Skip to main content

Table 6 Summary of selected literature relevant to oxidative status in all-cause mortality prediction

From: Systematic review regarding metabolic profiling for improved pathophysiological understanding of disease and outcome prediction in respiratory infections

First author, year, reference

Marker

Study type

Study population

Key findings

Limitations

Rodas et al., 2012, [118]

- Total glutathione

Single-center, prospective observational cohort study (6-month follow-up)

174 patients admitted to the ICU (without thoracic, neurosurgery and trauma cases) (>17 years of age)

- After division of patients into quartiles according to total concentration of glutathione and GSH in whole blood, no relationship to mortality was observed (data not shown); division of patients into quartiles according to GSH/total glutathione ratio showed a significant higher 6-month-all-cause mortality for the quartile with the highest ratio compared to the lower three quartiles (p = 0.026); the AUC of GSH/total glutathione ratio regarding all-cause mortality prediction was 0.56 (p = 0.18) with an optimal cut-off for the GSH/total glutathione ratio of 0.66; a stepwise multiple logistic regression analysis regarding 6-month-mortality prediction showed an OR of 2.35 (CI 1.02-5.41; p < 0.001) for the GSH/total glutathione ratio

- The study only represented the group actually studied, namely patients >17 years of age admitted to the ICU

- GSH

- Glutathione redox status (GSH/total glutathione ratio)

- Observational studies allow no statement about causal association

- Redox status of glutathione in plasma might not be representative for the intracellular status (Nevertheless, the plasma ratio of GSH/total glutathione could be useful as a biomarker to predict post-ICU mortality)

Herzenberg et al., 1997, [123]

- Total glutathione

Single-center, prospective observational cohort-study (2-3-year follow-up)

204 HIV-infected patients, who were screened for enrollment into a clinical trial designed to determine whether orally NAC replenishes GSH in subjects with low GSH levels; however, subjects who were finally enrolled in the NAC trial were not included in the present observational cohort study

- Logistic regression analyses demonstrated an association between increased baseline GSB and improved 2-3-year survival, whereas baseline GSB levels below normal values showed a higher mortality (p < 0.0001)

- Selection bias due to exclusion of patients who finally were trial subjects of the NAC interventional trial

- GSH

- GSB

- No data about plasma glutathione

- Kaplan-Meier survival curves demonstrated that patients with baseline GSB levels below 1.05 show a significantly increased 2-3-year all-cause mortality compared to subjects with higher baseline GSB levels (p = 0.005); the above named optimal treshold of 1.05 for separating survivors from non-survivors was determined using ROC analyses for 2-3-year survival

- In proportional hazard analyses with inclusion of GSB levels as well as CD4 T cell counts in the model, GSB has still been demonstrated to be a significant, independent mortality predictor (RR 1.6 [95 % CI 1.1–2.5], p = 0.009), despite its significant correlation with CD4 T cell counts (p < 0.0001)

Xiu et al., 2012, [124]

Homocysteine

Prospective observational cohort-study (up to 10-year follow-up)

1,412 elderly >65 years of age, selected from the Elderly Nutrition and Health Survey in Taiwan (NAHSIT) [142]

- In a model adjusted for sociodemographic, behavioral, and nutritional variables, plasma homocysteine levels >14.5 umol/l were associated with significantly higher mortality compared to those <9.3 umol/l (HR 1.8, 95 % CI 1.20–2.71; p = 0.002)

- Although adjustment for several known associations with homocysteine status has been undertaken, residual confounding is still possible

Wong et al. 2013, [125]

Homocysteine

Prospective observational population-based cohort study (follow-up of 5.1 +/− 1.3 years)

4,248 community-dwelling Australian men aged 70–88 years, selected from the Health in Men Study (HIMS) [143]

- After adjustment for frailty, age, education, living circumstances, smoking, cardiovascular disease, cardiovascular risk factors and renal function, high total plasma homocysteine levels (≥15 umol/l) were a significant all-cause mortality predictor (HR 1.25, 95 % CI 1.06–1.48, p < 0.05)

- Self-selection of study participants might have biased the findings toward lower homocysteine, lower age and fewer comorbidities compared to the non-respondents, which limits generalizability of study results and might have moved the results toward the null hypothesis with underestimation of associations

Swart et al., 2011, [126]

Homocysteine

Sub-study of a prospective observational population-based cohort study (11-year follow-up)

1,117 independently living elderly (mean age = 75.1 years), selected from the Longitudinal Aging Study Amsterdam (LASA)

- After adjustment for several confounders, women in the third and the fourth quartile of plasma homocysteine levels were associated with a significantly higher all-cause mortality risk compared to those in the first quartile (Q3 HR 1.70, 95 % CI 1.08–2.65; Q4 HR 1.91, 95 % CI 1.22–3.00), independently of vitamin B12 status; in men there was no significant difference in all-cause mortality rates between different quartiles of homocysteine levels in adjusted models

- Possible underestimation of the actual relationship between homocysteine levels and mortality due to selection bias, since excluded subjects were older, more likely to be cognitively impared, less physically active and had suffered CVD more often at baseline

Drewes et al., 2014, [127]

Homocysteine

Post-hoc subanalysis of the double-blind, randomized placebo-controlled study, stratified by plasma homocysteine levels (mean follow-up of 3.2 years)

3,522 subjects with history of or risk factors for CVD (aged 70–82), selected from the primary care setting in two of the three PROSPER study sites (Netherlands and Scotland)

- In the placebo group subjects with plasma homocysteine levels in the highest tertile showed a significantly higher all-cause mortality risk (HR 1.7, 95 % CI 1.2–2.5, p = 0.003) as well as a greater risk of fatal and nonfatal CHD (HR 1.8, 95 % CI 1.2–2.5, p = 0.001) compared to those with low plasma homocysteine levels

- Study not originally designed to collect blood samples for plasma homocysteine level assessment, which is why data could only be used from two of the three PROSPER study sites

- Some blood samples have possibly been stored at room temperature for up to 8 h, which could have contributed to artificially high plasma homocysteine levels, and thus misclassification

- Regarding all-cause mortality the pravastatin treatment group showed an absolute risk reduction of 4.6 % (95 % CI 0.78–8.4 %) in the subgroup with high homocysteine levels compared to -0.66 % (95 % CI -4.0–2.7) in the subgroup with low homocysteine levels (absolute risk reduction difference of 5.2 %, 95 % CI 0.19–10.3; p = 0.04)

- No data about the intakte of vitamin B, which may lower homocysteine levels

Waskiewicz et al., 2012, [128]

Homocysteine

Prospective observational, population-based cohort study (mean follow-up of 5.4 years)

7,165 Polish people aged 20–74 years

- In multivariable proportional hazards models adjusted for sex, age and cardiovascular risk factors, RR of all-cause long-term mortality was significantly higher in plasma homocysteine levels >10.51 umol/l compared to those <8.20 umol/l (HR 1.766, 95 % CI 1.197–2.605)

- The immunoenzymatic method used to determine homocysteine levels in this study has shown the lowest precision of measurements in comparative analyses of various homocysteine assay methods

- In Kaplan-Meier survival curves patients with homocysteine concentrations in the highest tercile (>10.51 umol/l) showed significantly increased long-term all-cause mortality compared to subjects with homocysteine levels in the lowest tercile (p = 0.0003)

- Impossible to perform analyses of mortality related to stroke and ischemic heart disease due to low number of deaths in these groups

Naess et al., 2013, [129]

Homocysteine

Prospective, observational population-based cohort study (mean follow-up of 12.4 years)

198 patients with first ischemic stroke living in Hordaland County (mean age of 47.8 years)

- After adjustment for age, sex and CRP levels, Cox regression analysis – excluding patients with stroke caused by dissection – showed high plasma homocysteine levels (>9 ug/l) to be significantly and independently associated with all-cause mortality (HR 1.04, p = 0.02)

- Small sample size

- Possible selection bias due to retrospective patient recruitment

- High plasma homocysteine levels have been shown to be significant all-cause mortality predictors only after exclusion of patients with dissection, suggesting homocysteine to be a potential confounder in mortality prediction

Vieira et al., 2010, [130]

Homocysteine

Single-center, prospective observational cohort-study (2-year follow-up)

95 predialysis patients with chronic kidney disease (mean age of 69.4 years)

- Kaplan-Meier survival curves demonstrated a significant lower survival in patients with total plasma homocysteine levels and nutritional status assessment (by mSGA) above the mean level compared to lower levels (p = 0.04; survival at 24 months = 50 %)

- Small sample size

Tekin et al., 2012, [144]

Homocysteine

Single-center, prospective observational cohort-study (1-year follow-up)

70 patients with HF (left ventricle ejection fractions < 35 %) (mean age 60 +/− 12)

- Serum homocysteine levels were significantly higher in non-survivors compared with survivors (20.8 +/− 5.8 vs. 16.9 +/− 5.1 umol/l, p = 0.029)

- Small sample size

- With an optimal cut-off value of >17.45 umol/l, the AUC of serum homocysteine levels with regard to mid-term mortality prediction was 0.855 (95 % CI 0.792–0.965, p < 0.001)

  1. AUC area under the receiver operating characteristic curve, CI confidence interval, CHD coronary heart disease, CKD chronic kidney disease, CRP C-reactive protein, CVD cardiovascular disease, GSB glutathione-S-bimane, GSH glutathione, HF heart failure, HIMS Health in Men Study, HIV human immunodeficiency virus, HR hazard ratio, ICU intensive care unit, LASA Longitudinal Aging Study Amsterdam, mSGA modified Subjective Global Nutritional Assessment, NAC N-acetylcysteine, NAHSIT Nutrition and Health Survey in Taiwan, OR odds ratio, p p-value are statistically significant at p < 0.05, PROSPER PROspective Study of Pravastatin in the Elderly at Risk, Q quartile, ROC receiver operationg characteristic, RR relative risk