Does MRI attract the iron in your blood?

Of course, the answer to any headline which asks a question is “No”. Try this with newspaper headlines. If the answer was “Yes!” the headline would read: “MRI attracts the iron in your blood!” But I digress.

Isn’t iron, well, iron? Why isn’t it attracted by the MRI magnet?

Iron in your body is not ferromagnetic. The strong attraction of iron to a magnet is due to ferromagnetism. This is magnetism in the colloquial sense. Ferromagnetism occurs in some materials which have molecules with a permanent magnetic dipole moment. When the material is placed in a magnetic field, molecules with a magnetic dipole moment tend to line up with the external magnetic field. In addition to this, ferromagnetic molecules also interact very strongly with one another (the exchange interaction: a quantum mechanical interaction in which the orbital motion of electrons from separate atoms are coupled), and this causes much more alignment with the external magnetic field. The interaction between molecules is strong enough for domains of aligned molecules to form, even when no external magnetic field is present. In an external magnetic field, the domains of molecules become partially aligned, producing an overall magnetisation in the material. If this material was a red blood cell, you might expect a magnet to attract or orient the red blood cell.

However, the exchange interaction which causes ferromagnetism does not occur unless iron is found in bulk form. In the body, iron is distributed in various chemical compounds such as haemoglobin (in red blood cells), serum ferritin (iron stores) and haemosiderin (e.g. in clotted blood). These compounds are not ferromagnetic; they are weakly paramagnetic.

Other types of magnetism. As well as ferromagnetism, there are a few other types of magnetism. One of these is paramagnetism, in which (like ferromagnetism) molecules act like a permanent magnet. But unlike ferromagnetism, the effect is very short range. This is because the aforementioned exchange interaction between molecules does not occur; the paramagnetic interaction is weaker than ferromagnetism by orders of magnitude. Thermal motion, which causes molecules to be randomly oriented, easily overcomes any tendency of molecules to align; the thermal motion manifests in vigorous molecular tumbling, and collisions with other molecules. In fact, even with the strongest magnetic field we could create, the temperature would have to be as low as a few kelvins to allow a high degree of molecular alignment. Paramagnetism occurs when unpaired electrons are present (that is, electrons without a notional partner in any orbital—see the Pauli exclusion principle).

Another type of magnetism is diamagnetism. All materials are diamagnetic. Orbiting electrons in a material are affected by an external magnetic field. Their orbits change and as a result they generate their own magnetic field which opposes the external magnetic field within the sample (this is Lenz’s Law). But diamagnetism is very, very weak—it is easily masked by paramagetism in molecules with a permanent magnetic dipole moment.

So what about magnet therapy? There is alot of advertising on the web for magnet therapy. The British Medical Journal recently published an editorial (also here) concerning magnet therapy, stating that “magnet therapy has no proved benefits”. This is in reference to trials in which magnets were used for therapy. Even this is quite a generous statement; from a physical (physics) point of view, a therapeutic effect is unrealistic, because any magnetic effect is entirely overwhelmed by the thermal motion, not to mention haemodynamic forces in flowing blood. (As we saw above, ferromagnetism does not feature in the body, and diamagnetism and any paramagnetism are swamped by thermal motion.) Magnets sold for these purposes are relatively weak, and their magnetic fields do not extend significantly through the skin; even if there was a plausible therapeutic effect of such magnets, they wouldn’t affect muscles or joints!

It would be sensible distinguish “magnet therapy” as discussed here from transcranial magnetic stimulation. The former refers to the effects of static magnetic fields from permanent magnets. The latter refers to generating a huge oscillating magnetic field very close to your head, which induces currents in your brain. I visited a Physics department many years ago in which this effect was demonstrated to my amusement. They put a probe near my head and turned the current on and my arms twitched down. They turned it over and my arms twitched up! I wonder if they had ethical approval for such “live” tests…!

What about BOLD? The Blood Oxygen Level Dependent susceptibility effect is a technique which uses the magnetic properties of blood, but note that they are not ferromagnetic properties. Oxyhaemoglobin is only diamagnetic, whereas deoxyhaemoglobin is paramagnetic (four unpaired electrons), and therefore due to an increase in spin dephasing, has a much shorter T2*. (This is a statement about the effect of hydrogen protons coming close to the paramagetic molecule due to thermal motion, which causes a change in the Larmor frequency the 1H experience. It is not a statement about any attraction or torque, which are overwhelmed by the themal motion.) BOLD can be used to locate brain activation; increased blood flow occurring in activated areas decreases deoxyhaemoglobin concentration. BOLD may also be used in certain circumstances to assess poor blood perfusion. During high oxygen demand, ischaemic tissue does not experience the same reduction in deoxyhaemoglobin levels as normally perfused tissue, which benefits from a greater increase in blood perfusion. The difference in the signal intensity change may be used, for example, to assess blood flow.

Perhaps it should also be stated that talk of “alignment” of 1H spins with the external magnetic field in MRI texts does not refer to the physical orientation of their parent molecules. That alignment refers to the direction of nuclear magnetism (ferromagnetism, paramagnetism and diamagnetism are concerned with atomic / electronic magnetism at the molecule level). Nuclear magnetic moments are about a thousand times smaller than atomic magnetic moments (for a given angular momentum).

What about clotted blood? Even in clotted blood, ferromagnetic molecules are not present. When blood leaves the vascular system, the haemoglobin content can rise quickly. This will cause a loss in the MRI signal due to a shorter T2*. Then haemosiderin forms, and the blood clots. Haemosiderin is approximately one-quarter Fe3+, which is highly paramagnetic (it has five unpaired electrons). This will cause signal loss in T2 and T2* weighted images. In a T1-weighted image, the shorter T1 will cause a signal increase. After this, methemoglobin forms which also contains Fe3+ (it is also highly paramagnetic). Finally the red blood cells break down.

Did you know? There’s only about 4g of iron in the average human body.

Further reading: JMRI