Utility of Cytomegalovirus Quantitative Polymerase Chain Reaction in Tissue Biopsy for the Diagnosis of Cytomegalovirus Gastrointestinal Disease Among Solid Organ Transplant Recipients

ABSTRACT

Background

Gastrointestinal (GI) cytomegalovirus (CMV) infection is an important cause of morbidity after solid organ transplantation (SOT), and diagnosis mainly relies on histopathology of GI tissue biopsies. CMV detection by quantitative polymerase chain reaction (qPCR) on tissue biopsy is not routinely performed, but potentially holds many practical advantages.

Methods

We compared the performance of CMV qPCR on fresh GI biopsies to histopathologic identification for the detection of CMV GI disease.

Results

Sixty‐one SOT patients with GI symptoms underwent endoscopic assessment, with tissue biopsies obtained. Eleven patients had proven CMV disease by histopathologic detection. Among them, all had a positive qPCR on tissue biopsy (median of 8.7 × 10⁷ IU/mL [interquartile range {IQR} 3.1 × 10⁷, 18.2 × 10⁷]). Of the 49 patients with negative histopathology, 27 (55%) had CMV qPCR‐positive tissue biopsy specimens (median of 43 604 IU/mL [IQR 2923, 497 570]). Receiver operating characteristic analysis for optimal threshold value for CMV qPCR on tissue biopsy for diagnosis of proven CMV GI disease was 147 906 IU/mL (sensitivity 100%, specificity 80%, area under the curve = 0.975).

Conclusion

Compared to histopathologic detection, CMV qPCR on GI tissue biopsy is highly sensitive for the diagnosis of CMV GI disease in SOT patients, making it a potentially useful adjunctive diagnostic tool for rapid diagnosis in this population.

Abbreviations

AUC

area under the curve

CMV

cytomegalovirus

GI

gastrointestinal

IHC

immunohistochemical

qPCR

quantitative polymerase chain reaction;

ROC

receiver operating characteristic

SOT

solid organ transplantation.

1. Introduction

Cytomegalovirus (CMV) is an important opportunistic viral infection following solid organ transplantation (SOT), ranging from asymptomatic viremia to severe tissue‐invasive disease [1, 2, 3]. Among SOT patients, gastrointestinal (GI) involvement, such as ulcerative gastritis or colitis, is the most common presentation of CMV disease, and can present with a wide variety of symptoms, ranging from nausea and vomiting to refractory, life‐threatening colitis [4]. Rapid and accurate diagnosis of CMV GI disease is thus central to the early initiation of appropriate therapy.

Currently, the diagnosis of CMV GI disease mainly relies on clinical and endoscopic signs, as well as histopathologic analysis or viral culture on formalin‐fixed or fresh tissue biopsies [5]. While histopathology is considered the gold standard for the diagnosis of CMV GI disease, it is subject to sampling variability, slow turnaround time, and inter‐observer variability and interpretation [6, 7]. The addition of immunohistochemical staining (IHC) and in situ DNA hybridization has improved the sensitivity of GI CMV detection by histopathology, but equivocal or atypical results remain a diagnostic challenge 6, 7]. Viral culture has low sensitivity, is technically cumbersome and slow, and is thus no longer routinely performed in most medical laboratories [6]. Quantitative polymerase chain reaction (qPCR) on serum can be a useful adjunctive diagnostic modality, but its sensitivity for the detection of CMV GI disease can be as low as 50% and thus remains a suboptimal diagnostic tool [6, 8, 9].

CMV qPCR on tissue specimens has been proposed as an adjunctive test for the diagnosis of CMV GI disease. CMV qPCR on tissue specimens offers several advantages over standard histopathology, such as providing a quantifiable automated result, a rapid turnaround time as low as 24–48 h, and it is unaffected by interobserver variability [10, 11, 12]. Moreover, the quantitative nature of the test can potentially document viral burden and may also have a role in prognosis and follow‐up [10, 11, 12, 13, 14]. However, the data supporting the utility of CMV qPCR on tissue biopsy for the diagnosis of CMV GI disease is limited. Thus, it is not routinely used, nor is it included in the definition of proven or probable CMV disease in SOT patients [5, 7]. In this study, we aimed to determine the diagnostic performance of CMV qPCR on fresh tissue biopsy for the diagnosis of CMV GI disease in SOT patients.

2. Materials and Methods

2.1. Study Design

We conducted a single‐center retrospective cohort study among SOT recipients (lung, kidney, and liver) followed at the Centre Hospitalier de l'Université de Montréal in Montreal, Canada. All SOT recipients who presented with GI symptoms and underwent endoscopic investigation with tissue biopsy for histopathology and CMV qPCR between 2017 and 2020 were included. Patients for whom both methods of detection were not performed were excluded. A retrospective chart review of electronic patient records was done to obtain demographic, clinical, laboratory, endoscopic, histopathologic, and therapeutic data.

2.2. Definitions

Patients were classified as having proven, probable, possible, or no CMV GI disease, based on the updated definitions of CMV disease in transplant patients [13, 15]. Proven disease was defined by the presence of upper or lower GI symptoms, macroscopic lesions on endoscopy, and evidence of CMV in targeted GI tissue samples (obtained through histopathologic analysis with or without IHC, DNA hybridization, or viral culture). Probable disease was defined by the presence of GI symptoms and evidence of CMV in targeted GI tissue samples (obtained through histopathology, IHC, DNA hybridization, or viral culture) without endoscopic lesions. Possible disease was defined by the presence of GI symptoms combined with the detection of CMV by qPCR in blood or GI tissue samples, regardless of endoscopic findings [13, 15]. To further define the diagnostic certainty of patients with possible CMV GI disease, a 6–8‐week assessment for clinical, endoscopy, and virology follow‐up was conducted. These patients were further categorized as clinically possible CMV if they met the three following criteria: 1) received CMV directed therapy, 2) presented clinical and endoscopic improvement at 4–8 weeks follow‐up, and 3) no alternative diagnosis was made (co‐diagnosis of CMV infection was retained if considered plausible by the treating physician). Exclusion criteria included patients who did not have endoscopic follow‐up to assess for response to therapy. Patients with a previous diagnosis of CMV with ongoing CMV‐directed therapy were excluded (initial diagnostic biopsy episode not included due to missing CMV PCR on tissue biopsy).

2.3. Analyses

2.3.1. Histopathologic Detection of CMV Infection

Sections of formalin‐fixed endoscopic biopsies were stained with hematoxylin and eosin (H&E) in search of active inflammation and viral cytopathic effect or inclusions. Immunohistochemical staining of 3 µm‐thick paraffin sections was performed using anti‐CMV antibody on a BenchMark Ultra automated stainer (Dako CCH2+DDG9 clone 1:100 dilution).

2.3.2. CMV qPCR in Blood Sample

Blood samples were collected in EDTA tubes for qPCR analysis. CMV DNA loads were measured in plasma samples using real‐time qPCR based on the COBAS AmpliPrep/COBAS TaqMan CMV Test manufacturer's instructions. Results were reported to clinicians in copies/mL and IU/mL.

2.3.3. CMV qPCR on Fresh Tissue Biopsy Samples

CMV qPCR on fresh tissue biopsy specimens was analyzed at the Centre Hospitalier Universitaire Sainte‐Justine virology laboratory, using a previously validated in‐house method that is currently applied to clinical samples.  GI tissue biopsies were collected and transported in sterile saline. Biopsy sizes varied between 3 and 4 mm and were limited to the mucosa with practically no muscle, thereby conferring some specimen harmonisation. Of note, a one‐log change on DNA load would require a biopsy size 10 times larger (30–40 mm); thus, the variability in biopsy size is unlikely to significantly impact DNA load. Nevertheless, biopsies could vary in terms of lymphocyte count in the specimen, inflammation, and necrosis, all of which may impact each sample's DNA content. Primers and TaqMan MGB Hydrolysis probes were designed for the UL83 gene of CMV AD169 strain, with an internal control, the Arabidopsis thaliana CSG4 gene, using Primer Express 2.0. Biopsies were manually ground with 1 mL of viral transport media, centrifuged, and extracted (supernatant input volume: 200 µL) using the Maxwell 16 instrument in combination with the Viral Total Nucleic Acid Purification Kit. Nucleic acids were eluted in 200 µL of PCR‐grade water for both methods. qPCR reactions are set up as a duplex using 5 µL of extracted nucleic acid in a total volume of 30 µL mastermix containing 15 µL of 2X QuantiTect Multiplex PCR NoROX Kit by Qiagen plus 50 nM ROX reference dye by Invitrogen. The analytical limit of detection of this method is 200 copies/mL. Standard curves were performed in triplicate on every run, along with positive and negative controls. The assay is routinely validated using the CAP external control for CMV DNA loads. Results were reported in copies/mL and IU/mL. Internal validation has shown that viral culture correlated with a viral load of between 1000 and 5000 copies/mL using the in‐house quantitative assay. The cost of CMV qPCR on tissue biopsy was 35 CAD per analysis.

2.3.4. Statistical Analysis

Descriptive statistics were used to characterize the study population. Values were expressed as the mean (with standard deviations) or median (with interquartile ranges) for continuous variables. A receiver operating characteristic (ROC) analysis was performed for tissue biopsy CMV qPCR using the definition of proven and probable CMV GI disease as the gold standard. Sensitivity analyses using different CMV DNA load values were performed to find the optimal diagnostic threshold. All analyses were performed using the STATA statistical software (version 17.0).

3. Results

Among the 728 SOT patients transplanted between 2017 and 2020, 8.4% (61/728) underwent endoscopic examination with biopsy to investigate GI symptoms. Population characteristics are shown in Table 1. The median age was 57 years, and 51% of participants were male. There were eight (13%) kidney, 41 (67%) liver, and 10 (16%) lung transplant recipients, as well as two dual transplant recipients. Among the 61 patients, 15 (25%) were seropositive D+/R+. while 24 (39%) were CMV sero‐mismatched D+/R‐ and 15 (25%) were sero‐mismatched D‐/R+. The median time of onset of GI symptoms was 196 days from the transplant date (IQR 79, 377). Twenty‐six (43%) patients presented with upper GI symptoms, while 19 (31%) presented with lower GI symptoms, and 16 (26%) presented with both.

TABLE 1.

Baseline patient population characteristics.

Characteristics	N (%)
All	61 (100)
Age, years (mean)	51
Age, years (median)	57
Sex
Male	31 (51)
Female	30 (49)
SOT type
Kidney	8 (13)
Liver	41 (67)
Lung	10 (16)
Kidney + Lung	1 (2)
Kidney + Liver	1 (2)
CMV serostatus at induction
D+/R+	15 (25)
D+/R‐	24 (39)
D‐/R+	15 (25)
D‐/R‐	7 (11)
Induction agent
Methylprednisolone alone	17 (28)
Basiliximab alone	44 (72)
Immunosuppressive therapy
Tacrolimus + Mycophenolate	34 (56)
Tacrolimus + Azathioprine	23 (38)
Other	4 (6)
Treatment for rejection in the first 3 months	13 (21)
Receipt of CMV prophylaxis in the 6 months post‐transplant	30 (49)
Median time of onset post‐transplant (days)	196 (IQR 79, 385)

Note: CMV: Cytomegalovirus, D: Donor, R: Recipient, SOT: solid organ transplantation.

Among the 61 patients who underwent investigation of GI symptoms, 18.0% (11/61) had proven CMV GI disease, 44.3% (27/61) had possible CMV GI disease, and 37.7% (23/61) had no CMV GI disease (Table 2). There were no cases of probable disease. Among 11 patients with proven CMV disease, two were positive by H&E and nine by CMV‐specific IHC. All had abnormal endoscopy and therefore met criteria for proven GI disease (Table 2). Forty‐nine biopsies were negative for CMV disease by histopathologic analysis. Among these, 54% (27/50) were considered to have possible CMV disease because of GI symptoms associated with a positive CMV qPCR on tissue biopsy. Fifty‐five percent (15/27) of these patients had abnormal endoscopic findings (Table 2). Finally, 23 patients had no evidence of CMV on tissue biopsy by either histopathology or CMV qPCR and were classified as not having CMV GI disease. Their final diagnoses included inflammatory bowel disease (IBD) (n = 3), Clostridioides difficile colitis (n = 1), unspecified viral syndrome (n = 1), diverticulitis (n = 1), Candida oesophagitis (n = 1), peptic ulcer disease (n = 3), drug side effects (n = 3), anastomotic bowel obstruction (n = 1), and inflammatory polyps (n = 1). Nine remained with an unclear diagnosis.

TABLE 2.

Clinical characteristics and description of cytomegalovirus gastrointestinal (CMV GI) disease by disease category.

Characteristics	Proven CMV Disease n (%)	Possible CMV Disease n (%)	No CMV disease n (%)
Total	11 (100)	27 (100)	23 (100)
Disease site
Upper GI	3 (27)	9 (33)	14 (61)
Lower GI	5 (46)	6 (22)	8 (35)
Both	3 (27)	12 (45)	1 (4)
CMV serostatus at induction
D+/R+	3 (27)	8 (30)	4 (18)
D+/R‐	7 (64)	11 (44)	5 (22)
D‐/R+	1 (9)	7 (26)	7 (30)
D‐/R‐	0 (0)	0 (0)	7 (30)
Endoscopic findings
Normal	1 (9)	12 (44)	7 (30)
Erythema	4 (36)	8 (30)	9 (39)
Ulcer (s)	3 (27)	6 (22)	2 (9)
Both	3 (27)	1 (4)	5 (22)
Positive CMV histopathology on GI tissue biopsy	11 (100)	0 (0)	0 (0)
Positive CMV immunohistochemistry on GI tissue biopsy	10 (91)	0 (0)	0 (0)
Positive CMV qPCR on GI tissue biopsy	11 (100)	27 (100)	0 (0)
Median (IU/mL)	8.7 x107	43 604	0
Positive CMV qPCR in serum	11 (100)	11 (41)	2 (9)
Median (IU/mL)	49 208	0	0

Note: CMV: cytomegalovirus; GI: gastrointestinal; qPCR: quantitative polymerase chain reaction.

Twenty‐seven patients classified as possible CMV GI disease were further assessed to determine their clinical outcome at 6–8 weeks after initiation of therapy. Of these, one patient was excluded because the current episode was a follow‐up biopsy of a case of proven CMV GI disease with ongoing therapy, and two were excluded due to lack of follow‐up endoscopic assessment. At 8 weeks follow‐up, 6 patients were diagnosed with an alternative diagnosis as a cause to explain their initial symptoms: gastroparesis (n = 3), polymedication (n = 1), esophageal candidiasis (n = 1), and ulcerative colitis (n = 1). As such, the remaining 18 patients were classified as clinically possible CMV GI disease. Their clinical characteristics are shown in Table 3.

TABLE 3.

Clinical characteristics of patients with clinically possible cytomegalovirus gastrointestinal (CMV GI) disease.

Case	Organ	Serology	Timing post‐transplant (days)	Symptoms	Endoscopic findings	CMV PCR on tissue biopsy (IU/mL)	CMV viremia (IU/mL)	Treated	Symptoms at follow‐up	Endoscopy at follow‐up
1	Kidney	D+/R‐	199	Epigastric pain	Gastric ulcer	32 324 143	57 135	Yes	Resolved	Resolved
2	Kidney	D‐/R+	169	Anal pain	Rectal ulcer	5489	1863	Yes	Resolved	Improved
3	Lung	D+/R‐	271	Epigastric pain	Gastric erosions	2136	0	Yes	Resolved	Resolved
4	Lung	D+/R‐	316	Epigastric pain	Gastric ulcer + erythema	245 203	12 516	Yes	Resolved	Improved
5	Liver	D+/R+	44	Abdominal pain	Normal mucosa	60 190	2467	Yes	Resolved	Normal
6	Liver	D+/R‐	425	Follow‐up; No symptoms	Normal mucosa	43 604	0	Yes	Resolved	Normal
7	Liver	D+/R‐	340	Nausea	Normal mucosa	10 221	0	Yes	Resolved	Normal
8	Liver	D+/R+	306	Anemia	Gastric punctiform erosions	1411	621	Yes	Resolved	Resolved
9	Liver	D+/R‐	200	Nausea	Gastric erosions	54 895	0	Yes	Resolved	Resolved
10	Liver	D+/R‐	34	Nausea	Normal mucosa	446 198	3085	Yes	Resolved	Normal
11	Liver	D‐/R+	56	Viremia screening	Antral erosions + ulcer	1 257 555	0	Yes	NA	Resolved
12	Liver	D‐/R+	39	Epigastric pain	Gastric erosions	30 166 645	0	Yes	Resolved	Resolved
13	Liver	D‐/R+	79	Diarrhea + abdominal pain	Rectal ulcerations	691 078	0	Yes	Resolved	Resolved
14	Liver	D+/R+	418	Epigastric pain	Normal mucosa	223 628	949	Yes	Resolved	Normal
15	Liver	D+/R‐	285	Diarrhea	Normal mucosa	890 280	1 081 139	Yes	Resolved	Normal
16	Liver	D‐/R+	393	Epigastric pain	Gastric patchy erythema	548 942	0	Yes	Resolved	Resolved
17	Kidney/Liver	D+/R‐	541	Nausea	GAVE	1 620 045	1871	Yes	Improved	GAVE
18	Lung	D+/R+	597	Diarrhea	Colonic ulcers	141 100	0	Yes	Improved	Resolved

3.1. Performance of CMV qPCR on Tissue Biopsy

The median value of tissue biopsy CMV qPCR among patients with proven, possible and clinically possible CMV GI disease was 8.7 × 10⁷ IU/mL (IQR 3.1 × 10⁷, 18.2 × 10⁷), 43 604 IU /mL (IQR 2923, 497 570) and 234 415 IU /mL (IQR 46 426, 840 479), respectively (Figure 1). As such, the sensitivity and specificity of CMV qPCR in tissue biopsy for proven CMV GI disease were 100% (11/11) and 46% (23/50), respectively. When combining proven (n = 11) and clinically possible CMV GI disease (n = 18) together, the sensitivity and specificity of the tissue biopsy PCR were 100% (29/29) and 72% (23/32), respectively.

FIGURE 1.

Distribution of cytomegalovirus (CMV) DNA load in gastrointestinal (GI) tissue biopsies in proven, clinically possible, and possible CMV GI disease.

ROC analysis determined the optimal threshold value to be 147 906 IU/mL for the diagnosis of proven CMV GI disease (Figure 2A). Using this cut‐off threshold value for tissue biopsy CMV qPCR, the sensitivity and specificity for the diagnosis of proven CMV GI disease were 100% and 80%, respectively (area under the curve [AUC] = 0.975). For proven and clinically possible CMV GI disease combined, ROC analysis determined the optimal threshold value to be 29 021 IU/mL (Figure 2B). Using this cut‐off threshold value for tissue CMV qPCR, the sensitivity and specificity for diagnosis of proven and clinically possible CMV GI disease were 86% and 100%, respectively (AUC = 0.990).

FIGURE 2.

(A) Receiver operating characteristic (ROC) curve analysis for cytomegalovirus (CMV) quantitative polymerase chain reaction (qPCR) on tissue biopsy for diagnosis of proven CMV gastrointestinal (GI) disease. (B) ROC curve analysis for CMV qPCR on tissue biopsy for diagnosis of proven and clinically possible CMV GI disease.

3.2. Performance of Serum CMV DNA Load

All patients with proven CMV GI disease had a positive serum CMV DNA load at the time of diagnostic endoscopy, with a median value of 49 208 IU/mL (IQR 3894, 255 206). Among the 27 patients with possible CMV GI disease, 40.7% (11/27) had positive serum CMV DNA load while exhibiting GI symptoms, with a median value of 0 IU/mL (IQR 0, 1866). Among patients with clinically possible CMV GI disease, 50% (9/18) had detectable serum CMV DNA load with a median value of 310 IU/mL (IQR 0, 2317). As such, the sensitivity and specificity of serum CMV DNA load for proven CMV GI disease were 100% and 75.5%, respectively; the sensitivity and specificity for proven and clinically possible CMV GI disease were 72% and 93%, respectively. The overlapping performance of histopathology, tissue biopsy, CMV qPCR, and serum CMV DNA load with regard to CMV GI disease is illustrated in Table S1 and Figures S1 and S2.

3.3. Test Performance According to CMV Serology

Among the 38 patients with proven or possible CMV GI disease, 50% (19/38) were sero‐mismatch: 63% (7/11) and 44% (12/27) among proven and possible CMV GI disease, and 154 713 IU/mL (IQR 4599, 974 316) among sero‐mismatch patients and seropositive patients, respectively (p = 0.02). Median serum CMV DNA load was 1 871 IU/mL (IQR 0, 53 171) and 620 IU/mL (IQR 0, 2776) among sero‐mismatch and seropositive patients, respectively (p = 0.11).

3.4. Turnaround Time

Histopathology reports were available at a median time of 8.5 days (ranging from 3 to 15 days). Tissue biopsy CMV qPCR results were available at a median time of 4 days (ranging from 2 to 11 days).

4. Discussion

This study demonstrated that CMV qPCR on fresh GI tissue biopsy is highly sensitive for the diagnosis of GI CMV disease in SOT patients. Using a cut‐off value of 147 906 IU/mL, the sensitivity and specificity of tissue biopsy CMV qPCR for the diagnosis of proven CMV GI disease in SOT were 100% and 80%, respectively. Moreover, tissue biopsy CMV qPCR provided faster turnaround time for results compared to histopathologic analysis, thereby allowing for possible earlier diagnosis and treatment initiation.

Our results are in line with previous studies in which the sensitivity of CMV qPCR on tissue biopsy ranged from 88% to 100% [12, 13, 15, 16, 17, 18]. The high sensitivity of the tissue qPCR is, however, offset by the lower specificity, ranging between 72% and 100%, and has contributed to the slow uptake of this test in clinical practice. Indeed, the main drawback of using CMV qPCR on tissue biopsy is the uncertainty of its significance in the absence of CMV‐positive histopathology, and thereby a lack of sufficient specificity. Importantly, the definition of disease is crucial in the determination of test performance (sensitivity and specificity). In previous studies on tissue CMV qPCR, its performance has been compared to histopathology as the gold standard. However, there is debate about whether histopathology is an imperfect gold standard. For example, in a cohort of 101 SOT patients with biopsy‐confirmed CMV GI disease (histopathology, IHC, or viral culture), 26% of patients were diagnosed by viral culture but with negative histopathology [6]. Interestingly, Ganzenmueller et al. also assessed the performance of tissue biopsy PCR compared to traditional histopathology as well as clinical CMV disease (signs and symptoms in the absence of another plausible etiology and/or the subsequent improvement after antiviral treatment); specificity was 73% when compared to histopathology but increased to 85% when compared to clinical CMV disease [9]. Thus, while histopathology is currently the reference gold standard, it may be imperfect and not accurately capture all patients with CMV GI disease, and thus negatively bias other tests toward lower specificity. In this present study, we defined the gold standard of disease based on histopathology to provide the most conservative estimate of sensitivity and specificity.

Based on current definitions, the detection of CMV qPCR in tissue biopsies alone is insufficient for the diagnosis of CMV GI disease [13, 15]. In the present study, 27 patients with negative histopathology had detectable CMV DNA load by qPCR on tissue biopsy, and the significance of these discordant cases remains unclear. It is possible that the highly sensitive qPCR may detect latent CMV or viral shedding without disease, leading to false‐positive results. One possible theory to explain this phenomenon is that latent CMV, which is known to reside predominantly in leukocytes, could theoretically be detected in colon biopsies with extremely robust inflammatory infiltrates, without necessarily infecting the colonic tissue itself [19]. On the other hand, there is evidence to suggest that these discordant results may represent early CMV GI disease. In a study by McCoy et al., CMV was detected by qPCR on GI tissue biopsies in 60% of histopathology‐negative biopsies from patients who underwent repeat biopsy within 7 days and subsequently developed histopathology‐positive biopsies. Moreover, tissue biopsy CMV qPCR was positive in 78% of IHC equivocal biopsies, 36% (5/14) of which were positive by viral culture [18]. Similarly, in another study by Suarez‐Ledo et al., three out of eight patients with histopathology negative/CMV qPCR positive tissue biopsies underwent a repeat endoscopic procedure seven days later; the samples taken then were CMV‐IHC positive [12]. Our study results demonstrated a trend of higher tissue CMV DNA load in proven versus possible GI CMV disease, suggesting that possible CMV disease could represent an earlier stage of disease. To further explore the significance of CMV PCR among patients with possible CMV GI disease, we conducted a follow‐up assessment at 6–8 weeks of antiviral therapy to determine the clinical likelihood of CMV GI disease. Among 27 patients with negative histopathology but positive tissue biopsy CMV qPCR, 66% (18/27) received antiviral therapy with improvement or resolution of GI symptoms, endoscopic findings, and virologic response at 6–8 weeks follow‐up, further supporting the diagnosis of clinical CMV GI disease. As such, tissue CMV qPCR may support the early diagnosis of end‐organ disease or reactivation and increase the sensitivity for detection of CMV GI disease before histopathologic changes are detected.

While test performance characteristics play an important role in determining the clinical utility of a test, the predictive value of a test, which accounts for the disease prevalence in a given population, is an important factor. Previous studies assessing tissue biopsy CMV qPCR included patients with diverse underlying diseases such as human immunodeficiency (HIV), cirrhosis, diabetes, IBD, hematopoietic stem cell transplant recipients (HSCTs), and SOT recipients [9, 12, 18, 20, 21, 22]. CMV GI disease in these patient populations presents differently in terms of incidence, concomitant GI disease, disease presentation and progression, and host response to therapy, making them difficult to compare. This is the first study to include only SOT patients, a population with a high incidence of CMV disease and less likely to develop confounding GI conditions, such as ulcerative colitis exacerbation or graft versus host disease, which can negatively impact diagnostic accuracy.

In our study, the sensitivity of serum CMV DNA load for the diagnosis of proven CMV GI disease was 100%. However, among patients with possible CMV GI disease, only 38% had detectable viremia. In the current CMV definition consensus [13], possible CMV disease is defined as a patient with diarrhea or upper GI symptoms with increasing serum CMV DNAemia.  Using the current CMV disease definition, nine patients in our study would have been considered as NOT having CMV disease. In clinical practice, these patients would not have undergone endoscopic investigation and would not receive CMV‐directed therapy. Our results suggest that serum viral load is not sensitive for CMV GI disease and that the current definition may not appropriately identify patients with CMV GI disease. Among 18 patients with clinically possible CMV, 50% (9/18) had undetectable CMV DNAemia. Nevertheless, there was enough clinical suspicion of CMV GI disease to undergo endoscopic evaluation. Positivity of tissue biopsy CMV PCR reinforced the clinical suspicion. Finally, these patients responded to CMV‐directed treatment, further confirming the clinical diagnosis of possible CMV disease. As such, in a patient with high pre‐test probability, a negative CMV DNAemia should not exclude the possibility of CMV GI disease, and patients should undergo further investigation with endoscopy and GI tissue biopsy PCR. Serum CMV qPCR was generally lower than tissue biopsy CMV qPCR, supporting the adage that CMV is a compartmental disease with mostly localized viral replication in affected organs. Previous studies have reported sensitivity as low as 50% for serum CMV qPCR for GI disease, particularly among seropositive SOT patients with late onset disease (>6 months after transplantation) [6]. Taken together, these findings highlight the limited value of CMV blood tests for the prediction and diagnosis of CMV invasive disease.

Our study has several limitations that should be acknowledged. First, this was a single‐center study, and the results may not be generalizable to other centers. The small sample size may also limit the accuracy of the test performance assessment. Next, the exclusion of patients who were treated empirically for CMV syndrome without GI biopsy confirmation may lead to selection of more severely ill patients and lead to selection bias. Moreover, tissue sampling may vary between endoscopists, resulting in varying sampling sites (ulcerated lesion, peri‐ulceration, erythematous patch, or normal mucosa) and biopsy size, which in turn may impact the actual quantification of CMV DNA load. Although it is not feasible to standardize tissue sampling, normalization of CMV qPCR results over biopsy weight or by beta globulin concentration would improve standardization of results. This, however, was not done at our center. Finally, unlike commercial serum CMV DNA load assays, in‐house tissue biopsy CMV qPCR lacks standardization. The cutoffs we have reported in this study, therefore, also cannot apply to other institutions, and each site would have to establish its own cutoff values to apply this technique. Calibration with the WHO International Standard was performed to improve standardization and reproducibility; however, it remains insufficient for interlaboratory comparisons due to extraction protocols and amplicon lengths [17, 23, 24]. However, all standardisation techniques have their pros and cons and can be difficult to incorporate into routine practice. Serum CMV NAAT testing itself has a wide interlaboratory variability; extraction efficiency on biopsies may also vary significantly between instruments. As such, routine standardization practices of such an assay can be cumbersome. Despite this limitation, we believe that our results highlight the potential utility of quantitative biopsy testing and lead to the development of a standardized commercial assay and improved processing of biopsy testing. More data trying to correlate visual aspects of mucosa, histopathology, and quantitative PCR are needed.

In conclusion, our data suggest that CMV qPCR on fresh GI tissue biopsies is a highly sensitive diagnostic tool for the rapid diagnosis of CMV GI disease among SOT recipients. By providing rapid and quantifiable results, this assay may facilitate early diagnosis and timely initiation of antiviral therapy.

Author Contributions

Lucila Baldassarre: methodology, data curation, investigation, formal analysis, writing – original draft, and writing – review and editing. Koray Demir: methodology, data curation, investigation, formal analysis, writing – original draft, and writing – review and editing. Camille Pelletier Vernooy: conceptualization, methodology, data curation, and investigation. Guillaume Butler Laporte: conceptualization, formal analysis, and writing – review and editing. Geneviève Huard: conceptualization and writing – review and editing. Catherine Girardin: conceptualization and writing – review and editing. Katarzyna Orlicka: conceptualization, methodology, investigation, and writing – review and editing. Bich Ngoc Nguyen: conceptualization, methodology, investigation, and writing – review and editing. Christian Renaud: conceptualization, methodology, investigation, and writing – review and editing. Charles Poirier: conceptualization, methodology, and writing – review and editing. Julie Morisset: conceptualization, methodology, and writing – review and editing. Me‐Linh Luong: conceptualization, methodology, data curation, investigation, validation, formal analysis, supervision, visualization, project administration, writing – original draft, and writing – review and editing.

Ethics Statement

The study was approved by the Research Ethics Board of our institution.

Consent

Permission was obtained to review and publish information from patient records. Informed consent was not required due to the retrospective nature of the study. All patients were assigned a study number for de‐identification. Data confidentiality was maintained at all times, and only members of the research team had access to patient data.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Figure 1: Positivity rates of pathology, tissue CMV qPCR, and serum CMV DNAemia among patients with at least one positive test (38 patients). Legend: CMV: cytomegalovirus, qPCR: quantitative polymerase chain reaction.

Supporting Figure 2: Positivity rates of pathology, tissue CMV qPCR and serum CMV DNAemia among patients with at least one positive test (38 patients).

Supporting Table S1: Clinical characteristics of patients with proven CMV GI disease.

Baldassarre L., Demir K., Vernooy C. P., et al. “Utility of Cytomegalovirus Quantitative Polymerase Chain Reaction in Tissue Biopsy for the Diagnosis of Cytomegalovirus Gastrointestinal Disease Among Solid Organ Transplant Recipients.” Transplant Infectious Disease 27, no. 4 (2025): 27, e70082. 10.1111/tid.70082