Case Report
Bacterial Pericarditis and Tamponade Due to Nonencapsulated Haemophilus influenzae Complicating a Case of Adult Community-Acquired Pneumonia

Praveen Garg, MD; Ruchi Gupta, MD; James E. Szalados, MD, JD, MBA, FCCP 

Medscape General Medicine.  2006;8(4):48.  ?2006 Medscape
Posted 12/07/2006

 

We report a case of bacterial pericarditis in an immunologically competent adult female caused by nonencapsulated Haemophilus influenzae (H influenzae) that was complicated by the acute development of life-threatening pericardial tamponade.

H influenzae is a gram-negative coccobacillus, a pathogen most frequently associated with childhood exanthema (otitis media, meningitis) and, less frequently, adult pneumonia. Encapsulated, type b, or typable H influenzae is the strain implicated in childhood infections. On the other hand, nonencapsulated or nontypable H influenzae is the specific strain most often associated with exacerbation of chronic obstructive airway disease.

Bacterial pericarditis caused by either subtype of H influenzae is exceedingly rare. We have located only 15 previously reported cases of H influenzae pericarditis occurring in adults in the world medical literature, the majority of which date back to the pre-antibiotic era. In 12 of these 15 cases (the only cases in which typing could be accomplished), the encapsulated strain of H influenzae was cultured from the pericardial fluid.

Thus, to the best of our knowledge, we are reporting here the first case of bacterial pericarditis caused by nonencapsulated H influenzae in an immunologically competent adult.


Readers are encouraged to respond to George Lundberg, MD, Editor of MedGenMed, for the editor's eye only or for possible publication via email: glundberg@medscape.net

Case Report

A previously healthy 78-year-old white female was admitted complaining of sore throat and dysphagia of 6 days' duration and 2 days of progressively worsening dyspnea. Review of systems was otherwise noncontributory; the patient specifically denied symptoms of cough, chest pain, or palpitations. The past medical history was significant only for mild asthma and osteoporosis, for which she was prescribed salmeterol/fluticasone and alendronate, respectively. Social history elicited no reporting of tobacco or alcohol abuse.

On initial clinical evaluation, the patient was afebrile (37?C), tachycardic (116 beats/min), and profoundly hypotensive (70 mm Hg/palpable); saturation by pulse oximetry was 94% with 2 L/min supplemental oxygen. Physical examination revealed an erythematous posterior pharyngeal wall without exudative discharge, normal-looking tonsils, no cervical adenopathy, and internal jugular venous distension 10 cm above the sternal angle in the sitting position. Chest auscultation revealed bi-basilar crackles and an unremarkable cardiac auscultatory examination; specifically, there were no audible murmurs, gallops, or rubs. A significant clinical finding was a clearly demarcated erythematous area over the mid-chest; but there was no associated cutaneous injury, fluctuance, discharge, or crepitus.

Laboratory data were remarkable for a white blood cell count of 12,000/mL with 57% bands, hematocrit 36%; platelet count 310,000/mL, normal liver function tests, serum bicarbonate 32 mmol/L; normal troponin I (< 0.10 pg/mL) blood urea nitrogen 34 mg/dL, serum creatinine 2.3 mg/dL, and a negative urinalysis and culture. Throat culture was negative for streptococcus.

An anteroposterior (AP) portable chest radiograph demonstrated early and diffuse bilateral lower lobe infiltrates with small bilateral pleural effusions. A lateral-view radiograph of the neck was unremarkable. Specifically, there was no evidence of posterior pharyngeal or laryngeal soft tissue abnormality and the prevertebral space was normal. A 12-lead electrocardiogram (ECG) showed diffuse but nondiagnostic ST-segment elevation (Figure 1).

Figure 1. 

Admission electrocardiogram demonstrating diffuse and nonspecific ST-segment elevation in leads I, II, aVL, aVF, and V2-V6. There is also PR-segment depression in lead II).

     

Transthoracic 2-dimensional echocardiography (TTE) demonstrated a globally reduced left ventricular systolic function, an estimated left ventricular ejection fraction of 35%, and a small pericardial effusion but with no associated compression of the right atrium.

The patient was admitted with the presumptive diagnosis of pericarditis and was initially empirically treated with naproxen and fluid restriction. Empiric intravenous antibiotic therapy was initiated with moxifloxacin, clindamycin, and ceftriaxone to treat the possibility of a coexisting community-acquired pneumonia as a potential cause of the sepsis syndrome.

However, the patient continued to deteriorate clinically. Progressive hypoxemia with a widened alveolar-arteriolar oxygen gradient, persistent refractory hypotension, tachycardia, metabolic acidosis (serum bicarbonate 16 mmol/L), and profound oliguria supported the clinical impression of evolving septic shock. Serial chest radiographs showed worsening interstitial infiltrates consistent with pulmonary edema. The patient was transferred to the critical care service, where a pulmonary artery catheter (PAC) was inserted to better determine the pathophysiology of the worsening shock, evaluate the relative contributions of pericarditis and sepsis, and guide fluid and pharmacologic management. The initial hemodynamic parameters were: central venous pressure (CVP) 8 cm of water, pulmonary arterial diastolic pressure (PAD) 10 cm of water, and pulmonary capillary wedge pressure (PCWP) 14 cm of water.

Hypoxia and imminent respiratory failure necessitated endotracheal intubation and initiation of mechanical ventilation. Significant intravascular volume expansion with crystalloid bolus infusion therapy was also initiated in the ICU and was thereafter titrated to cardiac filling pressures, thermodilution cardiac output, and mixed venous oximetry. The clinical state of shock improved after intravascular volume infusion with increased urine output, blood pressure, and heart rate; however, the patient remained profoundly hypotensive (systolic blood pressure 60 mm Hg) despite aggressive vasopressor support with both vasopressin and norepinephrine.

The following day (hospital day 2), a computer-assisted tomography (CT) scan of the chest revealed only bilateral pleural effusions; specifically, the CT did not show any significant pericardial fluid. A follow-up ECG revealed generalized low voltage complexes but a normalization of the previous ST-segment abnormalities (Figure 2).

Figure 2. 

ECG shows resolution of ST segment elevation, presence of tachycardia, and low voltage complexes.

     

Serum troponin levels remained negative. Blood cultures that had been drawn on admission tested positive for H influenzae. Later that day, the PAC data for the first time showed equalization of cardiac filling pressures (CVP: 14 cm water; PAD 14 cm water; PCWP 14 cm water), prompting an emergent follow-up TTE. The repeat TTE clearly demonstrated the interval development of a large pericardial effusion associated with both right atrial and right ventricular collapse in diastole, thereby confirming the diagnosis of acute pericardial tamponade (Figure 3).

Figure 3. 

Apical 4-chamber echocardiogram demonstrating a pericardial effusion. PE = pericardial effusion; LV = left ventricle; RV = right ventricle; LA = left atrium; RA = right atrium.

     

Emergent therapeutic pericardiocentesis was performed at the ICU bedside and approximately 500 mL of thick, brown, purulent, bacterial pericardial fluid was removed by needle aspiration (Figure 4).

Figure 4. 

Purulent material obtained after pericardiocentesis.

     

A pericardial drainage catheter was then left in situ. Immediately following pericardiocentesis, the patient's hemodynamic status progressively improved. A diagnostic bedside thoracocentesis was performed later the same day but produced only a minimal quantity of clear serous fluid. Pressors and volume replacement were rapidly tapered in the setting of improved blood pressure, heart rate, and urine output. Gram-stain specimens of both the pericardial and pleural fluid specimens revealed abundant neutrophils but no bacteria. The pericardial fluid cultured H influenzae, which was forwarded to the New York State Department of Health for serotyping using latex agglutination test. The results came back as the nontypable strain of H influenzae. Pleural fluid cultures were negative. Sputum cultures were negative but had > 25 polymorphonuclear neutrophils per high-power field.

The patient made a remarkable clinical recovery. Ceftriaxone was continued for a total of 4 weeks. Echocardiography was repeated 3 weeks later and showed only a small residual pericardial effusion without evidence of functional cardiac sequelae.

Discussion

The implications of this case are 3-fold: first, we present here, to the best of our knowledge, the first case of spontaneous bacterial pericarditis caused by nonencapsulated H influenzae occurring in an immunocompetent adult; second, we report on a rare complication of pericarditis -- purulent bacterial tamponade; and third, this case illustrates the utility of the PAC in the diagnosis and treatment of cardiac tamponade.

We describe an unusual case of H influenzae community-acquired pneumonia in an immunocompetent and previously healthy adult patient progressing to septicemia, bacterial pericarditis, and finally pericardial tamponade. Bacterial pericarditis caused by H influenzae has thus far been reported in only 15 adults.[1-8] H influenzae type b (encapsulated, typable) was the organism cultured from the pericardial fluid in 12 of these 15 patients.[9] These patients ranged in age from 18 to 59 years; 9 of the 15 patients presented with pharyngitis; pneumonia and empyema were present in 9 patients, and cardiac tamponade occurred in 7 of 15 patients.[10-13] On the other hand, the strain of H influenzae cultured from the pericardial fluid in the present case was nonencapsulated. Nonencapsulated or nontypable H influenzae is a well-recognized pathogen commonly implicated in community-acquired pneumonia, particularly in the elderly population. After comprehensive review of the world medical literature, we found 1 case of nonencapsulated H influenzae in an immunocompromised patient[14]; however, the case we describe here represents the first finding of nonencapsulated H influenzae causing bacterial pericarditis in an otherwise healthy adult.

Cardiac tamponade and obstructive shock due to the accumulation of pus in the pericardium is the most serious complication of bacterial pericarditis, thereby forming the basis for the term 'purulent' pericarditis. Prompt diagnosis and percutaneous drainage of the pericardial pus is essential. Bacterial seeding of the pericardium is known to occur following instrumentation such as pacemaker or catheter placement, cardiothoracic surgery, trauma, or direct local spread from a pulmonary, myocardial, or subdiaphragmatic focus. Spontaneous bacterial pericarditis may also occur in patients with immunologic suppression due to chemotherapy, malignancy, or acquired immunodeficiency syndrome.

In this case, the patient presented with a presumptive diagnosis of pericarditis. The classic clinical findings associated with acute pericarditis (chest pain, pericardial friction rub, and pulsus paradoxicus) are actually demonstrated in only 50% of patients.[15] Beck's triad of hypotension, increased venous filling pressure, and muffled heart sounds is more diagnostic but are also rare. Signs of inflammation such as fever, leukocytosis, and elevated erythrocyte sedimentation rate are present in virtually all patients, but these findings are also very nonspecific. In this patient, the signs of inflammation may have been suppressed by nonsteroidal anti-inflammatory agents that were initially administered for acute pericarditis. The majority of patients with bacterial pericarditis have typical ECG changes associated with acute pericarditis; however, the ECG may be normal in 10%-35% of the patients. Normalization of the diffuse ST segment elevation on our patient's ECG may have been attributable to increasing pericardial effusion.

Since pericarditis causes inflammation of the pericardium, a generalized state of systolic and diastolic dysfunction results, and manifests as cardiogenic shock. Patients with pericarditis do not tolerate elevated cardiac filling pressures and are likely to respond to fluid boluses by developing atrial fibrillation or pulmonary edema. Therefore, patients with pericarditis are typically treated with fluid restriction. However, patients with pericardial tamponade instead are in a state of obstructive shock and are extremely preload-dependent; they require significant volume resuscitation to maintain right heart filling and cardiac output. Pericardial tamponade is a result of rapid accumulation of fluid in the pericardial sac, which restricts diastolic filling of the right ventricle and consequently limits right ventricular stroke volume and cardiac output, subsequently reducing left ventricular preload and cardiac output.

The gold standard for the diagnosis of pericardial tamponade is echocardiographic demonstration of right heart chamber collapse due to accumulation of pericardial fluid.[16] However, well-recognized limitations of echocardiography include the interval nature and operator-dependence of data collection. In this case, early serial imaging did not reveal fluid accumulation or tamponade despite worsening clinical shock. The PAC was a key diagnostic tool in this case since the acute onset of cardiac tamponade would have been overlooked on the basis of data obtained from the TTE alone. Equalization of intracardiac pressures (CVP, right atrial, left atrial, PCWP, left ventricular)[17] confirmed the diagnosis of cardiac tamponade. Therefore, although the use of the PAC has become controversial,[18] this case supports the contention that it may improve clinical outcomes when used judiciously, correctly, and with due recognition of its potential limitations.[19,20]

It is interesting to speculate as to the pathogenesis of H influenzae pericarditis in our case. Presence of sepsis and positive blood cultures suggests hematogenous dissemination rather than direct extension from the lung.

In summary, H influenzae must be considered to be a possible causative agent of pyogenic pericarditis in immunocompetent elderly patients. Specifically, the confirmed isolation of nontypable H influenzae from the bacterial pericardial drainage from this patient may represent a shift in the virulence, pathogenicity, and epidemiology of this organism. Respiratory symptoms, pneumonia, systemic infection, and pericarditis in the setting of nontypable H influenzae must be considered to be prodromal to the development of bacterial pericarditis and pericardial tamponade. In the setting of refractory shock, the use of complementary adjunct modalities to evaluate cardiovascular function should be considered. Specifically, serial intermittent echocardiography may, at least in some instances, need to be supplemented by continuous measurement of hemodynamic indices using a pulmonary artery catheter.

References

  1. Hirschmann JV, Everett ED. Hemophilus influenzae infections in adults: report of nine cases and a review of literature. Medicine (Baltimore). 1979;58:80-94.
  2. Norden CW. Hemophilus influenzae infections in adults. Med Clin North Am. 1978;62:1037-1046.
  3. Sink JD, Spray TL, Rankin JS. Hemophilus influenzae pericarditis. Clin Cardiol. 1982;5:547-549.
  4. Graham BS, Reiss TF, Gretory DW. Pericarditis associated with Hemophilus influenzae type B pneumonia and bacteremia in two adults. Chest. 1983;84;48-50.
  5. Mier A. Bilateral empyema and bacterial pericarditis due to Hemophilus infuenzae capsular type b. Thorax. 1989;44:80.
  6. Jenning HS, Eskind JB, Savage AM, Harston WE. Hemophilus influenzae bacterial pericarditis with cardiac tamponade. South Med J. 1983;76:812-813.
  7. Buckingham TA, Wilnel G, Surgar SJ. Hemophilus influenzae pericarditis in adults. Arch. Intern Med. 1983;143:1809-1810.
  8. Yeh YH, Chu PH, Yeh CH, et al. Hemophilus influenzae pericarditis as the initial presentation of systemic lupus erythematosis. Int J Clin Pract. 2004;58:1045-1047.
  9. Rubin RH, Moellering RC Jr. Clinical, microbiologic, and therapeutic aspects of bacterial pericarditis. Am J Med. 1975;59:68-78.
  10. Sagrista-Sauleda J, Barrabes JA, Permaneyer-Miralda G, Soler-Soler J. Bacterial pericarditis: review of a twenty year experience in a general hospital. J Am Coll Cardiol. 1993;22:1661-1665.
  11. Iggo R, Higgins R. Bilateral empyema and purulent pericarditis and bacterial pericarditis due to Hemophilus infuenzae capsular type b. Thorax. 1988;43:582-583.
  12. Leggier RJ, Balsam D. Hemophilus sepsis leading to pericarditis despite antimicrobial therapy. Johns Hopkins Medical Journal. 1980;146:133-136.
  13. Welikovitch L, Knight JL, Burggraf GW, Sanfilippo AJ. Cardiac tamponade secondary to Hemophilus pericarditis: A case report. Can J Cardiol. 1992;8:303-305.
  14. Ligtenberg JM, van derWerf TS, Zijlstra JG, Tulleken JE, Fijen JW, Moller LVM. Non-surgical treatment of bacterial pericarditis, due to non-encapsulated Hemophilus influenzae, in an immunocompromised patient. Neth J Med. 1999;55:151-154.
  15. Pankuweit S, Ristic A, Seferovic PM, Maich B. Bacterial pericarditis: diagnosis and management. Am J Cardiovasc Drugs. 2005;5:103-112.
  16. Bjorkhem G, Lundstrom NR, Vitarelli A. Sequential study of echocardiographic changes in bacterial pericarditis. Pediatric Cardiol. 1984;5:317-321.
  17. Jones JW, Izzat NN, Rashad MN, et al. Usefulness of right ventricular indices in early diagnosis of cardiac tamponade. Ann Thorac Surg. 1992;54:44-49.
  18. Connors AF Jr, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA. 1996;276:889-897.
  19. Pinsky MR, Vincent J-L: Let us use the pulmonary artery catheter correctly and only when we need it. Crit Care Med 2005;33:1119-1122.
  20. Sakr Y, Vincent JL, Reinhart K, et al. Use of the pulmonary artery catheter is not associated with worse outcome in the ICU. CHEST 2005;128:2722-2731.

Praveen Garg, MD, Assistant Professor of Medicine, Department of Medicine, University of Florida, Jacksonville

Ruchi Gupta, MD, Housestaff, Department of Medicine, University of Florida, Jacksonville

James E. Szalados, MD, JD, MBA, FCCP, Attending Physician and Medical Director of Respiratory Care, Department of Anesthesiology and Critical Care Medicine, Unity Health System, Rochester, New York

Author's Email:
Praveen.Garg@jax.ufl.edu

Disclosure: Praveen Garg, MD, has disclosed no relevant financial relationships.

Disclosure: Ruchi Gupta, MD, has disclosed no relevant financial relationships.

Disclosure: James E. Szalados, MD, JD, MBA, FCCP, has disclosed no relevant financial relationships.