Pneumonia
Laennec described the traditional findings of lobar pneumonia back in in 1819 and what he noted was a progression of findings: Bronchial breath sounds and bronchophony appeared early and dullness to percussion appeared later. These were pre-antibiotic and pre-CXR findings. Today we probably diagnose milder PNA and antibiotics change the natural history of the disease.
So yet again we turn to JAMA’s Rational clinical exam and Steven McGee’s Evidence Based Physical Diagnosis for guidance.
In this JAMA Rational Clinical Exam article, the authors discusses the validity of diagnosing PNA without CXR and which clinical elements improve the yield of using CXR for PNA diagnosis.
History
There are NO elements from the history that reliably exclude or confirm a diagnosis of PNA.
A subjective fever has a +LR of 2.1 in one of the reviewed studies on outpatients, but has only a 5.5% PPV as the study had a very low PNA incidence (Diehr et al., JAMA 1997).
Physical Exam
Vital Signs
Of vital signs, the JAMA authors found low LR+’s for vital signs (~2) and per Mcgee’s Evidence Based Physical Diagnosis only having an SpO2 < 95% is the vital sign mildly indicative of PNA with a LR+ of 3.1.
If ALL vital signs are normal, there is a moderate decrease in the likelihood of PNA, with a LR 0.3 per Mcgee’s pooled findings. However, the JAMA authors note that, in a study where CXRs were not ordered on the basis of normal vital signs, 38% of radiographic PNAs were missed!
It should be noted that many studies evaluating exam findings were done on patients presenting with cough and fever and that CXR was the diagnostic standard.
Chest Examination
Chest exam findings tend to be highly specific for PNA but are not often present, so absence does not exclude disease.
Even if all of the described exam findings are negative, the LR- is only 0.57 (95% CI, 0.39-0.83) [JAMA 1997]. Lower (“better”) values for absence of all findings have been reported retrospectively but not validated.
| Finding | LR+ | LR- | Kappa |
| Asymmetric chest expansion | 44.1 | NS | 0.38-0.85 |
| Egophany | 4.1 | NS | ? |
| Bronchial breath sounds | 3.3 | NS | 0.19-0.32 |
| Crackles | 2.3 | 0.8 | 0.21-0.65 |
| Diagnostic Score | |||
| Heckerling Score 4-5 | 8.2 | – | N/A |
| Heckerling Score 0-1 | 0.3 | – | N/A |
The Table above is largely adapted from Mcgee’s Evidence Based Physical Diagnosis but includes JAMA’s findings as well. The major disagreement is with the agreement (Kappa) for abnormal chest expansion with 0.38 comping from JAMA and 0.85 from McGee.
The finding of crackles is included in the table above because of its often perceived association with PNA. The JAMA authors state “the presence or absence of crackles on examination [is not] sufficient to rule in or rule out” PNA.
This score has five historical and clinical findings:
- Absence of asthma
- Temperature >37.8°C (100°F)
- Heart rate >100 beats/min
- Decreased breath sounds
- Crackles
The JAMA authors are less enthusiastic about Heckerling’s score than McGee. Metlay et al. stress in JAMA that this rule would not perform well in settings with low pre-test probability for PNA, such as the outpatient setting.
Summary
- “There are no combinations of history and physical examination findings that confirm the diagnosis of pneumonia.”
We agree – but in a patient presenting with subjective fever and cough and is found to have an SpO2 <95%, the findings of Asymmetric chest rise, Egophany if present or a Heckerling score of 4-5 would be sufficient for a working diagnosis of PNA.
- “If diagnostic certainty is required in the management of a patient with suspected pneumonia, then [CXR] should be performed.”
This is likely more important in the outpatient setting, as ED and inpatients will likely have CXR performed regardless of presentation. At this point it seems nothing clinical can replace CXR. Except for maybe ultrasound…
