Nutrients. 2019 (Mar 27); 11 (4). pii: E708 ~ FULL TEXT
Harri Hemilä & Elizabeth Chalker
Department of Public Health,
University of Helsinki,
POB 41, FI-00014 Helsinki, Finland.
A number of controlled trials have previously found that in some contexts, vitamin C can have beneficial effects on blood pressure, infections, bronchoconstriction, atrial fibrillation, and acute kidney injury. However, the practical significance of these effects is not clear. The purpose of this meta-analysis was to evaluate whether vitamin C has an effect on the practical outcomes: length of stay in the intensive care unit (ICU) and duration of mechanical ventilation. We identified 18 relevant controlled trials with a total of 2004 patients, 13 of which investigated patients undergoing elective cardiac surgery. We carried out the meta-analysis using the inverse variance, fixed effect options, using the ratio of means scale. In 12 trials with 1766 patients, vitamin C reduced the length of ICU stay on average by 7.8% (95% CI: 4.2% to 11.2%; p = 0.00003). In six trials, orally administered vitamin C in doses of 1–3 g/day (weighted mean 2.0 g/day) reduced the length of ICU stay by 8.6% (p = 0.003). In three trials in which patients needed mechanical ventilation for over 24 hours, vitamin C shortened the duration of mechanical ventilation by 18.2% (95% CI 7.7% to 27%; p = 0.001). Given the insignificant cost of vitamin C, even an 8% reduction in ICU stay is worth exploring. The effects of vitamin C on ICU patients should be investigated in more detail.
KEYWORDS: antioxidants; artificial respiration; burns; cardiac surgical procedures; cardiovascular system; critical care; dietary supplements; oxidative stress; sepsis; systematic review
From the FULL TEXT Article:
For many centuries, scurvy was a severe disease most commonly afflicting sailors on long voyages [1–16]. In 1753, James Lind, a Scottish surgeon, reported that scurvy could be successfully treated with citrus fruit. However, it was not until the 1930s that Albert Szent-Györgyi discovered that vitamin C was the specific cure for scurvy . Since then, it seems that many people have also believed the converse to be true; that is, that scurvy is the only disease for which vitamin C is effective. This history has led to a widespread belief that vitamin C is only useful for the prevention and treatment of scurvy. However, there is strong evidence that vitamin C is also effective for many conditions other than scurvy.
In controlled trials, vitamin C has improved endothelial function [18–20], lowered blood pressure , increased left ventricular ejection fraction [22–26], decreased the incidence of atrial fibrillation [27–29], protected against contrast-induced acute kidney injury [30,31], decreased glucose levels in patients with type 2 diabetes , decreased bronchoconstriction [33–35], shortened the duration of colds [36–42], decreased the incidence of colds in physically stressed people [41–43], and it has prevented pain [44–47]. There is also evidence that vitamin C may have a beneficial effect on pneumonia [2, 42, 48, 49]. These findings demonstrate that the physiological effects of vitamin C are not limited to preventing and treating scurvy, although conclusions about the practical significance of these effects require further exploration.
In experimental vitamin C deficiency studies, scurvy is induced when dietary vitamin C intake and plasma vitamin C levels are below approximately 10 mg/day and 11 µmol/L, respectively [50–62]. In the USA, the average vitamin C intake is about 100 mg/day  and this leads to plasma levels of 50–60 µmol/L . These are important reference levels when considering vitamin C doses and plasma levels in intensive care unit (ICU) patients.
While scurvy is relatively rare nowadays, it has not disappeared altogether. A number of case reports have described patients with overt scurvy in diverse hospital settings, including ICUs [13, 64–81]. Furthermore, even though clinical scurvy is rare, very low vitamin C plasma levels, such as those found in scurvy patients, are not uncommon in hospitals. Several studies have been undertaken looking at this issue [82–89].
A survey of elderly Scottish patients hospitalized because of acute respiratory infections reported that 35% of patients had vitamin C plasma levels less than 11 µmol/L . One study in a hospital in Paris reported that 44% of patients had vitamin C plasma levels less than 6 µmol/L , and in another hospital in Paris, 17% of patients had levels less than 11 µmol/L . Another survey, of patients with advanced cancer in a hospice, found that 30% had vitamin C plasma levels less than 11 µmol/L . In a Canadian university hospital, it was found that 19% of patients had vitamin C plasma levels less than 11 µmol/L . In a study of surgical patients in Australia, it was found that 21% had vitamin C plasma levels less than 11 µmol/L . Finally, in another study, 3% of cancer patients had clinical scurvy and vitamin C plasma levels less than 6 µmol/L , and in a survey of 145 hospitalized elderly French patients, 18 had clinical symptoms of scurvy . Although these estimates of prevalence for vitamin C deficiency should not be extrapolated directly to other hospital settings, they demonstrate that low vitamin C levels are common in some contexts.
Given this, it would seem reasonable to screen plasma vitamin C levels in hospital patients when appropriate and administer vitamin C to patients with particularly low levels, irrespective of whether vitamin C is believed to be effective for conditions other than scurvy.
Furthermore, there is evidence that vitamin C metabolism is changed in many conditions that involve physiological stress, such as infections, surgery, traumas, and burns, in which case vitamin C levels can decline dramatically [90–101]. Although 0.1 g/day of vitamin C can maintain a normal plasma level in a healthy person , much higher doses (1–4 g/day) are needed for critically ill patients to increase plasma vitamin C levels to the normal range [102–105]. Therefore, high vitamin C doses may be needed to compensate for the increased metabolism to reach normal plasma levels.
Given that vitamin C has diverse effects on medical conditions [18–49], it is possible that, in some contexts, these effects may also lead to practical benefits. A previous meta-analysis of 11 trials on patients undergoing elective cardiac surgery found that vitamin C shortened hospital stay on average by 10% . The purpose of this current meta-analysis was to analyze controlled trials which have reported on the effect of vitamin C on the length of ICU stay or on the duration of mechanical ventilation.
The Complex Biochemistry of Vitamin C
Poor healing of wounds and swelling of the gums are the most well-known symptoms of scurvy. At the biochemical level, they are explained by the role of vitamin C in collagen metabolism. However, vitamin C has a number of biochemical roles unrelated to collagen metabolism. For example, it participates in the synthesis of norepinephrine, which is a neurotransmitter, carnitine, which is involved in energy metabolism, and nitric oxide (NO) [133–139]. Vitamin C participates in the terminal amidation of several neuropeptides [140,141]. There are over 100 known neuropeptides and over half of them require amidation to achieve full biological activity, such as vasopressin, oxytocin, gastrin, calcitonin, neuropeptide Y, and substance P.
Vitamin C also participates in the demethylation of DNA and histones and thereby influences the epigenome [142–147]. It has been estimated that vitamin C may demethylate over 1000 genes in embryonic stem cells . Vitamin C hydroxylates specific proline residues in hypoxia-inducible factor-I, which is a transcription factor that is a key to oxygen sensing, and may regulate several hundred genes [143–149]. In addition, vitamin C is a major water-soluble antioxidant, and thereby, can have a wide range of nonspecific effects. Such biochemical effects translate to vitamin C participating in higher-level physiological processes, for example, in the immune system [142, 147, 150, 151], endothelial functions [18–20, 143, 152, 153], the nervous system including pain sensation [44,45,46,47,136,139,154,155], and in the functioning of the heart [22–29, 156–159]. These are unrelated to collagen metabolism.
The numerous biochemical effects of vitamin C can explain the large variety of symptoms that have been reported in experimental vitamin C deficiency studies, such as depression, anxiety, irritability, confusion, shortness of breath, fatigue, oliguria with edema especially of the lower extremities, anemia, bruising, hemorrhages, joint pains, joint effusions, peripheral neuropathy, and common colds that last longer than usual [50–62].
According to case reports of scurvy, in addition to fatigue, anemia, bruising, hematomas, and different types of pain, vitamin C deficiency has been associated with various cardiovascular symptoms [67–76], dyspnea [64, 70–81], lower limb edema [13, 64, 66, 70, 72, 80, 160–175], infection-type symptoms [66–68, 73, 160, 161, 165, 167, 174–177], bleeding in the gastrointestinal and other regions [76, 164, 172, 175, 177–181], and neuropsychiatric symptoms [70, 71, 182–185].
Logically, low systemic vitamin C levels should be considered in the differential diagnosis of these symptoms observed in experimental vitamin C deficiency studies and in case reports of scurvy. Scurvy is a clinical diagnosis and not a diagnosis based on plasma vitamin C level. For example, Hodges concluded that “the subjects … had plasma values well above the deficiency level throughout the time they had frank scurvy. Furthermore, a comparison between plasma levels of ascorbate and pool sizes showed a very poor correlation… In general, it is fair to say that scurvy appeared when the body pool fell below 300 mg” (, p. 441). The most undisputable evidence of scurvy symptoms is that they disappear with vitamin C administration [13, 64–81, 160–185].
Early Evidence of Vitamin C Deficiency in Hospital Patients
In the 1930s, after vitamin C had been identified, there were reports stating that plasma vitamin C levels of hospital patients were much lower than those of healthy people [186–188]. Finkle (1937) stated that “a large percentage of the population encountered in the wards and outpatient department of the Hospital were found to have excretion levels of vitamin C in the urine considerably below that found in normal subjects living on an adequate mixed diet. When such individuals receive 100 mg of vitamin C intravenously, there is practically no rise in urinary excretion of vitamin C”  (p. 593). Decrease in vitamin C levels after surgery was reported by Lund (1939) who wrote that “plasma cevitamic acid determinations were made before and after operation on 43 patients coming to major operation. In nearly every case there was a prompt fall after operation from the original level”  (p. 127). Abbasy et al. (1937) wrote that “in common with other infectious conditions, osteomyelitis causes a diminished rate of excretion of vitamin C in the urine and a lowered response to test dose, indicative of an apparently increased usage of the vitamin during the infective process”  (p. 180). In addition, Evans (1938) reported that vitamin C increased the urinary output of patients with heart failure [189, 190].
At that time, the concept of oxidative stress was unknown, as were many biochemical and physiological functions of vitamin C that were identified much later. Thus, the biochemical concepts required to consider the potential physiological harms of very low vitamin C levels were not developed. Randomized trial methodology had also not evolved, to enable an unbiased scientific investigation of whether vitamin C administration actually benefits hospital patients with very low vitamin C levels. A low vitamin C level does not necessarily mean that administration of vitamin C would be beneficial. A particular vitamin C level may simply be a marker for the various pathological processes rather than being implicated in the actual processes themselves. If this was the case, there would be no benefit from administration of vitamin C. Nevertheless, low levels of vitamin C in hospital patients still occur [82–89], and there is modern evidence that vitamin C metabolism is changed under various forms of severe physiological stress such as surgery, sepsis, and trauma [90–105]. Hence, the question of the clinical importance of low vitamin C levels in hospital patients is as relevant today as it was in the 1930s.
Vitamin C and Pneumonia and the Common Cold
Before vitamin C was identified, scurvy was associated with an elevated risk of pneumonia [2, 48], which suggests that vitamin C might have clinically relevant effects on infections. After the isolation of vitamin C, a few German and US physicians proposed that it might be beneficial in the treatment of pneumonia [191–197]. However, to our knowledge, only two small controlled trials published decades ago have investigated the effect of vitamin C for treating pneumonia.
A Russian study reported a linear dose response over two vitamin C doses on the length of hospital stay for pneumonia patients [49, 198]. A Scottish study reported significant benefit on the respiratory symptoms of pneumonia patients, with only one death in the vitamin C group compared with five in the placebo group [49, 82]. A further study found that vitamin C reduced the incidence of pneumonia by 80% in influenza A patients [49, 199]. Although these studies were small and had methodological deficiencies, they provided significant justification for the need for further research into the effects of vitamin C on pneumonia . While this was over a decade ago, no new controlled trials on vitamin C and pneumonia have been published since.
The leap from the common cold to critical care is vast; however, the ubiquity and mild character of the common cold has led to a large number of studies of vitamin C for prevention and treatment being undertaken. These studies include a total of more than 10,000 participants and such a large data set allows for informative subgroup analyses and other comparisons [16, 41, 42]. There are some lessons from the common cold studies that may be informative in the ICU context.
Firstly, the common cold leads to a decline in vitamin C levels in white blood cells, and while 0.2 g/day of vitamin C did not return the levels to normal, administration of 6 g/day did . Thus, it seems that even mild infections can increase vitamin C metabolism such that systemic vitamin C levels are reduced.
Secondly, there is evidence that the effect of vitamin C supplementation on common cold incidence is context dependent. Context-dependency means that we should not expect a uniform effect in all populations. Although vitamin C has no effect on the incidence of colds in the normal population, it halved the incidence of colds in participants undergoing heavy physical activity [41–43, 202], and it decreased the incidence of colds in males who had particularly low dietary vitamin C intake [16, 202, 203]. Vitamin C administration has shortened the duration of colds which indicates that taking more vitamin C is beneficial when there is an infection ongoing 36–42, 204]. Heavy physical activity and infections such as the common cold lead to oxidative stress, which may explain the benefits of vitamin C administration in these contexts. Furthermore, vitamin C supplementation may be beneficial for those who have particularly low intake levels.
Thirdly, two common cold studies compared two different vitamin C doses, and both found linear dose-dependency [40, 42, 204]: 6 g/day was twice as effective as 3 g/day in shortening the common cold , and administration of 8 g on a single day was twice as effective as 4 g . This indicates that even for a mild infection such as the common cold, doses of 3–4 g per day of vitamin C may not be sufficient for maximum benefit. The common cold induces only low-level physiological stress, and so it is possible that more severe medical conditions may benefit from higher doses than 6–8 g/day of vitamin C. As noted above, dose-dependency was also reported in an old pneumonia study [49, 198].
Fourthly, some common cold studies have indicated that the effects of vitamin C may be greater for outcomes that reflect greater severity . In two large studies, vitamin C did not shorten the duration of runny nose, but significantly decreased the days off work in adults  and the days off school in children .
Finally, a few common cold studies have indicated that in some contexts vitamin C administration may lead to greater effects in males than in females [202, 209–212].
Vitamin C and the Cardiovascular System
A meta-analysis found that vitamin C reduced post-operative atrial fibrillation (POAF) in trials conducted outside of the USA, but not in studies carried out in the USA, indicating context dependency of the effect . For several treatments, substantial divergence in the size of effect has been found between more developed and less developed countries , and that may be the case also for vitamin C and POAF.
In patients undergoing cardiac surgery, vitamin C has increased cardiac perfusion after the operation  and decreased the level of markers of myocardial injury, such as CK-MB (creatine kinase MB isoenzyme) [22, 106, 214]. A meta-analysis of patients with atherosclerosis or heart failure found that vitamin C improved endothelial function . In a few trials, vitamin C increased the left ventricular ejection fraction [22–26]. A meta-analysis showed that vitamin C supplementation can lower blood pressure . Vitamin C may also protect against myocardial ischemia/reperfusion injury .
In one of the experimental scurvy studies, two participants developed acute cardiac complications, and a third developed shortness of breath and pain in the chest. Therefore, the trial was terminated for each of them, and all three showed striking improvements within a few days of receiving vitamin C . Several case reports have indicated that vitamin C deficiency can cause cardiovascular effects such as hypotension, hypertension, right heart failure, and ECG changes [67–76].
Other Possible Effects of Vitamin C Relevant in the ICU Context
Pneumonia is the most common cause for severe sepsis, accounting for about half of all cases . As described above, vitamin C may be effective in the prevention and treatment of pneumonia in some contexts.
Vitamin C has been beneficial in animal models of sepsis, for example, by decreasing mortality [152, 153, 217]. In one study, intravenous infusion of vitamin C increased the free radical form of vitamin C significantly more in sepsis patients than in healthy subjects, indicating markedly different handling of vitamin C in the two groups . Recent reviews have discussed the rationale for investigating the effects of vitamin C as a treatment for sepsis [140, 218–223].
The highest vitamin C levels in the body are in the adrenal glands and brain . Adrenal glands produce hormones that participate in stress responses. Vitamin C deficiency can lead to symptoms reflecting the effects of vitamin C in the nervous system [61, 70, 71, 155, 182–185]. In two controlled trials, vitamin C administration improved the mood of patients hospitalized with acute illnesses [224, 225]. Thus, in addition to somatic effects, vitamin C can have psychological effects that could be beneficial for critically ill patients and might also influence their duration of ICU stay.
Pharmacokinetics of Vitamin C
Oral and intravenous administration of vitamin C lead to substantially different plasma vitamin C levels. When 1.25 g of vitamin C was given intravenously, plasma peak level increased approximately six-fold compared with the level after oral administration of the same dose [226,227]. Because of this, intravenous vitamin C has been used and investigated by some physicians [106, 107, 110, 114, 116, 117, 119, 122, 192–196, 228–235]. Thus, the two methods of administration may lead to different clinical effects even if the dose is the same. Paradoxically, a previous meta-analysis of vitamin C on POAF and hospital stay found that the effects of the two routes were divergent . Oral vitamin C appeared to cause a greater decrease in the incidence of POAF, whereas intravenous vitamin C led to a greater decrease in hospital stay.
Current Findings: Vitamin C and Length of ICU Stay and Duration of Mechanical Ventilation
There are a variety of biochemical reactions in which vitamin C plays a part, a large number of physiological processes that are consequently influenced, and a large number of clinical effects of vitamin C that have been observed in controlled trials [18–49]. Given this, and the evidence for low vitamin C levels and increased metabolism of vitamin C in critically ill patients [90–105], it seemed reasonable to examine whether vitamin C influences practical outcomes, such as the length of ICU stay and the duration of mechanical ventilation without any restrictions on specific medical conditions.
In our meta-analysis of 12 controlled trials with 1766 patients, we found that vitamin C shortened the duration of ICU stay on average by 8% (Figure 3). Meta-analysis of six trials that reported the duration of mechanical ventilation found that vitamin C shortened the duration by 8% (Figure 4).
We did not restrict our study to randomized and placebo-controlled trials. Although methodological details should be reported so that readers can draw their own conclusions about the validity of studies , there is empirical evidence that often trials are better than the publications indicate [237–239]. Therefore, we decided to include studies without strict methodological requirements, and instead used sensitivity analysis to test whether the exclusion of the methodologically less satisfactory studies resulted in a decreased estimate of effect.
In empirical analyses, the placebo-effect has not been quantitatively meaningful when the outcomes were objective . In particular, in an ICU context, the patients are given a number of treatments and it is unlikely that getting one more tablet or infusion would generate a quantitatively relevant placebo effect. In addition, usually ICU patients are so unwell that they are quite often unaware of what treatment they are receiving.
Most of the included trials were randomized (Figure 2, Table S1). Two studies [114, 118] did not report the method of allocation. The trial arms had an identical size, and thus, it is possible that alternative allocation was used. Tanaka allocated patients to groups by the month ; however, it does not seem reasonable to assume that such an allocation necessarily led to systematic bias between study groups, when the study lasted for years. Mirmohammadsadeghi  allocated patients to groups by the personal number of the patients. The purpose of random allocation is to generate study groups that are balanced, and the reported baseline data were balanced in each of the four trials. In any case, in our sensitivity analysis on ICU stay, we excluded three of these studies [114, 118, 121]; however, there was no material difference in the point estimates of effect (Table 2). The fourth study did not report duration of ICU stay .
The calculation of the 8% average decrease in ICU stay is reasonable when testing the null hypothesis that vitamin C has no effect on ICU stay. However, that estimate cannot be generalized to all ICU contexts, since the patients treated in ICU are highly diverse. All except one of the 12 trials in our main meta-analysis on ICU stay were with patients undergoing elective cardiac surgery (Figure 3). However, two non-cardiac trials found a 27% decrease in the duration of mechanical ventilation (Figure 4), indicating that the effects of vitamin C are not limited to cardiac patients. Furthermore, two trials with sepsis patients found other benefits from vitamin C [115, 117]. In addition, a trial with lung contusion patients found that vitamin C was beneficial when patients were also given vitamin E (Table 3).
We did not find any meaningful differences between the effects of oral and intravenous vitamin C (Table 2); however, the confidence intervals are wide, and no firm conclusions should be drawn. Six oral administration trials found an 8% reduction in ICU stay with a mean dose of 2 g/day. The data was too limited to carry out meta-regression by vitamin C dosage; however, the possibility of greater benefits with higher doses should be considered, since linear dose-dependency for 6 to 8 g/day has been found in two common cold studies [40, 42, 204–206]. Common cold studies have indicated differences between males and females in some contexts [202, 209–212]; however, the range of variation in the proportion of males in the studies included in this meta-analysis is so small that a meta-regression was not meaningful.
The effect of vitamin C administration on the duration of mechanical ventilation was significantly greater for ICU patients who were more severely ill (Figure 4). We also found a greater effect of vitamin C in trials in which the length of ICU stay was longer, corresponding to sicker patients, compared with the effect in trials in which the length of ICU stay was shorter (Table 2). This is also consistent with vitamin C having greater effect on more severely ill patients.
Given that all people obtain vitamin C from their food, and there is great variation in vitamin C intake, it is surprising that only one of the included trials measured and reported baseline vitamin C levels. Therefore, the possible role of baseline vitamin C level as a variable defining the benefit of giving vitamin C could not be investigated.
Other Relevant Studies
There are studies not included in our systematic review that are nevertheless informative on this topic. Two controlled trials with critical care patients did not report the length of ICU stay or duration of mechanical ventilation but reported other benefits from vitamin C [241,242].
A controlled trial carried out in Bangladesh reported that 1 g/day intravenous vitamin C decreased mortality for tetanus by 100% in 62 patients aged 1 to 12 years, and by 45% in 55 patients aged 13 years and more . The greater benefit for younger children is consistent with dose dependency. Young children weigh less, and therefore, 1 g/day corresponds to a greater dose per weight. Although the study had various shortcomings, the benefits are not easily explained by the methodological deficiencies . Furthermore, the concept that vitamin C may influence tetanus is supported by animal studies in which vitamin C has protected against tetanus toxin and other clostridial toxins [244, 245].
A randomized trial of 84 patients with acute pancreatitis compared 1 g/day and 10 g/day intravenous vitamin C for five days . The higher dose significantly shortened the duration of fever by 26%, vomiting by 16%, and the length of hospital stay by 30%. Given that the control group was also given quite a high dose of vitamin C, this study indicates dose dependency even at higher doses.
In addition to trials with vitamin C alone, there are controlled trials of vitamin C given with other antioxidants. However, if there is a benefit from a set of several antioxidants, it is not possible to tell which one of them was effective, or whether an interaction between several antioxidants led to the benefit as in Table 3. Similarly, if there is no benefit in another study with several different antioxidants, we do not know whether some of the components alone might be beneficial; another component may have prevented the benefit. The latter is not pure speculation as vitamin E supplementation was harmful for middle aged males with high dietary vitamin C intake, suggesting harmful interaction in certain contexts . Nevertheless, certain studies have reported that mixtures containing vitamin C were beneficial for critically ill patients [247–253].
Furthermore, some before-after studies have reported benefits from vitamin C or from vitamin C together with other antioxidants. However, there can be various temporal changes and other confounders that could bias before-after studies. The concern regarding bias in observational studies is the reason that evidence-based medicine focuses on controlled trials. However, if the differences in a before-after comparison are very large, it is not easy to explain them by confounding. Some before-after studies were consistent with the notion that vitamin C may be beneficial for critically ill patients [254–260]. In one study with 4294 acutely injured patients , the correct number-needed-to-treat to prevent one death with vitamins C and E and selenium was 17 and not 1710 as reported by the authors .
Safety of Vitamin C
In the US nutritional recommendations, the ’tolerable upper intake level’ of adults for vitamin C was stated to be 2 g/day. However, the basis for this upper limit is the appearance of diarrhea  which is a trivial adverse effect. Furthermore, consistent with changes in vitamin C metabolism, patients with various diseases can take much higher doses before they get diarrhea .
Several studies have shown that 4 g/day of vitamin C increases plasma levels of critically ill patients to the levels of the general population [102–105]. Furthermore, all the studies in Figure 3 used doses of 3 g/day or less. Hence, there should not be any safety concerns if the goal of treatment is to reverse the particularly low vitamin C levels in ICU patients.
High doses of vitamin C, over 10 g/day, and in some cases even over 100 g/day, have been administered both orally and intravenously for patients with low frequency of adverse effects [228–235, 262]. However, there are cases of oxalate nephropathy caused by very high doses of vitamin C. In two cases, oxalate nephropathy appeared as the cause of death; however, one of the patients was administered 101 g of vitamin C intravenously in 18 h, and the other was administered 224 g in 20 h . The patients had burns to more than 60% of their bodies and it is not clear what role the vitamin C played in their deaths versus the very severe burns. In any case, it seems evident that doses substantially higher than 3 g/day should be administered with caution, e.g., by monitoring serum oxalate levels.
The rare cases of oxalate nephropathy caused by particularly high doses of vitamin C should not be used as a counterargument for raising very low vitamin C plasma levels of ICU patients to the range of normal healthy people with doses of about 4 g/day [102–105].
Preconceptions about Vitamin C: Why Is Vitamin C not Used Despite the Evidence?
In 1975, Thomas Chalmers published a meta-analysis in which he pooled the findings of eight vitamin C and common cold trials and calculated that colds were 0.11 (SE 0.24) days shorter in the vitamin C groups . Chalmers concluded from the small effect and the large standard error that there was no evidence that vitamin C was effective for reducing the duration of the common cold. That review was highly influential and has been cited in textbooks on nutrition and infectious diseases, and in nutritional recommendations [38, 265]. For example, in 1987, the American Medical Association based its official statement that “One of the most widely misused vitamins is ascorbic acid. There is no reliable evidence that large doses of ascorbic acid prevent colds or shorten their duration” wholly on Chalmers’ review  (p. 1934).
However, in 1995 the Chalmers review was shown to be seriously flawed. There were errors in the extraction of data, inconsistency in the selection of studies for the meta-analysis, and errors in the calculations [38, 265]. If the Chalmers review had included only studies in which the vitamin C dose was 1 g/day or greater, had extracted data correctly, and calculated the results correctly, then the estimate of effect would have been substantially higher: the effect of vitamin C would have been 0.93 (SE 0.22; p = 0.01) days reduction in the duration of colds . Unfortunately, the textbook authors, authors of the nutritional recommendations, and experts of the American Medical Association at that time did not look critically at the original scientific literature on vitamin C and the common cold, and thus missed the errors in Chalmers’ review.
Another influential and frequently cited review by Dykes and Meier from 1975 on vitamin C and the common cold  was also flawed [39, 265]. Linus Pauling also pointed out several problems in the Dykes and Meier review and submitted a manuscript to the Journal of the American Medical Association (JAMA); however, his paper was rejected even after Pauling twice made revisions to meet the suggestions of the referees. The manuscript was finally published in a minor journal [268, 269]. The rejection of Pauling’s review by JAMA was disappointing, since the readers were thereby prevented from seeing the other side of an important scientific controversy. In addition, a highly influential randomized trial on vitamin C and the common cold, also from 1975  was analyzed erroneously [40, 265, 270, 271].
The problems with the two reviews and with the analysis of the randomized trial may be partly explained by the preconceptions of the authors. Furthermore, the very wide acceptance of the two reviews and the randomized trial in mainstream medicine, e.g., reflected by their usage in textbooks and nutritional recommendations etc. , suggests that the readers probably also had strong preconceptions against vitamin C being beneficial; otherwise they might have read the texts with a more critical eye.
Prejudice against vitamin C is not limited to the common cold. Richards documented bias against vitamin C when she compared the attitudes and arguments of physicians to three putative cancer medicines: 5-fluorouracil, interferon, and vitamin C [272–274].
In an analysis discussing the preconceptions in academic medicine on micronutrient supplements, Goodwin and Tangum gave several examples to support the conclusion that there has been systematic bias against the concept that vitamins might be beneficial in levels higher than the minimum required to avoid classic deficiency diseases . In other papers, Goodwin and Goodwin reviewed several cases in which an effective method of treatment was erroneously rejected: the rejection seemed to be caused by the lack of understanding of the physiological mechanism of the effect [276, 277]. This may also apply to the prejudices against vitamin C being beneficial in higher doses for diseases other than scurvy. The effects of vitamin C on scurvy were explainable by its effects on collagen metabolism. However, with the early understanding of biochemistry it was not evident how vitamin C could have an effect on infections, cancer, or other diseases that did not involve collagen metabolism. Thus, it seems that the limited biochemical understanding may have been one reason why the early findings [2, 186–199, 244] were not pursued.
Using vitamin C to prevent or treat diseases is often classified in the field of alternative medicine. However, such a classification does not mean that vitamin C is ineffective for all conditions other than scurvy .
Conclusions: The Way Forward
We found statistically highly significant evidence that vitamin C can shorten the length of ICU stay. We consider that our finding is a proof of concept, strongly encouraging further research, rather than justifying recommendations for change in practice. In further studies, the dose-response relationship should be carefully investigated, and oral and intravenous administration should be compared directly. Given that some common cold studies found the benefit of vitamin C to be greater in males than females, the effects of vitamin C in the ICU context should be compared between sexes. ICU patients are a highly heterogeneous group, and evidently, one estimate of vitamin C effect should not be expected to apply to all patient groups. Our analysis on mechanical ventilation indicated that the benefits of vitamin C may be greater for patients with more severe illness.
Studies of arbitrary combinations of antioxidants teach us little, since they cannot be compared or pooled and drawing conclusions about a specific antioxidant is impossible. Instead, factorial trials such as the one shown in Table 3 are informative regarding individual effects and possible interactions. Although the dose of vitamin C may influence the size of effect, it is likely that the baseline level of vitamin C also has an impact on the benefits of vitamin C administration and should therefore be measured. In addition, it is important that any further studies are sufficiently powered to be able to detect a reasonable effect and the required sample size should be calculated accordingly.
Vitamin C costs only pennies per gram, whereas one day in the ICU may cost thousands of dollars; therefore, an 8% decrease in ICU stay from the administration of 2 g/day of vitamin C warrants further research.