Cases & Commentaries

Missed Pneumonia

Commentary By Jeffrey M. Rohde, MD, and Scott A. Flanders, MD

The Case

A 32-year-old man presented to the emergency department (ED) with 3 days of fever and right pleuritic chest pain. Review of systems was negative for cough or dyspnea, and medical, surgical, and social history were unremarkable. The patient had taken a 2-hour plane ride the day before onset of symptoms.

On arrival to the ED, the patient was tachycardic at 106 beats per minute, tachypneic at 24 breaths per minute, and febrile to 38.8° C orally. No heart or lung abnormalities were noted on exam. Initial workup included a normal complete blood count (CBC) and basic metabolic panel (BMP), an elevated D-dimer (666 μg/mL), and a moderate right pleural effusion on chest radiograph. Although the clinicians had a high suspicion for pulmonary embolism (PE), a computed tomography angiogram (CTA) chest demonstrated moderate right pleural effusion without evidence of PE or infiltrate. The patient was discharged home with prescriptions for oxycodone/acetaminophen and ibuprofen. Discharge diagnoses were "fever, pleural effusion, and chest wall pain."

The patient returned to the ED 3 days later reporting worsening pain and continued fever with new cough and dyspnea. Examination was significant for tachycardia at 105 beats per minute, mild tachypnea at 20 breaths per minute, and diminished right-sided lung sounds. Laboratory test results, including CBC and BMP, were unremarkable. However, a chest radiograph now showed a right-sided effusion with consolidative changes, and a thoracentesis was consistent with parapneumonic effusion. The patient was started on antibiotics and admitted for pneumonia with effusion. The patient quickly improved on antibiotics and was discharged 2 days later.

The Commentary

Every year, more than 4 million Americans are diagnosed with community-acquired pneumonia (CAP), resulting in significant morbidity, mortality, and a total cost of more than $10 billion.(1) Guidelines dictate that the diagnosis of CAP requires the presence of clinical features suggestive of acute respiratory infection coupled with an acute infiltrate demonstrable by chest radiograph or other imaging modality.(2) As this case demonstrates, however, the clinical signs and symptoms associated with CAP are nonspecific, and radiographic imaging is not always diagnostic, resulting in potential for misdiagnosis.

Most literature regarding misdiagnosis of pneumonia describes overdiagnosis: attributing a constellation of clinical signs as symptoms to pneumonia ultimately found to be caused by another condition. In one study, 27.3% of the patients admitted from the emergency department (ED) with a diagnosis of pneumonia ultimately had another diagnosis on discharge.(3) Rates of overdiagnosis may have increased over time, a trend that may be driven, at least in part, by the Centers for Medicare & Medicaid Services (CMS) core performance measures focusing on timely antibiotic administration for CAP.(4) Although equally important, underdiagnosis of pneumonia, such as this case, is not as well-described.

Diagnosing pneumonia requires integration of a number of symptoms and physical exam findings, as no one symptom or sign is sufficient to make the diagnosis. Instead, patients typically present with any number of respiratory symptoms combined with a variety of signs indicating infection. More than 90% of patients admitted with CAP have cough, two-thirds of which will be productive. Additionally, 66% of patients have dyspnea, and 50% have pleuritic chest pain. On physical examination, more than 90% of patients with CAP have a focal lung exam, while fever (80%), tachypnea (70%), and tachycardia (50%) are also commonly found.(5) Likelihood ratios for diagnosing pneumonia associated with common signs and symptoms are presented in the Table. Although these findings can be helpful, all are nonspecific and, especially in patients with comorbidities, are consistent with several non-infectious disease entities including pulmonary embolism (PE), congestive heart failure (CHF), asthma, or chronic obstructive pulmonary disease (COPD) exacerbation. A thorough history and physical exam must be combined with sound clinical judgment to differentiate these conditions from pneumonia and establish a pre-test probability of disease prior to further diagnostic testing.

Although radiographic imaging is a cornerstone in the diagnosis of CAP, studies examining its utility reveal a surprising lack of sensitivity. In one study, the sensitivity of chest radiographs in patients with autopsy-confirmed ventilator-associated pneumonia (VAP) was only 68%.(6) Another study found that 21% of patients hospitalized with a diagnosis of CAP had normal initial chest radiographs on admission.(7) Many of these patients later developed infiltrates on repeat radiographs or had infiltrates seen on computed tomography (CT) scan, reinforcing that each radiological study is only one picture in time. A developing infectious process dictates the need for continued monitoring. Although chest CT scans increase the accuracy of diagnosing pneumonia, even these high-resolution images do not always identify a diagnostic infiltrate, as evidenced by this case.

The patient described in this case was found to have a moderate pleural effusion but no infiltrate. Parapneumonic effusions develop in 20%–40% of patients admitted with bacterial pneumonia and are associated with a 3.7–6.5-fold increased risk of mortality.(8) Isolated effusions, occurring in the absence of an infiltrate, are associated with viral, tuberculous, and limited bacterial pulmonary infections, but the overall incidence is unknown. Because the differential diagnosis of pleural effusions is quite broad, in the absence of an obvious explanation for a moderate-sized effusion (as in this case), thoracentesis was indicated. Although it is possible that the pleural fluid analysis might have facilitated the diagnosis of pneumonia, up to 27% of pleural effusions remain undiagnosed.(9) However, thoracentesis could have ruled out other important diagnoses such as empyema, hemothorax, and malignancy.

Given the challenges associated with diagnosing patients presenting with respiratory complaints, there has been considerable interest in the utility of biomarkers. For example, several well-validated clinical prediction rules have been combined with measurement of D-dimer to assign pre-test probabilities in patients with possible PE. However, D-dimer levels, while very sensitive, are notoriously nonspecific, being normal in only 40%–68% of patients without PE.(10) Specificity further declines in the setting of another active disease, renal dysfunction, or increased patient age. This particular patient's Wells clinical prediction tool for PE results in a score of 1.5 (points are for tachycardia), resulting in a pre-test probability of 1.3% for PE.(11) With such a low probability, a negative D-dimer would have effectively ruled out a PE. A low Wells score combined with an elevated D-dimer is considered inconclusive, justifying further imaging, such as a PE-CT, which in addition to evaluating for a PE, may provide further clues to alternative diagnoses. This patient's elevated D-dimer, however (which was likely due to his undiagnosed infection), was erroneously attributed as further evidence for a PE and further masked the true diagnosis.

Inflammatory biomarkers have been evaluated as diagnostic tools in patients suspected of having pneumonia. The most promising one is procalcitonin (PCT), a peptide precursor of calcitonin produced in response to inflammation caused by bacterial infections. The addition of PCT measurements to physician estimates of the probability of pneumonia resulted in a diagnostic accuracy of more than 86% in all patients. The PCT levels were particularly useful in patients with a history of CHF and COPD.(12) Indeed, PCT has even been shown to be valuable in differentiating patients with CAP from those with PE.(13) Although an elevated serum PCT in this case might have suggested the diagnosis of pneumonia, falsely low serum PCT levels have been reported, even with parapneumonic effusions and empyema.(14) At this point, the role of PCT in the clinical diagnosis of pneumonia remains uncertain.

Although a careful history and physical, combined with radiographic imaging and selected use of biomarkers, may facilitate the diagnosis, there is no "holy-grail" diagnostic test, no foolproof clinical prediction rule, and no checklist that ensures an accurate diagnosis of pneumonia in every patient. We must rely on the diagnostic skills of clinicians and a certain mindfulness that not all diagnoses are obvious at first or even second pass, particularly for those infectious etiologies that can evolve over time. If diagnostic uncertainly exists, firm plans must be put in place to monitor the signs and symptoms for either evolution or resolution over time.

The current patient was "diagnosed" with "fever, pleural effusion, and chest wall pain," representing a description of the discovered signs and symptoms, not a diagnosis. A thoracentesis should have been performed to facilitate a diagnosis once the pleural effusion was discovered. If the thoracentesis had been nondiagnostic, a mindful approach would have been called for, including a realization that serious conditions, like CAP, had not been ruled out and a choice must be made to pursue either empiric therapy or careful observation.

Take-Home Points

  • Respiratory and infectious issues are commonly encountered in the clinic and ED. Providers must recognize the signs and symptoms that suggest bacterial pneumonia in the face of other etiologies.
  • No one symptom, sign, physical exam finding, or radiographic finding is sufficient to diagnose pneumonia.
  • Biomarkers can be helpful in differentiating respiratory processes, but they are nondiagnostic and must be considered in terms of the overall clinical context.
  • A mindful approach to patient care requires full evaluation of all serious abnormalities encountered, with processes put in place for continued monitoring should a diagnosis remain elusive.

Jeffrey M. Rohde, MD
Assistant Professor of Internal Medicine

University of Michigan Medical School

Scott A. Flanders, MD
Professor of Internal Medicine

Associate Division Chief of General Medicine for Inpatient Programs

Director, Hospital Medicine Program

University of Michigan Medical School

References

1. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med. 1997;336:243-250. [go to PubMed]

2. Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis. 2007;44(suppl 2):S27-S72. [go to PubMed]

3. Chandra A, Nicks B, Maniago E, Nouh A, Limkakeng A. A multicenter analysis of the ED diagnosis of pneumonia. Am J Emerg Med. 2010;28:862-865. [go to PubMed]

4. Wachter RM, Flanders SA, Fee C, Pronovost PJ. Public reporting of antibiotic timing in patients with pneumonia: lessons from a flawed performance measure. Ann Intern Med. 2008;149:29-32. [go to PubMed]

5. Sharpe BA, Flanders SA. Community-acquired pneumonia: a practical approach to management for the hospitalist. J Hosp Med. 2006;1:177-190. [go to PubMed]

6. Wunderink RG, Wolenberg LS, Zeiss J, Day CM, Ciemins J, Lacher DA. The radiologic diagnosis of autopsy-proven ventilator-associated pneumonia. Chest. 1992;101:458-463. [go to PubMed]

7. Hagaman JT, Rouan GW, Shipley RT, Panos RJ. Admission chest radiograph lacks sensitivity in the diagnosis of community-acquired pneumonia. Am J Med Sci. 2009;337:236-240. [go to PubMed]

8. Light RW. Parapneumonic effusions and empyema. Proc Am Thorac Soc. 2006;3:75-80. [go to PubMed]

9. Sahn SA. The diagnostic value of pleural fluid analysis. Semin Respir Crit Care Med. 1995;16:269-278. [Available at]

10. Stein PD, Hull RD, Patel KC, et al. D-dimer for the exclusion of acute venous thrombosis and pulmonary embolism: a systematic review. Ann Intern Med. 2004;140:589-602. [go to PubMed]

11. Wells PS, Anderson DR, Rodger M, et al. Excluding pulmonary embolism at the bedside without diagnostic imaging: management of patients with suspected pulmonary embolism presenting to the emergency department by using a simple clinical model and d-dimer. Ann Intern Med. 2001;135:98-107. [go to PubMed]

12. Maisel A, Neath SX, Landsberg J, et al. Use of procalcitonin for the diagnosis of pneumonia in patients presenting with a chief complaint of dyspnoea: results from the BACH (Biomarkers in Acute Heart Failure) trial. Eur J Heart Fail. 2012;14:278-286. [go to PubMed]

13. Kötürk N, Kanbay A, Bukan N, Ekim N. The value of serum procalcitonin in differential diagnosis of pulmonary embolism and community-acquired pneumonia. Clin Appl Thromb Hemost. 2011;17:519-525. [go to PubMed]

14. Christ-Crain M, Schuetz P, Müller B. Biomarkers in the management of pneumonia. Expert Rev Respir Med. 2008;2:565-572. [go to PubMed]

15. Metlay JP, Kapoor WN, Fine MJ. Does this patient have community-acquired pneumonia? Diagnosing pneumonia by history and physical examination. JAMA. 1997;278:1440-1445. [go to PubMed]

Table

Table. Likelihood ratio* for pneumonia given the presence of findings or symptoms.(15)

Exam Finding Positive Likelihood Ratio Symptoms Positive Likelihood Ratio
Respiratory rate > 25 1.5–3.4 Cough 1.8
Heart rate > 100 1.6–2.3 Dyspnea 1.4
Temperature > 37.8C 1.4–4.4 Fever 1.7–2.1
Crackles 1.6–2.7 Chills 1.3–1.7
Rhonchi 1.5–1.4 Myalgias 1.3
Bronchial breath sounds 3.5 Sore throat 0.78
    Rhinorrhea 0.78

*Likelihood ratios above one elevate the probability of the disease; those below one lower the probability.