https://c19early.org/dmeta.html
•Statistically significant improvements are seen in treatment studies for
mortality,
ICU admission,
hospitalization, and
cases.
56 studies from 52 independent teams in 20 different countries show statistically significant
improvements in isolation (40 for the most serious outcome).
•Random effects meta-analysis with pooled effects using the most serious
outcome reported shows 60% [40‑74%] and 37% [31‑42%] improvement for early treatment
and for all studies. Results are similar after restriction to
101 peer-reviewed
studies: 57% [36‑71%] and 37% [31‑42%], and for the 60 mortality results: 68% [39‑84%] and
36% [28‑44%].
•Late stage treatment with calcifediol/calcitriol shows greater
improvement compared to cholecalciferol:
73% [57‑83%] vs.
41% [27‑52%].
•Sufficiency studies show a strong association between
vitamin D sufficiency and outcomes. Meta analysis of the
157 studies using the most serious outcome reported
shows 53% [49‑57%] improvement.
•No treatment, vaccine, or
intervention is 100% effective and available. All practical, effective, and
safe means should be used based on risk/benefit analysis.
Multiple treatments are typically used
in combination, and other treatments
may be more effective.
Only 13% of vitamin D
studies show zero events with treatment.
The quality of non-prescription supplements can vary widely [Crawford, Crighton].
•All data and sources to reproduce this paper are in the appendix.
Other meta analyses for vitamin D treatment can be found in
[Argano, D’Ecclesiis, Hosseini, Nikniaz, Shah, Tentolouris, Varikasuvu, Xie], showing significant improvements for mortality, mechanical ventilation, ICU admission, hospitalization, severity, and cases.
| Early treatment | Late treatment | All studies | Studies | Patients | Authors | |
|---|---|---|---|---|---|---|
| All studies | 60% [40‑74%] **** | 46% [33‑56%] **** | 37% [31‑42%] **** |
107 | 182,747 | 1,057 |
| Randomized Controlled TrialsRCTs | 32% [8‑50%] * | 34% [13‑50%] ** | 31% [17‑42%] **** |
26 | 41,836 | 299 |
| Calcifediol/calcitriolCalcifediol | - | 73% [57‑83%] **** | 49% [26‑65%] *** |
11 | 8,891 | 135 |
| Mortality | 68% [39‑84%] *** | 45% [31‑57%] **** | 36% [28‑44%] **** |
60 | 62,080 | 561 |
| HospitalizationHosp. | 90% [-453‑100%] | 22% [6‑35%] ** | 20% [8‑31%] ** |
21 | 85,626 | 200 |
| RCT mortality | - | 31% [6‑50%] * | 31% [6‑50%] * |
15 | 1,949 | 174 |
Highlights
Vitamin D reduces
risk for COVID-19 with very high confidence for mortality, ICU admission, hospitalization, recovery, cases, viral clearance, and in pooled analysis, low confidence for ventilation, and very low confidence for progression.
We show traditional outcome specific analyses and combined
evidence from all studies, incorporating treatment delay, a primary
confounding factor in COVID-19 studies.
Real-time updates and corrections,
transparent analysis with all results in the same format, consistent protocol
for 50
treatments.
B
Loading..
C
Loading..
Figure 1.
A. Random effects
meta-analysis of treatment studies. This plot shows pooled effects, analysis for individual outcomes is below, and
more details on pooled effects can be found in the heterogeneity section.
Effect extraction is pre-specified, using the most serious outcome reported.
Simplified dosages are shown for comparison, these are the total dose in the
first five days for treatment, and the monthly dose for prophylaxis.
Calcifediol, calcitriol, and paricalcitol treatment are indicated with (c), (t), and (p).
For details of effect extraction and full dosage information see the appendix.
B. Scatter plot
showing the distribution of effects reported in sufficiency studies and treatment studies. Diamonds show the
results of random effects meta-analysis.
C. Scatter plot showing the most serious outcome in all studies in the
context of multiple COVID-19 treatments. Diamonds show the
results of random effects meta-analysis for each treatment.
D. Timeline of results in vitamin D treatment studies. The marked dates indicate the time when efficacy was known with a statistically significant improvement of ≥10% from ≥3 studies for pooled outcomes, one or more specific outcome, pooled outcomes in RCTs, and one or more specific outcome in RCTs. Efficacy based on RCTs only was delayed by 10.8 months, compared to using all studies. Efficacy based on specific outcomes in RCTs was delayed by 2.2 months, compared to using pooled outcomes in RCTs.
We analyze all significant controlled studies regarding vitamin
D and COVID-19. Search methods, inclusion criteria, effect extraction criteria
(more serious outcomes have priority), all individual study data, PRISMA
answers, and statistical methods are detailed in Appendix 1.
We perform random-effects meta analysis for all treatment studies, Randomized
Controlled Trials, peer-reviewed studies, studies using cholecalciferol,
studies using calcifediol/calcitriol, and for specific outcomes: mortality,
mechanical ventilation, ICU admission, hospitalization, and case results.
Results are presented for prophylaxis, early treatment, and late treatment.
Separately, we perform random-effects meta analysis for studies that analyze
outcomes based on vitamin D sufficiency (non-treatment studies).
Vitamin D undergoes two
conversion steps before reaching the biologically active form as shown in
Figure 2. The first step is conversion to calcidiol, or 25(OH)D, in
the liver. The second is conversion to calcitriol, or 1,25(OH)2D, which
occurs in the kidneys, the immune system, and elsewhere. Calcitriol is the
active, steroid-hormone form of vitamin D, which binds with vitamin D
receptors found in most cells in the body. Vitamin D was first identified in
relation to bone health, but is now known to have multiple functions,
including an important role in the immune system [Carlberg, Martens].
For example, [Quraishi] show a strong
association between pre-operative vitamin D levels and hospital-acquired
infections, as shown in Figure 3. There is a significant delay
involved in the conversion from cholecalciferol, therefore calcifediol
(calcidiol) or calcitriol may be preferable for treatment.
Figure 2. Simplified view of vitamin D sources and conversion.
Many vitamin D studies
analyze outcomes based on serum vitamin D levels which may be maintained via
sun exposure, diet, or supplementation. We refer to these studies as
sufficiency studies, as they typically present outcomes based on vitamin D
sufficiency. These studies do not establish a causal link between vitamin D
and outcomes. In general, low vitamin D levels are correlated with many other
factors that may influence COVID-19 susceptibility and severity. Therefore,
beneficial effects found in these studies may be due to factors other than
vitamin D. On the other hand, if vitamin D is causally linked to the observed
benefits, it is possible that adjustments for correlated factors could
obscure this relationship. COVID-19 disease may also affect vitamin D levels
[Silva], suggesting additional caution in interpreting results for
studies where the vitamin D levels are measured during the disease. For these
reasons, we analyze sufficiency studies separately from treatment studies. We
include all sufficiency studies that provide a comparison between two groups
with low and high levels. Some studies only provide results as a function of
change in vitamin D levels
[Butler-Laporte, Gupta, Raisi-Estabragh], which may not be indicative of results
for deficiency/insufficiency versus sufficiency (increasing already
sufficient levels may be less useful for example).
Some studies show the
average vitamin D level for patients in different groups
[Al-Daghri, Alarslan, Azadeh, Chodick, D'Avolio, Desai, Ersöz, Hosseini (B), Jabbar, Kerget, Latifi-Pupovci, Mansour, Mardani, Morad, Nicolescu, Qu, Ranjbar, Rathod, Saeed, Schmitt, Shannak, Sinnberg, Soltani-Zangbar, Takase, Vassiliou], most of which show lower D levels
for worse outcomes. Other studies analyze vitamin D status and outcomes in
geographic regions
[Bakaloudi, Jayawardena, Marik, Papadimitriou, Rhodes, Sooriyaarachchi, Walrand, Yadav], all
finding worse outcomes to be more likely with lower D levels.
Sufficiency studies vary widely in terms of when vitamin D
levels were measured, the cutoff level used, and the population analyzed (for
example studies with hospitalized patients exclude the effect of vitamin D on
the risk of hospitalization). We do not analyze sufficiency studies in more
detail because there are many controlled treatment studies that provide better
information on the use of vitamin D as a treatment for COVID-19. A more
detailed analysis of sufficiency studies can be found in [Chiodini].
[Mishra] present a systematic review and meta analysis showing that
vitamin D levels are significantly associated with COVID-19 cases.
For studies regarding
treatment with vitamin D, we distinguish three stages as shown in
Figure 4. Prophylaxis refers to regularly taking vitamin D
before being infected in order to minimize the severity of infection. Due to
the mechanism of action, vitamin D is unlikely to completely prevent
infection, although it may prevent infection from reaching a level detectable
by PCR. Early Treatment refers to treatment immediately or soon after
symptoms appear, while Late Treatment refers to more delayed
treatment.
Figure 4. Treatment stages.
6 In Silico studies support the efficacy of vitamin D [Al-Mazaideh, Chellasamy, Mansouri, Pandya, Qayyum, Song].
Preclinical research is an important part of the development of
treatments, however results may be very different in clinical trials.
Preclinical results are not used in this paper.
Table 1 summarizes the results for all stages combined, with different exclusions, for specific outcomes, and for sufficiency (non-treatment) studies.
Table 2 shows results by treatment stage.
Figure 5 plots individual results by treatment stage.
Figure 6, 7, 8, 9, 10, 11, 12, 13, and 14 show forest
plots for treatment studies with pooled effects, peer-reviewed studies,
cholecalciferol studies, calcifediol/calcitriol studies, and for studies
reporting mortality, mechanical ventilation, ICU admission, hospitalization,
and case results only.
Figure 15 shows a forest plot for random effects meta-analysis of
sufficiency (non-treatment) studies.
| Improvement | Studies | Patients | Authors | |
|---|---|---|---|---|
| All studies | 37% [31‑42%] p < 0.0001 **** | 107 | 182,747 | 1,057 |
| After exclusions | 39% [33‑45%] p < 0.0001 **** | 80 | 160,512 | 814 |
| Peer-reviewed studiesPeer-reviewed | 37% [31‑42%] p < 0.0001 **** | 101 | 181,129 | 1,009 |
| Randomized Controlled TrialsRCTs | 31% [17‑42%] p < 0.0001 **** | 26 | 41,836 | 299 |
| RCTs after exclusionsRCTs w/exc. | 34% [17‑47%] p = 0.0003 *** | 19 | 40,741 | 223 |
| Cholecalciferol | 36% [29‑41%] p < 0.0001 **** | 96 | 173,856 | 922 |
| Calcifediol/calcitriolCalcifediol | 49% [26‑65%] p = 0.00036 *** | 11 | 8,891 | 135 |
| Mortality | 36% [28‑44%] p < 0.0001 **** | 60 | 62,080 | 561 |
| VentilationVent. | 26% [-2‑46%] p = 0.068 | 17 | 7,852 | 184 |
| ICU admissionICU | 50% [33‑62%] p < 0.0001 **** | 25 | 40,098 | 274 |
| HospitalizationHosp. | 20% [8‑31%] p = 0.0022 ** | 21 | 85,626 | 200 |
| Recovery | 38% [23‑50%] p < 0.0001 **** | 9 | 545 | 75 |
| Cases | 15% [6‑23%] p = 0.0015 ** | 25 | 134,291 | 284 |
| Viral | 52% [30‑67%] p = 0.00014 *** | 4 | 200 | 26 |
| RCT mortality | 31% [6‑50%] p = 0.02 * | 15 | 1,949 | 174 |
| RCT hospitalizationRCT hosp. | 21% [-4‑40%] p = 0.087 | 8 | 39,713 | 107 |
| Sufficiency | 53% [49‑57%] p < 0.0001 **** | 157 | 182,257 | 1,349 |
| Early treatment | Late treatment | Prophylaxis | |
|---|---|---|---|
| All studies | 60% [40‑74%] **** | 46% [33‑56%] **** | 31% [23‑38%] **** |
| After exclusions | 68% [45‑82%] **** | 58% [45‑69%] **** | 28% [20‑35%] **** |
| Peer-reviewed studiesPeer-reviewed | 57% [36‑71%] **** | 45% [32‑56%] **** | 31% [23‑38%] **** |
| Randomized Controlled TrialsRCTs | 32% [8‑50%] * | 34% [13‑50%] ** | 22% [-129‑73%] |
| RCTs after exclusionsRCTs w/exc. | 65% [-65‑92%] | 34% [16‑47%] *** | 22% [-129‑73%] |
| Cholecalciferol | 60% [40‑74%] **** | 41% [27‑52%] **** | 31% [22‑38%] **** |
| Calcifediol/calcitriolCalcifediol | - | 73% [57‑83%] **** | 32% [6‑50%] * |
| Mortality | 68% [39‑84%] *** | 45% [31‑57%] **** | 21% [6‑33%] ** |
| VentilationVent. | 97% [56‑100%] * | 17% [-14‑40%] | 38% [-3‑63%] |
| ICU admissionICU | 87% [-143‑99%] | 52% [30‑67%] *** | 46% [22‑63%] ** |
| HospitalizationHosp. | 90% [-453‑100%] | 22% [6‑35%] ** | 13% [-4‑27%] |
| Recovery | 31% [7‑49%] * | 43% [24‑58%] *** | - |
| Cases | - | - | 15% [6‑23%] ** |
| Viral | 52% [24‑70%] ** | 53% [8‑76%] * | - |
| RCT mortality | - | 31% [6‑50%] * | - |
| RCT hospitalizationRCT hosp. | - | 29% [10‑44%] ** | -26% [-92‑17%] |
Loading..
Figure 5. Results by treatment stage.
Loading..
Loading..
Loading..
Figure 7. Random effects meta-analysis for peer-reviewed
treatment studies.
[Zeraatkar] analyze 356 COVID-19 trials, finding no significant
evidence that preprint results are inconsistent with peer-reviewed studies.
They also show extremely long peer-review delays, with a median of 6 months to
journal publication. A six month delay was equivalent to around 1.5 million
deaths during the first two years of the pandemic. Authors recommend using
preprint evidence, with appropriate checks for potential falsified data, which
provides higher certainty much earlier.
Effect extraction is pre-specified, using the most serious outcome reported,
see the appendix for details.
Loading..
Loading..
Loading..
Loading..
Figure 10. Random effects meta-analysis for
treatment mortality results only.
Loading..
Figure 11. Random effects meta-analysis for
treatment mechanical ventilation results only.
Loading..
Figure 12. Random effects meta-analysis for
treatment ICU admission results only.
Loading..
Figure 13. Random effects meta-analysis for
treatment hospitalization results only.
Loading..
Figure 14. Random effects meta-analysis for
treatment COVID-19 case results only.
Loading..
Figure 15. Random effects meta-analysis for
sufficiency studies. This plot pools studies with different effects,
different vitamin D cutoff levels and measurement times, and studies may be
within hospitalized patients, excluding the risk of hospitalization. However,
the prevalence of positive effects is notable.
Results restricted to Randomized Controlled Trials (RCTs),
after exclusions, and for specific outcomes are shown in
Figure 16, 17, 18, and 19.
RCTs
help to make study groups more similar and can provide a higher level of
evidence. However they are subject to many biases [Jadad]. For
example, the overhead may delay treatment, dramatically compromising efficacy;
they may encourage monotherapy for simplicity at the cost of efficacy which
may rely on combined or synergistic effects; the participants that sign up may
not reflect real world usage or the population that benefits most in terms of
age, comorbidities, severity of illness, or other factors; standard of care
may be compromised and unable to evolve quickly based on emerging research for
new diseases; errors may be made in randomization and medication delivery; and
investigators may have hidden agendas or vested interests influencing design,
operation, analysis, and the potential for fraud. All of these biases have
been observed with COVID-19 RCTs. There is no guarantee that a specific RCT
provides a higher level of evidence.
High quality RCTs for novel acute diseases are more challenging, with
increased ethical issues due to the urgency of treatment, increased risk due
to enrollment delays, and more difficult design with a rapidly evolving
evidence base. For COVID-19, the most common site of initial infection is the
upper respiratory tract. Immediate treatment is likely to be most successful
and may prevent or slow progression to other parts of the body. For a
non-prophylaxis RCT, it makes sense to provide treatment in advance and
instruct patients to use it immediately on symptoms, just as some governments
have done by providing medication kits in advance. Unfortunately, no RCTs have
been done in this way. Every treatment RCT to date involves delayed treatment.
Among the 50 treatments we have analyzed,
64% of RCTs involve very late treatment 5+ days after
onset. No non-prophylaxis COVID-19 RCTs match the potential real-world use of
early treatments (they may more accurately represent results for treatments
that require visiting a medical facility, e.g., those requiring intravenous
administration).
RCTs have a bias against finding an
effect for interventions that are widely available — patients that
believe they need the intervention are more likely to decline participation
and take the intervention. RCTs for vitamin D are more likely to
enroll low-risk participants that do not need treatment to recover, making the
results less applicable to clinical practice. This bias is likely to be
greater for widely known treatments, and may be greater when the risk of a
serious outcome is overstated. This bias does not apply to the typical
pharmaceutical trial of a new drug that is otherwise unavailable.
Evidence shows that non-RCT trials can also
provide reliable results. [Concato] find that well-designed
observational studies do not systematically overestimate the magnitude of the
effects of treatment compared to RCTs. [Anglemyer] summarized reviews
comparing RCTs to observational studies and found little evidence for
significant differences in effect estimates. [Lee] shows that only
14% of the guidelines of the Infectious Diseases Society of America were based
on RCTs. Evaluation of studies relies on an understanding of the study and
potential biases. Limitations in an RCT can outweigh the benefits, for example
excessive dosages, excessive treatment delays, or Internet survey bias could
have a greater effect on results. Ethical issues may also prevent running RCTs
for known effective treatments. For more on issues with RCTs see
[Deaton, Nichol].
Currently, 37 of 50 treatments we analyze show statistically significant efficacy or harm, defined as ≥10% decreased risk or >0% increased risk from ≥3 studies. Of the 37 treatments with statistically significant efficacy/harm, 23 have been confirmed in RCTs, with a mean delay of 4.4 months. For the 14 unconfirmed treatments, 4 have zero RCTs to date. The point estimates for the remaining 10 are all consistent with the overall results (benefit or harm), with 7 showing >20%. The only treatments showing >10% efficacy for all studies, but <10% for RCTs are sotrovimab and aspirin.
We need to
evaluate each trial on its own merits. RCTs for a given medication and disease
may be more reliable, however they may also be less reliable. For off-patent
medications, very high conflict of interest trials may be more likely to be
RCTs, and more likely to be large trials that dominate meta analyses.
Loading..
Figure 16. Random effects meta-analysis for
Randomized Controlled Trials only. Effect extraction is pre-specified, using the most serious outcome reported,
see the appendix for details.
Loading..
Figure 17. Random effects meta-analysis for
RCTs after exclusions. Effect extraction is pre-specified, using the most serious outcome reported,
see the appendix for details.
Loading..
Figure 18. Random effects meta-analysis for
RCT mortality results.
Loading..
Figure 19. Random effects meta-analysis for
RCT hospitalization results.
To avoid bias in the selection of studies, we include all
studies in the main analysis, with the exception of
[Espitia-Hernandez]. This study uses a combined protocol with another
medication that shows high effectiveness when used alone. Authors report on
viral clearance, showing 100% clearance with treatment and 0% for the control
group. Based on the known mechanisms of action, the combined medication is
likely to contribute more to the improvement.
Here we show the results after excluding studies with critical
issues.
[Murai] is a very late stage study (mean 10 days from
symptom onset, with 90% on oxygen at baseline), with poorly matched arms in
terms of gender, ethnicity, hypertension, diabetes, and baseline ventilation,
all of which favor the control group. Further, this study uses
cholecalciferol, which may be especially poorly suited for such a late stage.
[Cannata-Andía, Mariani] are also very late stage studies using
cholecalciferol.
The studies excluded are as follows, and the resulting forest
plot is shown in Figure 20.
[Abdulateef], unadjusted results with no group details.
[Asimi], excessive unadjusted differences between groups.
[Assiri], unadjusted results with no group details.
[Baykal], unadjusted results with no group details; significant confounding by time possible due to separation of groups in different time periods.
[Campi], significant unadjusted differences between groups.
[Cannata-Andía], very late stage study using cholecalciferol instead of calcifediol or calcitriol.
[Din Ujjan], based on dosages and previous research, combined treatments may contribute more to the effect seen.
[Elhadi], unadjusted results with no group details.
[Fairfield], substantial unadjusted confounding by indication likely.
[Guldemir], unadjusted results with no group details.
[Güven], very late stage, ICU patients.
[Holt], significant unadjusted confounding possible.
[Junior], unadjusted results with no group details.
[Khan], based on dosages and previous research, combined treatments may contribute more to the effect seen.
[Krishnan], unadjusted results with no group details.
[Leal-Martínez], combined treatments may contribute more to the effect seen.
[Lázaro], very few events; unadjusted results with no group details; minimal details provided.
[Mahmood], unadjusted results with no group details; substantial unadjusted confounding by indication likely.
[Mahmood], unadjusted results with no group details; substantial unadjusted confounding by indication likely.
[Mohseni], unadjusted results with no group details.
[Murai], very late stage, >50% on oxygen/ventilation at baseline; very late stage study using cholecalciferol instead of calcifediol or calcitriol.
[Pecina], unadjusted results with no group details.
[Shahid], minimal details provided.
[Shehab], unadjusted results with no group details.
[Singh], minimal details provided.
[Ullah], significant unadjusted confounding possible.
[Zurita-Cruz], randomization resulted in significant baseline differences that were not adjusted for.
Loading..
Figure 20. Random effects meta-analysis
excluding studies with significant issues. Effect extraction is pre-specified, using the most serious outcome reported,
see the appendix for details.
Heterogeneity in COVID-19 studies arises from many factors including:
The time
between infection or the onset of symptoms and treatment may critically affect
how well a treatment works. For example an antiviral may be very effective
when used early but may not be effective in late stage disease, and may even
be harmful. Oseltamivir, for example, is generally only considered effective
for influenza when used within 0-36 or 0-48 hours [McLean, Treanor].
Baloxavir studies for influenza also show that treatment delay is critical
— [Ikematsu] report an 86% reduction in cases for post-exposure
prophylaxis, [Hayden] show a 33 hour reduction in the time to
alleviation of symptoms for treatment within 24 hours and a reduction of 13
hours for treatment within 24-48 hours, and [Kumar] report only 2.5
hours improvement for inpatient treatment.
| Treatment delay | Result |
| Post exposure prophylaxis | 86% fewer cases [Ikematsu] |
| <24 hours | -33 hours symptoms [Hayden] |
| 24-48 hours | -13 hours symptoms [Hayden] |
| Inpatients | -2.5 hours to improvement [Kumar] |
Figure 21 shows a mixed-effects meta-regression for efficacy
as a function of treatment delay in COVID-19 studies from 50 treatments, showing
that efficacy declines rapidly with treatment delay. Early treatment is
critical for COVID-19.
Figure 21. Early treatment is more effective. Meta-regression showing efficacy as a function of treatment delay in COVID-19 studies from 50 treatments.
Details of the patient population including age and comorbidities may
critically affect how well a treatment works. For example, many COVID-19
studies with relatively young low-comorbidity patients show all patients
recovering quickly with or without treatment. In such cases, there is little
room for an effective treatment to improve results (as in
[López-Medina]).
Efficacy may
differ significantly depending on the effect measured, for example a treatment
may be very effective at reducing mortality, but less effective at minimizing
cases or hospitalization. Or a treatment may have no effect on viral clearance
while still being effective at reducing mortality.
There are many
different variants of SARS-CoV-2 and efficacy may depend critically on the
distribution of variants encountered by the patients in a study. For example,
the Gamma variant shows significantly different characteristics
[Faria, Karita, Nonaka, Zavascki]. Different mechanisms of action may be
more or less effective depending on variants, for example the viral entry
process for the omicron variant has moved towards TMPRSS2-independent fusion,
suggesting that TMPRSS2 inhibitors may be less effective
[Peacock, Willett].
Effectiveness may depend strongly on the dosage, treatment regimen, and the
form of vitamin D used (cholecalciferol, calcifediol, or calcitriol).
The use of other
treatments may significantly affect outcomes, including anything from
supplements, other medications, or other kinds of treatment such as prone
positioning.
The
quality of medications may vary significantly between manufacturers and
production batches, which may significantly affect efficacy and safety.
[Williams] analyze ivermectin from 11 different sources, showing
highly variable antiparasitic efficacy across different manufacturers.
[Xu] analyze a treatment from two different manufacturers, showing 9
different impurities, with significantly different concentrations for each
manufacturer.
Non-prescription supplements may show very wide variations in quality
[Crawford, Crighton].
We present both pooled analyses and specific outcome analyses. Notably, pooled
analysis often results in earlier detection of efficacy as shown in
Figure 22. For many COVID-19 treatments, a reduction in mortality
logically follows from a reduction in hospitalization, which follows from a
reduction in symptomatic cases, etc. An antiviral tested with a low-risk
population may report zero mortality in both arms, however a reduction in
severity and improved viral clearance may translate into lower mortality among
a high-risk population, and including these results in pooled analysis allows
faster detection of efficacy. Trials with high-risk patients may also be
restricted due to ethical concerns for treatments that are known or expected
to be effective.
Pooled analysis enables using more of the available
information. While there is much more information available, for example
dose-response relationships, the advantage of the method used here is
simplicity and transparency. Note that pooled analysis could hide efficacy,
for example a treatment that is beneficial for late stage patients but has no
effect on viral replication or early stage disease could show no efficacy in
pooled analysis if most studies only examine viral clearance.
While we present pooled results, we also present individual
outcome analyses, which may be more informative for specific use cases.
Currently, 37 of 50 treatments we analyze show statistically significant efficacy or harm, defined as ≥10% decreased risk or >0% increased risk from ≥3 studies. 100% of treatments showing statistically significant efficacy/harm with pooled effects have been confirmed with one or more specific outcomes, with a mean delay of 3.0 months. When restricting to RCTs only, 55% of treatments showing statistically significant efficacy/harm with pooled effects have been confirmed with one or more specific outcomes, with a mean delay of 2.9 months.
Loading..
Loading..
Figure 22. The time when studies showed that
treatments were effective, defined as statistically significant improvement of
≥10% from ≥3 studies.
Pooled results typically show efficacy earlier than specific
outcome results. Results from all studies often shows efficacy much earlier
than when restricting to RCTs.
Results reflect conditions as used in trials to date, these depend on the
population treated, treatment delay, and treatment regimen.
The
distribution of studies will alter the outcome of a meta analysis. Consider a
simplified example where everything is equal except for the treatment delay,
and effectiveness decreases to zero or below with increasing delay. If there
are many studies using very late treatment, the outcome may be negative, even
though early treatment is very effective. This may have a greater effect than
pooling different outcomes such as mortality and hospitalization. For example
a treatment may have 50% efficacy for mortality but only 40% for
hospitalization when used within 48 hours. However efficacy could be 0% when
used late.
All meta analyses combine heterogeneous studies, varying in
population, variants, and potentially all factors above, and therefore may
obscure efficacy by including studies where treatment is less effective.
Generally, we expect the estimated effect size from meta analysis to be less
than that for the optimal case.
Looking at all studies is valuable for providing an overview of all research,
important to avoid cherry-picking, and informative when a positive result is
found despite combining less-optimal situations. However, the resulting
estimate does not apply to specific cases such as
early treatment in high-risk populations with a specific form and dosage of vitamin D.
While we present results for all studies, we also present treatment time and
individual outcome analyses, which may be more informative for specific use
cases.
Vitamin D studies vary widely in all the factors above, which
makes the consistently positive results even more remarkable. A failure to
detect an association after combining heterogeneous studies does not mean the
treatment is not effective (it may only work in certain cases), however the
reverse is not true — an identified association is valid, although the
magnitude of the effect may be larger for more optimal cases, and lower for
less optimal cases. While we present results for all studies in this paper,
the individual outcome, form of vitamin D, and treatment time analyses are
more relevant for specific use cases.
For sufficiency
studies, different studies use different levels
as the threshold of sufficiency, vitamin D levels were measured at different
times, and some studies measure risk only within hospitalized patients, which
excludes the risk of a serious enough case to be hospitalized. However,
147 of 157 studies present positive effects.
Sufficiency studies show a strong correlation between low
vitamin D levels and worse COVID-19 outcomes, however they do not provide
information on vitamin D treatment. Studies with vitamin D levels measured
after admission may show lower levels because COVID-19 infection reduces
vitamin D levels. Studies with levels measured before infection also show
signficant benefit, however the cause could be one or more correlated factors.
For example, sunlight exposure increases vitamin D levels, but also increases
intracellular melatonin [Zimmerman], and melatonin shows significant
benefit for COVID-19 [c19melatonin.com]. Sun
exposure is also correlated with physical exercise, which also shows benefit
for COVID-19 [c19early.org].
91 of 107 treatment
studies report positive effects. Studies vary significantly in terms of
treatment delay, treatment regimen, patients characteristics, and (for the
pooled effects analysis) outcomes, as reflected in the high degree of
heterogeneity. However treatment consistently shows a significant benefit. The
treatment studies not showing positive effects are mostly prophylaxis studies
with unknown dosages. The only non-prophylaxis studies reporting negative
effects are a small unadjusted retrospective [Assiri],
[Zangeneh] with no details of treatment, and
[Cannata-Andía, Mariani, Murai] which are very late stage studies using
cholecalciferol. For [Murai], the result also has very low statistical
significance due to the small number of events, and the other reported
outcomes of ventilation and ICU admission, which have slightly more events and
higher confidence, show benefits for vitamin D. Calcifediol or calcitriol,
which avoids several days delay in conversion, may be more successful,
especially with very late stage usage.
Acute treatment shows greater efficacy than chronic prophylaxis for mortality
(and in pooled analysis).
One hypothesis is that long-term supplementation may affect normal biological
processing. A key component of vitamin D processing is regulation via the
enzyme CYP24A1, which breaks down active vitamin D. Long-term supplementation
may lead to upregulation of CYP24A1, and potentially lower availability of
active vitamin D where needed during infection. The prophylaxis RCTs to date
[Jolliffe, Villasis-Keever] are consistent with this possibility, with
the shorter-term supplementation in [Villasis-Keever] showing better
results compared to the longer-term high adherence daily supplementation in
[Jolliffe]. Specific forms and administration of vitamin D may
minimize upregulation of CYP24A1 [Petkovich].
[Bader] performed an RCT showing high-dose cholecalciferol (50,000
IU/week) significantly increased IL-6, however other studies have shown no
significant difference in IL-6 [El Hajj, Mousa] (30,000IU/wk and
100,000IU bolus + 4,000IU/day).
Other factors may be responsible for the observed lower
efficacy in prophylaxis studies. For example, analysis of hospitalized
patients is subject to selection bias because long-term accurate-dosage
supplementing individuals may be significantly less likely to be hospitalized.
Studies spanning higher-UV months are subject to confounding. Note that
prophylaxis studies include case results, whereas we may expect vitamin D to
be more effective against serious outcomes. Comparison of acute treatment
versus long-term supplementation should use the specific outcome analyses
rather than the pooled outcome analyses.
Publishing is often biased
towards positive results, however evidence suggests that there may be a negative bias for
inexpensive treatments for COVID-19. Both negative and positive results are
very important for COVID-19, media in many countries prioritizes negative
results for inexpensive treatments (inverting the typical incentive for
scientists that value media recognition), and there are many reports of
difficulty publishing positive results
[Boulware, Meeus, Meneguesso].
One method to evaluate bias is to compare prospective vs.
retrospective studies. Prospective studies are more likely to be published
regardless of the result, while retrospective studies are more likely to
exhibit bias. For example, researchers may perform preliminary analysis with
minimal effort and the results may influence their decision to continue.
Retrospective studies also provide more opportunities for the specifics of
data extraction and adjustments to influence results.
Figure 23 shows a scatter plot of
results for prospective and retrospective treatment studies.
Prospective studies show 51% [34‑64%] improvement in meta
analysis, compared to 32% [25‑38%] for retrospective
studies, suggesting possible negative publication bias, with a non-significant trend towards retrospective studies
reporting lower efficacy.
This gives us further confidence in the significant efficacy seen in all studies.
Loading..
Figure 23. Prospective vs. retrospective studies.
The diamonds show the results of random effects meta-analysis.
Funnel
plots have traditionally been used for analyzing publication bias. This is
invalid for COVID-19 acute treatment trials — the underlying assumptions
are invalid, which we can demonstrate with a simple example. Consider a set of
hypothetical perfect trials with no bias. Figure 24 plot A
shows a funnel plot for a simulation of 80 perfect trials, with random group
sizes, and each patient's outcome randomly sampled (10% control event
probability, and a 30% effect size for treatment). Analysis shows no asymmetry
(p > 0.05). In plot B, we add a single typical variation in COVID-19 treatment
trials — treatment delay. Consider that efficacy varies from 90% for
treatment within 24 hours, reducing to 10% when treatment is delayed 3 days.
In plot B, each trial's treatment delay is randomly selected. Analysis now
shows highly significant asymmetry, p < 0.0001, with six variants of
Egger's test all showing p < 0.05
[Egger, Harbord, Macaskill, Moreno, Peters, Rothstein, Rücker, Stanley].
Note that these tests fail even though treatment delay is uniformly
distributed. In reality treatment delay is more complex — each trial has
a different distribution of delays across patients, and the distribution
across trials may be biased (e.g., late treatment trials may be more common).
Similarly, many other variations in trials may produce asymmetry, including
dose, administration, duration of treatment, differences in SOC,
comorbidities, age, variants, and bias in design, implementation, analysis,
and reporting.
Figure 24. Example funnel plot analysis for simulated perfect trials.
Pharmaceutical drug
trials often have conflicts of interest whereby sponsors or trial staff have a
financial interest in the outcome being positive. Vitamin D for COVID-19
lacks this because it is an inexpensive and widely available supplement.
In contrast, most COVID-19 vitamin D trials have been run by
physicians on the front lines with the primary goal of finding the best
methods to save human lives and minimize the collateral damage caused by
COVID-19. While pharmaceutical companies are careful to run trials under
optimal conditions (for example, restricting patients to those most likely to
benefit, only including patients that can be treated soon after onset when
necessary, and ensuring accurate dosing), not all vitamin D trials
represent the optimal conditions for efficacy.
Other meta analyses for vitamin D treatment can be found in
[Argano, D’Ecclesiis, Hosseini, Nikniaz, Shah, Tentolouris, Varikasuvu, Xie], showing significant improvements for mortality, mechanical ventilation, ICU admission, hospitalization, severity, and cases.
The first version of
[Lakkireddy] was censored based on incorrect claims from an
anti-treatment researcher. For example, the author claims that the gender
difference between arms (7/44 vs. 15/43 female) indicates randomization
failure, however by simulation, using the group sizes and overall gender
ratio, the difference between the number of female patients in each arm is
expected to be ≥8 6.4% of the time (2.7% with ≥8 in the control arm, and 3.7%
with ≥8 in the treatment arm).
Author claims that the difference in CRP would only happen
about one in a billion times. This is incorrect. CRP is not normally
distributed, and the observed values could be due to a very small number of
outliers with very large CRP in one group.
A response from the study authors can be found at
[c19vitamind.com]. The study was
republished.
Summary statistics from
meta analysis necessarily lose information. As with all meta analyses, studies
are heterogeneous, with differences
in treatment delay, treatment regimen, patient demographics, variants,
conflicts of interest, standard of care, and other factors. We provide analyses by specific
outcomes and by treatment delay, and we aim to identify key characteristics in
the forest plots and summaries. Results should be viewed in the context of
study characteristics.
Some analyses classify treatment based on early or late
administration, as done here, while others distinguish between mild, moderate,
and severe cases. Viral load does not indicate degree of symptoms — for
example patients may have a high viral load while being asymptomatic. With
regard to treatments that have antiviral properties, timing of treatment is
critical — late administration may be less helpful regardless of
severity.
Details of treatment delay per patient is often not available.
For example, a study may treat 90% of patients relatively early, but the
events driving the outcome may come from 10% of patients treated very late.
Our 5 day cutoff for early treatment may be too conservative, 5 days may be too late in many cases.
Comparison across treatments is confounded by differences in
the studies performed, for example dose, variants, and conflicts of interest.
Trials affiliated with special interests may use designs better suited to the
preferred outcome.
In some cases, the most serious outcome has very few events,
resulting in lower confidence results being used in pooled analysis, however
the method is simpler and more transparent. This is less critical as the
number of studies increases. Restriction to outcomes with sufficient power may
be beneficial in pooled analysis and improve accuracy when there are few
studies, however we maintain our pre-specified method to avoid any
retrospective changes.
Studies show that combinations of treatments can be highly
synergistic and may result in many times greater efficacy than individual
treatments alone
[Alsaidi, Andreani, Biancatelli, De Forni, Gasmi, Jeffreys, Jitobaom, Jitobaom (B), Ostrov, Thairu].
Therefore standard of care may be critical and benefits may diminish or
disappear if standard of care does not include certain treatments.
This real-time analysis is constantly updated based on
submissions. Accuracy benefits from widespread review and submission of
updates and corrections from reviewers. Less popular treatments may receive
fewer reviews.
No treatment, vaccine, or intervention is 100% available and
effective for all current and future variants. Efficacy may vary significantly
with different variants and within different populations. All treatments have
potential side effects. Propensity to experience side effects may be predicted
in advance by qualified physicians. We do not provide medical advice. Before
taking any medication, consult a qualified physician who can compare all
options, provide personalized advice, and provide details of risks and
benefits based on individual medical history and situations.
Table 4 shows the reported results of physicians that use early
treatments for COVID-19, compared to the results for a non-treating physician.
The treatments used vary. Physicians typically use a combination of
treatments, with almost all reporting use of ivermectin and/or HCQ, and most
using additional treatments, including vitamin D.
These results are subject to selection and ascertainment
bias and more accurate analysis requires details of the patient populations
and followup, however results are consistently better across many teams, and consistent
with the extensive controlled trial evidence that shows a significant
reduction in risk with many early treatments, and improved results with the use of
multiple treatments in combination.
| LATE TREATMENT | ||||||
| Physician / Team | Location | Patients | HospitalizationHosp. | MortalityDeath | ||
| Dr. David Uip (*) | Brazil | 2,200 | 38.6% (850) | Ref. | 2.5% (54) | Ref. |
| EARLY TREATMENT - 38 physicians/teams | ||||||
| Physician / Team | Location | Patients | HospitalizationHosp. | ImprovementImp. | MortalityDeath | ImprovementImp. |
| Dr. Roberto Alfonso Accinelli 0/360 deaths for treatment within 3 days |
Peru | 1,265 | 0.6% (7) | 77.5% | ||
| Dr. Mohammed Tarek Alam patients up to 84 years old |
Bangladesh | 100 | 0.0% (0) | 100.0% | ||
| Dr. Oluwagbenga Alonge | Nigeria | 310 | 0.0% (0) | 100.0% | ||
| Dr. Raja Bhattacharya up to 88yo, 81% comorbidities |
India | 148 | 1.4% (2) | 44.9% | ||
| Dr. Flavio Cadegiani | Brazil | 3,450 | 0.1% (4) | 99.7% | 0.0% (0) | 100.0% |
| Dr. Alessandro Capucci | Italy | 350 | 4.6% (16) | 88.2% | ||
| Dr. Shankara Chetty | South Africa | 8,000 | 0.0% (0) | 100.0% | ||
| Dr. Deborah Chisholm | USA | 100 | 0.0% (0) | 100.0% | ||
| Dr. Ryan Cole | USA | 400 | 0.0% (0) | 100.0% | 0.0% (0) | 100.0% |
| Dr. Marco Cosentino vs. 3-3.8% mortality during period; earlier treatment better |
Italy | 392 | 6.4% (25) | 83.5% | 0.3% (1) | 89.6% |
| Dr. Jeff Davis | USA | 6,000 | 0.0% (0) | 100.0% | ||
| Dr. Dhanajay | India | 500 | 0.0% (0) | 100.0% | ||
| Dr. Bryan Tyson & Dr. George Fareed | USA | 20,000 | 0.0% (6) | 99.9% | 0.0% (4) | 99.2% |
| Dr. Raphael Furtado | Brazil | 170 | 0.6% (1) | 98.5% | 0.0% (0) | 100.0% |
| Dr. Heather Gessling | USA | 1,500 | 0.1% (1) | 97.3% | ||
| Dr. Ellen Guimarães | Brazil | 500 | 1.6% (8) | 95.9% | 0.4% (2) | 83.7% |
| Dr. Syed Haider | USA | 4,000 | 0.1% (5) | 99.7% | 0.0% (0) | 100.0% |
| Dr. Mark Hancock | USA | 24 | 0.0% (0) | 100.0% | ||
| Dr. Sabine Hazan | USA | 1,000 | 0.0% (0) | 100.0% | ||
| Dr. Mollie James | USA | 3,500 | 1.1% (40) | 97.0% | 0.0% (1) | 98.8% |
| Dr. Roberta Lacerda | Brazil | 550 | 1.5% (8) | 96.2% | 0.4% (2) | 85.2% |
| Dr. Katarina Lindley | USA | 100 | 5.0% (5) | 87.1% | 0.0% (0) | 100.0% |
| Dr. Ben Marble | USA | 150,000 | 0.0% (4) | 99.9% | ||
| Dr. Edimilson Migowski | Brazil | 2,000 | 0.3% (7) | 99.1% | 0.1% (2) | 95.9% |
| Dr. Abdulrahman Mohana | Saudi Arabia | 2,733 | 0.0% (0) | 100.0% | ||
| Dr. Carlos Nigro | Brazil | 5,000 | 0.9% (45) | 97.7% | 0.5% (23) | 81.3% |
| Dr. Benoit Ochs | Luxembourg | 800 | 0.0% (0) | 100.0% | ||
| Dr. Ortore | Italy | 240 | 1.2% (3) | 96.8% | 0.0% (0) | 100.0% |
| Dr. Valerio Pascua one death for a patient presenting on the 5th day in need of supplemental oxygen |
Honduras | 415 | 6.3% (26) | 83.8% | 0.2% (1) | 90.2% |
| Dr. Sebastian Pop | Romania | 300 | 0.0% (0) | 100.0% | ||
| Dr. Brian Proctor | USA | 869 | 2.3% (20) | 94.0% | 0.2% (2) | 90.6% |
| Dr. Anastacio Queiroz | Brazil | 700 | 0.0% (0) | 100.0% | ||
| Dr. Didier Raoult | France | 8,315 | 2.6% (214) | 93.3% | 0.1% (5) | 97.6% |
| Dr. Karin Ried up to 99yo, 73% comorbidities, av. age 63 |
Turkey | 237 | 0.4% (1) | 82.8% | ||
| Dr. Roman Rozencwaig patients up to 86 years old |
Canada | 80 | 0.0% (0) | 100.0% | ||
| Dr. Vipul Shah | India | 8,000 | 0.1% (5) | 97.5% | ||
| Dr. Silvestre Sobrinho | Brazil | 116 | 8.6% (10) | 77.7% | 0.0% (0) | 100.0% |
| Dr. Vladimir Zelenko | USA | 2,200 | 0.5% (12) | 98.6% | 0.1% (2) | 96.3% |
| Mean improvement with early treatment protocols | 236,564 | HospitalizationHosp. | 94.0% | MortalityDeath | 94.8% | |
Random effects meta-analysis with pooled effects using the most serious
outcome reported shows 60% [40‑74%] and 37% [31‑42%] improvement for early treatment
and for all studies. Results are similar after restriction to
101 peer-reviewed
studies: 57% [36‑71%] and 37% [31‑42%], and for the 60 mortality results: 68% [39‑84%] and
36% [28‑44%].
Statistically significant improvements are seen in treatment studies for
mortality,
ICU admission,
hospitalization, and
cases.
56 studies from 52 independent teams in 20 different countries show statistically significant
improvements in isolation (40 for the most serious outcome).
Acute treatment (early 60% [40‑74%], late 46% [33‑56%]) shows greater efficacy than chronic
prophylaxis (31% [23‑38%]).
Late stage treatment with calcifediol/calcitriol shows greater
improvement compared to cholecalciferol:
73% [57‑83%] vs.
41% [27‑52%].
This paper is data driven, all graphs and numbers are
dynamically generated. We will update the paper as new studies are released or
with any corrections. Please
submit updates and corrections at https://c19early.org/dmeta.html.
3/28: We added [Schmidt].
3/23: We added [Davran].
3/15: We added [Bucurica].
3/15: We added [Topan].
3/14: We added [Domazet Bugarin, Siuka].
3/4: We added [Şengül].
3/4: We added [Chen].
3/2: We added [Tan].
2/18: We added [Ortatatli].
2/8: We added [Arabi].
1/28: We added [Batur].
1/20: We added [Mostafa].
1/19: We added [Din Ujjan].
1/17: We added [Valecha].
1/8: We updated [van Helmond] to the journal
version.
1/7/2023: We updated discussion of acute treatment vs.
long-term supplementation.
12/31: We added [De Nicolò].
12/20: We added [Abdrabbo AlYafei].
12/20: We updated the discussion of heterogeneity and
RCTs.
12/12: We added [Vásquez-Procopio].
12/3: We added [Tallon].
11/27: We added [Guldemir (B)].
11/26: We added [Sharif].
11/13: We added [Gibbons].
11/8: We added [Said].
11/4: We added [Bychinin (B)].
10/28: We added [Álvarez].
10/26: We added [Hafezi].
10/15: We added [Charla].
10/8: We added [Karimpour-Razkenari].
10/1: We added [Singh].
9/20: We added [Shahid].
9/19: We added [van Helmond].
9/15: We added [Brunvoll].
9/11: We added [Zeidan].
8/25: We added [Hafez].
8/24: We added [Aldwihi, Sharif-Askari].
8/23: We added [Doğan].
8/21: We added [Reyes Pérez].
8/19: We added [Kalichuran].
8/16: We updated [Lakkireddy] to the new version (post
censorship of the previous version).
8/12: We added [Dana, Zurita-Cruz].
8/10: We added [Barrett].
8/5: We added [Bogliolo].
8/3: We added [Alzahrani].
7/27: We added [De Niet].
7/26: We added [Neves].
7/24: We added [Gholi].
7/19: We added [Baykal].
7/2: We added [Hunt].
6/24: We added [Karonova (D)].
5/28: We added [Mariani].
5/24: We added [Ghanei].
5/23: We added [Fiore].
5/20: We added [Hosseini (C)].
5/19: We added [Jabeen].
5/19: We added [Ozturk].
5/8: We added [Charkowick].
5/5: We added [Nguyen].
5/1: We added [Khan].
4/30: We added [Voelkle].
4/24: We added [Davoudi].
4/22: We added discussion of [Lakkireddy].
4/18: We added [Villasis-Keever].
4/17: We added a section on preclinical research.
4/15: We added [Parant].
4/12: We added [Martínez-Rodríguez].
4/5: We added preprint discussion based on
[Zeraatkar].
4/2: We added [Ferrer-Sánchez].
3/31: We added [Ramos].
3/27: We added [Pande].
3/25: We added [Elhadi].
3/23: We added [Jolliffe].
3/20: We added [Bushnaq].
3/19: We added [Shehab].
3/7: We added [Rodríguez-Vidales].
3/5: We added [Reis].
3/4: We added [Nimer].
3/3: We added [Karonova].
2/24: We added [Zidrou].
2/20: We added [Sanson].
2/19: We added [Cannata-Andía].
2/18: We added [González-Estevez, Junior].
2/17: We added [Mahmood].
2/15: We updated [Vanegas-Cedillo] to the journal version.
2/11: We added [Bychinin].
2/8: We added [Subramanian].
2/8: We added [Ranjbar].
2/6: We added [Bishop].
2/4: We added [Ahmed].
2/4: We updated [Dror] to the journal version.
1/30: We updated [Leal-Martínez] to the journal version.
1/29: We added [Ansari].
1/28: We added [Anjum].
1/25: We added [Saponaro].
1/23: We added [Juraj].
1/14: We added [Baguma (B)].
1/13: We updated [Israel] to the journal version.
1/8: We added [Seal].
1/5: We added [Pepkowitz].
1/3/2022: We added [Efird].
12/26: We added [Abdulateef].
12/21: We added [Beigmohammadi, Sainz-Amo].
12/20: We added [Galaznik].
12/17: We added [Seven].
12/16: We added [Parra-Ortega].
12/14: We added [Putra].
12/9: We added analysis of the number of independent research
groups reporting statistically significant positive results.
12/7: We added [Ma].
12/5: We added [Asgari].
12/3: We updated [Loucera] to the journal version.
12/3: We added [Fatemi].
12/3: We added [Kaur].
11/22: Added discussion related to sufficiency studies.
11/14: We added [Gönen].
11/12: We added [Asghar].
11/7: We added [Holt].
11/3: We added [Atanasovska].
11/1: We updated [Golabi] to the journal version.
10/31: We added [Assiri, Bianconi, Leal-Martínez].
10/19: We added [Jimenez].
10/18: We added [Mohseni].
10/16: We added a summary plot for all results.
10/15: We added [Ramirez-Sandoval].
10/15: We added [Maghbooli (B)].
10/14: We added [Arroyo-Díaz, Burahee] and analysis of treatment mechanical ventilation, ICU admission, and hospitalization results.
9/28: We added [Yildiz].
9/27: We added [Derakhshanian].
9/22: We added [Bagheri].
9/14: We added [Ribeiro].
9/14: We updated [Vasheghani (B)] to the journal version of the article.
9/14: We added [Elamir].
9/10: We added [Tomasa-Irriguible].
9/7: We added [Karonova (B), Pecina].
9/6: We added [Soliman].
9/1: We added [Golabi].
8/23: We corrected [Jain] to include the mortality outcome.
8/15: We added [Nimavat].
8/13: We added [di Filippo] and updated [Louca] to the journal version of the article.
8/12: We added [Alpcan].
8/10: We added discussion of the immune system and vitamin
D.
8/2: We added [Matin].
8/1: We added [Pimental].
7/28: We added [Israel (B)].
7/27: We added [Cozier].
7/26: We added [Güven].
7/25: We added [Asimi].
7/24: We added [Orchard].
7/21: We added [Savitri].
7/19: We added [Oristrell].
7/11: We added [Krishnan].
6/25: We added [Cereda (B)].
6/19: We added [Jude].
6/16: We added [Campi].
6/12: We added [Levitus].
6/11: We updated [Oristrell (B)] to the journal version.
6/9: We added [Fasano].
6/8: We updated [Nogués] to the journal version.
6/7: We added [Diaz-Curiel, Dror].
5/29: We added [Sánchez-Zuno (B)].
5/22: We added analysis restricted to cholecalciferol
studies.
5/21: We added [Alcala-Diaz, Li].
5/20: We updated [Lakkireddy] to the journal version.
5/19: We added [AlSafar].
5/10: We added additional information in the abstract.
5/9: We clarified terminology for prophylaxis and added
discussion of heterogeneity.
5/8: We added analysis for treatment studies restricted to
peer-reviewed articles.
4/30: We added [Loucera].
4/29: We corrected the treatment group counts for the early
treatment group in [Annweiler] (there was no change in the relative
risk).
4/24: We added analysis restricted to RCT studies and to
calcifediol/calcitriol studies. We have excluded [Espitia-Hernandez]
in the treatment analysis because they use a combined protocol with another
medication that shows high effectiveness when used alone.
4/14: We added [Blanch-Rubió].
4/13: We added [Lohia, Oristrell (B)].
4/12: We added [Barassi].
4/10: We added [Szeto].
4/9: We added [Ünsal].
4/5: We added [Bayramoğlu, Livingston].
4/4: We added event counts to the forest plots.
3/31: We added [Mendy].
3/30: We added [Macaya].
3/29: We added [Im].
3/28: We added [Freitas].
3/22: We added [Meltzer].
3/15: We added [Vanegas-Cedillo].
3/14: We added [Cereda].
3/12: We added [Charoenngam].
3/10: We added [Mazziotti].
3/6: We added [Ricci].
2/26: We added [Lakkireddy].
2/25: We added [Sulli (B)].
2/20: We added [Gavioli].
2/20: We added [Infante].
2/18: [Murai] was updated to the journal version of the
paper.
2/17: We corrected an error in the effect extraction for
[Angelidi], and we added treatment case and viral clearance forest
plots.
2/16: We added [Susianti].
2/10: We added [