3.1 Search Results
According to the search strategy (see PRISMA diagram, Fig 1), 2034 articles were found in the accessed databases. Of these, 550 articles were excluded after database screening and removal of duplicates and 1484 studies were read for titles and abstracts. The process of analyzing titles and abstracts resulted in the exclusion of 1473 studies for not meeting the eligibility criteria. The remaining eleven (11) records were read in full and four (4) of them were discarded for the following reasons: two (2) records had data published in poster format, which had data from two other studies included in this review, being excluded, in order to avoid duplication of data analysis. One (1) study published in trial format was excluded, as there were no results available; there was an attempt to contact the authors of this study in order to obtain information on partial data, unpublished or in preprint, however, there was no return of the email sent. After analyzing the risk of bias, a fourth article was excluded, as the methodological design was evaluated as fragile, determining a high risk of bias, according to the evaluation of ROBINS I [17].
3.2 Description of Articles Included in the Systematic Review
The studies included in this review were developed in the following countries: Turkey, Netherlands and Poland. Table 2 shows the summary of the characteristics and main results of the studies. In total, 909 subjects were evaluated: 321 with UC, 272 with CD and 316 healthy controls participated in the studies. Only two studies [7,1] exclusively evaluated subjects in remission (one study evaluated 20 subjects with UC in remission and the other evaluated 51 subjects with CD in remission, respectively). The studies included in this review had data collected between 2006 and 2018 and were published between 2008 and 2019. There was no standardization regarding the type of thiol measured as an expression of the oxidative state, nor regarding the method used for its dosage. Of the included studies, four (1, 2, 4, 8) used the method described by Elman´s et al. [21] for thiol measurement, but only three [1, 2, 4] corrected for plasma albumin, as described by Hu et al. [22] and Turel et al. [23]. The method proposed by Erel & Neselioglu [24] was used for thiol dosage in one study [6]. The other two studies [5, 7] included in this review did not explain the reagent used to measure thiols. All studies described the use of spectrophotometric analysis for thiol dosage.
This systematic review included seven (7) analytical observational studies, case-control type [1 - 2, 4 - 8] that measured thiol levels in individuals with CD/UC with the aim of evaluating the potential association with oxidative stress and DII. Two studies [4, 7] evaluated only subjects with UC and one study [1] evaluated only subjects with CD. Only one study [1] evaluated only individuals in remission and two studies [5, 7] evaluated individuals with active and remitting IBD, comparing the results between individuals in different phases of the inflammatory disease. in all seven studies [1 - 2, 4 - 8] included changes in thiol levels in individuals with CD/UC and the authors inferred that there was oxidative stress associated with IBD.
3.3 Type of Measured Thiol
Of the seven studies included in this systematic review, four [4 - 7] used the levels of native and/or total serum thiols as a marker of thiol oxidation; two studies [1 -2] used plasma free thiol levels corrected for albumin levels to assess systemic oxidative stress, and only one study [8] used plasma free thiol levels, but did not report whether there was a correction. by plasma albumin levels, despite the attempt to obtain this methodological detail, by contacting the research group's electronic address. Thus, there was a limitation of the grouped analyzes of the studies, considering that the lack of standardization of the type of thiol to be used prevents the comparison of results.
3.4 Thiol Expression and Relationship with Oxidative Stress in Inflammatory Bowel Disease
Baskol et al. [4] detected increased total thiol levels in subjects with UC when compared to controls. Akinci et al. [5], when conducting the study with the largest number of individuals with IBD, included in this review, of both phenotypes, at different stages of the disease, found total thiol levels positively associated only with active DC. Yuksel et al. [6] identified a negative association between the reduction of total and native thiols in active CD and between native thiols in active RCU, supported by a negative correlation between native thiol and EAI, CDAI, erythrocyte sedimentation rate (TSE) and C-reactive protein (CRP) in individuals with active CD and UC, when compared with healthy controls. Neselioglu et al. [7] found: I. negative association between: a: the levels of native and total thiols in individuals with active UC; b: native thiol and CRP; c: native thiol and TSE; d: thiol homeostasis and RCU activity/severity; II. positive correlation between native thiol and albumin; III. lower total thiol levels in subjects with UC when compared to healthy controls and IV. higher total and native thiol levels in subjects with UC in remission compared to subjects with active UC or healthy controls. From these results, these researchers related disease activity to thiol oxidation, and suggested the use of thiol as a serum marker to assess activity and predict the severity of the disease course.
In 2019, the study conducted by Bourgonje et al. [1] was the precursor, among the studies included in this review, to measure the plasma concentrations of free thiols adjusting them to albumin, considering the characteristic of circulation of thiols in the human body [1, 9, 23, 25]. This group detected decreased levels of thiols in individuals with CD in remission when compared to healthy controls, a negative correlation between plasma thiols and inflammation biomarkers, including CRP and IL-17A, enabling the correlation of subclinical CD activity to systemic oxidative stress. Another study, conducted by Bourgonje et al. [2] detected a strong correlation between plasma thiols and the degree of inflammatory disease activity evaluated endoscopically and a negative correlation between plasma thiol levels and fecal calprotectin (FC) levels in individuals with IBD, of both phenotypes, this time in activity, comparing them with healthy controls, allowing the discrimination, with high precision and in a significant way, of the degree of activity (mild, moderate or severe) of the disease, better than the FC. The Polish study conducted by Neubauer et al. [8] detected lower amounts of thiol in people with CD and UC, regardless of disease activity, when compared to healthy controls, but the thiol concentrations of individuals with CD and UC with active disease were inversely correlated with, respectively, CDAI and Rachmilewitz Index (RI), although this inverse relationship did not reach statistical significance. Based on these results, these authors suggested the use of plasma thiols as a therapeutic target to monitor IBD activity, as it is a minimally invasive strategy, presents an inverse correlation with the severity/severity of intestinal inflammation, and therapeutic modulation, through the administration of of antioxidants, considering that higher levels of plasma thiols would be associated with lower levels of inflammatory biomarkers and favorable systemic status and evolution in IBD.
It is important to consider that individuals with IBD treated with corticosteroids had lower thiol levels in the study conducted by Neubauer et al. [8]. Free thiol concentrations were decreased, mainly in the active DC, and were inversely related to inflammatory markers and oxidative stress, demonstrating depleted total antioxidant capacity, instrumentalizing these authors to conclude that the assessment of the total systemic antioxidant status can be useful in the evaluation not invasive of mucosal healing in individuals with IBD, and, additionally, that the assessment of serum thiol levels can provide relevant information about the adverse effects of corticosteroid therapy.
From the analysis of the studies included in this review, it was possible to identify relevant results related to systemic oxidative stress, measured by serum thiol levels, and IBD activity, and negative association with inflammatory markers. The findings of a strong correlation between the degree of endoscopic disease activity and a negative correlation between FC and serum thiols strengthen the justification for investigating the potential of thiols as a marker of oxidative stress in IBD.
3.5 Meta-analysis Results
Figures 1, 2, 3 and 4 present the results of the meta-analyses to identify the association between active DC and active UC and different thiol dosages. No statistically significant differences were identified between active DC and total thiol and native thiol. Also, no association was identified between active UC and total thiol and native thiol.
3.6 Methodological Quality and Risk of Bias
The assessment of the risk of bias of the selected studies is presented in Appendix 3. Eight studies were included in the assessment of the risk of bias.
At this stage, the study conducted by Von Martels et al. [26] was excluded from this systematic review, as it presents a domain (related to the outcome measure) with a high risk of bias, from the evaluation by ROBINS I [17], for intervention studies, as proposed by the Cochrane Handbook [19], which compromised the global assessment of the methodological quality of this study and led to its exclusion.
All studies that followed this systematic review were observational studies, of the case-control type, and, for this reason, had the risk of bias and methodological quality evaluated by the instrument The Newcastle Ottawa Scale of Case-Control Studies [28], in agreement with the recommendation of the Cochrane Handbook [27]. In According to the methodological quality assessment proposed by this scale, it was observed that most studies presented more than 77.7% of adequacy in terms of quality, with percentages that varied between 66.6% and 88.8%. The issue that most contributed to the reduction in the assessment of methodological quality and the increase in the risk of bias in these studies was the criteria for selecting controls, which were not described in some studies. None of the included studies assessed the outcome non-response rate, which determined the loss of points from all studies included in this review for this domain.
Based on the evaluation criteria of The Newcastle Ottawa Scale of Case-Control Studies [28], of the seven studies included in the review, two of them [1 -2] achieved the best evaluation and scored 8/9 stars, losing points for not presenting the outcome non-response rate assessment. Three studies [4, 7 -8] scored 7/9 stars; the studies by Baskol et al. [4] and Neubauer et al. [8] lost points due to the lack of characterization of the selection of controls and for not presenting the non-response rate in the studies. As for the study by Neselioglu et al. [7], the evaluation was reduced in the domain of comparability of cases and controls and by the absence of evaluation of the non-response rate of the outcome.
The two studies [5-6] rated as the lowest methodological quality included in this review scored 6/9 stars, with 66.6% adequacy. The domains that determined the downgrading of the evaluation of these two studies were the selection and definition of controls [5] and the comparability of cases and controls [6]; these two studies also lost points for not evaluating the outcome non-response rate.
3.7 Certainty of the Evidence
The assessment of the certainty of evidence proposed by GRADE [18] is presented in Appendix 4 and shows that the evaluated outcomes (serum levels of native and total thiols in CD and UC) determine low confidence in the evidence. Studies conducted by Baskol et al. [4], Ankinci et al. [5], Yuksel et al. [6] and Neselioglu et al. [7] were combined in the meta-analysis and revealed high inconsistency and imprecision, associated with potential confounding factors, losing points in the assessment of confidence in the evidence, characterizing themselves as insufficient for safe decision making. Studies conducted by Bourgonje et al. [1], Bourgonje et al. [2] and Neubauer et al. [8] could not be included in the meta-analysis, considering the impossibility of grouping in this systematic review. From the GRADE assessment, confidence in the effect estimate is limited, which suggests that the true effect may be substantially different from the effect estimate and that, despite the meta-analysis revealing that there is no association between active CD and UC and total thiol and native thiol and the correlation trend between serum thiols and IBD activity presented by the studies included in this review, especially those with a more robust methodological design (which had a lower risk of bias, but which could not be included in the meta-analyses, considering the absence of combinable parameters between these studies), there is no reliability to confirm or deny the potential association between thiol levels and IBD.
Some factors potentially influenced these results as the methodological design, the size of the samples, the inconsistency, the heterogeneity, the wide confidence intervals and the imprecision of the studies, in addition to the small number of studies included in this systematic review, have contributed to inconclusive results and a tendency for non-correlation, in studies with lower methodological quality and/or those that could not be included in the meta-analysis, determining a reduction in the strength of the evidence, and too prevented the carrying out quality assessment's additionals.