Stewardship of Molecular Diagnostics in Transplant Viral Infections

ABSTRACT

The transplant environment requires special considerations when testing for viral infections as immunosuppression results in atypical infection profiles. Microbes otherwise considered commensals or causing mild disease can lead to severe infections in transplant environments. Therefore, guidelines tend to recommend broader microbial testing in these populations. In parallel, advances in molecular diagnostics have led to the availability of a wide selection of tests, including highly multiplexed nucleic acid amplification tests (NAATs) and direct next generation sequencing (NGS) based options. These newer technologies may provide information on many potential pathogens simultaneously, more rapidly, and while avoiding invasive specimen collection procedures. However, they are generally more expensive than conventional methods such as culture, and nucleic acid detection of multiple potential pathogens may be nonspecific and confuse the diagnosis. Navigating the complexity of the available molecular test landscape in immunocompromised patients is an opportunity for diagnostic stewardship. Here we discuss the clinical value of different molecular testing strategies for diagnosis of viral infectious diseases in immunocompromised transplant patients using several common transplant infection syndromes as a framework.

1. Introduction

Infections in solid organ and hematopoietic stem cell transplant recipients (SOTRs; HSCTRs) are common and cause significant morbidity and mortality. In contemporary studies, 55% of SOTRs experienced infection in the first posttransplant year [1], accounting for 17% of SOTR deaths [2]. Advancements in pharmacotherapies improved graft rejection rates and infection‐related mortality versus prior eras [3], but SOTRs and HSCTRs experienced higher mortality during the COVID‐19 pandemic [4, 5].

Before transplant, immune dysfunction results from hematologic malignancies, autoimmune disease, decompensated end‐organ disease, diabetes mellitus, advanced age, or nutritional factors. Following transplant, rejection is averted by inducing graft tolerance with immunosuppressants whose narrow therapeutic indices require close monitoring. The ensuing cell‐mediated and humoral immunodeficiencies predispose patients to infection and malignancy. In both solid organ transplant and hematopoietic stem cell transplant (SOT, HSCT) graft tolerance induction protocols, immunosuppressive intensity decreases over time after successful engraftment.

A myriad of infections may follow transplantation—post‐procedural, primary, localized, reactivated, disseminated, opportunistic, antimicrobial‐resistant, and chronic—caused by a vast array of environmental and nosocomial bacteria, viruses, mycobacteria, fungi, or parasites. Specific transplant infections and pathogens have been reviewed extensively [6]. Infection risk varies with donor and recipient epidemiologic risk factors, latent infections, serologic status, transplanted tissue, time elapsed since transplantation, era and country of transplantation, and equilibrium between immunosuppressive and prophylactic strategies [7].

Given the high stakes and broad differential, attaining early, specific, and actionable diagnoses is imperative. However, immunosuppression masks localized and systemic inflammation, causing variable or absent clinical signs and symptoms. Clinical signs like diarrhea may be indistinguishable from graft rejection or drug‐toxicity [8]. Further, the timeline of early (< 30 days), intermediate (1–6 months) and late (> 6 months) posttransplant infections has become less predictable and opportunistic infections are less common, foiled by robust vaccination strategies, rigorous screening and treatment of latent infections, and routine antimicrobial prophylaxis [1]. Conversely, emerging antimicrobial resistance (AMR) has amplified the impact of nosocomial and community‐acquired bacterial and fungal infections [6]. In contemporary studies, 32% of SOTRs already harbor multidrug‐resistant organisms (MDROs) at transplantation [9], and prevalence and recurrence of Clostridioides difficile infection (CDI) in SOTRs are 7% and 20%, respectively [10].

Against this challenging landscape, transplant infections pose a unique opportunity to optimize patient outcomes through diagnostic stewardship. Molecular diagnostic methods provide benefits over conventional ones, including rapid turnaround time, increased sensitivity, and may obviate the need for invasive procedures. Indeed, current guidelines for viral diagnosis in transplant patients may acknowledge the use of other technologies, but tend to emphasize nucleic acid amplification test (NAAT)‐based methods as state‐of‐the‐art and of high sensitivity [11, 12]. Molecular diagnostics have important limitations however, including cost, healthcare waste through overuse, overdiagnosis and overtreatment from misunderstanding of analytical performance or misinterpretation of false positive or negative results [13, 14].

This review describes recent advances in the molecular armament and discusses the role of representative platforms in diagnosing transplant infections. Understanding their performance, benefits and limitations, enables appropriate use, correct interpretation of results, and maximizes the value added by each test. We highlight the role of diagnostic stewardship in delivering high‐value care to transplant recipients (summarized in Table 1). Ideally, ordering molecular diagnostic tests should be justified by the pretest probability of a suspected infection, and their diagnostic accuracy.

TABLE 1.

Summary of diagnostic stewardship opportunities.

Syndrome	Opportunities for clinical Dx stewardship	Opportunities for laboratory Dx stewardship
General considerations	Viral prophylaxis status Vaccination status D/R sero‐status Degree of immunosuppression Type of immunosuppression (e.g. T cell vs. B cell depletion) Sick contacts and local epidemiology Age of patient Travel history Pitfalls Timelines of classic early (< 30 days), intermediate (1–6 months) and late (> 6 months) posttransplant infections have become less predictable Opportunistic infections are less common Using differential of immunocompetent people in transplant patients Different kinetics regarding detection of pathogens relative to immunocompetent persons (e.g. length of illness, speed of onset, presentation)	Offering algorithms based on time of year or regional prevalences, nudging, framing, and preventing repeat orders within specific time frames
Upper and lower respiratory tract infections	Wide range of potential infectious etiologies warrant broad testing Temporal disease prevalence could warrant narrower first line testing (e.g., "flu season") Local or intra/inter hospital epidemiology could warrant narrower first line testing (e.g. SARS‐CoV‐2 outbreak on wards or within community) If first round testing is not fruitful; revise DDx to include noninfectious causes and systemic illness without direct infection of airways Consider potential for infectious synergies (e.g. influenza‐streptococcal coinfections) Pitfalls Systematic infection can lead to respiratory distress without direct infection of the airways (e.g. CMV systematic illness) Symptom durations are less specific in transplant patients.	NAAT panels with breadth determined by DDx. Limit highplex panel usage when limited set of pathogens are of highest prevalence (e.g., "flu season"). Lack of detection on smaller panels warrants use of larger panels ("casting a wider net") NGS may be worthwhile in transplant patients owing to potential for exotic etiological agents. Given current lack of evidence of utility, restrict NGS to last‐line testing. Pitfalls Higher viral loads and asymptomatic shedding kinetics can be red herring, particularly adenovirus and SARS‐CoV‐2, but other viruses possible in transplant patients Potential of co‐detections confound identification of true causative agent(s) Evidence supporting clinical relevance of organisms detected by NGS is mixed
Central nervous system infections	Wide range of potential infectious etiologies warrant broad testing Catastrophic consequences of missing life‐threatening, potentially treatable diagnoses. "Can't miss" bacterial infectious warrant rapid molecular diagnosis. Headache, neck stiffness, photophobia, or signs of brain or spinal cord involvement, like seizures, encephalopathy, or focal neurologic deficits trigger suspicion for CNS infection Acuity of onset Rash and skin lesions may help narrow first line testing but does not rule in or rule out any infection reliably Co‐presentation with GI symptoms can point toward enterovirus or adenovirus infection CSF opening pressure, cytology, WBC count and differential, CSF glucose, and CSF protein all suggest different microbial etiologies to guide second‐line testing Negative tests results using highly sensitive molecular tests should prompt the clinician to reevaluate the differential diagnosis and consider alternative causes, especially with failure of empiric therapy or progression of CNS infection Pitfalls Direct infection of CNS vs. secondary effects of systematic reactivations (e.g., EBV, JC virus) Non‐viral causes of CNS infections are considerable Imaging is not specific for CNS infections Rash and skin lesions may help narrow first line testing but is not sufficiently accurate for diagnosis. Lesion presentation may be atypical in immunocompromised persons.	Highplex NAAT panels are often warranted in immunocompromised persons More esoteric testing if warranted by clinical presentation (e.g. EBV, JC virus) Singleplex testing for HSV, etc. may be warranted if original testing does not yield a causative agent and suspicion is high NGS may be worthwhile in transplant patients owing to the potential for exotic etiological agents in the transplant population, particularly "aseptic" organisms (i.e. pathogens that cannot be identified using routine techniques) Pitfalls Singleplex options of significantly higher performance may warrant their use Nonmolecular diagnostic techniques of higher performance for some pathogens (WNV, cryptococcal antigen) Observed poor accuracy and/or discordance for HSV‐1 target for commercially available highplex tests Pleocytosis as pre‐gate criteria may exclude severely immunocompromised persons from testing inappropriately Blood cultures as a minimum. Perform CSF cultures unless precluded by strong contraindications.
Gastrointestinal infections	Community vs. hospital acquired Noninfectious diagnoses and medication toxicity may rise in DDx if molecular testing fails to detect source of infection Pitfalls Many noninfectious etiologies can mimic infection Systematic infection can lead to GI distress without direct infection of the gut, requiring non‐stool sample testing (e.g. CMV systematic illness) Ordering stool sampling without meeting clinical criteria for diarrhea can lead to false positive test results Differentiating Clostridioides difficile colonization from infection using molecular methods	Highplex NAAT panel for all immunocompromised patients with diarrhea is recommended with symptoms lasting more than 7 days For hospital acquired infections, only accept orders for stool pathogen testing within the first 3 days of hospitalization Acuity vs. chronicity can inform causative pathogen Pitfalls Observed inaccuracies around norovirus detection for commercially available highplex tests Large panels may lead to a very high level of organisms detected Detection of asymptomatic colonizers; persistent asymptomatic shedding of norovirus and adenoviruses Standard tests such as stool culture, antigen testing, and microscopy have been of low yield and time consuming
Undifferentiated systematic illness	CMV and EBV high on differential Consider geographic distributions Consider tissue tropisms when testing for tissue of origin Rash and skin lesions may help narrow first line testing but is not sufficiently accurate for diagnosis. Lesion presentation may be atypical in immunocompromised persons. Evaluate lifestyle, animal exposure, travel, and dietary habits Consider potential for infectious synergies (e.g. CMV co‐activation with other herpesviruses) Consider infection potentiating graft rejection or organ failures (consider specific organ biopsy and testing) Pitfalls Pleomorphic manifestations that overlap various syndromes Lesion presentation may be atypical in immunocompromised persons High HHV‐6 viral loads can indicate asymptomatic chromosomal integration	Quantitative, serial NAAT measurements are often required Systemic detection may not be diagnostic. Organ biopsy and perhaps in situ detection may be required to diagnose the source of infection (often the transplanted organ in SOT). Pitfalls Qualitative detection of many viruses, particularly DNA viruses, is insufficient for definitive diagnosis Highplex NAAT not available for viral infections Highplex quantitative NAAT not available

Abbreviations: CMV, cytomegalovirus; CNS, central nervous system; CSF, cerebrospinal fluid; D/R, donor/recipient; DDx, differential diagnosis; EBV, Epstein‐Barr virus; GI, gastrointestinal; HSV, herpes simplex virus; NAAT, nucleic acid amplification test; NGS, next generation sequencing; SARS‐CoV‐2, severe acute respiratory syndrome coronavirus 2; SOT, solid organ transplant; WBC, white blood cell; WNV, West Nile virus.

To illustrate the advantages of judicious use of molecular methods, we use common transplant infectious syndromes as a framework, and viral infections as model diagnoses. Viral infections in transplant recipients are generally more severe and result in worse outcomes, including treatment failure, prolonged hospitalization, and mortality [15]. This is partly attributed to atypical symptoms, prolonged illness course, delays in diagnosis and appropriate therapy, and inappropriate empiric therapies. Diagnosing viral infections facilitates de‐escalation of unnecessary empiric antimicrobials, justifies decreased immunosuppression, facilitates stewardship of resources, decreases healthcare costs and unintended consequences, and has public health implications [16, 17, 18].

2. Diagnostic Stewardship in Transplant Infections: A Clinical Approach

Indiscriminate “shotgun” testing for broad and rapid results has unintended consequences including misdiagnosis, unnecessary use of antimicrobials, hospitalizations, and additional testing [19, 20]. Institutions and physicians may face excess costs, diminished reimbursement, and loss of accreditation or reputation [21, 22]. While transplant recipients are indeed at risk of severe infections, indiscriminate testing is governable without compromising accurate and early diagnosis in this high‐risk population.

The value of diagnostic tests is proportional to their appropriateness and quality (accuracy) [23]. Diagnostic stewardship refers to coordinated interventions that optimize the process from diagnostic test selection (preanalytic) to interpretation (postanalytic) [24]. The objective is to ensure effective use of healthcare resources to deliver high‐quality patient care. Multidisciplinary diagnostic stewardship teams are guided by core principles including evidence‐based testing, efficiency, acceptance of cognitive limitations and biases, understanding test performance parameters, and avoidance of redundant testing, among others [25].

Diagnostic and antimicrobial stewardship programs enhance high‐value care by curtailing unnecessary testing and harm to transplant recipients during sample collection or from misdiagnosis [26]. Examples of effective initiatives in transplant include avoidance of screening and treatment of asymptomatic bacteriuria (ASB) in kidney SOTRs [27, 28], and guideline‐concordant C. difficile testing to increase specificity for active colitis [29]. These best practices are backed by clinical guidelines to prevent misdiagnosis of ASB as urinary tract infection, unnecessary hospitalizations and antibiotics, and to mitigate MDRO and C. difficile colonization [30, 31].

Our diagnostic approach entails probabilistic reasoning and stepwise targeted testing, beginning with patient symptoms or functional alterations with undefined causes. To characterize clinical problems, their onset, duration, and distinguishing features are synthesized, without anchoring bias, with predisposing risk factors: donor/recipient seropositivity, latent infections, time since transplant, vaccinations, prophylaxis, immunosuppression, geography, and recent exposures. Initial clinical reasoning suggests a tiered differential diagnosis of specific etiologies ranked by probability. Vital signs, physical examination, imaging, chemistry and hematology findings narrow and refine the differential. If a patient exhibits classic highly specific, pathognomonic signs corresponding to a syndrome, diagnosis is established, and therapy may commence without further confirmatory testing. However, initial clinical reasoning is not always successful at pinpointing exact diagnoses. More often, it localizes problems to organ systems and important transplant syndromes: (1) respiratory tract infections, (2) central nervous system (CNS) infections, (3) gastrointestinal (GI) infections, and (4) undifferentiated systemic infections. These have established differentials that may be ruled in or out by laboratory‐based diagnostic testing.

The following sections describe our process for defining diagnosis in important transplant infectious syndromes. Beginning with clinical presentation and features that suggest a differential diagnosis, we consider the pretest probability of potential diagnoses. We then describe the rationale for ordering or deferring molecular diagnostic testing (including single pathogen or lowplex tests versus highplex syndromic panels), reviewing the pertinent performance parameters and predictive values of different testing strategies.

2.1. Upper and Lower Respiratory Tract Infections

Upper and lower respiratory tract infections (URTI/LRTIs) are common in transplant recipients. Although URTIs like rhinosinusitis and laryngopharyngitis are typically localized and self‐limiting, their progression to lower tract invasion is facilitated by immunosuppression [32]. However, transplant URTIs may not present as classic influenza‐like illness [33]. Transplant recipients with LRTIs may present with cough, dyspnea, chest pain, or sputum production, or systemic symptoms like fever or chills. Unlike immunocompetent patients, acuity and symptom duration are less specific for viral infection in SOTRs and HSCTRs, in whom symptoms and viral shedding are prolonged [33]. Laboratory diagnosis and treatment of URTI may reduce the occurrence of LRTI. LRTI syndromes, tracheobronchitis and pneumonia, may occur without upper tract involvement, warranting a heightened index of suspicion.

Common LRTI viral pathogens in transplant include influenza A virus, influenza B, severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) and other coronaviruses, respiratory syncytial virus (RSV), human metapneumovirus (hMPV), adenoviruses, parainfluenza 1–4, rhinoviruses and enteroviruses. Clues to viral etiology include recent symptom onset, close contact with ill persons, lack of vaccination, time of year, local epidemiological activity of seasonal respiratory viruses and associated upper respiratory symptoms. Cytomegalovirus (CMV), herpes simplex virus (HSV)‐1, varicella zoster virus (VZV), and measles may cause pneumonitis as part of systemic infections. Detection of CMV DNA in sputum is associated with, but not specific for, CMV pneumonitis [34]. Adenovirus causes both direct primary respiratory tract infection and indirect pneumonitis. Importantly, the airways of transplant recipients are disproportionately colonized with antimicrobial‐resistant bacteria [9]. Important noninfectious LRTI mimics include malignancy, graft rejection, pulmonary edema, hypersensitivity pneumonitis and sirolimus toxicity [35, 36].

Diagnostic testing for LRTI centers on patient status and disease severity. In hemodynamically stable patients with acute LRTI, up‐front molecular diagnostic testing may include (based on season and travel history) a lowplex nasopharyngeal NAAT that detects prevalent circulating viruses with lung tropism (i.e., RSV, Influenza A and B, SARS‐CoV‐2), which offers several advantages over broader upper or lower respiratory panels, including lower cost [37]. RSV and influenza viruses are highly seasonal in most geographic regions. When their prevalence is near zero, and in the absence of recent travel to a region with viral activity, it may be more efficient to test up front with a broader multiplex panel. If a lowplex NAAT is positive, targeted antiviral treatment may be initiated and empiric antibacterials discontinued, unless other clinical factors suggest concurrent or post‐viral bacterial infection. Bacterial coinfections are relatively rare with SARS‐CoV‐2 [38], whereas influenza is well‐known to predispose to bacterial pneumonia [39].

Given the greater susceptibility of transplant patients, broader viral panel testing is usually warranted if an etiological agent is not detected using narrow testing. In addition, as immunocompromised patients are more susceptible to coinfections from multiple organisms, the use of highplex panels may lead to clinically relevant co‐detections more frequently in transplant patients. Recent guidelines recommend broad testing of common viral agents in transplant patients, including Influenza, RSV, PIV, hMPV, Rhinovirus, and human coronaviruses [40]. Still the value of detecting viral targets without available therapeutic interventions should be weighed against the extra cost of repeat testing with more expansive panels (e.g., epidemiological tracking, changes to prognosis and supportive care).

Severe LRTIs with respiratory failure or hemodynamic instability may require empiric antibacterial therapy before establishing a microbiologic diagnosis [41]. Guideline‐concordant empiric therapy for immunosuppressed transplant recipients often targets nosocomial drug‐resistant pathogens given prior healthcare encounters, parenteral antibacterials, and prevalent MDRO colonization. Broader multiplex NAAT pneumonia panels are appealing for diagnosis in patients with severe LRTIs, and testing is recommended when results are likely to be actionable [40], such as when therapeutic management is expected to change.

As most viruses causes of respiratory infections are not part of a normal human flora, nor persist as a part of a latent lifecycle, qualitative NAAT‐based panels are well suited for detection of viral respiratory infections in most circumstances [11, 42]. NAAT respiratory panels initially designed for URTIs using nasopharyngeal samples may act as acceptable surrogates when suspecting that infection has spread to the lower airways [11, 43, 44, 45]. However, the analysis of lower respiratory tract specimens has shown increased clinical sensitivity [37]. Newer highplex panels designed specifically for detection of pneumonia pathogens from lower‐lung specimens (e.g., sputum) are commercially available, but only one of these panels queries for viral pathogens [46]. Identification of viruses may facilitate antibacterial de‐escalation, but studies have shown that identification must be bundled with other antimicrobial stewardship interventions for the greatest impact [47, 48]. The role of procalcitonin in discriminating bacterial from viral LRTI is debated [49]. In contrast to procalcitonin, there are now tests that employ machine learning and multiplex host‐response signatures (proteins or host messenger RNAs) to differentiate bacterial and viral causes of infection, offering rapid, potentially actionable results. In addition to distinguishing viral from bacterial infections, these tests may assess illness severity. Clinical trials are underway to assess their application in immunocompromised hosts to determine impact on antibiotic use for viral cases, mitigate AMR, and reduce hospital length of stay or unwarranted hospital admissions [50]. Additionally, highly multiplexed pneumonia panels increase the detection rate of respiratory viruses relative to the older test modalities [44], which must be interpreted thoughtfully in clinical context, as nucleic acids may be detectable in the absence of viable infectious particles and may not necessarily represent the causative agent of infection (i.e., analytically true positive, but clinically false positive results). For example, a patient with LRTI and worsening clinical status may improve if hospital‐acquired influenza is detected and treated. However, a positive influenza NAAT would confound management if the same patient had been improving clinically (although it might facilitate infection control).

Prolonged shedding of many viruses is more common in transplant patients, with which may confuse diagnosis if lab results are interpreted outside of the clinical context. Adenoviruses and coronaviruses, including SARS‐CoV‐2, are especially notable [53, 54], with adenoviruses in particular are associated with prolonged asymptomatic, and sometimes intermittent, shedding [51]. Indeed, 2019 guidelines from the American Society of Transplantation Infectious Diseases Community of Practice note that “Recovery of adenovirus from urine, respiratory, or stool specimens by culture or PCR does not confirm adenovirus disease since patients can asymptomatically shed for prolonged periods of time”[42]. In addition (and related), detection of adenovirus in large multiplex NAAT panels has been problematic in the past, and notable differences in clinical specificity have been observed amongst different commercially‐available tests [52]. Singleplex NAATs targeting adenovirus alone have been shown to be more accurate than first generation multiplex NAATs [53]. More modern multiplex assays have improved accuracy of adenovirus detection, but differences in sensitivity and species coverage remain [54, 55].

Finally, commercially available NAAT respiratory panels do not include all pathogens capable of causing LRTI in transplant patients. Therefore, although broader pneumonia panel use may be warranted in certain circumstances, one‐off single‐target testing (e.g. Pneumocystis jirovecii, Nocardia) may still be indicated to fill the gaps in current panel comprehensiveness.

Inappropriate reasons to order multiplex syndromic NAAT panels include confirmation of a different positive test, “tests of cure,” or to appease patients or families and avoid litigation. Tests of cure are discouraged as nucleic acids alone are not specific for active infection. Research regarding the optimal high‐value use of multiplex LRTI and URTI panels in immunocompromised patients is ongoing [56]. One large academic center evaluated two different interventions across a broad patient population: a reflexive highplex NAAT URTI panel only after a negative influenza/RSV PCR test, and noninterruptive clinical decision support. The investigators determined that the noninterruptive clinical decision support interventions were more effective in reducing unnecessary highplex NAAT respiratory panel testing [57].

2.2. CNS Infections

Fever, accompanied by acute to chronic meningeal symptoms, headache, neck stiffness, photophobia, or signs of brain or spinal cord involvement, like seizures, encephalopathy, or focal neurologic deficits trigger suspicion for CNS infection. These are medical emergencies and the threshold for undertaking invasive or costly molecular diagnostic testing is appropriately decreased. Molecular panels that detect pathogens causing meningitis and encephalitis provide a faster diagnosis, improved antimicrobial stewardship and hospital cost savings compared to other technologies [58]. However, history and examination followed by targeted testing guided by clinical reasoning should not be entirely dismissed for indiscriminate testing.

When it comes to immunocompromised patients, the differential diagnosis of meningoencephalitis almost always warrants use of a syndromic panel. Reactivated and primary CNS infections with herpesviruses, including CMV, HSV‐1, HSV‐2, Human herpesvirus (HHV)‐6, and VZV may be seen in high‐risk recipients who are not on viral prophylaxis. Enterovirus infections are common in most populations particularly children [59, 60]. Human parechovirus (HPeV) is a common cause of meningoencephalitis in young children [61]. Human immunodeficiency virus (HIV), lymphocytic choriomeningitis virus (LCMV), and arboviruses are infrequent, but have specific risk factors that can be queried to assess pretest probability before testing. JC virus and Epstein–Barr virus (EBV) can also cause CNS complications in the form of progressive multifocal leukoencephalopathy (PML) and posttransplant lymphoproliferative disease (PTLD) respectively. But clinical presentation and time‐to‐symptoms are often quite distinct from other forms of virus‐mediated CNS disease, allowing clinical suspicion to guide more targeted testing. Detection of JC virus and EBV in the CSF alone is less specific for disease etiology relative to detection of viruses capable of causing direct acute infection [62, 63, 64, 65, 66].

Non‐viral causes of CNS infections are considerable and include Streptococcus pneumoniae, Listeria monocytogenes and Cryptococcus neoformans. Despite widespread pneumococcal and meningococcal vaccination, the incidence of bacterial meningitis in SOTRs and HSCTRs is elevated, associated with lack of vaccination, hypogammaglobulinemia, functional or anatomic asplenia, and corticosteroids.

When initial evaluation is suspicious for CNS infection, blood cultures and empiric antimicrobials are initiated according to clinical practice guidelines [67, 68]. The diagnostic process should include lumbar puncture with opening pressure for cerebrospinal fluid (CSF), unless precluded by strong contraindications [67, 68]. US guidelines recommend neuroimaging including computerized tomography (CT) for immunocompromised patients suspected of CNS disease [67]. However, magnetic resonance imaging (MRI) or CT are variably sensitive at detecting ischemia, mass lesions, inflammation, bleeding, demyelination or trauma, but are not specific for CNS infections.

Conventional CSF testing for glucose, protein, cell count and differential, combined with the high negative predictive value of a multiplex meningoencephalitis panel rules out many of the most common etiologies [69, 70, 71]. NAAT using CSF is the gold standard for diagnosis of viral meningoencephalitis, with both high sensitivity and specificity. NAAT detection in CSF has been the gold standard for diagnosis of HSV‐1 encephalitis for nearly three decades PCR [72].

Highly multiplexed panels have detected the presence of low prevalence etiological agents that were originally missed when more targeted standard‐of‐care testing was originally performed [73]. On the other hand, false negatives have been observed in the case of viral etiologies [70, 74, 75, 76], and singleplex NAAT for viruses may be a more sensitive option [74, 77]. The accuracy of HSV‐1 detection is questioned more broadly [78, 79], which may be concerning in immunocompromised populations where HSV‐1 false positives are more likely [80]. Results must be interpreted in context and additional testing considered based on pretest probability. While other multiplexed assays have been developed elsewhere [81, 82], only two large multiplex panels are available for detection of CNS infections in the United States [78, 83, 84]. Concordance between these two panels and traditional techniques can be low [78]. Given the uncertain performance of the current meningitis/encephalitis NAAT panels, some have recommended their use as front‐line tests, followed by smaller panel or single pathogen NAAT for confirmation [74].

Restricting molecular panel testing to CSF samples with abnormal white blood cell counts may improve diagnostic yield for non‐viral targets [85, 86], but the value of pleocytic selection is less well understood for viral etiologies [85, 86], particularly in the young [74]. As immunosuppressive treatments in the transplant setting can lead to abnormally low white blood cell counts [46] an exception should be made to the requirement for CSF pleocytosis when considering CSF NAAT.

Importantly, not all pathogens causing meningoencephalitis are included in panels, and not all molecular testing is superior to nonmolecular testing. For instance, West Nile virus (WNV) IgM from CSF has adequate diagnostic accuracy when the pretest probability of an arboviral neuroinvasive infection is high (i.e. in select geographies and seasons) [87], and detection of cryptococcal antigen from CSF performs with high accuracy [88].

Negative tests results using highly sensitive molecular tests should prompt the clinician to reevaluate the differential diagnosis and consider alternative causes, especially with failure of empiric therapy or progression of CNS infection. Agnostic cell‐free plasma metagenomic NGS may have a role if the diagnosis is uncertain after initial conventional and pathogen‐targeted molecular testing [89, 90].

2.3. GI Infections

Diarrhea is a common manifestation of GI infections, with a prevalence of 20%–50% in SOT recipients and 20%–44% in HSCT recipients [91]. There are both noninfectious and infectious causes of diarrhea. Immunosuppression‐related diarrhea is most common in the first month after transplant. Other noninfectious causes of diarrhea include graft versus host disease (GVHD), neutropenic enterocolitis, cord colitis and PTLD. Viral infections that contribute to diarrhea are seen in the 1‐to‐6‐month period after transplant. General viral causes of GI infection targeted by commercially available NAATs include adenovirus species F, astrovirus, norovirus genogroups I and II, rotavirus, and sapovirus. CMV is also a common etiological agent of GI complications in transplant recipients [91, 92], as well as a wide range of non‐F species adenovirus types [93, 94]. The non‐viral differential diagnosis of GI syndromes includes bacteria (diarrheagenic Escherichia coli [E. coli]: enteroinvasive E. coli, enterohemorrhagic E. coli, enteropathogenic E. coli, and enterotoxigenic E. coli, Shigella, Salmonella, Campylobacter, Aeromonas, Plesiomonas, and toxigenic C. difficile), parasites (Giardia, Cryptosporidium, Entamoeba), preformed toxins (Staphylococcus aureus, Bacillus cereus), and noninfectious (medications, postinfectious inflammatory bowel syndrome).

Often, the presentations of infectious and noninfectious etiologies overlap. Fever with diarrhea typically indicates an infectious cause but can also indicate GVHD. The morbidity and mortality associated with diarrheal illness in the immunocompromised host prompt evaluation is warranted.

Diagnosing infectious diarrhea should take into consideration travel history, food exposures, time of the year, recent exposures to antibiotics and/or healthcare settings and the chronicity of symptoms [95]. For example, even though norovirus infection can occur at any time of the year, outbreaks are frequent in the winter [96]. While patients infected with norovirus typically have acute symptoms of nausea, vomiting and diarrhea, chronic intermittent diarrhea and chronic shedding of the virus can occur in immunocompromised individuals [96].

Standard tests such as stool culture, antigen testing, and microscopy have been of low yield and time consuming [92]. Development of singleplex and multiplex NAAT panels has increased identification of potential microbial etiologies of GI distress [97, 98], particularly for norovirus [99, 100, 101]. Multiple NAAT panel options are commercially available [46, 102, 103, 104] and are being used with increasing frequency. They have high sensitivity and specificity, allowing for rapid diagnosis.

Without proper clinical context, the clinical validity of positive results of NAATs performed on stool samples can be questionable [105]; detection of asymptomatic colonizers can be an issue. In addition, C. difficile detection dominates within immunocompetent inpatients, making untargeted NAAT testing using panels containing C. difficile of questionable value in this population [106]. Although this might not always hold true in immunocompromised populations [97, 107]. Regardless of the source of symptoms, C. difficile colonization at the time of transplantation may be common, further confounding the value of C. difficile testing in transplant patients [108].

In addition to traditional GI pathogens, diarrhea in a transplant recipient should include evaluation for systemic CMV (i.e., plasma or whole blood viral load), especially if the patient is high risk (seropositive donor/seronegative recipient) and/or the recipient is not on CMV prophylaxis. GI infection due to CMV can present as gastritis, colitis, enteritis, and esophagitis [109]. CMV‐induced diarrhea is usually accompanied by disseminated disease with measurable CMV DNA in the blood. As such detection of CMV viremia via NAAT can help identifing CMV‐induced GI distress. However, lack of CMV viremia alone will not rule out tissue invasive GI infection. Therefore, we suggest that these tests should be interpreted based on clinical presentation and taking the above‐mentioned exposures and risk factors into consideration.

Commercially available panels, again, are not specifically tailored to transplant populations and do not account for the altered prevalences of infectious GI pathogens within this population, such as altered species of adenoviruses [110], EBV or CMV [111]. Discordance in norovirus results between NAAT GI panels exists [98, 112], including in transplant patients [113]. Thus, although evidence of clinical value continues to be generated in non‐transplant populations [114, 115, 116, 117], the true clinical utility of highly multiplex GI panels in transplant patients requires further evaluation.

Guidelines recommend diagnostic testing in the general population with severe disease, or symptoms lasting great than 7 days, and in the setting of outbreaks [112]. Guidelines specifically for transplant populations recommend multiplexed NAAT panels on all patients presenting with diarrhea, and take into consideration the persistence of some infections, such as those caused by norovirus and parasites [8].

In routine immunocompetent patients admitted with diarrhea, laboratories generally only accept orders for stool pathogen testing within the first 3 days of hospitalization, as diarrhea with hospital onset is usually due to other etiologies [46]. Another frequently applied diagnostic stewardship tool is to allow only one order of GI panel NAAT within a specified timeframe or during admission. This helps to prevent NAATs from being used inappropriately as a test of cure.

2.4. Undifferentiated Systemic Infections

Undifferentiated systemic infections in immunocompromised transplant recipients may present with nonspecific symptoms like fever, weight loss, fatigue, myalgia or functional decline. Some infections present with localizing features while others disseminate widely with pleomorphic manifestations that overlap various syndromes. Transplant infectiologists maintain a high index of suspicion and broad differentials that include fungi, bacteria, rickettsia, mycobacteria, parasites, and viruses [113]. The noninfectious differential of systemic syndromes in the transplant patient includes rejection, medication side effects, autoimmune disease, and malignancy. Before focusing on viral causes and molecular diagnostics, we must acknowledge the heightened degree of uncertainty and potential consequences with systemic infections.

Systemic undifferentiated infections are challenging, but thoughtful clinical reasoning may reveal patterns that permit rational testing. Deciding what test to order, when to order, and for whom, requires consideration of many variables. Systemic infections may appear undifferentiated but reflect the geographic distribution and unique tissue tropism of different pathogens. Similarly, lifestyle, animal exposure, travel, and dietary habits may be trivial, but is indispensable for diagnosis. Life‐threatening systemic infections, which can result in septic shock and multi‐organ system failure, warrant broad diagnostic testing ordered simultaneously, as sequential testing may delay diagnosis and essential treatment.

Viral causes of undifferentiated systemic infections include herpesviruses (CMV, EBV, VZV, HHV‐6, HHV‐7, and HHV‐8), adenoviruses and in rare cases other viruses such as measles virus, LCMV, WNV and dengue virus. JC and BK viruses can seed infection from the urinary tract, particularly in kidney transplant and HSCT recipients [114].

Among viral causes of undifferentiated systemic infections, CMV is the most important pathogen in transplant infectious diseases. Primary or reactivated CMV infections comprise a spectrum from mild flu‐like “CMV syndrome” with fever and leukopenia to severe systemic signs and organ‐invasive encephalitis, chorioretinitis, pneumonitis, hepatitis, pancreatitis, gastritis, or enterocolitis with ulceration [115]. Pretest probability for CMV is increased by compatible clinical features and pertinent risk factors including CMV serologic status, lymphocyte‐depleting or otherwise augmented immunosuppression, or discontinuation of CMV‐prophylaxis. CMV‐driven immune dysfunction is associated with coinfections with other herpesviruses, and elevated blood CMV DNA may increase their pretest probability [116, 117].

EBV is another important cause of undifferentiated systemic illness in transplant patients. Primary or reactivated EBV syndromes may be similarly pleomorphic, ranging from nonspecific febrile mononucleosis‐like illness to PTLD. PTLD may be systemic or localized and may mimic graft rejection or infiltrative disease. Pretest probability for EBV transplant infection is increased in EBV‐seronegative recipients of seropositive grafts, by CMV coinfection, and intensified immunosuppression.

Many of the pathogens associated with undifferentiated systemic illness can generate pathology through primary infection or reactivation from a latent lifecycle phase. Microbial nucleic acids are often detectable during latency, making qualitative molecular detection from blood, stool, or other secretions unhelpful in the absence of a determined focus of pathology, usually discovered via imaging and tissue biopsy (often a biopsy of the transplanted organ itself in the case of SOTRs). The inadequacy of qualitative NAAT detection may be particularly true for adenovirus infections [118, 119]. Skin and mucosal lesions can sometimes provide clues towards the underlying source of undifferentiated systemic illness [120, 121], but lack the required specificity and sensitivity for definitive diagnosis [122]. Lesion presentation may also be atypical in immunocompromised persons [120]. Quantitative NAAT (QNAT) assays capable of detecting increases in viral loads through time, or comparing viral loads to an established standard, can overcome these limitations [123, 124, 125].

Quantitative molecular detection of CMV DNA in blood has largely replaced antigenemia testing owing to advantages in throughput and the increase dynamic range for quantitation [126, 127, 128]. Quantitative EBV testing is helpful in diagnosis of conditions caused by nonneoplastic viral replication. However, the detection of high EBV viremia is not sufficient for bone fide diagnosis of PLTD, as EBV viral loads can remain consistently high post SOT without a clear risk for development of PTLD [129, 130]. Bone fide PTLD diagnosis is usually accomplished through observed morphological changes via histochemistry in combination with in situ EBV detection of nucleic acids, preferably RNA [131, 132].

Although international quantitation standards for CMV and EBV viral loads have been established to help unify test results across laboratories [133], the optimal sample type for diagnosis of CMV and EBV systemic disease is of debate [134, 135, 136, 137, 138]. Both plasma and whole blood appear to be acceptable, with whole blood slightly more sensitive, but less specific [136, 139, 140, 141, 142]. Commercial QNAT IVD products for quantification of CMV and EBV viremia are available [143, 144]. These tests are accurate and present results normalize to the WHO international standard, aiding interlaboratory and inter‐manufacture standardization [133, 145].

Systemic reactivation or primary infections with other viruses capable of latency (e.g. adenovirus, VZV, HSV‐1, HSV‐2, HHV‐6, HHV‐7, and HHV‐8, JC virus, BK virus) are considered when symptoms are unexplained by prior testing for common opportunistic infections, if the patient is considered “at‐risk” based on donor/recipient seropositivity, antiviral prophylaxis, and vaccination status. In the case of HHV‐6, persistent high levels HHV‐6 DNA detection within whole blood may indicate chromosomal integration of the HHV‐6 genome, a phenomenon that may be associated with increased risk of solid organ rejection and other infections [146].

Serial quantitative measurements of circulating viremia is again generally accurate [123, 124, 125, 147, 148, 149, 150, 151, 152]. For HSCTRs with adenovirus infections, NAAT‐based detection in the stool could precede detection in the blood, particularly in children [153, 154, 155, 156]. For BK virus and JC virus infections emanating from the urinary system, serial quantitation of viral load in the urine might be more sensitive, but less specific, with increases in viral load occurring early in reactivation [147, 148, 157].

Singleplex quantitative assays normalized to WHO standards are available for BK [158], however the universality of this standard has been questioned [159]. Commercial IVD NAAT products for JC virus detection from plasma, urine, and CSF have recently entered European markets [160, 161].

Unlike other syndromes discussed here, the inadequacy of qualitative NAAT detection precludes the use of currently available highplexed panels, as they are all quantitative in nature. Although the clinical picture may often point to a more limited number of infectious etiologies relative to other syndromes, the development of multiplexed QNAT panels may be of consideration for the future. Quantitative dualplex panels testing for HSV‐1 and HSV‐2 are commercially available IVD outside the United States.

3. Common Themes and Future Prespectives in Transplant Infectious Disease Testing and Stewardship of Molecular Methods

Although each syndrome has its own unique considerations, common themes emerge when considering diagnostic stewardship in the transplant environment.

Firstly, NAAT‐based diagnostic strategies will continue to grow in use and utility. Major advantages of NAAT technologies include rapid turnaround times and high analytical sensitivity [162], and the ability to be modified to account for novel variants with relative ease [163]. However, the high analytical sensitivity of NAAT can put into question clinical specificity when nucleic acid detection drives diagnosis without proper consideration for the clinical context. This is especially true in immunocompromised populations where microbial nucleic acids may be detected in the absence of symptoms, shed for long periods of time post convalescence, or be co‐detected along with other microbes of potential etiology.

Selecting the appropriate molecular test(s) and testing algorithms for maximum clinical and economic utility is an area of active research. Although broad testing strategies are often warranted in the transplant space, blindly testing for every possible infectious agent is ill‐advised (i.e., the “shotgun” or “kitchen sink” approach). The art of clinical medicine should always maintain its place in narrowing the diagnostic differential; the clinician is the first stage of diagnostic stewardship (Figure 1).

FIGURE 1.

Algorithm for test ordering.

The laboratory is the second stalwart of diagnostic stewardship and should help clinicians with test choices by offering algorithms based on time‐of‐year or regional prevalences, nudging, framing, and preventing repeat orders within specific time frames.

Singleplexed technologies are ideal when a particular pathogen is suspected to a higher degree, as singleplexed assays traditionally have higher accuracy relative to multiplexed assays [164, 165]. Mid‐sized NAAT panels (i.e., two to six target) are best used in situations where a limited number of pathogens are responsible for most cases of a given syndrome, in a given circumstance (e.g., “flu season”). Highplex panels many be most utile in situations where many distinct microbes can cause a particular syndrome, such as upper respiratory infections (outside “flu season” or in immunocompromised populations), undifferentiated systemic illness, signs of encephalitis/meningitis, or GI distress. However, immunocompromised persons may acquire atypical infections, and current commercially available highplex tests do not comprehensively contain all possible etiological agents within the immunocompromised hosts. As such, the full clinical utility of highplex technologies is not yet completely elucidated in transplant populations.

Microbe agnostic metagenomic NGS (mNGS) techniques may provide an additional line of testing when highplex panels fail to find a causative pathogen. NGS is increasingly being used in clinical practice, particularly in the field of oncology where it can enable tailored, and even personalized treatments. The diagnostic value of NGS strategies for infectious diseases is currently uncertain, although evidence and strategies continue to develop [89, 90, 166]. A recent systematic review in pediatric cancer patients concluded that mNGS identification of infectious disease resulted in accelerated pathogen‐directed therapy that supported antimicrobial stewardship, and enhanced outbreak surveillance [167]. mNGS can also facilitate the detection of novel microbes and/or microbe variants. Recent examples are a case of WNV meningoencephalitis detected in a renal transplant patient [168], and a case of St. Louis virus (SLV) encephalitis detected in patient with hematologic cancer [169]. Neither SLV nor WNV are detected in most highplex meningitis/encephalitis NAAT panels. Beyond clinical utility, detection of novel microbes/variants has public health and epidemic preparedness applications.

The higher resolution of NGS identification allows concomitant strain typing [170] and AMR with microbe identification [171]. Strain typing facilitates determination of causative organism through epi‐linkage [170] as well as epidemiological tracking during an outbreak [172, 173], while AMR profiling increases the effective actionability of results. NGS methods have been used to track emergence of variants in the UL56 gene of CMV porting resistance to letermovir in transplant patients [174]. Novel NGS assays capable of evaluating all CMV genes associated with antiviral resistance within a single test run are now seeing development [175]. PTLD can also be evaluated in EBV infection [176].

A fascinating potential of NGS is co‐monitoring of host factors within a single analytical run to gain a comprehensive assessment of a patient's status. In general populations, evaluating host immune signatures to determine if the source of infection is bacterial vs viral is often enough to drive relevant clinical decision‐making [177, 178]. In the transplant space, the emerging concept of “immune profile score” (IPS) involves evaluating biomarker signatures to gain insight on the total functionality of a patient's immune system. Although IPS has traditionally been performed using flow cytometry and immunoassay to evaluate T‐cell and cytokine populations, IPS scores incorporating NGS scores are in development, particularly in the field of hemo‐oncology [179, 180, 181], and many of these scoring methods are immediately applicable to the transplant space. Whole blood and tissue‐based sample types have been validated when in situ analysis is important [182]. Of particular note, NGS can profile T cell/B cell clonal expansion and evolution in a manner not available to other testing modalities there continued evaluation of the TCR/BCR regions or dominant cell populations.

Digital PCR (dPCR) is an attractive option when high resolution quantitation is warranted, such as distinguishing between chromosomally integrated HHV‐6 and active infection [183]. Novel dPCR techniques for quantitative testing are transitioning from RUO/LDT to commercialized IVD products [184], with CMV dPCR one of the first assays to see more regular clinical use. The ability of dPCR to provide absolute quantitation improves inter‐lab standardization. Although recent works have not found significant differences in accuracy compared to real‐time PCR techniques [185, 186]. Graph dysfunction—resulting from rejection, GVHD, or otherwise—can be measure through detection of donor‐derived cell free DNA (dd‐cfDNA) in blood circulation, a measure highly amenable to NGS, and particularly dPCR quantification [187].

Cost, turnaround time, and reimbursement remain important hurdles. When sequencing is performed off‐site, results typically require 3–5 days, or more, to become available to the treatment team. This may be too long for clinical relevance in transplant populations where pathogenesis can process quick. However advanced pipelines have been able to yield clinically actionable results with 3.5–7 h directly from the blood of septic patients [188, 189]. An mNGS pipeline for pathogen identification capable of turnaround of approximately 24 h has showed 40% greater pathogen detection relative to traditional culture, which led to improved diagnosis, tailored antibiotic therapy and halved the risk of early death in 7 days [190].

Costs can be prohibitive [89] and reimbursement for NGS is not currently common place in areas like the United States with private payer systems. However the cost effectiveness of NGS technologies is now being evaluated [191], and national adoption programs are being implemented in Europe [192]. Studies and programs of this nature will continue to generate data and may lead to evidence of value that could support an argument for test coverage by payors.

Finally, given the high sensitivity and wide breath of NGS testing, concerns around the potential for overcalling/overdiagnosing exist. These concerns are legitimate. As such we continue to advocate for NGS testing as a second line after more common causes are eliminated. Clinical picture and judgement used in concert with test results, as always, are essential.

4. Concluding Remarks

The transplant environment has unique challenges when evaluating for viral infectious diseases, including a broader differential of causative agents and different kinetics regarding detection of pathogens relative to symptoms when compared to fully immunocompetent persons. Advances in molecular diagnostics have helped address these challenges through the development of highplexed NAAT, dPCR and NGS anchored testing strategies. Indiscriminate, hypothesis‐devoid testing may create more questions than it answers. Moreover, even the largest commercially available syndrome‐based NAAT panels don't cover all the potential etiologies of viral infectious disease in transplant recipients. Some conditions require quantitative measurement of viral loads. As such, use of smaller panels and/or singleplexed tests may be appropriate for some clinical pictures. This complex landscape is an excellent opportunity for improved diagnostic stewardship. As new clinical utility evidence continues to emerge, we believe that responsible diagnostic stewardship in the transplant space is possible with mindful ordering of laboratory tests based on clinician observations coupled to laboratory‐driven guidance.

Sugden S., Reynoso D., Mathew R., and Loeffelholz M., “Stewardship of Molecular Diagnostics in Transplant Viral Infections.” Transplant Infectious Disease 28, no. 1 (2026): e70140. 10.1111/tid.70140

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study