01/12/2000 - Two large meta-analyses confirm increased risk of child leukaemia with power-frequency magnetic fields

[Post Comment] [View Comments (0)]

Share News: Slashdot Delicious Digg reddit Facebook StumbleUpon

A Powerwatch commented version of the UK NRPB Response Statement to the new UKCCS study is available.
We believe that this NRPB Response is both scientifically and morally flawed.

UK Childhood Cancer Study published their 'proximity' paper in British Journal of Cancer on 14th November 2000 [1].
"Childhood cancer and residential proximity to power lines" claims "no association" in the summary, but further reading of the paper reveals a 42% increase near to 275 kV and 400 kV high-voltage power lines that is not commented on in the Abstract/Summary at the start of the paper, which is both surprising and disappointing. The Abstract really should represent the important contents of the paper, whereas about half of this lists low ORs of less than unity and yet doesn't bother to print and comment on the OR of 1.42 found when looking for evidence to try to test for the Bristol theories. They were not able to fully check out the latest Bristol University findings as there were not enough cases near to power lines and they didn't have exact enough location data. What the paper shows, though, is that the number of childhood cancers in the UK is not significantly increased by people living near to power lines or other electrical installations. This may well be partly due to the balanced way most UK powerlines are configured which generally results in quite low magnetic fields (see next items re. magnetic fields and childhood cancers), or it may be due to other factors which are discussed below.

However, this study does seem out of line with other well conducted powerline proximity studies and there may be reasons for this. One might be that they only considered the home of the child for the year prior to diagnosis, whereas another (as yet unpublished) part of the study that is looking at the molecular genetics of the heel blood of newly born infants, has been finding genetic marker abnormalities that are associated with leukaemia AT BIRTH, i.e. usually quite a few years before the child actually develops leukaemia. So the relevant exposure period MAY be where the mother lived for the year BEFORE the birth of her child, rather than the year before the child was diagnosed with leukaemia.

The UKCCS claims that it was not able to test the Bristol aerosol up-wind/down-wind hypothesis as they did not have enough cases near to high-voltage overhead power lines, nor did they have precise enough location details. In this paper they list the following:
   132 kV lines: 25 cases and 24 controls living within 200 metres. No apparent effect on incidence.
   275 kV lines: 19 cases and 23 controls living within 400 metres. No apparent effect on incidence.
   400 kV lines: 31 cases and 22 controls living within 400 metres. This does show an effect for 400 kV lines, but the paper reports an adjusted Odds Ratio of 1.05 due to the way they analysed the results. They expected any effects to show up most nearest to the power lines and set the analysis to test this. When they did a separate analysis to partly test for the Bristol aerosol effect (albeit without the main up-wind/down-wind factor) they found a 42% increase within 400 metres of 275 kV and 400 kV power lines, but with most of the cases between 80 and 400 metres from the line and virtually no cases close to the line. This would, if anything, support the aerosol hypothesis of causation.

An earlier British Journal of Cancer Paper [2] confirms high power-frequency magnetic fields ARE associated with a doubling of childhood leukaemia. A major meta-analysis of the original data from a large number of competent studies, including the UKCCS, of 3,203 children with leukaemia and 10,338 children without showed that the  few (62) children exposed to residential power-frequency magnetic fields above 0.4 microtesla (4mG) have TWICE the chance of developing leukaemia compared with the unexposed control children. Relative Risk = 2.0 (1.27-3.13), p=0.002 showing a very high level of confidence in the result.

The UK results did not have enough cases above 0.3 microtesla to show an effect, but it is interesting to see the UKCCS used 'Geometric Mean' rather than 'Arithmetic Mean' (AM = normal average). Geometric Mean is an unusual metric and is the n^th root of all the numbers multiplied together. It tends to ignore relatively few large values whereas one single zero will take the GM to zero!  The more usual Arithmetic Mean is the normal 'average' where you add all the 'n' numbers up together and then divide by 'n'.  Most EMF studies have used A.M.or a time-weighted mean (for areas where the child spent time) and virtually none have used G.M. which always tends to be lower than the A.M.  The UKCCS used a Time Weighted G.M.

Interestingly, a mid 1990's UK National Grid paper by John Swanson and David Renew looked at A.M and G.M for two populations ~ those living within 100 m from 132 kV or above overhead power lines and those who lived further away. Note: uT = microtesla.
For the further away group the results were:
  P(5%) 0.012 uT,   G.M.= 0.036 uT,  A.M.= 0.051 uT,  P(95%)= 0.141 uT
For those living within 100 m the results were:
  P(5%) 0.025 uT,   G.M.= 0.153 uT,  A.M.= 0.973 uT,  P(95%)= 6.453 uT
So, for those living close to overhead power lines, G.M. seems to us to be especially inappropriate as it loses the peaks ~ in this case by a factor of 973/153 or divided by a factor of 6.4.

Another meta-analysis, this time of 15 studies, published in the November issue of Epidemiology [3] finds a 1.7 fold increase in childhood leukaemia at  0.3 microtesla.  This didn't include the UKCCS results ~ in fact the only UK paper was our Coghill, Stewart and Philips paper! 

There have been too few electric field studies published to carry out a meaningful meta-analysis, but electric fields are also under suspicion. The Coghill study and a few US studies have found significant associations with low levels of a.c. electric fields.  Against UK NRPB advice, the UKCCS did add the measurement of electric fields to the second part of the study (with funding provided by the Foundation for Children with Leukaemia), and we await the now overdue results of this with interest.

These meta-analyses provide strong evidence to cause us to DEMAND that a precautionary approach is taken and ambient residential power-frequency magnetic fields are kept below 0.3 microtesla. Percentagewise, very few people live in ambient fields as high as this. It would not cost that much to (i) either change the electricity system to reduce the fields, or (ii) subsidise the removal of these people from the areas of high fields. UK and European ambient power-frequency magnetic field levels from sources outside the home are around 0.03 to 0.05 microtesla, possibly rising to around 0.1 microtesla in cities and large towns. Think about that if people tell you living in high a.c. magnetic fields is OK.  It is estimated that less than 0.5% of Western people live in ambient power-frequency magnetic fields above 0.25 uT ~ do YOU really want to be in that exclusive group?

These results also raise large questions about adult cancers and EMFs. It is likely that fields of this level and above may well influence adult cancers. We already have repeated evidence that fields above one microtesla stop the anti-cancer action of the widely used breast cancer treatment drug Tamoxifen.

[1]   "Childhood cancer and residential proximity to power lines", Jane Skinner et al (UKCCS Investigators), in British Journal of Cancer (2000) 83 (11) 1573-1580 November 2000.
[2]   "A pooled analysis of magnetic fields and childhood leukaemia", Ahlbom, et al, British Journal of Cancer (2000) 83(5), 692-698.
[3]   "A pooled analysis of Magnetic Fields, Wire Codes, and Childhood Leukemia", Greenland, et al, Epidemiology, Vol.11 No.6, 624-634.