Shortcomings in the Cochrane review on zinc and the common cold by Nault et al. (2024)

Sho comings in he Coch ane e iew on zinc and he common cold by Naul e al. (2024)
Ha i Hemilä, MD, PhD
Depa men o Public Heal h, Uni e si y o Helsinki, Helsinki, Finland
h ps://o cid.o g/0000-0002-4710-307X
h ps://loop. on ie sin.o g/people/866706
h ps://www.scopus.com/au hid/de ail.u i?au ho Id=56223570700
h ps://pubmed.ncbi.nlm.nih.go /? e m=hemila+zinc
h ps://www.m .helsinki. i/home/hemila/zinc.h m
2024-9-14
A ailable a :
h ps://doi.o g/10.5281/zenodo.13762570
Summa y
O e a dozen placebo-con olled ials ha e examined he e ec s o zinc lozenges on common
cold symp oms. Se e al me a-analyses o he ials wi h di e en s a is ical app oaches ha e
demons a ed s ong e idence ha p ope ly composed zinc lozenges can sho en he du a ion
o colds. A Coch ane e iew on zinc o he common cold was published in 2013; howe e , i
was wi hd awn because o da a plagia ism. In 2024, a new Coch ane e iew was published on
zinc o he common cold. The e iew concluded ha “On he basis o his e iew, he cu en
e idence is insu icien o p o ide i m conclusions o ecommend zinc supplemen a ion o
he p e en ion o ea men o he common cold”. This conclusion is subs an ially di e en
om many o he me a-analyses which ha e indica ed bene i . The e o e, I c i ically app aised
he new Coch ane e iew (2024) and desc ibe he e se e al conce ns which explain he
di e en conclusions. The e a e e o s in he inclusion and exclusion o andomized ials, and
in da a ex ac ion and s a is ical analysis in he Coch ane e iew (2024). The Coch ane e iew
au ho s conside ha zinc lozenges a e a o m o die a y supplemen ; howe e , he e ec o
zinc lozenges is local and no explained by e ec s caused by inges ion. The analysis o
ad e se e ec s is misleading; based on he li e a u e, i seems highly unlikely ha he dosage
used in con olled ials, 80–92 mg/day o zinc o 1-2 weeks, would cause se e e ad e se
e ec s. In conclusion, he Coch ane e iew (2024) is no an adequa e analysis o he a ailable
da a on zinc lozenges o he common cold.
1
Con en s Page
S ong e idence ha p ope ly composed zinc lozenges can sho en he du a ion o colds 3
Flaws in he p eceding Coch ane e iew (2011/2013/2015) on zinc o he common cold 4
Conclusion o Coch ane e iew (2024) on he apeu ic zinc o he common cold 5
Re ision o Naul ’s main me a-analysis on zinc ea men o common cold du a ion 6
Pha macology o zinc lozenges is no conside ed in he Coch ane e iew (2024) 11
Zinc lozenges a e no all equal: many lozenges a e ine ec i e 14
Inclusion o lawed ials, and ials inconsis en wi h inclusion c i e ia 18
Igno ing a 48- old a ia ion in zinc dosage in he nasal adminis a ion ials 23
Unsound subg oup compa ison by zinc dosage and by he ype o zinc sal 26
Analysis o ongoing colds a he end o he ollow-up is a s a is ically unsound app oach 28
Expe imen al colds and na u al colds should no be pooled in he same me a-analyses 30
Misunde s andings abou RCT me hods in he Risk o Bias (RoB) assessmen 32
Analysis o ad e se e ec s is no app op ia e 37
PICO c i e ia we e no conside ed in he Coch ane e iew (2024) 43
Po en ial explana ions o he nega i e esul s in he Hemilä (2020) ial we e igno ed 46
Misleading Backg ound sec ion 48
P ac ical implica ions and guidance o u he esea ch 49
Flaws in da a ex ac ion 52
E o s in o he Coch ane e iews 57
Re e ences 59
2
S ong e idence ha p ope ly composed zinc lozenges can sho en he du a ion o colds
O e one dozen andomized ials (RCTs) ha e examined he e ec s o zinc lozenges on he
du a ion o he common cold. The indings ha e been a iable, howe e he e is a plausible
explana ion o he he e ogenei y. Eby demons a ed ha a subs an ial p opo ion o he
a ia ion in he esul s can be explained by he amoun o ee zinc ions eleased om he zinc
lozenges in ha low le els o ee zinc ions explained nega i e indings, and high le els o
ee zinc ions explained posi i e indings [1-5]. Baka also showed ha he le el o ee zinc
ions closely co ela ed wi h he obse ed e icacy o zinc lozenges [6].
Ebys’s analyses, based on solu ion chemis y, a e in o ma i e. Howe e , he impac o zinc
dosage le el can also be seen in a simple analysis by he o al zinc dosage. The e has been
mo e han a 7- old a ia ion in he daily dose o elemen al zinc in he zinc lozenge ials. In
2011, I showed ha 5 RCTs ha used low doses o zinc (<75 mg/day elemen al Zn) uni o mly
ound no bene i om zinc lozenges. In con as , 3 ials used zinc ace a e in daily doses o
≥75 mg elemen al Zn and he pooled esul indica ed a 42% (95% CI: 35-48%) educ ion in
he du a ion o colds [7]. Fi e ials used zinc glucona e in daily doses o ≥75 mg, and colds
we e sho ened on a e age by 20% (95% CI: 12-28%) [7]. Al hough he poin es ima e o
e ec was qui e di e en o ace a e and glucona e sal s in he 2011 analysis, a u he me a-
analysis di ec ly compa ing ials o zinc ace a e lozenges wi h ials o zinc glucona e
lozenges did no demons a e a signi ican di e ence [8]. The pooled es ima e om 7 ials o
ace a e and glucona e sal s wi h zinc dose ≥75 mg/day indica ed ha common colds we e
sho ened on a e age by 33% (95% CI: 21-45%; P = 0.000 000 05 = 10-7) [8]. Such a e y
na ow con idence in e al – as also indica ed by he P- alue – is e y s ong e idence ha
p ope ly o mula ed zinc lozenges can educe he du a ion o colds.
Because he e idence o zinc ace a e lozenges was s onge , he indings o he zinc ace a e
lozenge ials we e analyzed u he . In an Indi idual Pa ien Da a (IPD) me a-analysis, colds
we e sho ened by 36-40% (2.7-day educ ion o he a e age o 7-day colds in he placebo
g oup), and he e ec o zinc ace a e lozenges was no modi ied by age, gende , e hnic g oup,
alle gy s a us, smoking, o baseline se e i y o he common cold [9]. In ano he IPD me a-
analysis using Cox eg ession, zinc ace a e lozenges inc eased he eco e y a e om colds by
a e a io (RR) = 3.1 [10]. Based on da a a ailable o 3 zinc glucona e ials he inc ease in
eco e y a e was by RR abou 2 o 3 [10].
Zinc lozenges a e dissol ed in he o opha yngeal egion, hus i is possible he e a e
di e ences in he size o he e ec o zinc lozenges on he du a ion o pha yngeal symp oms
compa ed wi h nasal symp oms. Howe e , a me a-analysis o he 3 zinc ace a e lozenge ials
did no ind subs an ial di e ences in he e ec s on a ious espi a o y symp oms. Zinc
ace a e lozenges sho ened he du a ion o nasal discha ge on a e age by 34%, nasal
conges ion by 37%, sneezing by 22%, sc a chy h oa by 33%, so e h oa by 18%, hoa seness
by 43%, and cough by 46% [11]. Muscle ache was sho ened by 54%.
The quan ile ea men e ec (QTE) analysis o he 3 zinc ace a e lozenge ials i mly
demons a ed ha he a e age ea men e ec (ATE) o 2.7 days [9] was inconsis en wi h he
e ec s on sho and long colds. The ATE exagge a es he e ec on sho colds, and
unde es ima es he e ec on long colds [12-14]. The QTE analysis shows ha he ela i e
scale much be e cap u es he e ec o zinc ace a e lozenges, hence he 36% educ ion in
common cold du a ion is a mo e use ul es ima e o he e ec o zinc ace a e lozenges han
he 2.7-day educ ion in du a ion [12,14].
3
Flaws in he p eceding Coch ane e iew (2011/2013/2015) on zinc o he common cold
Coch ane e iews a e o en p omo ed as high-quali y sys ema ic e iews ha can be us ed:
“Coch ane: T us ed e idence. In o med decisions. Be e heal h.”
h ps://www.coch ane.o g (Accessed 2024-9-13).
In 2011, I ound ha he e we e se e e e o s in he Coch ane e iew “Zinc o he common
cold” (2011) [15]. I p o ided o mal eedback in which I poin ed ou se e al laws and
encou aged he au ho s o co ec hem [16]. Howe e , when he nex e sion o he Coch ane
e iew was published in 2013 [17], almos all he e o s emained. In he 2023 e sion, he e
also seemed o be plagia ism o ex and da a om ano he a icle as documen ed in a sepa a e
epo [18]. Un o una ely, he edi o s o he Coch ane Acu e Respi a o y In ec ions g oup had
no checked ha he laws demons a ed in 2011 we e co ec ed in he 2013 e ision.
E en ually, his p ocess ended wi h he wi hd awal o ha Coch ane e iew “Zinc o he
common cold” in 2015:
Hemilä iden i ied mul iple e o s in his Coch ane Re iew and made allega ions o
plagia ism o ex and da a om a p e iously published sys ema ic e iew… The Edi o
in Chie ca ied ou u he in es iga ion in o he alleged plagia ism o da a, wi h he co‐
ope a ion o he e iew au ho s, who p o ided supplemen a y in o ma ion in suppo o
hei wo k. The allega ions ela ed o he de i a ion o means and s anda d de ia ions o
da a om some o he included s udies. Al hough he au ho s acknowledge and ci e he
Hemilä 2011 e iew, he Edi o in Chie conside ed ha he au ho s’ explana ion
ega ding some simila i ies in p esen ed da a be ween he wo e iews was no
conclusi e. This e sion o he e iew will he e o e emain wi hd awn [19].
As a esul o he wi hd awal o he Coch ane e iew on zinc o he common cold (2013), he
associa ed JAMA Clinical E idence Synopsis was also e ac ed [20-22].
In e es ingly, he i s au ho o he Coch ane zinc o he common cold e iew [15,17,19],
which was wi hd awn because o plagia ism, s ill se es as an edi o o he Coch ane Acu e
Respi a o y In ec ions g oup despi e iola ing publica ion e hics:
“Con ac Edi o s: D Meenu Singh, India” h ps://a i.coch ane.o g/con ac -us (2024-9-10).
The e a e also laws in ce ain o he Coch ane e iews; see he las sec ion o his documen .
4
Conclusion o Coch ane e iew (2024) on he apeu ic zinc o he common cold
A new Coch ane e iew on zinc o he common cold by Naul e al. was published in 2024
[23]. The conclusions o ea men o he common cold we e a he nega i e:
On he basis o his e iew, he cu en e idence is insu icien o p o ide i m
conclusions o ecommend zinc supplemen a ion o he p e en ion o ea men o he
common cold.
[23, p.28: Implica ions o p ac ice].
This was based on Naul ’s calcula ion in Analysis 9.1 [23, p.163], which led o he es ima ed
e ec :
When zinc is used o cold ea men , he e may be a educ ion in he mean du a ion o
he cold in days (MD ‐2.37, 95% CI ‐4.21 o ‐0.53; I² = 97%; 8 s udies…
[23, p.2: Abs ac ].
This es ima e has a much wide con idence in e al han he abo e-desc ibed es ima e [8] (see
abo e), which indica es ha he e icacy o zinc may be doub ul.
Howe e , he e a e nume ous sho comings in he Coch ane (2024) me a-analysis on zinc and
he common cold by Naul [23]. This documen desc ibes he sho comings.
5

Re ision o Naul ’s main me a-analysis on zinc ea men o common cold du a ion
The e a e 4 majo sho comings in Naul ’s analysis o zinc o ea ing he common cold.
Fi s , in he gene al communi y he e is a la ge a ia ion in he du a ion o colds, which can
las o 1 day o 3 weeks and o e . Ob iously, he 1-day colds canno be sho ened by he 2.37
days calcula ed by Naul . Thus, he “2.37-day” ATE [23] is no applicable o e he wide
a ia ion in cold du a ion. Fu he mo e, gi en he ATE o 2.37 days, and he impossibili y o
1-day colds o be sho ened by mo e han 1 day, i is ob ious ha he e ec o zinc lozenges
mus be g ea e on some colds ha a e longe han 2.37 days. This is demons a ed wi h he
QTE analysis [12-14]. P e iously, i has been shown ha he ela i e scale much be e
cap u es he e ec o ea men s on many con inuous ou comes such as he du a ion o colds
[12-14,24-28]. The e o e, he ela i e scale (pe cen age) was used in ou Coch ane e iew on
i amin C and he common cold in he 2004 [29], 2007 [30,31] and 2013 e sions [32]. The
ela i e scale es ima e is applicable o he 1-day colds as well as o he 3-week colds. Me a-
analysis on he pe cen age e ec scale can be done easily in he Coch ane Re Man p og am
[29,30,32].
Second, Naul w i es ha :
We unde ook me a-analyses only whe e meaning ul, ha is, i he ea men s,
pa icipan s, and he unde lying clinical ques ion(s) we e simila enough o pooling o
make sense. Because he likely mechanisms and po en ial ad e se e ec s o o al and
in anasal zinc di e , we analysed ials o o al zinc (including able s, capsules, sy ups,
o lozenges) sepa a ely om ials o in anasal zinc (including sp ays o gels). I
s a is ical pooling was no app op ia e, we p oduced a na a i e summa y [23, p.10].
We pooled da a om s udies ha we judged o be clinically homogeneous [23, p.11].
Howe e , in he calcula ion o he abo e-men ioned 2.37-day ATE, Naul pooled 3 ials ha
adminis e ed zinc nasally [33-35] wi h 5 ials ha adminis e ed zinc as lozenges [36-40], see
Figu e 1. The wo ea men s a e no simila enough o pooling o make sense because
adminis e ing zinc o di e en ana omical egions migh cause di e en e ec s. Pooling he
zinc lozenge ials wi h nasal zinc ials in Analysis 9.1 is pa icula ly s ange since Naul
w i es in he Me hods sec ion ha hey did no pool, see he ex abo e. Gi en ha he e a e
10 au ho s in he Coch ane e iew [23], i is su p ising ha none o hem picked up his
inconsis ency be ween he epo ing o me hods and he ac ual me hods.
Fu he mo e, he e has been conce n o e decades ha nasal zinc adminis a ion may cause
long-las ing o pe manen anosmia in some pa ien s [41-44]. This conce n is also men ioned
by Naul [23, p.8], hough no p ope ly discussed. The e o e, nasal adminis a ion o zinc is
no app op ia e and es ima ing he e ec o nasal zinc adminis a ion is i ele an unless i
eme ges ha he conce ns a e un ounded. No conce ns abou anosmia ha e been exp essed
abou he adminis a ion o zinc lozenges.
Finally, in Analysis 9.1, he lowes zinc dose is 0.044 mg/day [34] while he highes zinc dose
is 190 mg/day [36]; see Figu e 1. When he in e en ion dose di e s by a ac o o 4300 he
ials a e no “simila enough”.
6
Figu e 1. Naul ’s Analysis 9.1.: Mean du a ion o colds (measu ed in days om s a o
esolu ion o he cold, as de ined by each s udy) - ea men , Ou come 1: Mean
du a ion o colds/URTIs ( ea men ): p ima y analysis.
Uppe : 3 o he included ials we e nasal zinc ials, and 5 we e zinc lozenge ials.
Lowe : he lowes dose o zinc was 0.044 mg/day, and he highes dose was 190 mg/day.
7
Nasal
zinc
O he s
zinc
lozenges
0.044
mg/day
190
mg/day
Thi d, Naul does no include he Mossad (1996) ial [45] in Analysis 9.1. Naul includes he
Mossad (1996) ial in a dozen o he Analyses (10.1, 14.1, 14.2, 14.3, 14.4, 14.6, 14.7, 14.8,
14.10, 14.11, 14.12, 14.13), which implies ha Naul has no conce ns wi h he ial me hods.
Mossad (1996) epo ed he du a ion o colds as su i al cu es. The IPD can be ex ac ed
om he cu es and he eby he mean and SD can be calcula ed. Tha has been published
p e iously, so he mean and he SD we e a ailable [7,8,12,27]. Ne e heless, calcula ing he
means and SDs is s aigh o wa d. Naul does no gi e any explana ion o he exclusion o
he Mossad (1996) ial om he main analysis on ea ing he common cold (Analysis 9.1).
This exclusion o he Mossad (1996) ial om Analysis 9.1 is inconsis en wi h he Me hods
sec ion which s a es ha missing da a we e impu ed:
I nume ical ou come da a we e missing, such as s anda d de ia ions (SD) o co ela ion
coe icien s, and we we e unable o ob ain hese om he ial au ho s, we calcula ed
hem om o he a ailable s a is ics… [23, p.11]
Fou h, one o he zinc lozenge ials included in he calcula ion o he 2.37 day ATE was
ca ied ou wi h child en: Macknin (1998) [38]. Naul does no discuss po en ial issues wi h
ials in child en. In some i amin C and common cold ials he e was s ong e idence ha
child en and adolescen s swi ched hei able s [46,47]. The e o e, common cold ials wi h
child en a e no as eliable as ials wi h adul s. In addi ion, he e can be di e ences in he size
o he e ec be ween adul s and child en. Consequen ly, adul s and child en we e sepa a ed in
he Coch ane e iews on i amin C o he common cold [29,30,32].
I e ised he me a-analysis co esponding o Naul ’s Analysis 9.1 by including only he zinc
lozenge ials, adding he Mossad (1996) ial, and using he ela i e scale. I calcula ed ha
zinc lozenges sho ened colds in adul s by
37% (95% CI: 27-46%; P = 0.000 000 0007 = 10-9); Figu e 2.
Thus, wi h he abo e-jus i ied modi ica ions, using he da a ha Naul collec ed, he
calcula ed es ima e p o ides e y s ong e idence ha zinc lozenges can sho en common
colds in adul s. This is consis en wi h he p e ious s a is ical analyses [1-14].
Inclusion o he ial wi h child en (median age 13) (Macknin 1998) [38] has a minimal e ec
on he es ima e (Figu e 2). Howe e , he e is a highly signi ican di e ence be ween he 5
adul ials and he Macknin (1998) ial wi h child en (P = 0.0001). The e o e, i is mos
in o ma i e o keep he adul and child ials sepa a e. So a he e is no e idence ha zinc
lozenges ha e an e ec on he du a ion o colds in child en. Naul w i es in he Me hods ha :
We conside ed he ollowing ac o s o be po en ial causes o he e ogenei y in he
e ec s o he in e en ion: … he age (child en, adul s, elde ly) o pa icipan s
Howe e , Naul did no sepa a e he Macknin (1998) ial om he adul ials wi h zinc
lozenges. In es iga ion would ha e e ealed ha he e is signi ican he e ogenei y be ween
adul and child ials (Figu e 2).
8
Thus, he 5 adul ials wi h zinc lozenges lead o a conclusion ha p ope ly composed zinc
ace a e and zinc glucona e lozenges can sho en common cold du a ion in adul s by 37% on
a e age.
Figu e 2. Pooling he zinc lozenge ials included in Naul ’s Coch ane e iew [23].
The child ial by Macknin (1998) is signi ican ly inconsis en wi h he 5 ials wi h
adul s. Based on he a io o means es ima e [25], zinc lozenges sho ened colds in
adul s by 37% (95% CI: 27-46%; P = 10-9). The da a a e om [23], excep o he
Mossad (1996) ial, which a e om [27]. The included ials we e pooled wi h he
me agen unc ion o he R package me a [48-50], using he in e se a iance, andom
e ec s op ions.
ROM, a io o means [25]; RoM = 1.0 indica es ha he mean du a ion o colds is
iden ical in he in e en ion and con ol g oups.
9
A en ion o hese omissions is c i ical i we wish o lea n he e ec s o ea ing
common colds wi h zinc lozenges, and i we wish o econcile he nega i e epo o
Tu ne e al. [63] wi h he e y posi i e epo s o Pe us e al. [37] and P asad e al.
[39]. Fo example, P asad e al. [39] showed ha 50% o zinc ace a e ecipien s we e
well in 3.8 days, compa ed wi h 7.7 days o 50% o placebo ecipien s. This du a ion
da a co esponds well wi h o he gene ally accep ed da a.
In he epo by Tu ne e al. [63], he ac ual zinc compound exposed o he o al mucosa
was no zinc ace a e, bu nonmiscible a complexes o zinc. The zinc ace a e lozenges
we e no desc ibed as p oducing a d y o as ingen eeling in he mou h; in all cases
whe e he ZIA alue is su icien ly high o allow Zn2+ ions o sho en he du a ion and
se e i y o common colds, he e has been and he e will be a d y o as ingen eeling in
he mou h. This d y eeling is iden ical o he “clean” mou h eeling p oduced by
swishing wa e in he mou h o 30–60 s.
The cumula i e e ec o he abo e omissions eaches eade s ha zinc ace a e lozenges
in gene al do no ha e e icacy agains common cold; howe e , p ope ly made zinc
ace a e lozenges wo k e y well in educing he du a ion o common colds.
Tu ne did no espond o Eby’s c i icism, which means ha Eby’s c i icism was no
challenged. Fu he mo e, he 30 mg/day dose is pa icula ly low. Finally, he ac ual usage o
lozenges was no epo ed, and he e o e i is no clea whe he he ac ual use o he highe
dosage was 69 mg/day o much less.
Failing o ci e Eby’s le e [64] o Eby’s e iews desc ibing he p oblems o Tu ne ’s (2000)
zinc ace a e lozenges [3,4], Naul includes he Tu ne (2000) ials wi h zinc ace a e lozenges
in Analyses 11.1, 11.2, 11.3, 11.4, 11.5. and in ex page 22.
Eby (2006) [65] Zinc o o a e lozenges
Eby’s commen ed [4, p.485]:
Zinc o o a e is igh ly bound (0 mg iZn [ ee zinc ions]) and essen ially insoluble, and
non-soluble compounds do no elease iZn… Lozenges we e nea ly insoluble and
equi ed mo e han 1 h o dissol e in he mou h. This s udy was he second componen
o ou 1984 clinical ial, and i s esul s we e published in 2 mid-90s books, bu we e
no published as a pee e iewed a icle un il 2006.
Naul ailed o ake his p oblem in o accoun when including he zinc o o a e ial by Eby
(2006) in Analyses 10.1, 11.1, 11.4, 11.5. I he goal is o es ima e he e ec o zinc ace a e
and/o zinc glucona e lozenges (Figu e 3), he Eby (2006) ial is i ele an .
This ield o p oblems in he composi ion o lozenges was igno ed by Naul . I is no
app op ia e o pool ials pu ely on he basis ha au ho s used he e m “zinc lozenge”. Fo
complex in e en ions he e needs o be jus i ica ion o conclude ha 2 e sions o he
in e en ion a e “simila enough o pooling o make sense” [23, p.10].
In addi ion o he chemical composi ion o zinc lozenges, he e has also been a ia ion in he
size o he lozenges and he speed a which hey a e dissol ed in he mou h [4, 66 able 4]. In
he wo P asad ials (2000, 2008), he lozenges dissol ed in abou 30 minu es. In he Al-
16

Nakib (1987) ial, he lozenges dissol ed in abou 20 minu es, and in he Pe us (1998) and
Eby (1986) ials, in abou 15 minu es. In he Hemilä (2020) ial, he lozenges dissol ed in 8
minu es, which was conside ed one po en ial cause o he lack o bene i in ha ial [66].
Al hough se e al andomized placebo-con olled double-blind ials ha e shown ha p ope ly
composed zinc lozenges can sho en common cold du a ion (Figu es 2 and 3) and [1-14], i is
no easy o membe s o he public o ind p ope zinc lozenges. In his 2010 e iew [4], Eby
w o e ha he had looked a he con en s o many dozens o zinc lozenges in he US ma ke [4,
67]:
Lack o egula o y e iew h ough use o homeopa hic laws and die a y supplemen
egula ions o he Uni ed S a es – e en hough h oa lozenges a e no allowed unde he
Uni ed S a es Die a y Supplemen Heal h and Educa ion Ac o 1994 – has esul ed in
comme cializa ion o zinc lozenges ha a e poo ly e ec i e o non-e ec i e…
A 2008 cold-season ma ke su ey by his au ho o zinc lozenges ound in na ional
chain s o es in Aus in, Texas, USA, showed ha none me he c i e ia o high iZn
con en and long dissolu ion imes, and nea ly all eleased ze o iZn…
Mos zinc lozenges also con ained ci ic acid…
Some zinc lozenges con ained non-ionizable (a pH 7.4) zinc compounds including zinc
oxide, aspa a e, a a e, picolina e, o o a e and a ious amino acid chela es, which a e
belie ed ine icacious agains colds. Choices o zinc compounds ha do no elease iZn
a pH 7.4 a e belie ed p edica ed on comme cial desi es o a oid he o ally as ingen
and d ying na u e o iZn, hus a cu e o he common cold is p ecluded by ma ke ing
o ces, no science. A cu en assessmen o mo e han 40 o e - he-coun e zinc
lozenges is main ained on he In e ne [no a ailable any mo e, see copy a [67]]. Zinc
lozenges ma ke ed in he Uni ed S a es appea o compe e based upon as e a he han
e icacy [4, p.490].
O he 40 di e en b ands o o e - he-coun e zinc lozenges and many a ia ions o
hem cu en ly a ailable in he US, e y ew – based upon his analysis and ing edien s
lis ed on hei labels – appea o elease use ul amoun s o iZn ega dless o o al zinc
con en , and none o hem can be conside ed as a cu e o common colds. Wi h se e al
excep ions, nea ly all appea likely o ha e a null e ec on colds [4, p.490].
17
Inclusion o lawed ials, and ials inconsis en wi h inclusion c i e ia
In addi ion o he abo e-desc ibed p oblems wi h he zinc lozenge composi ions in ce ain
ials, he e a e o he p oblems in some ials included by Naul [23].
Naul included he Weismann (1990) [68], Ku ugöl (2006) [69], Ku ugöl (2007) [70],Vakili
(2009) [71], Re ksuppaphol (2013) [72], Sánchez (2014) [73], and Somé (2015) [74] ials in
he Coch ane e iew al hough hese also ha e a ious issues.
In Analyses 11.1, 11.2, 11.4, Naul included he Weismann (1990) ial [68].
In he Me hods sec ion o he Coch ane e iew, Naul w i es: “Types o s udies: We included
andomised con olled ials (RCTs) and excluded s udies using quasi- andomisa ion because
hey a e suscep ible o selec ion bias” [23, p.8].
Howe e , he Weismann (1990) [68] ial did no use andomiza ion, ins ead i used
al e na i e alloca ion. Fo my i s me a-analysis o zinc lozenges, I con ac ed Kaa e
Weismann and in he Supplemen 2 [7, p.xiii], I documen ed:
The 1990 s udy epo did no desc ibe he me hod o alloca ion. D . Kaa e
Weismann desc ibed ha hey had used consecu i e alloca ion
(pe sonal communica ion by email 2 July 2010).
When Naul ’s inclusion c i e ia s ic ly equi e andom alloca ion, a ial should no be
included i he alloca ion me hod is unclea . Fu he mo e, when s a ing a new p ojec on zinc
o he common cold, Naul should ha e ead he p e ious analyses on he same opic and
would ha e ound he abo e in o ma ion abou Weismann’s alloca ion.
In Analyses 1.1, 3.1, 4.1, 7.1-7.9, 8.2, Naul included he Ku ugöl (2006) ial [69].
Ku ugöl epo ed ha he du a ion o colds was 5.3 (SD 0.7) days in he placebo g oup and
4.7 (SD 0.8) days in he zinc g oup, able III in [69]. Naul copies hese igu es o Coch ane
e iew Analysis 3.1. The weigh o he Ku ugöl (2006) ial is he g ea es in Analysis 3.1,
44.7%, so i is no a mino issue i he e a e laws in he Ku ugöl da a.
The e y na ow SD alues a e no belie able. In ou Coch ane e iew we impu ed some
missing SD alues on he basis o es ima ing he a io be ween SD/mean. We w o e [32, p.8]:
Some ials p esen ed he mean du a ion o se e i y o colds, bu no he espec i e
SD… we es ima ed SD as iden ical wi h he mean o he ea men g oup. This is based
on ou analysis ha o ials epo ing he SD, he a io o SD o mean is on a e age 0.7
so ha ou a io o 1.0 used in he SD impu a ion is somewha conse a i e. The
consequence o his is ha we a e pu ing sligh ly educed weigh in ou es ima es o
e ec on hese ials wi h missing SD alues, compa ed o he a e age.
In he Ku ugöl (2006) ial, he SD/mean a ios a e 0.13 and 0.17 in he placebo and zinc
g oups, espec i ely. This a io is e y small compa ed wi h he usual dis ibu ion o common
cold du a ion, so he “SD” alues a e doub ul.
In compa ison, Naul Analysis 3.1 also includes he Ka asu ya (2012) ial. In ha ial, he
SD/mean a ios a e 0.67 and 0.64 in he placebo and zinc g oups, espec i ely, which a e close
18
o he mean alue 0.7 in he la ge se we analyzed o ou Coch ane e iew on i amin C and
he common cold [32].
Fu he mo e, Ku ugöl’s able III epo s he SD o o al cold symp oms, which is 0.7 days in
he placebo g oup as men ioned abo e. The able also epo s he SD o cough (SD 2.0 days),
nasal d ainage (SD 1.8 days), nasal conges ion (SD 1.0 days). I is no easonable o assume
ha he a ia ion in he componen symp oms would sum up o he o al cold symp oms wi h
less a iance han he a iance o he componen symp oms.
Fu he mo e, in igu e 1, Ku ugöl epo s o al symp om se e i y sco es by day in bo h he
zinc and placebo g oups. He calcula ed ha P = 0.000 o day 2, al hough he SE-ba s o he
wo g oups subs an ially o e lap. Thus, Ku ugöl’s s a is ical analysis is ques ionable.
I was able o con ac D . Ku ugöl, who sen an email o me on 2015-1-16. Howe e , when I
desc ibed he p oblems in he SD alues in Ku ugöl’s 2006 pape and asked whe he he da a
se was s ill a ailable, he e was no esponse. The e o e he epo ed SD es ima es should no
be us ed.
E iden ly, a ial should no be included in a me a-analysis, i i is no clea whe he he
dispe sion is epo ed as SD o SE; o a conse a i e SD = mean impu a ion should be used
o common cold du a ion.
Naul included he Ku ugöl (2007) ial [70] in he ex sec ion [23, p.20-25].
The Ku ugöl (2007) epo does no e e o he Ku ugöl (2006) epo ; see abo e. This is e y
s ange since he p e ious zinc ial by he same au ho s is ob iously ele an o he new ial
epo . The Ku ugöl 2007 epo has simila p oblems o he 2006 epo .
Fo example, in igu e 1, Ku ugol epo s o al symp om se e i y sco es by day in he zinc and
placebo g oups. He calcula ed ha P = 0.000 o day 2, al hough he SE-ba s o he wo
g oups subs an ially o e lap. Thus, he s a is ical analysis is also no us wo hy in he 2007
epo .
In Analyses 2.1, 4.1, 8.2 Naul included he Vakili (2009) ial [71]. Vakili epo ed ha he
a e age common cold occu ence was 3.1 (SD 0.55) in he placebo g oup and 1.7 (SD 0.86)
in he zinc g oup, able 2 in [71].
E en s such as he occu ence o colds a e usually o e -dispe sed compa ed wi h a Poisson
dis ibu ion. The Poisson dis ibu ion has he ela ion a iance = mean, bu o na u al e en s
he dis ibu ion is usually wide so ha a iance > mean. The e o e, gi en he mean = 3.1 in
he placebo g oup, he SD should be ≥1.76. Vakili’s SD = 0.55 co esponds o Va = 0.30
which is no belie able o a Poisson ype ou come wi h mean 3.1.
Fu he mo e, Vakili s a es in he Me hods ha “a o al o 200 child en we e andomly
assigned o supplemen a ion wi h 10 mg elemen al zinc as a able (n = 100, 50 males, and 50
emales), o placebo (n = 100, 50 males, and 50 emales).”
Howe e , i is ex emely unlikely ha andom alloca ion o 200 pa icipan s leads o exac ly
he same numbe o bo h sexes wi hin bo h ea men g oups.
19
Finally, no d op-ou s we e epo ed e en hough 200 child en we e ollowed o 5 mon hs.
E iden ly, in such a la ge g oup some d op-ou s a e o be expec ed. Fo all he abo e easons
he esul s o he Vakili 2009 ial should be ques ioned.
In 2018, I ied o con ac D . Vakili and some o he colleagues whom I was able o iden i y
om PubMed pape s on he basis o hem being co-au ho s. I was no success ul in con ac ing
D . Vakili.
When he published SD is s a is ically impossible, he ial should no be included in a me a-
analysis; o a conse a i e SD should be impu ed.
In he Me hods sec ion Naul w i es: “Types o s udies: We included andomised con olled
ials (RCTs) and excluded s udies using quasi- andomisa ion because hey a e suscep ible o
selec ion bias” [23, p.8]. Howe e , gi en he iden ical dis ibu ion o 50 pa ien s in each cell
o he 2×2 able by ea men and sex, i seems unlikely ha he Vakili (2009) ial was
andomized.
In Analyses 1.1, 3.1, 4.1, 7.1-7.3, 8.1, Naul included he Re ksuppaphol (2013) ial [72].
Re ksuppaphol (2013) andomized 100 child en o he zinc and placebo g oups using blocks
o wo. Acco ding o hei Table 1 [72], he a e age age was 10.0 y (SD 0.5 y ) in he zinc
g oup, bu 11.4 y (SD 0.8 y ) in he placebo g oup. The au ho s calcula ed P = 0.0001 o he
di e ence in baseline age be ween he zinc and placebo g oups. Consis en ly, he heigh was
also highly signi ican ly di e en in he zinc g oup (135.5 cm) compa ed wi h he placebo
g oup (141 cm), wi h P = 0.0001. In hei pape , Re ksuppaphol e al. do no p opose an
explana ion o he highly signi ican imbalance in baseline age; hey jus w i e ha :
Despi e s ic andomiza ion and a double blind design, he e was an imbalance in he
s udy be ween he demog aphic p o iles o he wo g oups. This was mos likely
because o s a is ical p obabili y…
This is no a easonable explana ion o a P = 0.0001 di e ence in a baseline a iable,
pa icula ly o a highly essen ial a iable such as age. I 20 baseline a iables a e measu ed
in an RCT, and one o wo o hem gi e a P- alue somewha less han 0.05 o he baseline
di e ence in a - es , ha is consis en wi h success ul andomiza ion. Howe e , success ul
andomiza ion is e u ed when P = 0.0001 o a baseline a iable as essen ial in biology as
age.
As hei mos in e es ing inding, Re ksuppaphol epo ed ha coughs and unny noses we e
sho e in he zinc g oup, wi h P = 0.01. The median du a ion o cough was 1.0 days in he
zinc g oup, and 6.0 days in he placebo g oup. The o al du a ion o he ial was 3 mon hs and
i seems ob ious ha many pa icipan s had mo e han 1 cough episode pe 3 mon hs.
Howe e , he Me hods and Resul s sec ions do no desc ibe whe he he du a ion o cough
means du a ion pe episode o du a ion pe pe son. Thus he ou comes a e no well desc ibed.
Di e ences in cumula i e du a ion o cough can be 1) due o di e ence in incidence o cough
episodes o 2) due o di e ence in he du a ion o cough episodes, o 3) due o a combina ion
o bo h.
20
Gi en he epo ed highly signi ican baseline imbalance in age and heigh , and he lack o
anspa ency in he ou comes, i is di icul o us he epo s on coughs and unny noses.
In he RoB able, Naul assigns a “+” ma k o he “ andom sequence gene a ion” indica ing
ha Naul is sa is ied wi h he alloca ion e en hough he e is published e idence by he
au ho s ha he g oups we e highly signi ican ly biased a baseline wi h P = 0.0001.
Naul desc ibes he indings o he Sánchez (2014) ial [73] in he ex sec ion [23, p.18]:
Sánchez 2014 ound incidence a es o acu e espi a o y in ec ions (ARI) o hose
ecei ing chela ed zinc (1.42 pe 1000 child-days), zinc sul a e (1.57 pe 1000 child-
days), and placebo (3.3 pe 1000 child-days).
The Sánchez (2014) ial was no an indi idual-le el RCT. Sánchez w i es in he abs ac :
“Randomized iple-blind communi y ial wi h 301 child en be ween 2-5 yea s o age om six
child dayca e cen e s in Medellin, Colombia. Child en we e dis ibu ed in h ee g oups
ecei ing zinc amino acid chela e, zinc sul a e and placebo” [73].
Alloca ion is desc ibed in igu e 1 o [73]. Google T ansla e gi es he ollowing e sion o he
alloca ion p ocess [o iginal in Spanish]:
Randomiza ion and masking
The six clus e s we e andomly assigned by means o a ballo so ha wo child en’s
cen e s we e in each o he h ee s udy g oups. One g oup ecei ed zinc sul a e, ano he
g oup ecei ed aminochela ed zinc, and he las g oup ecei ed placebo [73, p.81].
Al hough 6 uni s can be “ andomized” o 3 g oups as 2 + 2 + 2, i is no easonable o
conside ha such “ andomiza ion” leads o meaning ul baseline simila i y. The e is no
jus i ica ion o belie e ha a e y small numbe o obse a ion uni s a e balanced on baseline
a iables. The p e alence o i uses a ies be ween geog aphic egions and be ween social
g oups, e c. Any di e ences be ween a ew child dayca e cen e s can be caused by di e ences
be ween exposu es o i uses o o he ac o s.
I he numbe o uni s is subs an ial, hen i is easonable o assume ha on a e age he
dis ibu ions a e simila o e he nume ous uni s so ha wo la ge g oups o uni s would gi e
simila means i he e is no ea men . Fo example, he Somé (2015) [74] ial included 25
illages, see below. In con as , he compa ison o 2 s 2 s 2 childca e cen e s is no
meaning ul because o he na u al a ia ions in he p e alence o i uses, e c.
Unde he i le “Mean numbe o colds/URTIs de eloped”, Naul desc ibes he indings o he
Somé (2015) [74] ial in he ex [23, p.18]. Naul w i es:
Somé 2015 epo ed hei longi udinal p e alence o URTI a 7.3 (6.2 o 8.4) and 8.1
(6.9 o 9.3) o he 5 mg and 10 mg zinc g oups, espec i ely, compa ed wi h he g oup
who ecei ed a placebo able ’s longi udinal p e alence o 7.5 (6.3 o 8.7).
Howe e , “longi udinal p e alence” is no a o m o “mean numbe o colds/URTIs
de eloped”.
21

In hei pape , Somé [74] de ines:
… longi udinal p e alence as he pe cen o o al days o obse a ion (o ‘ ecalled’
days) on which he disease was p esen (ie, he nume a o is he o al numbe o days
wi h a disease and he denomina o is he o al numbe o days o obse a ion).
Di e ences in “longi udinal p e alence” can be 1) due o di e ence in incidence o episodes
o 2) due o di e ence in he du a ion o episodes, o 3) due o a combina ion o bo h. In any
case, i is no a measu e o “Mean numbe o colds/URTIs de eloped”.
22
Igno ing a 48- old a ia ion in zinc dosage in he nasal adminis a ion ials
In Analysis 9.1, Naul pools 3 nasal zinc adminis a ion ials oge he wi h 5 zinc lozenge
ials.
In Analysis 9.2.1, Naul pools 3 nasal zinc ials as a sepa a e subg oup. In he la e me a-
analysis, he e is ex eme he e ogenei y be ween he 3 nasal zinc ials wi h I2 = 99%. The
maximum o he I2-he e ogenei y scale is 100% and he minimum is 0%, hus he
he e ogenei y o e he 3 nasal zinc ials is ex eme. When he e is subs an ial he e ogenei y,
he goal should be o sea ch o explana ions o he he e ogenei y, ins ead o pooling ials
which a e e y di e en . One e iden possibili y o explain he ex eme he e ogenei y in his
case is he dose used in he ials, which is no conside ed by Naul .
Belongia (2001) w o e ha he “ o al maximum daily dose was 0.044 mg elemen al zinc” [34,
p.104].
Mossad (2003) desc ibes a “ o al daily dose o elemen al zinc o abou 2.1 mg” [35, p.37].
Thus, he daily dose o zinc used by Mossad (2003) was 48 (= 2.1/0.044) imes g ea e han
he daily dose o zinc used by Belongia (2001). Mossad ound a s a is ically signi ican
educ ion in common cold du a ion o 2.4 days, whe eas Belongia ound no signi ican
di e ence be ween he zinc and placebo g oups. An ob ious explana ion is ha he dose used
by Belongia was simply so small ha no e ec is expec ed. Figu e 5 shows he compa ison o
he 2 ials as a o es plo . He e ogenei y be ween he 2 ials is e y high (P = 0.0022).
Figu e 5. Compa ison o he Belongia (2001) [34] and Mossad (2003) [35] ials in a
o es plo on he a io scale. The e is s ong e idence ha he ials by Belongia (2001)
and by Mossad (2003) do no es ima e he same ea men e ec because he P- alue o
he es o inconsis ency is e y small, P = 0.0022. The da a a e om [23].
ROM, a io o means [25].
Assuming ha Mossad’s 2.4-day (37%) educ ion in cold du a ion is a easonable es ima e o
he e ec o he 2.1 mg/day nasally adminis e ed zinc, and assuming a linea dose- esponse
ela ionship, we can in e pola e ha he e ec o he 0.044 mg/day zinc in he Belongia s udy
should be ha colds a e 0.05 days (o 0.8%) sho e in he zinc g oup (2.4 days/48 o
23
37%/48). Figu e 6 shows he ela ion be ween dosage and he pe cen age e ec in he wo
ials. The obse ed di e ence be ween he zinc and placebo g oups in he Belongia (2001)
ial is consis en wi h he expec ed e ec .
Figu e 6. Dosage o zinc and he di e ence be ween he zinc and placebo g oups in he
nasal zinc adminis a ion ials by Belongia (2001) [34] and Mossad (2003) [35]. The
ed diagonal line indica es in e pola ion o dose- esponse assuming linea i y wi h he
cons ain ha he zinc dose o 0 mg/day equals placebo wi h null e ec . The poin
es ima e o he Belongia (2001) ial is e y close o he expec ed null e ec . The
e ical lines indica e he 95% CI ange and he poin in he middle is he obse ed
e ec . The e ec es ima es a e om Figu e 5.
The hi d nasal zinc ial in Naul ’s Analysis 9.1 was Hi (2000) [33], who epo ed he
olume o zinc gel o bo h nos ils pe day (0.96 mL), bu no he concen a ion o zinc in he
gel. In hei in oduc ion hey s a e ha hey in ended o epea an ea lie s udy ha
adminis e ed gel con aining 33 mmol/L o elemen al zinc. I ha was he concen a ion o Hi
nasal zinc gel, ha would lead o zinc dose o 2.06 mg/day. Hi ound signi ican bene i om
zinc, consis en wi h he Mossad (2003) ial. The Hi (2000) ial is no included in Figu e 6
since he dose is no epo ed in he pape [33], hough i can be assumed.
24
Naoul w o e:
We unde ook me a-analyses only whe e meaning ul, ha is, i he ea men s,
pa icipan s, and he unde lying clinical ques ion(s) we e simila enough o pooling o
make sense [23, p.10].
When he e is a 48- old di e ence in he dosage o a d ug in 2 ials, i does no make sense o
assume ha he in e en ions a e “simila enough”.
Fu he mo e, Naul w i es in Me hods [23, p.11]:
We conside ed he ollowing ac o s o be po en ial causes o he e ogenei y in he
e ec s o he in e en ion: … dose o he zinc in e en ion …
Howe e , Naul did no publish any analysis compa able o Figu e 6.
Naul also w i es [23, p.21]:
When h ee s udies u ilising in anasal zinc ... we e compa ed wi h i e s udies u ilising
zinc lozenges … he e was no appa en di e ence be ween subg oups (I2 = 0%;
Analysis 9.2).
He e Naul igno es he lack o s a is ical powe in he compa ison o he zinc lozenge and
nasal zinc ials. The e ec es ima e in he “in anasal zinc” g oup is no accu a e enough o
meaning ul compa isons. The 95% CI o he es ima e o e ec o he nasal zinc g oup is
om -7.36 days o +1.05 days. This means ha i he es ima e in he zinc lozenge subg oup is
wi hin his ange, he di e ence be ween he poin es ima es o nasal zinc and zinc lozenges
would no be s a is ically signi ican . When he es ima e o e ec wi hin he nasal zinc g oup
is e y ex eme wi h I² = 99% (mos ly explained by he 48- old a ia ion in zinc dose), he
mean e ec o such a g oup is no a use ul basis o he compa ison wi h o he kinds o zinc
ea men s.
25
Misunde s andings abou RCT me hods in he Risk o Bias (RoB) assessmen
This sec ion conside s he alidi y o conclusions in he RoB assessmen [23, p.16-17: igu e
3].
Alloca ion concealmen
Naul pu a ques ion ma k “?” o “alloca ion concealmen (selec ion bias)” o he ollowing
7 ials: Caesa (2012), Ka asu ya (2012), Mossad (2003), Pe us (1998), Smi h (1989),
Tu ne (2000) challenge, Tu ne (2000) na u al.
Howe e , he same 7 ials ha e plus ma k “+” (accep able) on “Blinding o pa icipan s and
pe sonnel (pe o mance bias): All ou comes” and on “Blinding o ou come assessmen
(de ec ion bias): All ou comes”. This indica es ha he e is some con usion abou wha
“alloca ion concealmen ” means.
In he old li e a u e “double-blind” was used o indica e ha bo h pa ien s and esea che s did
no know which ea men g oup he pa ien was in un il he end o he ial. O e ime his
e ol ed o desc ibe blinding in mo e de ail. The e a e usually h ee aspec s: he pa ien , he
pe son(s) ca ying ou he in e en ion, and he pe son(s) assessing he ou come, and in he
ideal case all h ee pa ies do no know wha he ea men is un il he ial ends. Tha is
some imes desc ibed as iple-blind, hough mos ly he e m double-blind is used o indica e
comple e blinding.
“Alloca ion concealmen ” is a much mo e ecen concep a ising om he ac ha in many
s udies i is no possible o keep all pa ies blind un il he end o he ial. Fo example, in
su ge y i is di icul o ca y ou a ial ha is double-blind un il he ial ends. In 1983,
Thomas Chalme s showed ha he e we e signi ican di e ences in indings depending on
whe he “ he andomiza ion p ocess was blinded” [78]. This caused in e es in keeping he
alloca ion s age blinded e en i he la e s ages o a ial canno be blinded, e.g. in su ge y.
The name o his concep was la e changed o “alloca ion concealmen ”.
I Naul is sa is ied ha he e is “Blinding o pa icipan s and pe sonnel” and “Blinding o
ou come assessmen ” un il he ial ends, hen – logically – he e mus be blinding a all
p eceding ime poin s du ing he ial, including he p ocess o alloca ion. Thus, all 7 ials
should ha e a “+” o alloca ion concealmen .
Random sequence gene a ion
Fo nume ous ials, Naul pu a ques ion ma k “?” o he i em “Random sequence gene a ion
(selec ion bias)” e en hough he epo s s a e ha he ials we e andomized.
These ials ha e “?” on “ andomiza ion”
Al-Nakib (1987) “membe s o each g oup we e andomly alloca ed o ecei e zinc glucona e
lozenges o placebo” [62]
Eby (1984) “was gi en o each subjec , using a double-blind, andom me hod” [53]
Fa T ial 2 (1987) “In ial 2, 23 subjec s we e andomly chosen o ecei e zinc and 22 we e
andomly chosen o ecei e placebo” [54]
32

Pe us (1998) “The subjec s o his andomized, double-masked, placebo-con olled s udy”
[37]
P asad (2000) “A esea ch consul an p epa ed he andomiza ion code and he packages o
medica ion” [39]
P asad (2008) “A esea ch consul an p epa ed he andomiza ion code and he packages o
medica ion” [40]
Smi h (1989) “Upon en ollmen , subjec s we e andomly assigned o ecei e ei he zinc
glucona e o placebo” [61]
Tu ne (2000) na u al “we e andomized o ecei e 1 o he 3 ea men s” [63]
Tu ne (2001) “Volun ee s we e andomly assigned o ecei e ei he he ac i e p epa a ion o
placebo” [75]
The ials abo e wi h he ques ion ma k “?” can be compa ed wi h ials ha ha e a “+”.
These ials ha e “+” on “ andomiza ion”
Belongia (2001) “The ac i e and placebo nasal sp ay bo les we e andomly placed in blocks
o ou by he manu ac u e ” [34]
Douglas (1987) “ei he zinc ace a e o placebo andomly alloca ed o he sequence” [60]
Fa T ial 1 (1987) “we e assigned by p io compu e andomiza ion o ecei e ei he zinc
glucona e o placebo lozenge he apy in ial 1” [54]
God ey (1992) “Randomiza ion by a hi d pa y was used o assign he 87 pa icipan s o
ea men g oups” [36]
Ku ugöl (2006) “A s a is ical consul an p og ammed a compu e -gene a ed andomiza ion
code and p epa ed he packages o medica ion” [69]
Ku ugöl (2007) “A s a is ical consul an p og ammed a compu e -gene a ed andomiza ion
code and p epa ed he packages o medica ion” [70]
Malik (2014) “We andomized he ea men alloca ion by simple andomiza ion using
compu e gene a ed andom numbe s (Excel 2010)” [79]
Re ksuppaphol (2013) “Using a compu e ized p og amme (G aphPad QuickCals), he
en olled child en we e andomized o he zinc o placebo g oup using blocks o wo by a
s a is ical consul an who was no in ol ed in he implemen a ion phase o he s udy” [72]
Tu ne (2000) challenge “Subjec s we e andomized o ecei e s udy medica ion” [63].
The e a e se e al inconsis encies in Naul ’s assessmen o andomiza ion.
Naul ga e he Fa (1987) T ial2 a “?” bu he T ial1 a “+” al hough he ex s in he Fa
epo a e simila o bo h ials, compa e abo e. I is unlikely ha he andomiza ion me hods
in pa allel ials un by he same esea che s would di e subs an ially.
Naul ga e Tu ne (2000) na u al ial a “?” bu he challenge ial a “+” al hough he ex s in
he Tu ne epo a e also e y simila o bo h ials, compa e abo e. Again, i is unlikely ha
33
he andomiza ion me hods in pa allel ials un by he same esea che s would di e
subs an ially.
Simila ly, why is he
“ esea ch consul an p epa ed he andomiza ion code” in he (P asad 2000, 2008) ials
unsa is ac o y,
whe eas
“Randomiza ion by a hi d pa y” (God ey 1992) and
“A s a is ical consul an p og ammed a compu e -gene a ed andomiza ion code” (Ku ugöl
2006) and
“A s a is ical consul an p og ammed a compu e -gene a ed andomiza ion code” (Ku ugöl
2007) and
“ andomly alloca ed o he sequence” (Douglas 1987), e c. a e sa is ac o y?
Naul pu a “?” o he Pe us (1998) ial. I con ac ed Pe us and asked abou de ails o hei
me hods and ecei ed he ollowing esponse:
The bo les o he zinc lozenges and placebo we e sen by he manu ac u e and each
bo le was iden ical excep a sequen ial numbe . A egis a ion, a e quali ying o he
s udy each pa ien was gi en a bo le o 180 lozenges. A he conclusion o he s udy,
when he dia ies we e assembled, he code o he bo les was sen by he manu ac u e ,
and he pa ien s we e placed in he zinc o placebo ca ego y. Then he esul s we e
abula ed and he s a is ical analysis was unde aken (Edwa d Pe us 24 Ma ch 2016).
[8,9,10,27]
When s a ing a new analysis on zinc ea men o common cold, Naul should ha e ead he
p e ious me a-analyses on zinc lozenges and would ha e ound ha he e is addi ional
in o ma ion abou he Pe us (1998) ial.
Ano he ques ionable classi ica ion by Naul was o pu a “+” ma k o he RoB able o he
Re ksuppaphol (2013) ial which epo ed highly signi ican baseline di e ence o age and
heigh (P = 0.0001); see abo e. The goal o andomiza ion is o minimize selec ion bias.
When such a g ea imbalance is obse ed and epo ed, he bias does no disappea wi h a
“sa is ac o y” desc ip ion o andomiza ion me hods.
Selec i e epo ing
Naul pu a ques ion ma k “?” on nea ly all ials [23, igu e 3].
The Coch ane Handbook s a es [80]:
This domain add esses bias ha a ises because he epo ed esul is selec ed (based on
i s di ec ion, magni ude o s a is ical signi icance) om among mul iple in e en ion
e ec es ima es ha we e calcula ed by he ial au ho s.
The i em is ele an , o example, in ials on pain, in which case he e can be a dozen
di e en measu es o pain expe ience. In ac , pain is used as one example in he Coch ane
Handbook [80]:
Selec i e epo ing o a pa icula ou come measu emen (based on he esul s) om
among es ima es o mul iple measu emen s assessed wi hin an ou come domain.
34
Examples include: … use o mul iple measu emen ins umen s (e.g. pain scales) and
only epo ing da a o he ins umen wi h he mos a ou able esul ...
The e a e no simila conce ns abou he du a ion o he common cold which canno be
measu ed in a dozen di e en ways. I is no clea why Naul has gi en a “?” o he majo i y
o ials.
Po en ial o bias depends on wha was ound
I a ial inds a posi i e e ec , hen i is easonable o conside whe he sys ema ic biases
migh explain he obse ed di e ence, o pa o i . Howe e , i he e is no di e ence be ween
ial g oups, hen he conce n o bias is di e en . I he obse ed null e ec is hough o be a
esul o bias, hen he c i ic belie es ha he ue unbiased e ec is no null.
Weismann (1990) ound no di e ence be ween he zinc lozenge and placebo g oups.
Naul pu he “-” ma k on he “Random sequence gene a ion” i em o he ial in he isk o
bias able, which means “High isk o selec ion bias”. Thus, his indica es ha he null e ec
is biased in Naul ’s iew.
Howe e , Naul e al. do no discuss whe he hey belie e ha he unbiased es ima e in
Weismann’s ial indica es ha m o bene i om zinc lozenges, i hey conside ha he
obse ed null e ec is biased.
Smi h (1989) ound no di e ence be ween he zinc lozenge and placebo g oups.
Naul pu a “-” ma k on he “Incomple e ou come da a” i em o he ial in he isk o bias
able, which means “High isk o a i ion bias”.
Howe e , Naul e al. do no discuss whe he hey belie e ha he unbiased es ima e in
Smi h’s ial indica es ha m o bene i om zinc lozenges, i hey conside ha he obse ed
null e ec is biased.
Tu ne (2001) ound no di e ence be ween he nasal zinc and placebo g oups.
Naul pu a “?” ma k on he “Blinding o pa icipan s and pe sonnel” and “Blinding o
ou come assessmen ” i ems o he ial, which means “Unclea isk o pe o mance and
de ec ion bias”.
Howe e , Naul e al. do no discuss whe he hey belie e ha he unbiased es ima e in
Tu ne ’s (2001) ial indica es ha m o bene i om zinc lozenges, i hey conside ha he
obse ed null e ec is biased.
Thus, i is no clea ha Naul conside ed he ac ual po en ial o bias in he included ials.
Some no es on alid and in alid c i icisms
Commen s on me hodological e alua ion o RCTs by Meine (1986) a e hough ul [81]:
Valid and in alid c i icisms:
The e is no such hing as a pe ec s udy, only a ying deg ees o impe ec ion. The
p o essional c i ic can always ci e one o mo e o he c i icisms lis ed in Table 26-3
wi hou ea o con adic ion. Fo example, he can always a gue ha he esul s o he
ial should be igno ed because he in es iga o s s udied he “w ong” popula ion. O he
can challenge he choice o ea men s o he way in which hey we e adminis e ed. And
i is always possible o chide in es iga o s because hey ailed o collec “impo an ”
35
da a—a leas as iewed om he pe spec i e o he c i ic. The p oblem is no coming
up wi h c i icisms, bu in deciding whe he o no hey a e alid. The ouble wi h he
c i icisms lis ed in Table 26-3 is ha hey a e so b oad and sweeping as o be beyond
deba e.
A c i icism, o be alid, should:
• Ha e some basis in ac
• Be bu essed wi h suppo ing e idence
• Make a di e ence in he in e p e a ion o he esul s [i alics added]
All h ee es s should be me . Among he h ee, he hi d is he mos di icul one o
sa is y. Fo example, i is ai ly easy o c i icize a ial because o di e ences in he
baseline composi ion o he ea men g oups. Howe e , i is qui e ano he hing o show
how hose di e ences migh ha e accoun ed o he esul s obse ed. The a iabili y has
o be sizable and mus occu in connec ion wi h an impo an p edic o o ou come o
make any eal di e ences in he esul s.
Table 26-3 Uni e sal c i icisms [sho ened o a ew examples]
• T ea men p o ocol no ollowed in all cases
• T ea men s no p ope ly adminis e ed
• Design o he s udy lawed (w ong design, inadequa e s a i ica ion,
w ong me hod o andomiza ion)
• Execu ion o he ial aul y
• Resul s no de ini i e
Naul does no show how he ”bias” may ha e accoun ed o he lack o obse ed bene i in
he Weismann, Smi h, and Tu ne ials.
36
Analysis o ad e se e ec s is no app op ia e
The mos common ad e se e ec s o zinc lozenges a e bad as e and i i a ion and d yness in
he mou h, and gas oin es inal discom o .
Naul ’s conclusion on ad e se e ec s o zinc ea men in he Abs ac [23, p.2]:
The e is p obably an inc ease in he isk o non-se ious ad e se e en s when zinc is
used o cold ea men (RR 1.34 …); no ea men s udy p o ided in o ma ion on
se ious ad e se e en s.
Howe e , he e a e laws in Naul ’s analysis o ad e se e ec s.
Fi s , he a ia ion in lozenge composi ion is no aken in o accoun : uni o m ad e se e ec s
should no be expec ed.
Second, es ima ing he size o he ad e se e ec as a RR is no meaning ul when he ocus is
on mild ad e se e ec s.
In his e iews, Eby poin ed ou ha he as e p oblems o zinc lozenges la gely depend on he
composi ion o he lozenges:
Unlike some unhelp ul zinc glucona e lozenges, zinc ace a e lozenges a e la ou -s able
and do no become bi e ega dless o ime o s o age condi ions. P ope ly p epa ed
ZIA 100 zinc ace a e lozenges ha e no objec ionable as e o a e as e in common cold
ea men . They do no seem o p oduce side-e ec s p e iously associa ed wi h zinc
glucona e lozenges, pe haps because much less zinc ace a e is equi ed o p oduce
iden ical esul s han zinc glucona e o any o he sui able zinc compound [2, p.491].
ZG [zinc glucona e] lozenges manu ac u ed wi hou dex ose-based ca bohyd a es a e
bland as ing and p oduce a annic acid-like mou h eel. Howe e , when ca bohyd a es
(excluding uc ose) a e used in lozenges, a slow chemical eac ion occu s, which o e a
ew weeks o a ew mon hs ime esul s in a change in la o o ZG om bland o
noisomely bi e . Consequen ly, some means o p e en ing his eac ion was needed o
p oduce pleasan as ing lozenges con aining ZG.
Two o en moles o glycine ela i e o ZG p e en s he ad e se la o eac ion
acco ding o US pa en 4,684,528 licensed o he Quigley Co po a ion (Doyles own,
PA), he manu ac u e o Cold-Eeze b and zinc glucona e–glycine (ZGG) lozenges [3,
p.29].
Se e al ials used o he ood acids o la o -mask he bi e ZG/dex ose eac ion,
esul ing in loss o e icacy… [3, p.31].
Al hough pu e ZG is bland and chalky in as e, i eac s wi h dex ose and ela ed
ca bohyd a es (excluding uc ose) upon aging o lozenge composi ions o p oduce
noisome bi e ness and compliance- ela ed ine icacy. ZG eleases la ge amoun s o
neu ally cha ged hyd oxide species likely o c oss cell memb anes and causes o al
i i a ion. Bi e ness occu s in all ZG lozenges excep hose ha ei he do no con ain
ca bohyd a es (excluding uc ose), o ha con ain s ong ex amola zinc binding
agen s, which esul s in some hing o he han ZG. Fo hese easons, ZG is no longe
belie ed sui able o use in zinc lozenges o ea ing colds. [3, p.34-35].
On he o he hand, ZA [zinc ace a e] allows he p oduc ion o pleasan as ing, la o
s able lozenges eleasing la ge amoun s o iZn [ ee zinc ions] ei he in ha d candies o
37

comp essed able s wi hou la o o s abili y issues. The mou h- eel p oduced is
su icien ly like he mou h- eel o ea (sligh as ingency) o allow using annic acid
wi hou added bi e agen s as a placebo in clinical ials [3, p.35].
F uc ose is he only ca bohyd a e swee ene ha does no become bi e a e aging o
se e al weeks when combined wi h zinc glucona e [4, p.484].
Due o se ious as e issues zinc glucona e was a poo choice o ea ing colds. Zinc
glucona e o ms ex emely bi e complexes wi h all swee ca bohyd a es excep
uc ose… The o e iding sou ce o ailu e was equi emen by pha maceu ical
ma ke ing companies o pleasan as ing, candy-like, non-me allic, non-as ingen and
non-d ying zinc lozenges [4, p.488].
Gi en his impac o zinc lozenge composi ion on as e, i is e iden ha he composi ion
should be conside ed when conside ing non-se e e ad e se e ec s.
Fu he mo e, he pe inen ques ion is he equency and ype o ad e se e ec s om he
lozenge composi ions ha wo ked (Figu es 2 and 3), and no he equency and ype o
ad e se e ec s om he zinc lozenges ha did no wo k, such as lozenges con aining ci ic
acid, a a ic acid, manni ol/so bi ol, o o a e, e c.
In my RCT, I also poin ed ou ha
he apid dissolu ion o lozenges may ha e exace ba ed as e and o he ad e se e ec s.
I 80 mg/day o zinc is dissol ed in mou h o e 0.7 hou s pe day, he empo a y zinc
ion concen a ions in he o opha yngeal egion a e se e al imes highe compa ed wi h
he same zinc dose being dissol ed o e some 3 hou s pe day [66].
I is also e iden ha he dosage in luences, o example, gas oin es inal discom o , so he
occu ence o ad e se e ec s should be conside ed as a unc ion o dosage.
Finally, ega ding mino ad e se e ec s such as as e, i is no easonable o compa e zinc
lozenges wi h placebo lozenges, when he goal in an RCT is o o mula e a placebo ha as es
he same as he zinc lozenge. I he placebo is pe ec , he placebo lozenge and he zinc
lozenge canno be dis inguished by as e o o he e ec s un ela ed o he common cold.
Howe e , his means ha bo h he placebo and he zinc lozenges can as e so e ible ha no
pa ien wan s o con inue he in e en ion, e en hough he ad e se e ec s o zinc lozenges
would no di e om he placebo lozenges when measu ed as a RR. Thus, he lack o
di e ence be ween an ideal placebo and zinc lozenge does no indica e lack o ad e se e ec s
o he pa icula zinc lozenges.
Because o hese issues, Naul ’s calcula ion o a single es ima e o ad e se e ec such as RR
= 1.34 “when zinc is used o cold ea men ” [23, p.2] is scien i ically unsound. When he
ad e se e ec s a e mino and sho -las ing, such ha hey disappea wi h he e mina ion o
zinc lozenge adminis a ion, he RR is no a easonable measu e o he ad e se e ec s.
Naul w i es abou he analysis o ad e se e ec s o zinc ea men s as ollows:
We pooled 16 ea men s udies (Belongia 2001; Caesa 2012; Eby 1984; Eby 2006;
God ey 1992; Hemilä 2020; Hi 2000; Macknin 1998; Mossad 1996; Mossad 2003;
Pe us 1998; P asad 2000; P asad 2008; Tu ne 2000 challenge; Tu ne 2000 na u al;
Weismann 1990). The e is p obably an inc ease in he isk o expe iencing non-se ious
38
ad e se e en s o hose aking zinc compa ed wi h hose aking placebo (RR 1.34, 95%
CI 1.15 o 1.55; I2 = 44%; 16 s udies, 2084 pa icipan s; mode a e-ce ain y e idence;
Analysis 11.1). [23, p.22]
Belongia (2001) adminis e ed 0.044 mg/day zinc.
Eby (1984) adminis e ed 207 mg/day zinc, which is 4700 imes Belongia’s dose.
God ey (1992) adminis e ed 190 mg/day zinc, which is 4300 imes Belongia’s dose.
The e is no basis o conside ha ials wi h o e a 4000- old di e ence in zinc dosage a e
“simila enough o pooling o make sense”.
Caesa (2012) adminis e ed 20 mg/day zinc as powde , while powde is di e en om nasal
zinc and zinc lozenges, and he e is no basis o assume he same ad e se e ec s.
Fu he mo e, he e a e calcula ion e o s in he doses o Mossad (1996) and P asad (2000)
ials, see sec ion “Flaws in da a ex ac ion”. In Analysis 11.4. (Ad e se e en by dose o
zinc), Naul classi ies hese ials in he “high dose” >85 mg/day subg oup, whe eas he
co ec doses a e less han 85 mg/day and hus he ials should be in he “low dose” <85
mg/day subg oup.
As no ed abo e, he e is also no basis o assume ha he ad e se e ec s o zinc lozenges (Eby
1984; God ey 1992; Hemilä 2020; Macknin 1998; Mossad 1996; Pe us 1998; P asad 2000;
P asad 2008; Tu ne 2000; Weismann 1990) and nasal zinc (Belongia 2001; Hi 2000;
Mossad 2003) a e simila enough o pool hem oge he .
In ac , Naul w i es in he Me hods sec ion ha hey did no pool o al and in anasal zinc:
Because he likely mechanisms and po en ial ad e se e ec s o o al and in anasal zinc
di e , we analysed ials o o al zinc (including able s, capsules, sy ups, o lozenges)
sepa a ely om ials o in anasal zinc (including sp ays o gels).
Howe e , in Naul ’s Analysis 11.1 his desc ip ion o me hods was no ollowed. Gi en ha
he e a e 10 au ho s in he Coch ane e iew [23] and all a e expec ed o ead and accep he
ex , i is conce ning ha no one o hem wonde ed whe he he epo ed me hods and he
ac ual me hods we e consis en .
The mos ele an measu e o assessing mild ad e se e ec s o zinc lozenges seems o be he
p opo ion o pa ien s who do no expe ience discom o s which a e oo annoying, and he
p opo ion who do no discon inue he ea men . Wi h his easoning, I assembled ele an
da a abou he zinc lozenge g oups o he mo e impo an ials (Table 2).
39
Table 2. Ad e se e ec s o zinc lozenges.
P opo ion
wi hou
he ad e se e ec
T ial / Ad e se e ec s Pa ien s
wi h
Pa ien s
wi hou
God ey (1992)
Zinc glucona e, N = 35
Any ad e se e ec s 20 15 43%
Gas oin es inal discom o 13 22 63%
Mou h i i a ion 12 23 66%
Dizziness 3 32 91%
Mossad (1996)
Zinc glucona e, N = 49
Any ad e se e ec s 44 5 10%
Nausea 10 39 80%
Cons ipa ion 1 48 98%
Mou h i i a ion 12 37 76%
Mou h d yness 6 43 88%
Bad as e 39 10 20%
P asad (2000)
Zinc ace a e, N = 25
Nausea 0 25 100%
Cons ipa ion 6 19 76%
Mou h i i a ion 10 15 60%
Mou h d yness 18 7 28%
Bad as e 13 12 48%
P asad (2008)
Zinc ace a e, N = 25
Nausea 3 22 88%
Cons ipa ion 2 23 92%
Mou h i i a ion 1 24 96%
Mou h d yness 13 12 48%
Bad as e 15 10 40%
Swee as e 11 14 56%
Bi e as e 8 17 68%
Sou as e 7 18 72%
Hemilä (2020)
Zinc ace a e, N = 46
Any ad e se e ec s 29 17 37%
S omach ache 12 34 74%
Tas e p oblems 24 22 48%
The able shows he epo ed ad e se e ec s in he zinc g oups o he ials.
40
God ey (1992) [36] epo ed ha 43% o he zinc lozenge pa icipan s did no epo any
ad e se e ec s. In he zinc lozenge g oup, 1 wi hd ew o spo s inju y, 1 o in luenza, 1 o
bac e ial in ec ion, and 1 because o doub ing e icacy: hese 4 a e no wi hd awals because o
ad e se e ec s. In addi ion, in he zinc lozenge g oup, 1 wi hd ew o nausea and 3 ailed o
appea a a ollow-up. The la e 4 pa ien s may be ela ed o ad e se e ec o zinc lozenges,
which gi es 2% (1/43) as he minimum pe cen age ( hough nausea may also be caused by he
i us), and 9% (4/43) as he maximum pe cen age who wi hd ew because o ad e se e ec s.
Mossad (1996) [45] epo ed ha 10% o he 49 zinc lozenge pa ien s had no ad e se e ec s.
Mossad w o e: “One pa ien in he zinc g oup wi hd ew om he s udy on he i s day
because she could no ole a e he lozenges.” Howe e , Mossad con inued: “All o he
pa ien s, as di ec ly obse ed by he s udy nu se, indica ed ha hey had good ole ance o he
i s lozenge.” Fu he mo e, “ wo zinc ecipien s d opped ou a e 7 o 16 days” bu he e is
no in o ma ion on he easons. Thus, 2% wi hd ew because o no ole a ing he zinc lozenge,
while 94% con inued o he end (no making guesses o he 2 d opou s a he la e s age).
P asad (2000) [39] and P asad (2008) [40] did no publish how many pa icipan s epo ed no
ad e se e ec s. In bo h ials, issues wi h bad as e we e epo ed by abou hal he
pa icipan s. Howe e , none o he pa icipan s in he zinc g oups (50 = 25 + 25 in he zinc
g oups o he 2 ials) wi hd ew om he ial, indica ing ha he ad e se e ec s we e qui e
mild in he opinion o he pa ien s. In he 2000 ial, “ wo pa icipan s in he placebo g oup
d opped ou on day 2” bu ha is no explained as an ad e se e ec o he zinc lozenges.
Hemilä (2020) [66] ound ha 37% (95% CI: 23% o 53%) o zinc lozenge pa icipan s did
no complain o any ad e se e ec s.
In summa y, he e is g ea a ia ion in whe he , and how, di e en pa ien s expe ience
discom o om he es ed zinc lozenges. While a subs an ial p opo ion o pa ien s ha e no
expe ienced ad e se e ec s, a ew ha e discon inued ials because o hem.
All he epo ed ad e se e ec s in he zinc lozenge ials we e mino incon eniences. Thus, i
a pa ien s a s es ing whe he zinc lozenges a e use ul o his o he cu en common cold,
he pa ien su e ing om acu e ad e se e ec s such as bad as e can simply s op aking he
lozenges. Such an app oach would no be di e en om he o dina y use o o e - he-coun e
medicines. Fo example, i a pa ien akes a painkille o headache, bu ge s s omach ache as
an ad e se e ec , he pa ien can discon inue he medica ion any ime. A s omach ache o an
indi idual pa ien does no p e en o he people om using painkille s o hei headaches.
The p opo ion o pa ien s who don’ ge ad e se e ec s om zinc lozenges (Table 1) appea s
a much mo e use ul app oach o analyze he occu ence o mino ad e se e ec s om zinc
lozenges han he “ he isk o non-se ious ad e se e en s when zinc is used o cold ea men
(RR 1.34 …)” epo ed by Naul [23, p.2].
In he Plain Language Summa y, Naul w i es [23, p.3]:
S udies adminis e ing in anasal zinc did no epo any cases o anosmia (loss o sense
o smell)
Howe e , analysis o ad e se e ec s o zinc should no be limi ed o he small RCTs included
in he me a-analysis es ima ing he ea men e icacy. The e a e ele an da a ou side o he
ials included by Naul .
41
Misleading Backg ound sec ion
In he Backg ound sec ion, Naul w i es:
Whils many people belie e ha zinc may be help ul in p e en ing o ea ing colds,
he e is cu en ly no es ablished in e en ion o p e en colds o o sho en hei
du a ion.
This s a emen igno es ou Coch ane e iew on i amin C and he common cold [29-32]. We
showed ha egula i amin C supplemen a ion o ≥0.2 g/day sho ened he du a ion o colds
sepa a ely in adul s by 7.7% (P = 0.0002) and in child en by 14% (P = 0.000 05) [32]. We also
calcula ed ha in 5 ials wi h pa icipan s unde hea y sho - e m physical s ess, i amin C
educed he incidence o colds by RR = 0.52 (P = 0.000 001).
In ou 2023 me a-analysis on he e ec s o i amin C on he se e i y o common cold
symp oms, we es ic ed o 15 compa isons in which ≥1 g/day i amin C was adminis e ed o
pa icipan s ini ially in good heal h, and when a common cold occu ed, he se e i y was
dec eased on a e age by 15% (P = 0.000 002) in he i amin C g oups [47]. Thus, he e is
e y s ong e idence ha in ce ain con ex s i amin C can p e en colds and educe hei
du a ion and se e i y.
Bias agains using i amin C o p e en and ea he common cold has been demons a ed in
se e al pape s [24,108-128].
Naul also igno es 4 placebo-con olled RCTs on nasal io a-ca ageenan [129-132] and 2
me a-analyses based on hose RCTs [133,134]. In an Indi idual Pa ien Da a (IPD) me a-
analysis o 2 ials in which io a‐ca ageenan was adminis e ed nasally 3 imes pe day o 7
days o pa ien s wi h he common cold [129,130], we calcula ed ha nasal ca ageenan
inc eased he eco e y a e om all colds by 54% (95% CI: 15-105%; P = 0.003) [133]. The
inc ease in he eco e y a e was 139% o co ona i us in ec ions, 119% o in luenza A
in ec ions, and 70% o hino i us in ec ions.
In a Quan ile T ea men E ec analysis o he nasal ca ageenan ials, he e was no indica ion
o a meaning ul bene i o pa ien s who had colds sho e han one week; howe e , o
pa ien s who had colds las ing wo weeks o mo e, ca ageenan sho ened colds by
up o 5–7 days [14,133]. Ano he IPD me a-analysis also concluded ha he e was e idence
ha ca ageenan was e ec i e [134]. Fu he mo e, in addi ion o he 2 ca ageenan ials o
which IPD was a ailable, 2 u he RCTs epo ed ha nasal ca ageenan was bene icial o
common cold pa ien s [131,132].
Thus, independen o he zinc lozenge ials (Figu es 2 and 3), he e is s ong e idence om
RCTs ha i amin C and nasal io a-ca ageenan can be bene icial o colds.
48

P ac ical implica ions and guidance o u he esea ch
Implica ions o p ac ice
This Coch ane e iew ound a ied e idence o suppo he e ec i eness o zinc in he
p e en ion o ea men o he common cold. Zinc supplemen a ion may p o ide some
limi ed bene i s o people wi h he common cold. On he basis o his e iew, he
cu en e idence is insu icien o p o ide i m conclusions o ecommend zinc
supplemen a ion o he p e en ion o ea men o he common cold [23, p.28].
“Zinc supplemen a ion may p o ide some limi ed bene i s o people wi h he common cold…
… zinc supplemen a ion… ”
As desc ibed abo e, zinc lozenges and nasal zinc a e no o ms o die a y supplemen a ion o
zinc.
As o he RCT esul s in he ials on zinc lozenges, he s a emen “may p o ide some limi ed
bene i ” is an unde s a emen , see Figu es 2 and 3.
The s a emen “some limi ed bene i s” is obscu e. I a educ ion in he du a ion o colds by he
Coch ane es ima e o 2.37-days [23] is oo limi ed o be o p ac ical ele ance, how la ge an
e ec would Naul equi e be o e iewing he e ec as ele an o o dina y common cold
pa ien s.? In compa ison, painkille s a e o en used o help common cold symp oms, bu he e
is no e idence ha hey sho en colds a all.
In any case, wi h a mo e s a is ically- obus analysis (Figu es 2 and 3), Naul could ha e
concluded ha “on he basis o published ials on zinc ace a e and zinc glucona e lozenges,
hey can sho en common colds in adul s on a e age by 37%”. In doing so, he au ho s would
lea e i o he eade o decide whe he he e ec is la ge enough o jus i y he eade o es
zinc lozenges himsel o he sel .
Al hough he zinc lozenge ials ha e implica ions o p ac ice, he eade s should be
in o med o he p oblems o he lozenges on he ma ke [4,67], and should be guided o sea ch
o o mula ions ha a e expec ed o be e ec i e.
See also he sec ion “PICO c i e ia we e no conside ed”
49
I a esea che has ca e ully e iewed a opic, usually i is cons uc i e o sugges di ec ions
o u u e esea ch. Naul concludes:
Implica ions o esea ch
… Fu u e esea che s should no e he need o s anda dised me hods when p o iding
zinc supplemen a ion o he p e en ion o ea men o he common cold. Mo e
esea ch is needed o de e mine he exac zinc ype, dose, and du a ion o
supplemen a ion app op ia e o he p e en ion o ea men o he common cold. Fu u e
di ec ions o esea ch migh be o conside explo ing he subjec i e expe ience o
quali y o li e o pa icipan s while aking zinc, inc easing he di e si y o
cha ac e is ics in he pa icipan s ec ui ed ( a ied ace/e hnici y, age, gende , heal h
s a us), and u he examining ac o s ha may in e e e wi h he bioa ailabili y o zinc
[23, p.29]
This ins uc ion is no pa icula ly help ul o u he esea ch.
“S anda dized me hods” is obscu e wi hou u he de ini ion.
“… o de e mine he exac zinc ype, dose, and du a ion o supplemen a ion app op ia e o
he p e en ion o ea men o he common cold”
Wha ype o zinc sal should ha e p io i y? Wha is a easonable dose? O dina y able s o
zinc lozenges o nasal zinc? Gi en ha Naul has ead he epo s on zinc ials, ques ions
such as hese should be commen ed on.
To my knowledge he e is no e idence indica ing ha o dina y zinc able s ha a e in ended o
be swallowed whole ha e e ec s on colds in he Wes e n adul popula ion. Naul also does no
p o ide any e idence o bene i s o die a y zinc supplemen a ion. Nasal zinc adminis a ion
should no be a high p io i y un il he conce ns abou anosmia ha e been esol ed; see abo e.
The e is s ong e idence ha p ope ly composed zinc lozenges can educe common cold
du a ion as shown in Figu es 2 and 3 and in e e ences [1-14] and he e o e u he esea ch
on zinc lozenges should ha e a high p io i y.
My 2011 and 2017 me a-analyses indica ed ha he e was much s onge e idence o he
e ec s o zinc ace a e lozenges when compa ed wi h zinc glucona e lozenges [7,8], which is
also shown in Figu e 3. The e o e, he highes p io i y should be o u he examine zinc
ace a e lozenges. The e does no seem o be any pa icula bene i om using zinc glucona e
lozenges.
My 2017 dose- esponse analysis did no ind any di e ence in he e ec s o 5 ials using
elemen al zinc doses om 80 o 92 mg/day compa ed wi h 2 ials ha used subs an ially
highe doses o zinc, 192 and 207 mg/day [8]. Howe e , he Tu ne (2000) ial ound no
bene i om 80 mg/day zinc (ins uc ed dose, bu ac ual dose was no epo ed) as zinc
glucona e lozenges [63], and my own RCT wi h 65 mg/day zinc (ac ual used dose) as zinc
ace a e lozenges did no ind any bene i , hough my ial demons a ed he ebound e ec
when he 5-day zinc dosage was e mina ed [66]. On his basis, i seems ob ious ha 200
mg/day elemen al zinc is no needed, bu 80 mg/day may be oo low o lead o consis en
bene i s. The e o e, a dosage o a ound 120 mg/day migh be easonable in u he esea ch
when ying o epea he p e ious posi i e indings. The ea e he lowes e ec i e doses and
op imal lozenge composi ions could be sea ched o .
50
A u he aspec o he zinc lozenges is he ime aken o dissol e in he mou h, which has
been conside ed by Eby [4] and Hemilä [66].
In 2008, Eby w o e abou implica ions o esea ch on zinc lozenges [135]:
Fu u e s udies should ocus on plain zinc ace a e lozenges ha a e chemically iden ical
o he lozenges epo ed o be e ec i e by Pe us e al. [37] in 1998, by P asad e al.
[39] in 2000, and in he upcoming epo by P asad [40]. Such lozenges ha e been
p e iously desc ibed [5] and ha e been highly success ul in ea ing colds bo h in
clinical ials and in he gene al popula ion. Al hough billions o dolla s ha e been spen
on zinc lozenges o ea men o colds, money spen on zinc lozenges con aining
amino, ci ic, o asco bic acids may ha e been was ed.
51
Flaws in da a ex ac ion
I did no sys ema ically check da a ex ac ion o he Naul me a-analysis, bu happened o
no e some subs an ial e o s.
Naul w i es abou he Tu ne (2001) [75] ial:
“Zinc glucona e 1200 mg/day nasal gel” [23, p.142]
Howe e , Tu ne (2001) w o e:
“S udy medica ion… The ac i e p epa a ion (Zicam) consis ed o 33 mM zinc glucona e…
S udy medica ions we e adminis e ed as a single nasal sp ay o 120 μL pe nos il… 5 imes
each day.”
120 μL × 2 nos ils × 5 imes each day = 1200 μL.
Thus, i seems ha Naul con used be ween mg/day and μL.
The ac ual dose is:
0.033 mol/L × 0.00012 L × 2 nos ils × 5 imes each day × 455.7 g/mol = 0.0180 g = 18 mg
zinc glucona e pe day (2.59 mg/day elemen al zinc). Naul ’s Coch ane e iew [23] has 10
au ho s and i is conce ning ha no one o hem ead he manusc ip ca e ully enough o
wonde whe he i is possible o adminis e 1.2 g ams o zinc glucona e nasally pe day.
Naul w i es abou he Hi (2000) [33] ial:
“Zinc glucona e 960 mg/d nasal gel” [23, p.70]
bu on he same page:
“Dose pe day (elemen al zinc in mg): unclea ” [23, p.70]
120 μL × 2 nos ils × 4 imes each day = 960 μL.
Thus, Naul con used also he e be ween mg/day and μL.
Repo ing o he zinc dose in Hi (2000) [33] was no clea , gi ing jus i ica ion o he second
s a emen “unclea ”.
Howe e , in able 2, Cha ac e is ics o in e en ions, Naul gi es a hi d zinc dose in he Hi
(2000) ial:
33 mmol concen a ion/sp ay 1 × sp ay pe nos il 4 × pe day (8 o al doses pe day)
[leading o] “App oxima e zinc dose pe day” o 2.1 mg/d [23, p.191].
Thus, Naul gi es 3 di e en zinc doses o he Hi (2000) ial:
1) 960 mg/d, 2) unclea , and 3) 2.1 mg/d.
In he calcula ion o he 2.1 mg/d, Naul assumed “33 mmol concen a ion/sp ay” bu Hi did
no epo such a zinc concen a ion o hei gel [33]. Ne e heless, he 2.1 mg/day does ha e
indi ec jus i ica ion om he Hi (2000) epo [33, p.778]:
Recen ly, a new app oach o zinc he apy – an o e - he-coun e nasal gel o mula ion
(Zicam) – was he subjec o a p elimina y s udy (C.B. Hensley, PhD, and R. Da idson,
PhD, unpublished da a, 1999). Tha s udy demons a ed ha he di ec applica ion o he
nasal gel con aining 33 mmol/L o ionic zinc wi hin 24 hou s o he onse o common
52
cold symp oms signi ican ly sho ened he du a ion o hose symp oms. The esea che s
hypo hesized ha he deli e y o ionic zinc di ec ly o he si e o in ec ion should be
mo e e ec i e han o al deli e y.
The goal o ou s udy was o a emp o ep oduce hose esul s and independen ly
assess he e ec o zinc nasal gel on he du a ion o common cold symp oms.
Thus Hi did no s a e in hei me hods ha hey used 33 mmol/L zinc gel, bu he
in oduc ion sugges s so. The me hods sec ion o Hi desc ibes 120 μL ou imes pe day o
bo h nos ils, which would lead o 2.07 mg o elemen al zinc, i we assume ha he
concen a ion was 33 mmol/L:
0.033 mol/L × 0.00012 L × 2 nos ils × 4 imes each day × 65.4 g/mol = 2.07 mg/day
elemen al zinc
I did no check h ough all ials included by Naul , bu hese below a e selec ed u he
examples o da a ex ac ion and calcula ion e o s, in addi ion o he wo ials desc ibed
abo e.
Hemilä (2020)
Naul w i es [23, p.67]:
“F equency o dose: 6/d… Dose pe day (elemen al zinc in mg): 78 mg/d”
Howe e , Hemilä (2020) epo ed:
“The zinc lozenge was a comme cially a ailable zinc ace a e lozenge wi h 13 mg elemen al
zinc pe lozenge” [66, p.3]
“O e he pe iod om he second o he i h days, he zinc pa icipan s used on a e age 5.05
lozenges pe day” [66, p. 4]
This gi es he co ec dose o daily elemen al zinc as:
13 mg * 5.05 = 65.6 mg/d
Mossad (1996)
Naul w i es [23, p.98]:
“F equency o dose: 1 lozenge e e y 2 hou s while awake.
Dose pe day (elemen al zinc in mg): 13.3 mg* (max) 12 oz/day = 159.6 mg/d”
Howe e , Mossad (1996) epo ed:
“… lozenges ha weighed 4.4 g and con ained 13.3 mg o zinc.” [45, p.82]
“ he zinc g oup ook an a e age o 6 ± 2 lozenges pe day (median, 5 lozenges pe day)” [45,
p. 84]
This gi es he co ec dose o daily elemen al zinc as:
13.3 mg * 6 = 80 mg/d
Thus, he co ec dose is hal o ha calcula ed by Naul .
This e o has u he consequences in Naul ’s analysis.
53

In Analysis 11.4. (Ad e se e en s), Naul classi ies he Mossad (1996) ial in he “high dose”
>85 mg/day subg oup, whe eas he co ec dose is <85 mg/day and hus he ial should be in
he “low dose” <85 mg/day subg oup.
P asad (2000)
Naul w i es [23, p.113]:
“Dose pe day (elemen al zinc in mg): 12.8 mg x 8 = 102.4 mg (es ima ed numbe o lozenges
pe day)”
Howe e , P asad (2000) epo ed:
“Each lozenge con ained 12.8 mg o zinc.” [39, p.246]
“The a e age numbe o lozenges aken daily was 6.2 in he zinc g oup” [39, p.249]
This gi es he co ec dose o daily elemen al zinc as:
12.8 mg * 6.2 = 79 mg/d
This e o has u he consequences in Naul ’s analysis.
In Analysis 9.3. (T ea men e ec s), Naul classi ies he P asad (2000) ial in he “high dose”
>85 mg/day subg oup, whe eas he co ec dose is <85 mg/day and hus he ial should be in
he “low dose” <85 mg/day subg oup.
In Analysis 11.4. (Ad e se e en s), Naul classi ies he P asad (2000) ial in he “high dose”
>85 mg/day subg oup, whe eas he co ec dose is <85 mg/day and hus he ial should be in
he “low dose” <85 mg/day subg oup.
P asad (2008)
Naul w i es [23, p.119]:
“Dose pe day (elemen al zinc in mg): 13.3 mg*8 = 106.4 mg/d”
Howe e , P asad (2008) epo ed:
“The ac i e lozenges con ained 13.3 mg o zinc” [40, p.796]
“The a e age numbe o lozenges aken daily was 6.9 in he zinc g oup” [40, p.799]
This gi es he co ec dose o daily elemen al zinc as:
13.3 mg * 6.9 = 92 mg/d
54
Macknin (1998) ial wi h child en [38]
In Analysis 9.1, Naul s a es ha he mean du a ion o colds was 8.7 days (SD 8.5 days) in he
zinc lozenge g oup, and 8.7 days (SD 2.8 days) in he placebo g oup; see Figu e 1 o his
documen .
Howe e , in he Cha ac e is ics o Included S udies, Naul [23, p.85] w i es:
No es: du a ion was only epo ed as median ime + CI, so while CI could be con e ed
o SD, he median could no . ‘The median ime o esolu ion o all cold symp oms was
9.0 days (95% CI, 8-9 days) in he placebo g oup and 9.0 days (95% CI, 7-10 days) in
he zinc g oup (P=.71; igu e 2).’ They also epo ed he numbe whose symp oms
esol ed o e he cou se o he 21 days s hose who did no . Tha is epo ed he e.
The e is no desc ip ion whe e he SD 8.5 and 2.8 days o Analysis 9.1 come om.
The dis ibu ion o eco e y in he zinc lozenge and placebo g oups is closely o e lapping and
hus he e canno be a 3- old di e ence in he SD alues be ween he wo ial g oups, see
Figu e 9. Measu ing he cu es o each he IPD, impu ing 18 days o he censo ed
obse a ions a he igh -hand side, gi es he mean du a ion 9.2 days (SD 4.6 days) in he zinc
lozenge g oup, and 9.5 days (SD 4.6 days) in he placebo g oup. These SD alues a e qui e
di e en om he SD alues published by Naul in Analysis 9.1.
Figu e 9. The eco e y cu es in he Macknin (1998) ial wi h child en.
In Analysis 9.1, Naul used SD = 2.8 days in he placebo g oup and SD = 8.5 days in he
zinc g oup [23, p.163]; see Figu e 1 o his documen . Howe e , he dis ibu ions a e
closely simila and hus he e canno be 3- old di e ence in he SD in he wo g oups.
Naul does no desc ibe wha is he sou ce o he SD alues in he Analysis 9.1.
55
Naul also epo s he age in he Macknin (1998) ial as ollows [23, p.83]:
“Age (mean (SD)): 11.67 (7.5)” wi h he mean and SD iden ical in zinc and placebo g oups.
Howe e , Macknin (1998) ial epo s as ollows [38, p.1964]:
“Age, median (IQR), y 13 (6-16) [in Placebo] 13 (6-16) [in Zinc]”
Macknin (1998) did no epo he mean alues and Naul does no desc ibe he sou ce o he
medians.
Naul ’ s able 1: S udy cha ac e is ics.
Desc ip ions o “se ing” a e misleading.
Naul s a es ha he “se ing” o Al-Nakib (1987), Fa (1987), and Tu ne (2001) was
“communi y”. Howe e , all hese ials we e abou expe imen al colds, i.e., colds caused in a
labo a o y wi h hino i us inocula ion (see Table 1 o his documen ). They we e no abou
colds occu ing na u ally in he communi y.
Because expe imen al colds and na u al colds di e subs an ially, i is essen ial o analyze
hem sepa a ely; see abo e. A easonable column o desc ibe he ype o in ec ion is he
“se ings”, bu i is misleading o desc ibe ha expe imen al colds a e communi y colds.
Fo se e al ials, Naul w i es “Clinic/medical cen e” (Belongia 2001; Eby 2006; God ey
1992; Ka asu ya 2012; Ku ugöl 2007; Mossad 2003; P asad 2008), hough he
“Clinic/medical cen e” was pu ely a con ac place o ou pa ien s wi h he communi y-
o igina ed common colds. In mos o he ials he se ing was “Communi y” which mo e
accu a ely desc ibes ha he pa ien s we e ou pa ien s om he communi y. In his espec he
desc ip ions in Naul ’s able 1 a e inconsis en .
56
E o s in o he Coch ane e iews
Coch ane e iews a e o en p omo ed as high-quali y sys ema ic e iews ha can be us ed:
“Coch ane: T us ed e idence. In o med decisions. Be e heal h.”
h ps://www.coch ane.o g (Accessed 2024-9-13).
The Coch ane e iews by Naul e al. (2024) [23] and by Singh and Das (2011/2013/2015)
[15,17,19] a e no he only Coch ane e iews ha ha e been shown o be lawed and ha ha e
misled eade s. Gi en he p omises o Coch ane, i is in o ma i e o look a ce ain o he
Coch ane e iews as hey indica e ha he e may be wide p oblems in he p ocedu es o
e iew and upda e he Coch ane e iews.
In 2021, I ound laws in he Coch ane e iew on Vi amin C and pneumonia (2020) [136],
which alls wi hin he esponsibili y o he Coch ane Acu e Respi a o y In ec ions g oup
along wi h he e iews on zinc and he common cold.
Fo example, he au ho s o he i amin C and pneumonia e iew s a ed:
We included s udies in ol ing: 1. heal hy adul s and child en ecei ing i amin C
supplemen a ion o he p e en ion o pneumonia.
Howe e , wo ials (Coulehan e al. and Bancala i e al.) we e included in he e iew despi e
he ac ha hey we e common cold ials and nei he men ioned pneumonia e en as a
seconda y ou come [137]. Fu he mo e, acco ding o he Coch ane e iew, he numbe o
pneumonia “e en s” in he Bancala i ial was 21 in each g oup. Howe e , ha was no he
numbe o in ec ions, bu he numbe o child en who had ≥1 in ec ions. The e we e 46
in ec ion e en s in he placebo g oup (N=30) and 38 in he i amin C g oup (N=32).
Howe e , none o hese 84 espi a o y in ec ions was pneumonia; hey we e he common cold
[137,138].
In 2009, I ound ha he e we e se e al e o s in he Coch ane e iew Vi amin C
supplemen a ion o as hma (2009) [139]. Fo example, he e we e e o s in da a ex ac ion;
one ial had 20 pa icipan s, bu only 11 pa icipan s we e included in he Coch ane analysis.
The e we e e o s in he calcula ions - he unpai ed - es was used when he pai ed es should
ha e been used. My eedback was published in he (2010) e sion o he “2009” e iew
[140,141].
The managing edi o o he Coch ane Ai ways g oup, Emma Welsh, w o e a esponse o my
c i icism in he (2012) e sion o he “2009” Coch ane e iew [142]. Simul aneously, he
h ee analysis igu es, which we e he main ocus o my c i ique, we e emo ed om he
e iew. In hei absence, howe e , my eedback hough o iginally alid became i ele an .
Fu he mo e, he emaining igu es we e enumbe ed so ha he igu e numbe s in my
eedback indica ed eal bu di e en igu es; he e o e, eade s would conside ha I was
con used. Finally, many esponses by Emma Welsh did no seem alid [143]. I con ac ed
Da id To ey, Edi o -in-Chie o Coch ane, bu ou co espondence did no lead o any
cons uc i e e o s o co ec he p oblem. A e a yea , I con ac ed COPE which s a ed ha
he o iginal (2009) e sion [139] mus be made a ailable o eade s; see a b ie summa y o
he long p ocess [144].
57
75. Tu ne RB. Ine ec i eness o in anasal zinc glucona e o p e en ion o expe imen al hino i us colds.
Clin In ec Dis. 2001;33(11):1865-70.
h ps://doi.o g/10.1086/324347
76. Rao SS, Hendley JO, Hayden FG, Gwal ney JM. Symp om exp ession in na u al and expe imen al
hino i us colds. Am J Rhinol. 1995;9(1):49-52.
h ps://doi.o g/10.2500/105065895781874114
77. Tu ne RB, Wi ek TJ, Rike DK. Compa ison o symp om se e i y in na u al and expe imen ally
induced colds. Am J Rhinol. 1996;10(3):167-72.
h ps://doi.o g/10.2500/105065896781794888
78. Chalme s TC, Celano P, Sacks HS, Smi h H J . Bias in ea men assignmen in con olled clinical ials.
N Engl J Med. 1983;309(22):1358-61.
h ps://doi.o g/10.1056/nejm198312013092204
h ps://www. esea chga e.ne /publica ion/16561928
79. Malik A, Taneja DK, De asenapa hy N, Rajeshwa i K. Zinc supplemen a ion o p e en ion o acu e
espi a o y in ec ions in in an s: a andomized con olled ial. Indian Pedia . 2014;51(10):780-4.
h ps://doi.o g/10.1007/s13312-014-0503-z
80. Higgins JPT, Thomas J, Chandle J, Cumps on M, Li T, Page MJ, Welch VA (edi o s). Coch ane Handbook
o Sys ema ic Re iews o In e en ions e sion 6.4 (upda ed Aug 2023). Coch ane, 2023.
h ps:// aining.coch ane.o g/handbook/cu en /chap e -08#sec ion-8-7
h ps:// aining.coch ane.o g/handbook/cu en /chap e -02 (Accessed 2024-9-13)
81. Meine CL. Clinical T ials: Design, Conduc , and Analysis. Ox o d, 1986, pp.276-277.
h ps://doi.o g/10.1093/acp o :oso/9780195035681.003.0026
82. Ma his S, Le Masson G, Soulages A, Du al F, Ca la L, Valla JM, Solé G. Ol ac ion and anosmia:
F om ancien imes o COVID-19. J Neu ol Sci. 2021;425:117433.
h ps://doi.o g/10.1016/j.jns.2021.117433
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC9755649
83. Suzuki M, Sai o K, Min WP, Vladau C, Toida K, I oh H, Mu akami S. Iden i ica ion o
i uses in pa ien s wi h pos i al ol ac o y dys unc ion. La yngoscope. 2007;117(2):272-7.
h ps://doi.o g/10.1097/01.mlg.0000249922.37381.1e
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC7165544
84. Tian J, Pin o JM, Li L, Zhang S, Sun Z, Wei Y. Iden i ica ion o i uses in pa ien s wi h
pos i al ol ac o y dys unc ion by mul iplex e e se- ansc ip ion polyme ase chain eac ion.
La yngoscope. 2021;131(1):158-64.
h ps://doi.o g/10.1002/la y.28997
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC7436707
85. Ake lund A, Bende M, Mu phy C. Ol ac o y h eshold and nasal mucosal changes
in expe imen ally induced common cold. Ac a O ola yngol. 1995;115(1):88-92.
h ps://doi.o g/10.3109/00016489509133353
86. Sud e CH, Keshe A, G aham MS, e al. Anosmia, ageusia, and o he COVID-19-like symp oms in
associa ion wi h a posi i e SARS-CoV-2 es , ac oss six na ional digi al su eillance pla o ms:
an obse a ional s udy. Lance Digi Heal h. 2021;3(9):e577-e586.
h ps://doi.o g/10.1016/s2589-7500(21)00115-1
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8297994
87. Do y RL. The mechanisms o smell loss a e SARS-CoV-2 in ec ion. Lance Neu ol. 2021;20(9):693-5.
h ps://doi.o g/10.1016/s1474-4422(21)00202-7
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8324112
88. Ro hman KJ, G eenland S. Causa ion and causal in e ence in epidemiology.
Am J Public Heal h. 2005;95 Suppl 1:S144-50.
h ps://doi.o g/10.2105/ajph.2004.059204
89. Ro hman KJ. The w ong message om he w ong alk. Obse a ional S udies. 2020;6(2),30-2.
h ps://doi.o g/10.1353/obs.2020.0006
h ps://muse.jhu.edu/pub/56/a icle/793347/summa y
90. Po ies WJ, Henzel JH, Rob CG, S ain WH. Accele a ion o healing wi h zinc sul a e.
Ann Su g. 1967;165:432-6.
h ps://doi.o g/10.1097/00000658-196703000-00015
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc1617499
91. Ba cia PJ. Lack o accele a ion o healing wi h zinc sul a e. Ann Su g. 1970;172:1048-50.
h ps://doi.o g/10.1097/00000658-197012000-00019
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc1397149
64

92. Se jean GR, Galloway R, Gue i M. O al zinc sulpha e in sickle-cell ulce s. Lance . 1970;2:891-2.
h ps://doi.o g/10.1016/s0140-6736(70)92067-2
93. G ea es MW, Skillen AW. E ec s o long-con inued inges ion o zinc sulpha e in pa ien s
wi h enous leg ulce a ion. Lance . 1970;2:889-91.
h ps://doi.o g/10.1016/s0140-6736(70)92066-0
94. Hallböök T, Lane E. Se um-zinc and healing o enous leg ulce s. Lance . 1972;2:780-2.
h ps://doi.o g/10.1016/s0140-6736(72)92143-5
95. Cze winski AW, Cla k ML, Se a e inides EA, Pe ie C, Hube W. Sa e y and e icacy o
zinc sul a e in ge ia ic pa ien s. Clin Pha macol The . 1974;15(4):436-41.
h ps://doi.o g/10.1002/cp 1974154436
96. Simkin PA. O al zinc sulpha e in heuma oid a h i is. Lance . 1976;2:539–42.
h ps://doi.o g/10.1016/s0140-6736(76)91793-1
97. Lyckholm L, Heddinge SP, Pa ke G, e al. A andomized, placebo con olled ial o o al zinc o
chemo he apy- ela ed as e and smell diso de s. J Pain Pallia Ca e Pha maco he . 2012;26:111–4.
h ps://doi.o g/10.3109/15360288.2012.676618
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc4042409
98. Bam o d JT, Gesse CE, Halle IV, K uge K, Johnson BP. Randomized, double-blind ial o
220 mg zinc sul a e wice daily in he ea men o osacea. In J De ma ol. 2012;51:459–62.
h ps://doi.o g/10.1111/j.1365-4632.2011.05353.x
99. Hadwan MH, Almashhedy LA, Alsalman AR. O al zinc supplemen a ion es o es high molecula weigh
seminal zinc binding p o ein o no mal alue in I aqi in e ile men. BMC U ol. 2012;12:32.
h ps://doi.o g/10.1186/1471-2490-12-32
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc3503568
100. Ala A, Walke AP, Ashkan K, Dooley JS, Schilsky ML. Wilson's disease. Lance . 2007;369(9559):397-408.
h ps://doi.o g/10.1016/s0140-6736(07)60196-2
101. Członkowska A, Li win T, Dusek P, Fe enci P, Lu senko S, Medici V, Rybakowski JK,
Weiss KH, Schilsky ML. Wilson disease. Na Re Dis P ime s. 2018;4(1):21.
h ps://doi.o g/10.1038/s41572-018-0018-3
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc6416051
102. B ewe GJ, Aska i FK. Wilson’s disease: clinical managemen and he apy.
J Hepa ol. 2005;42 (Suppl 1):S13-S21.
h ps://doi.o g/10.1016/j.jhep.2004.11.013
103. Hoogen aad TU. Zinc ea men o Wilson’s disease. J Lab Clin Med. 1998;132:240-1.
h ps://doi.o g/10.1016/s0022-2143(98)90034-8
104. Ma cellini M, Di Ciommo V, Callea F, De i o R, Compa cola D, Sa o elli MR, Ca elli G, Nobili V.
T ea men o Wilson's disease wi h zinc om he ime o diagnosis in pedia ic pa ien s:
a single-hospi al, 10-yea ollow-up s udy. J Lab Clin Med. 2005;145(3):139-43.
h ps://doi.o g/10.1016/j.lab.2005.01.007
105. Coch ane Lib a y. Wha is PICO.
h ps://www.coch anelib a y.com/abou -pico (Accessed 2024-9-13)
106. Hemilä H, Chalke E. Ab up e mina ion o i amin C om ICU pa ien s may inc ease mo ali y:
seconda y analysis o he LOVIT ial. Eu J Clin Nu . 2023;77(4):490-4.
h ps://doi.o g/10.1038/s41430-022-01254-8
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc10115628
107. Hemilä H, Chalke E. Rebound e ec explains he di e gence in su i al a e 5 days
in a con olled ial on i amin C o COVID-19 pa ien s. F on . Med. 2024;11:1391346.
h ps://doi.o g/10.3389/ med.2024.1391346
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc11151746
108. Hemilä H. Vi amin C supplemen a ion and common cold symp oms: p oblems wi h inaccu a e e iews.
Nu i ion 1996;12:804–9.
h ps://doi.o g/10.1016/s0899-9007(96)00223-7
h ps://hdl.handle.ne /10138/225877
109. Hemilä H. Vi amin C, he placebo e ec , and he common cold: a case s udy o
how p econcep ions in luence he analysis o esul s. J Clin Epidemiol. 1996;49(10):1079-84.
h ps://doi.o g/10.1016/0895-4356(96)00189-8
h ps://hdl.handle.ne /10138/225872
110. Hemilä H. To he dissen by Thomas Chalme s. J Clin Epidemiol. 1996;49(10):1087.
h ps://doi.o g/10.1016/0895-4356(96)00191-6
h ps://hdl.handle.ne /10138/225873
65
111. Hemilä H. Do Vi amins C and E A ec Respi a o y In ec ions? Ph.D. Thesis.
Uni e si y o Helsinki, Helsinki, Finland, 2006.
h ps://hdl.handle.ne /10138/20335
h ps://zenodo.o g/doi/10.5281/zenodo.6395595
112. Hemilä H. The ole o i amin C in he ea men o he common cold.
Am Fam Physician. 2007;76(8):1111, 1115.
h ps://www.aa p.o g/pubs/a p/issues/2007/1015/p1111a.h ml
h ps://pubmed.ncbi.nlm.nih.go /17992770
113. Hemilä H. Vi amin C o he common cold should no be ejec ed on he basis o old and e oneous a icles.
J Alle gy Clin Immunol. 2009;124(4):859.
h ps://doi.o g/10.1016/j.jaci.2009.06.015
114. Hemilä H. Randomised ials on i amin C. B J Nu . 2011;105(3):485-7.
h ps://doi.o g/10.1017/s000711451000351x
115. Hemilä H. The e ec o i amin C on he common cold. J Pha m P ac . 2011;24(2):241-2.
h ps://doi.o g/10.1177/0897190010392376
116. Louhiala P, Hemilä H. Can CAM ea men s be e idence-based?
Focus Al e n. Complemen The . 2014;19:84-9.
h ps://doi.o g/10.1111/ c .12110
h ps://hdl.handle.ne /10138/228056
117. Hemilä H. Thomas Chalme s, i amin C and he common cold. J R Soc Med. 2016;109(2):46.
h ps://doi.o g/10.1177/0141076815606279
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC4793763
118. Hemilä H. Vi amin C in clinical he apeu ics. Clin The . 2017;39(10):2110-2.
h ps://doi.o g/10.1016/j.clin he a.2017.08.005
h ps://hdl.handle.ne /10138/228957
119. Hemilä H, Ca A, Chalke E. Vi amin C may inc ease he eco e y a e o ou pa ien cases o SARS-CoV-2
in ec ion by 70%: eanalysis o he COVID A o Z andomized clinical ial.
F on Immunol. 2021;12:674681.
h ps://doi.o g/10.3389/ immu.2021.674681
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8141621
120. Hemilä H, Chalke E. Vi amin C and zinc lozenges o COVID-19? J Am Pha m Assoc. 2021;61(5):e39.
h ps://doi.o g/10.1016/j.japh.2021.05.018
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8163692
121. Hemilä H, Chalke E. Commen a y: The long his o y o i amin C: om p e en ion o he common cold o
po en ial aid in he ea men o COVID-19. F on Immunol. 2021;12:659001.
h ps://doi.o g/10.3389/ immu.2021.659001
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8047412
122. Hemilä H, Chalke E. Bias agains i amin C in mains eam medicine:
examples om ials o i amin C o in ec ions. Li e (Basel) 2022;12(1):62.
h ps://doi.o g/10.3390/li e12010062
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc8779885
123. Hemilä H, Chalke E. Assessmen o i amin C e ec s on pneumonia and COVID-19 using
Mendelian andomiza ion: analysis may be misleading. Eu J Clin Nu 2022;76(9):1347-8.
h ps://doi.o g/10.1038/s41430-022-01091-9
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8853427
124. Pauling L. Asco bic acid and he common cold. Am J Clin Nu . 1971;24(11):1294-9.
h ps://doi.o g/10.1093/ajcn/24.11.1294
125. Pauling L. Asco bic acid and he common cold: E alua ion o i s e icacy and oxici y. Pa I.
Medical T ibune. 1976;17:18-9.
h ps://zenodo.o g/doi/10.5281/zenodo.11546757
126. Pauling L. Asco bic acid and he common cold. Pa II. Medical T ibune. 1976;17:37-8.
h ps://zenodo.o g/doi/10.5281/zenodo.11547088
127. Pauling L. Medical es ablishmen and i amin C. In: Vi amin C, he Common Cold, and he Flu.
F eeman; San F ancisco, CA, USA, 1976, pp.121–138.
128. Pauling L. O ganized medicine and he i amins. In: How o Li e Longe and Feel Be e .
F eeman; San F ancisco, CA, USA, 1986, pp.225–236.
66
129. Fazekas T, Eickho P, P uckne N, Vollnho e G, Fischmeis e G, Diakos C, Rauch M, Ve dianz M, Zoubek
A, Gadne H, Lion T. Lessons lea ned om a double-blind andomised placebo-con olled s udy wi h a io a-
ca ageenan nasal sp ay as medical de ice in child en wi h acu e symp oms o common cold.
BMC Complemen Al e n Med. 2012;12:147.
h ps://doi.o g/10.1186/1472-6882-12-147
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc3575307
130. Ludwig M, Enzenho e E, Schneide S, Rauch M, Boden eich A, Neumann K, P ieschl-G assaue E,
G assaue A, Lion T, Muelle CA. E icacy o a ca ageenan nasal sp ay in pa ien s wi h common cold: a
andomized con olled ial. Respi Res. 2013;14(1):124.
h ps://doi.o g/10.1186/1465-9921-14-124
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc3840586
131. Eccles R, Meie C, Jawad M, Weinmüllne R, G assaue A, P ieschl-G assaue E. E icacy and sa e y o an
an i i al io a-ca ageenan nasal sp ay: a andomized, double-blind, placebo-con olled explo a o y s udy in
olun ee s wi h ea ly symp oms o he common cold. Respi Res. 2010;11(1):108.
h ps://doi.o g/10.1186/1465-9921-11-108
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC2923116
132. Eccles R, Win he B, Johns on SL, Robinson P, T ampisch M, Koelsch S. E icacy and sa e y o io a-
ca ageenan nasal sp ay e sus placebo in ea ly ea men o he common cold in adul s: he ICICC ial.
Respi Res. 2015;16:121.
h ps://doi.o g/10.1186/s12931-015-0281-8
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC4595062
133. Hemilä H, Chalke E. Ca ageenan nasal sp ay may double he a e o eco e y om co ona i us and
in luenza i us in ec ions: Re-analysis o andomized ial da a. Pha macol Res Pe spec . 2021;9(4):e00810.
h ps://doi.o g/10.1002/p p2.810
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC8204093
134. Koenigho e M, Lion T, Boden eich A, P ieschl-G assaue E, G assaue A, Unge H, Muelle CA, Fazekas
T. Ca ageenan nasal sp ay in i us con i med common cold: indi idual pa ien da a analysis o wo
andomized con olled ials. Mul idiscip Respi Med. 2014;9(1):57.
h ps://doi.o g/10.1186/2049-6958-9-57
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC4236476
135. Eby GA. The apeu ic e ec i eness o ionic zinc o common colds. Clin In ec Dis. 2008;46(3):483-4.
h ps://doi.o g/10.1086/527479
136. Padhani ZA, Moazzam Z, Ash a A, Bilal H, Salam RA, Das JK, Bhu a ZA. Vi amin C supplemen a ion o
p e en ion and ea men o pneumonia. Coch ane Da abase Sys Re . 2020;4(4):CD013134.
h ps://doi.o g/10.1002/14651858.cd013134.pub2
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC7192369
137. Hemilä H, Chalke E. Commen a y: Vi amin C supplemen a ion o p e en ion and ea men o pneumonia.
F on Med (Lausanne). 2021;7:595988.
h ps://doi.o g/10.3389/ med.2020.595988
h ps://www.ncbi.nlm.nih.go /pmc/a icles/pmc7854566
Supplemen wi h he de ails o he p oblems:
h ps://www. on ie sin.o g/a icles/10.3389/ med.2020.595988/ ull#supplemen a y-ma e ial
138. Hemilä H, Chalke E. Vi amin C supplemen a ion o p e en ion and ea men o pneumonia (2020)
[Sho comings in a Coch ane e iew]. Zenodo
h ps://zenodo.o g/doi/10.5281/zenodo.11621606
139. Kau B, Rowe BH, A nold E. Vi amin C supplemen a ion o as hma.
[2009 e sion; all igu es p esen , no eedback by Hemilä].
Coch ane Da abase Sys Re . 2009;CD000993.
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC6176494
140. Kau B, Rowe BH, A nold E. Vi amin C supplemen a ion o as hma.
[2010 e sion; bo h he eedback by Hemilä and all he o iginal igu es; he eedback is on p.27].
Coch ane Da abase Sys Re . 2009;CD000993.
h ps://hdl.handle.ne /10138/296440
h ps://zenodo.o g/doi/10.5281/zenodo.10853677
141. Hemilä H. Vi amin C supplemen a ion o as hma [Feedback sepa a ely] e sion 2010.
Uni e si y o Helsinki Open Reposi o y, 2010.
h ps://hdl.handle.ne /10138/296440
h ps://zenodo.o g/doi/10.5281/zenodo.6405775
67
142. Kau B, Rowe BH, S o old E. Vi amin C supplemen a ion o as hma.
[2012 e sion; eedback by Hemilä p esen , 3 igu es emo ed].
Coch ane Da abase Sys Re . 2009;CD000993.
h ps://doi.o g/10.1002/14651858.CD000993.pub3
h ps://zenodo.o g/doi/10.5281/zenodo.10959542
143. Hemilä H. Vi amin C and exe cise-induced b onchocons ic ion: u he p oblems
in he Coch ane e iew “ i amin C o as hma”. Uni e si y o Helsinki Open Reposi o y, 2013.
h ps://hdl.handle.ne /10138/40816
h ps://zenodo.o g/doi/10.5281/zenodo.6405803
144. Hemilä H. Coch ane has no consis en ly ollowed he COPE guidelines.
Eu J Clin In es . 2020;50(4):e13216.
h ps://doi.o g/10.1111/eci.13216
h ps://hdl.handle.ne /10138/326823
145. Kau B, Rowe BH, S o old E. WITHDRAWN: Vi amin C supplemen a ion o as hma.
Coch ane Da abase Sys Re . 2013 Oc 23;2013(10):CD000993.
h ps://pubmed.ncbi.nlm.nih.go /24155047
h ps://doi.o g/10.1002/14651858.CD000993.pub4
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC6176495
146. Bonini M, Di Mamb o C, Calde on MA, Compala i E, Schünemann H, Du ham S, Canonica GW.
Be a₂-agonis s o exe cise-induced as hma. Coch ane Da abase Sys Re . 2013;(10):CD003564.
h ps://doi.o g/10.1002/14651858.cd003564.pub3
h ps://www.ncbi.nlm.nih.go /pmc/a icles/PMC11348701
147. Hemilä H, F ied ich JO. Many con inuous a iables should be analyzed using he ela i e scale: a case
s udy o β2-agonis s o p e en ing exe cise-induced b onchocons ic ion. Sys Re . 2019;8(1):282.
h ps://doi.o g/10.1186/s13643-019-1183-5
Addi ional ile 1 desc ibes he e o s in he Bonini (2013) Coch ane e iew, pp. 12-15:
h ps://s a ic-con en .sp inge .com/esm/a %3A10.1186%2Fs13643-019-1183-5/MediaObjec s/
13643_2019_1183_MOESM1_ESM.pd
148. Hemilä H. Be a2-agonis s o exe cise-induced b onchocons ic ion: da a ex ac ion inconsis encies and
e o s in he Coch ane e iew by Bonini e al. (2013). Zenodo 2024.
h ps://zenodo.o g/doi/10.5281/zenodo.10959750
149. Ioannidis JPA. Coch ane c isis: sec ecy, in ole ance and e idence-based alues.
Eu J Clin In es . 2019;49(3):e13058.
h ps://doi.o g/10.1111/eci.13058
h ps://pubmed.ncbi.nlm.nih.go /30520025
150. Newman M. Has Coch ane los i s way? BMJ. 2019;364:k5302.
h ps://doi.o g/10.1136/bmj.k5302
Cla i ica ion:
h ps://doi.o g/10.1136/bmj.l670
151. Jø gensen L, Gø zsche PC, Je e son T. The Coch ane HPV accine e iew was incomple e and igno ed
impo an e idence o bias. BMJ E id Based Med. 2018;23(5):165‐8.
h ps://doi.o g/10.1136/bmjebm-2018-111012
152. Gø zsche PC. Wha is he mo al collapse in he Coch ane Collabo a ion abou ?
Indian J Med E hics. 2019;4 (NS)(4):303-309.
h ps://doi.o g/10.20529/ijme.2019.064
h ps://pubmed.ncbi.nlm.nih.go /31791931
153. Vespe I. Mass esigna ion gu s boa d o p es igious Coch ane Collabo a ion. Na u e 2018-9-17.
h ps://www.na u e.com/a icles/d41586-018-06727-0 (Accessed 2024-9-13)
154. Nass M, Noble JH J . Whi he Coch ane? Indian J Med E hics. 2018;4(1):1-5.
h ps://doi.o g/10.20529/ijme.2018.092
h ps://pubmed.ncbi.nlm.nih.go /30683645
68