PAEDIATRIC RESPIRATORY REVIEWS (2007) 8, 195–203

MINI-SYMPOSIUM: MICROBIOLOGICAL DIAGNOSTIC PROCEDURES

IN RESPIRATORY INFECTIONS

Congenital and neonatal pneumonia

Michael D. Nissen*

Department of Infectious Diseases, Royal Children’s Hospital, Herston Road, Herston, Queensland 4029, Australia

Summary The greatest risk of death from pneumonia in childhood is in the neonatal
period. It is estimated that pneumonia contributes to between 750 000–1.2 million
neonatal deaths annually, accounting for 10% of global child mortality. Congenital and
neonatal pneumonias are often a difficult disease to identify and treat, with clinical
manifestations often being non-specific. Many of the normal lung defences are compro-
mised in the fetus and neonate, leading to an increased susceptibility to infection. The
aetiology and epidemiology of congenital and neonatal pneumonias will depend on the
clinical setting and population that the baby belongs to, the stage in the perinatal period,
the gestational age of the baby and the definition of pneumonia. Diagnosis, treatment and
prevention strategies are therefore also dependent on these factors, and will differ
depending on the clinical setting. This review summarizes the current knowledge
concerning congenital and neonatal pneumonia worldwide and discusses future directions
in the prevention of the disease .
! 2007 Published by Elsevier Ltd.

INTRODUCTION infectious causes such as meconium aspiration. Many of the

normal lung defences are compromised in the fetus and
The greatest risk of death from pneumonia in childhood is in neonate. These include non-specific barriers such as the
the neonatal period.1 At least one third of the annual glottis and vocal cords, cilary escalator, airway phagocytes,
10.8 million deaths in children worldwide occur in the first secretory antibodies, mucosal lymphoid tissue, antimicrobial
28 days of life,2 with a substantial proportion of these being proteins and opsonins.
due to pneumonia. It is estimated that pneumonia contri- The aetiology and epidemiology of congenital and neo-
butes to between 750 000 and 1.2 million neonatal deaths natal pneumonias will depend on: (1) the clinical setting and
annually, accounting for 10% of global child mortality.3 Of all population that the baby belongs to (e.g. developed/devel-
neonatal deaths, 96% occur in the developing world.2 oping world, tertiary/district hospital or community setting);
Congenital and neonatal pneumonias are often difficult to (2) the stage in the perinatal period; (3) the gestational age
identify and treat. Clinical manifestations are often non- of the baby, and (4) the definition of pneumonia.1 Diag-
specific, sharing respiratory and a range of non-inflammatory nosis, treatment and prevention strategies are therefore
processes. Laboratory findings also have limited value, with also dependent on these factors, and will differ depending
attempts to identify specific microbes often unsuccessful due on the clinical setting.
to the difficulty in sample recovery from intrapulmonary sites
without contamination. Many organisms are primarily uncul-
tivable or uncultivable due to antimicrobial therapy. Radi- EPIDEMIOLOGY
ological evidence of lung inflammation may result from non-
Lower respiratory tract infections (LRTI) in neonates can be
* Tel.: +61 7 33655021; Fax: +61 7 33655455. classified as congenital or neonatal in origin, and are defined
E-mail address: [email protected]. by when the infection or pathogen has been acquired.
1526-0542/$ – see front matter ! 2007 Published by Elsevier Ltd.
doi:10.1016/j.prrv.2007.07.001
196 M. D. NISSEN

Congenital pneumonias are usually part of a transplacental Table 1 Pathogens associated with congenital and neonatal
infection, while neonatal pneumonias can evolve from pneumonias
intrauterine or postnatal acquisition. Bacteria Viruses
Neonatal pneumonia can be classified as early and late
Escherichia coli Herpes simplex virus
onset.1 Early-onset neonatal pneumonia, in general, is Enterobacter aerogenes Respiratory syncytial virus
defined as a clinical presentation in the first 48 h up to 1 Group B streptococcus Human metapneumovirus
week of life, while late-onset neonatal pneumonia occurs in (S. agalactiae)
the next 3 weeks. Intrauterine pneumonia is a subgroup of Group A streptococcus Parainfluenzae viruses
early-onset neonatal pneumonia. It presents as a stillbirth, (S. pyogenes) 1, 2 and 3
low Apgar scores or severe respiratory distress and is Klebsiella spp Human cytomegalovirus
usually associated with maternal chorioamnionitis. Infected Pseudomonas aeruginosa Influenza A and B viruses
(maternal) amniotic fluid is then aspirated in utero, after Streptococcus viridans group Human adenovirus
prolonged rupture of the chorioamniotic membranes, or Staphylococcus aureus Human
during delivery of the affected neonate. immunodeficiency virus
Proteus spp
Congenital pneumonia occurs in the setting of a mater-
Streptococcus pneumoniae Fungi
nal systemic infection, which may or may not be sympto- Group D and G streptococci Candida albicans
matic in the mother. Enterococcus spp
Neonatal autopsy studies have determined that intrau- Haemophilus influenzae Atypical bacteria
terine and early-onset pneumonia occurs in 10–38% of (non-typeable)
stillborns and 20–63% of liveborn babies who subsequently Staphylococcus epidermidis Chlamydia trachomatis
died.4 Early investigations of the cause of neonatal deaths in Salmonella spp Ureaplasma urealyticum
the first 48 h of life found a pneumonia in 20–38% of cases, and U. parvum
with the highest incidence in lower socioeconomic groups.5 Acinetobacter spp Listeria monocytogenes
In an Indian study, >50% of all childhood pneumonia deaths Neisseria meningitidis Treponema pallidum
occurred in neonates.6 Morganella spp Mycobacterium tuberculosis
Serratia spp Pneumocystis jirovecii
Birth weight and age of onset strongly determined the
Bordetella pertussis
mortality risk from pneumonia.1 Case fatality rates are
higher for low birth weight infants,7 and for intrauterine
and early-onset pneumonia when compared with late-
onset disease.8,9 The epidemiology of postpartum and with neonatal pneumonia include numerous bacteria, fungi
late-onset neonatal pneumonias in general tend to be and viruses (Table 1). Bacterial pneumonia from infected
associated with nosocomial infections, with introduction amniotic fluid or colonization of the birth canal is linked with
of pathogens occurring transplacentally via maternal maternal chorioamnionitis and fetal asphyxia. It is assumed
chorioamniotitis or medical intervention. The true inci- that asphyxia leads to fetal gasping and aspiration of
dence of late-onset pneumonia in neonates is difficult to infected amniotic fluid. This hypothesis is based on the
determine since many series do not report the age of histological finding of amniotic fluid and/or maternal white
onset. blood cells in the affected neonatal lungs.1 The bacterial
The multi-country World Health Organization (WHO) aetiology of neonatal pneumonia is also influenced by
Young Infants Study provides useful data on community- nosocomial infection in neonatal intensive care units. In
acquired neonatal sepsis, including pneumonia. These data some areas of the world high rates of Streptococcus pneu-
however are skewed towards infants seen in hospital out- moniae have been detected in late-onset neonatal pneu-
patient departments and therefore presentations of intrau- monia. Also, since lung-derived samples are rarely obtained
terine pneumonia or early-onset pneumonia, which occur from neonates,5,8 studies pertaining to neonatal pneumonia
mainly in the first 48 h of life, may not be well repre- contain blood culture data which will underestimate the
sented.10 proportion of cases that are bacterial.1 A diagnosis of
pneumonia in a study of Ethiopian infants was not asso-
ciated with bacteraemia when compared with a clinical
AETIOLOGY AND PATHOGENESIS diagnosis of sepsis, ‘severe disease’ or death.11
The epidemiological features of neonatal pneumonia, in Atypical bacterial pathogens, e.g. Chlamydia trachomatis,
general, with their resultant implications for treatment and are well recognized as agents of late-onset pneumonia. Bor-
prevention are sufficiently similar to those of neonatal detella pertussis may present as early-onset or late-onset
bacteraemia and meningitis, and therefore can be used pneumonia, and is most commonly associated with close
to understand the aetiology of the disease.1 contact with an infected parent, sibling, relative or healthcare
The aetiology of maternal chorioamnionitis has also worker.
been used to understand the aetiology of intrauterine Viral neonatal pneumonias can either be associated
and early-onset pneumonia.1 The pathogens associated with intrauterine, early-onset or late-onset pneumonias,
CONGENITAL AND NEONATAL PNEUMONIA 197

and can be acquired from the birth canal [e.g. herpes The differentiation of neonatal pneumonia from non-
simplex virus (HSV)]), infected siblings, parents and/or infectious respiratory conditions such as hyaline membrane
healthcare workers [e.g. respiratory viruses such as respira- disease (HMD), transient tachypnoea of newborn and
tory syncytial virus (RSV)] with or without nosocomial meconium aspiration is problematical since the clinical radi-
involvement. ological appearance can be identical, and pathology and
Congenital pneumonia is caused by a mixture of bac- radiology services in some facilities may be rudimentary or
teria, viruses and fungi. unavailable.
Gestational age of the neonate and time to onset of
clinical signs together assist in the clinical diagnosis of
CLINICAL MANIFESTATIONS
pneumonia. However, HMD can occur close to term
The definition of neonatal pneumonia will depend on and intrauterine infection may result in the premature
clinical setting and geographical location, access to medical onset of labour. In addition, gestational age assessment
care and investigations, and public health preventative may be inaccurate and dependent on the clinical expertise
measures in place in the population. available, so treatment decisions for respiratory distress of
Neonatal pneumonia is suspected in any newborn infant the newborn based on the estimated gestational age are
with respiratory distress, the features of which include any not always practical, may be unsafe and may interfere with
of the following: rapid, noisy or difficult breathing, respira- the true contribution of pneumonia to mortality in the first
tory rate >60 beats/min, chest retractions, cough and/or week of life.1
grunting. The clinical risk factors and features of neonatal The incidence of factors predicting neonatal pneumonia
pneumonia are listed in Table 2. varies between clinical studies and with the methodology
WHO does not distinguish between neonatal pneumo- used to study the disease.9,12 Babies prospectively admitted
nia and other forms of severe sepsis, such as bacteraemia, to a UK neonatal unit with early-onset pneumonia (before
since there is much overlap in the clinical signs on pre- 48 h of age) had risk factors in 17% of cases, but in 41% of
sentation, and organ involvement and empiric treatment these babies early onset of labour was the only predictor of
regimens are similar.1 pneumonia. An Indian study, in comparison, found over
The sensitivity of clinical findings for radiologically diag- 50% of neonatal pneumonia cases had no known predis-
nosed pneumonia in neonates has been evaluated in position or risk factor.
developing countries.8,13,14 Tachypnoea appears to be Other non-infectious respiratory conditions in the
the most consistent sign present in 60–89% of cases. Other neonatal period must be considered in the assessment
signs appear to be less reliable and include chest recession of babies with respiratory distress. Pulmonary oedema
(36–91% of cases), fever (30–56%), inability to feed (43– secondary to congenital heart disease can be confused
49%), cyanosis (12–40%) and cough (30–84%). with late-onset pneumonia. Other rare presentations of
neonatal respiratory distress that create diagnostic dilem-
mas include pulmonary haemorrhage and pulmonary
Table 2 Clinical risk factors and features of neonatal pneu- infarct.
monia (modified from Mathur et al.12) The presence of an endotracheal tube for mechanical
1. Predisposing factors
ventilation or medical tubing causing breaches of the
a. Maternal fever (>38 8C) respiratory tract mucosal barrier and its local immune
b. Foul smelling liquor system makes bacterial invasion of the lower respiratory
c. Prolonged rupture of membranes (>24 hours) tract inevitable, clouding the true incidence of late-onset
pneumonia, particularly in developed countries with ready
2. Clinical picture of sepsis
a. Poor feeding
access to neonatal intensive care nurseries.
b. Lethargy In the UK neonatal study cited above, endotracheal tube
c. Poor reflexes bacterial colonization occurred in 94% of artificially venti-
d. Hypothermia or hyperthermia lated babies, most commonly by Gram negative organisms
e. Abdominal distension and Staphylococcus epidermitis.9 In only one of seven cases
3. Chest X-ray suggestive of pneumonia (nodular or
with simultaneous bacteraemia was the same organism
coarse patchy infiltrate, diffuse haziness or granularity, air grown from blood cultures. Significantly there was no
bronchogram, lobar or segmental consolidation); difference in the incidence or the timing of endotracheal
radiological changes not resolved within 48 hours tube colonization between babies who did or did not have
late-onset pneumonia, regardless of gestational age or the
4. Positive sepsis screen (any of the following):
a. Bands >20% of leucocytes
duration of mechanical ventilation.
b. Leucocyte count out of reference range The suspicion of staphyloccal infection as an aetiological
c. Raised C reactive protein agent for neonatal pneumonia is warranted when skin
d. Raised erythrocyte sedimentation rate pustules, cellulitis, oomphalitis (umbilical cord infection),
pneumatocoeles or empyema is present.
198 M. D. NISSEN

ATYPICAL NEONATAL ficiency virus (HIV) infection has been described in two
PNEUMONIAS southern African studies.23,24 Co-infections with Mycobac-
teriumtuberculosis, syphilis and cytomegalovirus were com-
Chlamydia trachomatis mon and contributed to the clinical presentation.
Chlamydia trachomatis pneumonia can occur at 1–3 Congenital HIV infection also increases the mortality rate
months of age, manifesting as a protracted onset of staccato from neonatal respiratory distress syndrome and sepsis
cough, usually without wheezing or fever. Findings on chest associated with S. pneumoniae and Staphylococcus aureus.
X-ray include hyperinflation and diffuse bilateral infiltrates,
with peripheral blood eosinophilia possibly present. Diag- Mycobacterium tuberculosis25
nostic testing is usually performed on a nasopharyngeal
specimen. Treatment is with a macrolide antibiotic (ery- Infants may acquire tuberculosis (TB) by transplacental
thromycin, clarithromycin or azithromycin). spread, aspiration or ingestion of infected amniotic fluid,
The estimated risk of C. trachomatis pneumonia devel- or airborne inoculation from close contacts (family mem-
oping in a neonate with maternal colonization is 7%.15 The bers or nursery personnel). Approximately 50% of children
incidence of C.trachomatis pneumonia in neonates exposed born to mothers with active TB develop the disease during
to C. trachomatis in large cohort studies varies from 3 to the first year of life if chemoprophylaxis or BCG vaccine is
16%.16,17 The interpretation of C. trachomatis-associated not given.
neonatal pneumonia is hampered because the ‘gold stan- The clinical presentation of neonatal TB is non-specific,
dard’ of diagnosis by percutaneous lung aspirate is rarely, if but is usually marked by multiple organ involvement. The
ever, performed. The ability now to perform C. trachomatis exposed neonate may present acutely or chronically ill with
polymerase chain reaction (PCR) testing on nasopharyngeal fever, lethargy, respiratory distress, hepatosplenomegaly or
or endotracheal aspirates from infants with neonatal pneu- failure to thrive. All cases of suspected congenital TB should
monia has increased the sensitivity of detection of this have a chest X-ray and cultures of tracheal aspirates, gastric
pathogen. It is assumed that C. trachomatis contributes washings, urine and cerebrospinal fluid (CSF) for acid-fast
significantly to neonatal pneumonia in countries where bacilli. Skin testing is not sensitive in neonates but should be
untreated sexually transmitted diseases in women are com- attempted, with biopsy of liver, lymph nodes, lung or pleura
mon. Studies from Ethiopa, Papua New Guinea and Kenya sometimes needed to confirm the diagnosis.
have detected C. trachomatis by direct immunofluorescent Pregnant women with a positive Mantoux skin test but
staining of the nasopharyngeal aspirates of neonates with normal chest X-ray should either start chemoprophylaxis
radiologically confirmed pneumonia in 16%, 22% and 46% of during pregnancy or after delivery, depending on the
cases, respectively.11,18,19 In Papua New Guinea there was likelihood of their being recently infected and their risk
also a borderline association of C. trachomatis pneumonia of progression to disease, as well as their clinical evidence
with S. pneumoniae bacteraemia.18 Whether the inflamma- of disease. Pregnant women with a positive Mantoux skin
tion of C. trachomatis pneumonia predisposes to secondary test and chest X-ray or symptoms indicative of active
bacterial pneumonia by mucosal injury is unknown. disease should be treated with non-teratogenic agents
during pregnancy; all household contacts should also be
screened. When TB is suspected around delivery, the
Ureaplasma urealyticum and U. parvum mother should be assessed by chest X-ray and sputum
U. urealyticum and U. parvum are part of the normal genital smear; separation of the mother and her offspring is
bacterial flora of both men and women. They are found in indicated only if the mother is non-adherent to medical
about 70% of sexually active humans. They are associated treatment or needs to be hospitalized, or when drug-
with non-specific urethritis (NSU) and infertility, but can resistant TB is involved.
also cause maternal and neonatal disease, including chor- Treatment is with a combination of routine anti-TB
ioamnioitis, stillbirth, premature birth and, in the perinatal drugs such as isonazid, rifampicin, pyrazinamide and strep-
period, pneumonia or meningitis. tomycin for up to 12 months depending on the extent of
the disease. According to the American Academy of
Pediatrics, treatment of latent infection with isoniazid for
Treponema pallidum 9 months is highly effective. This regimen may be extended
Fatal cases of congenital syphilis are usually associated with to 12 months for immunocompromised patients. When
severe pneumonitis (pneumonia alba) and hypoxaemia, drug resistance is suspected, combination therapies, which
especially in developing countries.20–22 usually consist of isoniazid with rifampin (rifampicin), are
administered until the results of susceptibility tests become
available. Organisms resistant to isoniazid only may be
Human immunodeficiency virus treated with rifampin alone for a total of 6–9 months.
A persistent neonatal pneumonia associated with a rapidly All infants with tuberculous disease should be started on
progressive presentation of congenital human immunode- four agents (isoniazid, rifampin, pyrazinamide and etham-
CONGENITAL AND NEONATAL PNEUMONIA 199

butol or streptomycin) until drug susceptibility is assessed. of cases. Mothers of neonates with HSV infection tend to
For susceptible intrathoracic TB, isoniazid, rifampin and have no history or symptoms of genital infection at the time
pyrazinamide are administered for a total of 2 months, of delivery.
at which point pyrazinamide is withdrawn and the other Neonates may present with either local or disseminated
two agents are continued for another 4–10 months disease. Babies with disseminated disease and visceral organ
depending on the severity of the disease. The same regi- involvement have pneumonitis, hepatitis and/or dissemi-
men may be applied in extrapulmonary TB with the nated intravascular coagulation with or without encephalitis
exception of skeletal, miliary and central nervous system or skin disease. Signs which can occur singly or in combina-
disease, which require daily administration of isoniazid, tion include temperature instability, lethargy, hypotonia,
rifampin, pyrazinamide and streptomycin for 1–2 months, respiratory distress, apnoeas and seizures.
followed by isoniazid and rifampin daily or twice weekly for Rapid diagnosis by HSV PCR, immunofluorescence of
another 10 months. When drug-resistant TB is suspected, a vesicle scrapings or viral culture is essential, with the most
regimen of isoniazid, rifampin and pyrazinamide plus either common site of retrieval being skin vesicles followed by
streptomycin or ethambutol should be prescribed initially, CSF, eyes and mouth. If no diagnostic virology facilities are
until the results of susceptibility tests become available. available, a Papanicolaou test of a smear from the vesicle
HIV-seropositive infants with pulmonary TB should receive base may show characteristic multinucleated giant cells and
isoniazid, rifampin, pyrazinamide and ethambutol or an intranuclear inclusions ! the Tzanck test.
aminoglycoside for 2 months, followed by isoniazid and The mortality of untreated disseminated disease is 85%.
rifampin for a total of at least 12 months. Apart from Treatment halves the mortality rate, and consists of high
conventional antimycobacterial agents, novel therapeutic dose aciclovir (20 mg/kg/q8h for 14–21 days), with vigor-
modalities, which stimulate the host immune system, such ous supportive therapy.
as interleukin-2 (IL-2), IL-12 and g-interferon, have been
tested with promising results.
Respiratory viruses
Listeria monocytogenes The role of respiratory viruses (RSV, influenza, parain-
Transplacental infection with Listeria monocytogenes can fluenza viruses, adenovirus and metapneumovirus) in neo-
result in fetal dissemination with granuloma formation (skin, natal pneumonia is well described by retrospective reports.
liver, adrenals, lymphatics, lungs and brain) called granulo- These viruses have been associated with seasonal late-
matosis infantisepticum. Aspiration of infected amniotic onset pneumonia where viral diagnostic techniques are
fluid or vaginal secretions can lead to perinatal L. monocy- accessible. Nosocomial outbreaks of respiratory viruses in
togenes pneumonia manifesting in the first week of life neonatal nurseries and co-infections with RSV and human
with respiratory distress, shock and a fulminant course. metapneumovirus (hMPV) have been described.26
Nosocomial acquisition has also been reported. Apnoea may be the sole presenting feature in neonatal
Neonates with early-onset L. monocytogenes infection viral pneumonias. The risk of death from neonatal pneu-
frequently are of low birth weight, have associated obstetric monia is higher in early-onset disease, hypoxaemia, low
complications and show evidence of sepsis with circulatory birth weight and absence of tachypnoea.
and/or respiratory insufficiency. Babies with delayed onset
are usually full-term, previously well neonates presenting
with meningitis or sepsis.
DIAGNOSIS AND INVESTIGATIONS
A sick neonate whose mother has listeriosis should be Chest X-rays should be performed in any patient with
evaluated for sepsis, including cultures of the umbilical cord, respiratory abnormalities. Blood should be collected for a
peripheral blood, CSF, gastric aspirate, meconium, mother’s complete blood count (CBC), C-reactive protein and
lochia and exudates from cervix and vagina, and grossly culture in all cases of neonatal pneumonia. While the
diseased parts of the placenta. Treatment should be yield from blood cultures is low, blood, if possible, should
initiated with ampicillin and an aminoglycoside such as be collected prior to antibiotic therapy to guide second-
gentamicin. Once a clinical response is seen the aminogly- line treatment in the event of first-line antibiotic failure.
coside can be discontinued. A minimum 14-day course of Blood cultures collected simultaneously with endotra-
ampicillin is usually necessary. cheal tube aspirates in mechanically ventilated neonates
may also assist in determining the significance of endo-
tracheal tube colonization. Pus from empyema or skin
Herpes simplex virus
pustules should be submitted for Gram staining and
HSV is usually transmitted during delivery through an bacterial culture.
infected maternal genital tract. Transplacental transmission The most useful diagnostic tests for congenital pneu-
of virus and hospital-acquired spread from one neonate to monia facilitate identification of the infecting microorgan-
another by hospital personnel or family may account for 15% ism.27
200 M. D. NISSEN

Bacterial culture been in place, the greater the likelihood that recovered
organisms represent colonizers rather than invasive patho-
Conventional bacteriological culture is used most widely and
gens; nonetheless, recovery of a single recognized pathogen
is currently the most helpful test. Aerobic incubation of
in large quantities may be helpful in the selection of
cultures is sufficient for recovery of most responsible patho-
antibiotic therapy, especially if culture results from normally
gens. Although the foul smell of amniotic fluid in the setting of
sterile sites are negative.
maternal chorioamnionitis is often attributable to anaerobes,
these organisms are seldom shown to be causative. The
culture of fungi, viruses, U. urealyticum and other unusual Culture from extrapulmonary sites
organisms often requires different microbiological processing
Detection of microorganisms at inflamed extrapulmonary
but may be warranted in suggestive clinical settings.
sites may be helpful, since concurrent involvement of the
lungs is not rare. Studies of abscesses, conjunctivitis, skin
Blood culture lesions and vesicles may be fruitful. Care should be taken to
Blood culture with at least 1 ml of blood from an appro- ensure that the specimen submitted is as free of contam-
priately cleaned and prepared peripheral venous or arterial ination as possible or that the test performed, such as
site is essential, since many neonatal pneumonias are organism-specific DNA probe or PCR-based assay, is less
haematogenous in origin and others serve as a focus for likely to be affected by such factors.
secondary seeding of the bloodstream. Blood culture
samples drawn through freshly placed indwelling vascular Culture from other respiratory sites
catheters may be helpful, but the possibility of contamina-
tion rises the longer the catheter is in place. Multiple In the presence of radiographically visible pleural fluid,
cultures of blood from different sites and/or drawn at careful positioning of the infant and cautious thoracentesis
different times may increase culture yield, but limited after sterile preparation of the sampling site may yield
circulating blood volume precludes this as the standard diagnostic findings on Gram stain, direct microscopy and/
of care in neonates. or culture. Sonography may reveal smaller fluid pockets and
facilitate safer sampling under direct visualization. Although
data from neonates are insufficient to draw conclusions,
Culture of specimens from lumbar puncture studies in older populations suggest a very high correlation
Routine culture and analysis of spinal fluid in infants in with culture of lung tissue and/or blood.
whom neonatal pneumonia is suspected is controversial, Quantitative culture techniques of bronchoscopic
since the yield is low and many infants with respiratory alveolar lavage (BAL) fluid have been reported to offer a
support requirements do not tolerate lumbar puncture specificity of >80% depending on the threshold selected
well. Spinal fluid may yield a pathogen when blood does not (values from >100 to 100 000 CFU/ml have been used).
(especially following maternal antibiotic pre-treatment). Non-directed non-bronchoscopic protected specimen
Presence of a pathogen in the spinal fluid may indicate brush specimens have been obtained through endotracheal
the need for alteration in the selection, dosage and duration tubes "3.0 mm in internal diameter and intuitively appear
of antibiotic therapy even if cultures from other sites yield to offer decreased probability of contamination. Data from
the same organism. neonates are sparse. Unlike bronchoscopically obtained
specimens, ensuring that sampling from a particularly
involved site has been accomplished is more difficult.
Urine culture Although used much less frequently than in previous
During the first 3 days of life, urine culture is unlikely to be decades, the lung puncture technique may still be useful in
helpful because most urinary tract infections seen at this age circumstances where pleural and subpleural lung surfaces
are haematogenous in origin. are visibly involved and can be well localized. The risk–
benefit ratio merits careful consideration given the risk of
such complications as pneumothorax, bronchopleural fis-
Culture of specimens from endotracheal
tula, haemothorax and sampling a non-diagnostic site.
aspiration
Culture and Gram stain of an endotracheal aspirate
Limitations of bacterial cultures
obtained by aseptic technique as soon as possible after
intubation may be useful. Under typical circumstances, A number of factors may interfere with the ability to
airway commensals take as long as 8 h to migrate down cultivate a likely pathogen from the sites noted, including
the trachea. At least one study demonstrated that culture (but not limited to) the following: (1) pretreatment with
of endotracheal aspirates obtained within 8 h of birth antibiotics that limit in vitro but not in vivo growth; (2)
correlates very well with blood culture results and probably contaminants that overgrow the pathogen; (3) pathogens
reflects aspirated infected fluid. The longer the tube has that do not replicate in currently available culture systems;
CONGENITAL AND NEONATAL PNEUMONIA 201

(4) sampling of an inappropriate site; and (5) patients in WHO recommends using ampicillin (50 mg/kg) every
whom the process is inflammatory but not infectious, such 12 h in the first week of life, then every 8 h from 2–4 weeks
as with meconium aspiration. of age, plus a single daily dose of gentamicin.28 First-line
Techniques that may help overcome some of these alternatives to ampicillin include benzylpenicillin or amox-
limitations include antigen detection, nucleic acid probes, icillin, and alternatives to gentamicin are tobramycin or
PCR-based assays or serological tests. Although once amikacin. If S. aureus is suspected, a penicillinase-resistant
widely used, tests such as latex agglutination for detection penicillin such as flucloxacillin or cloxacillin should be
of Group B streptococcal antigen in urine, serum or other substituted for ampicillin.
fluids have fallen into disfavour because of poor predictive A randomized trial of once daily gentamicin dosing in
value; however, new generations of non–culture-based Kenyan infants has confirmed a loading dose of 8 mg/kg
technologies continue to undergo development and may followed by 2 mg/kg (if weight <2 kg) or 4 mg/kg (if weight
be more accurate and widely available in the future. >2 kg) in the first week of life, followed by 4 mg/kg (weight
<2 kg) or 6 mg/kg (weight >2 kg) in week 2 of life and
Serological tests above.29
If a neonate fails to respond to first-line antibiotics,
Serological tests have limited use but may offer some WHO suggests changing to a third-generation cephalos-
insights. Serological tests for syphilis may suggest or confirm porin or chloramphenicol (only if the neonate is not
the presence of pneumonia alba, particularly in high risk premature and drug levels can be monitored).28
populations. The value of assessing antibody responses in Bacterial pathogens associated with neonatal pneumo-
acute and convalescent sera from infants using flora recov- nia are demonstrating increased antimicrobial resistance
ered from endotracheal aspirates has been suggested as worldwide. Patterns of antimicrobial resistance vary locally
being useful. This usually permits diagnosis only retrospec- and are dependent on whether cases are community or
tively, but may be useful in infants who fail to respond nosocomially acquired.1 Enteric Gram negative bacteria
adequately to empirical therapy or for epidemiological such as Escherichia coli show increasing resistance and all
purposes. Concerns persist regarding the specificity of such Klebsiella spp are intrinsically resistant to ampicillin, espe-
tests in distinguishing invasion from colonization. cially if this antibiotic is used for the treatment of maternal
fevers and urinary tract infections during pregnancy.
Markers of inflammation Multiresistant Pseudomonas spp and Serratia spp tend to
be associated with nosocomial environments and can cause
The use of markers of inflammation to support a diagnosis neonatal pneumonia. Antimicrobial resistance in Gram nega-
of suspected infection, including pneumonia, remains con- tive bacteria associated with neonatal sepsis ranges from 20
troversial. Various indices derived from differential leuko- to 84%, depending on the geographical locale.13,30–33
cyte counts have been used most widely for this purpose,
although non-infectious causes of such abnormal results are
numerous, and many reports have been published regard- Supportive care
ing infants with proven infection who initially had neutrophil The supportive care of neonates with pneumonia is linked
indices within reference ranges. Quantitative measure- with improved outcomes and lower case fatality rates.1 This
ments of C-reactive protein, cytokines (e.g. IL-6), and includes the use of oxygen, detection and treatment of
batteries of acute-phase reactants have been touted as hypoxaemia and apnoea, thermoregulation, detection and
more specific but are plagued by concerns regarding limited treatment of hypoglycaemia, and improved use of intrave-
positive predictive value because of: (1) lag from infection nous fluids and nutritional supplementation via nasogastric
to abnormalities; and (2) utility of serial measurements, feeding. The continuation of breast feeding is recommended
especially with a high negative predictive value. These tests unless there is a true contraindication, such as frequent
may be useful in assessing the resolution of an inflammatory vomiting, gastrointestinal intolerance or a high risk of aspira-
process, including infection, but they are not sufficiently tion. An intravenous preparation containing isotonic saline
precise to establish a diagnosis without additional support- with 5–10% dextrose infused at less than regular mainte-
ing information. Decisions about antimicrobial therapy nance rates is recommended, since free water excretion is
should not be based on inflammatory markers alone. decreased in infants with acute pneumonic infections.
Antimicrobial therapy for atypical respiratory pathogens
TREATMENT is based on the organism.
Antibiotics
There are few randomized treatment trials for neonatal
PREVENTION
pneumonia.1 Treatment decisions are therefore based on Strategies to prevent and treat neonatal pneumonia require
local antimicrobial resistance patterns and clinical cost– interventions at all levels of healthcare provision, i.e. com-
benefit ratios.1 munity, primary care, district and tertiary hospitals.1
202 M. D. NISSEN

Measures already shown to be effective in the preven- 4. Vitamin A supplementationatbirth. The precise mechan-
tion of neonatal pneumonia include: (1) the active manage- ism for this strategy is unknown, but is thought to reduce
ment of premature rupture of maternal chorioamniotic S. pneumoniae colonization in early infancy.49,50
membranes;34 (2) the early and exclusive use of breast 5. Proper management of apnoea and respiratory failure in
feeding;35,36 and (3) avoiding nosocomial pneumonia in neonates – the utilization of alternatives to mechanical
neonatal intensive care units where infections due to ventilation such as continuous positive airway pressure
enteric Gram negative bacilli (E. coli, Klebsiella, Enterobac- (CPAP), and respiratory stimulants such as theophylline
ter and Pseudomonas spp), coagulase negative staphylo- and caffeine.51,52
cocci and multiresistant S. aureus are common.37–39
Bacterial colonization of endotracheal tubes, humidifers,
ventilator tubing, intravenous lines, temperature probes, CONCLUSION
staff equipment (e.g. stethoscopes) and hands, all predate The global impact of neonatal pneumonia is significant and
the onset of neonatal infection. Handwashing is the simplest the epidemiology and aetiology are complex when com-
and most effective intervention for preventing nosocomial pared to the pneumonias in older children. Management
neonatal sepsis.40,41 The identification and cleaning of and prevention strategies for neonatal pneumonias cross
contaminated equipment also prevents nosocomial infec- multiple levels of the population and healthcare provision,
tion.1 and have broader based effects that are sometimes difficult
Restricted antibiotic policies have also been shown to to measure.
prevent neonatal nosocomial infection. The use of benzyl- The growing prevalence of antibiotic resistance to
penicillin and tobramycin or gentamicin for early-onset common and affordable antibiotics will eventually impact
sepsis, and flucloxacillin and tobramycin or gentamicin morbidity and mortality rates for neonates, especially in the
for late-onset sepsis, instead of cefotaxime and ampicillin developing world, and emphasize the importance of con-
resulted in lower neonatal colonization rates with resistant tinuing to develop universal maternal and preventative
bacteria.42 health programmes.

Future directions REFERENCES

1. Duke T. Neonatal pneumonia in developing countries. Arch Dis Child
Other preventative measures for neonatal pneumonia that
Fetal Neonatal Ed 2005; 90: F211–F219.
require ongoing and further investigation include: 2. Black RE, Morris SS, Bryce J. Where and why are 10 million children
dying every year?. Lancet 2003; 361: 2226–2234.
1. Maternal immunization. Maternal immunization with 3. The Child Health Research Project. Reducing perinatal and neonatal
pneumococcal polysaccharide vaccine and the subse- mortality: report of a meeting, Baltimore, Maryland. Baltimore 1999; 3:
quent induction of vaccine protective antibodies for S. 6–12.
4. Barnett ED, Klein JO. Bacterial infections of the respiratory tract. In:
pneumoniae in neonates has been studied in the Gam- Remington JS, Klein JO, ed. Infectious Diseases of the Fetus and
bia, Bangladesh, Philippines and Papua New Guinea.43– Newborn Infant. Philadelphia: WB Saunders, 2001; pp. 1006–1018.
46
Subsequent response to vaccine by children whose 5. Naeye RL, Dellinger WS, Blanc WA. Fetal and maternal features of
mothers were immunized during pregnancy, does not antenatal bacterial infections. J Pediatr 1971; 79: 733–739.
appear to be suppressed.46 6. Bang AT, Bang RA, Tale O et al. Reduction in pneumonia mortality and
total childhood mortality by means of community-based intervention
Regular antenatal monitoring of maternal immunity trial in Gadchiroli, India. Lancet 1993; 336: 201–206.
for rubella antibodies and boosting of rubella immunity 7. Lehmann D, Heywood P. Effect of birth weight on pneumonia-specific
following pregnancy by vaccination is already established and total mortality among infants in the highlands of Papua New
and has lead to a considerable drop in the incidence of Guinea. P N G Med J 1996; 39: 529–537.
8. Shakunthala SKV, Rao GM, Urmila S. Diagnositic lung puncture
congenital rubella in many developed countries. A
aspiration in acute pneumonia of the newborn. Indian Pediatr 1978;
further expansion of this strategy to prevent neonatal 15: 39–44.
pneumonia by maternal immunization will occur when 9. Webber S, Wilkinson AR, Lindsell D et al. Neonatal pneumonia. Arch
maternal vaccines are developed for HSV, cytomegalo- Dis Child 1990; 65: 207–211.
virus and HIV.47 10. The WHO Young Infants Study Group. Bacterial etiology of serious
2. Prevention and treatmentofmaternalsexually transmitted infections in young infants in developing countries: results of a multi-
centre study. Pediatr Infect Dis J 1999; 18: S17–S22.
diseases. Programmes to prevent, detect and treat 11. Muhe L, Tilahun M, Lulseged S et al. Etiology of pneumonia, sepsis and
syphilis, C. trachomatis and HIV would reduce neonatal meningitis in infants younger than three months of age in Ethiopia.
respiratory disease and improve neonatal outcomes in Pediatr Infect Dis J 1999; 18: S56–S61.
areas where maternal rates of these sexually transmitted 12. Mathur NB, Garg K, Kumar. Respiratory distress in neonates with
diseases are high. special reference to pneumonia. Indian Pediatr 2002; 39: 529–537.
13. Misra S, Bhakoo ON, Ayyagari A et al. Clinical and bacterial profile of
3. Cleansing the birth canal with antiseptic solution (chlor- neonatal pneuomonia. Indian J Med Res 1991; 93: 366–370.
hexidene 0.25%) at every vaginal examination before 14. Singhi S, Singhi PD. Clinical signs in neonatal pneumonia. Lancet 1990;
delivery.48 336: 1072–1073.
CONGENITAL AND NEONATAL PNEUMONIA 203

15. Rosenman MB, Mahon BE, Downs SM et al. Oral erythromycin pro- 34. Kenyan S, Boulvain M, Neilson J. Antibiotics for preterm rupture of
phylaxis vs watchful waiting in caring for newborns exposed to Chlamy- membranes. Cochrane Database Syst Rev 2001; 4: CD001058.
dia tracheomatis. Arch Pediatr Adolesc Med 2003; 157: 565–571. 35. Cesar JA, Victora CG, Barros FC et al. Imapct of breast feeding on
16. Preece PM, Anderson JM, Thompson RG. Chlamydia trachomatis admission for pneumonia during postneonatal period in Brazil: nested
infection in infants: a prospective study. Arch Dis Child 1989; 64: 529. case-control study. BMJ 1999; 318: 1316–1320.
17. Prevedoros HP, Lee RP, Marriot D. CPAP, effective respiratory 36. Kirkwood BR, Gove S, Rogers S et al. Potential interventions for the
support in patients with AIDS-related Pneumocystiscarnii pneumonia. prevention of childhood pneumonia in developing countries: a sys-
Anaesth Intens Care 1991; 19: 561–566. temic review. Bull World Health Organ 1995; 73: 793–798.
18. Lehmann D, Sanders RC, Marjen B et al. High rates of Chlamydia tra- 37. Petdachai W. Nosocomial pneumonia in a newborn intensive care
cheomatis infections in young Papua New Guinea infants. Pediatri unit. J Med Assoc Thai 2000; 83: 392–397.
Infect Dis J 1999; 18: S62–S69. 38. Ng SP, Gomez JH, Lim SH et al. Reduction of nosocomial infection in a
19. Were FN, Govedi AF, Revathi G et al. Chlamydia as a cause of late neonatal intensive care unit (NICU). Singapore Med J 1998; 39: 319–323.
neonatal pneumonia at Kenyatta National Hospital, Nairobi. East Afr 39. Pawa AK, Ramji S, Prakash K et al. Neonatal nosocomial infection:
Med J 2002; 79: 476–479. profile and risk factors. Indian Pediatr 1997; 34: 297–302.
20. Duke T, Michael A, Mgone J et al. Etiology of child mortality in Goroka, 40. Webster J, Faoagali JL, Cartwright D. Elimination of methicillin-
Papua New Guinea: a prospective two-year study. Bull World Health resistant Staphylococcus aureus from a neonatal intensive care unit
Organ 2002; 80: 16–25. after hand washing with triclosan. J Paediatr Child Health 1994; 30:
21. Frank D, Duke T. Congenital syphilis at Goroka Base Hospital: 59–64.
incidence, clinical features, and risk factors for mortality. P N G Med 41. Sharek PJ, Benitz Abel NJ et al. Effect of an evidence-based hand
J 2000; 43: 121–126. washing policy on hand washing rates and false-positive coagulase
22. Duke T, Blaschke AJ, Sialis S et al. Hypoxaemia in acute respiratory and negative staphylococcus blood and cerebrospinal fluid culture rates in
non-respiratory illness in neonates and children in a developing a level III NICU. J Perinatol 2002; 22: 137–143.
country. Arch Dis Child 2002; 86: 108–112. 42. de Man P, Verhoeven BA, Verbrugh HA et al. An antibiotic policy to
23. Pillay T, Adhikari M, Makili J et al. Severe, rapidly progressive human prevent emergence of resistant bacilli. Lancet 2000; 355: 973–978.
immunodeficiency virus type 1 disease in newborns with coinfections. 43. O’Dempsey TJ, McArdle T, Ceesay SJ et al. Immunization with
Pediatr Infect Dis J 2001; 20: 404–410. pneumococcal polysaccharide vaccine during pregnancy. Vaccine
24. Aiken CG. HIV-1 infection and perinatal mortality in Zimbabwe. Arch 1996; 14: 963–970.
Dis Child 2004; 67: 595–599. 44. Shahid NS, Steinhoff MC, Hoque SS et al. Serum, breast milk, and
25. Skevaki CL, Kafetzis DA. Tuberculosis in neonates and infants: epi- infant antibody after maternal immunisation with pneumococcal.
demiology, pathogenesis, clinical manifestations, diagnosis, and man- Lancet 1995; 346: 1252–1257.
agement issues. Paediatr Drugs 2005; 7: 219–234. 45. Quaimbao BP, Nohynek H, Kayhty H et al. Maternal immunization
26. Semple MG, Cowell A, Dove W et al. Dual infection of infants by with Pneumococcal polysaccharide vaccines in the Phillippines. Vaccine
human metapneumovirus and human respiratory syncytial virus is 2003; 21: 3451–3455.
strongly associated with severe bronchiolitis. J Infect Dis 2005; 191: 46. Lehmann D, Pomat WS, Riley ID et al. Studies of maternal immuniza-
382–386. tion with pneumococcal polysaccharide vaccine in Papua New Guinea.
27. Faix RG. Congenital pneumonia. eMedicine. www.emedicine.com/ Vaccine 2003; 21: 3446–3450.
ped/topic2765.htm (accessed 29 May 2007). 47. Munoz FM, Englund JA. A step ahead. Infant protection through
28. World Health Organization. Management of the Child with a Serious maternal immunization. Pediatr Clin North Am 2000; 47: 449–463.
Infection or Severe Malnuitrition: Guidelines for Care at First-Referral Level 48. Taha TE, Biggar RJ, Broadhead RL et al. Effect of cleansing the birth
in Developing Countries. Geneva: WHO, 2000. canal with antiseptic solution on maternal and newborn mortality in
29. English M, Mohammed S, Ross A et al. A randomized, controlled trial Malawi: clinical trial. BMJ 1997; 315: 216–219.
of once daily and multi-dose gentamicin in young Kenyan infants. Arch 49. Humphrey JH, Agoestina T, Wu L et al. Impact of neonatal vitamin A
Dis Child 2004; 89: 665–669. supplementation on infant morbidity and mortality. J Pediatr 1996;
30. Musake RN, Revathi G. Emergence of multi-drug resistant gram- 128: 489–496.
negative organisms in a neonatal unit and the therapeutic implications. 50. Rahmathullah L, Tielsch JM, Thulasiraj RD et al. Impact of supplement-
J Trop Paediatr 2000; 46: 86–91. ing newborn infants with vitamin A on early infant mortality: com-
31. Aurangzeb B, Hameed A. Neonatal sepsis in hospital-born babies: munity based randomised trial in southern India. BMJ 2003; 327: 254–
bacterial isolates and antibiotic susceptibility patterns. J Coll Physicians 259.
Surg Pak 2003; 13: 629–632. 51. Pieper CH, Smith J, Maree D et al. Is nCPAP of value in extreme
32. Rahman S, Hameed A, Roghani MT et al. Multidrug resistant neonatal preterms with no access to neonatal intensive care? J Trop Paediatr
sepsis in Peshawar, Pakistan. Arch Dis Child Fetal Neonatal Ed 2000; 87: 2003; 49: 148–152.
F52–F54. 52. Henderson-Smith DJ, Subramaniam P, Davis PG. Continuous
33. Duke T, Michael A. Increase in sepsis due to multi-resistant enteric gram positive airway pressure versus theophylline for apnea in
negative bacilli in Papua New Guinea. Lancet 1999; 353: 2210–2211. preterm infants. Cochrane Database Syst Rev 2001; 4: CD001072.