Objectives: The human cytomegalovirus is a notorious pathogen in the pediatric transplant setting. Although studies on factors in complicity with cytomegalovirus infection abound, the roles of age, sex, allogeneic hematopoietic stem cell transplant modality, and type of underlying disease (malignant vs nonmalignant) with regard to cytomegalovirus infection and viral load in children are seldom explored. Our aim was to examine the significance of these factors on cytomegalovirus infection and viral load in Serbian pediatric recipients of allogeneic hematopoietic stem cell transplant.
Materials and Methods: Thirty-two pediatric recipients of allogeneic hematopoietic stem cell transplant to treat various malignant and nonmalignant disorders were prospectively monitored for cytomegalovirus infection. The real-time quantitative polymerase chain reaction was used for pathogen detection and quantitation. Demographic and virologic parameters were statistically analyzed with SPSS statistics software (version 20).
Results: Cytomegalovirus DNA was detected in 23 patients (71.9%). Infection occurred significantly more often (P = .015) in patients with haploidentical donors. The opposite was noted for matched sibling grafts (P = .006). Viral load was higher in female patients (P = .041) and children with malignant diseases (P = .019). There was no significant relationship between viral infection or load and medical complications.
Conclusions: Transplant recipients presented with a high incidence of cytomegalovirus viremia. The modality of allogeneic hematopoietic stem cell transplant was associated with the frequency of cytomegalovirus infection. Age, sex, type of underlying disease, and medically relevant events were not conducive to occurrences of viremia. Notably, we observed substantial viral loads in female patients and patients with neoplastic diseases. Studies comprising larger populations are needed to better understand these results.
Key words : Haploidentical, Herpesviridae, Pediatric, Viral load
Cytomegalovirus (CMV) is a double-stranded DNA virus that belongs to the Herpesviridae family. It is a ubiquitous agent of infection, and there is a broad range seroprevalence, from 40% to 95%.1 After initial contact, the virus establishes persistent infection within its host. Whereas primary infection may occasionally manifest as a mononucleosis-like syndrome, as fever, or as hepatitis, it is usually asymptomatic in the general population. This is in stark contrast with CMV infection in immunosup-pressed patients, for whom it may have a fatal outcome.
Viral infections are a major concern in recipients of allogeneic hematopoietic stem cell transplant (allo-HSCT). Cytomegalovirus is a common pathogen in the pediatric HSCT setting2 and may cause a panoply of life-threatening conditions. Immunosuppression that follows allo-HSCT is often associated with reactivation of CMV and other herpesviruses3 and may result in substantial morbidity and mortality. Cytomegalovirus infection in children typically appears within the first 100 days after the allo-HSCT.4 Although CMV infection can be asymptomatic, a wide range of salient diseases occur such as pneumonia,5 hepatitis, hemorrhagic cystitis, bone marrow suppression, and graft failure.3 This pathogen is also implicated in the development of posttransplant lymphoproliferative disorder.6,7 Systematic virologic surveillance is crucial for pediatric recipients after allo-HSCT. In this context, real-time quantitative polymerase chain reaction (PCR) has become a principal laboratory diagnostic technique to effectively guide effective therapy.
The etiology of CMV infection in a pediatric population with impaired immune response is multifaceted, and a variety of factors influence the outbreaks of viremia. Most of these factors have been reported in both adult and pediatric patients. Regardless, the roles of age, sex, allo-HSCT modality, and the underlying disease (malignant vs nonmalignant) with regard to CMV infection and viral load remain seldom explored. To the best of our knowledge, no similar research has been hitherto carried out in a Serbian patient population.
We initiated a prospective study with real-time PCR monitoring to investigate factors related to the emergence of CMV infection and the extent of viral load in a single-center group of pediatric allo-HSCT recipients.
Materials and Methods
The study was approved by the ethical committee of the Medical Faculty, University of Belgrade, as well as the ethical committee of the Mother and Child Health Care Institute of Serbia Dr Vukan ?upi? from which the samples were obtained. Informed consent was obtained for all patients. Protocols conformed to the ethical guidelines of the 1975 Helsinki Declaration.
Recommendations by the European Society for Blood and Marrow Transplantation suggest 22 transplants per 1 million population. There are between 22 and 26 allo-HSCT procedures performed at our center each year, and the total of 32 patients in our study represents a respectable population sample in the study period.
Thirty-two patients (18 male, 14 female) were children and adolescents (mean age, 8.57 years; range, 2-21 years) treated at the Mother and Child Health Care Institute of Serbia Dr Vukan ?upi? in Belgrade, Serbia. The patients received allo-HSCT from human leukocyte antigen (HLA)-matched sibling donors (MSD), haploidentical donors, HLA-matched unrelated donors (MUD), and a cord-blood stem cell donor. The conditioning regimens were chosen according to existing protocols or guidelines by the European Society for Blood and Marrow Transplantation workgroups.8 Before transplant, all patients were serologically tested for CMV and were found to be positive for immunoglobulin G and negative for immunoglobulin M. Prophylactic acyclovir was obligatory for all patients. Therapeutic doses of acyclovir were introduced if viral reactivation was detected, namely, a positive test for herpes simplex virus DNAemia. Ganciclovir and valganciclovir were administered solely in cases where CMV infection was detected via real-time PCR. Clinical characteristics of the study participants are summarized in Table 1.
Patients were prospectively screened for active viral infection. The initial blood sample was taken as part of the pretransplant patient workup, with follow-up sampling typically performed once a week after allo-HSCT. Peripheral blood samples were drawn from January 2017 to October 2018 into vacutainers with either sodium citrate or EDTA. After the samples were centrifuged at room temperature, plasma was separated from whole blood, subjected to DNA extraction, and subsequently analyzed at the Virology Laboratory of the Institute of Microbiology and Immunology, Faculty of Medicine, University of Belgrade, Serbia. The criterion for diagnosis of CMV infection was the detection of viral DNA in blood via real-time PCR.9,10
Viral nucleic acid was isolated from 200 ?L of plasma with a QIAamp DNA Blood Mini Kit (Qiagen). Detection and quantitation of the CMV DNA were performed with a TaqMan real-time PCR assay on an Applied Biosystems 7500 Real-Time PCR system as previously described.11 Briefly, primers were modeled for amplification of a CMV immediate-early region gene segment, with sequences 5?-CGC-TCA-CAT-GCA-AGA-GTT-AAT-CTT-C-3? and 5?-AAC-TCG-GTA-AGT-CTG-TTG-ACA-TGT-ATG-3?. The TaqMan probe was fluorescence labeled at 5? and 3? ends with 6-carboxyfluorescein (FAM) and 6-carboxytetramethyl-rhodamide (TAMRA) dyes, respectively. The probe sequence was 5?-(FAM)-CTC-TAT-CTG-ACA-TAC-ACA-AGT-AAA-TCC-ACG-TCC-CA-(TAMRA)-3?. The FAM functioned as the reporter dye. The standards used for quantitation were plasmids that included viral amplicons for the CMV immediate-early antigen region (Clonit). The standard viral load range was 103 to 106 copies/?L. The quantity was determined according to standard curve calculations, and extrapolated values were recorded as number of copies per microliter. Values were then multiplied by 150 to correct for sample dilution and convert to copies per milliliter of sample material. We used RNase-free water in place of the sample for negative controls. All PCR mixtures were performed in a final volume of 25 ?L.
Statistical calculations were performed with SPSS statistics software (version 20). The chi-square test was used for categorical variables. Scalar and nominal data were compared via the Wilcoxon-Mann-Whitney test. Associations between numeric variables, patient age, and viral load were analyzed by Pearson correlation. P < .05 was considered significant.
Cytomegalovirus DNA was detected in 23 patients (71.9%). Clinical characteristics of patients with CMV infection are presented in Table 1. Infections developed significantly more often (P = .015) in recipients (n = 9, 39.1%) of haploidentical HSCT (haplo-HSCT). We observed a CMV reactivation in 8 recipients of MUD transplants (34.8%) and in 5 recipients (21.7%) of HLA MSD transplants. Also, we observed virus reactivation in a single recipient (4.3%) of cord-blood HSCT. Cytomegalovirus reactivation was less frequent (P = .006) in patients who received the MSD graft versus other types of transplant.
Patients with malignancies developed CMV more often than did patients with nonmalignant underlying conditions, although the difference was not statistically significant. However, significantly higher (P = .019) viral loads were noted in subjects with malignancies.
The mean viral load was 277?329 copies/mL (range, 1 × 102 to 2.9 × 106 copies). Viral loads were lower in recipients of haplo-HSCT and MUD HSCT, whereas the opposite was noted in recipients of MSD transplants. These variations could not be distinguished statistically.
Although the difference between male and female patients in terms of infection was not significant, CMV positivity was more common in male patients. In contrast, virus copy numbers were significantly higher in female patients (P = .041). We observed that CMV DNAemia was predominant in older patients, but there was no correlation between age or viral load and CMV infection.
Cytomegalovirus infection was not significantly related to early posttransplant complications such as mucositis, febrile neutropenia, sepsis, veno-occlusive disease, pneumonia, and hemorrhagic cystitis. Patients with these complications did not carry a significantly higher viral load.
Although CMV incidence and the graft-versus-host disease (GVHD) did not covary significantly, the CMV events were more frequently detected in patients with GVHD versus patients without GVHD. Albeit higher in subjects with GVHD, the difference in viral loads between these 2 groups was not statistically significant.
This is the first prospective study on Serbian pediatric recipients of allogeneic transplants regularly monitored for CMV infection. A wide spectrum of CMV incidents has been reported in published studies of patient cohorts from geographically diverse parts of the world.12-16 In our cohort, we observed CMV viremia in approximately 72% of the patients. This agrees with the study of Bonon and colleagues,12 who reported an incidence rate of roughly 68% in Brazilian HSCT recipients (Campinas). A similar percentage (78%) was observed by Dieamant and colleagues13 in Brazilian (Campinas) allo-HSCT recipients with malignant and nonmalignant disorders. However, they studied a mixed population of pediatric and adult patients. Alternatively, the same authors (de Campos Dieamant and colleagues) studied a cohort that consisted solely of children and have reported a slightly lower rate of infection (60.6%), which is similar to our results.14 In contrast, Qayed and colleagues17 and Paris and colleagues18 have reported that CMV infection in pediatric patients occurred at rates of ~33% (Atlanta, Georgia, USA) and ~25% (Santiago, Chile), respectively.17,18 Significantly lower rates of CMV infection were noted in pediatric patients from Fukuoka, Japan (45.4%),15 and Seoul, Korea (24%).16
Data from cohorts comprising adults or mostly adult individuals also cover a broad range of occurrence. Higher rates of CMV DNAemia (84.3% in patients from Kurume, Fukuoka, Japan) were observed in recipients of allo-HSCT for various hematologic disorders.19 In contrast, lower rates of CMV DNAemia were reported in other studies and ranged from ~50% in the Birmingham-London region of the UK20 to 39.1% in Westmead, New South Wales, Australia.21
Cytomegalovirus DNA was detected in plasma of all haplo-HSCT recipients in our present study. A much lower rate was noted by Goldsmith and colleagues,22 who reported CMV-positive test results in 53.4% of T-cell-depleted haplo-HSCT recipients (in St. Louis, Missouri, USA). In our study, we observed significantly higher rates of active CMV infection in recipients of haplo-HSCT than in recipients of MSD HSCT and MUD HSCT. This is likely a consequence of a greater level of immunosuppression and slower recovery of the immune system in patients with haploidentical donors.
Furthermore, we observed a significantly lower rate of CMV DNAemia in MSD recipients (21.7%). Similarly, Jaing and colleagues23 have reported a significantly higher number of CMV-free patients (30.9%) who received MSD grafts versus unrelated/mismatched donor HSCT recipients (in Taoyuan, Taiwan). In a study from Houston, Texas, USA, the MSD recipients without acute GVHD had the lowest rate of reactivation (31%),24 although these were all late-stage CMV reactivations. It is known that recipients of MSD transplants recover immunity faster than recipients of other transplants, and this may lower their risk of CMV reactivation. Moreover, most of these pediatric patients had an underlying malignancy. Reduction of immunosuppressive therapy is recommended in such patients because a subsequent GVHD could confer resistance due to a graft-versus-leukemia effect and thereby abrogate viral reactivation rates. However, the opposite effect was noted in a cohort of pediatric patients with acute leukemia in whom there was no significant difference in the rate of CMV reactivation with regard to the type of matched donor.15 Of note, in our present study we observed fewer episodes of viremia in patients whose donors were related to the recipient and for whom a good HLA match was obtained.
There are several explanations for disparities in the published reports on CMV incidence rates, the most important of which is cohort composition, ie, age and number of patients, solid-organ transplants versus HSCT modalities, administration of antiviral prophylaxis or lack thereof, and diversity of transplant modalities. These variances may also appertain to a specific methodology. Quantitative PCR may detect CMV infection earlier than the antigenemia assay23; moreover, neutropenia is a clear limiter for the antigenemia assay.25
There was no significant correlation between CMV infection and sex or age, although most CMV-positive samples were found in male pediatric patients and in older pediatric patients. Our results are similar to a retrospective study of children with allo-HSCT for whom sex and age were not good predictors of CMV infection,23 as well to results from Inagaki and colleagues.15 Further evidence to corroborate our results was shown in a group of patients of mixed age (pediatric and adult),26 as well as in a cohort of adult patients.22 Conversely, in a previously published study of haplo-HSCT, the authors reported a higher rate of CMV viremia in female recipients.27
Interestingly, Miller and colleagues28 reported a higher tendency of older patients (mean age 17.5 years) toward CMV infection, albeit not statistically significant. These results were attributed to a higher expectation of CMV seropositivity as patients aged. Significance between patient age and incidence of CMV antigenemia was demonstrated by univariate analysis in another study, although the multivariate analysis showed no statistical significance.29 Goldsmith and colleagues described significantly higher rates of CMV viremia in subjects older than 60 years,27 and similar results for patients 50 years or older have been reported elsewhere.30
Of note, younger donor age (as well as active disease at the time of transplant) was related to a significantly higher rate of late CMV reactivation in a study of patients who received allo-HSCT in response to hematological malignancies.24 In a study of CMV DNAemia in critically ill but immunocompetent patients, in multivariate models, male sex was associated with a higher risk for reactivation,31 although no such connection was shown with respect to patient age. Notably, in healthy populations, rates of virus reactivation are substantially higher in female patients.32 Higher rates of CMV DNAemia in older children may be explained by a greater seroprevalence in older populations. Likewise, different treatment modalities, fluctuating degrees of immunosup-pression, and nutritional stress33 could account for patterns in variability of age and rates of CMV reactivation. The full causative interplay between sex and CMV reactivation has not yet been demonstrated.
Most previous studies have reported an association between GVHD and CMV infection, although exceptions do exist. In the present study, GVHD was not significantly linked to CMV infection. Likewise, Goldsmith and colleagues did not observe this association in their patients.22 This is in contrast to reports in which the CMV infection correlated with grade 3 and grade 4 acute GVHD23 and in which patients with GVHD had a significantly higher risk for CMV infection.26 The GVHD was also identified as a factor in the late recurrence of CMV in a multivariate analysis.24 A significant association was demonstrated between chronic GVHD and CMV positivity by Inagaki and colleagues.15 Even higher correlation between acute GVHD and CMV infection was provided by Yoon and colleagues.16 Our results contrast with these conclusions because of our limited number of patients, different demog-raphics of patient groups, and the inconsistencies in patient group configuration.
Our results in immunosuppressed patients contrast with other studies because of intricate differences in within-the-host demographic processes of the CMV populations themselves (eg, virion population expansion in size that followed severe bottlenecks, periodic admixture events during superinfection, and gene flow between compartments) as reviewed by Sackman and colleagues.33
In the present study, significant preference for CMV DNAemia to develop was observed neither in malignant nor nonmalignant conditions. Nonetheless, virus infections were detected more often in patients with malignancies. This is in agreement with similar studies.16,23,26
We failed to demonstrate a connection between the viral load and different HSCT modalities. However, recipients of haplo-HSCT and MUD HSCT had smaller viral loads, and the opposite was true for recipients of MSD transplants. Likewise, no associations between viral load and transplant modality, HLA matching, and donor relatedness have been reported in a cohort of recipients of peripheral blood and bone marrow HSCT.34 To our knowledge, our study is the first to prospectively appraise CMV viral load with regard to categories of graft donors in the youngest allo-HSCT recipients in Serbia. More studies are warranted in order to examine the connection, if any, between degrees in the HLA match or mismatch of a donor and the allograft recipient to confirm our results.
In this study, age was not a significant correlate for the level of viremia; however, there was sex patterning, with higher rates of viremia in female pediatric patients versus male pediatric patients (P = .041). As yet, we are not aware of any reports to confirm such significance. The basis for our observation of the sex-related difference in viral loads remains unexplained. The data regarding age and viral load in pediatric allo-HSCT recipients are few. There is evidence, however, that in healthy populations the CMV copy number increases markedly over approximately 70 years of age.35
The principal disease proved a noteworthy factor to influence differences in viral copy numbers. Namely, patients with malignancies tended to show higher viral loads after transplant than those with nonmalignant disorders (P = .019). There are few extensive studies on the viral load relative to the type of underlying morbid condition. Habib and colleagues investigated the increase in antigenemia with absence of concomitant increase in viral copies and did not find significant relationships between active CMV infection episodes and age, underlying disease (malignant vs nonmalignant), and the incidence of acute or chronic GVHD.36 The higher rates of CMV infection in patients with malignant diseases could be explained by a greater degree of immunosuppression in these children; this alone, however, most likely does not account for the observed significance. Further studies are warranted for the association to be fully elucidated.
Finally, the GVHD presented another variable that lacked an association with the level of viremia despite higher viremia noted in patients with this complication. Similar results were provided by Howden and colleagues in the context of acute GVHD.34 However, there is significant support for an association of increased DNA loads and acute GVHD as observed in a cohort of HSCT recipients.37
Here, we sought to increase our understanding of the factors involved in CMV infection and viral load in pediatric allo-HSCT recipients. Cytomegalovirus remains the most common viral pathogen in these patients. Although incidence rates vary, rates of reactivation are high and compulsory monitoring is crucial for judicious patient care. Age and sex are not crucial to predict higher risk of viral incidence. Higher levels of viremia in female pediatric patients and in nonmalignant settings after HSCT justify more scrutiny. The association between viral load and type of received graft remains to be studied in detail; the same holds true for the comparison of viral load with patient age. Regulation of various evidentiary details of patient cohorts, together with large cooperative multicenter studies, should facilitate future reports of reliable data.
Volume : 19
Issue : 11
Pages : 1156 - 1162
DOI : 10.6002/ect.2021.0282
From the 1Faculty of Medicine, University of Belgrade; the 2Department of Virology, Institute of Microbiology and Immunology, Faculty of Medicine, University of Belgrade; and the 3Mother and Child Health Care Institute of Serbia Dr Vukan ?upi?, Belgrade, Serbia
Acknowledgements: This work was supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia (MPNTR grant agreement, project No. 1750-73). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Other than described, the authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no declarations of potential conflicts of interest.
Corresponding author: Marko Jankovi?, Faculty of Medicine, University of Belgrade, 8 Dr Suboti?a starijeg Street, Belgrade 11000, Serbia
Phone: +381 63 7357799
Table 1. Patient Characteristics and Cytomegalovirus DNA Status Information