Coch ane
Lib a y
Coch ane Da abase o Sys ema ic Re iews
Vi amin C supplemen a ion o as hma (Re iew)
Kau B, Rowe BH, S o old E
Kau B, RoweBH, S o oldE.
Vi amin C supplemen a ion o as hma.
Coch ane Da abase o Sys ema ic Re iews 2009, Issue 1. A . No.: CD000993.
DOI: 10.1002/14651858.CD000993.pub3.
www.coch anelib a y.com
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
T A B L E O F C O N T E N T S
HEADER......................................................................................................................................................................................................... 1
ABSTRACT..................................................................................................................................................................................................... 1
PLAIN LANGUAGE SUMMARY....................................................................................................................................................................... 2
BACKGROUND.............................................................................................................................................................................................. 3
OBJECTIVES.................................................................................................................................................................................................. 3
METHODS..................................................................................................................................................................................................... 3
RESULTS........................................................................................................................................................................................................ 4
Figu e 1.................................................................................................................................................................................................. 6
Figu e 2.................................................................................................................................................................................................. 7
DISCUSSION.................................................................................................................................................................................................. 8
AUTHORS' CONCLUSIONS........................................................................................................................................................................... 8
ACKNOWLEDGEMENTS................................................................................................................................................................................ 8
REFERENCES................................................................................................................................................................................................ 9
CHARACTERISTICS OF STUDIES.................................................................................................................................................................. 11
DATA AND ANALYSES.................................................................................................................................................................................... 17
Analysis 1.1. Compa ison 1 O al i amin C s placebo (single-dose s udies), Ou come 1 Change in FEV1 (L) - pos -exe cise
challenge................................................................................................................................................................................................
17
Analysis 1.2. Compa ison 1 O al i amin C s placebo (single-dose s udies), Ou come 2 FVC (L) - pos -exe cise challenge........... 17
Analysis 1.3. Compa ison 1 O al i amin C s placebo (single-dose s udies), Ou come 3 PEFR (L/min) - pos -exe cise challenge.... 18
Analysis 2.1. Compa ison 2 O al i amin C s placebo (sho e m s udies), Ou come 1 FEV1 (% d op) pos -exe cise................... 18
Analysis 2.2. Compa ison 2 O al i amin C s placebo (sho e m s udies), Ou come 2 Symp om sco es (As hma Quali y o Li e
Ques ionnnai e).....................................................................................................................................................................................
18
Analysis 3.1. Compa ison 3 O al i amin C s placebo (long- e m s udies), Ou come 1 IgE (IU/ml se um) - absolu e alues......... 19
Analysis 3.2. Compa ison 3 O al i amin C s placebo (long- e m s udies), Ou come 2 FEV1 mL 4 mon hs................................... 19
Analysis 3.3. Compa ison 3 O al i amin C s placebo (long- e m s udies), Ou come 3 Peak Flow (L/min) 4 mon hs.................... 19
Analysis 3.4. Compa ison 3 O al i amin C s placebo (long- e m s udies), Ou come 4 Geome ic mean dec ease in inhaled
co icos e oid use (μg)..........................................................................................................................................................................
20
ADDITIONAL TABLES.................................................................................................................................................................................... 20
FEEDBACK..................................................................................................................................................................................................... 20
WHAT'S NEW................................................................................................................................................................................................. 22
HISTORY........................................................................................................................................................................................................ 22
CONTRIBUTIONS OF AUTHORS................................................................................................................................................................... 23
DECLARATIONS OF INTEREST..................................................................................................................................................................... 23
SOURCES OF SUPPORT............................................................................................................................................................................... 23
INDEX TERMS............................................................................................................................................................................................... 23
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
i
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
[In e en ion Re iew]
Vi amin C supplemen a ion o as hma
Bal inde Kau 1, B ian H Rowe2, Elizabe h S o old3
1Depa men o P ima y Ca e and Public Heal h, Facul y o Medicine, Impe ial College London, London, UK. 2Depa men o Eme gency
Medicine, Uni e si y o Albe a, Edmon on, Canada. 3Popula ion Heal h Sciences and Educa ion, S Geo ge's, Uni e si y o London,
London, UK
Con ac add ess: Emma J Welsh, Popula ion Heal h Sciences and Educa ion, S Geo ge's, Uni e si y o London, C anme Te ace,
Too ing, London, SW17 0RE, UK. [email p o ec ed].
Edi o ial g oup: Coch ane Ai ways G oup.
Publica ion s a us and da e: S able (no upda e expec ed o easons gi en in 'Wha 's new'), commen added o e iew, published in
Issue 8, 2012.
Ci a ion: Kau B, RoweBH, S o oldE. Vi amin C supplemen a ion o as hma. Coch ane Da abase o Sys ema ic Re iews 2009, Issue 1.
A . No.: CD000993. DOI: 10.1002/14651858.CD000993.pub3.
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
A B S T R A C T
Backg ound
Vi amin C is one o he key an ioxidan i amins which is abundan in he ex acellula luid lining he lung and low i amin C in ake has
been associa ed wi h pulmona y dys unc ion.
Objec i es
To e alua e he e idence o he e icacy o i amin C in he ea men o as hma.
Sea ch me hods
The Coch ane Ai ways Re iew G oup as hma egis e was sea ched and bibliog aphies o s udies iden i ied we e also checked o u he
ials. This e iew has been upda ed by sea ches o Augus 2008.
Selec ion c i e ia
Only andomised con olled ials we e eligible o inclusion. S udies we e conside ed o inclusion i hey deal wi h he ea men o
as hma using i amin C supplemen a ion. Two independen e iewe s iden i ied po en ially ele an s udies using p e-de ined c i e ia and
selec ed s udies o inclusion.
Da a collec ion and analysis
Da a we e abs ac ed independen ly by wo e iewe s. In o ma ion on pa ien s, me hods, in e en ions, ou comes and esul s was
ex ac ed using s anda d o ms.
Main esul s
Nine s udies me he e iew en y c i e ia, andomising a o al o 330 pa icipan s. S udy design a ied and he epo ing was gene ally
poo . Fi e ials con ibu ed nume ical da a o he e iew. They p o ided ou come da a on lung unc ion, symp om sco es, IgE le els and
inhaled s e oid use. One small s udy showed a signi ican di e ence in % d op in FEV1 pos -exe cise.
Au ho s' conclusions
A p esen , e idence om andomised-con olled ials is insu icien o ecommend a speci ic ole o i amin C in he ea men o
as hma. Fu he me hodologically s ong and la ge-scale andomised con olled ials a e needed in o de o add ess he ques ion o he
e ec i eness o i amin C in as hma.
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
1
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
P L A I N L A N G U A G E S U M M A R Y
Vi amin C supplemen a ion o as hma
As hma is a ch onic in lamma o y disease o he ai ways cha ac e ised by wheeze and b ea hlessness. One heo y o he obse ed inc ease
in he numbe o people wi h as hma is he 'wes e n' die wi h i 's lack o nu ien s om esh ood. We e iewed e idence om nine ials
o he an ioxidan i amin C as a ea men o as hma. In gene al he ials we e small, a ied g ea ly in hei design and he epo ing was
poo . F om he a ailable e idence i is no possible o ecommend ei he he use o a oidance o i amin C supplemen s in as hma.
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
2
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
B A C K G R O U N D
As hma is now ecognised as a ch onic in lamma o y disease
esul ing in e e sible ai ways b onchocons ic ion (Holga e 1990).
The incidence and p e alence o as hma has inc eased in many
coun ies o e he pas ew decades. The mos ma ked inc ease
has been obse ed in child en (S achan 1999; Lewis 1996). This
may be due o changing en i onmen al exposu es o he inc eased
suscep ibili y o popula ions wi h educed hos esis ance (Sea on
1994).
One hypo hesised cause o his inc ease is ha changes in
"wes e n" die ha e p oduced a educ ion in hos esis ance
o e ime. In pa icula ecen in e es has ocused on he
associa ion be ween an i-oxidan s in die s and heal h ou comes.
C oss-sec ional s udies show ha in equen ui consump ion is
associa ed wi h educed lung unc ion, bo h in child en (Cook 1997)
and adul s (Bu land 1999). In addi ion, he Na ional Food Su ey
has documen ed a d op in he consump ion o an ioxidan ood
sou ces such as esh ui and ege ables in coun ies such as G ea
B i ain since he 1950's (Sea on 1994).
Vi amin C is one o he key an ioxidan i amins which is abundan
in he ex acellula luid lining he lung. Low i amin C in ake
is associa ed wi h pulmona y dys unc ion (Schwa z 1994). Bo h
adul s (Olusi 1979) and child en (Ade ele 1985) wi h as hma ha e
been ound o ha e lowe concen a ions o i amin C when
compa ed o no mal subjec s. Pa ien s wi h as hma may ha e low
supplies o i amin C o an inc eased demand o i amin C in he
ace o an oxidan load esul ing in deple ion. The e is a need o
cla i y whe he supplemen a ion wi h i amin C may b ing bene i s
in educing mo bidi y, imp o ing pulmona y unc ion o quali y o
li e in pa ien s wi h as hma.
The e ha e been h ee ecen e iews o he li e a u e on he ole o
i amin C in as hma (Bielo y 1994; Ha ch 1995; Mon eleone 1997).
Howe e , he e iew by Bielo y e al (Bielo y 1994) only sea ched
he English language li e a u e using MEDLINE and ga e no u he
de ails as o how he s udies had been loca ed. The o he wo
e iews did no speci y hei me hodology. These e iews eached
di e en conclusions. Bielo y e al (Bielo y 1994) concluded ha
he ole o i amin C in as hma was unclea and ha cu en
li e a u e did no suppo i s use. The e iew by Ha ch e al
(Ha ch 1995) ound 7 ou o 11 s udies indica ed ha i amin
C supplemen a ion migh e e se o imp o e as hma symp oms.
The e iew by Mon eleone e al (Mon eleone 1997) o e s he
opinion ha i amin C p o ides a sho e m p o ec i e e ec on
ai way esponsi eness, bu less clea impac on o he objec i e
lung unc ion measu emen s. All h ee e iews ecommend u he
s udies in o he ole o i amin C in as hma. As a i s s ep, a e iew
using he Coch ane me hodology is needed o sys ema ically weigh
he quali y o he exis ing e idence be o e ecommending any
u u e s udies.
O B J E C T I V E S
To de e mine he o e all e icacy o i amin C supplemen a ion in
pa ien s wi h s able ch onic as hma.
M E T H O D S
C i e ia o conside ing s udies o his e iew
Types o s udies
To be eligible, all s udies needed o be andomised-con olled ials
(RCTs). Double-blinded ials we e p e e ed, bu single blind and
open s udies we e also e iewed o possible inclusion.
Types o pa icipan s
S udies we e conside ed o inclusion i hey ec ui ed adul s and/
o child en wi h ch onic s able as hma, seasonal as hma o hose
wi h exe cise-induced b onchospasm. S udies o o he alle gic
condi ions such as hay e e , alle gic hini is and eczema we e only
conside ed i he esul s o subjec s wi h as hma we e p esen ed
sepa a ely. Vi amin C s udies, which epo ed ou comes on pa ien s
wi h as hma sepa a ely as a sub-g oup, we e also conside ed o
inclusion.
Types o in e en ions
Vi amin C supplemen a ion compa ed o placebo o "s anda d
ca e". We conside ed s udies ha adminis e ed i amin C ia any
ou e, dosage o dose in e al. Bo h single dose and longe - e m
s udies we e conside ed o inclusion.
Types o ou come measu es
P ima y ou come measu es
1) Lung unc ion (e.g. FEV1, PEFR)
2) Symp oms (e.g. symp om sco es)
Seconda y ou come measu es
3) Func ional ou comes (e.g. quali y o li e, sickness absence,
exe cise capaci y)
4) Non-speci ic b onchial hype - eac i i y (BHR) o his amine o
me hacholine
5) Immunological ma ke s (IgE le els)
6) As hma medica ion equi emen s (e.g. addi ional s e oid o
b onchodila o usage)
7) Heal h se ice u ilisa ion (e.g. GP a endance, hospi al
admissions)
8) As hma exace ba ions
Sea ch me hods o iden i ica ion o s udies
Elec onic Sea ches
T ials we e iden i ied using he Coch ane Ai ways G oup
Specialised Regis e o ials, which is de i ed om sys ema ic
sea ches o bibliog aphic da abases including he Coch ane
Cen al Regis e o Con olled T ials (CENTRAL), MEDLINE, EMBASE,
CINAHL, AMED and PsycINFO, and handsea ching o espi a o y
jou nals and mee ing abs ac s (please see he Ai ways G oup
Module o u he de ails). All eco ds in he Specialised Regis e
coded as 'as hma' we e sea ched using he ollowing e ms:
asco bic* o " i amin c" o an ioxid*
The mos ecen sea ch was conduc ed in Augus 2008.
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
3
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
O he sou ces
Re e ence lis s o all p ima y s udies and e iew a icles we e
e iewed o addi ional e e ences. Au ho s o iden i ied ials we e
con ac ed.
Da a collec ion and analysis
Re ie al o s udies
All ials ha appea ed po en ially ele an we e assessed by wo
independen e iewe s o ele ance using abs ac and i le om
he elec onic sea ch. Using ex s om all po en ially ele an
a icles, inal inclusion was also de e mined independen ly by wo
e iewe s. Disag eemen s abou s udy inclusion we e esol ed wi h
discussion.
Assessmen o me hodological quali y
The quali y o included s udies was assessed using he Coch ane
app oach. A isk o bias able was comple ed o each s udy
assessing he epo ing o me hod o andomisa ion, alloca ion
concealmen and blinding. Each i em was judged as being
adequa e, unclea o inadequa e. Any disag eemen s be ween
e iewe s we e esol ed by discussion.
Da a abs ac ion
Da a we e ex ac ed independen ly by wo e iewe s and en e ed
in he Coch ane Collabo a ion So wa e, Re iew Manage (Re Man).
Fo s udies whe e he o iginal da a we e no p esen ed, when
possible, hey we e ex ac ed om g aphically ep esen a ions.
S udy ou comes ha we e epo ed pos -b onchial challenge
(e.g. exe cise o his amine) we e analysed sepa a ely om
ou comes ha did no in ol e b onchial challenge o when esul s
we e epo ed be o e such challenges (bu pos -dose i amin C
adminis a ion).
S a is ical conside a ions
Ou comes om included ials we e combined using Re Man. Fo
con inuous ou comes he weigh ed mean di e ence (WMD) wi h
ixed e ec was used o es ima e he indi idual e ec sizes and
95% con idence in e als (95% CI). I he e we e any dicho omous
ou comes he Pe o ixed o andom e ec model was o be used o
es ima e he pooled odds a io (OR) and 95% CI.
The main planned compa ison o s a is ical conside a ion was
any o m o dose o i amin C supplemen a ion e sus placebo
o "s anda d ca e". I he e we e adequa e included s udies, he
ollowing p e-de ined sub-g oup analysis we e planned:
1. Single dose e sus ch onic adminis a ion o i amin C
2. Die a y ad ice o inc ease i amin C consump ion e sus no
in e en ion
3. O al i amin C e sus in a enous i amin C supplemen a ion
4. Adul s e sus child en
5. Males e sus emales
R E S U L T S
Desc ip ion o s udies
De ails o he sea ch his o y and esul s can be ound in Table 1.
Fo he 2008 upda e, 39 new e e ences we e iden i ied and he ull
ex o 15 o hese we e e ie ed. One addi ional included s udy
(Tecklenbu g 2007), an ex ension o a p e iously included s udy
(Foga y 2003), and 5 u he excluded s udies we e iden i ied. The
e iew now con ains a o al o nine s udies which mee he inclusion
c i e ia. Fu he de ails can be ound in he able, "Cha ac e is ics
o included s udies". A o al o wel e s udies we e excluded
a e examining he ull- ex pape . Please see "Cha ac e is ics o
excluded s udies" o u he de ails.
The included s udies we e conduc ed in he USA (Ko dansky 1979;
Schach e 1982,Tecklenbu g 2007), Nige ia (Anah 1980), Sou h
A ica (Ande son 1983), Canada (Malo 1986), Is ael (Cohen 1997)
and he UK (Foga y 2003; O'Sulli an 2000).
Th ee s udies examined he impac o i amin C supplemen a ion
on exe cise challenge es s in subjec s wi h a con i med
diagnosis o exe cise-induced as hma (Cohen 1997; Schach e
1982, Tecklenbu g 2007). Two s udies examined he impac o
i amin C adminis a ion on b onchial hype esponsi eness o
his amine challenge es s in pa icipan s wi h as hma (Malo 1986;
O'Sulli an 2000). Fou s udies examined he impac o i amin
C on b onchial hype esponsi eness o alle gen challenge in
subjec s sensi i e o agweed alle gen (Ko dansky 1979), equency
o as hma exace ba ions due o in ec ion (Anah 1980), lung unc ion
and immunological ma ke s in pa ien s wi h as hma (Ande son
1983) and clinical con ol o as hma in p ima y ca e (Foga y 2003).
Th ee o he s udies (Anah 1980; Ande son 1983; Foga y 2003)
ollowed a pa allel s udy design and he emaining 6 used c osso e
designs. Da a om he wo ypes o s udy designs we e p esen ed
sepa a ely in Re Man. No usable da a could be ex ac ed om
he epo s o ou s udies (Anah 1980;Ko dansky 1979; Malo 1986;
O'Sulli an 2000), despi e a emp s a au ho con ac .
Se en s udies (Anah 1980; Foga y 2003; Ko dansky 1979; Malo
1986; O'Sulli an 2000; Schach e 1982, Tecklenbu g 2007) in ol ed
adul pa ien s, one s udy (Ande son 1983) in ol ed only child en
and one (Cohen 1997) had bo h adul s and child en. The smalles
s udy had six pa icipan s (Ko dansky 1979). O he s anged om
8 o 41. The la ges was Foga y 2003 wi h 210 pa icipan s. This
e iew con ains a o al o 330 andomised pa icipan s.
All ea men s we e adminis e ed o ally, ei he as able s o
as an o al solu ion. Th ee s udies (Anah 1980; Ande son 1983;
Foga y 2003) ea u ed long- e m supplemen a ion wi h 1 g i amin
C daily o 14 weeks, 6 mon hs and 16 weeks, espec i ely.
Ano he long- e m s udy (Ko dansky 1979) used 500 mg i amin
C supplemen a ion daily o se en days. One s udy looked a
supplemen a ion wi h 1500mg o e a sho - e m pe iod o wo
weeks (Tecklenbu g 2007) and he emaining ou s udies (Cohen
1997; Malo 1986; O'Sulli an 2000; Schach e 1982) used single
doses o i amin C 2g, 2g, 2g and 500 mg, espec i ely.
Two o he c osso e s udies (Cohen 1997, Tecklenbu g 2007)
men ioned a washou pe iod. None o he o he c osso e s udies
(Ko dansky 1979; Malo 1986; O'Sulli an 2000; Schach e 1982)
epo ed a washou pe iod. I has been sugges ed ha a e a single
o al dose o i amin C, a leas 1-2 days is equi ed o exc e ion
depending on p e-exis ing body le els (Ba es 2001).
Included s udies epo ed dispa a e ou come measu es, which
made agg ega ion o he pu pose o a me a-analysis di icul . Mos
o he s udies did no epo he ac ual da a in he published pape s
o did no p o ide su icien da a o a me a-analysis, al hough
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
4
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
a emp s we e made o con ac he au ho s o da a. Ou come
measu es included a a ie y o lung unc ion es s, symp oms and
symp om sco es, immune ma ke s and educ ion in he use o
inhaled s e oids.
Risk o bias in included s udies
In gene al, he epo ing quali y o he s udies was poo . All he
included s udies we e epo ed as being andomised, howe e , only
one s udy (Foga y 2003) epo ed he me hod o andomisa ion
and jus h ee o he nine s udies epo ed he me hod o alloca ion
concealmen (Anah 1980, Foga y 2003, Malo 1986). Fou s udies
epo ed he me hod o blinding (Anah 1980, Malo 1986, Schach e
1982, Tecklenbu g 2007), while one s udy was inadequa ely blinded
(Ande son 1983) and he emaining ou we e unclea .
None o he s udies adequa ely epo ed all h ee me hods o
andomisa ion, alloca ion concealmen o blinding. Anah 1980,
Foga y 2003 and Malo 1986 ga e he mos de ailed accoun . An
o e iew o ou judgmen s o a e p esen ed in Figu e 1 and Figu e 2.
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
5
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Figu e 1. Me hodological quali y summa y: e iew au ho s' judgemen s abou each me hodological quali y i em
o each included s udy.
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
6
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Figu e 2. Me hodological quali y g aph: e iew au ho s' judgemen s abou each me hodological quali y i em
p esen ed as pe cen ages ac oss all included s udies.
E ec s o in e en ions
O he nine included s udies, i e ha e con ibu ed nume ical
da a o he e iew: Ande son 1983, Cohen 1997, Foga y 2003,
Schach e 1982 and Tecklenbu g 2007. Fou s udies (Anah 1980;
Ko dansky 1979; Malo 1986; O'Sulli an 2000) did no epo da a
in an manne ha pe mi ed u he analysis and au ho con ac
has been unsuccess ul. Howe e , none o hese s udies ound a
signi ican di e ence o he e ec o i amin C on lung unc ion o
symp oms.
S udies could no be combined s a is ically, because hose which
add essed simila compa isons used di e en in e en ions o
ou come a iables. Fo example, he e we e h ee s udies whe e
he p o ec i e e ec s o i amin C we e in es iga ed using exe cise
challenge. All h ee epo ed pulmona y unc ion ou comes, bu
Cohen 1997 only epo ed esul s om 11 o he 20 andomised
pa icipan s (who we e ound o ha e a all in FEV1 o less
han 15% ollowing i amin C), Schach e 1982 epo ed absolu e
change pos -exe cise and Tecklenbu g 2007 was a wo week
in e en ion s udy a he a single dose s udy, and epo ed
maximum pe cen age all om baseline.
P ima y ou comes
Lung unc ion
Single dose s udies
Change in FEV1 (L) - pos -exe cise challenge: (MD 0.13; 95% CI
-0.05 o 0.31) (, Schach e 1982). Cohen epo ed ha 11 o he 20
pa icipan s had a change o FEV1 a e exe cise o less han 15%
ollowing i amin C bu all 20 pa icipan s had a change o 15% o
mo e ollowing placebo. Mean changes we e only p esen ed o he
11 pa icipan s who had be e esul s wi h i amin C, and hese
a e no included as his is a biased es ima e o he ue expec ed
ea men e ec .
Change in FVC (L) - pos -exe cise challenge: (MD 0.13; 95% CI -0.03
o 0.29) (Schach e 1982)
Change in PEFR (L/min) - pos -exe cise challenge: (MD 0.49; 95% CI
-0.07 o 1.05) (Schach e 1982)
Sho e m s udies
FEV1 (%) d op pos -exe cise: a signi ican di e ence was shown in
a ou o i amin C: (MD 6.50%; 95% CI 0.05 o 12.95) (Tecklenbu g
2007)
Long e m s udies
FEV1 mL a ou mon hs:(MD -11.00; 95% CI -91.36 o 69.36) (Foga y
2003)
Peak low L/min (mo ning and e ening) a 4 mon hs: (Mean
di e ence am 0.90; 95% CI -11.74 o 13.54) and (Mean di e ence pm
2.20; 95% CI -9.95 o 14.35) (Foga y 2003)
Symp om Sco es
One s udy (Tecklenbu g 2007) epo ed da a on symp om sco es
(As hma Quali y o Li e Ques ionnai e). The e was no signi ican
di e ence (MD 0.50; 95% CI -0.24 o 1.24).
Seconda y ou comes
IgE (IU/ml se um) - absolu e alues a one mon h: (MD 4.00; 95% CI
-140.42 o 148.42) (Ande son 1983)
IgE (IU/ml se um) - absolu e alues a h ee mon hs: (MD -312.00;
95% CI -628.21 o 4.21) (Ande son 1983)
IgE (IU/ml se um) - absolu e alues a six mon hs: (MD -143.00; 95%
CI -425.38 o 139.38)(Ande son 1983)
Dec ease in inhaled co icos e oid use (µg): no signi ican
di e ence (Foga y 2003)
The e we e no da a om any o he included s udies o heal h
se ice u ilisa ion.
Da a on acu e exace ba ions was p o ided by one s udy (Anah
1980). The e we e nine exace ba ions in he in e en ion g oup
which had 22 pa ien s and 35 exace ba ions in he placebo g oup
which had 19 pa ien s. Thus some pa ien s had mo e han one
exace ba ion. Da a conce ning he numbe o pa ien s who had
one o mo e exace ba ions would be mo e meaning ul since i
would no be subjec ed o bias om a ew pa ien s wi h ecu en
exace ba ions.
Ad e se e ec s
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
7
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Ou comes PD20 FEV1, PD35 SGaw, es ed on day 7, 3h s a e dose o placebo/ i amin C
No es Da a no p esen ed in an abs ac able o ma , no eply om au ho o da e.
Risk o bias
Bias Au ho s' judgemen Suppo o judgemen
Random sequence gene a-
ion (selec ion bias)
Unclea isk Me hod o andomisa ion no desc ibed
Alloca ion concealmen
(selec ion bias)
Unclea isk Me hod o alloca ion no desc ibed
Blinding (pe o mance
bias and de ec ion bias)
All ou comes
Unclea isk S udy desc ibed as double blind bu me hod o blinding (e.g. iden ical placebo
pill) no desc ibed
Ko dansky 1979(Con inued)
Me hods Randomised double blind placebo con olled c osso e s udy.
Pa icipan s 16 adul s (3M; 13F) wi h as hma ha me he ATS c i e ia. Age ange 19-59, mean 43.1 (SD 7.7) y s, mean
du a ion o as hma 10.5 y s (SD 14.6).
In e en ions The subjec s we e s udied on 4 di e en days. Subjec s ecei ed ea men o placebo which consis ed
o 250ml o a anspa en and odou less swee liquid in which was dissol ed ei he 2g asco bic acid o
placebo. One hou la e spi ome y measu ed and his amine challenge done un il PC20 eached.
Ou comes FEV1, FVC, PC20
No es Da a no p esen ed in an abs ac able o ma . Values epo ed o di e en days a he han di e en
g oups. No eply om au ho o da e.
Risk o bias
Bias Au ho s' judgemen Suppo o judgemen
Random sequence gene a-
ion (selec ion bias)
Unclea isk Me hod o andomisa ion no desc ibed
Alloca ion concealmen
(selec ion bias)
Low isk Concealmen o ea men s we e done using codes. The o al solu ions we e
p epa ed by hospi al pha macy.
Blinding (pe o mance
bias and de ec ion bias)
All ou comes
Low isk O al solu ions we e desc ibed as being o simila as e
Malo 1986
Me hods Randomised double-blind, c oss-o e placebo con olled s udy.
O'Sulli an 2000
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
14
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Pa icipan s Ten mild (ATS c i e ia) as hma ic pa icipan s.
In e en ions Each pa icipan comple ed wo ea men pe iods wi h inges ion o ei he 2g o asco bic acid o place-
bo 45 minu es p io his amine b onchop o oca ion.
Ou comes Spi ome y was measu ed be o e, du ing and a e he his amine challenges.
No es Abs ac only published, da a no p esen ed in an abs ac able o ma , no eply om au ho o da e.
Risk o bias
Bias Au ho s' judgemen Suppo o judgemen
Random sequence gene a-
ion (selec ion bias)
Unclea isk Me hod o andomisa ion no desc ibed
Alloca ion concealmen
(selec ion bias)
Unclea isk Me hod o alloca ion no desc ibed
Blinding (pe o mance
bias and de ec ion bias)
All ou comes
Unclea isk S udy desc ibed as double blind bu me hod o blinding (e.g. iden ical placebo
pill) no desc ibed
O'Sulli an 2000(Con inued)
Me hods Randomised double-blind con olled ial, c osso e design.
Pa icipan s 12 adul s (5 male, 7 emale) wi h exe cise-induced as hma, ne e on co icos e oids o admi ed o hos-
pi al.
In e en ions Single dose o 500 mg i amin C o ally o suc ose placebo. S udy done on 2 subsequen days. 90 min-
u es pos does subjec s unde wen exe cise challenge. No washou indica ed. Exe cise challenge in in-
c emen al wo kload and un il subjec s hea a e each 170bpm o he subjec a igued. Pulmona y
unc ion was measu ed be o e & a e o al dose and a e exe cise.
Ou comes FVC, FEV1, PEFR be o e and a e exe cise challenge on cyclegome e o 170bpm o exhaus ion.
No es
Risk o bias
Bias Au ho s' judgemen Suppo o judgemen
Random sequence gene a-
ion (selec ion bias)
Unclea isk Me hod o andomisa ion no desc ibed
Alloca ion concealmen
(selec ion bias)
Unclea isk Me hod o alloca ion no desc ibed
Blinding (pe o mance
bias and de ec ion bias)
All ou comes
Low isk Iden ical placebo capsule
Schach e 1982
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
15
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Me hods Randomised, double-blind, c osso e ial o e 5 consecu i e weeks
Pa icipan s Eigh pa icipan s (2 male, 6 emale) wi h physician diagnosed mild- o-mode a e as hma and docu-
men ed exe cise-induced b onchocons ic ion. Pa icpan s we e ec ui ed om Uni e si y popula ion
and he local communi y and we e ac i e.
In e en ions Asco bic acid supplemen 1500mg/day (3x500mg capsules) o placebo (suc ose) (3 capsules). Manu-
ac u ed by NOW Foods. Pa icpan s we e andomised o ac i e ea men o placebo o wo weeks.
The e was a wash-ou pe iod o one week and hen he pa icipan s c ossed o e .
Pa icpan s we e ad ised o a oid oods ha we e high in i amin C du ing he s udy.
Ou comes Pulmona y unc ion (FEV1) p e-and pos -exe cise; Exhaled ni ic oxide (FENO) p e- and pos -exe cise;
symp om ques ionnai e
No es S udy conduc ed in he USA
Risk o bias
Bias Au ho s' judgemen Suppo o judgemen
Random sequence gene a-
ion (selec ion bias)
Unclea isk Me hod o andomisa ion no desc ibed
Alloca ion concealmen
(selec ion bias)
Unclea isk Me hod o alloca ion concealmen no desc ibed.
Blinding (pe o mance
bias and de ec ion bias)
All ou comes
Low isk Ma ching placebo manu ac u ed by he same company as he ac i e ea -
men .
Tecklenbu g 2007
Cha ac e is ics o excluded s udies [o de ed by s udy ID]
S udy Reason o exclusion
Cuomo 2004 S udy used a combined supplemen o i amin C, i amin E and o he an ioxidan s
Fo as ie e 2000 No a andomised con olled ial (be o e and a e ques ionnai e su ey).
G ozdjako a 2005 S udy used a combined supplemen o i amin C, Coenzyme Q10 and α- ocophe ol.
Konge ud 2003 No a andomised con olled ial and in e en ion no es ing e icacy o i amin C. S udy exam-
ined he le els o asco bic acid in induced spu um o as hma ic pa ien s compa ed o heal hy ol-
un ee s.
Mi ic 1991 No a andomised con olled ial. All subjec s gi en placebo i s han ollowed la e wi h all pa-
ien s ecei ing i amin C.
Mohsenin 1983 S udy did no ha e a placebo a m (only had be o e and a e e ec s o i amin C adminis a ion).
Mohsenin 1987 S udy used heal hy subjec s and excluded subjec s who had as hma.
Mu phy 2002 S udy used a combina ion o i amin C and α- ocophe ol
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
16
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
S udy Reason o exclusion
Omenaas 2003 Pos al ques ionnai e no a andomised con olled ial.
Panina 2002 S udy used a complex o o al an ioxidan s. I is no clea ha he s udy was andomised.
Romieu 2002 S udy used bo h i amin C and E in he in e en ion g oup.
Ting 1983 S udy did no ha e a placebo a m and was no andomised (only had be o e and a e e ec s o i -
amin C).
D A T A A N D A N A L Y S E S
Compa ison 1. O al i amin C s placebo (single-dose s udies)
Ou come o subg oup i le No. o
s udies
No. o
pa ici-
pan s
S a is ical me hod E ec size
1 Change in FEV1 (L) - pos -exe cise chal-
lenge
1 Mean Di e ence (IV, Fixed, 95% CI) To als no selec ed
2 FVC (L) - pos -exe cise challenge 1 Mean Di e ence (IV, Fixed, 95% CI) To als no selec ed
3 PEFR (L/min) - pos -exe cise challenge 1 Mean Di e ence (IV, Fixed, 95% CI) To als no selec ed
Analysis 1.1. Compa ison 1 O al i amin C s placebo (single-dose
s udies), Ou come 1 Change in FEV1 (L) - pos -exe cise challenge.
S udy o subg oup Vi amin C g oup Placebo g oup Mean Di e ence Mean Di e ence
N Mean(SD) N Mean(SD) Fixed, 95% CI Fixed, 95% CI
Schach e 1982 12 0.2 (0.2) 12 0.1 (0.3) 0.13[-0.05,0.31]
Fa ou s Placebo
10.5-1 -0.5 0
Fa ou s Vi amin C
Analysis 1.2. Compa ison 1 O al i amin C s placebo (single-
dose s udies), Ou come 2 FVC (L) - pos -exe cise challenge.
S udy o subg oup Vi amin C g oup Placebo g oup Mean Di e ence Mean Di e ence
N Mean(SD) N Mean(SD) Fixed, 95% CI Fixed, 95% CI
Schach e 1982 12 0.1 (0.2) 12 -0 (0.2) 0.13[-0.03,0.29]
Fa ou s Placebo
10.5-1 -0.5 0
Fa ou s Vi amin C
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
17
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Analysis 1.3. Compa ison 1 O al i amin C s placebo (single-
dose s udies), Ou come 3 PEFR (L/min) - pos -exe cise challenge.
S udy o subg oup Vi amin C g oup Placebo g oup Mean Di e ence Mean Di e ence
N Mean(SD) N Mean(SD) Fixed, 95% CI Fixed, 95% CI
Schach e 1982 12 0.6 (0.5) 12 0.1 (0.8) 0.49[-0.07,1.05]
Fa ou s Placebo
42-4 -2 0
Fa ou s Vi amin C
Compa ison 2. O al i amin C s placebo (sho e m s udies)
Ou come o subg oup i le No. o
s udies
No. o
pa ici-
pan s
S a is ical me hod E ec size
1 FEV1 (% d op) pos -exe cise 1 Mean Di e ence (Fixed, 95% CI) To als no selec ed
2 Symp om sco es (As hma Quali y o Li e
Ques ionnnai e)
1 Mean Di e ence (Fixed, 95% CI) To als no selec ed
Analysis 2.1. Compa ison 2 O al i amin C s placebo (sho e m s udies), Ou come 1 FEV1 (% d op) pos -exe cise.
S udy o subg oup Expe imen al Con ol Mean Di -
e ence
Mean Di e ence Mean Di e ence
N N (SE) IV, Fixed, 95% CI IV, Fixed, 95% CI
Tecklenbu g 2007 0 0 6.5 (3.29) 6.5[0.05,12.95]
Fa ou s placebo
10050-100 -50 0
Fa ou s i amin C
Analysis 2.2. Compa ison 2 O al i amin C s placebo (sho e m s udies),
Ou come 2 Symp om sco es (As hma Quali y o Li e Ques ionnnai e).
S udy o subg oup Vi amin C Placebo Mean Di -
e ence
Mean Di e ence Mean Di e ence
N N (SE) IV, Fixed, 95% CI IV, Fixed, 95% CI
Tecklenbu g 2007 0 0 0.5 (0.38) 0.5[-0.24,1.24]
Fa ou s Placebo
10050-100 -50 0
Fa ou s Vi amin C
Compa ison 3. O al i amin C s placebo (long- e m s udies)
Ou come o subg oup i le No. o
s udies
No. o
pa ici-
pan s
S a is ical me hod E ec size
1 IgE (IU/ml se um) - absolu e al-
ues
1 Mean Di e ence (IV, Fixed, 95% CI) To als no selec ed
1.1 1 mon h 1 Mean Di e ence (IV, Fixed, 95% CI) 0.0 [0.0, 0.0]
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
18
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
Ou come o subg oup i le No. o
s udies
No. o
pa ici-
pan s
S a is ical me hod E ec size
1.2 3 mon hs 1 Mean Di e ence (IV, Fixed, 95% CI) 0.0 [0.0, 0.0]
1.3 6 mon hs 1 Mean Di e ence (IV, Fixed, 95% CI) 0.0 [0.0, 0.0]
2 FEV1 mL 4 mon hs 1 Mean Di e ence (Fixed, 95% CI) To als no selec ed
3 Peak Flow (L/min) 4 mon hs 1 Mean di e ence (Fixed, 95% CI) To als no selec ed
3.1 Mo ning 1 Mean di e ence (Fixed, 95% CI) 0.0 [0.0, 0.0]
3.2 E ening 1 Mean di e ence (Fixed, 95% CI) 0.0 [0.0, 0.0]
4 Geome ic mean dec ease in in-
haled co icos e oid use (μg)
1 Mean Di e ence (IV, Fixed, 95% CI) To als no selec ed
Analysis 3.1. Compa ison 3 O al i amin C s placebo (long-
e m s udies), Ou come 1 IgE (IU/ml se um) - absolu e alues.
S udy o subg oup Vi amin C g oup Placebo g oup Mean Di e ence Mean Di e ence
N Mean(SD) N Mean(SD) Fixed, 95% CI Fixed, 95% CI
3.1.1 1 mon h
Ande son 1983 7 246 (124.9) 9 242 (169.7) 4[-140.42,148.42]
3.1.2 3 mon hs
Ande son 1983 7 211 (93.1) 9 523 (472.4) -312[-628.21,4.21]
3.1.3 6 mon hs
Ande son 1983 7 253 (147) 9 396 (398.8) -143[-425.38,139.38]
Fa ou s Vi amin C
1000500-1000 -500 0
Fa ou s Placebo
Analysis 3.2. Compa ison 3 O al i amin C s placebo (long- e m s udies), Ou come 2 FEV1 mL 4 mon hs.
S udy o subg oup Vi amin C Placebo Mean Di -
e ence
Mean Di e ence Mean Di e ence
N N (SE) IV, Fixed, 95% CI IV, Fixed, 95% CI
Foga y 2003 95 106 -11 (41) -11[-91.36,69.36]
Fa ou s placebo
1000500-1000 -500 0
Fa ou s Vi amin C
Analysis 3.3. Compa ison 3 O al i amin C s placebo (long- e m s udies), Ou come 3 Peak Flow (L/min) 4 mon hs.
S udy o subg oup Vi amin C Placebo Mean di -
e ence
Mean di e ence Mean di e ence
N N (SE) IV, Fixed, 95% CI IV, Fixed, 95% CI
3.3.1 Mo ning
Fa ou s placebo
10050-100 -50 0
Fa ou s Vi amin C
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
19
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
S udy o subg oup Vi amin C Placebo Mean di -
e ence
Mean di e ence Mean di e ence
N N (SE) IV, Fixed, 95% CI IV, Fixed, 95% CI
Foga y 2003 95 106 0.9 (6.45) 0.9[-11.74,13.54]
3.3.2 E ening
Foga y 2003 95 106 2.2 (6.2) 2.2[-9.95,14.35]
Fa ou s placebo
10050-100 -50 0
Fa ou s Vi amin C
Analysis 3.4. Compa ison 3 O al i amin C s placebo (long- e m s udies),
Ou come 4 Geome ic mean dec ease in inhaled co icos e oid use (μg).
S udy o subg oup Vi amin C Placebo Mean Di e ence Mean Di e ence
N Mean(SD) N Mean(SD) Fixed, 95% CI Fixed, 95% CI
Foga y 2003 29 49 (156.6) 32 11 (37.5) 38[-20.46,96.46]
Fa ou s placebo
10050-100 -50 0
Fa ou s i amin C
A D D I T I O N A L T A B L E S
Sea ch da es Resul s
Janua y 2001 Thi y- i e abs ac s we e iden i ied om he sea ch o he Coch ane Ai ways G oup egis e , o
which 6 me he inclusion c i e ia (Anah 1980, Ande son 1983, Cohen 1997, Ko danksy 1979, Malo
1986, Schach e 1987) and 5 we e added as excluded s udies.
Janua y 2001 - Ap il 2004 Twen y- ou abs ac s we e iden i ied by he upda ed sea ch. Two addi ional included s udies
(Foga y 2003, O'Sulli an 2000) and 4 excluded s udies we e added o he e iew.
Table 1. Sea ch his o y
F E E D B A C K
Feedback submi ed by Ha i Hemila, 24 Ma ch 2009
Summa y
"The Coch ane e iew i amin C o as hma (2009 e sion) has e o s in he ex ac ion o da a and in he analysis.
Schach e 1982 ca ied ou a ial wi h pa icipan s who had exe cise-induced b onchocons ic ion (EIB) so ha each o he 12 pa icipan s
was adminis e ed placebo and i amin C a di e en imes. Thus, each pa icipan se ed as his o he own con ol (c oss-o e ). In Table
III Schach e epo ed p e-pos -exe cise change o FEV1, so ha he la e FEV1 was measu ed 5 minu es a e he exe cise. Because wo
obse a ions a e measu ed om he same pa icipan , he placebo pe iod and i amin C pe iod di e ence in FEV1 change should be
analysed using he pai ed - es . The FEV1 da a in Schach e 's Table III gi es he mean di e ence be ween he i amin C and placebo pe iods
as 0.20 (SD 0.33) li es/s. Schach e 1982 calcula ed = 2.13 in hei pape , co esponding o P[1- ail] = 0.028.
The e iew p esen s Schach e 's FEV1 changes in Analysis 1.2. Howe e , da a in Analysis 1.2 we e ex ac ed om Schach e 's Table II, which
p esen s pos -exe cise FEV1 alue measu ed immedia ely a e he exe cise. In EIB he all in FEV1 occu s 5 o 20 minu es a e he end o
exe cise (Rundell 2009), and e en Schach e epo ed ha , on he sc eening day, he e was no all in FEV1 immedia ely a e exe cise, bu
a signi ican all 5 minu es a e he exe cise (Schach e 1982 Fig. 2). The e o e, ex ac ing he FEV1 changes om Schach e 's Table II (FEV1
immedia ely a e he exe cise) is no easonable i he pu pose is o examine he e ec o i amin C on EIB.
Cohen 1997 ca ied ou an EIB ial wi h 20 pa icipan s who we e adminis e ed placebo and i amin C a di e en imes (c oss-o e ). Pos -
exe cise FEV1 was measu ed 8 minu es a e he end o he exe cise. The obse a ions a e pai ed also in his case and he esul s should
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
20
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
be analysed using a pai ed es . 9 pa icipan s had FEV1 dec ease >15% on bo h i amin C and placebo ea men s. 11 pa icipan s had
>15% FEV1 dec ease on placebo bu <15% FEV1 dec ease on i amin C (Cohen 1997 Fig. 2). None o he pa icipan s had he opposi e
e ec : <15% FEV1 dec ease on placebo and >15% FEV1 dec ease on i amin C. In he pai ed 2x2 able analysis, he ques ion is whe he he
di e ence be ween he co ne s (he e 11 and 0) is s a is ically signi ican . This di e ence gi es z = (11-0)/sq (11+0) = 3.31, co esponding
o P[1- ail] = 0.0005.
A basic p inciple in con olled ial analysis equi es ha all andomised pa icipan s should be included in he analysis ( he ITT p inciple).
Howe e , he e iew does no gi e he esul s o all o Cohen's 20 pa icipan s (Cohen 1997 Fig. 2); Analysis 1.2 gi es he esul s o only
he 11 pa icipan s who had bene i o i amin C (Cohen 1997 Table 2).
Fu he mo e, he e iew p esen s he a e age o pos -exe cise FEV1 alues and no he p e-pos -exe cise di e ence in FEV1 in analysis 1.2.
The pos -exe cise a e ages o Cohen's Table 2 a e 1.66 (SD 0.80) li es/s in he placebo pe iod and 1.93 (SD 0.78) li es/s in he i amin C
pe iod (P = 0.42). Howe e , gi en ha he EIB is de ined by he p e-pos change in FEV1, he measu emen o he e ec on EIB should be
based on he p e-pos -exe cise di e ence in FEV1 (Rundell 2009). Fu he mo e, he ela i e e ec calcula ed by Cohen (Table 2; in %uni s) is
a be e measu e han he absolu e alue (in li es/s) because he ela i e e ec adjus s o he g ea a ia ion in baseline FEV1; he ela i e
dec ease in FEV1 is also used in guidelines (Rundell 2009). Cohen epo s ha he a e age ela i e all in FEV1 is 25% in he placebo pe iod
and 5% in he i amin C pe iod (Cohen 1997 able 2). Because he obse a ions a e pai ed, he pai ed - es should be used. The a e age o
he di e ences is 20% (SD 12%, SE 3.7%), which gi es = 5.57, co esponding o P[1- ail] = 0.00012. Thus, al hough he e iew p esen ed
only he 11 pa icipan s in which i amin C was bene icial, he calcula ion sugges s ha e en in his subg oup i amin C was wi hou e ec
(P = 0.42), whe eas a co ec calcula ion gi es a much smalle P- alue.
In hei EIB ial, Tecklenbu g 2007 s udied 8 pa icipan s who we e adminis e ed i amin C and placebo a di e en imes. They measu ed
pos -exe cise FEV1 a 1, 5, 10, 15, 20, and 30 min a e he exe cise. Tecklenbu g 2007 epo ed ha he dec ease in FEV1 in he i amin C
pe iod was 6.4% (SE 2.4%) and dec ease in he placebo pe iod was 12.9% (SE 2.4%). Tecklenbu g did no publish he pai ed compa ison,
no o iginal da a so ha he pai ed - es could be calcula ed. Ne e heless, hese a e ages gi e unpai ed = 1.91, co esponding o P[1-
ail] = 0.038, which is conse a i e, he pai ed es P- alue would be smalle .
Thus, h ee ials included in he e iew ound bene i o i amin C supplemen a ion agains EIB a 5 and 8 minu es a e he exe cise (Cohen
1997; Schach e 1982), o a he ime o maximum all in FEV1 (Tecklenbu g 2007). The h ee P- alues calcula ed abo e (0.028, 0.0005,
0.038) can be combined by using he Fishe me hod (Fishe 1948). The combined P[1- ail] = 0.00007 p o ides e idence ha he e ec s o
i amin C on EIB in hese h ee ials a e no explained by andom luc ua ions.
Analyses 1.1, 1.3 and 1.5 p esen baseline da a o wo EIB ials discussed abo e (Cohen 1997; Schach e 1982). Howe e , when a ial
speci ically examines he e ec o i amin C on EIB, he ele an ou come is he di e ence be ween he baseline and he 5-10 minu es pos -
exe cise FEV1 alues ( he p e-pos change), and no he baseline FEV1 alue alone.
Finally, diagnosis o EIB by he change in FEV1 is well es ablished (Rundell 2009) and he au ho s should ha e conside ed whe he he e is
any bene i o eade s om making addi ional analyses o he FVC and PEFR alues o he oldes ial by Schach e 1982. The mo e ecen
ials by Cohen 1997and Tecklenbu g 2007 did no epo changes in FVC and PEFR."
Reply
This commen on he ials ela ing o exe cise induced b onchocons ic ion (EIB) was submi ed in Ma ch 2009 and published alongside
he e iew in No embe 2010.
We hank D Hemilä o he eedback, bu do no hink ha he echnical issues aised o e he analysis o da a om he h ee small c oss-
o e ials (including a o al o 40 pa icipan s), subs an i ely al e he s eng h o di ec ion o he esul s, he quali y o he e idence, o
he conclusions o he e iew.
We ag ee ha c osso e ials a e bes analysed using pai ed - es s, bu do no ag ee wi h he p esen a ion o one- ail P alues abo e.
A wo- ailed pai ed - es did no show a s a is ically signi ican di e ence in change in FEV1 ei he immedia ely a e exe cise (shown in
analysis 1.2) o i e minu es la e in Schach e 1982 (P = 0.18 and 0.057 om Table II and Table III espec i ely). The e o e he au ho 's
choice no o include he la e obse a ion does no mislead he eade in ou opinion.
We ag ee ha he mean di e ences in FEV1 epo ed om only 11 o he 20 pa icipan s in Cohen 1997 should no be included in he e iew,
and his has been emo ed om he analyses.
The au ho s en e ed da a o he all in FEV1 om Tecklenbu g 2007, using a s anda d e o de i ed om a conse a i e es ima e o he
P alue based on he pai ed - es ( epo ed in he pape as P < 0.05). We see no compelling eason o o e u n his app oach since he
a e age e ec is unal e ed and he da a come om a s udy o only eigh pa icipan s.
We ag ee ha he baseline lung unc ion is no a use ul ou come o his e iew and ha e emo ed he p e-exe cise ou comes.
We do no ag ee wi h he sugges ed app oach o combining P alues om Cohen 1997, Schach e 1982 and Tecklenbu g 2007 in iew o he
clinical he e ogenei y be ween he s udies and ou comes unde conside a ion. Such an app oach ocuses a en ion on whe he any e ec
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
21
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
obse ed is a ibu able o chance. This is i sel po en ially misleading since i does no ake accoun o he magni ude o e ec ac oss
he s udies. The analyses p esen ed in he e iew ha e now been amended so ha only mean di e ences and con idence in e als o he
s udies a e p esen ed, and no associa ed P alues.
We a e con en o eade s o conside he commen om D Hemilä alongside ou esponse, and o make up hei own minds ega ding
he au ho s' app oach o he analysis o da a and he conclusions o he e iew.
Pos ed by Emma Welsh, Mangaging Edi o o he Coch ane Ai ways G oup, on behal o he au ho and edi o ial eams.
Con ibu o s
Ha i Hemilä, Depa men o Public Heal h, Uni e si y o Helsinki, Helsinki, Finland
W H A T ' S N E W
Da e E en Desc ip ion
13 June 2012 Amended Feedback inco po a ed,
We a e awa e o a new ele an s udy, his has been added o
s udies awai ing classi ica ion.
13 June 2012 Feedback has been inco po a ed In ligh o he eedback, we ha e emo ed h ee ins ances o e-
po ing o baseline lung unc ion alues and dele ed s a is ical
da a om a ial who only epo ed da a on pa icipan s who
bene i ed om ea men . These changes ha e no al e ed he
conclusions o he e iew and we do no belie e he e iew will
mislead he eade .
13 June 2012 Re iew decla ed as s able The me hods used in he e iew a e somewha ou da ed and
he e o e a new e iew is equi ed in his opic. Applica ions o
egis e his i le will be subjec o ou p io i isa ion p ocedu e.
H I S T O R Y
P o ocol i s published: Issue 3, 1996
Re iew i s published: Issue 1, 1999
Da e E en Desc ip ion
5 No embe 2010 Amended Feedback has been published alongside he o iginal ex o he
e iew.
1 Decembe 2008 Amended Con ac de ails o B Kau al e ed
29 Oc obe 2008 New ci a ion equi ed bu conclusions
ha e no changed
One new included s udy, one ex ension o a p e iously included
s udy (Foga y 2003) and i e excluded s udies we e iden i ied.
Conclusions emain unchanged. Change in au ho ship.
29 Augus 2008 New sea ch has been pe o med New sea ch.
14 Augus 2008 Amended Con e ed o new e iew o ma .
1 Ap il 2004 New ci a ion equi ed and conclusions
ha e changed
Subs an i e amendmen
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
22
Coch ane
Lib a y
T us ed e idence.
In o med decisions.
Be e heal h.
Coch ane Da abase o Sys ema ic Re iews
C O N T R I B U T I O N S O F A U T H O R S
Felix Ram and BK conduc ed he o iginal e sion o his e iew in 2001. BR was he assigned edi o con ibu ing o he p o ocol and e iew
edi ing. FR upda ed he e iew in Ap il 2004. ES upda ed he e iew in Augus 2008.
D E C L A R A T I O N S O F I N T E R E S T
The e a e no known con lic s o in e es .
S O U R C E S O F S U P P O R T
In e nal sou ces
•S Geo ge's, Uni e si y o London, UK.
Ex e nal sou ces
•No sou ces o suppo supplied
I N D E X T E R M S
Medical Subjec Headings (MeSH)
*Die a y Supplemen s; An ioxidan s [* he apeu ic use]; Asco bic Acid [* he apeu ic use]; As hma [*d ug he apy]; Randomized
Con olled T ials as Topic
MeSH check wo ds
Humans
Vi amin C supplemen a ion o as hma (Re iew)
Copy igh © 2012 The Coch ane Collabo a ion. Published by John Wiley & Sons, L d.
23
