108N J. Environ. Monit., 1999, 1

Focus

Genetically modified organisms and monitoring

The genetic modification of organisms forfood use has raised serious concern aboutthe potential for adverse effects on theenvironment, ecosystems and on thehealth of humans and animals. As arelatively new technology, its impactsremain uncertain but could range fromdisturbances to the genetic functioning ofindividual organisms to a reduction in thebiodiversity of farmland. As a result, thequestion of how to monitor for potentialimpacts is beset with problems. The factthat genetic modification can be used on arange of organisms for a variety ofpurposes means that those developingmonitoring systems will need to be asimaginative as those developing GMOs.

In the case of genetically modifiedorganisms (GMOs) for food use, concernhas focussed on the transfer of genes toother organisms, the potential for effectson non-target organisms, or on the healthof humans and animals, and thelikelihood of adverse effects on wildlifedue to changes in farming practice. Aswith other new and unfamiliartechnologies, genetic modification is alsoplagued by the problem of uncertainty.Novel genes are inserted randomly intothe genome of the host organisms, andthis leads to the possibility of unexpectedeffects. Unanticipated environmentaldisasters, such as the concentration ofpersistent organic pollutants inecosystems at high latitudes, havehighlighted the need for monitoringdespite the obvious difficulties inherent inmonitoring for unexpected effects.

Monitoring as a legal requirement

The regulation of the experimental andcommercial release of GMOs is coveredin the European Union by Directive90/220/EEC. This is currentlyundergoing revision, and the revisedDirective will require anyone seekingconsent to market a GM crop or food toproduce and carry out a ‘post-marketing’ monitoring plan. At present,it is proposed that monitoring will beused as a means of evaluatingassumptions made in the riskassessment, and also to identifyunanticipated effects as they occur. Thesecond objective poses a seriouschallenge to both the developers and the

regulators of GM crops and foods,begging the question of whether it ispossible to establish monitoring systemsto look for unanticipated effects.

There are currently 20 applications forconsent to market GMOs in theEuropean Union, only a small number ofwhich have actually gained consent. Butas more GMOs are released into theenvironment, the question of how tomonitor for adverse effects and whatresources this requires will becomepressing. The range of organisms that canbe modified and genes that can beinserted, combined with the difficulty ofpredicting what might happen, meansthat monitoring of GM crops must beclearly defined. The purpose of anymonitoring must be to act as an earlywarning system for adverse effects, and asa basis for action rather than an end initself.

Monitoring at present

It is stated in the proposed revision toDirective 90/220/EEC that ‘experienceand data gained through the monitoringexperience of experimental releases ofGMOs may assist in designing the postmarket monitoring regime’.1 However,an examination of the current procedureused for monitoring experimentalreleases shows that there is unlikely tobe much data of use for post marketmonitoring. Oilseed rape (Brassicanapus) has been shown to be capable ofcross breeding with several species ofwild and cultivated plant found growingin the UK.2 In particular, wild turnip(Brassica rapa) and feral oilseed rape arecommonly found growing in, or adjacentto, oilseed rape fields and gene transferoccurs relatively easily. As a result, therisk of gene transfer from GM oilseedrape is significantly higher than for othercrops, such as maize, which have no wildrelatives in the UK. Despite this, theGovernment’s Advisory Committee onReleases to the Environment (ACRE)has not made monitoring for such eventsa prerequisite for consent to conductGM trials. In fact, consent holders haveonly been required to monitor for GMvolunteers emerging in following years,while monitoring for gene flow has beenentirely at their discretion.

In the case of GM oilseed rape testsites, the majority of monitoring reportssubmitted to ACRE read like agronomicstudies with detailed provision ofinformation about the growth,development and pest infestations of thecrop, but with little to indicate thatpotential environmental impacts wereroutinely monitored for. Typicalcomments from these reports are that‘The agronomic performance of the lineswas satisfactory’3 or that ‘thetransformed crop did not look anydifferent to the non transformed crops’.4Out of 25 monitoring reports for GMoilseed rape releases held on the publicregister at the beginning of 1999, onlyone consent holder mentionedundertaking botanical monitoringaround the release site and only thisconsent-holder mentioned removingplant species related to oilseed rape fromthe vicinity of the trial site. If monitoringprogrammes for GM crops givencommercial marketing consent are to bebased on experience gained frommonitoring around test sites, then thequality and range of this monitoringmust improve significantly.

Test sites cover a tiny area incomparison with the area that wouldneed to be monitored if GM crops startbeing grown on a commercial basis andto undertake such monitoring on morethan a limited scale will requireconsiderable resources. Although thelegal responsibility for monitoring plansis likely to fall upon those developingGM crops, at the very least theregulatory authorities will have toauthenticate and audit them. However,at present the Health and SafetyExecutive does not even have theresources to inspect all of the GM testsites for compliance with consentconditions, a simple task whencompared to overseeing monitoringprogrammes.

The difficulties of monitoringGMOs

It is suggested in the revision ofDirective 90/220 that monitoringprogrammes be developed on a case bycase basis, but effective monitoring will

Publ

ishe

d on

01

Janu

ary

1999

. Dow

nloa

ded

on 2

2/10

/201

4 12

:04:

46.

View Article Online / Journal Homepage / Table of Contents for this issue

J. Environ. Monit., 1999, 1 109N

Focus

have to take into account the possibleinteractions of the increasing range ofcrops that are being modified and thevariety of novel genes which are beinginserted. Returning to the example ofmonitoring for gene flow from a GMcrop, the increasing number of genesbeing inserted will inevitably complicatethe task. In particular, certain crops arebeing modified for a range of traits. Inthe UK, trials have been conducted ofGM oilseed rape varieties that have beenmodified for tolerance to three differentherbicides, for altered oil composition,reduced seedpod shattering, and reduceddisease susceptibility. So far, the fewmonitoring programmes that have beenconducted have concentrated onherbicide tolerance, a fairlystraightforward trait to screen for, buttraits such as disease resistance orreduced pod shattering are not so easyand it may be that the only suitablescreening technique will be DNAanalysis. This is a costly and time-consuming procedure at present. Finally,compatible plant populations are likelyto be exposed to pollen from more thanone GM variety of a crop, furthercomplicating monitoring requirements.

Similar difficulties are likely to befaced in attempts to monitor for otherimpacts. For example, a recentlaboratory study by Cornell Universityfound that pollen from maize modifiedto express a toxin from the bacteriumBacillus thuringiensis (Bt) slowed thedevelopment and increased the mortalityof the larvae of the Monarch butterfly.The Monarch larvae feed on milkweed, aplant commonly found growing adjacentto maize fields in the United States. Thestudy highlights the potential of GMcrops to affect organisms other than thepests that they are designed to target.Similar research into ‘Bt’ maize hasindicated that there may be impacts onbeneficial insects or invertebratesinvolved in the decomposition of plantmaterial in soil.

Crop plants are host to a range ofpest, beneficial and neutral organismsranging from the more obvious, such asthe Monarch, to the virtually unknown,such as the micro-organisms whichoccupy the surface of leaves and roots.Some of these will be more economicallyimportant or ecologically vulnerablethan others and many of them arepoorly understood. It will not bepossible to monitor every organism thatcomes into contact with all the differentGM crops that are being developed.Instead, it may be necessary to establish

indicator species to be monitored ineach region that GM crops are released.However, as more crops are released, itwill become progressively more difficultto establish which crop is causingchanges that are observed.

Monitoring for health effects onhumans looks likely to be an equallycomplex task. To produce anymeaningful results, it will be necessary toestablish population groups withvarying levels of exposure and then toestablish that differences in healthindicators are due to the consumption ofGM foods. GM products have thepotential to be found in a wide range offoods, for example soya is found in 60per cent of processed foods, and soexposure to GM soya could come fromconsumption of a range of products. Itwill be necessary therefore to establishextremely detailed information aboutpurchasing behaviour in order todetermine the level of exposureindividuals or groups receive. The UKgovernment’s Advisory Committee onNovel Foods and Processes (ACNFP)has been considering methods ofundertaking such monitoring since 1998,such as using market survey data toestablish consumption patterns.However there are problems with this,such as how representative this data maybe and whether it would cover a largeenough section of the population toprovide meaningful results. In addition,such data only provides brandinformation and so for the scheme to besuccessful, manufacturers would have toco-operate and provide accurateinformation about which GM productswere included in their brands. It isproposed that monitoring of foodconsumption patterns could becorrelated with systems already in placewhich routinely monitor health eventssuch as cancer, congenital anomalies,still births and birth weights.

A basic problem facing this proposalis that confounding factors such associo-economic status or the localenvironment strongly affect health.Although statistical analysis can to someextent account for these, it will still beextremely difficult to relate health eventsto consumption of specific GM productsand show that one is the cause of theother. As the proposal stresses, this is a‘notoriously difficult and complex’ task.The potential for long time lags betweenexposure and health impacts adds to thedifficulty. The paper presented to theACNFP concludes that the main valueof such a system would be that it ‘could

be rapidly interrogated if some potentialhealth effect came to light, or if somecluster of health events were thought toreflect exposure to novel foods’.5 Inother words, due to the inherentdifficulties, a system for monitoring thehealth impacts of the consumption ofGM foods would quite possibly onlyserve to provide corroboration forhealth effects which had becomeapparent by other means.

Some requirements for futuremonitoring

For the proposed monitoring plans to beof any value, they must be based on abetter understanding of the potentialrisks of GM crops and foods. To startwith, a broader scope needs to be givento the risk assessments presented inapplications for marketing consent. Thiswill at least provide monitoringprogrammes with a better starting point.In the case of ‘Bt’ maize, the laboratoryresearch that highlighted potentialadverse effects was undertaken after theGM crop was given consent formarketing in the European Union. It isessential for the protection of theenvironment that this scenario is notrepeated. Companies wishing tointroduce GM crops must be more openwith their own research and examinemore avenues for potential impacts oftheir crop before it is marketed.

The possible adverse affects of GMcrops may take years to becomeapparent. For example, the transfer ofgenes from GM crops to natural plantpopulations will probably take severalyears, and any resulting changes innatural populations are unlikely to beimmediately apparent. In the UK,populations of birds such as the skylarkand hedge sparrow have declined inconventional crop farming areas. Thesedeclines took decades to becomeapparent and the exact causes are still amatter of contention. In the case ofhuman health effects, it is quite possiblethat exposure of foetuses and neonatescould lead to health impacts later in life.Based on previous experience and thelikely time scale of effects, anymonitoring programmes for GM cropsthat are put into place will have to bemaintained for years, if not decades.

It is essential that monitoringprogrammes do not examineenvironmental or health impacts inisolation. The patterns of uptake of GMcrops by farmers, or of consumption ofGM foods will determine where effects

Publ

ishe

d on

01

Janu

ary

1999

. Dow

nloa

ded

on 2

2/10

/201

4 12

:04:

46.

View Article Online

110N J. Environ. Monit., 1999, 1

are likely to become apparent.Monitoring of the use and consumptionof GMOs can be used to determinewhich species, habitats or groups of thepopulation are especially vulnerable. Forexample, data showing a high relativeconsumption of GM foods by certainsectors of the population would allowhealth monitoring to be strategicallyemployed. Another facet is the necessityto track the changes in agriculturalpractice and food consumption patternsthat are bound to occur regardless ofwhether GM crops are introduced.Existing monitoring schemes, such as theCountryside Survey, will be invaluableas baselines against which to measurechange.

Finally, it must be accepted thatmonitoring of GM crops and foods willnot guarantee their safety, or even thatadverse effects will be detected beforeharm is done. With such a range of GMcrops and foods being developed, it isinevitable that exposure will come fromdiffuse and diverse sources. Even if anadverse effect on health or theenvironment is detected by a monitoringscheme, establishing its cause or relatingit back to a specific GMO is unlikely tobe a simple or speedy process. As ishighlighted by the Government

at present effective monitoring is evenpossible, and the question must bewhether it is appropriate to releaseGMOs into the environment and ourfood chain in the first place.

References

1 Common position paper on the revisionof Directive EEC 90/220, Annex VII.

2 A. J. Gray and A. F. Raybould,Environmental Risks of Herbicide-tolerantOilseed Rape: a Review of the PGS HybridOilseed Rape, Report produced for theAdvisory Committee on Releases to theEnvironment, December, 1998.

3 Monitoring reports provided by PlantGenetic Systems to the Department of theEnvironment, Transport and Regions,covering consents to release GM oilseedrape, reference numbers 95/R15/14,96/R15/16, 95/R15/17 and 95/R15/18.

4 Monitoring report provided byPerryfields Holdings to the Department ofthe Environment, Transport and Regions,covering consent to release GM oilseedrape, reference 96/R27/1.

5 Advisory Committee on Novel Foods andProcesses (MAFF) Post-marketMonitoring of Novel Foods, Minutes ofSub-Group Meeting, 2 September, 1999.

Emily DiamandResearch Officer

Real Food CampaignFriends of the Earth

proposals to monitor for health effects,monitoring will not necessarily preventserious impacts, and may achieve nomore than providing confirmation of anevent after it has occurred.

Conclusion

GM crops and foods are highlycontroversial, and considered by manyto pose a serious threat to human healthand the environment. It is essential thatany monitoring undertaken is fullyresourced, and that it is used as a basisfor decision-making and action, ratherthan as a fig leaf to cover inaction. Theremust also be acceptance by thedevelopers of GMOs, the public andregulators that monitoring in no wayensures prevention of harm from therelease of GM crops and food. Thedifficulties posed by the monitoring GMcrops and foods, combined with thepoor understanding of what the resultsmight actually mean, provide a goodargument for taking a precautionaryapproach. As is common practice forfood additives or food contaminants,appropriate animal or in vitro testingmay have to be conducted prior toformal approval to assess any potentialhuman risk. It is by no means clear that

Focus

Publ

ishe

d on

01

Janu

ary

1999

. Dow

nloa

ded

on 2

2/10

/201

4 12

:04:

46.

View Article Online