Nickel is commonly known to cause contact dermatitis, with estimates ranging from 10% to 34% of the population [Katta2014, Schnuch2012, Thyssen2009] and rising [Rietschel2008]. It is now known that some individuals are sensitive enough to nickel that they have a systemic dermatological reaction to the nickel introduced by their diet, in doses as low as 220 µg/day [Jensen2006]. For some the reaction goes beyond dematological and also includes headache, gastrointestinal and respiratory symptoms [Ricciardi2014, Calogiurui2016, Tammaro2011, Minelli2010, Schiavino2006] and other symptoms consistent with fibromyalgia syndrome [Goldenberg2015]; this is known as Systemic Nickel Allergy Syndrome (SNAS). Nickel hypersensitivity has also been implicated in fibromyalgia, chronic fatigue, and other chronic diseases through the corrosion of dental metals [Sterzl1999, Lindh2002, Muris2006, Sjursen2011, Stejskal2013, Stejskal2014], where the combined oral intake from dental metals and diet can lead to a systemic immune hypersensitivity reaction.
Advice on lowering dietary nickel intake is available on the Internet (e.g. [Athena2016, PennState2016, WhatAllergy2016, Tehrani2016, Riefler2016, Kerns2016]), but it is inconsistent and sources are rarely cited. Tea, baking powder, tomatoes and margarine are examples of foods that are inconsistently listed as high nickel foods. This stems partly from inconsistencies in the food samples themselves: nickel levels can vary greatly among samples of a single food, and if cooked, the choice of cooking method affects the nickel content.
To gain a better understanding of the nickel content of foods, we've collected data from around the world into a single spreadsheet so that we can assess the nickel content of each food as an average over all samples, and discuss the effect of cooking methods. We express the nickel content in terms of a typical serving size to keep comparisons realistic. By doing this, we can give consistent advice based directly on scientific data, and enable sound recommendations for a low nickel diet.
The primary sources of nickel data are the food composition studies carried out by scientists in national health organizations. Nickel data is freely provided by the United States, Denmark, Estonia, Canada, Australia, New Zealand, Hong Kong, India, and the UK. Data for prepared meals such as fast food items or highly mixed foods was excluded, both to manage the size and because of the huge potential for variation in the ingredients and preparation methods. The scientific literature was also searched for any openly available data on the nickel content of foods.
These sources were imported into a common spreadsheet, with each food placed into one of 26 categories: Dairy products and substitutes; Grain products and bakery; Whole grains, flours and starches; Fruits (berries, citrus, or other); Vegetables (root, leafy, other or cooked); Meat, poultry and eggs (raw or cooked); Fish and seafood (raw or cooked); Beans and bean products; Nuts and seeds; Beverages; Herbs and spices; Fats and oils; Snacks; Sugars and sweets; Sauces, dips, condiments and spreads; or Miscellaneous. Health Canada's serving sizes were used to scale the data to a typical serving size.
New data is added as it is discovered; a complete list of the included data sources is supplied within the workbook. Please be sure to tell me if you have some data to contribute.
Since the nickel content of a food varies depending on several factors, every food will have a range that its nickel content will lie within. Sometimes the range is quite large. In the summary data tables, where <Ni> is the mean nickel content per serving, the standard deviation σ is a statistic that means that 68% of the time, a food will have a nickel content between <Ni> - σ and <Ni> + σ. The last two columns show the minimum and maximum reported values, which gives a better indication of just how big the spread can be.
The number of samples is important. Many foods have only one or a few samples. Because of the wide variation in nickel content, a single sample could appear on the very low or very high end of the range, which would be misleading. Only the foods with a large number of samples can be trusted.
What does it mean to say that a food is low, medium or high nickel? For treating nickel hypersensitivity, the recommended daily upper limit is 150 µg nickel [Mislankar2013]. If a person eats 3-4 types of food at each of 3 meals, plus 2-3 snacks of 1 or 2 different types of food, for them to stay under 150 µg each food would on average have to have less than 10 µg of nickel. Therefore we consider foods with less than 10 µg to be low nickel, foods with between 10 and 20 µg to be moderate nickel, and anything higher to be high nickel.
Dairy is consistently the lowest nickel food group. Cream, butter and cream cheese have less than 1 µg per serving. Milk and buttermilk have a few large measurements, but they are outliers, far from the average.
For dairy substitutes, rice beverages are the best option. Soy beverages are most often high nickel and should be avoided. Although there is no data, quinoa beverage may be an acceptable, nutrient-rich option, based on the low nickel levels in the grain itself (see next section).
Whole grains are the guilty culprit in this category. The foods with the least nickel per serving are made with white wheat flour, corn and white rice, while the foods with the highest nickel content are made with buckwheat and oat products and amaranth. The exception is quinoa, which although it is a whole grain has about the same nickel content as white wheat products according to two separate studies. Whole wheat, brown rice, rye and amaranth products fall somewhere in between.
For breads, the ingredients can vary greatly. Whole wheat breads are not always made entirely with whole wheat flour, so their nickel values may be closer to that of a white bread. Likewise, rye breads often contain more white flour than rye, particularly in North America. Nuts, seeds and whole grains have become popular additions. It is important to read the list of ingredients for premade products.
Wheat, barley and rye flours are low in nickel if they do not include the whole grain.
If you’re gluten-free, a blend of white rice flour and potato or corn starch should make a good low-nickel blend. A small amount of potato flour has been suggested to add crispiness and density without adding nickel.
Unfortunately tapioca starch has not been tested, but there is one measurement of tapioca root that is moderate nickel.
Bean flours (including soy and chickpea) and nut flours are high nickel, as are millet, sorghum and oat flours.
Coconut flour is likely not good since coconut meat is high nickel.
Cherries, apples and mangoes are common low-nickel fruits. Citrus fruits are on average low in nickel, but there is some variability. The same goes for blueberries. Unfortunately, banana and grapes on average have low nickel levels, but are highly variable and not to be trusted. Common fruits to avoid for their high average nickel content and/or high variability include raspberries, blackberries, pineapples, dates, stone fruits, melons, pears, coconut and avocado.
Canning has no obvious effect on nickel content for fruits, even though they are often acidic.
Common low nickel root vegetables include onions, carrots, beets and potatoes. Earlier versions of this document concluded that the skin of the potato contributes more nickel, but this has been proven wrong since more data was added.
Of the leafy vegetables, cabbage and lettuces (aside from head/iceberg lettuce) are reasonably low in nickel on average, but there is a wide variation, up to 20 µg in a serving. All others, including spinach, kale, and lettuce, have on average more than 10 µg in a serving. Iceberg lettuce is the worst type of lettuce, with a moderate average and huge variability. Avoid seaweed.
Other common low-nickel vegetables include tomatoes, zucchini, celery, squash, corn, broccoli, cauliflower, cucumber, peppers. Mushrooms have a low average value but a higher variability, so your content may vary. Peas, green beans and bean sprouts are high nickel. Brussels sprouts have a moderate average and high variability.
Canning significantly increases the average nickel content for tomatoes, peas and corn, yet decreases it for beets and green beans.
In food studies, cooking is typically carried out by boiling, baking or frying, often using stainless steel pots and utensils. This may increase the nickel content, as seen for squash, broccoli, rutabaga and potato, but in some cases cooking decreases it significantly, as seen for spinach and collard greens. Which cooking methods were used would probably explain the difference (e.g. boiling vs. frying), but we aren't given that level of detail in the data for the most part. In all cases the variability in nickel content is larger for cooked vegetables.
Meats and eggs are on par with dairy for overall nickel content, when raw, with the exception of goat. Organ meats are on average low nickel, but with some variability.
Cooking has a significant effect on nickel content, most notably on the variability for all meats. The mean value increased for beef and pork; the largest increases are seen in the Canadian food studies, which may be because nickel-containing cookware was used for these samples, or it may be because these samples were higher in nickel even in their raw state. Cooking had a much smaller effect on eggs, poultry and organ meats, regardless of the location. These may be safer choices for eating out.
Fish and many shellfish are, on average, a low nickel food whether they are canned or raw, with the stand-out exception of molluscs (e.g. oysters, mussels and scallops). Cooking molluscs reduces their nickel content slightly, possibly from releasing nickel into the boiling water, but they are still moderately high nickel.
There are no beans that can be recommended for a low-nickel diet. Cooking reduces the nickel content, possibly releasing nickel into the boiling water, but they are still over 25 µg/serving. The lowest-nickel beans appear to be chickpeas and fava beans, and the highest soy beans, however there are too few data points to differentiate between them with any certainty. Noodles made from beans may be low nickel, but there is only one sample.
There are no nuts or seeds that can be recommended for a low-nickel diet. Peanuts, pistachios and almonds are on the lower end, while cashews and brazil nuts are very high. The variability of nuts and seeds is large.
Beverages are an important category, because we need to stay hydrated to support the body's detoxification systems, but even water contains nickel. Many municipalities produce annual reports on contaminant levels in their water supply and should be contacted for their testing results. The age and composition of the water pipes inside your home can also influence the nickel content. If you have misgivings about your water supply, bottled water may be the better option.
It appears that tea brewed from loose leaves has less nickel than tea brewed from a bag, and that green tea has slightly more nickel on average than black tea. In teas, the effect of brewing time (5 minutes vs. 10 minutes) on the content of other trace metals was found to be negligible [Sembratowicz2014]; presumably the same trend would hold for nickel. [Berg2000] found that kettles can introduce large amounts of nickel, whereas coffee machines did not. Coffee, at about 7 µg/cup, also has less nickel than tea, on average. So coffee may be the better choice of hot drink, especially when out and about. Energy drinks (canned) are about the same as coffee. Hot chocolate has high nickel levels.
Of drinks containing alcohol, beer is the least favourable; it has an average of 7 µg in a 12 ounce serving, but the standard deviation is so large that this value could easily reach 30 µg. Liquor (vodka and Cachaça) appears to be the best of the alcohols, with less than 1 µg per serving at most. Wine and cider have around 5 µg.
Citrus, grape and apple juices are low-nickel drinks. Vegetable based beverages are moderate nickel, and "superfruit" juices appear to have high nickel levels, based on few measurements. Pineapple and prune juices are high nickel, as expected based on the values for the fruits themselves. Juices from concentrate tend to be higher in nickel. Canning increases the variability of nickel content in vegetable juices, but interestingly not in fruit juices.
Other high-nickel drinks include coconut water and all chocolate beverages. Canned carbonated drinks have a much larger variability than non-canned.
Due to the small amounts used, most herbs and spices are nearly nickel-free. However, more data is needed in this category since many of them have only one sample, which may be misleading. Be mindful of how much will end up in a serving of your meal; it adds up quickly.
Salt is the best-tested with 22 studies, and it has the lowest nickel of all at 0.05 µg in 1/8 teaspoon (1 g). On the other hand, black pepper is near the top of the list with over 5 µg in 1/4 teaspoon (0.5 g). Hot peppers top the chart at 10 µg for 40 g (about 3 jalepenos, for example). Garlic has about 1 µg in 1/2 tablespoons fresh or 1/8 teaspoon garlic powder. Cinnamon has 2.5 µg in 1/8 teaspoon, so do not use it too liberally. Sage and oregano have higher nickel content, 3.5 and 5 µg in only 1/8 teaspoon, which may add up to a significant amount per serving of a recipe.
Cooking oils are all fairly low in nickel on average, since the serving size is only 2 teaspoons (10 g), but only a few have enough data to draw conclusions about them. The relative average nickel content of the oils for which we have 3 or more values is: olive (0.25 µg) ≈ corn (0.25 µg) < sunflower (0.4 µg) < soybean (0.8 µg) < canola (1.2 µg). Anecdotally, soybean oil has been problematic for some people.
Lard and butter have about 0.5 µg in 2 tsp, and margarine comes in only slightly higher at 0.8 µg, probably because of the tendency for margarines to contain soy or canola oils.
The snacks that are highest in nickel are those that include chocolate, nuts, oats or dried fruits as ingredients. Any bean-based snacks also have elevated nickel content, such as chick pea snacks. Savoury snacks such as potato chips and popcorn are highly variable. Low-nickel options for a sweet treat include flavoured gelatin, applesauce, popsicles and cookies that don’t contain chocolate or nuts.
Gelatin is not permitted in some low nickel diets [Sharma2013, Antico2015] despite testing as having low nickel. There is little doubt that it causes flare-ups in some people, however it is not clear that nickel is at the root of it.
Syrups and confections that contain chocolate are high nickel. Candy is very variable, so it should be avoided unless home-made. Refined sugars are low nickel, white sugar having half that of brown. Honey is well-tested and has only 3.8 µg in 4 teaspoons, and maple syrup (for which we have just one measurement) has only slightly more, though we tend to use more of it at a time. Molasses has less nickel than honey, but only based on one measurement.
Be conservative with condiments, for although they are usually low nickel in small amounts, it adds up quickly. As expected, hummus, pesto made with nuts, and hazelnut chocolate spread are high nickel. Tartar sauce and fruit pastes are unexpectedly high. Soy mayonnaise has a much higher nickel level than regular mayonnaise, which is consistently very low nickel. A tablespoon of soy sauce has around 5 µg of nickel.
It is also worth mentioning that there is a giant geographical variation in the salad dressing data. In New Zealand, Eta brand salad dressings were measured as having zero nickel, and Kraft brand dressings were extremely high. Elsewhere in the world, salad dressings of various brands came in at <2.5 µg per 30 g serving.
This category contains items often used in cooking and baking. The important thing to remember about these ingredients is that they usually create a finished product that has many servings, for example 2 teaspoons of baking powder or soda serves an entire recipe of perhaps 12 muffins or 30 cookies. The contributions of the ingredients should be added up and divided by the number of servings.
A cup of ready-made broth has about 10 µg of nickel. The data implies that the concentrated powders have less, but this seems unlikely; a closer look at the raw data shows a lot of repetition in the Estonian data which may be biasing the average. It would be useful to know whether home-made broth could be made to contain less nickel.
The serving size for baking powder or soda is 0.6 g or 1/8 of a teaspoon, which is not the amount typically used in a recipe. Increasing this to 1 teaspoon, baking powder has about 6 µg of nickel and baking soda about 8 µg, on average.
Tomato paste and sauce are higher nickel than fresh tomatoes.
Cocoa is high in nickel. There is only one value for coconut milk, and it says that it is quite low in nickel; however, this does not fit with the pattern since fresh coconut, dried coconut, and coconut water are all high nickel. More data is needed to confirm this.
For the majority of the data, if a measurement was marked as being lower than the limit of detection (LOD), we used the LOD as the sample measurement. For the Canadian data, some of these values were extraordinarily high and were left out as outliers. For the New Zealand data, a measurement is flagged if it is lower than the LOD, but the LOD isn’t given, so those were also omitted. In some research articles, values differed from the food composition studies by as much as two orders of magnitude. These were excluded from the data.
In some data, a single country can contribute more than one measurement of a food. For example, New Zealand contributes 4 measurements for seeded bread which are all quite high. As well, the Estonian and Danish databases have the same values for several foods; the labels differ but they may be from the same source. These cases both affects the final result by biasing the average.
The data themselves are prone to error introduced by the measurement method. The Inductively Coupled Plasma Mass Spectrometry (ICPMS) method is often used to find the nickel content of foods, but this machine has a part in it that is sometimes made of nickel and can introduce nickel into the sample.
Sometimes assumptions are made about a food based on misunderstandings. For example, some believe that margarine is very high in nickel because nickel is used as a catalyst in the hydrogenation process. The definition of a catalyst in a chemical reaction is an atom or molecule that participates in the reaction to help it along, but is not part of the final product. The data presented here confirms that margarine has low nickel content; while it is slightly higher than that of butter, it isn’t a high-nickel food. Likewise, baking powder has been listed as a food to avoid, but the data doesn't support it. This misconception may exist because portion size hasn’t been accounted for; a typical recipe calls for 2 tsp (10 g) of baking powder whereas measurements are typically reported in units of µg/100g or mg/kg. One does not often use a kilogram of baking powder in a muffin recipe.
Cooking methods can make a big difference, as seen in the lowering of nickel content for foods that are boiled (e.g. beans) and elevated nickel for foods that are fried (e.g. cooked meats). A literature review by the Finnish Institute of Occupational Health [Santonen2010] found that the amount of nickel released into foods through cookware or kettles can vary substantially; this is confirmed in our collection of data, where cooked foods often have a large standard deviation. Any food that is cooked in a nickel-based stainless steel container is at risk of having high nickel content; for example, tomatoes were shown in [Kamerud2013] to contain a 26-fold increase in nickel content when cooked for 6 hours in a nickel-based stainless steel. [Bassioni2015] shows that leaching from stainless steel in fruit juices increases with time, so it can be assumed that any process that requires a long processing time will be at risk of being high in nickel. Acidity is also a factor.
Canning appears to only significantly increase nickel for tomatoes, corn and carbonated drinks; these foods have a few large outliers. It is possible that the outliers are caused by cooking prior to canning, unlined cans, or are the result of an unusually high nickel content in the ingredients themselves.
Popular dietary advice instructs us to improve our health by increasing whole grains, fruits, vegetables, nuts and seeds, and warns us of the danger of eating too much meat and dairy. Since for a portion of the population these guidelines have the opposite effect of decreasing health and well-being, it is clear that they do not apply to everyone. This collection of data can be used as a tool to design a healthy diet. We have applied the Low Nickel Diet Scoring System [Mislankar2013] to this data to create the Rebelytics Low Nickel Diet Scoring System, which provides colour-coded score sheets that are biased toward a user's location. After a diagnosis of nickel hypersensitivity, individualized advice from a nutritionist or doctor can help to achieve a truly healthy diet.