I have noticed a growing number of companies offering to measure nutrient levels and then offering a personalized treatment approach to address deficiencies identified. I have also been sent individual blood results from members of the public and asked whether the results can be used to direct the best treatment. Others contact me and tell me their nutrient levels are “normal” so their doctor told them there was no need for additional nutrients.
It is a reasonable question because there are many studies that suggest that people with psychological problems such as ADHD have lower levels of nutrients in their blood relative to the nonclinical population. What we don’t know is whether it is necessary to be deficient in order to benefit from additional nutrients than what you can get out of your diet.
So, what does the research say?
Our lab at the University of Canterbury in New Zealand is one of a few that has specifically investigated whether nutrient levels are predictive of response to a broad-spectrum micronutrient intervention [1, 2, 3]. It is important to note that not many labs take this approach, which is giving a combination of nutrients together and then assess treatment response. Many researchers assume that one must be deficient to benefit from nutrients, and therefore select people for the deficiency and only treat them. We treat everyone, regardless of identified deficiency, and then assess whether the deficiency predicts who will respond and who won’t.
Overall, our research shows that the effect, if there is one, is weak, and certainly not useful at an individual level as a good predictor of treatment outcome.
Here is what we did: We assessed some key nutrients pre-treatment via serum/plasma. We measured vitamin B12, vitamin D, zinc, copper, folate, ferritin, potassium, sodium, calcium, and homocysteine. We have looked at two data sets – an adult study and a child study, both comparing vitamins/minerals to placebo in the treatment of symptoms associated with ADHD.
Findings from the adult ADHD trial:
Participants improved significantly on all outcome measures after exposure to the micronutrients for 8 weeks; 61% were identified as responders.
But, there was no relationship between baseline functioning and baseline nutrient levels. This was a bit surprising given that studies have identified deficiencies in magnesium, zinc and iron in children with ADHD. Surprisingly, we didn’t find that these nutrient levels were highly correlated with ADHD symptoms.
Very few predictors were identified. We found that greater pre-treatment with ferritin predicted who would be an ADHD responder. We wondered if those with higher ferritin had higher inflammation and therefore responded more rapidly to the treatment as the micronutrients may have improved inflammation.
Lower pre-treatment vitamin D predicted greater change on a measure of mood. This finding is not unexpected as low vitamin D levels have been associated with low mood. Pre-treatment copper gave us a signal, but it was weak and mixed.
It is important to note that while there were these small signals, there were still many people with normal levels of these nutrients who benefitted from the nutrient approach, only there were fewer relative to those with vitamin D and copper deficiencies.
No other relationships between baseline nutrient levels and treatment response were identified. In other words, zinc, iron and vitamin B12 pre-treatment did not predict who would benefit and who would not. Further, there were no specific demographic variables (age, socio-economic status, gender, marital status, education) that contraindicated micronutrient treatment for ADHD in adults.
Findings from the child trial:
We identified that 49% of the children responded to the micronutrient intervention. Substantial nutrient deficiencies pre-treatment were observed only for vitamin D (13%) and copper (15%), otherwise, most children entered the trial with nutrient levels falling within expected ranges. Lower pre-treatment folate and B12 levels, being female, greater severity of symptoms and co-occurring disorders pre-treatment, more pregnancy complications and fewer birth problems were identified as possible predictors of greater improvement for some but not all outcome measures although predictive values of all of them were weak. Lower IQ and higher BMI predicted greater improvement in aggression.
It is important to note that levels of folate pre-treatment for ADHD responders were within the normal reference range for folate (>8nmol/L). In other words, the blood tests did not identify responders as deficient in folate, just lower relative to non-responders. Note though, that there were many children with higher B12 and folate who did benefit from the nutrient treatment. No other relationships between pre-treatment nutrient levels and treatment response were identified.
It is also important to point out that across two studies, replication did not occur and any findings we did observe were incredibly modest. As such, they could not be used at an individual level to reliably identify who might benefit from this treatment approach. We see this as good news as it means people don’t have to feel they need to get expensive testing done before trying nutrients. The bad news is that the search is still on to figure out why some people respond and some don’t.
Although not reported in these trials, we have also looked at the predictive value of nutrient levels recorded from hair samples and similarly, the levels were also not overly helpful at predicting treatment response.
Do nutrient levels have to change for benefit to occur?
Now, this is a tricky question. But we have now published a study looking at this very question, that is, whether the change in a nutrient biomarker is correlated with improvement in mental health. Our overall findings were that they were not.
I think this type of question stems from research in medicine such as physicians tracking cholesterol levels in order to determine whether they are associated with the progression of the disease (such as incidence of stroke). Changes in cholesterol levels are used to estimate risk for future cardiovascular events.
In the mental health world, at best, they are weakly correlated with improvement in symptoms and probably not that helpful. We investigated whether changes in serum nutrient levels mediate clinical response to a micronutrient intervention for ADHD. Data were compiled from two ADHD trials (8-10 weeks), one in adults (n = 53) and one in children (n = 38). Seven outcomes included change in ADHD symptoms, mood, overall functioning (all clinician-rated) as well as response status. Changes in serum/plasma nutrient levels (vitamins B12 and D, folate, ferritin, iron, zinc, and copper) were considered putative mediators.
We found that a decrease in ferritin and an increase in copper were weakly associated with a greater likelihood of being identified as an ADHD responder; none of the other nutrient biomarkers served as mediators. Perhaps we need to look to see if other tissue (like hair or microbiome samples) might be more useful. Monitoring these biomarkers is unlikely helpful in understanding clinical response to a broad-spectrum micronutrient approach.
Blood levels don’t necessarily tell us what is going on in the brain and what nutrients are being used and what isn’t being used. We didn’t look at ALL nutrients so we may have missed an important biomarker. It may be ratios that are more important. But next time a professional is keen to track nutrient levels as a proxy for a response, perhaps be a bit skeptical about whether the data support such testing.
Is the term deficiency accurate?
The term “deficiency”, as is often used in the ADHD literature when discussing nutrient levels, may be problematic. Although research shows that the ADHD group mean nutrient levels are often below control group means, the ADHD means are typically still falling within the normal reference range, potentially challenging the use of the term “nutrient deficiency” when attempting to investigate causes of ADHD and in relation to predicting response to nutrients. Given that reference ranges are generally defined as the set of values that 95 percent of the normal population falls within, this does not necessarily mean that these ranges are best equipped to identify what is required for optimal health for any particular individual.
Had functional ranges (the range used to assess risk for disease before the disease develops) been used in these studies, many more would have been identified with “deficiencies”. An important hypothesis that requires further investigation is that some individuals may have suboptimal nutrition for brain health despite having nutrient levels within the reference range. In other words, they might have a nutrient deficiency relative to their metabolic needs rather than relative to general population levels.
It is exciting that the Eat2beNice consortium will be looking at nutrient levels alongside other biomarkers so we can confirm whether these results are replicable. Hopefully, some of the other biomarkers will prove more useful at predicting treatment response. After all, it is a valid clinical question to wonder – when treatment works, who does it work for and why? These types of data inform clinical practice and can help the consumer decide whether you should go for that expensive testing, or not bother. At this stage, I wouldn’t bother.