Quantitative PCR for the Diagnosis of HCMV Pneumonia in HSCT Recipients and Other Immunocompromised Hosts

: Pneumonia is among the most serious manifestations of HCMV infection, with high morbidity and mortality. Probable pneumonia is deﬁned as the detection of HCMV in bronchoalveolar lavage (BAL) by viral isolation or DNA quantiﬁcation (qPCR) combined with symptoms and/or signs of respiratory infection. However, currently, there is no reproducible and well-deﬁned viral load (VL) from BAL that can reliably differentiate patients with pneumonia from the much more common detection of viral DNA in seropositive patients without true HCMV pneumonia. Several studies have been published with the aim of establishing an optimal VL for differentiating pneumonia from viral lung shedding. The aim of this review is to collect and analyze the methodology and the conclusions obtained in studies whose objectives included the correlation between HCMV VL in BAL and/or the plasma and the occurrence of HCMV pneumonia. For this purpose, a total of 14 articles have been included. There are some conclusions on which they all agree. PCR techniques were more sensitive and had a higher NPV than culture techniques but were less speciﬁc and had a low PPV. The mean HCMV loads in both BAL and the plasma were signiﬁcantly higher in patients with pneumonitis than in those without. The HCMV load in patients with pneumonitis was higher in BAL than in the plasma, making qPCR in BAL a better predictor of HCMV pneumonitis than in the plasma. Nevertheless, this review highlights the difﬁculty of establishing a universal VL value, both in BAL and in the blood, to differentiate patients with HCMV pneumonia from those without. To complete the information available in these studies, prospective multicentre studies would be required. Methodologically, a large number of patients with HCMV pneumonitis would have to be included, and a subclassiﬁcation of the type of immunosuppression of each patient should be made in order to obtain an optimal VL threshold in different host groups.


Introduction
Human Cytomegalovirus (HCMV) is a double-stranded DNA virus of the family Herpesviridae, subfamily Betaherpesvirinae. It is one of the largest viruses known to cause clinical disease (230 kb). The primary infection usually occurs in the first years of life, by direct contact with secretions of infected individuals, such as saliva, breast milk or urine. The seroprevalence in adults is between 50 and 98%, being higher in developing countries. In the immunocompetent host, the primary infection is usually asymptomatic or presents as a self-limited mononucleosis-like syndrome characterized by fever, lymphadenopathy and lymphocytosis. After primary infection, HCMV remains latent for life in different cells. In immunocompromised hosts (IMC), such as hematopoietic stem cell transplantation (HSCT) or solid organ transplant (SOT) recipients or patients with HIV, uncontrolled viral replication can occur, both after a primary infection, but, most commonly, after reactivation of the latent virus. In these patients, HCMV can cause severe clinical disease, mainly pneumonitis, colitis or retinitis. The incidence of HCMV disease has decreased in recent years thanks to strategies such as pre-emptive therapy and, more recently, due to prophylaxis in HSCT recipients. Nevertheless, HCMV remains one of the most feared opportunistic pathogens in IMC, especially in those with impaired T-cell mediated immunity [1,2], and is certainly one of the viruses for which HSCT clinicians screen for the most both before and after transplantation.
Pneumonia is among the most serious manifestations of HCMV infection, with high morbidity and mortality, even with antiviral treatment. Allogeneic HSCT recipients are particularly at high risk of HCMV pneumonia, with an extremely variable incidence nowadays (from <1% to 30% of patients with HCMV reactivation) and a mortality of up to 70% [1][2][3][4]. The radiological findings and symptoms are nonspecific, which makes diagnosis difficult and delayed. This is paradoxically a more important clinical conundrum nowadays since true HCMV pneumonia has decreased while the development of pneumonias of unknown origin has increased. In chest radiographs, HCMV pneumonitis typically manifests as diffuse interstitial infiltrates, although a nodular pattern may be observed, and completely nonspecific lung infiltrates may also be due to HCMV pneumonia [1,5].
To define and unify the concepts of HCMV infection and disease for use in clinical trials, a consensus report was developed at the Fourth International Conference on HCMV in Paris in 1993. Since then, significant advances have been made in the diagnosis and treatment of HCMV infection, leading to several guideline updates, the latest in 2017 [6]. It differentiates between proven and probable HCMV pneumonia. For proven pneumonia, symptoms and/or signs of respiratory infection are needed with the presence of HCMV identified in lung tissue (i.e., biopsy or autopsy) by virus isolation, rapid culture, histopathology, immunohistochemistry, or DNA hybridization techniques. Probable pneumonia is defined as the detection of HCMV in bronchoalveolar lavage (BAL) by viral isolation or DNA quantification (qPCR) combined with symptoms and/or signs of respiratory infection. However, currently, there is no reproducible and well-defined viral load (VL) from BAL to differentiate patients with pneumonitis from the much more common detection of viral DNA in seropositive patients without true HCMV pneumonia. This problem is, of course, not limited to HCMV but is also the case for other human herpesviruses, especially HHV-6, and Pneumocystis jirovecii, to cite only a few of the BAL "positive PCR" difficultto-interpret pathogens. This makes it difficult to interpret the results of qualitative reversetranscriptase PCR. At least in theory, a quantitative PCR (qPCR) would be more helpful in identifying patients at high risk of having true HCMV pneumonia for clinical decisionmaking. However, qPCR from BAL is far from being standardized.
Definitive pneumonia requires analyzing lung tissue, but obtaining lung biopsies for the detection of HCMV by immunohistochemistry requires invasive techniques, which are nearly always contraindicated, especially in HSCT recipients. Currently, the use of BAL samples for the diagnosis of pneumonia in these patients is already widely extended and accepted [7].
In the last 20 years, the laboratory methods used to detect the virus have also been changing with the development and introduction of molecular techniques. However, one should never forget that proving the presence of a viable, replicating virus is the only way of extrapolating active disease in the clinical sample tested. Thus, viral culture is considered the gold standard for the diagnosis of active HCMV disease, either by traditional tube culture or by culture through centrifugation (shell vial). These allow the recovery of the virus that is replicating and has a high correlation with the occurrence of disease [1,8]. The main limitation of traditional culture is the time needed to observe the characteristic cytopathic effect of HCMV replication, which can range from 1 to 4 weeks, which limits its usefulness in diagnosis and clinical decision-making. However, thanks to the centrifugation of the culture in a shell vial, it is possible to obtain results within 24-48 h, which increases its usefulness in clinical practice, as well as being more sensitive than the traditional tube [1,8]. The detection of the HCMV pp65 antigen expressed on leukocytes during the early period of infection (antigenemia) has also been a widely used technique for the diagnosis and monitoring of HCMV infection. This technique has many limitations, such as a lack of standardization, laboriousness, and the need for samples with an adequate number of leukocytes in the blood (difficult in transplant patients) [8]. Both culture-based techniques and antigenemia have been displaced in most laboratories by simpler, faster, and more sensitive techniques, such as DNA detection by PCR. It should be kept in mind that HCMV persists latently in cells after primary infection without causing disease, and this non-replicating DNA may be detected by PCR techniques. This makes the detection of HCMV DNA a less specific indicator of active virus replication and disease than culture. The widespread use of PCR techniques and the possible misinterpretation of their results (possible detection of non-replicating DNA) may lead to the over-diagnosis of HCMV disease and, consequently, over-treatment [9,10].
Today, DNA quantification using qPCR is the most widely used PCR method for the diagnosis of HCMV infection or monitoring the patients' VL, especially in peripheral blood samples. Blood screening by qPCR for HCMV DNAemia is recommended at least once a week during the high-risk period after transplantation and to monitor response to preemptive antiviral treatment. The major limitation of qPCR is a lack of well-established VL thresholds to guide various clinical applications, a consequence, above all, of the variability between the assays marketed for this determination, despite their calibration to the WHO International Standard. This limitation applies to any sample analyzed, but in the case of HCMV pneumonia, its diagnosis using BAL faces further issues that complicate its interpretation. Thus, in addition to the lack of standardization in obtaining BAL samples, their subsequent processing also introduces variability when it comes to quantifying HCMV DNA. Finally, clinically relevant threshold BAL VL certainly varies depending on the type of the patient [8,11].
Several studies have been published with the aim of establishing the optimal VL of HCMV by qPCR in BAL samples for differentiating patients with pneumonia from those without. The quantification of HCMV in the blood (DNAemia) has also been studied for its potential role in diagnosing patients with pneumonia. In both cases, a universal conclusion has not been reached.
The aim of this review is to collect and analyze the methodology and the conclusions obtained in studies whose objectives included the correlation between HCMV VL in BAL and/or the plasma and the occurrence of HCMV pneumonia.

Materials and Methods
For this purpose, a search was carried out in the Pubmed database. This search consisted of three components: "cytomegalovirus", "bronchoalveolar lavage", and "viral load". "cytomegalovirus" [Title/Abstract] AND "bronchoalveolar lavage" [Title/Abstract] AND "viral load" [Abstract].
Only articles with full-text availability that were written in English between 1992 and October 2022 were included. This search generated 47 results as of 14 November 2022. After reviewing these articles, we excluded those that dealt with ventilator-associated pneumonia or nosocomial pneumonia and those whose primary objectives did not include the study of HCMV pneumonia. This resulted in 14 studies whose objective(s) included the relationship between HCMV DNA load in BAL and HCMV pneumonia.

Results
The results of the 14 studies analyzed are shown in detail in Table 1.

Conclusions
As shown in this review, in recent years, several studies have been carried out to determine the relationship between the HCMV VL in BAL and the plasma with HCMV pneumonia.
There are some conclusions on which they all agree. PCR techniques were more sensitive and had a higher NPV than culture techniques but were less specific. HCMV DNA detection in BAL does not always imply disease, so quantifying VL should, in theory, be important for clinical interpretation. The mean HCMV load in both BAL and plasma was significantly higher in patients with pneumonitis than in those without. The HCMV load in patients with pneumonitis was higher in BAL than in the plasma, making qPCR in BAL a better predictor of HCMV pneumonitis than in plasma.
Nevertheless, this review highlights the difficulty of establishing a universal VL value, both in BAL and in blood, to differentiate patients with HCMV pneumonia from those without (HCMV DNA shedding). The results and conclusions obtained in these studies are quite different, partly because of the methodological variability of the different studies. In addition, only a few studies used the reference techniques (HCMV detection in biopsy or isolation by culture) to compare viral load results and define HCMV pneumonia cases [16,20]. In each of them, a different value of VL in BAL indicative of active replication in the lung was obtained, ranging from 500 IU/mL [18] to >500,000 copies/mL [20]. On the other hand, some authors concluded that it is not possible to obtain a VL threshold that allows the identification of patients with HCMV pneumonitis [11,22]. Regarding the value of DNAemia testing in the diagnosis of HCMV pneumonia, there are also discrepancies. Some authors found that in up to one-third of patients with pneumonia, HCMV DNA was not detected in the blood [9,19]. Conversely, others claim that in patients in whom HCMV is not detected in plasma, pneumonia can be ruled out [12].
These authors reflect in their work that there are many factors that may influence the VL value in BAL. The first and most important of these is the dilutional effect of bronchoscopy. The serum instilled in the procedure is distributed heterogeneously, and both the volume of lung bathed and the amount of fluid recovered are highly variable among individuals. This makes the quantification in the BAL highly variable. Another important factor is related to the lack of standardization of the tests available in the different laboratories for the quantification of HCMV DNA. In the studies included in the review, it can be observed that different protocols and tests are used in each of them for this purpose. Aspects such as assay performance (limits of detection and quantification), the method for DNA extraction, gene target, and amplicon size contribute to VL variability. In addition, clinically relevant VL values are likely to differ depending on the type of patients and their risk profiles.
Additional caveats include the small number of patients with HCMV pneumonitis in each study. In addition, not all studies used direct diagnostic tests such as viral culture or histology in lung biopsies to establish the level of evidence for CMV pneumonia. Finally, some studies used qPCR assays that were not fully automated and required previous sample handling prior to the extraction of the DNA.
To complete the information available in these studies, prospective multicenter studies would be required. Methodologically, a large number of patients with HCMV pneumonitis would have to be included, and a subclassification of the type of immunosuppression of each patient should be made in order to obtain a VL value in BAL according to the risk of each group. Such a study is, of course, logistically impossible to perform. In addition, in an ideal world, the BAL-related variability would have to be reduced by using a common BAL procedure, another impossible objective. Finally, uniform virologic tests would be needed with all centers participating in the study using the same qPCR assay, which should be as sensitive and automated as possible, and HCMV cultures should be performed on all BAL samples. Of course, concomitant plasma qPCR should be performed, which is the only feasible test currently available in all centers. However, certainly, the most problematic issue would be the very few cases of histologically proven disease that would be available due to the very low rate of lung biopsies performed, especially in alloHSCT recipients.