The most commonly measured marker of oxidative DNA damage is 8-oxo-7,8-dihydroguanine (8-oxoGua) or its deoxyribonucleoside (8-oxodGuo). Published estimates of the concentration of 8-oxoGua/8-oxodGuo in DNA of normal human cells vary over a range of three orders of magnitude. Analysis by chromatographic methods (GC-MS, HPLC with electrochemical detection (ECD) or HPLC-MS/MS) is beset by the problem of adventitious oxidation of guanine during sample preparation. An alternative approach, based on the use of the DNA repair enzyme formamidopyrimidine DNA N-glycosylase (FPG) to make breaks in the DNA at sites of the oxidised base, gives much lower values. ESCODD, the European Standards Committee on Oxidative DNA Damage, has been testing the ability of different laboratories using a variety of methods to measure 8-oxoGua in standard samples of 8-oxodGuo, calf thymus DNA, pig liver, oligonucleotides, and HeLa cells, and in lymphocytes isolated from blood of volunteers. HPLC-ECD is capable of measuring 8-oxodGuo induced experimentally in calf thymus DNA or HeLa cells with high accuracy. However, there is no sign of consensus over the background level of this damage, suggesting that, even though standard extraction procedures were used, variable oxidation of Gua is still occurring. GC-MS failed to detect a dose response of induced 8-oxoGua and cannot be regarded as a reliable method for measuring low levels of damage. HPLC-MS/MS as yet has not proved capable of measuring low levels of oxidative DNA damage. FPG-based methods seem to be less prone to the artefact of additional oxidation. Although they can be used quantitatively, they require careful calibration and standardisation if they are to be used in human biomonitoring. The background level of DNA oxidation in normal human cells is likely to be around 0.3-4.2 8-oxoGua per 10(6) Gua. An effort should be made to develop alternative, validated methods for estimating oxidative DNA damage.