Urinary 8-OHdG as a Biomarker for Oxidative Stress: A Systematic Literature Review and Meta-Analysis

1Oxidative stress reflects a disturbance in the balance between the production and accumulation of reactive oxygen species (ROS), and an overproduction of ROS has negative consequences for cell physiology. When ROS concentration is in excess, oxidative damage to proteins, lipids, and DNA occurs, thus causing structural and functional cellular changes. DNA damage is usually repaired primarily via the base excision repair pathway, and oxidized products are excreted in urine. 8-hydroxy-2-deoxyguanosine (8-OHdG) is one of the most widely studied oxidized metabolites and is considered as a biomarker for oxidative damage of DNA. The formation of 8-OHdG by oxygen radicals was first reported in 1984 by Kasai and Nishimura.

2The interaction of the hydroxyl radical, the most important oxygen-free radical, with the nucleobases of the DNA strand, such as guanine, leads to the formation of 8-OHdG (Figure 1).

3Some diseases, such as cardiovascular or chronic obstructive pulmonary diseases (COPD), have been associated with excessive concentrations of 8-OHdG. 8-OHdG levels also increase due to smoking, aging, or occupational exposure to physical, chemical, or biological substances.

4A recent study suggested that 8-OHdG had high intraclass correlation coefficients (0.96), reproducible measurements, and low coefficients of variation and was the most suitable biomarker of oxidative stress in spot urine samples. Concentrations of urinary oxidative stress biomarkers have been proposed as an effect biomarker to survey populations exposed to xenobiotics such as particulates, oxidizing agents, and lately, engineered nanomaterials.

5Measuring urinary 8-OHdG has some advantages as it is very stable in urine, it is noninvasive, and its excretion is likely to reflect the oxidative DNA damage and can be assessed by two main analytical techniques: mass-based methods (using either gas (GC) or liquid (LC) chromatography) and immunological methods. Another source of 8-OHdG in urine is DNA polymerase-dependent incorporation of 8-oxodGTP from the nucleotide pool. Chromatographic methods are considered to be the gold standard; however, immunological techniques, which are less costly and time-consuming, are widely used because enzyme-linked immunosorbent assay (ELISA) kits have been developed for rapid detection and quantification of 8-OHdG.

6A background range for 8-OHdG has been reported in different studies for healthy persons. However, these studies reported a wider range of values, making the identification of background cut-off values challenging.

7Therefore, the systematic review and meta-analyses of the reported values appears the most appropriate approach to bypass this issue. Our objective was to assess background ranges for urinary 8-OHdG concentrations in healthy adults.

8Chemical methods were used in 44 of the 128 study subgroups, and immunological techniques were used in 84 (Table 1). We decided to stratify the subgroups by body mass index (BMI), gender, and smoking status reported.

9We retrieved 1246 articles, included 84 articles, and considered 129 study subgroups (Figure 2, Table 2, Table 3, Table 4 and Table 5) in the quantitative synthesis, which we stratified by main quantification techniques: immunological and chemical methods. For subgroups evaluated with the chemical methods, 31 studies had participants with a mean BMI between 18 and 25 (14 study subgroups of nonsmokers and 2 study subgroups of smokers) (Figure 3, Table 2). Nine studies had participants with a mean BMI > 25 (three study subgroups of nonsmokers and two study subgroups of smokers) (Figure 4, Table 3). The mean BMI was unknown for four study subgroups.

10For subgroups analyzed with immunological techniques, 47 studies had participants with a mean BMI between 18 and 25 (24 study subgroups of nonsmokers, no study subgroups of smokers and 6 study subgroups with unknown smoking status) (Figure 5, Table 4). Twenty-six studies had participants with a mean BMI > 25 (13 study subgroups of nonsmokers and 6 study subgroups of smokers) (Figure 6, Table 5). The mean BMI was unknown for 11 study subgroups. Supplementary material provides detailed information on the criteria used for the quality assessment (S1) and on the quality level of each included study subgroup (S2). Overall, two study subgroups (1.8%) were classified as low quality, 66 (58.4%) as moderate quality, and 45 study subgroups (39.8%) were of high quality.

11As between-study heterogeneity was much larger than the between-subject heterogeneity, we decided to use a mixed model with study ID as a random effect. The IQR of subgroup-specific GM in subgroups with a mean BMI ≤ 25 with 8-OHdG measured using chemical methods was 3 to 5.5 ng/mg creatinine (Table 1). IQR of subgroup-specific GM in subgroups with a mean BMI > 25 measured using immunological methods was 5.9 to 19.8 ng/mg creatinine (Table 1).

12We compared urinary 8-OHdG concentrations by smoking status within the study subgroups analyzed with chemical techniques and found that for study subgroups with mean BMI ≤ 25, smokers were 2.84 ([2.56, 3.16], p < 0.0001) times greater compared to nonsmoker study subgroups.

13For study subgroups with mean BMI > 25, smokers were 1.61 ([1.17, 2.23], p = 0.004) times greater compared to the nonsmoker study.

14No consistent effects of BMI and gender were observed in our mixed model either for chemical or immunological methods. Gender and BMI seem to not influence urinary 8-OHdG concentrations.

15We found that urinary 8-OHdG concentrations in smokers were greater than in nonsmokers when analysis was conducted with chemical techniques. However, in the population with mean BMI between 18 and 25, this finding was mainly due to one study and needs to be confirmed. The absence of BMI effect on 8-OHdG in urine is in line with data from Lee et al. 2010.

16The IQR range for 8-OHdG in urine given in this meta-analysis is in line with two other studies trying to define reference values for the Italian population (female: 3.25–6.85 ng/mg creatinine; male: 2.9–5.5 ng/mg creatinine). The absence of gender effect observed for 8-OHdG in this study is in line with data from the Italian population but in contradiction with two others.

17The analysis of the data was difficult due to the diversity in study design, analytical methods (chemical or immunoassay techniques), statistical analysis, and data presentation in studies included.

18The heterogeneity in techniques used to quantify urinary 8-OHdG makes it more difficult to compare data between laboratories.

19Chemical techniques are superior to immunological techniques due to their sensitivity and specificity. Chemical techniques require expensive instruments and trained users, but we recommend using chemical quantification methods as standard methods for future studies of biomonitoring.

20Most studies used spot urine samples for 8-OHdG rather than 12- or 24-h collection. However, 8-OHdG levels showed fluctuation during the day under oxidative states, but good correlations have been observed between levels of 8-OHdG in spot morning urine and levels of 8-OHdG in the 24-h urinary collection. Therefore, we included studies reporting spot morning urine, 12- or 24-h urinary samples. The first morning urine void is particularly valuable because it provides a time average for biomarker concentrations that may occur during the hours of sleep (approximately 8 h) and is also relatively free of dietary, physical, and environmental exposures. A significant increase in time in the urinary 8-OHdG during the first part of the day was recently reported among smokers. To make it easier to compare results between studies, we recommend collecting spot morning urine.

21We confirm that smokers have a significantly greater concentration of urinary 8-OHdG, as has been previously reported in the literature. The concentration differences need to be quantified, but with only a few studies in smokers available, this cannot be done at the present time.

22We emphasize here that the values we report are for a healthy population. We were not able to analyze parameters previously reported to influence 8-OHdG concentrations such as occupation, pregnancy, special diet, vitamin, and physical activity due to the limited number of studies with such data.

23The fluctuation in urine flow rate could in fact affect the assessment of urinary 8-OHdG. The urinary 8-OHdG concentrations need to be normalized by urinary creatinine concentrations for healthy adults. Different studies indicated a correlation between excretion of creatinine and 8-OHdG. In addition, normalization with creatinine for spot urine can be considered as a surrogate for the 24-h excretion of 8-OHdG.

24To reach consensual background of urinary 8-OHdG values, harmonization of the unit (ng/mg creatinine) is needed. Harmonization of the statistical reporting of the results is also recommended (geometric means (GM) and geometric standard deviations (GSD)). We suggest reporting the median and the 1st and 3rd quartile as GSDs are not easy to interpret.

25We established a standardized protocol for systematic review and meta-analysis for a set of biomarkers of oxidative stress. This protocol was registered in the International Prospective Register of Systematic Reviews (registration number CRD 42020146623) and described in detail by Hemmendinger et al.. The protocol was then adapted for each biomarker depending on the biological matrix focused, here the urinary 8-OHdG. The methods and results of this study are reported following recommendations from Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations.

26The search strategy was done with a medical librarian. The MeSH (Medical Subject Headings) terms from the PubMed database and free text words were combined. The complete search string was: (“Smoking/urine”[Mesh] OR “Urine”[Mesh] OR Urine*[tiab] OR Urinary[tiab] OR Urinal*[tiab]) AND (“8-oxo-7-hydrodeoxyguanosine”[Supplementary Concept] OR 8-OHdg[tw] OR 8ohdg[tw] OR 8-oh-dg[tw] OR 8-ohg[tw] OR 8-OH-2dG[tw] OR 8-hydroxydeoxyguanosine[tw] OR 8-hydroxyguanine[tw] OR 8-hydroxy-g[tw] OR 8-hydroxy-dg[tw] OR 8-hydroxy-guanine[tw] OR 8-hydroxy-2-deoxyguanosine[tw]) NOT ((“Child”[Mesh] OR “Infant”[Mesh] OR “Adolescent”[MeSH]) NOT “adult”[MeSH]) NOT (animals[mh] NOT humans[mh]).

27The search was performed on 7 May 2019. Rayyan, a systematic review web application, was used for title and abstract screening. We selected the studies in a stepwise process as depicted in Figure 6. To be included in the analysis, a study had to be in English and to provide urinary 8-OHdG concentrations in healthy adults (ages 18—no upper age limit) populations. We excluded non-human studies, in vitro studies, reviews, letters, expert opinions, and editorials. We read the eligible articles in depth, and only studies with original data from healthy (no known disease) adult populations were included in the statistical analysis. All techniques used for the quantification of 8-OHdG were included and classified accordingly. We excluded studies with coefficient variation <10% or >200%. We also excluded data suspected to have unit or reported value mistakes (more than three orders of magnitude higher than the median levels).

28We extracted the following information: first author name, publication year, study type, country, analytic method, sample time, sample size, gender, mean age, mean BMI, smoking status, season, occupation, pregnancy, diet, vitamin, exercise, outcome (8-OHdG concentration), references, and article DOI. We extracted all subgroup-specific data when data on several subgroups were available in a given paper. Then, we excluded all subgroups selected based on disease status (e.g., cardiovascular disease) and all subgroups selected based on an exposure status (e.g., bus drivers). If data on the same subgroup were reported for different times (e.g., different seasons), only the data at the time of participant inclusion were included. In a third round, we excluded duplicate data (e.g., control population reported in more than one study) and retained the most complete and the most recent study.

29First, we analyzed the values of urinary 8-OHdG measured in original studies in view of establishing the background ranges using meta-analysis. Measured values were generally log-normally distributed. We therefore computed geometric means (GM) and geometric standard deviations (GSD) as the basis for the meta-analysis or equivalently and. Further details on the data treatment and analyses are available elsewhere.

30We could not compute standard errors on the geometric (or arithmetic) scale when neither standard deviation (SD), GSD, IQR, nor confidence interval (CI) were reported. As a consequence, we excluded these studies from the meta-analysis. We converted all the concentration values to the same units (ng/mg creatinine) before computing GM and GSD. We used 113.12 g/mol for the molecular weight for creatinine and 283.24 g/mol for 8-OHdG. We regrouped the data according to analytical techniques used; immunological techniques and chemical techniques. The data were analyzed separately.

31We followed standard practice in meta-analysis and represent the data as forest plots including the I-squared. This is an estimate of the between-study heterogeneity in percentage. If the between-study heterogeneity is much larger than the between-subject heterogeneity, then I2 is large. In this case, any attempt of obtaining a background value for individual participants will not be valid. In our case, a mixed model with study ID as a random effect is a more relevant analysis model. This yields results on the study subgroup level rather than at the individual level. Data management and statistical analyses were performed in STATA version 16 software.

32We report pooled GM values for urinary 8-OHdG in healthy adults, separately for chemical and immunological methods. We observed a significant positive association between smoking status and urinary 8-OHdG concentrations when measured by chemical analysis. No gender effect was shown. Urinary 8-OHdG can potentially be used to quantify excess oxidative stress due to external exposures when background values have been established in different populations. We recommend adjusting urine samples with creatinine, quantifying 8-OHdG with chemical methods, and reporting results as median and quartiles. Comparing values across studies will then be feasible.