of inorganic arsenic substantially in wild winged kelp. Differences were seen between wild and cultivated plants, but the differences were inconsistent between the two species and between the different minerals.

The amount of data on season, geography, exposure etc. are still scarce and more studies are needed. The variation even within the same locality (data not shown) for cultivated kelp is surprisingly high, and more knowledge of the factors causing such variation is needed to allow more predicable product quality for the seaweed industry.

3.3 - Radioactivity

Macroalgae are known to effectively concentrate radionuclides from seawater and are therefore widely used as a bio-indicator for radioactive pollution in the marine environment (e.g. Keogh, 2006; Kershaw et al., 2005) . For example, Ascophyllum nodosum are known to have a high uptake of the radionuclide ⁹⁹ Tc (Sjøtun et al., 2011) , with increasing concentrations from young to older growth segments (Heldal and Sjøtun, 2010) . Nevertheless, few publications describe food safety aspects of radionuclides in macroalgae. Tuo et al. (2016) found no higher levels of either natural ( ²³⁸ U, ²²⁶ Ra, ²²⁸ Ra, ⁴⁰ K) or anthropogenic ( ¹³⁷ Cs) radionuclides in different seaweed species compared to other seafood, vegetables or meat products. Similarly, Moreda-Piñeiro et al. (2011) concluded that radiation levels in typical Japanese and Korean foodstuff, which include seaweed, are safe and at the same level as other countries.

The Norwegian marine monitoring programme Ra dioactivity in the M arine E nvironment (RAME) ( www.dsa.no ) charts trends of radionuclides in the marine environment. The results show that levels of the beta emitter ⁹⁹ Tc in F. vesiculosus collected along the Norwegian coast in the period 2012-2014 did not exceed 70 Bq/kg dry weight (d.w.) (Skjerdal et al., 2017). Further, the levels of the gamma emitter ¹³⁷ Cs, in F. vesiculosus collected in the same area and period ranged from 0.17 Bq/kg (d.w.) to 3.2 Bq/kg (d.w.). Levels of the alpha emitter ^239,240 Pu were below the detection limit in samples collected in the same area and period (Skjerdal et al., 2017) . There are no maximum permitted levels for ⁹⁹ Tc and ^239,240 Pu in foodstuffs in Norway. The maximum permitted level for ¹³⁷ Cs in foodstuffs set by Norwegian authorities after the Chernobyl accident is 600 Bq/kg. The ¹³⁷ Cs-levels in macroalgae along the Norwegian coast are far below this and should be of little or no concern to seafood consumers. In general, no radioactivity levels of concern with respect to food safety is found in seafood. Levels of natural radionuclides are generally higher than levels of anthropogenic radionuclides. For example, Tuo et al. (2016) found that the natural radionuclide ⁴⁰ K accounted for around 87% of the total dose from the radionuclides ²³⁸ U, ²²⁶ Ra, ²²⁸ Ra, ⁴⁰ K and ¹³⁷ Cs in foodstuff from coastal areas of China.

3.4 - Other components

A few samples of kelp have been analysed for dioxins and PCB’s and show low concentrations, in accordance with the 2016 report. 13 samples of dulse were analysed for kainic acid at the Danish Technical University and showed a range from 5 to 180 µg/g dry weight and hence confirm relatively low levels of this toxin. The toxicity of kainic acid has not yet been determined. However, in order to reach the hazardous levels of kainic acid similar to what was dosed in mice and rat experiments, a total amount of about 30 kg dry dulse of the 130 µg/g dulse would have to be eaten (Mouritsen et al., 2013) . Recently, higher values up to 500 µg/g dry weight has been found in Danish dulse, but still almost 10 kg would be needed for toxic responses.

3.5 - Microbiology

Since 2016 IMR worked with microbiology in an industry project. Blikra et al. (2019) describe the food quality and microbial safety of the two brown macroalgae winged kelp (Alaria esculenta) and sugar kelp (Saccharina latissima) harvested and processed in Norway. Samples included raw and frozen kelp. The authors found for all samples low microbial numbers (1–3 log colony forming units/g) for total aerobic count, cold adapted bacteria and spore-forming bacteria. Furthermore, there were no detection of indicators of faecal contamination as enterococci and coliforms, nor pathogenic vibrios or Listeria monocytogenes .

However, in several of the examined samples, Bacillus spp. were isolated and seems able to pose a challenge if not processing and storage conditions take their possible presence into account.

Bacteria in the genus Bacillus may grow under aerobic and anaerobic conditions. The bacteria and their spores are widely distributed in the environment and has been isolated from a wide variety of foods, especially of plant origin, but also from meat, fish, and dairy products. The most important species in food microbiology is B. cereus , and bacteria in this group have been involved in several food borne infections and intoxications.

Most strains of B. cereus are able to grow in low-acid foods at temperatures down to 10 ^o C and up to 55 ^o C (optimum 30 to 40 ^o C). During the last decades, some strains of B. cereus have been found able to grow at temperatures down to 4 ^o C. Bacillus species has been isolated from sous vide cod fillets at 5 ^o C and in many other sous vide products. Food containing more than 10⁴ B. cereus cells per g may not be safe for consumption. This number is far above what was seen by Blikra et al., 2019.

Several foods like milk and rice are also known to contain Bacillus spp., and it is common routine that heat-treated food products containing milk or rice are stored cold to prevent Bacillus spp. from growing. Heat treatment usually gives a temperature high enough to kill vegetative bacterial cells and any competitive microbiota, but not Bacillus spores. After such heat treatment, spores may be reactivated and give multiplication of Bacillus without competing bacteria present. Control of Bacillus cereus is efficiently obtained by chilling, except for some few cold-adapted strains that mainly pose a challenge for dairy products. The same precautions should hence be taken for heat treated products containing macroalgae due to possible presence of Bacillus spp. spores.

3.6 - Bioavailability, geographical variation and effect of cooking on sugar kelp composition

This project has not yet been finalised and reported, but summary of the findings this far, is given below. The project as financed by the Norwegian Seafood Research Fund (FHF) and carried out in collaboration with the Technical University of Denmark (DTU) and the kelp producer Ocean Forest.

Original project title: «På sporet av ny mat»

Experimental diet for rats added sugar kelp (Saccharina latissima) 

We have performed a 13-week feeding trial with female Wistar IGS rats to compare the health effects and bioavailability of iodine from farmed sugar kelp with potassium iodine. No signs of toxic responses were found in animals with 0.5 and 5% sugar kelp or comparable amounts of potassium iodine in the diet. No difference in body weight, feed intake, heart-, kidney- or live weight were observed in the animals fed dried kelp in the diet at 0.5 and 5%. Biomarkers for liver and kidney damage measured in plasma were not increased in rats given potassium iodide or iodine from sugar kelp. No significant differences in serum concentrations of the thyroid stimulating hormone (TSH) and thyroid hormone T3 concentrations were observed. However, a small but significant reduction in serum T4 was observed in the rats fed the highest concentration of sugar kelp and potassium iodide.

Collectively, the observed high tolerance of iodine in rats agree with previous studies (Calil-Silveira et al., 2016; Yoshida et al., 2014). Yoshida et. al (2014) reported no significant effects on serum T3 and T4 concentration in rats despite an iodine intake up to 3500 µg iodine per day from kelp. 

Based on available literature and the results from our experiment, it seems clear that rats have a high tolerance for iodine, even at the high intake of sugar kelp (5% of the diet) in the present study. The rat model used was hence not suited for evaluation of negative effects of iodine. On the other hand, the high tolerance of the iodine allowed to conclude that none of the other components of the kelp had a negative impact on the health of the rats. The high tolerance of iodine in rats also made it possible to study the bioavailability of iodine from kelp at very high concentrations.

Bioavailability of iodine from sugar kelp in rats 

From previous studies both in animals (Yoshida et al., 2014) and humans (Aquaron et al., 2002; Combet et al., 2014), based on excretion of iodine in urine, there are indications that the bioavailability of iodine is lower from kelp compared to the salt form of iodine, potassium iodide. Most of absorbed iodine is excreted in the urine, and only a small fraction is retained in the thyroid gland or found in circulation. The amount of iodine excreted in urine is hence used as a measure for the amount of absorbed iodine over a time period. By monitoring urinary iodine excretion in rats from the 13-week feeding trial, the bioavailability of iodine from sugar kelp was evaluated. Urine excretion of iodine reflects the dietary intake in all experimental groups. 94-95 % of total iodine intake was excreted in urine in both the low and high dose of potassium iodide. Urine excretion in rats given iodine from sugar kelp, show a significantly lower excretion level in urine with 73-78% bioavailability of iodine. This is still relatively high, but significantly lower compared to potassium iodide. These results are in agreement with data reported on bioavailability of iodine in humans using both potassium iodide (96.4%) and seaweed (93 and 75%) (Aquaron et al., 2002). A lower bioavailability of iodine from seaweed were also observed when comparing three iodine-rich foods in a human intervention pilot study. Iodine excretion in urine were 86%, 87% and 60% with 36 hours collection after consumption of fish, milk or seaweed, respectively (Redway et al., 2018). The rat model used in our study was probably well suited for evaluation of iodine availability, although it was not suited for evaluation of toxic effects as discussed above.

In contrast to earlier reports we quantified iodine in feces. A lower urine excretion of iodine in rats given sugar kelp was accompanied by a significant increase in iodine found in feces. Diets supplemented with potassium iodide resulted in 0.5-1 % of the iodine in feces on average. In contrast, rats given kelp in the diet had on average 8% of total iodine in feces, representing iodine not absorbed in the intestine from the kelp diet.
 

Bioavailability of total and inorganic arsenic, cadmium and copper from sugar kelp in rats

More than 70 % of ingested total arsenic in rats fed the a diet with 5 % sugar kelp was found in the 24 hours feces collection. This is in accordance with a high proportion of arsenic being organic, hence having a high availability. For cadmium, the amounts in the 24 hours feces collection were almost identical to the ingested amount, indicating very low availability. This is in accordance with a generally low availability of inorganic metals forms. Some cadmium was assimilated, though, as seen from the significant increase in cadmium in liver and kidney in the 5% diet group. Regarding copper, the concentrations in the kelp were too low to give differences in the feed between the groups, and there were no differences in the liver or kidney samples.

Regarding inorganic arsenic, rats and other rodents are not suitable models since these animals metabolise and excrete inorganic arsenic at high rates. Inorganic arsenic was only detected in the 5 % kelp group feed, and no concentrations above limit of quantification (LOQ) could be found in the liver or kidneys samples from any of the groups.

The effect of cooking on iodine levels in sugar kelp

Iodine is water soluble, and iodide (I ⁻ ), probably the main form in kelp, reacts with water forming the volatile hydrogen iodide. Several forms of cooking were performed to assess the effect of iodine content of the sugar kelp. Boiling in water for 15 minutes released 50-90 % of iodine to the water, and of this iodine 50% was released to air. Continued simmering of the stock reduced further 50 % over 15 minutes. Frying released in average 50 % (25-80) of the iodine. Drying of sugar kelp increases concentrations of all ellements ten times just by removing the water, which makes 90 % of the wet weight. However, when comparing iodine concentrations on dry weight, it was seen that about 25 % of the iodine evaporated during drying.

Geographic variation of cadmium in farmed sugar kelp

In cooperation with SINTEF (Macrosea project), a standardised growth trial was performed on localities all along the coast of Norway, and samples were analysed for metals at the IMR. The results showed a clear trend with increasing concentrations of cadmium from south to north. The levels in the north (68-70N) were above 1.0 mg/kg dry weight which is the limit for inclusion in fish feed, and more than four times higher than the southern samples (58-60N). Inorganic arsenic and iodine did not show a similar trend.

3.7 - Student projects on iodine

There are three relevant master theses at the IMR with main focus on iodine. Two of these are based on purchased seaweed products from Norwegian web shops. The products are analysed for iodine, total arsenic, mercury, lead and cadmium. This work also includes blood and urine samples from some seaweed eaters and are analysed for iodine status and thyroid markers. The third study has included four of the purchased seaweed products in a vegan week menu. These daily menus are analysed with and without seaweed for iodine and metals. All three master theses were defended in June 2020 and the main results will be published in peer reviewed journals in 2020-21.

3.8 - Nutrient analyses

The trace minerals, iron (Fe), zinc (Zn) and selenium (Se), were analysed together with the heavy metals and metalloids and hence with the same number of analyses. The macro minerals, calcium (Ca), potassium (K), magnesium (Mg), sodium (Na) and phosphorus (P), were analysed separately with in total 104 analyses. Seven species had three or more samples. The results are presented in Table 7 and Table 8.