|
|
||||
|
|
The national clinical
examination. Does this infant have pneumonia? CLINICAL
SCENARIO A mother brings her 8-month-old infant to
your office in midwinter with a bad cough. She reports that the illness began 4 days ago
with a runny nose. Two days
ago, the baby developed a fever. Now the baby's symptoms are
getting worse. The baby has become more irritable, is eating less, and seems
to be having more difficulty breathing. This is the third child you have seen
today with a cough. Why Is This an Important Question to Answer With Clinical Examination?
Acute respiratory illnesses are among the most common conditions of
infants seen in primary care offices. Although the majority of respiratory
illnesses involve infections of the upper respiratory tract, most infants will experience a lower
respiratory tract illness (LRI) in the first year of life. Of those with LRIs, a]
about 30% visit a physician, and about 2% are hospitalized. Lower respiratory
tract illnesses can be defined simply as infections at an anatomic level
below the vocal cords. The majority of LRIs in infants are caused by viruses;
only a small proportion are due to bacteria. The
differential diagnosis for cough is long (Table 1). Therapies are available to treat a variety of manifestations of lower
respiratory tract disease, so it is important to diagnose these complaints
accurately and estimate their severity to deliver the appropriate treatment.
Identifying infants at lower risk of bacterial disease may help clinicians :1.void the unnecessary use of antibiotics.
This may reduce the risk of subsequent bacterial infection and slow the
emergence of resistant strains of bacteria within the population.4 Greater
certainty about the presence of a viral LRI may also help clinicians avoid
additional testing such as radiography or blood culture. This overview
focuses on the history and physical examination findings of infants that
distinguish pneumonia from other LRIs. METHODS We conducted a
MEDLINE search from 1982 to 1995 to identify articles about the diagnosis of
pneumonia in children. We searched for articles with any of the following
Medical Subject Heading (MeSH) terms: pneumonia, diagnostic tests,
sensitivity and specificity, reproducibility of results, physical
examination, or medical history taking. This search was further
limited to studies published in English about humans that involved children.
This search strategy identified 38 articles. Four more articles were
identified by reviewing a compendium of references prepared by the World
Health Organization.5 Etiologic studies, which did not include a chest x-ray
examination as part of the "gold standard," involved only
inpatients. Studies of illness in families' homes, rather than in clinical
settings, were excluded (n=29).
All the articles were reviewed by the authors and disagreements were
re- solved by discussion. We used the methods developed for this series to
assess the quality of the articles. The highest- quality studies are
emphasized in the "Results" section. We did not aggregate
statistically the results of the studies be- cause of differences in the ages
of the study samples and differences in cut off points of key variables, such
as respiratory rate. Confidence intervals were calculated according to the
method suggested by Koopman.6.7 Reference Standard for Diagnosing Pneumonia The reference
standard for diagnosing pneumonia is an aspirate from the lower respiratory
tract obtained by bronchoaveolar lavage or lung puncture. The use of
bronchial lavage is appropriate in guiding antibiotic choice in patients with
refractory or complicated pneumonia. In general practice, chest x-ray
films are readily obtained and can be considered a pragmatic reference
standard for pneumonia. A number of studies
evaluated the accuracy of the chest x-ray film in differentiating viral from
bacterial disease in children. It is difficult to determine the accuracy of
the chest x-ray film from these studies because of methodologic limitations, as well as challenges to study design
F introduced by the biology of pneumonia. (;1 It is
not possible to obtain cultures from a til lung in most patients. Therefore,
investigators have had to use combinations of other clinical features as a
proxy for bacterial pneumonia. Reliance on less than ti perfect gold
standards for diagnosing bacaterial pneumonia may produce over or
underestimates of the association of a SI positive chest x-ray finding with
bacterial pneumonia. Two studies used the same definition of bacterial
pneumonia (duration of symptoms <2 days, temperature >39.5°C, total white
blood cell count > 15.0x 109/L}. Both found the sensitivity of the chest
x-ray film for diagnosing o bacterial vs. viral pneumonia to be approximately
75%. However, one reported a specificity of 100%, the other, a specificity of
63%. The reported sensitivity for studies with varying definitions ranges
from 42% to 80% and the specificity from 42% to 100%. Studies of the accuracy
of el chest x-ray films have also been compromised by other methodologic
problems, el such as interobserver variability in the al interpretation of
the x -ray film, oversampling patients with relatively severe disease, and
the relatively small numbers of patients with bacterial pneumonia. Such
problems make estimates of chest x-ray film accuracy unreliable. Variation in the biologic manifestations of bacterial pneumonia also
presents challenges in the interpretation of el published studies. For
example, bacterial bi pneumonia is classically associated with 10- pi bar
consolidation on the x-ray film. However, studies report that bacterial
pneumonia may be associated with infiltrates that are lobar, perihilar,
segmental, interstitial, or nodular infiltrates}4-16 Consolidation can also
be seen with viral pneumonia, but it is unclear if this radiologic m
appearance is due to segmental consolidation, atelectasis, or bacterial
coinfection. Such variability in the radiographic appearance of bacterial
pneumonia may produce over- or underestimates of the association of a
positive chest x-ray finding with bacterial pneumonia. Clinicians should be aware that the chest x-ray finding may be
negative in patients with early bacterial pneumonia. The sensitivity of the
chest x-ray film will be reduced in this group. The implications of this
observation are important for studies of the clinical examination. For the
purposes of this systematic review, we included studies that used the chest
x-ray film as the reference standard. Studies that combined the clinical
diagnosis with the chest x-ray film results as the reference standard were
excluded because inclusion of the diagnostic test in the reference standard
may overestimate the accuracy of clinical findings. The significance of
clinical findings of pneumonia in the absence of a positive chest x-ray
findings remains to be studied. Normal Anatomy and Pathophysiology of PneumoniaLower respiratory tract infections occur at or below the larynx and
include epiglottitis, laryngitis, laryngotracheo- bronchitis (croup),
bronchiolitis, and pneumonia (Figure 1). Pneumonia typically follows an upper
respiratory tract illness in which the lower respiratory tract is invaded by
bacteria, viruses, or other pathogens that trigger the immune response and
produce inflammation. Histamines, leukotrienes, and chemotactic factors are
released that recruit white blood cells to the area. This response fills the air spaces of the lower
respiratory tract with white blood cells, fluid, and cellular debris. This
process reduces lung compliance, increases resistance, obstructs smaller
airways, and may result in collapse of distal air spaces. The resultant physical findings vary with the site of infection,
ranging from coarse breath sounds or rhonchi in bronchopneumonia to crackles
in the alveoli in cases of pneumonia or bronchiolitis. Crackles are the
result of the explosive 41 equalization of gas pressure between the terminal
bronchiole and the alveoli.1B Wheezes result from the oscillation of air
through a narrowed airway that produces a musical sound likened to a
vibrating reed. Decreased breath sounds may also be noted in areas of
consolidation. How to Elicit the Relevant
Symptoms and Signs There are 2 goals of the history and physical examination in a child
who presents with a cough. The first is to identify the clinical syndrome;
the second is to estimate the severity of the illness. In taking the history,
the examiner should ask the parent about symptoms associated with pneumonia,
as well as those that may discriminate pneumonia from other lower respiratory
tract diseases. In addition to cough, symptoms that may increase the likelihood
of pneumonia include trouble breathing, rattling in the chest, noisy
breathing, trouble feeding, fever, rapid breathing, anxiety, or restlessness.
Clinicians working in different regions; or with different cultures, need to
familiarize themselves with local terminology for lower respiratory tract
symptoms. It may also be useful to ask bout prior episodes of these chest
symptoms because recurrent bouts of pneumonia or bronchitis may suggest
reactive airway disease. In early infancy (<2 months), infants of mothers
who had clamydia during pregnancy may develop afebrile pneumonia. Clinicians
should note that infants only rarely produce sputum. In older infants,
foreign body ingestion and salicylate poisoning should be considered.
Although clinical experience suggests that the history of pneumonia may be of
acute or gradual onset and that bacterial pneumonia tends to be associated
with fever, we were unable to find any studies substantiating these
observations. The physical examination should include an assessment of
the child's general appearance, measurement of the respiratory rate,
evaluation of the work of breathing, and auscultation of the chest. The
child's general appearance may provide important clues about the presence , bacterial illness and its severity. Infants can
exhibit a wide range of behaviors and mood changes during the parental
interview, while being undressed, and during the physical examination.
Therefore, it is important to take a non-threatening approach with the young
child. Infants should be observed initially at a distance, while they are
comforted, usually the caretaker's lap. The assessment of general appearance
should include an evaluation of a number of factors: attentiveness to the
environment, ability to 'breast-feed or drink, ability to sustain sucking,
vocalization, smiling, movement, color, and consolability. If there is
uncertainty about particular findings, it may be helpful to try to elicit
them. For example, encouraging the child to smile, having the mother offer the
breast, or showing the child a toy to engage his or her attention. Respiratory
rates change considerably the first year of life, decreasing from a mean
in awake babies of about 50/min at 1 week of age to about 40/min at 6 months
of age. The respiratory rate in children can also vary over brief intervals
as the child's level of interest in the environment changes, or while the
child is asleep or feeding. Polygraphic studies of infants younger than 6
months have demonstrated that mean respiratory rates were 4/min to 13/min
higher in active sleep (rapid eye movement) than in quiet sleep. Fever can
also increase an infant's respiratory rate by 10/min per degree centigrade in
children without pneumonia. However, the effects of fever in the presence of
pneumonia have not been studied. The respiratory rate is best measured by observing chest wall
movements over 1 minute.Listening to the chest with a stethoscope may
stimulate the child and lead to a falsely elevated measurement. Measurement
errors in counting the respiratory rate are greater when children are
agitated or crying compared with when they are calm, feeding, or sleeping.
The examiner should count the respiratory rate before conducting other parts
of the examination. Respiratory rate cutoffs that are commonly used to
indicate an elevated rate are greater than 60/min in infants youngerthan2
months, greater than 50/min in infants 2 through 12 months of age, and
greater than 40/min in children older than 12 months. Assessing an infant's work of breathing is important to estimate the severity of
LRI. This assessment includes evaluation of chest wall movements, nasal
flaring, and grunting. Chest wall movements include retractions or chest
indrawing, best observed with the chest fully exposed. Supraclavicular retractions
may be observed as indrawing of the soft tissue above the clavicle or above
the sternal notch. Intercostal retractions are seen as indrawing of skin
between the ribs. Subcostal retractions occur on or just below the costal
margin. Many experts suggest that these types of retractions, involving only
the soft tissue, should be distinguished from chest wall indrawing, defined
as an inward movement of the lower chest wall (ie, ribs) when the child
breathes in. Chest indrawing is more likely to be observed in infants younger
than 18 to 24 months whose chest walls are more compliant. The finding may be
appreciated best by viewing the chest laterally and looking for indrawing of
the ribs and/or lower sternum with inspiration, relative to a fixed point beyond
the child's chest that is set as a mental reference point (Figure 2).
Normally, the costal margin moves little during quiet breathing. If it does,
it moves up and outward because the normal diaphragm lifts the costal margin
outward. In disease states, the depressed diaphragm may apply an inward
traction on the chest resulting in paradoxical movement of the chest wall
during inspiration. Therefore, in airway obstruction, the costal margin tends
to move paradoxically (ie, down and inward). Sometimes, the abdomen moves
outward while the chest moves inward during inspiration. This has also been
called
Nasal flaring is enlargement of both openings of the nose
during inspiration. It is due to constriction of anterior and posterior
dilators naris muscles. Grunting is a repetitive short upper
respiratory tract sound produced by partial vocal cord closure during
expiration. Grunting slows expiratory flow and increases lung volume and
alveolar pressures. It can be a sign of severe disease and suggests impending
respiratory failure. Examiners should be aware that the presence of signs of
increased work of breathing may change with the state of the child. For
example, chest wall indrawing may be present only when the child is awake or
more active. Adventitious
sounds that can be
appreciated on auscultation include discontinuous or popping sounds, sounds
that occur throughout the inspiratory or expiratory phase, or continuous
sounds. Discontinuous sounds have been called crackles, rales, or
crepitations. They typically occur at the end of inspiration. Continuous
sounds include wheezes and rhonchi and can be musical, high- or low-pitched,
inspiratory or expiratory, short or long, or monophonic or polyphonic. Clinicians
should try to distinguish whether sounds are continuous or discontinuous
before applying a name. Many clinicians differentiate continuous sounds that
are whistling or high-pitched (usually called wheezes) from low-pitched,
snoring or rattling sounds (usually called rhonchi). Many experts consider
wheezes to reflect small airway obstruction (ie, bronchioles), while rhonchi
reflect obstruction of the large airways (ie, bronchi). It is interesting to note that the language used to describe
auscultatory findings can be a source of confusion. For example, rhonchi and
rales are, respectively, the Latin and French words for crackles. Indeed,
Laennec (the inventor of the stethoscope) distinguished 6 types of
crackles.He believed that only 1 of these was associated with pneumonia. Auscultation of the chest is often more difficult in infants when they
are crying. For this reason, it should be performed after the visual
inspection of the child. It is important to listen to the front, back, and
sides of the infant's chest because adventitious sounds may only be heard in
1 location. Even when the infant is crying, adventitious sounds may be heard
at the end of inspiration when the infant is quiet and about to take a
breath. Examiners should also be aware that wheezes can often be appreciated
by listening to the sounds of breaths from infants
mouths (audible wheezing). Finally, infants may have several types of
adventitious sounds present although this is more common in reactive airway
disease or viral LRI). Text- books do not recommend percussion of the hest in
infants because it is difficult to get infants to cooperate with this
maneuver.
Are These Symptoms or Signs Ever Premature infants and neonates may app ear to have chest indrawing
during normal breathing or exertion. Grunting and groaning noises occur from
time to time in normal healthy infants. An infant who is very playful may
demonstrate increased respiratory rate, intercostal retractions, and
increased work of breathing. RESULTS The Precision of Symptoms and Signs A total of 56 patients with lower respiratory tract symptoms were
examined by pairs of general pediatricians from a group that included
academic pediatric generalists, pediatric residents, and pe- iatricians in
community practice. Agreement was good for most signs on physical examination
that could be observed by inspection, including the social interaction
markers of attentiveness K, 0.49), smiling (K, 0.51), quality of cry K,
0.63); physical appearance and movement (K, 0.54), color (K, 0,66),
respiratory effort retractions (K, 0.48), and use of accessory muscles (K,
0.59). There was only fair agreement about most auscultatory findings:
prolonged expiratory phase (K, .22), adventitious sounds (K, 0.3), and
inspiratory wheezing (K, 0.29). Agreement was good for audible wheezing (K,
.7) and for expiratory wheezing (K, .63). In general, physicians agreed more
often that a finding was present than when it was absent. A second study
indicated that observers are less likely to agree about the severity of
findings than about their presence or absence. Several studies of
the precision of the respiratory rate suggest that respiratory rates counted
over 30 seconds average 2/min to 4/min faster than respiratory rates counted
over 60 seconds. Counting the respiratory rate over 30 seconds will lead to
more abnormal rates and may spuriously increase the number of children
diagnosed as having pneumonia. More accurate results are obtained t if the
average of two 30-second counts is t taken or one 60-second count. Observer agreement
is good for most signs on the physical examination. There is better agreement
about signs that can be observed than signs that require auscultation of the
chest. The Accuracy of Signs of Pneumonia The reported accuracy of clinical findings varies considerably among
studies because of methodologic limitations and (
differences in the spectrum of illness severity among sites in which
the studies were conducted. In most reports, chest x-ray films were used as
the gold standard and children who had clinical findings suggestive of
pneumonia were more likely to have had an x-ray examination than those who
did not (Table 2). Although this approach makes sense clinically, it
introduces verification bias that tends to overestimate a test's sensitivity
and underestimate its specificity.4° Two studies, both of which were con- ducted in developing countries,
at- tempted to overcome the problem of selective ordering of the gold
standard by obtaining chest x -ray films on all children with abnormal
clinical findings (eg, elevated respiratory rate), as well as a sample of
children without abnormal findings. The reported accuracy was then adjusted
statistically for the fraction of patients sampled in each group. These 2 studies
found that there was no single sign that could be used to rule in or rule out
pneumonia definitively .In these studies, children with elevated respiratory
rates were about twice as likely to have pneumonia (positive likelihood ratio
[LR+ ], 1.5-2.1) as children without elevated respiratory rates (Table 3).
Conversely, those without elevated respiratory rates were only about 0.36 to 0.5 times as likely
to have pneumonia. These studies also found that the presence of chest
indrawing (retractions) increased the likelihood of pneumonia (LR+, 2.4- .n
2.5). However, normal chest movements did not rule out pneumonia(negative
likelihood ratio [LR- ],0.7-0.78). Other useful findings that increased the
likelihood of
pneumonia included nasal flaring (LR+, 1- 3.0), and crepitations
(LR+, 3.5). Once again, the absence of nasal flaring and crepitations did not
effectively lower the likelihood of pneumonia: no nasal flaring, LR-, 0.71;
no crepitations, LR-, 0.69. Other studies in developing countries, even
though less methodologically sound, found the accuracy of clinical signs to
be more or less in the same range as that found in the 2 more well-designed
investigations (Table 3). The lower prevalence of bacterial disease and severe pneumonia found
in developed countries might suggest that the accuracy of physical
examination signs would be lower than that reported in studies from
developing countries. However, the few studies performed in developed
countries reported results similar to those cited above. These studies may have
overestimated the accuracy of clinical findings because chest x-ray films
were more likely to be obtained in patients with signs and symptoms of
disease. In a study by Leventhal,so the absence of
tachypnea, as observed by the clinician examining the patient, was useful for
ruling out pneumonia (LR-, 0.32), while the presence of tachypnea somewhat
increased the odds of pneumonia (LR+, 2.03). Grunting and crepitations were
more useful in ruling in disease (LR+,3.17 and 2.1,
respectively). Their absence did not appreciably lower the likelihood of
disease (LR-, 0.86 and 0.73). The study by Taylor et al reported a somewhat
higher LR+ for tachypnea (LR+, 3.22), but this study included only febrile
children and chest x-ray films were not obtained for all study patients. A study by Crain et al included only infants with fever younger than 8
weeks who were seen in an emergency department. The authors reported that
tachypnea (LR+, 8.0; 95% confidence interval (CI], 5.3-12.1) and chest
indrawing (LR+, 26.0; 95% CI, 2.7-118.8) substantiallyin- creased the
likelihood of pneumonia. Although these likelihood ratios are quite high, the
number of patients with pneumonia in this study was small and the reported
estimates are imprecise (as indicated by the wide 95% CIs). In addition, the
high likelihood ratios also reflect the high specificity of tachypnea and
indrawing in a particular group of patients (early infants). The value of the
clinical examination may differ in this group of children. As in other
studies, the absence of these findings did not dramatically lower the
likelihood of disease for tachypnea (LR-, 0.55) or for indrawing (LR-, 0.75).
Accuracy
of Combinations of Findings
Clinicians typically evaluate the presence of many findings
simultaneously to rule in or
rule out pneumonia. Despite the large number of studies, few have examined
the value of clinical findings when they are used together. Two studies
assessed the value of combinations of clinical findings. Leventhal found that
the absence of pulmonary findings defined as respiratory distress (nasal
flaring, grunting, retractions), tachypnea, rales, or decreased breath sounds
ruled out pneumonia (LR-, 0.0; 95% CI, 0.0- 0.4). When present, these
findings raised the likelihood of pneumonia to 1.6 (95% CI, 1.3-31.3). In
this study, information about the presence or absence of respiratory symptoms
was used in the decision to obtain the gold standard examination (a chest
x-ray examination). Thus, the reported data are likely to overestimate the
diagnostic accuracy of these combinations of findings so that the true LR- is
not as good as reported and the LR+ is better than reported. In a study of
children younger than 2 months, Grain et al found that the absence of any
respiratory findings (rhinorrhea, cough, adventitious sounds, or retractions)
lowered substantially the likelihood of a positive chest x-ray finding (LR-,
0.10; 95% CI, 0.03-0.4). The presence of any of these findings increased the
likelihood of pneumonia to 3.4 (95% CI, 2.6-4.3). Again, since this study
included only infants younger than 8 weeks, it is not clear how well the
results apply to older age groups. Grain et al also found that as the number
of positive respiratory findings increased, so did the probability of an abnormal
chest x-ray finding. To summarize,
physical examination findings can help primary care physicians be more
certain that an infant does or does not have pneumonia. In developed
countries, where the prevalence of bacterial pneumonia is lower, pneumonia is
unlikely if all signs are negative. The presence of a positive sign will be
more useful in increasing clinicians' certainty that an infant has pneumonia
in developing countries as compared with developed countries because the
prevalence of bacterial pneumonia is higher. In developed countries,
clinicians will be more certain if multiple findings are positive. Further
studies are needed to examine the diagnostic accuracy of the chest x-ray
examination, the value of certain signs (such as fever and toxic appearance),
as well as how to best take advantage of combinations of clinical findings.
THE BOTTOM LINE First, the initial observation of the infant may be the most critical
component of the examination. Observation is important before interacting
with a child. Secondly, because
of its moment-to-moment variability, the respiratory rate should be counted
by observing the chest while the child is quiet over tow 30-second intervals
or over a full minute. Clinicians need to be especially aware of the
variability of the examination as the child’s level or activity changes. . Third, auscultation is relatively unreliable when examining infants.
Clinicians need better training and better terminology to describe abnormal
chest sounds. The overall clinical appearance may be accurate but the
delineation of its value needs more study. Fourth, the best individual finding for ruling out pneumonia is the
absence of tachypnea. Chest indrawing and other signs of increased work of
breathing (eg, nasal flaring) and abnormal auscultatory findings are better
for ruling in pneumonia. In developed countries, multiple findings must be
present to be more certain about the presence of pneumonia. Fifth, if all
clinical signs (respiratory rate, auscultation, and work of breathing) are
negative, the chest x-ray finding is unlikely to be positive. Margolis P, Gadomski A. Does this infant have pneumonia?
JAMA 1998 Jan 28;279( ):308-313. PubMed
|
|||
