RESEARCH ARTICLE Open Access
Fac o s associa ed wi h in-hospi al
mo ali y om communi y-acqui ed
pneumonia in Po ugal: 2000–2014
Ezequiel Pessoa
1*
, C is ina Bá ba a
2,3,4
, Lau a Viegas
1
, And eia Cos a
1,3
, Ma ilde Rosa
3
and Paulo Noguei a
3,5
Abs ac
Backg ound: Communi y-acqui ed pneumonia (CAP) is one o he leading causes o mo bidi y and mo ali y
wo ldwide, o en leading o hospi al admissions. In Po ugal, he ac o s associa ed wi h in-hospi al mo ali y due o
CAP a e no ully documen ed. The aim o his s udy was o cha ac e ize he ends o CAP hospi aliza ion in all age
g oups and he ac o s associa ed wi h hei mo ali y be ween 2000 and 2014.
Me hods: We conduc ed a c oss-sec ional s udy using CAP hospi aliza ion da a in all age g oups, in Po ugal
Mainland. Logis ic eg ession was used o iden i y he ac o s associa ed wi h in-hospi al mo ali y.
Resul s: Be ween 2001 and 2011, CAP hospi aliza ion a e inc eased om 2.8 o 4.3 pe 1000 popula ion.
Hospi aliza ion a es we e highe in he ex eme ages ( ≤4 and ≥75 yea s). Howe e , a dec ease in he
hospi aliza ion a e and i s mo ali y was obse ed, in he younge ages.
A o al o 548,699 hospi aliza ion CAP episodes, be ween 2000 and 2014, we e analyzed, wi h male (56.2%) and
elde ly ≥65 yea s (91.7%) p edominance, esul ing in 101,740 dea hs (18.5%). Men had a signi ican ly lowe mean
age (64.3 ± 26.4 yea s e sus 67.9 ± 27.5 yea s; p< 0.001). Du ing he s udied 15 yea s, he e was an inc ease o 45.2%
in he numbe o annual hospi aliza ions, concomi an wi h he admission inc ease o indi iduals aged o e 75
yea s. Since 2012 a dec ease in hospi aliza ions and associa ed dea hs we e de ec ed.
The inc ease in age ep esen ed a p og essi e and signi ican ise in he p obabili y o dea h, excep o he age
g oup 1–4 yea s. The age g oup ≥85 yea s old (Adjus ed OR = 124.256; 95%CI: 97.838–157.807) and males (Adjus ed
OR = 1.261; 95%CI: 1.243–1.280) we e signi ican ly associa ed wi h dea h isk o CAP hospi aliza ion. A e 2010, his
isk dec eased (Adjus ed OR = 0.961; 95%CI: 0.940–0.982). The main ac o s a ec ing mo ali y we e age, sex,
unemploymen a e, numbe o pe o med p ocedu es and admission quinquennia.
Conclusions: Despi e a end o dec ease in CAP hospi aliza ions and associa ed dea h since 2012, he numbe s o
in-hospi al mo ali y showed, in he 15 yea s unde analysis, an o e all inc ease o e ime, mainly associa ed wi h
age, in pa icula e y old people ( ≥75 yea s), males and a highe pa ish unemploymen a e. The e o e, he
implemen a ion o CAP p e en i e measu es should be ein o ced in hese ulne able g oups.
Keywo ds: Communi y-acqui ed pneumonia, In-hospi al mo ali y, Po ugal
© The Au ho (s). 2020 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0
In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o
he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e
(h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed.
* Co espondence: [email p o ec ed]
1
Escola Supe io de En e magem de Lisboa, Lisbon, Po ugal
Full lis o au ho in o ma ion is a ailable a he end o he a icle
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18
h ps://doi.o g/10.1186/s12890-019-1045-x
In oduc ion
Backg ound
Communi y-acqui ed pneumonia (CAP) is one o he
main causes o mo bidi y and mo ali y wo ldwide [1,2].
Globally, pneumonia, oge he wi h o he lowe espi a-
o y ac in ec ions a e he leading cause o dea h om
in ec ious disease causing 3.0 million dea hs wo ldwide
and being he ou h cause o dea h om all causes in
2016 [2]. CAP mainly a ec s he elde ly in de eloped
coun ies and child en in de eloping coun ies [2]. Also,
in he Eu opean Union (EU) pneumonia emains he
mos equen cause o dea h om in ec ion and was e-
sponsible o nea ly 140,000 dea hs in 2015, accoun ing
o o e 30% o all espi a o y disease mo ali y [3].
The e a e la ge a ia ions in mo ali y a es ac oss EU
coun ies: Po ugal, he Slo ak Republic and he Uni ed
Kingdom ha e he highes a es o pneumonia mo ali y
[3]. Despi e his epidemiological in o ma ion, i is di i-
cul o de e mine he eal bu den o CAP in Eu ope, be-
cause p ecise da a a e sca ce [4]. The highes incidence
o CAP occu s among young child en and olde adul s
[5], whe eas mo ali y is highes among olde adul s [6–
8]. Many CAP cases equi e hospi aliza ion and some
s udies epo admission a es be ween 20 and 30% [9,
10]. Se e al s udies ha e shown ha he incidence a es
o hospi aliza ion and he isk o dea h, a y wi h age,
wi h he highes a es occu ing in he mos ex eme age
g oups, mainly o e 75 yea s old [1,10].
Popula ions wo ldwide a e expec ed o age signi i-
can ly in he coming decades. In Eu ope, 25% o he
popula ion is al eady aged 60 yea s o o e and ha
p opo ion is p ojec ed o each 35% in 2050 and 36% in
2100 [11]. This scena io implies ha CAP ela ed ou -
come may con inue o ha e a high clinical and economic
impac by he expec ed inc ease o hospi al admissions,
cos s and po en ially dea hs [9–11].
Despi e many s udies ca ied ou , in Eu ope, i emains
di icul o unde s and disc epancies be ween coun ies and
geog aphic egions [4,12–16].Theanalyseso da a om
one o a ew hospi als ha e he po en ial o in oducing
bias.Thislimi a ion,howe e ,canbeo e comebyusing
da a om la ge na ional da abase samples, making possible
amo e obus analysisabou he ealbu deno CAP.
In Po ugal, he e a e some s udies ha highligh he
e olu iona y pe spec i e o CAP hospi aliza ions on
mainland hospi als o Na ional Heal h Se ice, be ween
2000 and 2009 [13,17,18]. Howe e , in hese s udies,
he phenomenon was explo ed, a decade ago and only in
he adul popula ion.
The p esen s udy aimed o cha ac e ize he ends o
CAP hospi aliza ion episodes in all age g oups and o
iden i y he ac o s associa ed wi h hei mo ali y,
du ing a 15-yea pe iod, be ween 2000 and 2014, in
Po ugal Mainland.
Me hods
S udy design and da a collec ion
A c oss-sec ional s udy was conduc ed using da a om
2000 o 2014, on hospi aliza ions due o CAP. CAP
hospi aliza ion was de ined as an episode wi h he p in-
cipal diagnosis coded as 480–486, om he In e na ional
Classi ica ion o Diseases (ICD), 9 h Re ision Clinical
Modi ica ion (ICD-9-CM). Pa ien s, in whom pneumo-
nia was no he main diagnosis, we e excluded. Da a
we e ob ained om he Na ional Hospi al Discha ge
Da abase owned and managed by he Cen al Adminis-
a ion o he Heal h Sys em o he Po uguese Minis y
o Heal h. This da abase con ains adminis a i e and
clinical da a o all hospi al admissions o he public Na-
ional Heal h Se ice (NHS) in Po ugal (up o 2014, i
only included Po ugal Mainland public hospi als).
These da a we e eco ded h ough he WebGDH com-
pu e applica ion (a webpage applica ion wi h es ic ed
access) and coded acco ding o diagnoses and pe o med
p ocedu es, du ing he hospi al s ay, by specially ained
heal h s a using he ICD-9-CM.
S udy popula ion encompassed indi iduals om all age
g oups, in Po ugal Mainland. Fo each case, in o ma ion
was ob ained conce ning yea o hospi al admission, age,
sex, unemploymen a e o he pa ish o o igin o he
pa ien ( alue eco ded acco ding o he in o ma ion om
he demog aphic census su ey o he Po uguese popula-
ion [19,20], ch onologically close o he hospi aliza ion
episode), leng h o s ay (LOS), numbe o p ocedu es
pe o med du ing hospi aliza ion, and discha ge ou come.
The anonymi y o he subjec s was main ained in all he
analyzed da a, hus in o med consen was no equi ed.
The s udy was app o ed by he E hics Commi ee o he
Lisbon Facul y o Medicine (107/19).
Pa ien s ‘pa ish unemploymen a e was calcula ed as
he numbe o unemploymen pe sons pe 100 ac i e
pe sons egis e ed in he ci il pa ish o o igin o he
pa ien , acco ding o census in o ma ion ch onologically
close o he admission episode (2001 Census and 2011
Census) [19,20].
S a is ical analysis
Sociodemog aphic and clinical a iables we e analyzed
h ough desc ip i e s a is ics. A compa a i e analysis
was pe o med o iden i y associa ions be ween a iables
o in e es h ough S uden ’s - es o independen
samples in nume ical a iables o al e na i ely he non-
pa ame ic U Mann-Whi ney es . Fo he compa a i e
analysis o he ca ego ical a iables, he Chi-squa e and
he Fishe exac es we e used. The s a is ical analysis
assumed a signi icance le el o 5%.
To iden i y he ac o s associa ed wi h in-hospi al mo al-
i y a bina y logis ic eg ession was used. All a iables whose
bi a ia e analysis (Odds Ra io) e ealed s a is ical in e es
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 2 o 10
(p< 0.20) we e included in a mul i ac o ial analysis. The
a iables in he inal model we e chosen using he o wa d
condi ional me hod, ob aining he adjus ed Odds Ra io.
The quali y o he adjus men o he mul i ac o ial
model was calcula ed by he a ea unde he ROC cu e
(Recei e Ope a ing Cha ac e is ic).
To compa e he e olu ion o hospi aliza ion episodes
o e ime, a hospi aliza ion a e pe 1000 popula ion,
was calcula ed as he a io be ween he numbe o
hospi aliza ion episodes wi h he main diagnosis o CAP
and he numbe o inhabi an s o he same yea , mul i-
plied by 1000. The yea s o 2001 and 2011 we e chosen
because i was possible o de e mine accu a ely he
o icial numbe o he esiden popula ion in Po ugal
Mainland, based on ha yea census’su eys o he
Po uguese popula ion [19,20].
S a is ical analysis was pe o med using he S a is ical
Package o he Social Sciences (SPSS) e sion 23 (SPSS
Inc., Chicago, IL, USA).
Resul s
T ends in CAP hospi aliza ion a e
Based on da a om 2001 and 2011 census, conce ning
he popula ion esiden in Po ugal Mainland [19,20], a
CAP popula ion based hospi aliza ion a e was de e -
mined, o hese wo- ime poin s. Ac oss his pe iod,
CAP hospi aliza ion a e inc eased om 2.8 o 4.3 pe
1000 popula ion, conce ning he o e all age g oups
(Table 1). In bo h ime poin s, and ega ding he s udied
age g oups, hospi aliza ion a es we e highe in he
ex eme age g oups ( ≤4 and ≥75 yea s). Howe e , when
compa ing 2001 wi h 2011, i should be no ed ha a
dec ease in CAP hospi aliza ion a e was de ec ed in he
age g oup 0–4 yea s. In adul s, in bo h ime poin s, he
hospi aliza ion a e a ia ion inc eased p og essi ely
wi h age, wi h special ele ance in he elde ly o e 75
yea s, highligh ing he obse ed inc ease in he absolu e
numbe o hospi aliza ions in 2011, compa ed o 2001.
Be ween 2001 and 2011, in he age g oup ≥65 yea s old,
he CAP hospi aliza ion a e inc eased om 10.3 ( ange:
5.2–17.6) o 16.5 ( ange: 6.3–27.6) pe 1000 popula ion.
T ends in CAP hospi aliza ions and in-hospi al mo ali y
Be ween 2000 and 2014, a o al o 548,699 hospi aliza ion
episodes due o CAP we e obse ed in Po ugal Mainland
(Figs. 1and 2). Du ing his ime pe iod, he e was an in-
c ease in he numbe o annual hospi aliza ion episodes,
om 29,065 in 2000 o 42,213 in 2014, co esponding o
an o e all inc ease o 45.2%. In he analysis by age g oup,
unlike o he global end o inc eased hospi aliza ions, a
pa e n o dec eased hospi aliza ions was de ec ed in he
g oup o child en unde 4 yea s o age. We also highligh
he inc ease obse ed in CAP hospi aliza ions in he age
g oups o e 75 yea s (Fig. 2). Howe e , since 2012, a sligh
dec ease in he numbe o hospi al admissions was
de ec ed in people o e 65 yea s old.
Du ing he 15-yea s udy pe iod, he mean age o CAP
hospi alized pa ien s inc eased om 58.1 ± 29.7 yea s o
70.7 ± 24.0 yea s (da a no shown), wi h 70% o
hospi aliza ion episodes occu ing o e 65 yea s old,
mainly in he age g oup ≥75 yea s old (Table 2).
Conce ning he absolu e numbe o dea hs pe yea ,
he e was a s eady inc ease un il 2012, ollowed by a de-
c ease in he wo consecu i e ollowing yea s (Fig. 1).
The analysis by age g oup showed ha , in 2008, he
dea hs in he popula ion wi h mo e han 85 yea s old
exceeded o he i s ime, he numbe o dea hs in he
age g oup 75–84 yea s, becoming since hen, he age
g oup wi h he highes numbe o in-hospi al dea hs
(Fig. 3). Un il he end o he pe iod unde analysis, des-
pi e an obse ed dec ease in he hospi al mo ali y since
2012, he gap be ween he absolu e numbe o dea hs, in
he age g oups o e 85 and 75–84 yea s, was s ill
Table 1 CAP hospi aliza ion a e based on census popula ion and in-hospi al mo ali y, by age g oups, in 2001 and 2011
Age
G oup
Numbe o
hospi aliza ion
episodes
CAP hospi aliza ion
a e pe 1000
popula ion
Ra io o CAP
hospi aliza ion
a es
In-hospi al mo ali y p- alue*
% (n)
2001 2011 2001 2011 2011/2001 2001 2011
0–4 3418 2465 6.7 5.4 0.805 0.3% (11) 0.2% (4) < 0.001
5–14 1163 1097 1.1 1.1 0.964 0.9% (10) 0.4% (4) 0.376
15–24 444 469 0.3 0.4 1.344 5.2% (23) 2.4% (11) < 0.001
25–44 2013 2036 0.7 0.7 1.029 4.9% (98) 5.6% (113) 0.314
45–64 3562 5393 1.5 2.0 1.331 9.4% (333) 10.5% (568) < 0.001
65–74 4910 6311 5.2 6.3 1.214 15.2% (746) 15.8% (995) < 0.001
> 75 11,906 25,571 17.6 27.6 1.562 25.1% (2986) 26.7% (6835) < 0.001
To al 27,416 43,342 2.84.3 1.550 15.4% (4207) 19.7% (8530) < 0.001
*p- alue e e s o he compa ison o in-hospi al mo ali y p opo ion
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 3 o 10
inc easing, seconda y o a highe dec ease o dea hs in
he age g oup 75–84 yea s (Fig. 3).
Mo e han a qua e o CAP in-hospi al mo ali y
occu ed wi hin 72 h. In ac , 10.7% o he mo ali y was
obse ed wi hin 24 h, 27.5% wi hin 72 h, and 35.9%
wi hin 4 days.
Table 2p esen s he compa a i e analysis o in-
hospi al mo ali y acco ding o socio-demog aphic and
heal h a iables.
In-hospi al mo ali y, o he o e all pe iod unde ana-
lysis, was 18.5%, wi h a p edominance o males (56.2%)
and o he age g oups ≥65 yea s (91.7%).
The numbe o hospi alized men was almos always
highe han ha obse ed in women h oughou he li e
cycle, excep in he popula ion o e 85 yea s, whe e
women ep esen ed 56.9% (da a no shown). Men had a
signi ican ly lowe mean age (64.3 ± 26.4 yea s e sus
67.9 ± 27.5 yea s; p< 0.001).
Ac oss ime, he pe cen age o dea hs inc eased o e
each 5 yea s’ ime pe iods. Deceased pa ien s we e sig-
ni ican ly olde han su i ing pa ien s (80.2 e sus 65.9
yea s). Also, deceased pa ien s we e submi ed o a
g ea e numbe o p ocedu es and p esen ed a sligh ly
lowe LOS (Table 2).
Fig. 1 Numbe o CAP hospi al admissions and dea hs pe yea , be ween 2000 and 2014
Fig. 2 E olu ion o he numbe o CAP hospi al admissions by age g oup, be ween 2000 and 2014
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 4 o 10
Table 2 Dis ibu ion o hospi aliza ion episodes due o CAP as he main diagnosis (n= 548,699) and pe cen age o in-hospi al
mo ali y, acco ding o sociodemog aphic and clinical a iables, be ween 2000 and 2014
CAP hospi al admission In-hospi al mo ali y
% dis ibu ion (n) o Mean (± SD) / median % o Dea h (n) o Mean (± SD) / median p- alue *
Age (yea s) 65.9 (27.0) / 76 80.2 (12.0)/82 < 0.001
Age G oup (yea s) < 0.001
0–24 12.5 (68486) 0.7% (502)
25–44 5.3 (28962) 4.9% (1411)
45–64 12.4 (67870) 9.7% (6568)
65–74 15.8 (86576) 15.8% (13649)
75–84 30.4 (167038) 23.1% (38568)
≥85 23.7 (129762) 31.6% (41042)
Sex < 0.001
Male 55.1 (302219) 18.9% (57136)
Female 44.9 (246479) 18.1% (44604)
Unemploymen a e
a
(%) 10.6 (4.7)/10.4 10.8 (4.6)/10.7 < 0.001
Yea o admission < 0.001
2000–2004 27.7 (152069) 16.0% (24342)
2005–2009 33.5 (183659) 18.7% (34359)
2010–2014 38.8 (212971) 20.2% (43039)
Leng h o s ay (days) 10.6 (11,6)/8 10.0 (15.68)/6 < 0.001
Pe o med p ocedu es (n) 7.9 (3,7)/8 8.5 (4.01)/8 < 0.001
Legend: * - Mann-Whi ney es o nume ical a iables and Chi-squa e independence es o ca ego ical a iables;
a
- Numbe o unemploymen pe sons pe 100
ac i e pe sons egis e ed in he ci il pa ish o o igin o he use , acco ding o census in o ma ion ch onologically close o he episode (2001 Census and 2011
Census) [19,20]
Fig. 3 E olu ion o he numbe o in-hospi al CAP mo ali y by age g oup, be ween 2000 and 2014
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 5 o 10

Fac o s associa ed wi h in-hospi al CAP mo ali y
The sociodemog aphic and heal h a iables ha showed
g ea e associa ion wi h in-hospi al mo ali y we e
inse ed in o a mul iple logis ic eg ession analysis
model, hus ob aining he magni ude o he associa ion
h ough he de e mina ion o Odds Ra io and adjus ed
Odds Ra io (Table 3).
Acco ding o he adjus ed Odds Ra io analysis, he
inc ease in age ep esen ed a p og essi e and signi ican
aise in he p obabili y o dea h, excep o he age g oup
1–4 yea s, who showed a lowe p obabili y o dea h. The
age g oup ≥85 yea s old (Adjus ed OR = 124.256; 95%CI:
97.838–157.807) and males (Adjus ed OR = 1.261;
95%CI: 1.243–1.280) we e signi ican ly associa ed wi h
dea h isk o CAP hospi aliza ion. Males had a 26%
inc eased p obabili y o dea h, compa ed o emales.
The inc ease o 1 pe cen age poin in he pa ien s ‘pa -
ish unemploymen a e ep esen ed an inc ease o 0.5%
in dea h p obabili y. Conce ning he yea o admission,
i was obse ed a dec ease in p obabili y o dea h i he
admission occu ed in he las quinquennium (2010–
2014). A e 2010, his isk dec eased (Adjus ed OR =
0.961; 95%CI: 0.940–0.982).
LOS was associa ed wi h he dea h e en and each day
o hospi aliza ion ep esen ed a dec ease o 3.3% in he
p obabili y o dea h. The numbe o p ocedu es pe -
o med du ing hospi aliza ion was also associa ed wi h
an inc eased isk o dea h, wi h an inc ease o 3.5% o
each addi ional p ocedu e.
Figu es 4and 5 ep esen he compa ison o Odds
Ra io and Adjus ed Odd-Ra ios o p obabili y o dea h
a e hospi al admission (sex, yea o hospi al admission,
unemploymen a e, numbe o days o admission, num-
be o p ocedu es). A e he de e mina ion o he ad-
jus ed Odds Ra io, all a iables main ained he same isk
end e i ied in he analysis o he c ude Odds Ra io,
excep o he quinquennium o admission, since hospi-
aliza ions a e 2010 showed a lowe isk o dea h.
The de e mina ion o he a ea unde he ROC cu e
esul ed in 72.2%, which means a good model was
ob ained [21].
Discussion
Ou s udy showed ha , in he 15 yea s unde analysis,
despi e a end o dec ease in CAP hospi aliza ions and
associa ed mo ali y a e 2012, he numbe o
Table 3 Adjus ed odds a io o “dea h”e en acco ding o sociodemog aphic and clinical a iables be ween 2000 and 2014
a
(p
alue < 0.05)
CAP In-hospi al mo ali y
OR (95% CI) Adjus ed OR (95%CI)
Age G oup (yea s)
<1 1 1
1–4 0.410 (0.294–0.572) 0.564 (0.404–0,787)
5–14 1.201 (0.892–1.619) 1.697 (1.257–2.290)
15–24 7.149 (5.464–9.354) 9.746 (7.435–12.776)
25–44 9.317 (7.310–11.875) 11.862 (9.286–15.154)
45–64 19.492 (15.364–24.729) 26.098 (20.530–33.175)
65–74 34.050 (26.857–43.169) 47.238 (37.183–60.011)
75–84 54.616 (43.097–69.214) 78.158 (61.545–99.256)
≥85 84.156 (66.406–106.651) 124.256 (97.838–157.807)
Sex
Male 1.055 (1.041–1.070) 1.261 (1.243–1.280)
Female 1 1
Unemploymen a e
a
1.013 (1.011–1.014) 1.005 (1.003–1007)
Yea o admission
2000–2004 1 1
2005–2009 1.208 (1.186–1.230) 1.047 (1.025–1.069)
2010–2014 1.329 (1.306–1.352) 0.961 (0.940–0.982)
Leng h o s ay 0.994 (0.993–0.995) 0.977 (0.976–0.978)
Numbe o p ocedu es 1.049 (1.047–1.050) 1.035 (1.033–1.037)
Da a was also adjus ed o Region o Po ugal Mainland and o Majo Diagnos ic Ca ego ies ( esul s no p esen ed);
a
- Numbe o unemploymen pe sons pe 100 ac i e pe sons egis e ed in he ci il pa ish o o igin o he use , acco ding o census in o ma ion ch onologically
close o he episode (2001 Census and 2011 Census) [19,20]
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 6 o 10
admissions and in-hospi al mo ali y inc eased s eadily
o e ime. In-hospi al CAP mo ali y was mainly associ-
a ed wi h he e y old indi iduals (> 75 yea s), males,
and pa ien s’pa ish unemploymen a e.
This ans e sal la ge-scale s udy, o e a pe iod o 15
yea s, based on a e ospec i e da a analysis o hospi al
admission episodes and associa ed mo ali y is one o
he la ges s udies o his na u e ca ied ou in Po ugal,
co e ing all ages, all public hospi als, and he en i e
mainland popula ion, allowing obus esul s and
compa isons.
Repo s on CAP incidence ocus mainly on hospi ali-
za ions and show wide a iabili y ac oss coun ies,
anging om 1.48 pe 1000 popula ion in England, 2.75
in Ge many o 6.27 in Spain [6,22–24]. Despi e ou
hospi aliza ion a es (2001/2011) a e highe han o he
Eu opean coun ies [25], ou da a a e in line wi h a p e-
ious s udy in ou coun y [13] and close o epo s om
Spain [23,24]. Fu he mo e, ou age g oup-speci ic inci-
dence a es inc eased wi h ad ancing age, wi h incidence
a es in olde adul s e y simila o hose epo ed in
Spain [24]. This c oss-coun y a iabili y may be ela ed
no only o clima e di e ences and accine co e age o
in luenza and pneumococcal accines, bu also o di e -
ences in heal h sys em s uc u e and o he ac o s such
as he e ogenei y in admission c i e ia [26].
The s eady inc ease o CAP hospi aliza ions and asso-
cia ed mo ali y is pa icula ly ela ed o he e y old
people (Figs. 2and 3) and e lec s mos p obably he
demog aphic changes ha occu ed in he Po uguese
popula ion du ing he la ge s udy pe iod, wi h a mo e
aged popula ion acco ding o 2001 and 2011 census
Fig. 4 Compa ison o Odds Ra io and Adjus ed Odd-Ra ios o p obabili y o dea h a e hospi al admission (age g oup)
Fig. 5 Compa ison o Odds Ra io and Adjus ed Odd-Ra ios o p obabili y o dea h a e hospi al admission (sex, yea o hospi al admission,
unemploymen a e, numbe o days o admission, numbe o p ocedu es)
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 7 o 10
da a. Howe e , i we look a he numbe s associa ed wi h
he g ow h o he elde ly popula ion and he numbe o
CAP hospi aliza ions, he inc emen s a e disp opo ion-
a e, aking ha in he co esponden pe iod, he g ow h
o popula ion o e 75 yea s was 37.5%, con as ing wi h
an inc emen o 115% o he hospi aliza ion episodes.
This sugges s ha he e a e o he esponsible ac o s
ega dless o he change in he popula ion s uc u e.
Ac ually, his disp opo iona e inc ease may e lec no
only, he weigh o ageing, bu also he o e all impac o
he associa ed como bidi ies in he elde ly, as docu-
men ed in se e al s udies [25–28]. Fu he mo e, apa
om he known heal h consequences o he ageing
p ocess, i has been obse ed ha he olde Po uguese
adul s, epo a low household income and a e a ulne -
able g oup in e ms o unheal hy li es yle beha io s and
poo socioeconomic condi ions [29]. In pa icula , olde
people o e 75 yea s o age who li e alone ep esen a
g oup ha is subjec o g ea economic and social
ulne abili y [30].
Du ing he analyzed pe iod, he e was a dec ease in
hospi al admissions and co esponding mo ali y in he
g oup o child en unde 4 yea s, which can be explained
by he in oduc ion in Po ugal o 7- and 13- alen
pneumococcal conjuga e accine in 2001 and 2009, e-
spec i ely. Al hough no included in he Na ional
Immuniza ion P og am un il 2015, hei use was widely
ecommended by pedia icians o all heal hy child en,
allowing inc easing accina ion co e ages, 33–72%
be ween 2001 and 2013 [31], e en ually jus i ying he e-
duc ion o hospi al admissions and hei mo ali y aking
in o accoun ha S ep ococcus pneumoniae is he main
e iological agen o CAP [4].
Rega ding he in-hospi al mo ali y associa ed wi h CAP,
ou explo a o y model o he ac o s associa ed wi h he
p obabili y o dea h e ealed ha he inc ease in age was
he ac o ha con ibu ed mos o he ise in dea h p ob-
abili y, mainly in he e y old (> 85 yea s). In ac , 91.7% o
all dea hs occu ed in he age g oup ≥65 yea s old.
The o e all mo ali y a e was 18.5% o e he 15 yea s
analyzed. This alue is highe han ha ound in se e al
s udies anging om 4.8 o 14.4% [3,25,27,28], bu is
simila o o he se ies o a ious s udies conduc ed in
Eu ope [4]. A possible explana ion he obse ed highe in-
hospi al mo ali y a e o CAP may be ela ed o he p o-
g essi e ageing o Po ugal elde ly popula ion whe e e y
olde people (aged 85 and o e ) a e obse ed o be g ow-
ing a a as e a e han any o he age g oup. Indeed,
Po ugal s ands ou om he es o he EU coun ies, as i
is one o he as es ageing coun ies, conside ing ha by
2050 hose aged 55 and o e should ep esen almos hal
(47.1%) o he o al popula ion [32]. In addi ion, Teixei a-
Lopes hypo hesized ha he ea o dying a home, o
pa ien s and amily membe s, could explain he inclusion
o many pa ien s wi h end-o -li e pneumonia, especially in
olde age g oups, and we belie e his may ha e impac on
obse ed hospi al mo ali y [33].
In one s udy whe e h ee wo ld egions we e analyzed,
CAP mo ali y was 13.3% in La in Ame ica, 9.1% in
Eu ope and 7.3% in US/Canada [34]. A e adjus men
o con ounding a iables, es ima ed di e ences in mo -
ali y be ween he h ee egions we e signi ican ly e-
duced. The ac o s ha con ibu ed o he di e ences
ound we e he incidence o H1N1 in ec ion, como bidi-
ies (ce eb o ascula disease), abno mal labo a o y
alues (ele a ed blood u ea ni ogen), an imic obial he -
apy (mac olide use, luo oquinolone use) and accina-
ions (in luenza, pneumococcal). Some o hese ac o s,
may explain he numbe s obse ed in ou coun y, how-
e e , gi en he limi a ions ela ed o he use o an
adminis a i e da abase we we e unable o include
accu a e da a conce ning he abo e-men ioned ac o s.
A e 2012, he o e all end o g ow h in CAP hospi-
aliza ions and associa ed mo ali y changed. This
change, in pa icula he dec ease in he numbe o
dea hs in he las 2 yea s o he s udy and a educ ion in
he isk o dea h be ween 2010 and 2014, may be ela ed
o he c ea ion and implemen a ion o a na ional p o-
g am o ch onic espi a o y diseases in 2012 [35,36].
This p og am unde he guidance o he Di ec o a e-
Gene al o Heal h was esponsible and encou aged he
c ea ion and implemen a ion o espi a o y guidelines
and heal h policies in ou coun y. We highligh he ee
in oduc ion o he in luenza accine in he popula ion
o e he age o 65 and in o he a - isk g oups since he
2012/2013 win e season. In ac , wi h his public heal h
policy in luenza accina ion co e age a e in he popula-
ion aged 65 and olde inc eased om 43 o 50%,
be ween 2012 and 2014 [37,38]. P e ious s udies ha e
documen ed an associa ion be ween in luenza accin-
a ion and a less se e e clinical cou se o CAP and an im-
p o ed su i al in hospi alized pa ien s wi h his disease,
du ing in luenza season [39,40]. Mo eo e , his educ-
ion in CAP hospi aliza ions associa ed wi h he
obse ed isk o dea h educ ion in he i e-yea pe iod
2011–2014 compa ed o he p e ious i e-yea pe iods,
may also be ela ed o he in oduc ion in Po ugal o
he 13- alen pneumococcal conjuga e accine, as
e e ed in he s udy o Kislaya, whe e a educ ion o
pneumococcal pneumonia admissions in Po uguese
elde ly was a ibu ed o he hypo he ical indi ec e ec
(He d e ec ) o pneumococcal child accina ion [31].
Ou esul s showed a sex gap, wi h men p esen ing
highe numbe s o CAP hospi aliza ions, mo ali y and a
highe isk o dea h du ing hospi aliza ion. These esul s
a e in acco dance wi h o he s udies ha show a sex in-
equali y in heal h indica o s associa ed wi h hospi aliza-
ions [3,4,12,29], wi h men p esen ing consis en ly
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 8 o 10
highe hospi aliza ion a es o CAP, which may be e-
la ed o a g ea numbe o isk ac o s, such as smoking,
alcoholism, and exposu e o oxic occupa ional exposu e
[3,29]. The g ea e a io o male o emale mo ali y
may also e lec highe a es o smoking in males p e-
disposing o espi a o y disease [3,16].
Ou s udy e ealed a mean LOS o 10.6 days and a me-
dian o 8 days which was in line wi h o he epo s [3,29,
34]. E en ual di e ences be ween s udies and coun ies
a e ela ed o he a ailabili y o con inuing ca e ou side
he hospi al en i onmen [41]. In ou model o each day
o hospi aliza ion, he isk o dea h dec eased 2.3%, which
we in e p e as a e lex o he in e up ion o
hospi aliza ion by dea h, gi en ha mo e han one- hi d
o he dea hs (35.9%) occu ed wi hin he i s 4 days o
admission.
The e a e ew epo s on he associa ion be ween he
unemploymen a e and heal h s a us [42]. In ou s udy,
pa ien s’pa ish unemploymen a e was posi i ely asso-
cia ed wi h CAP mo ali y. This is he i s s udy ha
links CAP mo ali y wi h unemploymen a e as a p oxy
o he egional economic si ua ion. Gi en ha a highe
unemploymen a e may gene ally be ela ed o a mo e
ad e se economic condi ion and his has been associa ed
wi h inc eased bac e ial pneumonia hospi aliza ion a es
[28], ou hospi al mo ali y a es may ha e been in la ed
by he impac o he economic c isis. This ela ionship
poin s o he need o speci ic socioeconomic, egional
and popula ion policies in he mos ulne able g oups,
such as he elde ly in social dep i a ion. I also hin s ha
common spa ial p ocesses (en i onmen al-le el ac o s)
may be in ol ed in pneumonia ou come, de e mining
pa e ns and geog aphic a ia ions and jus i ying u u e
in es iga ions.
The e a e limi a ions o his s udy. Fi s , i was conduc ed
using e ospec i e adminis a i e da a, condi ioning he
lack o in o ma ion abou como bidi ies, mic obiologic
e iology, he p ocess o ca e and in luenza/pneumococcal
accina ion s a us. Also, i was no possible o exclude
pa ien s who li ed in nu sing homes o in long- e m ca e
ins i u ions in which i was possible, ha pneumonia cases
we e heal hca e ela ed. Second, he ICD-9 codes did no
allow o he exclusion o nosocomial in ec ions as no code
speci ic o his diagnos ic ca ego y exis s, wi h he conse-
quen o e es ima ion o CAP hospi aliza ion a es. Thi d,
changes, ac oss he ime pe iod in analysis, conce ning
admission guidelines, diagnosis and coding p ac ices migh
be p obable sou ces o bias.
Conclusions
Ou s udy e eals a huge bu den o CAP hospi aliza ions
and mo ali y in he olde popula ion, showing ha CAP
can be a majo public heal h conce n, e en in a de eloped
coun y, such as Po ugal. In-hospi al mo ali y was
pa icula ly ela ed o he ageing p ocess and dep i ed
socioeconomic condi ions. The e o e, implemen a ion o
CAP p e en i e measu es should be ein o ced in hese
ulne able g oups, conce ning smoking cessa ion, lu, and
pneumococcal accina ions o educe he bu den, pa icu-
la ly in he olde popula ion. Also, ou indings highligh
ha e en small shi s in unemploymen a e a e associa ed
wi h in-hospi al CAP mo ali y, sugges ing ha unemploy-
men migh be an impo an social de e minan .
Abb e ia ions
95% CI: 95% Con idence in e als; ACSS: Cen al Adminis a ion o he
Heal h Sys em; CAP: Communi y-associa ed pneumonia; CI: Con idence
in e als; EU: Eu opean Union; ICD: In e na ional Classi ica ion o Diseases;
LOS: Leng h o s ay; OR: Odds a io; ROC: Recei e Ope a ing Cha ac e is ic
Acknowledgemen s
The au ho s acknowledge D . Rica do Mes e, om ACSS, o he a ailabili y
o da a.
Au ho s’con ibu ions
S udy design (EP, LV, MR, PN), da a acquisi ion (EP, LV, PN), da a analysis (EP,
LV, PN), manusc ip d a ing (EP, CB, LV, AC, PN), and c i ical manusc ip
e ision (CB, AC, PN). All au ho s ead and app o ed he inal manusc ip .
Funding
This publica ion was unded by Fundação pa a a Ciência e a Tecnologia.
A ailabili y o da a and ma e ials
The da a ha suppo he indings o his s udy a e a ailable om [ACSS] bu
es ic ions apply o he a ailabili y o hese da a, which we e used unde
license o he cu en s udy, and so a e no publicly a ailable. Da a a e
howe e a ailable om he au ho s upon easonable eques and wi h
pe mission o [ACSS].
E hics app o al and consen o pa icipa e
This s udy was app o ed by he E hics Commi ee o he Lisbon Facul y o
Medicine (107/19) and he need o in o med consen was wai ed.
Consen o publica ion
No applicable.
Compe ing in e es s
The au ho s decla e ha hey ha e no compe ing in e es s.
Au ho de ails
1
Escola Supe io de En e magem de Lisboa, Lisbon, Po ugal.
2
Di eção-Ge al
de Saúde, Lisbon, Po ugal.
3
Ins i u o de Saúde Ambien al (ISAMB), Faculdade
de Medicina, Uni e sidade de Lisboa, Lisbon, Po ugal.
4
Cen o Hospi ala
Uni e si á io Lisboa No e, Lisbon, Po ugal.
5
Labo a ó io de Bioma emá ica,
Faculdade de Medicina, Uni e sidade de Lisboa, Lisbon, Po ugal.
Recei ed: 29 Ap il 2019 Accep ed: 30 Decembe 2019
Re e ences
1. B a NK, Niede man MS. Managemen o communi y-acqui ed pneumonia:
a e iew and upda e. The Ad Respi Dis. 2011;5(1):61–78.
2. Wo ld Heal h O ganiza ion. The op 10 causes o dea h [In e ne ]. Fac
Shee . 2018 [ci ed 2019. A ailable om: h ps://www.who.in /news- oom/
ac -shee s/de ail/ he- op-10-causes-o -dea h
3. O ganiza ion o Economic Co-ope a ion and De elopmen . Heal h a a
Glance: Eu ope 2018—S a e o Heal h in he EU Cycle. 2018.
4. Wel e T, To es A, Na hwani D. Clinical and economic bu den o
communi y-acqui ed pneumonia among adul s in Eu ope. Tho ax.
2012;67(1):71–9.
5. Woodhead M. The Eu opean ision o communi y-acqui ed pneumonia.
InSemina s in espi a o y and c i ical ca e medicine 2009 Ap (Vol. 30, No.
02, pp. 136–145). © Thieme Medical Publishe s.
Pessoa e al. BMC Pulmona y Medicine (2020) 20:18 Page 9 o 10