Performance of a molecular viability assay for the diagnosis of Pneumocystis pneumonia in HIV-infected patients

Diagnostic Microbiology and Infectious Disease 57 (2007) 169 – 176 www.elsevier.com/locate/diagmicrobio

Virology

Performance of a molecular viability assay for the diagnosis of Pneumocystis pneumonia in HIV-infected patients Ana de Oliveiraa, Thomas R. Unnascha,4, Kristina Crothersb,c, Shary Eiserc, Patrizia Zucchic, Jonathan Moirc, Charles B. Beardd, Gena G. Lawrencee, Laurence Huangb,c a Division of Geographic Medicine, BBRB Box 7, University of Alabama at Birmingham, Birmingham, AL 35294, USA Division of Pulmonary and Critical Care Medicine, San Francisco General Hospital, University of California San Francisco, San Francisco, CA 94110, USA c HIV/AIDS Division, San Francisco General Hospital, University of California San Francisco, San Francisco, CA 94110, USA d Bacterial Zoonoses Branch, Division of Vector-borne Infectious Diseases Centers for Disease Control and Prevention, Fort Collins, CO 80521, USA e Division of Parasitic Diseases, National Center of Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, GA 30341, USA Received 15 May 2006; accepted 17 August 2006

b

Abstract Pneumocystis pneumonia (PCP), caused by infection with Pneumocystis jirovecii, remains an important opportunistic infection in humans. A reverse transcriptase polymerase chain reaction assay has been shown to specifically detect viable P. jirovecii organisms. In the current study, we evaluated this assay on different types of respiratory samples. The assay had a diagnostic sensitivity of 100% and a specificity of 86% when applied to bronchoalveolar lavage samples. The assay’s performance declined when applied to less invasive induced sputum and oropharyngeal wash (OPW) samples. The sensitivity, when applied to OPWs, was improved by examining multiple sequential OPW samples and was affected by clinical sampling parameters that could increase or decrease the number of potential organisms in the oropharynx. When used in conjunction with an optimized clinical sampling protocol, this assay may become a useful tool for detecting and monitoring P. jirovecii in minimally invasive clinical samples. D 2007 Elsevier Inc. All rights reserved. Keywords: PCP; Opportunistic infections; HIV/AIDS; RT-PCR

1. Introduction Despite significant advances in the treatment of HIV infection, such as combination antiretroviral therapy, pneumocystis pneumonia (PCP) remains the most common AIDS-defining opportunistic infection in the United States and a frequent HIV-associated pneumonia worldwide (Fisk et al., 2003; Kaplan et al., 2000). PCP is caused by infection with Pneumocystis jirovecii (Miller and Huang, 2004; A report of preliminary data leading to this manuscript was presented at the 8th International Workshop on Opportunistic Protists held at Hilo, HI, from July 23 to 29, 2003, and was included in a report of the proceedings of this meeting [Huang L, Crothers K, DeOliveira A, Eiser S, Zucchi P, Beard CB, and Unnasch TR. Application of an mRNA-based molecular viability assay to oropharyngeal washes for the diagnosis of Pneumocystis pneumonia in HIV-infected patients. A pilot study. J Eukaryot Microbiol 50 (2003) (Suppl) 618–620]. 4 Corresponding author. Tel.: +1-205-975-7601; fax: +1-205934-5600. E-mail address: [email protected] (T.R. Unnasch). 0732-8893/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2006.08.015

Stringer et al., 2002). The diagnosis of PCP currently relies on microscopic visualization of the characteristic cysts and/ or trophic forms on stained respiratory specimens obtained from induced sputum (IS) or bronchoscopy with bronchoalveolar lavage (BAL). The requirement for invasive procedures to obtain respiratory specimens for microscopic examination together with the inability to culture P. jirovecii renders PCP an attractive target for molecular diagnosis. A 2nd potential application of a molecular assay for PCP would be to assess treatment response and to address the question of drug resistance. Numerous studies have demonstrated that PCP prophylaxis using sulfamethoxazole (in combination with trimethoprim) or dapsone, both inhibitors of the pneumocystis dihydropteroate synthase (DHPS), is associated with the presence of PCP that contains DHPS mutations (Helweg-Larsen et al., 1999; Huang et al., 2000, 2004; Kazanjian et al., 1998, 2000; Ma et al., 2002; Nahimana et al., 2003). Although different studies suggest that these DHPS mutations are associated with increased

170

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

mortality, increased trimethoprim-sulfamethoxazole PCP treatment failure, and perhaps also increased requirement for mechanical ventilation, these studies cannot directly attribute these outcomes to the presence of DHPS mutations and the existence of drug resistance (Crothers et al., 2005; Helweg-Larsen et al., 1999; Kazanjian et al., 2000). DNA-based polymerase chain reaction (PCR) assays have been shown to be more sensitive than conventional microscopic examination for the diagnosis of PCP. Given this increased sensitivity, PCR assays have been applied to oropharyngeal wash (OPW, gargle) specimens for PCP diagnosis. OPW offers the benefit of being a rapid noninvasive procedure that can be performed in any clinical setting. In several studies, the sensitivity of DNA-based PCR assays to diagnose PCP has ranged from 50% to 94%, and the specificity has ranged from 73% to 96% (Atzori et al., 1998; Fischer et al., 2001; Helweg-Larsen et al., 1998; Larsen et al., 2002, 2004; Matos et al., 2001). False-positive PCR results may limit the clinical application of these DNA assays to diagnose PCP and assess response to therapy. Because these PCR assays target genomic DNA, which is a relatively stable molecule, they cannot distinguish viable from nonviable organisms (Josephson, 1993). Therefore, P. jirovecii DNA detected in some individuals may be derived from nonviable organisms and may represent, for example, prior rather than current infection. Similarly, DNA detected during PCP therapy cannot be interpreted to indicate the presence of viable organism and possible PCP treatment failure. A molecular assay that could specifically detect viable P. jirovecii could allow for more sensitive and specific noninvasive diagnosis of PCP and assessment of treatment effectiveness. A reverse transcriptase (RT) PCR assay targeting the mRNA derived from a gene that is a member of the heat shock family of proteins of P. jirovecii (the Phsb1 gene) has been reported, which only detects viable P. jirovecii organisms (Maher et al., 2001). Because mRNA is an unstable molecule that degrades shortly after the death of the organism, the detection of mRNA can be taken as an indirect indicator of the presence of viable P. jirovecii organisms. In an earlier pilot study of 6 HIV-infected patients, this assay, when applied to BAL specimens, had a sensitivity of 67% and a specificity of 100% for PCP (Maher et al., 2001). The overall objective of this study was to evaluate the sensitivity and specificity of this assay when applied to less invasive clinical samples, and to test the hypothesis that variations in the clinical sampling protocol might affect the sensitivity of the assay when applied to such samples.

2. Materials and methods 2.1. Study population Eligible subjects were HIV-infected adult patients who were admitted to the San Francisco General Hospital

(SFGH) with clinically suspected PCP during the period spanning August 2002 to May 2004. As part of their clinical care, patients underwent an initial IS sample collection, which was examined for pneumocystis and other respiratory pathogens (Huang et al., 1995; Ng et al., 1989). Patients underwent a subsequent bronchoscopy with BAL if no alternate diagnosis was established by microscopic examination of IS or other testing (e.g., blood cultures). Study subjects were consecutive patients who were prospectively enrolled and provided written informed consent. The institutional review boards at the University of California at San Francisco, the University of Alabama at Birmingham, and the Centers for Disease Control and Prevention approved the study protocol. Subjects agreed to have their remaining IS and/or BAL specimens sent for mRNA and DNA analysis, to provide daily OPW specimens during their hospitalization for serial mRNA analysis, and to have their medical records abstracted at enrollment and again after 6 weeks to assess response to therapy and treatment outcome. 2.2. Clinical sample collection The clinical samples examined in this study consisted of IS, BAL, and OPW specimens. IS and BAL specimens were collected as a part of clinical care, using well-established protocols (Huang et al., 1995; Ng et al., 1989), and the IS and BAL specimens examined in this study were the portions of the sample remaining after clinical use. Occasionally, there was insufficient or limited sample remaining. OPW specimens were collected for this study, and the 1st OPW (gargling 10 mL of 0.9% NaCl for 60 s) was collected before sputum induction. Subsequently, serial OPW specimens were collected daily according to subject availability and without regard to time of day or meals. OPW specimens were subjected to low-speed centrifugation for 10 min at room temperature, and the resulting supernatant was discarded. The IS, BAL, and OPW pellets were stored at 70 8C until shipping. 2.3. mRNA assay analysis The IS, BAL, and OPW specimens were shipped to the University of Alabama at Birmingham for RT-PCR mRNA assay testing. RNA was extracted from the pelleted material (representing approximately 200 mg of material) using the RNAeasyk mini kit from Qiagen (Valencia, CA). The pelleted material was resuspended in 600 AL of the guanidinium lysis buffer provided in the kit (buffer RLT). Roughly, 200 AL of 0.5-mm glass beads were then added to the sample, and cells were lysed for 3 min at 5000 rpm in a Bead BeaterR (Biospec Products, Bartlesville, OK). The glass beads were allowed to settle, and the homogenate was transferred to a fresh tube. RNA was then extracted from the homogenate following the manufacturer’s instructions, and the RNA was resuspended in 30 AL of nuclease-free water. Clinical samples were tested for the presence of viable P. jirovecii using a heminested version of the previously

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

developed molecular viability assay (Maher et al., 2001). This assay is based upon specific amplification of mRNA derived from the Phsb1 gene of P. jirovecii. The limit of detection for the P. jirovecii assay is not precisely known because of the difficulty in quantitating the number of viable P. jirovecii organisms in clinically derived samples (Maher et al., 2001). However, studies on a similar assay targeting the transcripts derived from the homologue of the Phsb1 gene in rat Pneumocystis carinii exhibited a limit of detection of approximately 100 organisms (Maher et al., 2000). As previously described, the primers for this assay were designed from sequential exon sequences. Thus, the primer sequences are interrupted by introns in the genomic DNA, permitting amplification from cDNA templates (from which the intron sequences have been removed) but preventing amplification from genomic DNA templates (where the intron sequences are present). A total of 3 AL of RNA prepared as described above was mixed with 25 pmol of a primer spanning intron 5 of the Phsb1 gene (Phsb1nc566: 5VCAGCAGTGGCTTTAACTGAA 3V) in a total volume of 5.75 AL. The sample was heated at 70 8C for 15 min and cooled at 0 8C for 5 min. The solution was brought to a final volume of 10 AL in reverse transcription reaction buffer. The final composition of the RT reaction consisted of 50 mmol/L Tris–HCl (pH 8.3); 75 mmol/L KCl; 3 mmol/L MgCl2; 10 mmol/L dithiothreitol; 1 mmol/L each of dATP, dCTP, dGTP, and dTTP; 100 U of RT (Superscript IIk; Life Technologies, Bethesda, MD); and 10 U of pancreatic RNAse inhibitor (RNASink; Promega, Madison, WI). The reaction was incubated at 45 8C for 50 min and then at 70 8C for 10 min. The reaction was then brought to a final volume of 50 AL for the 1st step amplification. The primers that were used in the amplification were Phsb1nc566 and Phsb1c161 (5VTGTTAAAAAAGACATGAAAATG3V). Phsb1c161 was designed to span intron 3 of the Phsb1 gene. The final composition of the 1st PCR reaction was 60 mmol/L Tris– HCl (pH 8.5); 15 mmol/L (NH4)2SO4; 3.5 mmol/L MgCl2; 200 Amol/L each of dATP, dCTP, dGTP, and dTTP; 0.5 Amol/L of each primer; and 2.5 U of AmpliTaq DNA polymerase (Applied Biosystems, Branchburg, NJ). PCR conditions consisted of 40 cycles, each consisting of 1 min at 94 8C, 1 min at 50 8C, and 1 min at 72 8C. A total of 1 AL of the 1st PCR amplification product was then used as a template in the heminested PCR amplification reaction. The primers that were used in the heminested amplification were Phsb1c161 and Phsb1nc284 (5VTTTCTTTCATCTTTCCTAAT 3V). Phsb1nc284 was designed to span intron 4 of the Phsb1 gene. The composition of the reaction and the cycling conditions were identical to those that were used in the 1st step amplification. Products were analyzed by electrophoresis on a 1.5% agarose gel, and samples producing a band of the predicted size (162 bp) were judged to be positive. The identity of amplicons derived from the initial positive sample collected from each patient was confirmed by DNA sequence analysis. All of the sequenced amplicons were identical to the P. jirovecii Phsb1 mRNA sequence.

171

2.4. Mitochondrial large subunit ribosomal RNA DNA PCR analysis Specimens were also subjected to DNA PCR analysis employing an assay targeting a portion of the P. jirovecii mitochondrial large subunit rRNA (mt LSU) gene. DNA extraction was performed using the Wizard Genomic DNA Purification Kit (Promega). The procedures for specimen processing, PCR amplification, and DNA sequencing have been described previously (Beard et al., 2000; Wakefield et al., 1990). The identity of all amplicons was confirmed by DNA sequence analysis. 2.5. Patient classification and clinical data Subjects were classified as PCP positive or PCP negative by clinical investigators who were blinded to the results of the experimental assays. Patients were considered PCP positive if they had Pneumocystis organisms microscopically visualized in their IS or BAL samples. Patients were considered PCP negative if they had no Pneumocystis seen microscopically, recovered from the acute pneumonic process without specific PCP treatment, were discharged from the hospital, and remained well without subsequent PCP treatment for 6 weeks after initial enrollment into the study. Several clinical sampling parameters were recorded to evaluate their effect on the RT-PCR assay’s performance. These included the time (in days) from the start of PCP treatment to the OPW sample collection, a qualitative assessment of the effectiveness of the gargle, and when the patient had last eaten, drank, or brushed their teeth. To evaluate the effect of coughing on assay performance, we asked a subset of confirmed PCP-positive patients (n = 9) to provide 2 sequential gargle samples. The 1st sample was collected as described above, whereas the 2nd sample was collected after the subject coughed 5 times before gargling. 2.6. Statistical analysis The sensitivity, specificity, and positive and negative predictive values were calculated for the RT-PCR assay using the PCP-positive and PCP-negative classifications described above. Data on time since last eating, drinking, or brushing teeth, and days on PCP treatment at time of OPW specimen collection were divided into dichotomous categories to examine factors that affect the performance of the RT-PCR assay (Table 2). Univariate analysis was then performed to assess the impact of various clinical sampling parameters on RT-PCR positivity. Calculations were done using Statcalc (EpiInfo version 6). Categoric variables were compared using a v 2 test, and a 2-sided P value of b .05 was considered to be statistically significant. 3. Results Fifty-eight HIV-infected patients, representing 61 separate hospital admissions for clinically suspected PCP,

172

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

it is possible that the amount and/or the quality of the IS specimen were inadequate for both microscopy and RT-PCR. The performance of the RT-PCR mRNA assay when applied to noninvasive OPW samples varied depending upon the number of samples examined. When based on a single initial OPW, 13 of 37 OPW samples from PCPpositive patients were RT-PCR positive (sensitivity of 35%), whereas 19 of 21 OPW samples from PCP-negative patients were RT-PCR negative (specificity of 90%) (Table 1). The sensitivity improved when the results from multiple OPW samples collected from the same individual on different days were combined. For example, classifying the RT-PCR assay as positive if 1 or more of the 1st 3 sequential OPW samples collected was positive improved the sensitivity of the assay to 61%, similar to that found with the IS samples, while maintaining a specificity of 92% (Table 1). When the number of sequential OPW samples included in the analysis was increased to between 4 and 7, the sensitivity of the assay increased to 85%. However, the specificity dropped to 50%, because a single OPW collected from 1 of 2 PCP-negative patients from whom it was possible to collect this large a number of sequential OPW samples was also RT-PCR positive (Table 1). Samples were also analyzed using a standard DNA PCR assay targeting the mt LSU gene of P. jirovecii. Results from the mt LSU DNA PCR assay when applied to OPWs were similar to that seen in other studies employing DNA PCR assays to detect evidence for P. jirovecii in OPW samples (Atzori et al., 1998; Fischer et al., 2001; HelwegLarsen et al., 1998; Larsen et al., 2002, 2004; Matos et al., 2001). When only data from the initial OPW were considered, the sensitivity of the DHPS DNA PCR was 85%. However, the specificity of the assay was relatively poor, at 46%. When multiple OPWs were analyzed, the sensitivity improved to 97%, but the specificity of the assay declined further, to just 25% (Table 1). When the data from

were enrolled. Most of the 58 patients were male (87.9%) and white non-Hispanic (41%), and the mean age was 42.3 years (range, 27–61 years). Overall, 39 patient episodes were classified as PCP positive and 22 patient episodes were classified as PCP negative. These subjects provided 361 samples (17 BAL, 33 IS, and 311 OPW), which were examined using the RT-PCR mRNA assay and DNA PCR. The RT-PCR mRNA assay performed well when applied to BAL samples. Ten PCP-positive patients had Pneumocystis organisms microscopically visualized on their BAL specimen (microscopy positive), and the RT-PCR assay was positive on all 10 BAL samples. Seven PCPnegative patients had no evidence of Pneumocystis on microscopy (microscopy negative), and the RT-PCR assay was negative on 6 of 7 BAL samples. Thus, the sensitivity and specificity of the assay on BAL samples was 100% and 86%, and positive and negative predictive values were 91% and 100%, respectively (Table 1). The sensitivity of the RT-PCR mRNA assay decreased when tested on less invasive IS samples. The RT-PCR assay was positive in 15 of 23 IS samples from PCPpositive patients and negative in 8 of 10 IS samples from PCP-negative patients, leading to a sensitivity of 65%, a specificity of 80%, and positive and negative predictive values of 88% and 50%, respectively (Table 1). However, the RT-PCR assay results appeared to parallel the microscopic examination results. Of the 23 PCP-positive patients who underwent sputum induction, 15 had a positive IS microscopic examination and 8 had a negative IS microscopic examination but had the diagnosis of PCP subsequently established by a positive BAL microscopic examination. The RT-PCR assay was positive in 11 of the 15 (73%) microscopy-positive IS samples and 4 of the 8 (50%) microscopy-negative IS samples. The precise reason why microscopy failed to demonstrate Pneumocystis organisms in these IS specimens but then demonstrated organisms in subsequent BAL specimens is unclear, but

Table 1 Performance of the P. jirovecii RT-PCR mRNA assay on different types of clinical samples Gold standard

Test assay

Sample type

No. tested

Sensitivity

Specificity

PPV

NPV

Microscopy Microscopy Microscopya Microscopya Microscopya Microscopya Microscopya Microscopya Microscopya DNA PCRb DNA PCRb

RT-PCR RT-PCR RT-PCR RT-PCR RT-PCR mt LSU PCR mt LSU PCR mt LSU PCR mt LSU PCR RT-PCR RT-PCR

BAL IS Initial OPW N 1 and b 3 OPWs N 1 and b 7 OPWs BAL IS Initial OPW N 1 and b 3 OPWs Initial OPW N 1 and b 3 OPWs

17 33 58 45 28 26 42 52 34 57 31

1.00 0.65 0.35 0.61 0.85 0.79 0.88 0.85 0.97 0.21 0.41

0.86 0.80 0.90 0.92 0.5 0.75 0.44 0.46 0.25 0.93 0.50

0.91 0.88 0.87 0.95 0.96 0.79 0.68 0.83 0.91 0.90 0.92

1.00 0.50 0.44 0.46 0.2 0.75 0.73 0.50 0.50 0.28 0.06

Abbreviations: PPV = positive predictive value; NPV = negative predictive value. a Positive if either BAL or IS microscopy was positive. b Positive if any specimen was mt LSU PCR positive.

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

the RT-PCR were analyzed using the data collected from the mt LSU DNA PCR as a gold standard, the RT-PCR exhibited a low initial sensitivity of 21%, which increased to just 41% when the assay was applied to multiple OPWs (Table 1). The presented data suggest that the sensitivity of the RT-PCR mRNA assay is increased by collecting and analyzing multiple serial OPW samples. To determine whether certain clinical parameters might affect the number of organisms present in the OPW sample (and therefore the performance of the assay), we evaluated the impact of several different variables on the sensitivity of the RT-PCR assay when applied to a single OPW sample (Table 2). OPW samples collected from PCP-positive individuals who last ate more than 3 h before the collection of the OPW were 2.3 times more likely to be RT-PCR positive than OPW samples collected from individuals who had eaten within the 3 h before providing the OPW ( P = .026). A similar statistically significant association was noted when the time between drinking and OPW collection was examined. OPW samples collected from PCP-positive individuals who last drank more than 3 h before the collection of the OPW were almost 3 times more likely to be RT-PCR positive than OPW samples collected from individuals who drank within the 3 h before providing the OPW ( P = .005). In contrast, the time between tooth brushing and OPW collection was not found to be significantly associated with OPW positivity. Table 2 Effect of clinical sampling parameters on the performance of the P. jirovecii RT-PCR mRNA assay Clinical parameter

No. tested

Food V3 h 129 N3 h 64 Drink V3 h 147 N3 h 44 Brushing teeth V3 h 43 N3 h 46 Cough None 20 5 before 20 sampling Gargle quality Poor 34 Good 6 Time posttreatment V 1 day 29 2–5 days 42 6–20 days 56

Odds ratio

95% CIa

P

1.00 2.27

1.03–5.01

.026b

1.00 2.94

1.27–6.80

.005b

1.00 1.36 1.00 4.64

.620b

0.85–28.11

1.00 1.07 1.00 0.58 0.23

.041d

.999c

0.19–1.81 0.07–0.78

.296b .007b P trend = .007

a Cornfield 95% confidence interval (CI). CIs are provided for only those parameters found to have a significant effect ( P b .05). b Mantel-Haenszel v 2. c Fisher’s exact v 2. d Fisher’s exact v 2 (1-sided P value).

173

We also tested the effect of coughing before OPW collection on the sensitivity of the RT-PCR assay, using paired pre- and post-cough OPW samples from PCPpositive individuals as described in Materials and methods. OPW samples collected after the patient coughed vigorously were 4.6 times more likely to be RT-PCR positive than those collected before coughing ( P = .041, Table 2). In contrast, the subject’s ability to gargle was not found to be significantly associated with OPW positivity. The time between the start of PCP treatment and OPW collection was also a significant factor in determining if a given OPW would be RT-PCR positive. Thus, OPW samples collected later than 24 h after the start of treatment were significantly less likely to be RT-PCR positive than those collected within 24 h of treatment initiation, and the probability of obtaining a positive result in the RT-PCR declined as the time from treatment initiation to sample collection increased (Table 2). 4. Discussion BAL is the most sensitive procedure to obtain specimens for the diagnosis of PCP, and BAL samples generally contain greater numbers of organisms than do samples collected by other methods (Bartlett and Smith, 1991). Thus, it could be predicted that the RT-PCR assay would have the highest sensitivity (100%) when applied to BAL samples. However, bronchoscopy with BAL is an invasive procedure and requires subspecialty trained personnel, a specifically engineered room, and expensive equipment, and it is associated with significant potential complications (Huang et al., 1995). Furthermore, bronchoscopy is generally limited to hospital settings and is unavailable or limited in availability in resource-limited countries where most of the HIV-infected persons reside. Therefore, less invasive procedures, paired with a molecular assay, would be important diagnostic tests if proven to be both sensitive and specific. At select institutions, IS and less commonly OPW are 2 other procedures used to collect respiratory specimens for the diagnosis of PCP. In this study, we found that the RTPCR assays had a lower sensitivity on these less invasive specimens than on BAL. Our OPW sensitivity in the RT-PCR was comparable to a recent study from South Africa that examined OPW specimens using a DNA-based PCR assay and reported an OPW sensitivity of 40% (Nyamande et al., 2005). The RT-PCR assay results appeared to parallel the microscopic examination results, and the sensitivities of the RT-PCR assay and microscopic examination of IS specimens were identical (65%). This finding is comparable to prior studies at SFGH, which reported that the sensitivity of microscopy is 89% to 98% on BAL specimens but just 56% to 81% using IS specimens (Bigby et al., 1986; Broaddus et al., 1985; Huang et al., 1995; Ng et al., 1989). These results suggest that the relative insensitivity of the RT-PCR assay when

174

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

applied to IS samples may be more of a factor of the sample collection process than of the molecular technique. One study that examined multiple IS specimens found that the IS sensitivity increased when multiple specimens from a single individual were analyzed and the use of multiple sputum inductions was advocated as an approach in settings where bronchoscopy was unavailable or limited in availability (Narasimha et al., 2004). The sensitivity of the RT-PCR was improved when multiple OPWs were analyzed, reaching a maximum of 85% when up to 7 OPWs were analyzed. This level of performance was similar to that obtained when the RT-PCR was applied to IS. Together, these data suggest that analysis of 2 to 3 sequential OPWs by RT-PCR might be as useful in diagnosing PCP as IS and should be considered in settings where sputum induction and bronchoscopy are unavailable. However, the RT-PCR assay is relatively expensive and technically complex. RNA preparation and RT-PCR require some technical expertise and specialized equipment, and the assay costs several dollars per sample in reagent expenses alone. These factors may present a hurdle to the adoption of this method in the developing world. Despite this, the assay may represent a potentially useful diagnostic procedure in settings that do not have the equipment or trained personnel necessary to use bronchoscopy but where the cost is not considered prohibitive. The increase in sensitivity associated with the analysis of multiple OPW samples by RT-PCR suggests the hypothesis that microscopy of serial OPWs might be a cheap and noninvasive method to diagnose P. jirovecii infection in resource poor settings. Microscopy of individual OPWs has not proven to be a useful diagnostic technique, because organisms in OPW specimens are rare and hard to visualize among nonspecific debris derived from the oral cavity (unpublished data). However, it is possible that microscopic examination of sequential OPWs might prove to be an effective diagnostic procedure, especially if this process was coupled with simple methods to purify and concentrate the organisms (e.g., by disruption of the mucus in the OPW sample by treatment with dithiothreitol followed by low-speed centrifugation to concentrate the organisms). The OPW sensitivity of both the RT-PCR and mt LSU DNA PCR assays increased when multiple OPWs from a single individual were analyzed. However, the specificity of the mt LSU PCR was quite poor when applied to sequential OPWs, declining to just 25%. The poor level of specificity was also reflected in the poor sensitivity of the RT-PCR when its performance was assessed using the mt LSU PCR as the gold standard. The low specificity of the DNA PCR when applied to noninvasive samples was in keeping with recent studies documenting the presence of P. jirovecii DNA in immunocompetent individuals with no clinical signs of PCP (Durand-Joly et al., 2003; Leigh et al., 1993; Medrano et al., 2005; Nevez et al., 1997; Sing et al., 1999, 2001). These data have been taken to suggest that

colonization with (or at least exposure to) P. jirovecii in immunocompetent individuals may occur at a relatively high prevalence. However, data derived from the DNA PCRs cannot distinguish between the presence of material derived from dead organisms and that derived from viable organisms. Thus, the low specificity of the DNA PCR assay that was seen in this and other studies may reflect the presence of dead organisms. In this regard, the increased specificity of the RT-PCR may reflect the fact that it specifically detects viable organisms, thereby possibly representing a clinically more useful method of diagnosing the presence of viable P. jirovecii organisms in noninvasive samples. However, the presence of viable organisms does not necessarily mean that an individual is actually infected with P. jirovecii, that is, a positive result in the RT-PCR does not indicate that the detected viable organisms are actively multiplying. Development of a quantitative version of the RT-PCR, together with serial sampling, might be useful in answering this question. The fact that analysis of multiple OPW samples increased the sensitivity of the RT-PCR assay suggested that variations in the clinical sampling procedure might affect the performance of the assay. We identified several parameters that appeared to affect the sensitivity of the assay when it was applied to OPW specimens. It appears that activities that result in a clearance of potential organisms from the oral cavity by swallowing (e.g. food and liquid intake) reduce the probability of obtaining a positive OPW sample. Similarly, vigorous coughing before sampling, which could serve to bring organisms from the lower respiratory tract up to the oropharynx, increased the probability of obtaining a positive OPW sample. Another significant factor in determining the probability that a given OPW would be positive in the RT-PCR assay was the amount of time between the start of PCP treatment and the collection of the OPW sample. Thus, the greater the period between the start of treatment and the collection of the OPW, the less likely it was that the OPW would prove to be positive in the RT-PCR. This is what would be expected if PCP treatment was resulting in a decrease in the number of viable organisms. This finding is consistent with previous studies, which have documented a decrease in amplifiable P. jirovecii DNA in oral samples after treatment (Vargas et al., 1995). These data suggest that the RT-PCR assay, when combined with an optimized clinical sampling protocol, may prove to be capable of monitoring the effectiveness of treatment of PCP. If so, the assay may be a useful tool for the early detection of treatment failure and/or drug resistance in individuals with patent PCP.

Acknowledgments The authors would like to thank Dr. Naomi LangUnnasch for critically reading the manuscript. This work

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

received financial support from grants from the National Institutes of Health (NIH K23 HL 072117) and the University-wide AIDS Research Program (UARP IDEA Award ID04-SF-026), both to LH.

References Atzori C, Agostoni F, Angeli E, Mainini A, Orlando G, Cargnel A (1998) Combined use of blood and oropharyngeal samples for noninvasive diagnosis of Pneumocystis carinii pneumonia using the polymerase chain reaction. Eur J Clin Microbiol Infect Dis 17:241 – 246. Bartlett MS, Smith JW (1991) Pneumocystis carinii, an opportunist in immunocompromised patients. Clin Microbiol Rev 4:137 – 149. Beard CB, Carter JL, Keely SP, Huang L, Pieniazek NJ, Moura INS, Roberts JM, Hightower AW, Bens MS, Freeman AR, Lee S, Stringer JR, Duchin JS, del Rio C, Rimland D, Baughman RP, Levy DA, Dietz VJ, Simon P, Navin TR, 2000. Genetic variation in Pneumocystis carinii isolates from different geographic regions: implications for transmission. Emerg Infect Dis 6:265–272. Bigby TD, Margolskee D, Curtis JL, Michael PF, Sheppard D, Hadley WK, Hopewell PC (1986) The usefulness of induced sputum in the diagnosis of Pneumocystis carinii pneumonia in patients with the acquired immunodeficiency syndrome. Am Rev Respir Dis 133: 515 – 518. Broaddus C, Dake MD, Stulbarg MS, Blumenfeld W, Hadley WK, Golden JA, Hopewell PC (1985) Bronchoalveolar lavage and transbronchial biopsy for the diagnosis of pulmonary infections in the acquired immunodeficiency syndrome. Ann Intern Med 102: 747 – 752. Crothers K, Beard CB, Turner J, Groner G, Fox M, Morris A, Eiser S, Huang L (2005) Severity and outcome of HIV-associated pneumocystis pneumonia containing Pneumocystis jirovecii dihydropteroate synthase gene mutations. AIDS 19:801 – 805. Durand-Joly I, Soula F, Chabe M, Dalle JH, Lafitte JJ, Senechal M, Pinon A, Camus D, Dei-Cas E (2003) Long-term colonization with Pneumocystis jirovecii in hospital staffs: a challenge to prevent nosocomial pneumocystosis. J Eukaryot Microbiol 50:614 – 615. Fischer S, Gill VJ, Kovacs J, Miele P, Keary J, Silcott V, Huang S, Borio L, Stock F, Fahle G, Brown D, Hahn B, Townley E, Lucey D, Masur H (2001) The use of oral washes to diagnose Pneumocystis carinii pneumonia: a blinded prospective study using a polymerase chain reaction-based detection system. J Infect Dis 184:1485 – 1488. Fisk DT, Meshnick S, Kazanjian PH (2003) Pneumocystis carinii pneumonia in patients in the developing world who have acquired immunodeficiency syndrome. Clin Infect Dis 36:70 – 78. Helweg-Larsen J, Jensen JS, Benfield T, Svendsen UG, Lundgren JD, Lundgren B (1998) Diagnostic use of PCR for detection of Pneumocystis carinii in oral wash samples. J Clin Microbiol 36:2068 – 2072. Helweg-Larsen J, Benfield TL, Eugen-Olsen J, Lundgren JD, Lundgren B (1999) Effects of mutations in Pneumocystis carinii dihydropteroate synthase gene on outcome of AIDS-associated P. carinii pneumonia. Lancet 354:1347 – 1351. Huang L, Hecht FM, Stansell JD, Montanti R, Hadley WK, Hopewell PC (1995) Suspected Pneumocystis carinii pneumonia with a negative induced sputum examination. Is early bronchoscopy useful? Am J Respir Crit Care Med 151:1866 – 1871. Huang L, Beard CB, Creasman J, Levy D, Duchin JS, Lee S, Pieniazek N, Carter JL, del Rio C, Rimland D, Navin TR (2000) Sulfa or sulfone prophylaxis and geographic region predict mutations in the Pneumocystis carinii dihydropteroate synthase gene. J Infect Dis 182:1192 – 1198. Huang L, Crothers K, Atzori C, Benfield T, Miller R, Rabodonirina M, Helweg-Larsen J (2004) Dihydropteroate synthase gene mutations in pneumocystis and sulfa resistance. Emerg Infect Dis 10:1721 – 1728.

175

Josephson KL (1993) Polymerase chain reaction detection of non-viable pathogens. Appl Environ Microbiol::3513 – 3515. Kaplan JE, Hanson D, Dworkin MS, Frederick T, Bertolli J, Lindegren ML, Holmberg S, Jones JL (2000) Epidemiology of human immunodeficiency virus-associated opportunistic infections in the United States in the era of highly active antiretroviral therapy. Clin Infect Dis 30:S5 – S14. Kazanjian P, Locke AB, Hossler PA, Lane BR, Bartlett MS, Smith JW, Cannon M, Meshnick SR (1998) Pneumocystis carinii mutations associated with sulfa and sulfone prophylaxis failures in AIDS patients. AIDS 12:873 – 878. Kazanjian P, Armstrong W, Hossler PA, Burman W, Richardson J, Lee CH, Crane L, Katz J, Meshnick SR (2000) Pneumocystis carinii mutations are associated with duration of sulfa or sulfone prophylaxis exposure in AIDS patients. J Infect Dis 182:551 – 557. Larsen HH, Masur H, Kovacs JA, Gill VJ, Silcott VA, Kogulan P, Maenza J, Smith M, Lucey DR, Fischer SH (2002) Development and evaluation of a quantitative, touch-down, real-time PCR assay for diagnosing Pneumocystis carinii pneumonia. J Clin Microbiol 40:490 – 494. Larsen HH, Huang L, Kovacs JA, Crothers K, Silcott VA, Morris A, Turner JR, Beard CB, Masur H, Fischer SH (2004) A prospective, blinded study of quantitative touch-down polymerase chain reaction using oralwash samples for diagnosis of pneumocystis pneumonia in HIVinfected patients. J Infect Dis 189:1679 – 1683. Leigh TR, Kangro HO, Gazzard BG, Jeffries DJ, Collins JV (1993) DNA amplification by the polymerase chain reaction to detect subclinical Pneumocystis carinii colonization in HIV-positive and HIV-negative male homosexuals with and without respiratory symptoms. Res Med 87:525 – 529. Ma L, Kovacs JA, Cargnel A, Valerio A, Fantoni G, Atzori C (2002) Mutations in the dihydropteroate synthase gene of human-derived Pneumocystis carinii isolates from Italy are infrequent but correlate with prior sulfa prophylaxis. J Infect Dis 185:1530 – 1532. Maher N, Vermund S, Lasbury M, Lee C, Bartlett M, Unnasch TR (2000) Development and evaluation of a molecular viability assay for Pneumocystis carinii. J Clin Microbiol 38:1947 – 1952. Maher NH, Vermund SH, Welsh DA, Dillon HK, Awooda A, Unnasch TR (2001) Development and characterization of a molecular viability assay for Pneumocystis carinii f sp hominis. J Infect Dis 183:1825 – 1827. Matos O, Costa MC, Lundgren B, Caldeira L, Aguiar P, Antunes F (2001) Effect of oral washes on the diagnosis of Pneumocystis carinii pneumonia with a low parasite burden and on detection of organisms in subclinical infections. Eur J Clin Microbiol Infect Dis 20:573 – 575. Medrano FJ, Montes-Cano M, Conde M, de la Horra C, Respaldiza N, Gasch A, Perez-Lozano MJ, Varela JM, Calderon EJ (2005) Pneumocystis jirovecii in general population. Emerg Infect Dis 11:245 – 250. Miller R, Huang L (2004) Pneumocystis jirovecii infection. Thorax 59:731 – 733. Nahimana A, Rabodonirina M, Helweg-Larsen J, Meneau I, Francioli P, Bille J, Hauser PM (2003) Sulfa resistance and dihydropteroate synthase mutants in recurrent Pneumocystis carinii pneumonia. Emerg Infect Dis 9:864 – 867. Narasimha PD, Prabha MR, Harsha AB, Thomas RJ, Shenoy S, Chakrapani M (2004) Role of multiple induced sputum examination in the diagnosis of Pneumocystis carinii pneumonia. Southeast Asian J Trop Med Public Health 35:702 – 704. Nevez G, Jounieaux V, Linas MD, Guyot K, Leophonte P, Massip P, Schmit JL, Seguela JP, Camus D, Dei-Cas E, Raccurt C, Mazars E (1997) High frequency of Pneumocystis carinii sp.F. hominis colonization in HIVnegative patients. J Eukaryot Microbiol 44:36S. Ng VL, Gartner I, Weymouth LA, Goodman CD, Hopewell PC, Hadley WK (1989) The use of mucolysed induced sputum for the identification of pulmonary pathogens associated with human immunodeficiency virus infection. Arch Pathol Lab Med 113:488 – 493.

176

A. de Oliveira et al. / Diagnostic Microbiology and Infectious Disease 57 (2007) 169–176

Nyamande K, Lalloo UG, York D, Naidoo M, Irusen EM, Chetty R (2005) Low sensitivity of a nested polymerase chain reaction in oropharyngeal washings for the diagnosis of Pneumocystis pneumonia in HIV-infected patients. Chest 128:167 – 171. Sing A, Roggenkamp A, Autenrieth IB, Heesemann J (1999) Pneumocystis carinii carriage in immunocompetent patients with primary pulmonary disorders as detected by single or nested PCR. J Clin Microbiol 37:3409 – 3410. Sing A, Geiger AM, Hogardt M, Heesemann J (2001) Pneumocystis carinii carriage among cystic fibrosis patients, as detected by nested PCR. J Clin Microbiol 39:2717 – 2718.

Stringer JR, Beard CB, Miller RF, Wakefield AE (2002) A new name (Pneumocystis jirovecii) for pneumocystis from humans. Emerg Infect Dis 8:891 – 896. Vargas SL, Hughes WT, Wakefield AE, Oz HS (1995) Limited persistence and subsequent elimination of Pneumocystis carinii from the lungs after P. carinii pneumonia. J Infect Dis 172:506 – 510. Wakefield AE, Pixley FJ, Banerji S, Sinclair K, Miller RF, Moxon ER, Hopkin JM (1990) Amplification of mitochondrial ribosomal RNA sequences from Pneumocystis carinii DNA of rat and human origin. Mol Biochem Parasitol 43:69 – 76.