Article Text

other Versions

Download PDFPDF
Pulmonary toxoplasmosis in immunocompromised patients with interstitial pneumonia: a single-centre prospective study assessing PCR-based diagnosis
  1. Guillaume Desoubeaux1,2,
  2. Églantine Cabanne1,
  3. Claire Franck-Martel1,
  4. Martin Gombert1,
  5. Emmanuel Gyan3,4,
  6. Séverine Lissandre3,
  7. Marc Renaud3,
  8. Hélène Monjanel3,
  9. Caroline Dartigeas3,
  10. Éric Bailly1,
  11. Nathalie Van Langendonck1,
  12. Jacques Chandenier1,2
  1. 1Service de Parasitologie—Mycologie—Médecine tropicale, CHU de Tours, Tours, France
  2. 2Faculté de Médecine, Université François-Rabelais, CEPR—INSERM U1100/Équipe 3, Université François-Rabelais Tours, France
  3. 3Service d'Hématologie et Thérapie Cellulaire, CHU de Tours, Tours, France
  4. 4Faculté de Médecine, Université François-Rabelais, N2C—INSERM U1069, Université François-Rabelais Tours, France
  1. Correspondence to Dr Guillaume Desoubeaux, Service de Parasitologie—Mycologie—Médecine tropicale, CHU de Tours, Hôpital Bretonneau, Pôle de Biologie Médicale, Bâtiment B2A—1er étage, 2 boulevard Tonnellé, 37044 CHU de TOURS Cedex 9, France; guillaume.desoubeaux{at}univ-tours.fr

Abstract

Aims Pulmonary toxoplasmosis has become a very rare parasitic infection since the advent of highly active antiretroviral therapies. It is generally diagnosed by the direct microscopic observation of Toxoplasma gondii tachyzoites in bronchoalveolar lavage fluid (BALF). The aim of this study was to assess possible improvements in diagnostic performance associated with the use of real-time PCR.

Methods This prospective study was carried out on BALFs obtained from immunocompromised patients over a 2-year period. We systematically compared the results of conventional staining with those of molecular detection.

Results Two cases of pulmonary toxoplasmosis were diagnosed for a total of 336 samples. PCR did not detect any additional cases and was more time-consuming than conventional staining.

Conclusions Conventional staining is a reliable technique and is probably the most appropriate method for experienced microbiology laboratories, whereas T. gondii-specific PCR may be useful for laboratories with less experience in parasitology.

Trial registration number 2015_030, May 27th 2015.

  • PCR
  • TOXOPLASMA
  • DIAGNOSIS
  • SEROLOGY
  • LUNG

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Toxoplasmosis is a zoonotic infection caused by the protozoan Toxoplasma gondii.1 Despite its high prevalence worldwide, estimated at about 25%–30%,2 toxoplasmosis is rarely symptomatic in human beings. Severe manifestations of toxoplasmosis, such as encephalitis,3 myocarditis and chorioretinitis,1 ,4 are usually seen in highly immunocompromised patients, particularly those with CD4+ T-cell counts below 0.1×109/L.5 In rare cases, the lungs may be the only organ affected. In such cases, T. gondii infection may then mimic interstitial pneumonia, closely resembling that caused by more common opportunistic agents, such as Pneumocystis jirovecii and cytomegalovirus,6 in the same context. In the absence of reliable diagnostic tools, it is easy to confuse these aetiologies, particularly as serological testing for T. gondii gives poor results in immunocompromised patients. For instance, antibody production may be totally abolished in patients undergoing haematopoietic stem cell transplantation (HSCT) and serological status is frequently uninterpretable in solid organ transplant (SOT) recipients unless there has been regular, thorough monitoring, both before and after transplantation.7 ,8 Molecular detection methods are becoming increasingly common. PCR detection techniques are commonly used to detect P. jirovecii DNA in bronchoalveolar lavage fluids (BALFs), and molecular detection is also routinely used for T. gondii detection,9 but mostly only at a small number of specialist centres.1 ,10 Furthermore, T. gondii-specific PCR is frequently reserved for prenatal diagnosis or analyses of cerebrospinal fluid (CSF), blood or aqueous humour.11 Direct observation of the parasite in respiratory fluids after conventional staining is assumed to be time-consuming and requires a certain level of expertise on the part of the microbiologist, but it remains the most widely used method for detection of the parasite in the lungs.1 Thus, the prevalence of pulmonary toxoplasmosis is probably underestimated, particular when symptoms are limited to the lungs and the parasite burden is low. BALFs are not routinely screened by PCR for T. gondii in our teaching hospital, but two cases of pulmonary toxoplasmosis were recently diagnosed by microscopy in patients initially thought to have P. jirovecii pneumonia. Following the diagnosis of these two cases, we decided to carry out real-time PCR systematically on BALFs from patients considered to be at high risk of opportunistic infections, to determine whether this molecular tool for the diagnosis of pulmonary toxoplasmosis presented any advantage over conventional staining techniques.

Methods

Selection of the study population

A single-centre prospective study was carried out from January 2013 to December 2014. Individuals with pulmonary symptoms and radiological signs of interstitial pneumonia, associated with a lymphocyte count below 1.0×109/L and/or a CD4+ T-cell count below 0.2×109/L (when determined) were included, as soon as they had undergone bronchoalveolar lavage. The exclusion criteria were: patient under the age of 18 years, or too little BALF collected for additional investigations.

Direct microscopic examination

We initially used microscopy (magnification ×1000) to search for T. gondii tachyzoites. We deposited 100 µL of BALF in a Cytofunnel double device (Thermo Fisher Scientific, Asnières, France) for Cytospin3 (Thermo Scientific Shandon, Illkirch, France) centrifugation (80×g, 5 min). The sample was then stained with May-Grunwald Giemsa stain (MGG, Microscopy Hemacolor, Millipore, Billerica, Massachusetts, USA). The two stained spots on each glass slide were read by a senior microbiologist and a laboratory technician.

Tests for parasite DNA

We then carried out real-time quantitative PCR (qPCR) on all samples, to detect the SSU of the T. gondii 18S rDNA. This analysis was carried out on all BALFs, regardless of the microscopy results obtained. Briefly, samples were centrifuged (1730×g, 10 min) and DNA was extracted from the supernatant with the QiAamp DNA Mini kit (Qiagen, Courtaboeuf, France), according to the manufacturer's recommendations. We then carried out real-time qPCRs on a final volume of 25 µL, with the Platinum Quantitative PCR SuperMix-UDG from InVitrogen (Life Technologies SAS, Saint-Aubin, France) and 10 µL of each DNA extract. The following oligonucleotides were used: TGIII/5′-GGCATTCCTCGTTGAAGATT-3′ and TGIIB/5′-CCTTGGCCGATAGGTCTAGG-3′ for the primers and 5′ FAM-TGCAATAATCTATCCCCATCACGATGCATACTCAC-TAMRA 3′ for the specific TaqMan probe.12 The final concentrations of primers and probe were 0.5 and 0.18 µM, respectively. MgCl2 concentration was adjusted to 0.5 mM. Amplification was conducted in duplicate on a LightCycler 480 machine (Roche Applied Science, La Rochelle, France) under the following conditions: initial decontamination for 2 min at 50°C, followed by denaturation for 5 min at 95°C, then 50 cycles of denaturation for 20 s at 95°C, annealing and elongation for 1 min at 65°C and a final extension phase at 40°C for 30 s. Inhibition was assessed with an exogenous positive internal control (Universal Inhibition Control Cy5, Diagenode, Seraing-sur-Meuse, Belgium) and an endogenous in-house positive internal control (T. gondii DNA extract). The cycle of quantification (Cq) is reported for all positive results.

Statistical analysis

Statistical analysis was performed with XLStat V.2014.6.04 software (Addinsoft, Paris, France). Epidemiological data were provided by the Service d'Information Médicale, Épidémiologie et Économie de la Santé of Tours University Hospital.

Results

Study population

We prospectively collected 336 BALFs from 333 patients over the study period. Only 97 of these samples strictly fulfilled the inclusion criteria (table 1). There were many more male patients (77.3% of the sample) than female patients (22.7%). The mean and median ages were 59 years and 1 month and 60 years and 6 months, respectively. Most of the patients were hospitalised in intensive care units (ICUs) (54.6%) and oncology departments (14.4%). Mean lymphocyte count in the patients was 0.5×109/L (the median lymphocyte count was identical). We did not calculate the mean CD4+ T-cell count, as values for this parameter were recorded in the notes of only 18% of the patients.

Table 1

Patient characteristics

Direct microscopy

The mean time required for this technique, from staining of the preparation to complete observation of the slide under the microscope, was 45 min. Two cases of pulmonary toxoplasmosis were diagnosed following the observation of T. gondii tachyzoites in BALFs (patients 1 and 2: see the Description of patient 1 and Description of patient 2 sections and table 2).

Table 2

Characteristics of the two cases of pulmonary toxoplasmosis

Polymerase chain reaction

All the PCR runs were successful (ie, no inhibition). Individual runs lasted an estimated 100 min (210 min from the extraction step). Two PCRs, for patients 1 and 2, gave positive results, with Cq values of 29.73 and 20.77, respectively. No additional cases of pulmonary toxoplasmosis were diagnosed by PCR.

Case report

Description of patient 1

Patient 1 was a 70-year-old woman with small lymphocytic lymphoma for which treatment had been unsuccessful. After third-line treatment with fludarabine, cyclophosphamide and rituximab, her total lymphocyte count had fallen to 0.13×109/L. Fever (40°C) and disseminated infiltrates on chest X-ray (figure 1A, B) and CT scan (figure 1C, D) led to an initial suspicion of P. jirovecii pneumonia. The patient was treated with intravenous pentamidine diisethionate, corticosteroids and a probabilistic antibiotic therapy. Following correction of the diagnosis by both the staining method (figure 2A) and PCR, the patient was treated with cotrimoxazole, and then with pyrimethamine and clindamycin, due to the myelosuppressive adverse effects of sulfonamide. Secondary drug treatment with atovaquone was administered to prevent immediate relapse. The serological status of this patient for toxoplasmosis was unclear, due to the recurrent intravenous administration of polyvalent immunoglobulins. Nevertheless, primary infection was considered most likely in this context (table 2).

Figure 1

Examples of radiological findings consistent with pulmonary toxoplasmosis. (A) Patient 1 chest X-ray 1 month before the onset of pulmonary symptoms. (B) Patient 1 chest X-ray at the onset of pulmonary symptoms. (C) Patient 1, chest CT scan 1 month before the onset of pulmonary symptoms; (D) Patient 1, chest CT scan at the onset of pulmonary symptoms.

Figure 2

Toxoplasma gondii tachyzoites in bronchoalveolar lavage fluids (BALFs) (May-Grunwald Giemsa, MGG, magnification ×1000). (A) One isolated extracellular tachyzoite (dark arrow) in the BALF of patient 1. Note the hook-like shape of the cytoplasm and the purple colour of the nucleus. The entire parasite is about 5 µm long. (B) Numerous intracytoplasmic tachyzoites (encircled) multiplying in a mononuclear cell in the BALF of patient 2. Also note the presence of miscellaneous bacteria probably originating from the commensal oropharyngeal flora.

Description of patient 2

Pulmonary toxoplasmosis was diagnosed in a 66-year-old man admitted to the ICU for invasive aspergillosis occurring 1 week after a second liver transplantation consecutive to mixed alcoholic and non-alcoholic steatohepatitic cirrhosis. The patient was treated with methylprednisolone, tacrolimus, mycophenolate mofetil and voriconazole. His total lymphocyte count was 0.04×109/L. Tachyzoites were found in the BALF (figure 2B) 7 days after the start of antifungal treatment. At this time, the patient presented persistent respiratory distress syndrome and chest X-ray showed signs consistent with the onset of interstitial pneumonia. Subsequent PCR analysis of BALF confirmed the diagnosis of pulmonary toxoplasmosis. The patient died 4 days later, before the introduction of antiparasitic medication. The patient's serological status for T. gondii infection was consistent with cystic reactivation, because this patient was known to have specific antibodies before his first liver transplant (table 2).

Discussion

Toxoplasmosis is a widespread infection that is rarely symptomatic, except in immunocompromised patients, who may suffer disseminated disease or toxoplasmic encephalitis.1 These two conditions have been widely described in patients with HIV infection and in individuals with haematological malignancies.4 ,5 ,14 ,15 Pulmonary toxoplasmosis is a debateable concept as several authors consider this condition to be a disseminated form. However, this condition is thought to be rarer,15–17 and has been reported in far smaller numbers of SOT8 ,18 and HSCT patients.19 ,20 Pulmonary toxoplasmosis should be treated as a genuine emergency, but it can be difficult to diagnose: its clinical signs are not specific and serological testing is not entirely reliable in immunocompromised individuals.7 ,20

In this prospective survey, we diagnosed two cases of pulmonary toxoplasmosis (incidence=0.017/1000 inpatient-days). This finding is consistent with assertions that pulmonary toxoplasmosis has become a rare event. Both cases occurred in the absence of specific primary chemoprophylaxis. This may have resulted in a larger parasite burden in the lung than would otherwise have been the case, facilitating direct diagnosis by microscopy. One of the two diseases was probably due to cystic reactivation, as it occurred in a liver transplant recipient. Indeed, the incidence of donor-acquired toxoplasmosis is less frequent in transplant patients receiving livers than in those receiving hearts,21 ,22 with only nine cases reported for mismatched patients.23 Likewise, such a mode of disease transmission appears highly unlikely in cases of HSCT, because this would require the donor to be parasitemic at the time of bone marrow donation.24 Chronic infections, or even primary infections via the usual route of contamination, are much more likely in these contexts.25 The two strains isolated in this study were of genotype 2 (data provided by the Centre National de Référence de la Toxoplasmose).26 This genotype is the most common among European strains, although it has been reported elsewhere that severe cases of toxoplasmosis may be linked to imported atypical isolates or variant strains.27

Direct microscopy examination of stained BALF for the detection of T. gondii tachyzoites remains an important tool providing a rapid diagnosis in cases in which the pulmonary parasite burden is high.1 Real-time PCR is a highly sensitive, easy-to-use technique and we hypothesised that it might be more reliable for diagnosis. However, the two cases reported here were first diagnosed by microscopy. This may reflect the considerable experience of the microbiologists of our teaching hospital, who are frequently confronted with difficult microscopic diagnoses (T. gondii is a very small parasite). It is also not possible to rule out the possibility that the T. gondii tachyzoites were observed purely by chance. At our centre, the use of PCR did not increase the efficiency of diagnosis for pulmonary toxoplasmosis, so the negative predictive value of direct sample examination was excellent. To our knowledge, this is the first systematic comparison of the use of these two methods on BALF. Another similar study differed from ours in that it was purely descriptive and focused solely on patients with HIV infection before the highly active antiretroviral therapies era.6 Furthermore, at the time of this previous study, molecular detection was carried out retrospectively to confirm the diagnosis of pulmonary toxoplasmosis, on only 6% of the BALFs. Furthermore, conventional PCR was used rather than real-time techniques. Other studies assessing the sensitivity of PCR were based on the use of blood and CSF samples in cases of disseminated or cerebral toxoplasmosis,28–31 but the results obtained were highly variable, with sensitivities ranging from 16% to 83%.

Our results, obtained by experienced staff at a teaching hospital, suggest that there is no advantage of systematic PCR over direct observation. However, the molecular detection of T. gondii in respiratory liquids may provide better results in laboratories with less microbiological experience.

Take home messages

  • PCR efficiently detects Toxoplasma gondii DNA in bronchoalveolar lavage.

  • Experienced microbiologists can obtain a rapid diagnosis by direct observation.

  • The incidence of pulmonary toxoplasmosis remains low.

  • Pulmonary toxoplasmosis mostly results from cystic reactivation in the context of immunosuppression.

Acknowledgments

We thank the technicians at the Parasitology Laboratory of Bretonneau Hospital, Tours: Béatrice Yzon-Champion, Ophélie Derouard and Lucie Chopin, together with Dr Josette Maheut-Lourmière for her precious help in interpreting the epidemiological data. The authors would also like to thank Dr Estelle Cateau for providing some of the serological data. We thank the French CNR for toxoplasmosis for the genotyping of the Toxoplasma gondii strains. Alexandra Farrell translated this manuscript from French into English. A professional scientific editing and translation company (Alex Edelman & Associates) then carried out a final revision of the English version.

References

Footnotes

  • Handling editor Slade Jensen

  • Contributors GD was involved in planning the study, writing the manuscript and reflecting about the data; EC, MG and CF-M were involved in the data collection, reflecting about the results and the practical works; the other authors were involved in the manuscript preparation and the reflecting.

  • Funding The data were obtained during routine work at Tours University Hospital, France.

  • Competing interests None declared.

  • Ethics approval The biological samples had been stored in the hospital collection No DC-20100-1216, approved by the French Ministry of Research. The study registration number No 2015_030 was issued by the Technology and Freedom National Committee (Commission de l'Informatique et des Libertés) on 27 May 2015. Final approval No 2015 23 was given by the Ethics Committee of Tours University Hospital (Espace de Réflexion Ethique, Région Centre, France).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Data related to PCR results are available. We state that we do agree to share this information if necessary.