The Effects of Ascorbic Acid Over Sepsis: Meta-analysis With Trial Sequential Analysis

Background: This meta-analysis is performed to evaluate the effects of AA on the mortality over sepsis patients, focusing on the courses and initiation of treatment as well as AA doses. Methods: Randomized controlled trials concerning sepsis patients treated with intravenous AA were included when searching the database. The meta-analysis was performed using the random (M-H heterogeneity) model to produce summary odds ratio with 95% CI. Trial sequential analysis was applied to evaluated the effect of random errors. Results: The included 12 trials enrolled a total of 1232 patients. Intravenously administration of AA could not lower 28-day mortality over sepsis patients (OR = 0.81; 95% CI (0.54-1.23); p = 0.326). Subgroup analysis demonstrated that when administrating AA alone, in a dose ≥ 10 g/d, or within 6 h of admission, the result may turn to positive (OR = 0.36; 95% CI (0.15-0.86); p = 0.020, OR = 0.50; 95% CI (0.27-0.92); p = 0.025, OR = 0.49; 95% CI (0.27-0.89); p = 0.019, relatively). The quality of evidence is moderate. Conclusion: IV AA may have no effects to lower mortality over sepsis patients. However, when administrating AA alone, in a dose ≥ 10 g/d, or within 6 h of admission, the result may turn to positive. Due to a moderate GRADE certainty of evidence, further studies are required to fully elaborate the effectiveness of AA during the management of the sepsis patients. 95% were performed for the assessment of continuous data. We used random (M-H heterogeneity) model to assess the effects since clinical heterogeneity including setting, regimen, co-interventions A begger test was conducted to assess the publication Subgroup analyses were pre-specied and performed to explore the correlation between mortality and the treatment course, doses, initiation of treatment, population and combination with co-interventions. A p value was considered statistically signicant, except when otherwise specied.

1. Performed on adults in condition of sepsis 2. IV AA vs. placebo or no-intervention groups 3. RCTs.
Exclusion criteria 1. Performed on children 2. AA administered orally, enterally or were permitted to change to an enteral dosage form once enteral access was established 3. Lack of mortality data 4. Lack of control group 5. The way of AA administration was unclear 6. Study design is observational studies including cohort studies and case-control studies.

Data extraction
Data were independently extracted by the rst and the second authors. Extracted data consisted of the name of rst author, year of publication, study population, number of patients, AA dose, AA course, initiation of treatment, co-interventions, clinical parameters and adverse events. We resolved disagreements through discussions until a consensus was reached. We didn't contact the authors for missing data including unreported data or additional details.

Outcome measurements and de nitions
The primary outcome was 28-day mortality. Secondary outcomes included ICU and in-hospital mortality, the length of ICU stay, duration of vasopressor requirement, patients suffering from AKI or demanding RRT and the change of Sequential organ failure assessment (SOFA) score.
Courses of AA treatment 3 d were de ned as short term, ≥ 3 d as long term. Doses < 3 g/d were de ned as low, ≥ 10 g/d as high, and 3-10 g/d as medium.

Assessment of risk of bias
We applied the Cochrane Collaboration tool to assess the risk of bias in RCTs [32,33] , the remaining non-RCTs were assessed via the ROBINS-I tool [34] . We rated each domain of the trials as low risk, unclear, or high risk. Trials were considered low risk when each independent domain was rated as low risk. Any domain rated as unclear or high risk increased the overall risk score. Risk were independently rated by the rst and the second authors. We resolved disagreements through discussions until a consensus was reached.

Statistical analysis
Data were analyzed using Statistics/Data Analysis 15.1. The results of dichotomous data were presented as forest plots through the odds ratios (ORs) with 95% con dence intervals (CIs). Forest plots using Weighted Mean Difference (WMD) with 95% CI were performed for the assessment of continuous data. We used random (M-H heterogeneity) model to assess the effects since clinical heterogeneity including setting, AA regimen, cointerventions was high. A begger test was conducted to assess the publication bias. Subgroup analyses were pre-speci ed and performed to explore the correlation between mortality and the treatment course, doses, initiation of treatment, population and combination with co-interventions. A p value ≤ 0.05 was considered statistically signi cant, except when otherwise speci ed.
Assessment of the certainty of the evidence Grading of Recommendations Assessment, Development and Evaluate system (GRADE) was used to evaluating the quality of evidence, based on the ve GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) [35] . The rst and the second authors independently rated the quality of evidence for primary outcome. The level of evidence was classi ed into four categories as high, moderate, low or very low. We resolved disagreements through discussions until a consensus was reached.
Trial sequential analysis (TSA) TSA was applied to assess the robustness of the results on RCTs, and to provide sample size needed in further trials [36] . We constructed the analysis via TSA version 0.9.5.10 Beta software with a 20% relative risk reduction. The control event rate was calculated from the control group of RCTs in this meta-analysis. The type I error rate was de ned as 5%, and the type II error rate was set as 20% (a statistical power of 80%). The heterogeneity correction was based on model variance.

Study characteristics
The characteristics of each trial are listed in Table 1. There were 8 trials exploring the effects of AA combined with co-intervention, such as thiamine [18][19][20][21][22][23][24] , N-acetylcysteine [28] , and α-tocopherol [28] . The longest intervention course of AA lasted for 10 days [18] , while the shortest was only 1 day [28] ; only 1 records lack of speci c data of the treatment course [29] . The articles included was published in a long span from 1997 to 2020. The sample sized ranged from 20 to 211. The included trials enrolled a total of 1232 patients of which 571 received AA administration, and 661 were controlled subjects.

Risk of bias and quality of evidence
The risk of bias was summarized in Fig. 2. The trial performed by Balakrishnan [24] and Jose [21] were deemed a high risk of bias as incomplete outcome data over mortality. Studies by Fujii [18] , Wani [23] , and Chang [19] were not performed in a double-blinded way and were rated as high risk. In total, 7 trials were de ned as an unclear risk and 5 trials were deemed high risk.

IV AA and mortality
The meta-analysis, recruited 12 RCTs, indicated that IV AA may have no in uence on 28-day mortality of sepsis patients (OR = 0.81; 95% CI (0.54-
Publication bias and sensitivity analysis Sensitivity analysis was performed through the removal of each individual trial and reanalysis the remaining trials. When excluding the trial performed by Wani [23] , the impact of AA on change of SOFA score altered (WMD = 1.19; 95% CI (0.49-1.89); p = 0.001; I 2 = 0.0%).

Overall certainty of evidence
The evidence for 28-day mortality was downgraded to moderate certainty, due to study limitations (including high or unclear risk of bias over studies, clinical heterogeneity), and serious concerns about impression (including TSA results) ( Table 2). Future evidence is likely to change the estimated effect.
Trial sequential analysis TSA were performed based on included RCTs using a random-effects model, and were constructed for a heterogeneity adjusted information size of 2589 patients corresponding to a relative risk reduction of 20% with an α = 0.05 and β = 0.20 (power 80%). TSA indicated lack of solid evidence for a bene cial effect on AA for mortality and further RCTs are needed to testify the effects (Fig. 4).

Discussion
Main ndings IV AA and in-hospital mortality Same as other meta-analysis [38,39] which observed no signi cantly reduction in mortality after IV AA intervention, the major nding of this study suggested that IV AA may have no effects to lower mortality rates over sepsis patients. Notable, when administrating AA alone, in a dose ≥ 10 g/d, or within 6 h of admission, the result may turn to positive.
AA was an essential endogenous trace element and a cofactor of biosynthetic enzymes, which plays an important role in anti-oxidative stress, maintaining the function of endothelia, enhancing the effects of immune system. For patients with COVID-19, early and high IV dose of AA alone or in combination with steroids may have a bene t, especially in a dose about 12 g/d, and at a duration of at least 7 days [40] . Investigators are engaged to nd out the in uences of AA on COVID-19 at a dose of 24 g/d for 7 days [41] .
The degradation of AA that occurs during preparation and storage, and an increased requirement may contribute to the low level of AA in critically ill patients [42] . Previous study has shown that in the post-injury period, only supraphysiologic doses (3000 mg/d) for 3 or more days could approached normal plasma levels [43] . A trial explores the optimal dose of AA demonstrated that 10 g/d dose was associated with supranormal plasma concentrations [44] . While Jackson and colleagues [45] declared that AA is insu cient to compete effectively with nitric oxide (NO) for superoxide at anything less than supraphysiologic concentrations, and a 10-100 mmol/L of AA is an e cient scavenger for preventing the interaction of NO with superoxide. According to the previous studies, we de ned a dose lower than 3 g/d as low, higher than 10 g/d as high, and 3-10 g/d (inclusive of 3 and 10 g/d) as medium dose. Our analysis revealed that high doses may have an ability to reduce mortality in sepsis condition, conversely low and medium doses don't. This is different from our previous results [46] in which after IV high dose of AA, no loss of mortality was observed. In a randomized pharmacokinetic trial, de Grooth and co-works [44] emphasized that high plasma concentrations require stained therapy. In this trial, after 48 h infusion of AA, a varying decline in plasma concentrations was observed even at the dose of 10 g/d. Therefore, they indicated a supplementation of AA more than 48 h, possibly as long as patients remain critically ill. Given the factors mentioned above, we de ned a treatment course less than 3 days as short-term, and the others as long-term. IV AA at a long-term course (≥ 3 d) could not lower mortality in patients with sepsis, which is inconsistent with previous studies. Micah and co-workers [47] stated that APAHCHE-adjusted ICU mortality was lowest when AA was initiated within 6 hours, which is same as our outcomes but still needs further exploration.
As an antioxidant, AA plays vital parts in assisting in the recycling of other antioxidant agents [67] . This seems to be inconsistent with what we have observed in this meta-analysis, of which AA alone rather than co-intervention has a better survival bene t. In view of the absence of relevant RCTs comparing the prognosis of AA monotherapy with combined-therapy, we hope the result can attract the further attentions.

Duration of vasopressor requirement
The result suggested that IV AA may reduce the duration of vasopressor requirement, but the quality of evidence was graded as "very low". Studies have proven that enzymes via which endogenous norepinephrine and vasopression are synthesized required AA as a cofactor for optimal activity [14] . An RCT recruited 53 patients with ST-segment elevation myocardial infarction observed an amelioration of the left ventricular ejection fraction after administration of AA [49] . Furthermore, in the trial performed by Galley and colleagues [28] , AA reduced the systemic vascular resistance index.
Sum together, all mentioned above could contribute to a decreased duration of vasopressor.
The incidence of AKI and RRT AA was reported to promotes the incidence of acute renal failure [50] . A randomized, crossover and controlled design trial observed an increased urinary oxalate and Tiselius Risk Index for calcium oxalate kidney stones [51] . Contrarily, this meta-analysis observed that the incidence of AKI and RRT did not grow after AA treatments (a maximum dose of 66 mg/kg/h), indicating that IV high dose of AA in sepsis patients may be safe. However, reports over the safety of AA in the condition of critical illness are scare in the literature. Cautions should be paid when IV high doses of AA in patients with hemochromatosis, glucose-6-phosphate dehydrogenase de ciency, renal dysfunction, kidney stone, oxaluria and pediatrics [52] .

Comparison with other studies
The major strength of this meta-analysis was to investigate the effects over different AA treatment course and initiation as well as doses contribute to mortality in patients with sepsis. In addition, we focused on the adverse consequences of AA, such as the incidence of AKI and RRT.

Limitation
This meta-analysis had several weaknesses that should be noted. Firstly, 5 trials are under high risks, which may lead to a bias. Secondly, except for 5 multicenter studies [18,[20][21][22]27] , 7 of the included studies were single-center background. Thirdly, the initiation of AA treatment, the treatment courses and doses, and the role of single versus co-intervention therapy still requires clari cation in future studies. Lastly, the missing data handling remained major limitations for this paper.

Conclusion
Based on the current available evidence, IV AA may have no effects to lower mortality over sepsis patients, the quality of evidence was graded as