The following symptoms and complaints did show a statistical negative correlation in occurrence". 

breathlessness

readiness to cry

A negative correlation means that these symptoms were actually less likely to occur with higher serum concentrations of mercury.  This should not be taken to mean that mercury should be used to "cure" these symptoms.  Even though the correlation was considered significant, the actual correlation was still very small (p=0.004 and p-0.016 respectively).  These correlations disappeared when the patient's age and education level were included in the analysis.

In 1980-81, the same correlation analyses were carried out with similar results.  The following symptoms and complaints showed no statistical increase or decrease in occurrence since the study began in 1968:

The following statistical complaints did show another negative correlation in occurrence:

Over-exertion

Poor appetite

Once again, a negative correlation means that these symptoms are less likely to occur in a population with a higher serum mercury.  Unlike the negative correlations seen in 1975, these correlations did not disappear when the patient's age and education level were included in the analysis.  Even so, this should not be taken to mean that mercury should be used to "cure" these symptoms.

2. Blood serum and urine variables in relation to serum mercury levels

In the 1968-69 study, several blood related variables showed statistically significant differences that varied with the serum mercury levels, while the 1980-81 study showed the following:

• Blood hemoglobin concentration showed a positive correlation in 1968-69, and a negative correlation in 1980-81.  Hemoglobin is essentially a measure of iron in the blood, and a positive correlation means that the presence of higher levels of mercury in the serum of these study subjects tended to be correlated with MORE iron in the blood.  In other words, these patients with higher serum levels of mercury were were less likely to be anemic than those with lower levels.  Once again, mercury does NOT cure anemia.  This correlation, while significant, was still relatively small.  In the 1981 study, this correlation reversed itself and became a negative correlation.  The meaning of these contradictory results is not apparent.

• Erythrocyte Sedimentation Rate (ESR) was negatively correlated with serum mercury levels in 1968-69,but not correlated in 1980-81.  Sedimentation rate is a clinical laboratory test that is roughly correlated with the level of inflammation found in various disease processes (especially autoimmune diseases).  This test is non specific in nature, and is used in conjunction with other tests to measure the severity of inflammatory diseases.  Negative correlation suggests that higher levels of mercury tend to reduce the level of non specific inflammation.  Once again this result does not suggest that mercury acts as an anti-inflammatory drug.  since the correlation disappeared in the later study, the meaning of the earlier correlation is not apparent.

• Serum levels of potassium were negatively correlated with serum mercury levels in the 1968-69 study, but not correlated in the 1980-81 study.  The actual decrease in in potassium levels caused by higher serum mercury levels is extremely small and the practical significance of this finding is uncertain.

• Serum levels of triglycerides were negatively correlated with higher levels of serum mercury in the 1968-69 study, but not correlated in the 1980-81 study.  Triglycerides are a measure of lipids (soluble fats) in the serum.  Triglycerides are medically associated with cholesterol, which is also a form of soluble fat.  A negative correlation suggests that the levels of triglycerides decreases with increasing levels of serum mercury.  Once again, the decrease in the earlier study is of little practical significance, especially since the correlation disappeared in the later study.

• Serum levels of vitamin B-12 were positively correlated with the level of serum mercury.  This factor was measured only in the 1968-69 study.  Vitamin B-12 is actively involved in the metabolism of mercury, and the authors assumed that this fact accounts for this correlation. 

• Serum levels of IgE and IgA showed positive statistical correlation with the level of serum mercury, however, this correlation was somewhat confused disappearing in one case when background variables were included in the analysis, and appearing in the other case only when  background variables were included.  IgE and IgA are different types of antibodies and as such are an indication of the function of the immune system.  The correlations are small, and judging by the lack of consistent correlation between serum levels of mercury versus the list of symptoms and complaints above, they do not appear to have a measurable practical effect on the general health of the subjects. 

3. Endpoints (disease states developed) during the follow-up period

A total of 253 women (out of the original 1462) died during the 24 year study and follow-up. 87 women developed myocardial infarction (heart attack) and of those, 39 died.  new signs of stroke were observed in 77 women, 77 developed diabetes, and 208 were diagnosed with cancer. 

There was no statistically significant correlation between serum mercury levels and the different diseases (heart attack, stroke, diabetes and cancer).  On the other hand there was a statistically significant negative correlation between mortality (death) and serum mercury levels when background factors were included in the analysis.  This means that higher serum levels of mercury were correlated with a lesser likelihood of death during the study period.  Again, we are talking relatively weak correlations and readers should not infer that mercury somehow prolongs life or immunizes against death!

Discussion

The following discussion is composed of quotes from the original paper:

• The present study does not indicate any unfavorable influence of serum mercury on symptoms experienced, incidence of disease, or mortality in the general population. 

• In three previous reports, the authors have related the number of amalgam fillings to different symptoms, incidence of early death and certain diseases, and to laboratory variables without finding any adverse effects of amalgam fillings as far as these variables are concerned. 

• The only positive statistically significant correlations with symptoms and complaints were those observed between over-exertion and poor appetite, on the one hand, and serum mercury on the other, both being negative, thus indicating a lower prevalence of the symptom with higher serum mercury. 

• One possible reason for the correlation between higher serum mercury and lower mortality may be the result of the protective effect from fish consumption. 

• The positive correlation between increased serum levels of mercury and increased levels of serum vitamin B-12 could be a result of vitamin B-12's roll in the metabosism of mercury, or it may reflect the fact that fish is a dietary source of both mercury and vitamin B-12.

• IgA is responsible for mucosal immunity and is thus the first line of defense for the majority of infections.  The fairly weak association between serum IgA concentration and serum mercury levels in the present study may be a chance finding, but it could also indicate an immunological response to mercury from amalgam fillings.  However, from the authors' data, they cannot reveal any impact of clinical significance of mercury uptake from amalgam fillings as reflected by serum mercury or number of fillings. 

• We conclude that there were some associations between biological variables (serum concentrations of IgA and B-12) and serum mercury levels.  Our main conclusion from our prospective population study, however, is that it gives no support for an unfavorable influence of the serum mercury levels on experienced symptoms, mortality, or incidence of diseases in a general population of middle-aged or older women.