Effect of vitamin C deprivation on the duration of colds in the Sheffield study (1953): a statistical analysis

Effect of vitamin C deprivation on the duration of colds in the Sheffield study (1953): a statistical analysis Harri Hemilä Department of Public Health, University of Helsinki, Helsinki, FINLAND. [email protected] https://orcid.org/0000-0002-4710-307X https://pubmed.ncbi.nlm.nih.gov/?term=hemila+h https://researchportal.helsinki.fi/fi/persons/harri-hemil%C3%A4 https://scholar.google.fi/citations?user=2mkomzUAAAAJ Location of this file: https://doi.org/10.5281/zenodo.14717361 ver. 2025-01-22 Summary The Sheffield study on the vitamin C requirement of human adults was carried out in the UK from 1944 to 1946. The main goal was to establish the minimum dose of the vitamin that cures scurvy. The participants were 19 men and 1 woman, aged 21-34; they lived a normal life without strenuous physical work. Evidently the conclusions are limited, given that the majority were male, the age range was small, and the lifestyles were sedentary. The Sheffield study is small and old, but it is a study of great importance. For example, it is the first reference in the current UK recommendations for vitamin C. For ethical reasons it is highly unlikely that a similar trial will be carried out in the future. Therefore, a critical analysis of the data is important even now. A recent (2022) reanalysis showed that the original conclusions about the impact of vitamin C on scar strength were flawed. In the study, participants were divided into 3 groups: 10 participants were “deprived” and not given vitamin C supplements (diet contained ~1 mg/day), 7 participants were administered 10 mg/day vitamin C as a supplement, and 3 participants were administered 70 mg/day. One of the recorded outcomes was the duration of common cold episodes. The statistician for the study (C. H. Jowett) calculated that colds lasted on average 6.4 days during vitamin C deprivation, compared with 3.3 days on the 10-70 mg/day dosage. Thus, deprivation nearly doubled the duration of colds. Jowett concluded that “such evidence as there is, however, definitely confirms the hypothesis that the absence of vitamin C tended to cause colds to last longer” and “the data support the hypothesis that colds of deprived subjects lasted longer, but do not establish it.” In this paper, the data on common cold duration is reanalyzed. There were 18 cold episodes during vitamin C deprivation, and 21 cold episodes when participants were administered 10-70 mg/day vitamin C. Vitamin C deprivation increased the duration of colds on average by 77% (P = 0.014). Vitamin C deprivation decreased the recovery rate from colds by 60% (P = 0.008). Vitamin C deprivation extended the duration of 1-day colds by 2.2 days (95% CI 1.0 to 5.4 days). The finding that vitamin C deprivation extended the duration of colds was not reported in the summaries of the trial published in the Lancet (1948) and in the Proceedings of the Nutrition Society (1953). In addition, this finding is also not mentioned in the current UK recommendations for vitamin C. As a consequence, readers have not been fully informed about the common cold results of the Sheffield trial for several decades. 1 Contents Page Background 3 Observations on the common cold 4 References to the Sheffield study 6 Reanalysis of the data on common cold duration in the Sheffield study A. Repeat of the Jowett (1953) analysis 7 B. Difference in the proportion of short 1-day colds 8 C. Relative effect of vitamin C deprivation on common cold duration 10 D. Recovery rate from the common cold using Cox-regression 11 E. Quantile treatment effect (QTE) of vitamin C deprivation 12 Discussion Ethical issues with intentionally causing scurvy in experimental subjects 14 The common colds of vitamin C deprived subjects lasted longer 16 Vitamin C and respiratory infection research before 1948 17 Flawed conclusions about causal effects 19 The Sheffield study has had great influence on the vitamin C recommendation in the UK 20 Bias against vitamin C for infections 21 Conclusions 22 Other references 23 Appendix 1. Common colds in the Sheffield study [5] 30 2. Data set based on Appendix 1 36 3. Findings in the Glazebrook and Thomson (1942) trial 37 4. Statistical issues 42 5. Replication of the Jowett (1953) analysis 44 6. Relative effect 46 7. Cox regression 47 8. QTE analysis 48 9. Vitamin C deficiency and pneumonia in Alfred Hess monograph on scurvy (1920) 50 10. Findings in the Cowan et al. (1942) trial 55 11. Benefit from vitamin C for common colds in UK trials after the Sheffield study 57 12. Possible socio-political roots for the bias against vitamin C supplements 58 2 Background The Sheffield study on the vitamin C requirement for human adults was carried out at the Sorby Research Institute at Sheffield, England, under the supervision of Hans Krebs (1953 Nobel Prize in Physiology or Medicine) from October, 1944, to February, 1946 [1-15]. A 6-page summary of the findings was published in the Lancet in 1948 [3]. In 1953, another 16-page summary by Hans Krebs was published in the Proceedings of the Nutrition Society [4], along with the detailed 179-page final report [5-8]. The selection of participants and the methods were summarized by Krebs as follows [4]: The main plan was to induce scurvy by a vitamin C-deficient diet and to establish the minimum dose of the vitamin that cures scurvy. A subsidiary aim was to study the clinical signs and symptoms of scurvy and to correlate them with laboratory findings (p.238). Nineteen men and one woman, aged 21-34, volunteered for the experiment. They lived a normal life without strenuous physical work. Their basal diet was designed to be as low as possible in vitamin C but complete in every other respect. It was sufficiently varied to be acceptable. From chemical analyses it was calculated that on the average a volunteer obtained not more than 1 mg vitamin C daily from the diet (p.238). The majority were conscientious objectors [pacifists] who had convictions which did not allow them to carry arms, but who did not wish to eschew hardship and danger. They offered themselves to serve a cause they considered good. They were fully aware of the nature of the trial and of the risks it involved (p.246). To obtain base-line data the trial began with a preliminary period, in most cases of 6 weeks, on a complete diet including about 70 mg vitamin C daily. At the end of the period all the volunteers were given the basal deficient diet and divided into three groups, ten having no supplements, seven 10 mg of vitamin C daily, and three 70 mg vitamin C daily. The group receiving 70 mg was intended to serve as a positive control and the group receiving 10 mg as a prophylactic test (p.238). The volunteers did not know to which group they belonged, nor did the physicians responsible for the clinical investigation. All the volunteers were daily given seven supplementary tablets of identical taste and appearance, some containing vitamin C, the others being dummies. Investigations, made on the volunteers at regular intervals, included general clinical examinations, chemical analyses of blood and urine, haematological examinations, capillary-fragility tests, capillaroscopy, radiography, electrocardiography, studies of fatigue, and studies of experimental wounds (p.238). Important clinically relevant data were collected and reported. For example, 2 of the 10 participants with vitamin C deprivation (Milburn [5,8; p.80-81] and “Another” [5,8; p.87]) suffered from acute cardiac events, which were relieved with immediate large doses of vitamin C [3-5,8]. A third deprived person (Williams, D) had pain in chest and shortness of breath after half a year of vitamin C deficiency. X-ray of chest revealed paravertebral abscess and an erosion of the 8th dorsal vertebra, which was interpreted as an active tuberculosis lesion. The condition was relieved with 500 mg/day vitamin C for 10 days. The study authors concluded that “although the process probably began before the deficient diet started it is likely that its development was precipitated by the [vitamin C] deprivation” [5,8; p. 84-85]. A fourth deprived man (Drake) developed effusions into left knee joint after a long walk after half a year of vitamin C deprivation [5,8; p.78]. Swelling of the knee reduced after dosing 10 mg/day vitamin C. 3 Observations on the common cold Throughout the Sheffield study the volunteers recorded the number and duration of the colds they experienced. The collected data are published in Table 44 of the main report [5, pp.134-135], with the discussion of the findings on pages 43-44 [5]; see Appendix 1 and 2. The study authors described their impression of the results in the trial report as follows [5, p.43]: At a glance [the data] seemed to indicate that the average number of colds in the deprived and non-deprived groups did not differ markedly, but that they lasted longer in the deprived group. The material was accordingly submitted to a statistician (C. H. Jowett) for analysis, who summarized his investigation as follows: The statistician C. H. Jowett wrote in the final report (p.43-44): The lengths of all colds were subjected to the transformation y = 20 log x. It was considered that this would make the various distributions on which the statistical tests depended more close to ‘normal’. The data analysed consisted of the following: Colds of members of deprived group before dosing. Colds of members of supplemented group up to and including July 1945. … Such evidence as there is, however, definitely confirms the hypothesis that the absence of vitamin C tended to cause colds to last longer. The geometric mean length of colds of non-deprived subjects = 3.3 days The geometric mean length of colds of deprived subjects = 6.4 days … Conclusion The data support the hypothesis that colds of deprived subjects lasted longer, but do not establish it. The Lancet report [3] and the 1953 short report [4] did not mention the doubling of common cold duration in the deprived subjects. They mentioned only the negative findings on common cold incidence. The former stated that in volunteers deprived of vitamin C there was “no increased incidence of infection” [3, p.854]. In the latter, Krebs wrote that “Negative findings during the period of deprivation included … no increased incidence of infection” [4, p.240]. It is evident that the comparison of 10 vitamin C deprived volunteers with 10 volunteers receiving vitamin C does not have the statistical power to test modest size effects on the incidence of colds. In comparison, Glazebrook and Thomson (1942) compared 335 boys in a boarding school (~200 mg/day vitamin C) with 1100 control boys (10-15 mg/day vitamin C from foods) and found a 17% lower incidence of colds (P = 0.047) and a 100% decrease in the incidence of pneumonia (P = 0.005); see Appendix 3. Obviously, with the 33to 110-fold smaller size of the Sheffield study, it is uninformative about the possible effects of vitamin C deprivation on the incidence of the common cold or pneumonia. It is a matter of opinion whether a potential 17% decrease in common cold incidence is large enough to be of practical interest. However, such an effect may also indicate broader physiological effects of vitamin C, such as influence on pneumonia and rheumatic fever, which are much more rare events, and days spent in the sick room due to various infections; see Appendix 3. 4 Neither of the short summaries noted that there were so few participants in the Sheffield study that it was uninformative about the incidence of colds yet both of them made firm negative statements [3,4]. Furthermore, neither of the summaries mentioned that colds were twice as long in the vitamin C deprived participants even though in the text section the long report states “The data support the hypothesis that colds of deprived subjects lasted longer” [5, p.44]. When the observation indicates that vitamin C deprivation may double the duration of colds, it would seem relevant to encourage further research to corroborate the finding, and to investigate dose dependency and potential variation in effects between different life-style contexts. Unfortunately, such encouragement was not expressed in the short reports [3,4], and not even in the summary or concise account of the long report [5,6; pp.6-21,143-4]. This report reanalyzes the duration of common colds in the Sheffield study. 5 References to the Sheffield study 1. Sorby Research Institute. Wikipedia 2025-1-16. https://en.wikipedia.org/wiki/Sorby_Research_Institute also: https://archives.shef.ac.uk/agents/corporate_entities/143 https://en.wikipedia.org/wiki/Kenneth_Mellanby 2. Rasmussen, L. (eds) Human Guinea Pigs, by Kenneth Mellanby: A Reprint with Commentaries. Philosophy and Medicine, vol 134; Springer, Cham. [Describes the context for the trial]. https://doi.org/10.1007/978-3-030-37697-0 3. Vitamin C requirement of human adults; experimental study of vitamin-C deprivation in man. [Summary]. Lancet. 1948;1(6510):853-8. https://doi.org/10.1016/S0140-6736(48)90572-8 4. Krebs H. The Sheffield experiment on the vitamin C requirement of human adults. [Summary]. Proceedings of the Nutrition Society. 1953;12(3):237-46. https://doi.org/10.1079/PNS19530054 5. Vitamin C requirement of human adults. Spec Rep Ser Med Res Counc (GB). 1953;280:1-179. https://pubmed.ncbi.nlm.nih.gov/13119061 6. Introduction; Historical Account; Concise Account of the Experiment; Summary. In: Vitamin C requirement of human adults. Spec Rep Ser Med Res Counc (GB). 1953;280:1-21,143-4. https://doi.org/10.5281/zenodo.14628064 7. Diet. In: Vitamin C requirement of human adults. Spec Rep Ser Med Res Counc (GB). 1953;280:56-69. https://doi.org/10.5281/zenodo.14628199 8. Introduction; Historical Account; Case Reports; References. In: Vitamin C requirement of human adults. Spec Rep Ser Med Res Counc (GB). 1953;280:1-8,73-88,176-9. https://doi.org/10.5281/zenodo.7661782 9. Experimental scurvy. Lancet. 1954;263:197-8. https://doi.org/10.1016/S0140-6736(54)91268-4 10. Waife SO. Man's requirement for vitamin C. Am J Clin Nutr. 1954;2(4):273-4. https://doi.org/10.1093/ajcn/2.4.273 11. Pemberton J. The BMJ’s Nuremberg issue. Nobody died during experiments on vitamin C and vitamin A intakes in Sheffield. BMJ. 1997;314(7078):440. https://pmc.ncbi.nlm.nih.gov/articles/PMC2125893 see also: https://pubmed.ncbi.nlm.nih.gov/8973237/ (Weidling) https://pubmed.ncbi.nlm.nih.gov/9133906/ (Addis) 12. Pemberton J. Medical experiments carried out in Sheffield on conscientious objectors to military service during the 1939-45 war. Int J Epidemiol. 2006;35(3):556-8. https://doi.org/10.1093/ije/dyl020 13. Commentary: guinea-pigs’ private war. Int J Epidemiol. 2006;35(3):558-60. https://doi.org/10.1093/ije/dyl031 14. Pemberton J. Unrecognised scurvy. Signs and requirements. BMJ. 2010;340:c590. https://doi.org/10.1136/bmj.c590 15. Hujoel PP, Hujoel MLA. Vitamin C and scar strength: analysis of a historical trial and implications for collagen-related pathologies. [Reanalysis of the scar strength data]. Am J Clin Nutr. 2022;115(1):8-17. https://doi.org/10.1093/ajcn/nqab262 6 Reanalysis of the data on common cold duration in the Sheffield study A. Repeat of the Jowett (1953) analysis I repeated the analysis of variance by Jowett. The data set in Appendix 2) collects the observations selected by Jowett [5, p.43]: “colds of members of deprived group before dosing” “colds of members of supplemented group up to and including July 1945” The current analysis yields similar results to those found by Jowett, as detailed in Appendix 5. The differences may be explained by less accurate pencil and paper calculations in the late 1940s. The same data set was used in analyses B to E. The duration of colds has a skewed spread and the log transformation used by Jowett is insufficient to lead to a distribution approximating the normal distribution (see Appendix 5). There is a high proportion of 1-day colds when participants were administered vitamin C, and no transformation can make such a cluster approach a normal distribution. On the other hand, the 1-day colds can be analyzed as a separate outcome; see analysis B. Assuming a uniform difference of 3.1-day longer colds (6.4 – 3.3 days calculated by Jowett, see above) may be overly simplistic. For example, 2-day colds in the vitamin C deprived participants cannot be shortened by 3.1 days by taking vitamin C. The relative effect is a more informative approach; see analysis C. Cox regression is also useful for the analysis of duration data and does not assume a uniform difference such as 3.1 days over the whole distribution; see analysis D. The number of participants in the Sheffield study is so small that subgroup analyses cannot be undertaken. Nevertheless, the quantile treatment effect (QTE) analysis does not assume a uniform effect and can be more informative than assuming a uniform effect; see analysis E. 7 B. Difference in the proportion of short 1-day colds This analysis compared the proportion of short 1-day colds between the vitamin C deprived participants and those who were administered 10-70 mg/day vitamin C. This approach does not require the assumption that the data are normally distributed. If a substantial proportion of colds are shortened due to vitamin C, that is relevant even if there are no effects on the distribution of colds that are longer. Among the 10 men administered 10-70 mg/day vitamin C, there were 21 colds. See the data set in Appendix 2. The maximum number of colds per person was 6 (Golding), and 4 participants had 3-4 colds per person. Four participants had no colds during the vitamin C periods. In this group, 7 of the 21 colds (33%) were short 1-day colds. Among the 10 men deprived of vitamin C, there were 18 colds during the deprivation period. The maximum number of colds per person was 7 (Tridgell), and 2 participants had 3 colds per person. Four participants had no colds during the follow-up. In this group, 0 of the 18 colds (0%) were short 1-day colds. Ignoring the fact that there were several colds per person for some participants, there was a significant difference in the proportion of 1-day colds between the participants with vitamin C deprivation and those who were administered vitamin C 10-70 mg/day with P = 0.004; see below. The mid-P and 1-tail were used in the calculation of the P value; see Appendix 4. In this comparison, the Number Needed for Harm (NNH) [16,17] is 3; based on 1/(33% – 0%). This means that 1 in 3 colds in the deprived group was not a short 1-day cold because of the vitamin C deprivation. > riskratio.small(SheffieldJowett$Deprived, SheffieldJowett$Days_1) $data Outcome Predictor 0 1 Total 0 14 7 21 1 18 0 18 Total 32 7 39 $p.value two-sided Predictor midp.exact fisher.exact chi.square 0 NA NA NA 1 0.00756 0.009629 0.006848 > (P_1tail=0.00756/2) [1] 0.00378 8 A similar calculation was undertaken using 1 cold per person, the shortest cold, which means that there is a single observation for each participant who had colds. In this calculation, the units of analysis are independent. Among the 6 participants administered 10-70 mg/day vitamin C who had colds during the followup, there were 4 participants whose shortest colds were just 1 day. Among the 6 participants deprived of vitamin C who had colds, none had colds shorter than 2 days, see Appendix 2. For the difference in the proportions of 1-day colds between the vitamin C deprived participants and those administered vitamin C, P = 0.015. > riskratio.small(SheffieldShortest) $data Outcome Predictor Disease1 Disease2 Total Exposed1 2 4 6 Exposed2 6 0 6 Total 8 4 12 $p.value two-sided Predictor midp.exact fisher.exact chi.square Exposed1 NA NA NA Exposed2 0.0303 0.0606 0.0143 > (P_1tail=0.0303/2) [1] 0.0152 9 the vitamin C experiment. After spending 229 days on a diet as deficient in vitamin C as we could make it he had clear evidence of scurvy, including bleeding gums and haemorrhages round the hair follicles of the legs. One night, after these signs had appeared, he developed all the major symptoms of coronary thrombosis. He was treated at once with intravenous vitamin C and recovered completely. Although it is great that Milburn did not die, the context in the study seems to have been close to the Russian roulette. In his summary of earlier literature, Lind (1757) described that many scurvy patients just fell down dead without obvious anticipatory symptoms, see above and [25,27]. Thus, it may have been good luck that neither of the 2 cases of cardiac emergencies died. Given the extensive evidence that vitamin C deficiency can lead to acute cardiac events and death, it seems highly unlikely that in future there will be any trial comparable to the Sheffield study in which the test group is administered ≤1 mg/day vitamin C for half a year and longer. Given this, it is worthwhile analyzing the data collected in the Sheffield study as extensively as reasonable. Recently Hujoel and Hujoel reanalysed the data on scar strength in the Sheffield study [15]. They showed that vitamin C intake that averaged 10 mg/day over a mean follow-up of 11.5 months was associated with a 42% weakened scar strength when compared with a vitamin C intake of 80 mg/day (P < 0.001). They concluded that: the observed dose-response curve between scar strength and vitamin C intake suggests that the daily vitamin C intake needed to prevent collagen-related pathologies is in the range recommended by the National Academy of Medicine and the European Food Safety Authority (75 to 110 mg/day), not the WHO recommendation (45 mg/day). They also concluded that “the prior lack of statistical analyses of a landmark trial may have led to a misleading narrative on the vitamin C needs for the prevention and treatment of collagen-related pathologies.” Thus, the original statistical analysis was insufficient, and the widely repeated conclusion that the Sheffield study showed that 10 mg/day leads to maximal recovery of wounds is false. It was timely to also reanalyze the data on common cold duration. The common colds of vitamin C deprived subjects lasted longer The current analysis calculated that vitamin C deprivation increased the duration of colds on average by 77% and decreased the recovery rate from colds by 60% (Fig. 1). In addition, vitamin C deprivation extended the duration of 1-day colds by 2.2 days (Fig. 2). These findings are consistent with Jowett’s calculations, which led to the conclusion that “The data support the hypothesis that colds of deprived subjects lasted longer, but do not establish it”. The Cochrane review on vitamin C and the common cold (2013) pooled the results of 31 placebocontrolled trials in which ≥200 mg/day vitamin C was administered regularly over the follow-up period [66,67]. There was very strong evidence that vitamin C shortened the duration of the common colds in the included trials with P = 0.000 000 07 (Z = 5.28 [66: Analysis 2.1]). Thus, the Sheffield study indicated that the duration of colds is an outcome on which vitamin C intake has an effect, and the Cochrane review (2013) confirmed that vitamin C dose has an impact on common cold duration. Nevertheless, the relationship between vitamin C dose and the duration 16 of common colds is complex and there seem to be differences between populations groups; see e.g. [66,68,69]. The doubling of common cold duration due to vitamin C deprivation in the Sheffield study should have led to explicit suggestions to investigate in more detail the relationship between the vitamin C intake level and the duration of the common cold with larger populations, and with wider ranges of vitamin C intakes, but such encouragement was not published [3-5]. The Lancet summary (1948) defined [3, p. 857]: The term requirement is here used to mean the amount of a dietary essential which must be eaten to maintain full health. The final report (1953) defined [5,6; p.20]: The term “requirement” is here used to mean the amount of a dietary essential which must be eaten to maintain full health. Given this definition, it is illogical to ignore the observed doubling of common cold duration during periods of vitamin C deprivation. Shorter duration of colds is relevant to a discussion of “full health”. Therefore, the outcome “common cold duration” is important when considering the vitamin C “requirement”. The Sheffield study researchers did not define “scurvy”. If scurvy refers to pathological effects caused by vitamin C deficiency, then increased incidence of pneumonia and the common cold, extension of the duration of colds, and the occurrence of cardiac emergencies in 20% of participants are relevant issues. However, it seems that the researchers were primarily interested in the dermatological effects of vitamin C deficiency, which is a very narrow view. Vitamin C and respiratory infection research before 1948 The Sheffield study was not carried out in a vacuum, as there were several previous papers in the medical literature on vitamin C and respiratory infections. However, they were not properly discussed in the long report [5] and not mentioned at all in the short reports that simply stated that vitamin C did not influence the number of colds [3,4]. Thomas Barlow (1894), who identified infantile scurvy (Barlow's disease), stated in his Bradshaw lecture that “Eleven years ago I submitted to the consideration of members of our profession the results of an analysis of thirty-one cases of what I believed to be infantile scurvy… If the cachexia is very profound the supervention of bronchitis, pleuro-pneumonia, severe diarrhoea, or an intercurrent exanthem may bring about a fatal issue” [70,71]. Thus, Barlow considered that scurvy can cause elevated risk of severe respiratory infections. Alfred Hess was an American pediatrician who appears the most important clinical researcher on scurvy in the 20th century. In his monograph (1920) [27,43] and papers [72-74], Hess described that vitamin C deficiency increases the risk of pneumonia and other infections in guinea pigs and humans; see Appendix 9. Albert Sabin reported on his studies with rhesus monkeys that “… it was found that monkeys on a scorbutic diet died of spontaneous acute infections, chiefly pneumonia and enterocolitis, while their mates receiving an adequate diet remained well” [75]. Although findings in guinea pigs and rhesus monkeys do not directly translate to effects in humans, 17 they should have provoked a deeper consideration of the role of vitamin C in human infections. In addition, Hess also treated children and not just guinea pigs. In the introduction to their trial report, Glazebrook and Thomson (1942) [76, p.4] wrote: There is evidence that it [vitamin C] is of value in pneumonia, particularly in hastening convalescence, and the claims made do not appear to have been contradicted (Gander & Niederberger, 1936 [77]; Vogl, 1937; Bonnholtzer, 1937 [78]; Hochwald, 1937 [79]; Günzel & Kroehnert, 1937 [80]; Sennewald, 1938; Szirmai, 1940). Szirmai (1940) noted that while tissue saturation is necessary to obtain maximal benefit in pneumonia, cases of typhoid fever and diphtheria were improved by daily supplements of vitamin C without producing saturation. In addition to the listed papers in German, benefit from vitamin C for pneumonia was also reported in the USA by Frederick Klenner [81]. Before 1948 there were also reports stating that vitamin C was beneficial against the common cold [82-84]. Though reports of uncontrolled observations do not demonstrate that vitamin C is definitively beneficial for the common cold, they too should have provoked proper consideration of the positive findings in the placebo-controlled Sheffield study. The Glazebrook and Thomson (1942) trial included 1435 boarding school boys as participants compared with just 20 participants in the Sheffield study. Glazebrook and Thomson described that they examined boys “from the lower wage-earning classes” with “the food distribution [at the school] was badly managed… Often 8 hr. elapsed between the time the food was cooked and its arrival on the dining tables…” [76]. The control group received 10-15 mg/day in foods; nevertheless, administration of ~200 mg/day vitamin C had significant effect on various clinically relevant outcomes; see Appendix 3. The final report wrote about the Glazebrook and Thomson (1942) [76] trial as follows [5, p.44]; see Appendix 1: In connexion with this result mention should be made of the observation of Glazebrook and Thomson (1942) who studied the incidence and duration of infectious diseases in groups of adolescents living in an institution where the dietary level of vitamin C was very low. The incidence of the common cold and tonsillitis or the average duration of illness due to the common cold was not affected by vitamin C supplements, but the average duration of illness due to tonsillitis was longer in the unsupplemented group. This is a greatly biased description about the findings; see Appendix 3. The final report could have stated: Glacebrook and Thomson (1942) found that schoolboys who were administered vitamin C had 23% decrease in the number of common colds needing treatment in the sick quarters (P < 0.05), 57% decrease in the days in hospital for tonsillitis patients (P < 0.01), 100% decrease in the incidence of pneumonia (P < 0.01), 100% decrease in the incidence of rheumatic fever (P < 0.01), and 50% decrease in “days spent in the sick room per boy due to infective conditions” (P < 0.00001). Given that the control group obtained 10-15 mg/day vitamin C in foods, the study findings are very strong evidence against 10 mg/day providing “sufficient” dose of vitamin C to maintain “full health”. 18 Thus, the Sheffield study final report ignored many significant benefits from vitamin C in the Glazebrook and Thomson (1942) trial. In addition, the description in the Sheffield report that “adolescents living in an institution where the dietary level of vitamin C was very low” is illogical. Glazebrook and Thomson (1942) estimated that dietary vitamin C intake was 10-15 mg/day. The Sheffield study concluded that 10 mg/day of vitamin C is a sufficient dose and logically 10-15 mg/day should also be sufficient and not “very low”. Appendix 10 shows the findings in the Cowan et al. (1942) trial [85] in the USA which recruited 363 schoolchildren compared with the 20 participants in the Sheffield study. In the Cowan trial, vitamin C intake level in the control group was not estimated, but it is unlikely to have been as low as for the Glazebrook Thomson trial. Cowan stated: “we did no analyses to see whether or not the students were suffering from vitamin deficiency. We assumed that they were not, since most of them were on a reasonably adequate diet” (p.1271), and Diehl stated: “the reason for making this study was not that we thought these students exhibited vitamin deficiencies but that the average population is buying millions of dollars’ worth of these vitamins every year” (p.1271). Thus, there is no basis to assume that food was as poor in vitamin C as the food in the Glazebrook and Thomson (1942) trial. In the Cowan et al. trial, the supplement of 200 mg/day vitamin C significantly decreased the incidence of colds by 14% and the absence from school by 31%. This report, published in JAMA, was also ignored in the Sheffield final report [5], and in the short summaries [3,4]. Thus, the findings in the Glazebrook and Thomson (1942) trial and the Cowan et al. (1942) trial were inconsistent with the Sheffield study (1953) interpretation that 10 mg/day of vitamin C leads to “full health”, but both controlled trials were not considered in the Sheffield report or the short summaries [3-5]. The low level of vitamin C intake generated in the deprived participants in the Sheffield study (~1 mg/day) is extremely rare in the community. However, research even before the Sheffield study indicated that the effects of vitamin C on infections are not limited to the 1 mg/day vs. 10 mg/day comparison, but there seem to be effects by doses much higher than 10 mg/day. Flawed conclusions about causal effects The Lancet 1948 summary wrote: “Some other negative findings are worth recording. There was… no increased incidence of infection…” [3, p.854]; “Many signs listed as scorbutic in the classical description of the disease —e.g., pallor, dryness of the skin, anæmia, and night-blindness—were not observed. It is probable that classical scurvy was often a multiple deficiency” [3, p.857]. In the 1953 summary, Krebs wrote: “Negative findings during the period of deprivation included… no increased incidence of infection…” [4, p.240] “Many signs listed as scorbutic in the classical description of the disease, e.g. pallor, dryness of the skin, anaemia, and night-blindness, were not observed. It is probable that classical scurvy was often a multiple deficiency” [4, p.240]. This is not a valid argument. First, the number of deprived participants was very small. Only 10 participants were vitamin C deprived while the control group also had just 10 participants. If the frequency of a particular symptom is low, it is missed simply because of the small number of participants. As an illustration 19 of this argument, smoking increases the risk of lung cancer greatly, by 10-fold. However, if 10 smokers are followed up for a period, it is unlikely that any of them will get lung cancer because the baseline rate is very low. The great majority of smokers don’t get lung cancer and observing that no-one in a group of 10 smokers gets lung cancer is not a valid argument against the statement that smoking causes lung cancer. For an example of a study with appropriate statistical power, the Doll and Peto study of British doctors recruited 34,440 men in 1951 and they were followed for 2 decades [86]. As to infections, the Glazebrook and Thomson (1942) trial recruited 1435 boarding school boys and the Cowan et al. (1943) trial recruited 363 schoolchildren. The Sheffield study had just 20 participants who were aged 21-34, and not undertaking strenuous physical work. Thus, for the outcome “incidence of infection”, the Sheffield study was minuscule. In addition, the participants were not children, and they had a sedentary lifestyle, whereas the majority of schoolchildren in the 1930s and 1940s had a physically active lifestyle. In fact, Glazebrook and Thomson wrote that “Physical training and games occupied much of the day, and it was found that youths at rest in bed required approximately half the quantity of vitamin C, i.e. 2000 mg., to produce full saturation” (p.16). Thus, even if the Sheffield study had been much larger, the participants and contexts were so different that the study might not necessarily have measured the same effects as the 2 trials in schoolchildren. Second, even if there was possibly a multiple vitamin deficiency in some old cases of scurvy, that does not mean that vitamin C deficiency could not be “a cause” for symptoms of these cases. For example, in 2 recent case reports of scurvy, hemoglobin level doubled during vitamin C administration [53,60]. It is evident that low vitamin C intake was a cause of anemia in these 2 cases. Anemia does not need to occur in all people with scurvy as noted in the previous point. In addition, emergence of anemia might require other simultaneous factors, but the simultaneous factors do not refute that vitamin C deficiency can be a cause of anemia. Rothman and Greenland formulated this issue as follows [87, p.S145]: The importance of multicausality is that most identified causes are neither necessary nor sufficient to produce disease. Nevertheless, a cause need not be either necessary or sufficient for its removal to result in disease prevention. If a component cause that is neither necessary nor sufficient is blocked, a substantial amount of disease may be prevented. That the cause is not necessary implies that some disease may still occur after the cause is blocked, but a component cause will nevertheless be a necessary cause for some of the cases that occur. That the component cause is not sufficient implies that other component causes must interact with it to produce the disease, and that blocking any of them would result in prevention of some cases of disease. Thus, one need not identify every component cause to prevent some cases of disease. In the law, a distinction is sometimes made among component causes to identify those that may be considered a “proximate” cause, implying a more direct connection or responsibility for the outcome. The Sheffield study has had great influence on the vitamin C recommendation in the UK The current (2025) UK nutritional recommendations for vitamin C were formulated in 1991 [88], with the Sheffield study being the first reference. However, the UK recommendation does not mention that colds were twice as long in vitamin C deprived participants in the Sheffield study. Neither does the recommendation mention the Glazebrook and Thomson (1942) trial although it was 71-fold larger than the Sheffield study, was also carried out in the UK, and found significant benefit from vitamin C administration to schoolchildren who had baseline vitamin C intake at the level of “requirement”. 20 Furthermore, after the publication of the Glazebrook and Thomson (1942) trial and the Sheffield (1953) study there were several further trials before 1991 that found vitamin C to be beneficial for the common cold, i.e., indicating that ordinary diets in the UK may have been too low in vitamin C to “maintain full health”; see Appendix 11. A systematic review published in 1992 identified 18 placebo-controlled trials and all of them found shorter duration or less severity of colds in groups administered 1 g/day of vitamin C compared with the placebo groups, with P = 0.000 004 in a binomial distribution [89]. All of these 18 trials were published before 1991, and all were not considered in the UK recommendations. Another systematic review published in 1997 showed that vitamin C administration decreased the incidence of colds in 4 trials with British males, on average, by 30% (P = 0.000 001) [90]. A fifth trial found a statistically significant reduction in recurrent colds in males during the follow-up period after 2½-day vitamin C treatment for the first cold (P = 0.018). See Appendix 11 for the positive findings in the UK trials between 1953 and 1991. These trials would seem to appear particularly relevant when considering vitamin C intake levels in the UK. However, these trials were also ignored in the UK recommendations [88]. Chris Bates criticized the 1997 meta-analysis [90,91]. He was the Vice-Chariman of the Panel on Dietary Reference Values Working Group on Vitamins, and a Member of the Committee on Medical Aspects of Food Policy Panel on Dietary Reference Values for the 1991 recommendations [88]. However, his criticism was ill conceived [91]. Bias against vitamins Given that the Sheffield report published Jowett’s conclusion “Such evidence as there is, however, definitely confirms the hypothesis that the absence of vitamin C tended to cause colds to last longer… The data support the hypothesis that colds of deprived subjects lasted longer, but do not establish it”, it is puzzling that this was ignored in the full report [5], in the summaries [3,4], and in the UK vitamin C recommendation (1991) [88]. As described above, this was not an isolated observation, instead there were many positive reports on the effects of vitamin C against respiratory infections after the pure substance became available in the 1930s [77-84,92-94]. A few controlled studies reported benefit from vitamin C administration, Appendix 3 and 10. A long time ago, vitamin C deficiency was identified as the explanation for scurvy, which was largely considered a disease of the connective tissues. This is illustrated also by the focus of the Sheffield study, with the main interest being on the timing and scope of dermatological changes and the recovery of wounds. It seems that the 20% incidence of cardiac emergencies, the 10% incidence of tuberculosis, and the 10% incidence of effusions in a knee joint were considered more as nuisances disturbing the dermatological observations, rather than clinically more important observations with greater relevance on “full health”. Evidently, it seemed illogical to consider that a substance that participates ‘only’ in connective tissue metabolism might also have effects on the common cold, which may explain that the observations on common cold duration were also ignored. This backgrounds may partly explain the rejection of the series of positive findings on vitamin C for infections over decades [95,96]. There can also be socio-political roots to the bias against vitamin C supplements, see Appendix 12. 21 Conclusions The Sheffield study observed that vitamin C deprivation nearly doubled the duration of colds. The reanalysis in this paper found that vitamin C deprivation increased the duration of colds on average by 77% and decreased the recovery rate from colds by 60%. Intake of vitamin C in the deprived participants was particularly low (~1 mg/day) and such levels are very rare. However, there was no basis to assume in the Sheffield study that the effect of vitamin C on the common cold was limited to the very low dose range. In fact, 2 placebo-controlled trials carried out before the Sheffield study had found significant benefits from vitamin C with higher doses. These findings were however ignored in the 179-page final report and in the major summaries of the Sheffield study. The 2 cases of cardiac emergencies among the 10 scurvy patients were mentioned but downplayed in the reports. There was no conclusion that the effects of vitamin C on cardiac function in the low intake ranges should have been studied in more detail. The bias against the positive observations on the common cold and on heart function may be explained by the narrow focus on dermatological effects of vitamin C, though we know the effects of vitamin C are not limited to the skin. Acknowledgments. The author is grateful to Elizabeth Chalker for critically reading the manuscript. 22 References to the Sheffield study are listed on page 6. Other references: 16. Number needed to harm. Wikipedia. https://en.wikipedia.org/wiki/Number_needed_to_harm 17. Mendes D, Alves C, Batel-Marques F. Number needed to treat (NNT) in clinical literature: an appraisal. BMC Med. 2017;15(1):112. https://pubmed.ncbi.nlm.nih.gov/28571585 18. Callaway B. Quantile Treatment Effects in R: The qte Package (2022-08-31). R project. https://cran.r-project.org/web/packages/qte/vignettes/R-QTEs.html (Accessed 2025-1-16) 19. Hemilä H, Chalker E, Tukiainen J. Quantile treatment effect of zinc lozenges on common cold duration: a novel approach to analyze the effect of treatment on illness duration. Front Pharmacol. 2022;13:817522. https://doi.org/10.3389/fphar.2022.817522 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8844493 20. Hemilä H, Chalker E, Tukiainen J. Response: Commentary: Quantile treatment effect of zinc lozenges on common cold duration: a novel approach to analyze the effect of treatment on illness duration. Front Pharmacol. 2024;15:1335784. https://doi.org/10.3389/fphar.2024.1335784 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11035776 21. Hemilä H, Pirinen M. Estimating quantile treatment effect on the original scale of the outcome variable: a case study of common cold treatments. arXiv 2023-10-27. https://doi.org/10.48550/arXiv.2310.17917 22. Pirinen M, Hemilä H. bqte: R code to estimate back-transformed quantile treatment [computer program, version 2023-6-26]. GitHub. https://github.com/mjpirinen/bqte 23. Hemilä H, Carr A, Chalker E. Vitamin C may increase the recovery rate of outpatient cases of SARS-CoV-2 infection by 70%: reanalysis of the COVID A to Z randomized clinical trial. Front Immunol. 2021;12:674681. https://doi.org/10.3389/fimmu.2021.674681 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc8141621 24. Bootstrapping (statistics). Wikipedia. https://en.wikipedia.org/wiki/Bootstrapping_(statistics) 25. Lind J. A Treatise on the Scurvy in Three Parts. A Treatise on the Scurvy, in Three Parts. Containing an Inquiry into the Nature, Causes, and Cure, of that Disease. Together with a Critical and Chronological View of what has been published on the Subject. The Second Edition. 1757. https://wellcomecollection.org/works/xc2t7736/items?canvas=161 [p.137: Krebs extract “Persons…”] https://wellcomecollection.org/works/xc2t7736/items?canvas=9 [p.v-vi: Hawkins] https://wellcomecollection.org/works/xc2t7736/items?canvas=459 [p.435: Anson] https://archive.org/details/treatiseonscurvy00lind/page/n164/mode/1up [p.137 Krebs extract] https://archive.org/details/treatiseonscurvy00lind/page/n10/mode/1up [p.v-vi: Hawkins] https://archive.org/details/treatiseonscurvy00lind/page/n462/mode/1up [p.435: Anson] See extracts of Lind’s texts in Ref. 27 below. 26. Drummond JC, Wilbraham A. William Stark, M.D.: an eighteenth century experiment in nutrition. Lancet. 1935;226:459–63. https://doi.org/10.1016/S0140-6736(00)94709-3 27. Hemilä H. The symptoms of vitamin C deficiency. 1. Observations and descriptions by Hess, Lind, Trotter and Blane. Zenodo. 2024. https://doi.org/10.5281/zenodo.10685194 28. Seidlitz. On haemorrhagic pericarditis. British and Foreign Medical Review. 1836;1(1):259–63. https://doi.org/10.5281/zenodo.11175282 https://pmc.ncbi.nlm.nih.gov/articles/PMC5556493 https://archive.org/details/s2571id1380017/page/258/mode/2up 23 29. Kyber A. On pericarditis scorbutica, and its treatment by paracentesis. Ranking's Abstract. 1848;7:64-5. https://doi.org/10.5281/zenodo.11175404 https://books.google.fi/books?id=nPJYAAAAMAAJ 30. Darling S. The pathologic affinities of beriberi and scurvy. JAMA. 1914;63(15):1290-4. https://doi.org/10.5281/zenodo.10102994 https://doi.org/10.1001/jama.1914.02570150046011 31. Hess AF. Cardiorespiratory involvement in infantile scurvy. Proc Soc Exp Biol Med. 1916;14(1):4-5. https://doi.org/10.3181/00379727-14-3 https://www.biodiversitylibrary.org/item/156376#page/14/mode/1up 32. Hess AF. Subacute and latent infantile scurvy: the cardiorespiratory syndrome (a new sign). JAMA. 1917;68(4):235-9. https://doi.org/10.5281/zenodo.10102042 https://doi.org/10.1001/jama.1917.04270010235001 33. Erdheim J. The heart in infantile scurvy [Über das Barlow-Herz]. [in German; translated to English]. Wiener Klinische Wochenschrift. 1918;31(49):1293-5. https://doi.org/10.5281/zenodo.7756493 34. Bierich R. Scurvy [Über Skorbut]. [in German; translated to English]. Deutsches Archiv Für Klinische Medizin. 1919;130(26 Sept):151-71. https://doi.org/10.5281/zenodo.10653559 https://archive.org/details/deutschesarchivfurklinischemedizin.v.129130.1919/page/n547/mode/2up 35. Comrie JD. Scurvy in North Russia. Edinb Med J. 1920;24(4):207-15. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5294660 36. Swarbreck A. Avitaminosis. Br Med J. 1932;2:123. https://doi.org/10.1136/bmj.2.3732.123 37. Barton WE, Freeman W. Pericardial hemorrhage complicating scurvy. N Engl J Med. 1934;210(10):529-31. https://doi.org/10.5281/zenodo.10103536 https://doi.org/10.1056/NEJM193403082101005 38. Platt R. Scurvy as the result of dietetic treatment. Lancet. 1936;228:366-7. https://doi.org/10.1016/S0140-6736(00)48407-2 39. Evans W. Vitamin C in heart failure. Lancet. 1938;231:308-9. https://doi.org/10.1016/S0140-6736(00)62412-1 40. Follis RH. Sudden death in infants with scurvy. J Pediatr. 1942;20(3):347-51. https://doi.org/10.1016/S0022-3476(42)80190-0 41. Shaffer CF. The diuretics effect of ascorbic acid: preliminary report on its use in cardiac decompensation. JAMA. 1944;124(11):700-1. https://doi.org/10.1001/jama.1944.02850110024006 42. Editorial. Ascorbic acid as a diuretic. Lancet. 1944;244:186. https://doi.org/10.1016/S0140-6736(00)42799-6 43. Hess AF. Scurvy: Past and Present. Lippincott; Philadelphia, PA, 1920. https://www.gutenberg.org/ebooks/40505 https://digital.library.cornell.edu/catalog/chla2903792 https://archive.org/search?query=creator%3A%22Hess+Alfred+Fabian%22 See biographies: https://doi.org/10.1053/j.spid.2005.01.003 https://doi.org/10.1007/bf01955048 https://doi.org/10.2105/ajph.65.9.977 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc1775957 https://doi.org/10.1093/jn/71.1.1 https://doi.org/10.1016/s0022-3476(55)80182-5 24 https://www.ncbi.nlm.nih.gov/pmc/articles/pmc1965595 https://doi.org/10.1001/archpedi.1934.01960100161022 https://doi.org/10.1126/science.79.2039.70 https://www.jstor.org/stable/1660244 https://en.wikipedia.org/wiki/Alfred_Fabian_Hess 44. Medical and surgical history of the British Army: which served in Turkey and the Crimea during the war against Russia in the years 1854-55-56 (Volume 2). London, UK, 1858. https://archive.org/details/62510370RX2.nlm.nih.gov/page/176/mode/1up https://babel.hathitrust.org/cgi/pt?id=nnc1.cu04838203&seq=198 45. Rozemeijer S, Hemilä H, van Baaren M, de Man AME. Vitamin C may reduce troponin and CKMB levels after PCI and CABG: a meta-analysis. BMC Cardiovasc Disord. 2023;23(1):475. https://doi.org/10.1186/s12872-023-03459-6 https://pmc.ncbi.nlm.nih.gov/articles/PMC10512653 46. Hemilä H, Suonsyrjä T. Vitamin C for preventing atrial fibrillation in high risk patients: a systematic review and meta-analysis. BMC Cardiovasc Disord. 2017;17(1):49. https://doi.org/10.1186/s12872-017-0478-5 https://pmc.ncbi.nlm.nih.gov/articles/PMC5286679 47. Hemilä H, Chalker E, de Man AME. Vitamin C may improve left ventricular ejection fraction: a meta-analysis. Front Cardiovasc Med. 2022;9:789729. https://doi.org/10.3389/fcvm.2022.789729 https://pmc.ncbi.nlm.nih.gov/articles/PMC8913583 48. Hemilä H, de Man AME. Vitamin C deficiency can lead to pulmonary hypertension: a systematic review of case reports. BMC Pulm Med. 2024;24(1):140. https://doi.org/10.1186/s12890-024-02941-x https://pmc.ncbi.nlm.nih.gov/articles/PMC10949735 49. Shafar J. Rapid reversion of electrocardiographic abnormalities after treatment in two cases of scurvy. Lancet. 1967;290:176-8. https://doi.org/10.1016/S0140-6736(67)90004-9 50. Singh D, Chan W. Cardiomegaly and generalized oedema due to vitamin C deficiency. Singapore Med J. 1974;15:60-3. https://smj.sma.org.sg/1501/1501smj11.pdf 51. Meisel JL, McDowell RK. Case 39-1995: a 72-year-old man with exertional dyspnea, fatigue, and extensive ecchymoses and purpuric lesions. N Engl J Med. 1995;333:1695-702. https://doi.org/10.1056/NEJM199512213332508 52. Kieffer P, Thannberger P, Wilhelm JM, Kieffer C, Schneider F. Multiple organ dysfunction dramatically improving with the infusion of vitamin C: more support for the persistence of scurvy in our welfare society. Intensive Care Med. 2001;27(2):448. https://doi.org/10.1007/s001340000830 53. Kupari M, Rapola J. Reversible pulmonary hypertension associated with vitamin C deficiency. Chest. 2012;142(1):225-7. https://doi.org/10.1378/chest.11-1857 54. Abbas F, Ha LD, Sterns R, von Doenhoff L. Reversible right heart failure in scurvy: rediscovery of an old observation. Circ Heart Fail. 2016;9(10):e003497. https://doi.org/10.1161/circheartfailure.116.003497 55. Bennett SE, Schmitt WP, Stanford FC, Baron JM. Case 22-2018: a 64-year-old man with progressive leg weakness, recurrent falls, and anemia. N Engl J Med. 2018;379:282-9. https:doi.org/10.1056/NEJMcpc1802826 56. Alnaimat S, Oseni A, Yang Y, Melvani V, Aronson A, Harris K, Panaich S. Missing vitamin C: a case of scorbutic cardiac tamponade. JACC Case Rep. 2019;1(2):192-6. https://doi.org/10.1016/j.jaccas.2019.07.006 57. Penn EH, Olenchock BA, Marston NA. A shocking deficiency. Circulation. 2019;140:613-7. 25 32 33 34 35 Appendix 2: Data set based on Appendix 1. This is the dataset used by Jowett. This is collected from data in Appendix 1. This is used in all analyses A to E. > SheffieldJowett ID Dose Days Days_1 Deprived Cured 1 Bartley 70 2 0 0 1 2 Bartley 70 2 0 0 1 3 Bartley 70 5 0 0 1 4 Garling 70 15 0 0 1 5 Garling 70 5 0 0 1 6 Garling 70 3 0 0 1 7 Hill 70 1 1 0 1 8 Hill 70 6 0 0 1 9 Hill 70 1 1 0 1 10 Golding 10 14 0 0 1 11 Golding 10 13 0 0 1 12 Golding 10 8 0 0 1 13 Golding 10 11 0 0 1 14 Golding 10 1 1 0 1 15 Golding 10 11 0 0 1 16 Jackson 10 1 1 0 1 17 Jackson 10 1 1 0 1 18 Woodhouse 10 3 0 0 1 19 Woodhouse 10 1 1 0 1 20 Woodhouse 10 1 1 0 1 21 Woodhouse 10 5 0 0 1 22 Drake 0 2 0 1 1 23 Drake 0 16 0 1 1 24 Hudson 0 16 0 1 1 25 Hudson 0 9 0 1 1 26 Hudson 0 11 0 1 1 27 Robinson 0 2 0 1 1 28 Robinson 0 3 0 1 1 29 Sanderson 0 24 0 1 1 30 Sanderson 0 17 0 1 1 31 Sanderson 0 3 0 1 1 32 Tridgell 0 8 0 1 1 33 Tridgell 0 4 0 1 1 34 Tridgell 0 14 0 1 1 35 Tridgell 0 15 0 1 1 36 Tridgell 0 3 0 1 1 37 Tridgell 0 9 0 1 1 38 Tridgell 0 8 0 1 1 39 Wodeman 0 3 0 1 1 table(SheffieldJowett$Deprived) 0 1 21 18 > table(SheffieldJowett$Deprived, SheffieldJowett$Days) 1 2 3 4 5 6 8 9 11 13 14 15 16 17 24 0 7 2 2 0 3 1 1 0 2 1 1 1 0 0 0 1 0 2 4 1 0 0 2 2 1 0 1 1 2 1 1 36 Appendix 3: Findings in the Glazebrook and Thomson (1942) trial [76] Variable Placebo Vitamin C Difference P a) Participants 1100 335 Vitamin C dose (mg/day) 10-15 50-300 Common colds 286 72 -17% 0.047 Colds treated in the Sick Quarters 253 59 -23% b) 0.017 Days in hospital 6.4 6.32 Tonsillitis c) 94 29 0% Admitted to hospital 83 18 -30% b) 0.086 Days in hospital, mean 16.7 10.1 -40% 0.0026 Days in hospital, SD 11.86 6.96 Days in hospital per infected d) 14.7 6.3 -57% 0.0021 d) Pneumonia 17 0 -100% 0.0053 Rheumatic fever 16 0 -100% 0.007 Days spent in the sick room due to ‘infective conditions’ per boy, mean 4.98 2.5 -50% <0.00001 e) The data are from [3.1]. a) P(1-tail) is shown. For the 2×2 tables, the mid-P is calculated. See statistical issues in Appendix 4. b) Difference and P-value calculated from the all participants population. c) Authors: “The term ‘tonsillitis’ is used here to be an index of haemolytic streptococcal disease of the nose and throat, and covers all such terms as ‘tonsillitis’, ‘sore throat’, ‘otitis media’, ‘pharyngitis’ and ‘cervical adenitis’, as nearly all these cases are of haemolytic streptococcal origin. Throat swabs were taken of large numbers of cases of tonsillitis to determine that the haemolytic streptococcus was the causative organism” [76, p.12]. d) Days in hospital per infected schoolboy. In the placebo group: 83×16.7/94 = 14.7 days. In the vitamin C group: 18×10.1/29 = 6.3 days. Combination of the tonsillitis P-values using Fisher’s method: https://en.wikipedia.org/wiki/Fisher%27s_method > (1pchisq(-2*(log(0.086) + log(0.0026)),4)) [1] 0.0021 e) The SD values were not reported for “days spent in the sick room”. In most vitamin C common cold trials the SD for the duration has been on average 70% of the mean duration of colds [3.2]. Here I use a conservative imputation of SD being 100% of the mean duration, see below the calculation. 37 3.1. Glazebrook AJ, Thomson S. The administration of vitamin C in a large institution and its effect on general health and resistance to infection. J Hyg (Lond). 1942;42:1-19. https://doi.org/10.1017/s0022172400012596 https://pmc.ncbi.nlm.nih.gov/articles/PMC2199803 Related papers: https://pmc.ncbi.nlm.nih.gov/articles/PMC2239225 [Description of the trial] https://pmc.ncbi.nlm.nih.gov/articles/PMC2162337 https://pmc.ncbi.nlm.nih.gov/articles/PMC2164503 https://pmc.ncbi.nlm.nih.gov/articles/PMC2164703 https://pmc.ncbi.nlm.nih.gov/articles/PMC2210931 https://pmc.ncbi.nlm.nih.gov/articles/PMC2286225 https://doi.org/10.5281/zenodo.14609923 Pauling’s analyses of the Glazebrook (1942) trial: https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-103.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/33/33-034.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/33/33-035.html 3.2. Hemilä H, Chalker E. Vitamin C for preventing and treating the common cold. Cochrane Database Syst Rev. 2013;2013:CD000980. https://pmc.ncbi.nlm.nih.gov/articles/PMC8078152 https://hdl.handle.net/10138/225864 https://www.researchgate.net/publication/273209193 https://doi.org/10.1002/14651858.cd000980.pub4 3.3. Hemilä H, Louhiala P. Vitamin C for preventing and treating pneumonia. Cochrane Database Syst Rev. 2013;(8):CD005532. https://hdl.handle.net/10138/225862 https://www.researchgate.net/publication/6549713 https://doi.org/10.1002/14651858.cd005532.pub3 Calculating the P-value for “Days in hospital due to tonsillitis” and “Days spent in the sick room”, using the Ratio of Means approach, see Appendix 4. Study Ne Me Se Nc Mc Sc ROM LnROM SDe SDc SElnROM z p 1 Friedrich (test) 9 213.0 67.0 10 177 40 1.2 0.19 0.011 0.00511 0.127 1.5 9.3e-01 2 Glazebrook_SickTonsillitis 18 10.1 7.0 83 17 12 0.6 -0.50 0.026 0.00608 0.180 -2.8 2.6e-03 3 Glazebrook_SickRoom 335 2.5 2.5 1100 5 5 0.5 -0.69 0.003 0.00091 0.062 -11.0 1.2e-28 > riskratio.small(GTcolds) $data Outcome Predictor Disease1 Disease2 Total Exposed1 814 286 1100 Exposed2 263 72 335 Total 1077 358 1435 $p.value two-sided Predictor midp.exact fisher.exact chi.square Exposed2 0.093467 0.097693 0.095063 > (P_1tail=0.09347/2) [1] 0.046735 > riskratio.small(GTColdsSick) $data Outcome Predictor Disease1 Disease2 Total Exposed1 847 253 1100 Exposed2 276 59 335 Total 1123 312 1435 $measure risk ratio with 95% C.I. Predictor estimate lower upper Exposed2 0.76342 0.59123 0.98575 $p.value two-sided Predictor midp.exact fisher.exact chi.square Exposed2 0.03436 0.040917 0.036332 > (P_1tail=0.03436/2) [1] 0.01718 38 > riskratio.small(GTpneumonia) $data Outcome Predictor Disease1 Disease2 Total Exposed1 1083 17 1100 Exposed2 335 0 335 Total 1418 17 1435 $p.value two-sided Predictor midp.exact fisher.exact chi.square Exposed2 0.010581 0.017995 0.022082 > (P_1tail=0.01058/2) [1] 0.00529 > riskratio.small(GTrheumfever) $data Outcome Predictor Disease1 Disease2 Total Exposed1 1084 16 1100 Exposed2 335 0 335 Total 1419 16 1435 $p.value two-sided Predictor midp.exact fisher.exact chi.square Exposed2 0.013852 0.031415 0.02643 > (P_1tail=0.01385/2) [1] 0.006925 In the Glazebrook and Thomson trial [3.1], allocation to treatment groups was carried out by institute ‘divisions’ and not on the basis of individual boys. Therefore, in the Cochrane review (2013) the trial was also analysed using the ’division’ as the unit of observation [3.3]. Distribution of pneumonia cases in the five control divisions was 5, 3, 2, 4 and 3 (mean 3.40 cases per division) and in the two vitamin C divisions it was 0 and 0. I assume that the mean of the control divisions was a suitable estimate for the Poisson distribution mean and I used that assumption as a basis for the statistical analysis. The size of the individual divisions was not stated in the paper but the two vitamin C divisions had on average 167 boys (335/2) and the five control divisions 220 boys (1100/5), thus the size of the vitamin C divisions was 0.761 times the size of the control divisions. I adjusted the mean incidence by this ratio, so that I expected 2.59 pneumonia cases per vitamin C division, assuming the same average incidence as for the control divisions. With this Poisson mean, I calculated the probability that there were no cases of pneumonia in two separate vitamin C divisions as having a P-value of 0.0056. Accordingly, using a ’division’ as the unit of observation does not change the conclusions. > (Control <- mean(c(5,3,2,4,3))) [1] 3.4 > (RatioBoys <- (335/2)/(1100/5)) [1] 0.761 > (lambda <- Control*RatioBoys) [1] 2.59 > dpois(0, lambda)^2 [1] 0.0056 39 Methods of the Glazebrook and Thomson (1942) trial are described in the report [3.1]. The structure of the paper is quite different from modern trial reports. Extracted here are the main descriptions: In a large training school under our observation there were some 1500 youths aged 15-20 years. For the most part they were drawn from the lower wage-earning classes, and a large proportion came from Scotland and the North Midlands, where economic conditions are probably below the average for the country. It is a reasonable assumption that the previous dietary of the recruits had been somewhat deficient in vitamin C judged by the standards already quoted (p.4). The food distribution [at the school] was badly managed. Electric ovens were used to reheat the food, and to keep it hot whilst awaiting distribution. Often 8 hr. elapsed between the time the food was cooked and its arrival on the dining tables. The minimum time that heat was applied to the food, including the original cooking and the subsequent reheating, was 2 hr (p.4). The total intake of vitamin C varied from about 10 to 15 mg per student per day (p.5). The calcium and vitamin B content of the dietary of the institution could perhaps be criticized, but the only outstanding deficiency, according to modern standards, was in vitamin C. As far as this one factor was concerned, the boys were almost certainly worse off, subsisting on the institution diet, than they would have been at home (p.17). Pure ascorbic acid powder was added to… the morning cocoa, and an evening glass of milk. The mixing was done in bulk in the kitchens before issue. The powder dissolved quickly and easily, and did not alter the appearance or taste of the vehicle (p.7). Thus, the trial corresponds functionally to a placebo-controlled trial because the participants were unable to identify the treatment, although no inactive powder was added to the cocoa and milk of the control group. The establishment was divided into seven groups or divisions for administrative purposes. The youths of one division worked as a unit, and occupied certain tables in the dining hall. To some extent each division occupied particular dormitories but this separation was not absolute, and there was a fair amount of mixing of divisions in the sleeping quarters. Sleeping and feeding conditions were, of course, the same for all divisions (p.12). Careful records had been kept of the incidence of all infections for 1½ years before the observations described here were begun. In the preceding year there had been an epidemic of tonsillitis, which had affected all the divisions uniformly, so that they could not be regarded as separate units within the larger populations (p.12). The above suggests that exposure to infectious agents was reasonably uniform over the study population. The observations were made by supplying vitamin C in the form of pure ascorbic acid to one or more divisions. This was considered to be the only practical method of carrying out the observations without introducing unnecessary complications. For example, it was not possible to choose boys at random as it would have been impossible to supply them with vitamin C-treated cocoa or milk in the dining room. With the method actually chosen, all that was necessary was to add vitamin C to the supplies of cocoa or milk serving the tables for the appropriate divisions (p.12). Moreover, all of the divisions had a population more or less the same as regards duration of stay in the establishment (‘institution age’). Infectious diseases were more common amongst those who had more recently joined the institution (p.12). When a youth fell ill he was admitted to Sick Quarters unless his complaint was very mild. In the latter case he was placed on the out-patients list and excused all duties except attendance at school instruction. Most of the cases of common cold and tonsillitis were admitted to Sick Quarters. In analysing the durations of illnesses, observations were restricted to the cases in the Sick Quarters. The 40 number of days spent there was obviously a more reliable index of the duration of illness, since the patient was under constant medical supervision. Frequently when a youth was discharged from the Sick Quarters he was put on the out-patients list, and this ‘convalescent period’ was neglected. The admission to and discharge from the hospital was not under our control (p.13). The diet in the Sick Quarters was basically similar to that of the healthy boys. It was modified, of course, to suit the needs of the sick, but was prepared in the central kitchens and suffered an equally drastic loss of its vitamin C. When a student from the experimental division fell ill and was admitted to Sick Quarters, his dosage of ascorbic acid was continued there (p.14). The period of treatment of cases of tonsillitis and common cold in the Sick Quarters was completely outside our control, and no biased attitudes influenced these durations from which we have drawn our conclusions (p.16). [pneumonia… rheumatic fever] … These cases were-subjected to special investigations by us (X-rays, etc.) to establish certain criteria for the diagnosis. There was, however, in our opinion a relationship between these conditions (p.16). It was not stated whether the diagnosis of pneumonia was carried out by the trial authors of the paper or by the physicians at the Sick Quarters. Although the method of diagnosing pneumonia was not described in detail in the paper, with the given descriptions and the severe pathological processes occurring in pneumonia it seems unlikely that administration of vitamin C in cocoa and milk could have led to detection bias. 41 Appendix 8. QTE analysis Fig. 2 was calculated with the bqte program [22]. A. Using the vitamin C administration (10-70 mg/day) as the control on the x-axis. Treatment = subset(SheffieldJowett, Deprived == 1)$Days > table(Treatment) Treatment 2 3 4 8 9 11 14 15 16 17 24 2 4 1 2 2 1 1 1 2 1 1 > Control = subset(SheffieldJowett, Deprived == 0)$Days > table(Control) Control 1 2 3 5 6 8 11 13 14 15 7 2 2 3 1 1 2 1 1 1 > res.bqte = bqte(Treatment = Treatment, Control = Control, tails = TRUE, at = c(1,2,3,4,5,6,7,8)) #Estimate BQTE Running bqte() with the following parameters. length(Treatment): 18 length(Control): 21 K: 21 B: 2000 bqte.conf: 0.95 at: 1, 2, 3, 4, 5, 6, 7, 8 bagging: TRUE tails: TRUE If you choose to change 'at' values, a recommended interval is [1, 8.71]. > > #Check output of bqte() function > > res.bqte at bqte bqte.low bqte.up 1 1 2.23 1.000 5.42 2 2 3.96 0.818 10.47 at = 1 gives the estimate and 95% CI (low, up) for the effect on the control group 1-day colds. at = 2 gives the estimate and 95% CI (low, up) for the effect on the control group 2-day colds. Thus, according to this estimation, an 1-day cold in participants administered vitamin C (10-70 mg/day) becomes 2.23 days longer when a participants is deprived of vitamin C, leading to a colds of 3.23 days. 48 B. Using the vitamin C deprivation as the control on the x-axis. > Treatment = subset(SheffieldJowett, Deprived == 0)$Days > table(Treatment) Treatment 1 2 3 5 6 8 11 13 14 15 7 2 2 3 1 1 2 1 1 1 > Control = subset(SheffieldJowett, Deprived == 1)$Days > table(Control) Control 2 3 4 8 9 11 14 15 16 17 24 2 4 1 2 2 1 1 1 2 1 1 > set.seed(1) #Set random seed for reproducible results > res.bqte = bqte(Treatment = Treatment, Control = Control, tails = TRUE) #Estimate BQTE Running bqte() with the following parameters. length(Treatment): 21 length(Control): 18 K: 18 B: 2000 bqte.conf: 0.95 at: 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 bagging: TRUE tails: TRUE If you choose to change 'at' values, a recommended interval is [3, 14.3]. > > #Check output of bqte() function > > res.bqte at bqte bqte.low bqte.up 1 3 -1.68 -2.00 0.2461 2 4 -1.92 -3.00 1.2656 49 Appendix 9. Vitamin C deficiency and pneumonia in the Alfred Hess monograph on scurvy [43] Alfred Hess – a pediatrician in the USA – was the most important clinical investigator of vitamin C deficiency in the early 20th century. He wrote a monograph about scurvy, in which he stated in several parts that there is an increased risk of pneumonia in scorbutic animals and humans. These statements are extracted below with their contexts; bold added. Hess AF. Scurvy: Past and Present. Philadelphia PA: Lippincott USA; 1920. https://www.gutenberg.org/ebooks/40505 https://digital.library.cornell.edu/catalog/chla2903792 https://archive.org/search?query=creator%3A%22Hess+Alfred+Fabian%22 More relevant parts of Hess’ book are extracted in: Hemilä H. The symptoms of vitamin C deficiency. 1. Observations and descriptions by Hess, Lind, Trotter and Blane. Zenodo. 2024. https://doi.org/10.5281/zenodo.10685194 p.88 [Chapter IV, Pathology of scurvy] Lungs.—… Edema of the lungs is not uncommon, as we should expect, especially as a terminal condition. Pneumonia, lobular or lobar, is one of the most frequent complications and causes of death. Active tuberculosis is a not uncommon secondary manifestation. p.99 [Chapter IV, Microsopic pathology of scurvy] Lungs.—Hemorrhages of various size occur in the tissue of the lung or in the air spaces. Hemorrhagic infarcts also have been described, and Sato and Nambu report hyaline degeneration of the blood-vessel walls. Secondary pneumonias, usually broncho-pneumonic in type, are of common occurrence, and in many epidemics constitute the prevailing cause of death. Tuberculous lesions are also frequently present, and are stated to assume fresh activity as the result of the nutritional disorder. Edema occurs frequently, the fluid in the acini often containing red blood-cells. Subpleural hemorrhages, thickening of the pleura, purulent or fibrinous pleurisy are common lesions. p.123 [Chapter V, Experimental scurvy; Guinea pigs] On opening the chest, slight hemorrhages may be noted in the pericardium and in the visceral and costal pleuræ. The heart is frequently enlarged, and the pericardial sac contains an excess of serum; the right ventricle, however, is not found disproportionately hypertrophied. Pneumonia is met with very frequently and constitutes a common terminal infection. p.179 [Chapter VII, Symptomatology and diagnosis of scurvy] Nowadays, the disease usually does not reach this stage, and rarely progresses further. If, however, the patient remains untreated, he becomes progressively weaker and more lethargic; there is frequent palpitation, shortness of breath, and increasing loss of weight. The pains in the limbs render him helpless and an object of pity. Marked edema may be added to the picture as the result of starvation, so that the legs become swollen, and even the face becomes bloated. Hemorrhages into the skin as large as the palm of the hand appear on different parts of the body. The gums swell to such an extent that they overlap and may even hide the teeth and protrude from the mouth as foul fungoid growth. Death comes about in various ways. Frequently sudden and fatal syncope occurs, due to heart weakness or to the pouring out of fluid into the pleural or the pericardial cavities. Another frequent cause of death is secondary infection, resulting in pneumonia, which finally ends the suffering of the patient. The fatal outcome is thus described in the narrative of Lord Anson’s voyage… 50 p.182 [Chapter VII, Symptomatology and diagnosis of scurvy] Pericarditis, hydrothorax, pleurisy with effusion, pneumonia, are common complications of severe forms of scurvy. Lind reports that the dominant complication varies in different epidemics; that on one cruise many cases of diarrhœa would occur and on another many pulmonary infections. [see below Lind’s section to which Hess refers] p.202 [Chapter VII, Symptomatology and diagnosis of scurvy] The main involvement of the respiratory system in scurvy is the polypnœa just described in connection with the cardiorespiratory syndrome. There is no aphonia, a sign so typical of adult and of infantile beriberi, although at times the voice is abnormal and whining. The lungs frequently show some dullness posteriorly, which may be due to engorgement or to the pressure of the enlarged heart. Pneumonia is a frequent complication and edema a terminal event. Hydrothorax associated with hydropericardium is of frequent occurrence, and was noted in the early description of this disease in adults and in the first account of Barlow. These effusions rarely progress to what may be termed the clinical degree and under antiscorbutic treatment are rapidly absorbed. p.217 [Chapter VII, Symptomatology and diagnosis of scurvy] We have already considered numerous complications of scurvy, and shall therefore not go over this ground again. Many of them are due to hemorrhages or to serous effusions in various parts of the body. Another large group in adults as well as in infants are the result of infection. The respiratory tract is particularly susceptible, pneumonia constituting the most common cause of death. In infants we meet with frequent attacks of “grippe,” widespread occurrence of nasal diphtheria, furunculosis and torpid ulcers of the skin, pyelitis, otitis, adenitis, etc. We have encountered nasal diphtheria—with typical bloody mucous discharge—so frequently in connection with scurvy, that where this local infection occurs among a group of infants they should be carefully examined for latent or mild scurvy. Aschoff and Koch recently have laid emphasis on the frequency with which diphtheria complicated scurvy among adults (soldiers). Dysentery is another complication resulting from an invasion of bacteria. Local infections occur more often in adults than in infants—cervical adenitis following gingival pyorrhœa, “bubo” of the groin following infection of the lower extremity, abscess of the calf of the leg following hemorrhage into this region. p.227 [Chapter VIII, Prognosis of scurvy] In adults the heart may be weakened by scurvy, and death may result from cardiac failure. Cardiac disturbances occur also in infantile scurvy. This involvement might be expected, in view of the tachycardia (cardiorespiratory phenomenon) which is so frequent a symptom of infantile scurvy. The heart may be rapid for months or even for years after the disorder, and tachycardia may develop on the occasion of even a mild infectious disease. For example, a fever of 101°, due to a common coryza, may cause the heart-beat to rise to perhaps 180 a minute. Children so affected succumb readily to infection, especially to pneumonia, which may lead to sudden collapse followed by death. An important factor in the prognosis of scurvy, as in that of other disorders due to a lack of vitamines, is the marked susceptibility to infection. Even latent or subacute scurvy causes a peculiar susceptibility to diphtheria (especially the nasal type), to coryza, bronchitis, and pneumonia. A perusal of the literature shows that this susceptibility was noted by the older authors in relation to adults. 51 Later comments on vitamin C and infections by Hess in 1932: Hess AF. Recent advances in knowledge of scurvy and the antiscorbutic vitamin. JAMA. 1932;98(17):1429-33. https://doi.org/10.5281/zenodo.14609694 https://doi.org/10.1001/jama.1932.02730430005002 ASSOCIATION OF SCURVY WITH INFECTIONS Another point that has been brought into prominence in connection with adult as well as infantile scurvy is its intimate association with the infectious process. In 1917 I stated that “one of the striking and important symptoms of scurvy is a susceptibility to infection (furunculosis, nasal diphtheria, grippe, etc.).” In this connection it may be mentioned that nasal diphtheria was noted among a group of scorbutic infants in spite of the fact that many of them gave a negative Schick reaction. Findlay, attacking the problem from an experimental standpoint, showed that guinea-pigs which suffered from chronic scurvy and showed but few clinical symptoms manifested a decreased resistance to bacterial infection; this he attributed to degenerative changes in the bone marrow. About this time Cramer and Kingsbury emphasized the importance of vitamin A in warding off infections, associating this susceptibility with a decrease in the number of platelets of the blood. Abels, in a monograph on scurvy, stressed this relationship of the scorbutic state to infection, giving the name “dysergie” to the nutritional disturbance which occasions the heightened susceptibility. Recently, Minot and his colleagues came to the conclusion that adult scurvy can be precipitated by infectious processes; in other words, that latent scurvy can by this means be changed to manifest scurvy. In general, therefore, investigations in the laboratory as well as clinical observations are in agreement in stressing the interrelationship of scurvy and bacterial infection. The 1917 paper: Hess AF. Infantile scurvy. Part V. A study of its pathogenesis. Am J Dis Child. 1917;14:337-53. https://doi.org/10.5281/zenodo.14609522 https://doi.org/10.1001/archpedi.1917.01910110018002 52 Hess AF. Diet, nutrition and infection. N Engl J Med. 1932;207:637-48. https://doi.org/10.5281/zenodo.14609616 https://doi.org/10.1056/NEJM193210132071501 In regard to vitamin C, I shall refer only to the bearing of this vitamin on infections, more particularly of the respiratory tract. In 1917, in a paper on the pathogenesis of infantile scurvy, I emphasized the fact that a lack of the antiscorbutic factor which leads to scurvy, at the same time predisposes to infections. This enhanced susceptibility has been confirmed by Abels, Ludwig Meyer and many others. It exists even before the scorbutic signs are manifest in the stage which is better termed “latent scurvy” than “Praeskorbutus” as the abnormal scorbutic state already exists. Similar susceptibility to infections goes hand in hand with adult scurvy. This was pointed out years ago by Lind and others in connection with scurvy in the mercantile marine and among the soldiers in times of war, for example, in our Civil War and in the Crimean War. But I wish to emphasize quite another aspect of infection in connection with infantile scurvy. In 1917, and again in 1920, I called attention to the “widespread occurrence of nasal diphtheria in infantile scurvy”, remarking that “we have encountered nasal diphtheria—with typical bloody mucus discharge—so frequently in connection with scurvy that where this local infection occurs among a group of infants they should be carefully examined for latent or mild scurvy”. At the same time I drew attention to the fact that “clinical tests showed that the blood contains sufficient antitoxin (diphtheria) to afford protection.” These were the days previous to the use of toxin-antitoxin. Much to our surprise, some of these cases gave a negative Schick test in spite of the definite clinical signs of nasal diphtheria. In two instances the diphtheria bacilli were tested on guinea pigs and found to be virulent. Not long after these observations an infant died from diphtheria of the larynx which developed although the Schick test was negative. At postmortem a typical membrane was found on the larynx. Since this time, there has been little opportunity to investigate this subject, as diphtheria has been banished from our institution by the routine use of toxin-antitoxin. Recently, however, we have met with three cases of nasal diphtheria which developed soon after their admission to the institution. These cases were characterized by the typical bloody nasal discharge. In two instances the cultures showed avirulent bacilli, in one which was obtained in December, 1930, the bacilli from the nose were virulent. In spite of this fact, not only was the Schick test negative, but tests carried out in February, 1931, with increasing doses of toxin 1/50-1/40-1/301/20-1/10-1/5 M.L.D. all failed to induce a skin reaction. As the result of these experiences we infer that a lack of the antiscorbutic vitamin exerts a local effect on the mucous membrane which diminishes its immunity, but at the same time may not be accompanied by a lowering of systemic immunity. It is probable that a similar phenomenon holds true in connection with a deficiency of vitamin A and that the marked changes in the epithelium, described by Wolbach, bring about a local diminution in resistance. Susceptibility to infections of the skin and of the respiratory tract which occur when this deficiency is marked may be largely a manifestation of a local pathological change. 53 James Lind and respiratory infections in vitamin C deficiency James Lind’s treatise on scurvy is the most important text on the topic over the ages. However, at this time there was no knowledge of bacteria and X-rays were not available, and for reasons such as these there was no concept of “pneumonia” in the same sense as was known at Hess’ times and currently. See links to several versions of Lind’s treatise through: Hemilä H. The symptoms of vitamin C deficiency. 1. Observations and descriptions by Hess, Lind, Trotter and Blane. Zenodo. 2024. https://doi.org/10.5281/zenodo.10685194 Also: https://archive.org/search?query=creator%3A%22Lind+James%22+scurvy In our current context, the section below to which Hess referred may indicate that in 1747 there were respiratory viruses circulating on the ship, and because of the very low vitamin C levels the sailors did not properly recover from the viral infections and some of them may have suffered from lower respiratory tract infections as a complication. On the other hand, in 1746 there may have been viral gastroenteritis circulating so that scurvy exacerbated their symptoms. This is the section to which Hess referred on p. 182, see above. Lind’s text; bold added. p.110-111 in Lind’s 1st edition (Lind’s 3rd edition p.105) I observed a considerable difference in the genius of the disease in the two cruises ann. 1746 and 1747. In the latter, when fevers from cold of the pleuritic and peripneumonic sort prevailed, it tended chiefly to affect the breast with a tightness, oppression, and a hard bound cough, by which a very viscid phlegm was with great difficulty brought up. The fits of coughing were not constant, but extremely fatiguing ; and this was a universal complaint. Several at this season were feverish ; we had none in a salivation, and the fluxes were mild and manageable. Whereas in the year 1746, when a different species of diseases prevailed, occasioned by the unwholesome newness of the ship’s timbers, and diarrhœas were frequent, the scurvy proved more virulent and fatal. Its worst, most common, and troublesome symptoms, were salivations and dysenteries, especially the latter ; in which one Nichols died, and eight or ten more were landed at Plymouth in a very low and exhausted condition by it. I did not at that time remark any of them to be feverish, and their breasts were but slightly affected. 54 Appendix 10. Findings in the Cowan et al. (1942) trial [85] Variable Placebo Vitamin C Difference P Participants 155 208 Vitamin C dose (mg/day) ? 100-200 Common colds per person, mean a) 1.9 2.2 -14% 0.0032 SD a) 1.00 1.01 SE 0.08 0.07 Percentage who had no colds 8.4% 11.5% 3.1 pp Days lost from school 1.6 1.1 -31% 0.0002 SD b) 1.6 1.1 pp, percentage point The data are from [9.1]. a) The authors reported the SE value, from which the SD is calculated. The authors also reported difference between groups as “0.3 (SE 0.11)”, which gives > (z <- 0.3/0.11) [1] 2.7 > (p = pnorm(-z)) [1] 0.0032 b) The authors did not report the SD value for “days lost from school”. The SD values were not reported. In most vitamin C common cold trials the SD for the duration has been on average 70% of the mean duration of colds [9.2]. Here I use a conservative imputation of SD being 100% of the mean duration. Calculation with the RoM approach gives: Study Ne Me Se Nc Mc Sc ROM LnROM SDe SDc 1 Friedrich (test) 9 213.00 67.00 10 177.00 40.00 1.203 0.185 0.01099 0.005107 5 Cowan 208 1.10 1.10 155 1.60 1.60 0.688 -0.375 0.00481 0.006452 SElnROM z p 1 0.1269 1.46 9.28e-01 5 0.1061 -3.53 2.07e-04 Reduction in “days lost from school” was also reported in 1977 by Ludvigsson [9.3,9.4] Cowan et al. wrote (1942) [9.1, p.1269]: … The actual difference between the two groups [in the incidence of colds] during the year of the study amounts to one third of a cold per person. Statistical analysis of the data reveals that a difference as large as this would arise only three or four times in a hundred through chance alone. One may therefore consider this as probably a significant difference, and vitamin C supplement to the diet may therefore be judged to give a slight advantage in reducing the number of colds experienced. 55 9.1. Cowan DW, Diehl HS, Baker AB. Vitamins for the prevention of colds. JAMA. 1942;120(16):1268-71. https://doi.org/10.5281/zenodo.14674396 https://doi.org/10.1001/jama.1942.02830510006002 Pauling’s analyses of the Cowan et al. (1942) trial: https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-096.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-097.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-098.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-099.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-100.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/31/31-101.html https://scarc.library.oregonstate.edu/coll/pauling/rnb/33/33-029.html The Cowan (1942) trial was included in Pauling’s (1971) meta-analyses: Pauling L. The significance of the evidence about ascorbic acid and the common cold. Proc Natl Acad Sci USA. 1971; 68: 2678-2681. https://doi.org/10.1073/pnas.68.11.2678 https://pmc.ncbi.nlm.nih.gov/articles/PMC389499 https://profiles.nlm.nih.gov/101584639X101 9.2. Hemilä H, Chalker E. Vitamin C for preventing and treating the common cold. Cochrane Database Syst Rev. 2013;2013:CD000980. https://pmc.ncbi.nlm.nih.gov/articles/PMC8078152 https://hdl.handle.net/10138/225864 https://doi.org/10.1002/14651858.cd000980.pub4 https://www.mv.helsinki.fi/home/hemila/CC [Links to references] 9.3. Ludvigsson J, Hansson LO, Tibbling G. Vitamin C as a preventive medicine against common colds in children. Scand J Infect Dis. 1977;9(2):91-8. https://www.researchgate.net/publication/22256528 https://doi.org/10.3109/inf.1977.9.issue-2.07 9.4. Hemilä H, Chalker E. Vitamin C reduces the severity of common colds: a meta-analysis. BMC Public Health. 2023;23(1):2468. https://doi.org/10.1186/s12889-023-17229-8 https://pmc.ncbi.nlm.nih.gov/articles/PMC10712193 56 Appendix 11. Benefit from vitamin C for common colds in UK trials after the Sheffield study Trial Outcome Effect P Charleston and Clegg 1972 [10.1,10.2] Incidence of colds in males -68% 0.00004 Elwood et al. 1976 [10.3,10.6] Incidence of ‘chest colds’ in females -18% 0.014 Tyrrell et al. 1977 [10.4,10.7] Incidence of recurrent colds in males a) -40% 0.018 Baird et al. 1979 [10.5,10.7] Incidence of colds in males -37% 0.00003 a) Tyrrell et al. (1977) administered “10 g of ascorbic acid taken during the first 2½ days on the symptoms of the common cold”. The outcome here is the incidence of second colds for the participants who were treated with vitamin C or placebo for the first cold. These trials were included in a 1997 systematic review of UK trials [10.8]. The meta-analysis was commented on, and comments were responded to [10.9]. 10.1. Charleston SS, Clegg KM. Ascorbic acid and the common cold. Lancet. 1972;1(7765):1401-2. https://doi.org/10.1016/s0140-6736(72)91143-9 https://pubmed.ncbi.nlm.nih.gov/4113614 10.2. Clegg KM. Studies associated with ascorbic acid. Acta Vitaminol Enzymol. 1974;28:101-2. https://doi.org/10.5281/zenodo.13946103 https://pubmed.ncbi.nlm.nih.gov/4615590 10.3. Elwood PC, Lee HP, St Leger AS, Baird M, Howard AN. A randomized controlled trial of vitamin C in the prevention and amelioration of the common cold. Br J Prev Soc Med. 1976;30:193-6. https://pmc.ncbi.nlm.nih.gov/articles/PMC478963 https://doi.org/10.1136/jech.30.3.193 https://www.jstor.org/stable/25565911 10.4. Tyrrell DA, Craig JW, Meada TW, White T. A trial of ascorbic acid in the treatment of the common cold. Br J Prev Soc Med. 1977;31:189-91. https://pmc.ncbi.nlm.nih.gov/articles/PMC479021 https://www.jstor.org/stable/25565969 https://doi.org/10.1136/jech.31.3.189 10.5. Baird IM, Hughes RE, Wilson HK, Davies JE, Howard AN. The effects of ascorbic acid and flavonoids on the occurrence of symptoms normally associated with the common cold. Am J Clin Nutr. 1979;32:1686-90. https://doi.org/10.1093/ajcn/32.8.1686 https://pubmed.ncbi.nlm.nih.gov/463806 10.6. Hemilä H, Chalker E. Vitamin C for the common cold and pneumonia. Pol Arch Intern Med. 2025 Jan 13:16926. https://doi.org/10.20452/pamw.16926 10.7. Hemilä H. Vitamin C and sex differences in respiratory tract infections. Respir Med. 2008; 102: 625-626. https://doi.org/10.1016/j.rmed.2007.12.011 https://hdl.handle.net/10138/228090 https://www.researchgate.net/publication/5628489 10.8. Hemilä H. Vitamin C intake and susceptibility to the common cold. Br J Nutr. 1997;77:59-72. https://doi.org/10.1017/s0007114500002889 https://hdl.handle.net/10138/223362 https://www.researchgate.net/publication/14153016 10.9. Bates CJ, Schorah CJ, Hemilä H. Vitamin C intake and susceptibility to the common cold: Invited commentaries and response. Br J Nutr. 1997;78:857-66. https://doi.org/10.1079/BJN19970201 https://hdl.handle.net/10138/223364 https://www.researchgate.net/publication/28369859 57