House of Commons - Science and Technology

Select Committee on Science and Technology Written Evidence

APPENDIX 18

Supplementary evidence from Dr Stephen Keevil, Consultant Physicist, Head of Magnetic Resonance Physics Guy's and St Thomas' NHS Foundation Trust

EU PHYSICAL AGENTS (ELECTROMAGNETIC FIELDS) DIRECTIVE

1. What indications were given by the MR community that it was content with the Directive following the removal of static fields; and when concerns about the inclusion of time-varying fields were first raised with the HSE or NRPB/HPA? (Q 868)

To the best of my knowledge no such indications were given. Indeed the question did not arise, since the community was advised that the HSE would seek to retain the static field limit in UK legislation, even after it was removed from the Directive (see Q 905).

Concerns about the time-varying field limits were first raised by the European MR equipment manufacturers in a communication to the Social Questions Working Party in April 2003. I believe that the HSE represented the UK on this working party and therefore would have become aware of the issue at that time. The earliest written record I can find of communication directly between the UK MR community and government agencies on this issue is in a letter from IPEM to MHRA in June 2004, but there were a number of earlier meetings with the HSE, for which there is no written record, where the issue may have been raised (see original submission from MR community, Annex A). The time-varying field issue was certainly discussed at the HSE stakeholder meeting in July 2004 and on a number of subsequent occasions when I was present.

There appear to have been communication problems within the HSE, so that senior policy staff were not aware of what the MR community was being told about the static field issue by HSE representatives, or what the same HSE representatives were being told about time-varying field issues by the MR community at European level.

2. You undertook to let the Committee have any comments on the transcript of the evidence session the Committee held in Brussels on 11 May.

Comments are attached as Annex A. I have also taken the opportunity to expand briefly on my answers to some of the questions posed during the session on 17 May, and this is attached as Annex B. I am enclosing with this letter a copy of a manuscript submitted to the Institution of Engineering and Technology Seminar on the Physical Agents (EMF) Directive which addresses the central question of whether there is evidence that the Directive will impact on MRI practice.

Finally, I am aware that the various ways in which EMF in different frequency bands are used in MRI, compounded by the different ways in which exposure limits are set in each of these bands, make detailed discussion of the impact of the limits complex. This situation has not been helped by some confusing statements made in Brussels. I hope the following table may help.

Frequency

Exposure limit Action value for magnetic flux density Estimated maximum occupational exposure in the UK

Static magnetic field (present at all times for mid and high field systems) 0 Hz None 0.2 T 3 T (clinical)
7 T (research)

<1 Hz (typical) (generated by movement of subject) Current density 40 mAm-2 to head and trunk 0.2 T Possibly up to several hundred mAm-2

Switched gradients (present during imaging) 500 Hz (typical) Current density 10 mAm-2to head and trunk 50 T 2,000 T (to head) (60 mAm-2 approx?—possibly higher)

RF field (present during imaging) 10-400 MHz SAR 0.4 Wkg-1 whole body average, averaged over six minutes.
SAR 20 Wkg-1 to the limbs, averaged over six minutes. 0.2 T <0.4 Wkg-1 whole body average in most conceivable situations.
Local SAR may approach limit in some instances.

MRI uses EMF in three distinct frequency ranges—a static magnetic field, time-varying magnetic fields in the frequency range 100s-1000s Hz (known as switched gradients), and radiofrequency (RF) fields (10s-100s MHz). Typical frequency values have been shown in the table in the case of time varying fields, as the limits are frequency dependent in this part of the spectrum. In the case of the static field, there is no exposure limit in the Directive but concern arises because movement through the field exposes staff to a slowly time-varying field which induces currents that almost certainly exceed the limits at 3 T and quite possibly also at 1.5 T. Where possible, references supporting the figures in the right hand column are given in Annex B in relation to question Q 810.

May 2006

Annex A

Comments on uncorrected transcript of oral evidence taken before the Science and Technology Select Committee, 11 May 2006

The expertise of the MR community

1. In a number of places (Q 678, 683, 777, 788, 794, 797) Mr Biosca seeks to draw a distinction between, on the one hand, an expert scientific community that supports the ICNIRP limits and, on the other, a community of medical users of MRI who he implies are ignorant of the issues around exposure to EMF in their practice.

2. In fact the MR community is a diverse body of medical practitioners, clinical and basic scientists with a wide range of relevant expertise and experience. Those who have made representations about the Physical Agents Directive in the UK comprise physicists, medical physicists, biochemists, physiologists and other biological scientists, many of them of professorial status. They include a Nobel Prize winning physicist (Sir Peter Mansfield) and both the present Chief Executive of the MRC (Professor Colin Blakemore, a professor of physiology with directly relevant research interests) and his immediate predecessor (Sir George Radda, a professor of biochemistry). [23]

3. Regarding input by manufacturers, Mr Biosca states that:

In my view, manufacturers have not said anything at all, they are very happy with the Directive so far (Q 683).

So far, neither from Siemens nor Philips. We do not have any complaints to the Commission so far from manufacturers of magnetic resonance equipment (Q 684).

4. Although these comments are contradicted later in the evidence (Q 742-749), it is perhaps worth reiterating that Siemens and Philips made representations about the impact of both the static and time-varying field limits on MRI to the Commission's Social Questions Working Party as early as April 2003. I can provide the Committee with a copy of this material if required.

The nature of the ICNIRP guidelines and the evidence behind the limits

5. Mr Biosca presents the ICNIRP 1998 guidelines as justification for the exposure limits contained in the Directive (eg Q 752), implying that the limits are supported by a large quantity of peer-reviewed publications (Q 712-715) and are the settled opinion of the relevant expert community. It is true that there is a large body of published research on EMF exposure, but this literature reveals a wide margin of uncertainty rather than agreement. Most published studies report negative results and many of the positive results, some of which were published over 20 years ago, lack replication. There are some established effects, such as magneto-phosphenes, but these relate to biological effects rather than adverse health effects.

6. Even in 1998, the view of ICNIRP was rather more equivocal than the Directive and Mr Biosca imply.

7. By 2004 ICNIRP accepted that:

8. And in the same year NRPB described adoption of the ICNIRP limits as:

. . . a cautious approach . . . to indicate thresholds for adverse health effects that are scientifically plausible[26]

9. The MR community has consistently maintained that, contrary to the view presented by Mr Biosca, the ICNIRP guidelines represent a cautious interpretation of limited scientific data, and should be treated as guidelines to consider alongside other factors as part of a wider risk assessment, rather than used as the basis for statutory exposure limits. [27]In his evidence to the Committee on 17 May, Dr McKinlay (a former Chair of ICNIRP) vindicated this position, stating that

ICNIRP does exercise caution in coming to its advice on guidelines; that is intrinsic in the way ICNIRP operates. It is dealing with the health of people and it does exercise caution, both in interpreting the signs [sic] and in arriving at the guidelines that we give. I would emphasise that they are guidelines. We do not recommend legislation and we do not recommend—such as in the EC directive—regulations (Q 866).

10. The MR community maintains that there is no evidence that adverse health effects occur at the exposure limits stated in the Directive, and that this assertion is supported by the fact that hundreds of millions of patients have been exposed to MRI over the past 25 years, at gradient field amplitudes up to 100 times the occupational exposure limit, with no evidence whatsoever of harm. This evidence relates to patient exposure, not exposure of workers, but it would be perverse to suppose that the physiology of the two groups differs fundamentally!

11. Later in the session comparison is drawn between the IEEE exposure guidelines (not IEE as stated) used in the USA (set by the International Committee on Electromagnetic Safety—ICES) [28]and those of ICNIRP. Mr Biosca says of the ICES limits that

They follow the same basic restrictions as ICNIRP but they set what they call the maximum permissible exposure . . . (Q 803).

12. The ICNIRP basic restrictions are referred to as exposure limit values in the Directive. As I stated in oral evidence (Q 828-829), it is not true that the basic restrictions set by the two organisations are the same over the frequency range of concern to the MR community. ICNIRP and ICES express their basic restrictions in different ways, but Reilly[29] has shown that the two sets of limits are quite different in the frequency range up to 3 kHz. For example, at 1000 Hz the Directive exposure limit (ICNIRP basic restriction) is an induced current density of 10 mAm-2, but the lowest ICES basic restriction at that frequency corresponds to approximately 180 mAm-2. The range of uncertainty in the scientific evidence is demonstrated by the fact that two expert bodies working with essentially the same evidence base are able to propose limits differing by a factor of almost 20.

13. ICE's Maximum Permissible Exposure (MPE) is analogous to the ICNIRP reference levels (action values in the Directive), and only of importance as a tool to demonstrate compliance with the exposure limits (basic restrictions). Hence discussion of the differences between how MPE and reference levels are calculated is something of a red herring in terms of exposure limitation.

Static magnetic fields

14. In relation to exposure to strong (7-8 T) static magnetic fields, Mr Biosca states:

They are doing experiments with volunteers and in order to walk the room with the patient—it is a room like this—they had to spend 20 minutes because if you move fast within the magnetic field you get induced currents in the body and you have disturbances in the brain because there is a migration of calcium ions in the neurons and you become really, really sick and you can fall.

15. I have discussed this point with Professor Penny Gowland, a professor of physics who works with the 7 T MR system in Nottingham (not Norfolk, as stated in the evidence). She assures me that this is completely untrue and staff move in the scanner room at perfectly normal speed without ill effects.

16. It is true to say that currents are induced in the body and the head on moving through a static magnetic field, and the limits contained in the Directive in this regard will be a serious problem for MRI. It is also true that many people report feelings of vertigo on moving through strong magnetic fields. This is a transient and harmless phenomenon, often compared to motion sickness. It is believed to be due to interaction with the organ of balance in the inner ear. [30]I am not aware of any evidence that it is due to migration of calcium ions in neurons as stated, and to suggest that the effect is so serious that it takes 20 minutes to walk across the scanner room is a gross exaggeration.

Time-varying fields

17. Mr Biosca states that the MR community was initially concerned only about the static field limit but has now changed its position to one of concern about pulsed fields in the frequency range 100-500 kHz (Q 770-774).

18. In fact, concerns were raised about the time-varying field limits as early as April 2003. The frequencies involved are up to about 3,000 Hz, not 100,000-500,000 Hz as stated.

19. Mr Biosca states that in this frequency range:

. . . the only thing you have is electro-stimulation which can go very, very high and induce a current of 1,000 amperes a square metre and you would have fibrillation of the heart (Q 774).

20. It is perfectly true that electrical stimulation occurs at very high induced current densities—peripheral nerve stimulation (PNS) at around 1 Am-2 (I assume this is what Mr Biosca is referring to, and that 1,000 Am-2 is an error in the transcript), cardiac stimulation at higher levels still. Obviously these effects are highly undesirable in either patients or staff, and MR scanners are designed according to international standards to prevent them. [31]The MR community agrees that PNS is the only effect for which there is evidence in this frequency range, and would have no issue with limits set at the PNS threshold. However, the limit set in the Directive (0.01 Am-2 over most of the relevant frequency range) is two orders of magnitude below this level.

21. He also states that:

At the highest frequencies in the range of 300 KHz and 500 KHz you have burns (Q 774).

22. EMF at these frequencies is not used in MRI. Radiofrequency fields at 10s to 100s of MHz are used, and at sufficiently high intensities they can certainly cause burns. Again, systems are already designed to minimise this risk for patients and hence also for staff. Burns do not occur anywhere near the specific absorption rate (SAR) limit in the Directive—0.4 Wkg-1, which corresponds to a rise in temperature of approximately 0.1oC.

23. Comments were made to the effect that exposure to time-varying fields could be made compliant with the Directive by limiting the time a worker uses MRI equipment (Q 683, 763, 764, 779, 780). It is important to make clear that this is an option only for radiofrequency fields. The MR community's greatest concern is in relation to lower frequency time-varying fields. At these frequencies, up to about 3 kHz, the Directive allows no scope for time averaging, so the exposure limits represent absolute limits that cannot be satisfied in this way. [32]

Evidence that the Directive will impact on MRI

24. In his evidence Mr Biosca repeatedly and vehemently denied that the Directive will affect MRI practice (Q 678, 683, 690, 691, 693, 717, 733, 735, 759, 794, 797). He claims that the reason adverse health effects have not been observed in MR workers is that they are invariably exposed below the limits (Q 733). The MR community believes that there is substantial evidence that exposure in MRI exceeds the limits, which is summarised in Annex B (additional response to Q 810) and in the attached paper. [34]

25. Mr Biosca said in evidence that ICNIRP have stated "No. In no circumstances do medical personnel with currently installed magnetic resonance equipment in hospitals, which goes up to three teslas, get exposures over the limit values" (Q 678).

26. ICNIRP does not have expertise on MR system design or use. In his evidence to the Committee on 17 May Dr McKinlay stated that:

I know of no evidence like that, or advice that ICNIRP gave to the Economic [sic] specifically on MRI. As I said at the beginning, we do not concern ourselves with exposure to the particular device with a particular frequency (Q 887).

27. In relation to interventional MRI specifically, Mr Biosca states:

Even in the operating theatre when they use these machines, the machines that are used for brain surgery or heart surgery are very low powered machines, a maximum of 0.4 teslas. You can go to the manufacturers and they will tell you that (Q 691).

28. This is untrue, as I am sure the manufacturers will readily confirm. There are two 1.5 T MR systems used for intervention in the UK (at Guy's and St Thomas' Hospitals), with two more being installed in the near future (at Great Ormond Street and Queen Square). Similar systems are in place in other European countries, and in the USA intervention is performed using magnets up to 3 T. There are also low field systems (0.2-0.3 T) with open architecture dedicated to interventional use, but these have relatively poor performance and the trend now is towards higher field (1 T) systems even in the open scanner market.

29. Describing use of interventional MR systems, Mr Biosca states:

Secondly, there are machines that when they are permanently installed in operating theatres they are hidden below the table and when the doctor needs to see an image after doing a certain operation it comes out from under the table and it is switched on (Q 691).

30. I am not aware of any installation that fits this description—there may be one somewhere in the world, but it is not the case generally. Most MR systems are based around superconducting magnets, which are always switched on. The switched gradients and radiofrequency field are only present during imaging, but in many cases the intervention is performed directly under imaging guidance.

31. Regarding exposure of interventionalists, he states that:

There is very limited exposure. Do not ever think that the doctors in operating theatres use the machines continually throughout the whole operation because that is not true. They use it for a very limited amount of time, a maximum of five minutes and no more (Q 691).

I am not sure on what basis Mr Biosca believes he knows more about interventional MRI practices than the professionals who have developed and perform these procedures. For the record, in my own institution, interventionalists are exposed to time-varying fields continuously for up to around 20 minutes. However, since the time-varying field exposure limits that we are most concerned about are absolute, without scope for time averaging, the length of exposure is irrelevant.

MR safety in hospitals

32. Mr Biosca acknowledges that hospitals have stringent MR safety procedures in place (Q 690), and of course these procedures include measures to limit risk due to attraction of ferromagnetic objects (Q 680, 731) (not `any metallic material' (Q 680)—many metals are not ferromagnetic and are quite safe in the MR environment). Safety procedures also include use of hearing protection by staff who remain close to the scanner during imaging (Q 690). Establishment of these safety precautions in UK hospitals is the responsibility of medical physicists, who have been prominent in the MR community's representations about the Directive.

Annex B

Supplementary material relating to oral evidence taken before the Science and Technology Select Committee, 17 May 2006

Q810 Dr Harris: . . . Can you help us by identifying that the opinion you give, that these action limits and maximum limits would impact on diagnostic and therapeutic interventional procedures, are published somewhere in the form of evidence, in peer-reviewed scientific journals, with a conclusion to pass muster which states: "These limits would interfere"?

In answering this question I omitted to mention a number of published, peer reviewed papers reporting the results of simulations of exposure to time-varying magnetic fields in and around MR scanners. Because the exposure limits are expressed in terms of induced current density, which cannot be measured directly, simulation is the only means of investigating compliance with the limits. Work by Crozier's group[35][36] demonstrates that exposure to very low frequency fields due to movement through the static field will almost certainly exceed the relevant exposure limit at 3 T and very likely also at 1.5 T. This will potentially affect all clinical and non-clinical activities in the vicinity of MR scanners. There is also some work on exposure to the switched gradient fields which, while less directly relevant, demonstrates that the occupational limits are breached for patients inside the scanner and therefore probably also for staff standing close to the bore during imaging. [37][38] I am appending a recent manuscript that explores this evidence in more detail. [39]The impact of the Directive in MRI has also been discussed in two other refereed publications, [40][41]

Q826 Mr Flello: I want to pick up on something you said a few months [sic] ago in terms of the dizziness effects that have been noted. You said there was some suggestion that it was to do with calcium ions but then I think you said, "We think it has more to do with the inner ear." Do you have any evidence on which this conclusion is based?

The reference for this is work by Schenck, [42][43] cited in the very recent WHO monograph on static field effects. [44]However, while there appears to be consensus that the organs of balance in the inner ear are involved, the precise mechanism of interaction remains uncertain. [45]