The electromagnetic-fields directive explicitly states that it is protecting against "established health risks" - i.e. it does not protect against interference with medical implants or against any as-yet undetected long-term effects, such as carcinogenesis. Unfortunately, it was only late in the drafting of the directive that the EU became aware that MRI was an employment sector that involved exposure of staff to large magnetic fields, albeit with no detectable adverse effects. In other words, the enormous detrimental effect of the directive on MRI and the knock-on implications for healthcare were not considered in the initial impact assessments.
The directive's exposure limits are largely based on figures provided by the International Commission on Non-Ionizing Radiation Protection (ICNIRP). ICNIRP established itself from within the International Radiation Protection Association in the early 1970s and is now recognized as the authority in this field by the World Health Organization. The relevant occupational exposure limits for the magnetic fields associated with MRI can be summarized as follows:
• RF: specific absorption rate of 0.4 W/kg whole-body average over 6 min.
• Gradients: induced current density of 10 mA/m2 (frequency-dependent).
• Static field: 2 T to the head, and an induced current-density limit of 40 mA/m2 at low frequencies (a limit that can be exceeded by movement in the stray field of a whole-body magnet, even at 1.5 T).
These limits were derived by ICNIRP using a very precautionary approach, extrapolating known effects between frequencies and amplitudes, and considering both short- and long-term potential effects (whereas, as stated above, the EU directive was supposed not to consider long-term effects). However, it is now clear that ICNIRP never intended its limits to be used as absolute occupational exposure limits, but only as guidelines.
It's my opinion that, although data remain limited, a rational interpretation of the latest ICNIRP literature reviews, along with similar reviews from the UK Health Protection Agency (formally the National Radiological Protection Board), is that there is no proven health risk associated with occupational exposure to magnetic fields up to the patient exposure limits in the range 0-10 kHz.
The case for intervention
The initial draft of the directive considered electric and magnetic fields from 0 Hz to 300 GHz. The limit for static magnetic fields was effectively 2 T, which would have immediately halted the development and exploitation of high-field MRI - particularly regrettable given the number of 3 T scanners being installed across Europe at the time. Thanks to last-minute lobbying, though, the MR community managed to get static magnetic fields removed from the directive. This was possible because ICNIRP informed the EU that relevant scientific data had been published since its previous review of the situation. ICNIRP, for its part, is currently updating its static-field advice and it is anticipated that the new limit will be incorporated into the directive when it is revised in 2008.
The other PAD limits remain in place, however. The RF limit will not restrict MR workers, because we would never normally be exposed to RF fields, except perhaps in an interventional procedure. The gradient-frequency limit is a problem, though - and especially so for interventional MRI, where the interventionalist is very likely to be positioned within the gradient coil field so as to be able to reach the patient. Put simply, the PAD will kill off interventional MRI in Europe, at an early stage of its development and before its value has been properly established.
Yet it is also clear that the PAD will affect many more MR workers than just this small group of interventionalists. This is because the gradient fields actually spill beyond the end of the magnet bore for most scanner designs. If you are standing approximately within arm's length of the bore end of most scanners during imaging, you will probably exceed the action value and the exposure limit. In practical terms, this means it will no longer be possible for an anaesthetist to stand by a patient who is unconscious, or for a nurse to stand by a nervous child.
The ultralow frequency limit will even prevent staff moving in the vicinity of the scanner magnets, as such movements will cause them to experience low-frequency magnetic fields and, in turn, potentially to exceed the current-density exposure limit. This means that engineers will not be able to climb into scanners to maintain them without costly ramp-downs of the magnets, and staff may even be prevented from walking around most magnets, or standing by magnets and turning their heads rapidly.
The bottom line is this: if the PAD exposure limits are enforced, interventional MRI patients (and staff) will have to be exposed to ionizing radiation, rather than to a novel procedure that is apparently without risk. Similarly, more children will be exposed to the risks of CT rather than to MRI, because it will no longer be possible to accompany them during the scan.
This situation - where new, safer technologies will have to be replaced by less safe, older ones - has arisen because, despite its known risk, the regulatory position is tipped in favour of ionizing radiation. Of course, for ionizing radiation, exposure limits are based on the "ALARP: as low as reasonably practicable" principle - in contrast with MRI, where the limits seem to be based on the precautionary principle.
Prudent yet practical
A good feature of the electromagnetic fields directive is the use of "action values" as distinct from "exposure limits". The action values are easy-to-measure parameters, and employers can confidently assume that if they are not exposing staff above the action values, then they will not be exceeding the exposure limit (i.e. the absolute limits above which no member of staff can ever be exposed). However, this approach was necessary because the exposure limits are given in terms of parameters - such as induced current density (at lower frequencies) or SAR (at higher frequencies) - which cannot be measured directly in vivo. This is particularly the case for induced current density.
Consequently, to determine whether a certain working practice complies with the exposure limit, employers will have to commission physicists or electronic engineers to mathematically model the current density resulting from a particular magnetic-field exposure. One can imagine that this will be a lucrative scenario for lawyers (and physicists), with different sides in a dispute using different models and getting different results.
By way of summary, I would like to stress that I do not think MRI should be free from any exposure limits. It is clear at this stage that there is no proven risk associated with the technology at the levels to which people are exposed occupationally. Yet it is sensible to remain prudent.
I believe that absolute exposure limits should be set at levels that will avoid known risks, such as RF burns, but that will not necessarily avoid all biological effects (such as dizziness, which has been associated with varying magnetic fields, and which must be considered in risk assessments for certain jobs, of course). I also accept that because of the limited extent of the scientific literature in this field, some time-average limits might be reasonable, but such limits would have no scientific basis, and so would have to be practical and consistent with current working practice and the ALARP principle.
• The House of Commons Science and Technology Committee in the UK has investigated the history of how the electromagnetic fields PAD came into force. Its report, Watching the Directives: Scientific Advice on the EU Physical Agents (Electromagnetic Fields) Directive, takes this as a test case of how scientific advice is used in government. The report was published on 29 June 2006.