Malondialdehyde, an Oxidative Stress Marker in Oral Squamous Cell Carcinoma—A Systematic Review and Meta-Analysis

Objective: To qualitative and quantitatively review published literature assessing the oxidative stress marker malondialdehyde (MDA) in oral squamous cell carcinoma (OSCC). Methodology: Pubmed (MeSH), Science Direct, Scopus, Web of Science, Willey Online Library, Cochrane, and Cross Reference were searched for studies assessing MDA levels in OSCC samples. Results: From the 1008 articles identified, 849 were excluded based on title and abstract screening due to duplication and irrelevance to the topic of interest. Full-text assessment of the remaining 159 articles led to the inclusion of only 46 articles that satisfied the selection criteria. Of these, only 26 studies had data compatible for quantitative analysis. The MDA levels in OSCC groups are significantly increased (p < 0.00001) in plasma, serum, and saliva samples in the majority of the studies evaluated. In contrast, MDA levels in OSCC tissue samples are significantly attenuated (p < 0.00001) compared to healthy controls, supported by fewer studies. Conclusions: The augmented MDA levels in plasma, serum, and saliva samples of the OSCC reflect the heightened oxidative stress level accurately. Further studies are required to understand the attenuated MDA levels in the tissue samples of OSCC. Correlation analysis between MDA levels with established clinicopathological prognostic markers could aid in formulating oxidative stress-based prognostication and treatment planning.


Introduction
Squamous cell carcinoma (SCC) is one of the most common oral malignancies. The incidence of oral cancer varies greatly. The annual worldwide report states the incidence

Focused Question
Is there any significant difference in the MDA level of biological samples between oral squamous cell carcinoma patients and the control group?
Based on the objective of the present meta-analysis and the research question, the following components were focused: (i) Population: patients with OSCC (ii) Exposure or Diagnostic marker: mean and standard deviation value of MDA (iii) Comparison: between patients with oral squamous cell carcinoma and healthy subjects (iv) Outcome: assessment of MDA in various biological samples of patients with OSCC (v) Study: identify related cross-sectional and case-controlled studies investigating the status of MDA in OSCC and control from 1999 to 2020.

Electronic Search Identification
Electronic databases, including PubMed (MeSH), Science Direct, Scopus, Web of Science, Willey Online Library, Cochrane, and Cross Reference, were searched for published articles addressing oxidative stress in oral squamous cell carcinoma using MDA assay between the years 1999-2020. The following keywords, 'oral squamous cell carcinoma,' 'oxidative stress,' and 'Malondialdehyde was employed.'

Screening for Relevance
Articles discussing oxidative stress in OSCCwere identified and shortlisted based on the titles and abstracts screening for relevance and duplication.

Inclusion Criteria
(a) Studies discussed the oxidative status of OSCC using lipid peroxidation marker-Malondialdehyde (MDA); (b) Studies involving various biological samples and expressed the MDA data in mean, standard deviation along with p-value; (c) Papers provided sufficient data to allow comparison of OSCC and control groups.

1.
Articles with the unmatched objective and abstract; 2.
Being literature reviews and systematic reviews; 3.
Studies used other oxidative stress markers as a marker of evaluation; 4.
The works provided inadequate data for the comparison between control and OSCC groups; 5.
Studies related to head and neck squamous cell carcinoma

Retrieval of Full-Text Articles and Evaluation
K.M., U.S., and T.B. screened the titles/abstracts of all the studies and excluded studies at high risk of bias from the evidence synthesis based on pre-specified criteria. K.M., S.P., and A.T.R., have independently screened each included study's full texts. K.M., M.M.A.A, M.A.A, H.S.A.D, Z.K., and A.T.R., have checked and discussed the relevant factors considered in each included study. After assessing all the particulars, the authors have considered the articles for eligibility criteria. The authors resolved disagreements by consensus. Finally, K.M., U.S., and S.P., have performed the data collection procedure.

Data Extraction
The extracted data from full-text articles were author, publication year, age groups, sample size, MDA measurements in OSCC, and control group expressed as the mean and standard deviation along with specific units. Collected data were tabulated separately in a specified format.

Statistical Analysis
The Forest plot was derived using the mean difference, and standard mean difference method to carry out a meta-analysis using comprehensive meta-analysis software version 3 (Biostat Inc. Englewood, NJ, USA). The overall mean difference or standardized mean difference value of MDA in OSCC was analyzed at a 95% confidence interval (CI). A random-effects model was used in the analysis due to the presence of significant heterogeneity. The articles, which expressed the MDA levels in similar units in each sample, only were included in the meta-analysis.

Results
Pubmed search yielded 517 papers; Science direct search yielded 292 papers; Scopus search yielded 141 papers; Web of Science yielded seven papers; Willey online library yielded 26 papers, and Cross-reference search yielded 25 papers. After search refinement, 849 articles were excluded due to unmatched titles and abstracts, including four duplicated data reports and one animal study. After extraction of these articles, 159 articles had their titles relevant to the present work. Full-text was retrieved for the screened articles. Articles with un-matched objectives (n = 84), systematic reviews (n = 1), critical reviews (n = 2), reviews (n = 25) and letter to the editor (n = 1) were excluded. Forty-six articles with matched objectives were included in the systematic review. Only 26 articles had data compatible for a meta-analysis ( Figure 1).

Data Extraction
The extracted data from full-text articles were author, publication year, age groups, sample size, MDA measurements in OSCC, and control group expressed as the mean and standard deviation along with specific units. Collected data were tabulated separately in a specified format.

Statistical Analysis
The Forest plot was derived using the mean difference, and standard mean difference method to carry out a meta-analysis using comprehensive meta-analysis software version 3 (Biostat Inc. Englewood, NJ, USA). The overall mean difference or standardized mean difference value of MDA in OSCC was analyzed at a 95% confidence interval (CI). A random-effects model was used in the analysis due to the presence of significant heterogeneity. The articles, which expressed the MDA levels in similar units in each sample, only were included in the meta-analysis.

Results
Pubmed search yielded 517 papers; Science direct search yielded 292 papers; Scopus search yielded 141 papers; Web of Science yielded seven papers; Willey online library yielded 26 papers, and Cross-reference search yielded 25 papers. After search refinement, 849 articles were excluded due to unmatched titles and abstracts, including four duplicated data reports and one animal study. After extraction of these articles, 159 articles had their titles relevant to the present work. Full-text was retrieved for the screened articles. Articles with un-matched objectives (n = 84), systematic reviews (n = 1), critical reviews (n = 2), reviews (n = 25) and letter to the editor (n = 1) were excluded. Forty-six articles with matched objectives were included in the systematic review. Only 26 articles had data compatible for a meta-analysis ( Figure 1).  Newcastle-Ottawa quality assessment scale was employed to grade the quality of included studies in the systematic review (Table 1). Collected MDA assessment data along with other findings of included articles in various biological samples were tabulated ( Table 2). Few studies compared the MDA level concerning clinical stages of OSCC in various samples (Table 3) and changes in varying histopathological grades ( Table 4). The analysis of MDA levels according to different clinical stages and histopathological grades could not be performed due to the scarcity of the reported studies.               The meta-analysis presented high heterogeneity, reflected by the I 2 values 92.648, 86.785, 97.769, and 64.792 of Figures 2-4, respectively. The different methodologies utilized to measure MDA levels could be the reason for the high heterogeneity.

Discussion
Lipid peroxidation is a sequential reaction providing a constant supply of free radicals that initiate further peroxidation and free radicals accumulation, resulting in OS [77].

Discussion
Lipid peroxidation is a sequential reaction providing a constant supply of free radicals that initiate further peroxidation and free radicals accumulation, resulting in OS [77]. The endogenous formation of MDA during lipid peroxidation serves as a suitable biomarker of endogenous DNA damage [12]. MDA interacts with cellular DNA and forms MDA deoxyguanosine (M1-dG), a DNA-MDA covalently bonded adduct, resulting in DNA damage that causes interference in repair [78]. This mutagenic transformation within the DNA alters their chemical behavior and possibly contributing to carcinogenesis. These reactive aldehydes (MDA) also bind to membrane proteins. They cause profound changes in their function, tonicity, permeability, rigidity, structural integrity, and enhancing neoplastic transformation of the affected tissues. Thus, the developed OS affects the cell membrane's essential constituents, which ultimately increases cell proliferation and actively influences cancer initiation, promotion, and progression [79].
The present systematic review included the research articles that involve 1307 patients diagnosed with OSCC and 1217 healthy volunteers for MDA analysis in various biological samples.
Previous studies demonstrated enhanced lipid peroxidation and malondialdehyde in patients with OSCC. The included studies had found a statistically significant increase in plasma or serum MDA levels in OSCC patients compared with controls (p < 0.001) [8,12,[19][20][21][22]24,30,32,[35][36][37]39,40,43,46,47,49,65]. Similarly, other studies also observed a significant rise compared with the control group (p < 0.05) [8,17,25,31,34,44,52,53,67]. Other studies also reported MDA rise in erythrocytes with statistical significance (p < 0.001) [20,26], (p < 0.01) [38] and (p < 0.05) [5,17,27]. On the contrary, one report did not show any change in blood MDA level in OSCC patients than in control [28]. In the present metaanalysis, the plasma samples showed an overall mean difference of 2.79 with a 95% CI (2.26-3.32). The serum samples showed an overall mean difference of 7.43 with 95% CI (5.99-8.87). The serological changes are consistent even though they are secondary to the tissue changes taking place anywhere in the body. A few studies had also reported higher salivary MDA levels in OSCC compared with healthy subjects with statistical significance (p < 0.001) [12,29,33,35,36,54] and (p < 0.05) [23,41,42,50]. However, three included studies expressed that the increase in the MDA level in saliva and mitochondria was insignificant (p > 0.05) [48,51]. In the present work, the saliva samples showed an overall mean difference of 0.91 with a 95% CI (0. 63-1.18). The increased levels could be due to the disintegration of polyunsaturated fatty acids of bio-membranes due to oxidative lipid damage [19]. The evaluation of tissue MDA level also showed a rise in OSCC patients than the control group with statistical significance (p < 0.001) [36], (p < 0.01) [38], and (p < 0.05) [27]. On the contrary, few authors differently reported the tissue MDA levels of the OSCC group [5,14,16,18,45]. Their studies in tissue displayed a decrease in mean MDA level in OSCC patients compared to the control group with statistical significance. (p < 0.001) [55][56][57][58] and (p < 0.05) [5]. In the present analysis, the tissue samples showed an overall mean difference of −37.08 with 95% CI (−41.25 to −32.92). The decrease in MDA levels observed in the tumor tissues of oral cancer patients reflects a decreased susceptibility of oral tumor tissue to lipid peroxidation. Srivastava 2016 et al. hypothesized that serum biology compared to tissue poses a considerable threat and produces free radicals in excess amounts [45]. They are readily diffused inside the cell to cause various mutations, favoring carcinogenesis. On the other hand, the tissue produces a relatively lesser amount of free radicals and, at the same time, is capable of counteracting them with the available enzymes. Therefore, Srivastava et al. stated that the external environment and the internal factors influence the selective growth of the tumor cells [45].
There is a gradual increase in the MDA level in plasma and erythrocyte when the clinical stage of OSCC advances on further analysis. According to severity, the difference in the rise of plasma MDA levels between the advancing stages was statistically significant within all the clinical grades (p < 0.01) [20] and (p < 0.001) [32]. Arya et al. observed a significant increase in serum MDA value from T1 to T3 group, and the p-value was <0.05 [8]. Therefore, a positive relationship between serum MDA level and tumor size was found. The authors stated that lipid peroxidation increases with the disease severity. Therefore, serological levels are reflecting the extent of tissue injury [24].
In contrast, Babiuch et al. observed decreasing salivary MDA value when the tumor progresses from T1 to T4 in size, statistically insignificant [51]. Two reported studies in tissue displayed a decreasing mean MDA level when the clinical stage of OSCC advances, which is statistically significant in one study (p < 0.01) [18] and insignificant in another report (p > 0.05) [45].
Few studies reported an increase in plasma and serum MDA level when histological grades of the disease advance with statistical significance (p < 0.001) [40] and (p < 0.01) [34]. On the contrary, three studies stated that lipid peroxidation level was inversely proportional to the degree of differentiation of OSCC as the grade advances. However, the change was statistically non-significant (p > 0.05) [8,12,25]. These results correlated with Salzman et al. 2009, who showed a negative correlation of MDA and tumor grade [80]. Thus, there was no definitive correlation pattern in lipid peroxidation between degrees of differentiation of malignant oral lesions. The expression of serum MDA levels in different histopathological grades exhibits a complex relationship. The present meta-analysis showed the MDA levels are significantly increased (p < 0.00001) in OSCC in all the samples of plasma, serum, and saliva except the tissue samples where MDA levels are significantly decreased (p < 0.00001) in OSCC compared to healthy tissues. The tissue-level changes with advancing clinical stages of the tumors were also very poorly explored. The authors used different methodologies to assess MDA levels in various biological samples [55,[57][58][59][60][61][62][63][64]66,[68][69][70][71][72]. The reported studies utilized different clinical staging systems [73,81] and histopathological grading systems [74][75][76] to categorize the OSCC group patients. It will be worthwhile if future studies consider these facts in the MDA assessment of the OSCC group to evaluate the effect of oxidative stress on tumors. Although various treatments have been proposed to manage this type of cancer, its aggressiveness and ability to metastasize make this cancer one of the most difficult to treat, so early diagnosis is crucial when facing this condition [82,83]. Therefore, the studies evaluating the OS will improve the understanding of the anti-oxidant enzyme activity in the early diagnosis and treatment of oral cancer [15].

Conclusions
The oxidant/anti-oxidant equilibrium is a critical step toward developing more effective strategies for prevention, early detection, and treatment of oral cancer. Estimating lipid peroxidation by-products in the OSCC group could assess the degree of oxidative stressrelated tissue injury. Therefore, the assay of malondialdehyde level in oral cancer may be helpful to evaluate the disease severity for both preventive and clinical intervention. Most studies revealed the significant elevation of malondialdehyde levels in oral squamous cell carcinoma patients than healthy controls. Therefore, there is a requirement of large-scale studies with better-matched controls and equal distribution of samples among different clinical stages and histological grades of OSCC to conclude MDA as a potential biomarker for oxidative stress and valid prognostic marker of OSCC.