Tracheal Aspirate as a Substrate for Polymerase Chain Reaction Detection of Viral Genome in Childhood Pneumonia and Myocarditis
Background—Infectious respiratory disorders are important causes of childhood morbidity and mortality. Viral causes are common and may lead to rapid deterioration, requiring mechanical ventilation; myocardial dysfunction may accompany respiratory decompensation. The etiologic viral diagnosis may be difficult with classic methods. The purpose of this study was to evaluate polymerase chain reaction (PCR) as a diagnostic method for identification of causative agents.
Methods and Results—PCR was used to amplify sequences of viruses known to cause childhood viral pneumonia and myocarditis. Oligonucleotide primers were designed to amplify specific sequences of DNA virus (adenovirus, cytomegalovirus, herpes simplex virus, and Epstein-Barr virus) and RNA virus (enterovirus, respiratory syncytial virus, influenza A, and influenza B) genomes. Tracheal aspirate samples were obtained from 32 intubated patients and nucleic acid extracted before PCR. PCR results were compared with results of culture, serology, and antigen detection methods when available. In cases of myocarditis (n=7), endomyocardial biopsy samples were analyzed by PCR and compared with tracheal aspirate studies. PCR amplification of viral genome occurred in 18 of 32 samples (56%), with 3 samples PCR positive for 2 viral genomes. Amplified viral sequences included RSV (n=3), enterovirus (n=5), cytomegalovirus (n=4), adenovirus (n=3), herpes simplex virus (n=2), Epstein-Barr virus (n=1), influenza A (n=2), and influenza B (n=1). All 7 cases of myocarditis amplified the same viral genome from heart as found by tracheal aspirate.
Conclusions—PCR is a rapid and sensitive diagnostic tool in cases of viral pneumonia with or without myocarditis, and tracheal aspirate appears to be excellent for analysis.
Infectious disorders of the respiratory tract are important causes of morbidity and mortality in children.1 In the immunocompetent host, viral and bacterial pneumonias are predominant,2 whereas opportunistic infections are common in those children who are immunocompromised.3 In both subgroups of patients, rapid respiratory and metabolic deterioration requiring intubation and mechanical ventilation may occur. Cardiac dysfunction caused by myocarditis may accompany respiratory decompensation.4 5 6 7 Frequently, the causative agent is unknown at the time of intubation. Traditionally, bacterial and viral cultures of blood and respiratory secretions, serologic studies, and rapid antigen tests have been used to identify causative agents. In many cases, disease origin cannot be determined from these studies. Even with more invasive methods such as bronchoalveolar lavage (BAL),8 lung biopsy,9 and endomyocardial biopsy, identification of the causative agent my be unsuccessful.
For this reason, more sensitive diagnostic methods have been sought. Polymerase chain reaction (PCR) has been shown to be useful in the identification of causative agents for several infectious disorders (including myocarditis, meningitis, AIDS, and others) through the use of samples from different tissues (including heart and lung), blood, and many body fluids.10 11 12 In the present study, tracheal aspirates obtained from intubated children were found to be useful in the rapid diagnosis of viral pneumonitis with or without accompanying myocarditis. The study has 2 primary hypotheses: that tracheal aspirate PCR results are predictive of findings by conventional diagnostic methods and that tracheal aspirate PCR results are predictive of PCR results from myocardial specimens.
All study patients were intubated and admitted to the Pediatric Intensive Care Unit at Baylor College of Medicine (Houston, Tex). Informed consent was obtained before enrollment. Standard diagnostic studies (chest radiography, nasal-wash ELISA) viral/bacterial culture of body fluids and secretions, serology, BAL, and lung or cardiac biopsy) were ordered at the discretion of the patient’s physician.
Tracheal aspirate samples were collected in a sterile fashion after instillation of 4 to 10 mL sterile normal saline into the endotracheal tube by suctioning into a sterile trap (closed system). Phlegm samples (nonintubated) were obtained after forced cough and collected in a sterile trap. All samples were transported on ice and stored at 4°C.
In cases of suspected myocarditis, right ventricular endomyocardial biopsy (EMB) was performed. Specimens were cultured, fixed in formalin for histopathological evaluation with the “Dallas criteria,”13 or snap-frozen in liquid nitrogen immediately. In all cases of myocarditis, EMB and tracheal aspirate specimens were obtained within 1 hour of each other.
PCR was performed by an investigator blinded to patient information. All samples were analyzed for the presence of nucleic and sequences specific for adenovirus,14 cytomegalovirus (CMV),15 enterovirus,16 respiratory syncytial virus (RSV),17 Epstein-Barr virus (EBV),18 herpes simplex virus (HSV),19 and influenza A and B20 with primers (Table 1⇓) on the basis of published sequences. Results were confirmed by Southern blotting21 and hybridization22 and/or DNA sequencing of PCR amplimers in some cases.23 The HSV, CMV, and EBV primers were shown to have sequence specificity (ie, no cross-reactivity).
The principal study group consisted of consecutively identified patients 1 day to 16 years of age who recently were endotracheally intubated for respiratory decompensation. A subset of these patients (n=8) also had myocarditis.
Three control groups were used for comparison of PCR findings. Patients intubated for open heart surgery or after trauma with no evidence of viral-induced diseases or recent viral illness were used as control group 1 (n=20). This group provided tracheal aspirate samples for PCR analysis. A second group of subjects (control group 2) had recent upper respiratory tract illnesses that had resolved 2 or 3 weeks before study (n=30). These children were included to test whether viral genome is amplified from the phlegm of previously infected individuals (ie, latency) and were selected on the basis of clinical findings. Inclusion criteria included the combination of rhinorrhea, nasal stuffiness, sore throat, cough, and fever (100.5°F to 102°F). None had evidence of pneumonia or pharyngeal exudate, and none was treated with antibiotics. Control group 3 included children with no recent infectious history (n=30) who were hospitalized and intubated before death. Lung tissue obtained after autopsy provided the samples for PCR analysis in control group 3.
Primer Design and Synthesis
Six pairs of primers were designed to detect viruses implicated as causes of viral pneumonitis, including enteroviruses (consensus sequence),16 CMV,15 adenovirus,14 HSV,19 EBV,18 RSV,17 and influenza A and B20 as previously described11 12 14 15 16 17 18 19 20 (Table 1⇑). One primer pair was designed to amplify a constitutive product of all cells, the K-ras oncogene,18 which was used to demonstrate adequate nucleic acid extraction and exclude false-negative results resulting from a lack of extracted nucleic acid.
Template Preparation and PCR
Total RNA and DNA were isolated simultaneously from tracheal aspirate, phlegm, lung, and EMB specimens by use of a modification of the RNAzol method24 with Tris-saturated phenol (pH 6.6) RNAzol solution.11 12 25 26
Reverse transcriptase–PCR was used to evaluate the RNA viruses (enteroviruses, RSV, influenza A, and influenza B); PCR was used to evaluate DNA viruses (adenovirus, CMV, HSV, and EBV).11 12 25 26 We analyzed 10 μL of each reaction on a 2% moderate EEO (ME) agarose gel (FMC Biochemicals) or 3% Nu Sieve agarose (FMC Biochemicals) and 0.5% ME agarose gel containing 0.5 μg/mL ethidium bromide (Sigma Chemical Co). The gels were placed under UV light for visualization of amplified products.
All samples were run with simultaneous positive and negative controls (ie, reaction mixture without sample nucleic acid) for the virus analyzed. If a band was visualized in the negative control lane, the PCR sample was considered contaminated and reanalyzed. For the PCR amplimer to be considered positive, reproducibility of the product was required. Control PCR amplification to verify the presence of amplifiable nucleic acid extracted from each sample was performed with K-ras primers.
Contingency table analyses and 2-tailed Fisher’s exact tests were used to test the 2 primary hypotheses. To address the hypothesis that tracheal aspirate PCR results are predictive of findings by conventional diagnostic methods, the latter were considered the “gold standard,” and the sensitivity and specificity of PCR were determined. To address the hypothesis that tracheal aspirate PCR results are predictive of PCR results from endomyocardial biopsy specimens, the latter were considered the gold standard, and the sensitivity and specificity of tracheal aspirate PCR were determined.
Viral cultures were transported, processed, and inoculated into cell culture through standard virological techniques as previously described.11
Indirect immunofluorescent antibody analysis against the viral capsid antigen of EBV was used on paired sera to identify recent EBV infection. Antigen detection was performed for identification of RSV (Abbott Test Pack), influenza viruses (Becton Dickinson Microsystems), HSV, and CMV (Syva Microtrak, Syva Co). Complement fixation studies were performed to identify elevated enteroviral titers. The Bartels Viral Respiratory Screening and Identification Kit (Baxter Diagnostics) was also used to identify adenovirus, influenza viruses, and RSV (direct fluorescent antibody analysis) from tracheal aspirates.
Table 2⇓⇓ presents the primary diagnosis, culture, and PCR results for the principal study group and each control group. Tracheal aspirate PCR was predictive of etiologic results obtained from conventional diagnostic methods (P<0.0001). Assuming that the conventional diagnostic methods represent the gold standard, the sensitivity and specificity of tracheal aspirate PCR were 79% and 96%, respectively. Although the number of subjects with both tracheal aspirate and endomyocardial biopsy samples was small (n=10), the sensitivity of tracheal aspirate PCR for predicting the PCR results on EMB specimens was 100% (P<0.05).
Results of culture of respiratory secretions were available for 31 of 32 patients (all except patient 17; Table 2⇑⇑) in the principal study group, with 9 samples (9 of 31, 29%) having positive results (1 influenza A, 4 enterovirus, and 5 CMV); 1 patient (patient 28; Table 2⇑⇑) had a tracheal aspirate culture positive for both CMV and enterovirus. No control subject had positive viral culture. In all cases in which EMB specimens were obtained, EMB culture was negative except for patient 3 (Table 2⇑⇑), in whom culture of the EMB identified enterovirus.
A rapid screen for RSV was performed on nasal washes from all patients in the principal study group with clinical presentation of bronchiolitis and reactive airways disease (n=7), with 5 of the 7 patients (71%) positive (patients 1, 2, 7, 8, and 30; Table 2⇑⇑).
EMB results were available for 10 patients. Seven showed changes consistent with acute myocarditis (patients 3, 4, 20 through 22, 25, and 28; Table 2⇑⇑), 2 had histopathological features of dilated cardiomyopathy (patients 14 and 23; Table 2⇑⇑), and 1 from a patient whose status was posttransplantation had allograft rejection.
In the principal study group, PCR amplification of a viral genome was identified in 18 of 32 tracheal aspirate samples (56%), 3 of which were positive for 2 viruses (patient 9, adenovirus and CMV; patient 13, HSV and CMV; patient 1, adenovirus and enterovirus; Table 2⇑⇑). Analysis of the 21 positive PCR amplimers in these 18 patients demonstrated the following: RSV (n=3), enterovirus (n=5), CMV (n=4), adenovirus (n=3; Figure 1A⇓), HSV (n=2), EBV (n=1; Figure 2A⇓), influenza A (n=2), and influenza B (n=1). All samples were positive for the presence of K-ras, providing evidence that no inhibitors were present that would compromise assay sensitivity. Therefore, samples that were PCR negative for viral genome but positive for K-ras were assumed to be without viral genome in the sample, or the viral load was assumed to be below the sensitivity of the method. Myocardial PCR results were identical to those obtained from tracheal aspirate PCR in all cases. In all control group samples, PCR was negative (control group 1, 0 of 20; group 2, 0 of 30; and group 3, 0 of 30).
Correlation Between Clinical Diagnosis, Culture, Serology, ELISA, and PCR
Tracheal aspirates from 11 of 14 patients (79%) with culture, serology, or ELISA identification of a viral origin were positive by PCR for the identical virus. In addition, 4 patients in whom conventional methods were negative had PCR-positive results. In 3 children with a single virus identified by conventional methods, PCR also identified a second viral sequence. Comparison of the specific viruses is given below by clinical diagnosis.
Of the 5 patients with nasal washes positive for RSV by ELISA, 3 were also positive by PCR. All 5 had bronchiolitis with respiratory failure, but the 2 with negative PCR results were late in the course of disease when PCR testing was undertaken (ie, 12 to 14 days after intubation). A sixth patient with bronchiolitis also had a hypoplastic left ventricle and double-outlet right ventricle (patient 12; Table 2⇑⇑) and was negative for RSV by rapid screen and viral culture; PCR was also negative.
Reactive Airways Disease
Influenza A virus was isolated from nasal wash and tracheal aspirate from 1 patient with reactive airways disease (patient 6; Table 2⇑⇑). This patient was also found to have influenza A viral genome by PCR (Figure 3⇓).
Of 7 patients with pneumonia, 2 had positive respiratory viral cultures, both for CMV (patients 13 and 32; Table 2⇑⇑). PCR results in each case agreed with culture results. PCR on a third patient with pneumonia was positive for enterovirus, but tracheal aspirate culture was negative. BAL from 1 patient with acute myelocytic leukemia and pneumonia was culture negative, but tracheal aspirate PCR was positive for HSV; the same patient subsequently had herpetic skin lesions identified. One patient with Down’s syndrome, reactive airways disease, and pneumonia (patient 13; Table 2⇑⇑) had a tracheal aspirate that was positive by culture only for CMV but was positive by PCR for both CMV and HSV. This patient had herpetic skin lesions; HSV was isolated from nasal wash, whereas serology was positive only for CMV.
Adult Respiratory Distress Syndrome
Culture of nasal wash from 1 of 3 patients with adult respiratory distress syndrome (ARDS) was positive for CMV (patient 9; Table 2⇑⇑), although tracheal aspirate culture was negative. PCR of tracheal aspirate was positive for CMV. CMV IgM in the patient’s serum was also significantly elevated.
Respiratory viral culture results were positive for enterovirus in 3 of 5 patients presenting with myocarditis and pneumonia. Tracheal aspirates on all 3 were also positive for enterovirus by PCR (patients 3, 20, and 21; Table 2⇑⇑). Culture of lung biopsy was negative for 1 of these patients (patient 3; Table 2⇑⇑). A fourth patient who was positive for EBV by tracheal aspirate PCR (Figure 2A⇑) but negative by culture was also positive for EBV by convalescent serology. PCR of an EMB specimen performed 1 week later also was positive for EBV genome (Figure 2B⇑). In fact, all cases of myocarditis in which biopsy was performed had PCR results of heart specimens identical to the results of tracheal aspirate screening. The fifth patient (patient 22; Table 2⇑⇑), although PCR positive for adenovirus from both samples (tracheal aspirate and EMB) obtained 12 days apart, was negative by culture and serology. DNA sequencing of the PCR amplimers was consistent with adenovirus type 2 (Figure 1B⇑). Culture of tracheal aspirate was positive for both CMV and enterovirus in another patient with acute myocarditis (patient 28; Table 2⇑⇑); PCR of tracheal aspirate and EMB was also positive for CMV and enterovirus.
Nasal-wash culture from 1 patient with acute myocarditis (patient 4; Table 2⇑⇑) was positive for CMV, but tracheal aspirate was negative for any organism. PCR on the tracheal aspirate specimen was positive for enterovirus, as was the EMB PCR.
All 8 patients were negative by culture and PCR.
The rapidity, sensitivity, and specificity of PCR analysis of respiratory secretions for viral diagnosis have previously been well documented. However, each of these previous studies was done to detect a specific viral genome—RSV,27 influenza viruses,28 CMV,29 and HIV30 —and only 4 few analyzed a purely pediatric population. Three of these studies describe detection of RSV in nasal washes, and a fourth reports amplification of HIV-1 from pediatric BAL specimens. Our study differs in 2 respects. First, tracheal aspirates from intubated children were used as the substrate for PCR in the present report. Because this is the most easily obtained specimen from the lower respiratory tract, we believe this technique has advantages over other more invasive sampling methods. Second, on each specimen, we used a cocktail of primers specific for a variety of common respiratory tract viruses. The results described here demonstrate that PCR can be used on tracheal aspirates for the rapid screening of presumed virus-induced injury of the lung. In addition, tracheal aspirate PCR was extremely sensitive and specific for predicting EMB PCR results in children with combined respiratory and cardiac disease or cardiac disease alone (in which myocarditis was a likely cause of cardiac dysfunction).
Three of 5 patients with RSV bronchiolitis were positive by PCR, slightly less than expected on the basis of previous studies that showed a higher sensitivity.31 All these prior studies report storing the clinical specimens at −70°C, and 1 reports adding RNase inhibitor to the specimens.27 The 2 PCR-negative specimens in our study were obtained late in the clinical course (ie, 3 days) and were left at room temperature for 48 hours before placement at 4°C for 2 weeks (because of storage by the clinical caregivers); the PCR-positive specimens were stored immediately at 4°C or −20°C and analyzed within 3 days. Thus, it is possible that our negative results in these 2 cases were due to the relative lability of single-stranded RNA in the nuclease-rich samples under the storage condition used and the late sampling in these cases. Although we do not intend to suggest PCR as a replacement for the excellent method of RSV antigen detection currently used, our results and those of others highlight the specificity of the method, its potential use in epidemiological studies, and the proper sample storage and recruitment procedures necessary.
All 4 patients having PCR-positive results for CMV were also CMV positive by culture, with 2 also positive by convalescent serology. Of these 4 patients, 1 had had liver transplantation (and developed ARDS) and 1 had had heart transplantation. Myerson et al29 have previously shown that PCR on BAL specimens can be used to advantage in the rapid diagnosis of CMV pneumonia.32 Our results suggest that at least in pediatric patients, tracheal aspirates could be equally useful. It should be noted, however, that CMV may be ubiquitous in critically ill patients, such as the child with ARDS, and for this reason it is not certain whether CMV is responsible for the clinical syndrome or simply a bystander.
Interesting results were also obtained from patients presenting with myocarditis and presumed pneumonia. Of the 7 PCR-positive patients (from tracheal aspirates), 4 were also positive by aspirate culture (enterovirus) and 1 was positive by EBV serology. In all cases, PCR performed with EMB specimens demonstrated identical results as those obtained by tracheal aspirate PCR. In a child diagnosed by tracheal aspirate PCR to have EBV (patient 25; Table 2⇑⇑), EMB PCR also identified this relatively uncommon cause of pneumonitis and myocarditis. Confirmation of this diagnosis was later provided by convalescent serology. Another patient who presented clinically with myocarditis/pneumonitis was positive by PCR for adenovirus from 2 consecutive tracheal aspirate samples and had adenovirus PCR-positive results from EMB. We have previously shown that adenovirus may be a more common cause of myocarditis than traditionally implicated, and serology or cultures may not be helpful in the etiologic diagnosis.25 26 This could conceivably be 1 such case of adenoviral myocarditis in which the diagnosis could have been elusive without PCR. In another case of myocarditis with pneumonia, the PCR, in addition to amplifying the same agent as isolated by culture (enterovirus), amplified adenovirus genome. Adenovirus respiratory tract infections are common in children and in this case may have contributed to myocardial injury.
A positive PCR could be expected for both CMV and HSV from our patient with Down’s syndrome and pneumonia, who had a tracheal aspirate culture growing CMV and HSV isolated from nasal wash and a vesicular finger lesion. Although it is not certain whether the development of skin lesions is coincidental or the result of invasive HSV disease resulting in pneumonia in this case, that a correlation exists with various methods is intriguing. It should be noted that in another patient in this series (patient 24 with leukemia and pneumonia; Table 2⇑⇑) in whom HSV PCR of tracheal aspirate was positive but BAL culture and vesicular fluid culture from skin lesion were negative, diagnostic association is merely suggestive.
Our results suggest that tracheal aspirates are a useful substrate for PCR analysis in intubated pediatric patients with suspected viral pneumonitis with or without myocarditis. Tracheal aspirate PCR may provide a safer means of arriving at an etiologic diagnosis in viral myocarditis than EMB, especially when the right ventricular free wall and outflow tracts are pathologically thinned. It is possible that tracheal aspirate PCR may become the method of choice in the etiologic diagnosis of viral myocarditis and in cases of viral pneumonia. However, these results should not be generalized to include, for example, any unselected patient with intubated respiratory disease or children with known cardiac dysfunction and recurrent cardiac decompensation. Confirmation of these findings is needed before changes in diagnostic methodology are embraced.
Reprint requests to Jeffrey A. Towbin, MD, Pediatric Cardiology, Baylor College of Medicine, One Baylor Plaza, Room 333E, Houston, TX 77030.
Guest Editor for this article was Welton M. Gersony, MD, Babies Hospital, New York City, NY.
- Received March 18, 1998.
- Revision received January 25, 1999.
- Accepted January 25, 1999.
- Copyright © 1999 by American Heart Association
Feigin RD, Matson DO. The compromised host. In: Feigin RD, Cherry JD, eds. Textbook of Pediatric Infectious Diseases. 3rd ed. Philadelphia, Pa: WB Saunders Co; 1992:960–989.
Woodruff JF. Viral myocarditis. Am J Pathol. 1980;101:427–478.
Gardiner AJS, Short D. Four faces of acute myocarditis. Br Heart J. 1973;35:433–442.
Henson D, Mufson MA. Myocarditis and pneumonitis with type 21 adenovirus infection. Am J Dis Child. 1981;121:334–336.
Baselski VS, Wunderink RG. Bronchoscopic diagnosis of pneumonia. Clin Microbiol Rev. 1994;7:533–558.
Winn WC Jr, Walker DH. Viral infections. In: Dail DH, Hammer SP, eds. Pulmonary Pathology. 2nd ed. New York, NY: Springer-Verlag; 1994:429–464.
Clewley JP, ed. The Polymerase Chain Reaction (PCR) for Human Viral Diagnosis. Boca Raton, Fla: CRC Press; 1995.
Martin AB, Webber S, Fricker J, Jaffe R, Demmler G, Kearney D, Zhang Y-H, Bodurtha J, Gelb B, Ni J, Bricker JT, Towbin JA. Acute myocarditis: rapid diagnosis by PCR in children. Circulation. 1994;90:330–339.
Allard A, Girones R, Juto P, Wadell G. Polymerase chain reaction for detection of adenovirus in stool samples. J Clin Microbiol. 1990;322:2659–2667.
Chapman NM, Tracy S, Gauntt CJ, Fortmueller U. Molecular detection and identification of enteroviruses using enzymatic amplification and nucleic acid hybridization. J Clin Microbiol. 1990;28:843–850.
Okamoto Y, Shirotori K, Togawa K, Nakazawa M, Patel JA, Kudo K, Ito E, Ogra PL. Detection of genomic sequences of respiratory syncytial virus in otitis media with effusion in children. Ann Otol Rhinol Laryngol. 1992;101:7–10.
Griffin LD, Kearney D, Ni J, Jaffe R, Fricker FJ, Webber S, Demmler G, Gelb BD, Towbin JA. Analysis of formalin-fixed and frozen myocardial autopsy samples for viral genome in childhood myocarditis and dilated cardiomyopathy with endocardial fibroelastosis using polymerase chain reaction (PCR). Cardiovasc Pathol. 1995;4:3–11.
Paton AW, Paton JC, Lawrence AJ, Goldwater PN, Harris RJ. Rapid detection of respiratory syncytial virus in nasopharyngeal aspirates by reverse transcription and polymerase chain reaction amplification. J Clin Microbiol. 1992;30:901–904.
Clarke LM, Sierra FM, Shahabuddin M, Cummings RV, Pellegrino MG, Steiner P, Volsky DJ, Nowakowski M. Markers of human immunodeficiency virus infection in pediatric bronchoalveolar lavage samples. Ann Clin Lab Sci. 1992;22:290–299.