Herpesvirus infections are a serious problem at the present time due to their ubiquity, high infection rate of the population, and the possibility of their reactivation in the development of immunodeficiency conditions having a high risk of adverse outcomes [1–3]. In recent years, increasing attention has been paid to herpesviruses due to the increasing case numbers of active herpesvirus infections, including human betaherpesviruses 6A/B, in people with infectious diseases caused by the SARS-CoV-2 virus (COVID-19) [4–6], as well as in children with recurrent respiratory diseases (RRD). Although human betaherpesvirus 6A/B (HHV-6A/B) was first detected in 1985 in immunocompromised patients [7], it is currently known as one of the most common herpesviruses with the level of seropositive individuals to HHV-6A, HHV-6B or both types in the adult population reaching 95% [8]. Like other representatives of the family, HHV-6A/B is characterized by polymorphism of clinical manifestations and acts as a trigger of autoimmune and lymphoproliferative diseases [9–13].

The high level of infection with HHV-6A/B leads to the frequent detection of markers of betaherpesvirus HHV-6A/B infection, difficulties in differential diagnosis (DD) of active and latent forms, especially in children with lymphoproliferative and catarrhal syndromes, as well as RRD, overdiagnosis of active forms of infection and subsequent administration of immunotropic therapy with antiviral purpose [8][14–16].

The aim of our study is to propose a modern DD method for active and latent forms of HHV–6A/B infection in children as a means of optimizing patient management tactics.

The study, carried out in the Children’s Scientific and Clinical Center for Infectious Diseases of the Federal Medical and Biological Agency of Russia from 2021 to 2023, included 152 patients aged 1 month to 17 years 11 months and 29 days, 112 of whom made up a training sample for training the classifier, while 40 comprised a test sample for validating the classifier. Inclusion criterion for the training sample: detection of HHV-6A/B DNA in whole blood by qualitative polymerase chain reaction (PCR) with subsequent assessment of the levels of threshold amplification cycles (cycle threshold, Ct). Exclusion criteria: disagreement of legal guardians of patients to participate in the study; severe somatic diseases in the decompensation stage; detection of SARS-CoV-2 RNA in the upper respiratory tract by PCR.

Depending on the severity of the condition, The patients were admitted to a 24-hour or day hospital. Upon admission, anamnestic data was collected, including the data of medical documentation of the outpatient stage (history of the child’s development — form No. 112/U or an extract from it). During the examination, special attention was paid to assessing the presence and severity of lymphoproliferative syndromes typical of herpesvirus infection, namely: adenoiditis, tonsillitis, lymphadenopathy, hepatosplenomegaly.

On the first day of hospitalization, clinical and biochemical blood tests were performed to assess the level of C-reactive protein (CRP). To interpret the results of the laboratory examination, reference values of manufacturers of test systems were used, taking into account age characteristics. All patients underwent a whole blood examination using high-quality real-time PCR to detect HHV-6A/B DNA with an assessment of Ct level (kits “AmpliSens EBV/CMV/HHV6-screen-FL” (Central Research Institute of Epidemiology, Russia). The Ct level is the value at which the threshold line intersects the S-shaped signal accumulation curves in the test samples and control samples. Thus, this indicator indicates the number of amplification cycles required to start detecting the virus DNA in the sample. A low Ct value indicates a significant amount of isolated viral DNA in the sample, which indicates a high viral load, while a high Ct value indicates a low viral load [17].

Statistical processing was carried out using Microsoft Excel modules, the StatSoft Statistica 7.0 software package. Using a training sample, a model was built for the dependence of the patient’s belonging to one of two groups according to the value of factors in various combinations, where 1 is the latent form of HHV-6A/B infection (n = 89), 2 is the active form of HHV-6A/B infection (n = 23). The group of active infection included children with sudden exanthema caused by laboratory-confirmed HHV-6A/V [18–20]. 27 anamnestic, clinical and paraclinical signs were considered as potential predictors in the initial training model. The discriminant model was constructed using the step-by-step inclusion of predictors using the Fisher F-criterion (the value of the F–criterion is 4.0, the lower limit of tolerance is 0.01). The significance level p < 0.05 was used for all types of statistical analysis.

In order to improve the accuracy and timeliness of diagnosis based on a comprehensive clinical and laboratory examination, including a PCR study, a predictive model using discriminant analysis was developed. Of the 27 studied features, 4 parameters were identified as demonstrating the greatest statistically significant differences between the groups — p < 0.001 (Table 1).

The obtained model of DD of latent and active forms of HHV-6A/B infection in children demonstrated high statistical significance (p < 0.001) and classificatory ability at the level of 93.75%. The model, which takes into account such parameters as severity of fever, the presence of a cough, the absolute number of neutrophils and the Ct DNA level of HHV-6A/B, has high sensitivity (91.3%) and specificity (94.4%).

On this basis, a diagnosis can be carried out as follows. When seeking medical help for a child with a suspected active form of HHV-6A/ B infection (fever, catarrhal and lymphoproliferative syndromes, clinical manifestations of infectious mononucleosis, the presence of RRD), a clinical blood test is performed on an automated hematology analyzer with an assessment of the absolute value of neutrophils, a whole blood study by PCR with Ct DNA determination of HHV-6A/V. The obtained data of predictor features are used for calculations using formulas (1, 2) of the linear discriminant function (LDF):

LDF1 (the latent form of HHV-6A/B infection): (1)

42.7×Х1 – 49.2×Х2 – 6.2×Х3 + 5.2×Х4 – 865,

LDF2 (the active form of HHV-6A/B infection): (2)

45.1×Х1 – 53×Х2 – 6.7×Х3 + 4.7×Х4 – 940,

where Х1 — the maximum severity of fever, °С;

Х2 — the presence of a cough: no — 0, yes — 1;

Х3 — absolute neutrophil count, ×109/L;

Х4 — threshold cycle of HHV-6А/В DNA amplification.

The obtained values of LDF1 and LDF2 are compared as follows: if LDF1 is greater than LDF2, the latent form of HHV-6A /B infection is diagnosed; with LDF2 more than LDF1, the active form is diagnosed.

A classification of the test sample (n = 40) was performed. Discriminant analysis successfully identified 8 out of 10 cases of the active form of HHV-6A/B infection and 29 out of 30 cases of latent infection. The total percentage of correct diagnoses was 92.5%, which is comparable to the results in the training sample. The patent for invention No. 2817089 dated 09/04/2024 was obtained [21].

For a better understanding of the proposed model, we present our own clinical observation.

Clinical case No. 1

The girl Ksenia, 8 years and 2 months old, was admitted to the Children’s Scientific and Clinical Center for Infectious Diseases of the Federal Medical and Biological Agency on a planned basis due to RRD accompanied by febrile fever, for examination and selection of therapy. The child has a history of Schinz disease, atopic dermatitis. The epidemiological history is not burdened. Upon admission, condition satisfactory, body temperature falling within the normal range (the value of the predictor X1 is “36.5”). Skin of normal color, moist, without an infectious rash. Rhinitis and cough absent (the value of the predictor X2 is “0”). Pharynx calm, tonsils not enlarged, no plaque. Peripheral lymph nodes not enlarged. Heart tones clear, rhythmic. Vesicular breathing carried out in all parts of the lungs; no wheezing heard. Abdomen soft, painless; liver and spleen not enlarged. Diuresis preserved. Following a clinical blood test, all parameters within the reference values (the value of the predictor X3 (absolute neutrophil count) is “3.35”). No inflammatory changes detected in the general urine analysis. An increase in normobiota was observed in the crops of the discharge from the nasopharynx and oropharynx to the microflora. Taking into account the anamnestic data, in order to exclude reactivation of herpesviruses, high-quality whole blood PCR for HHV-6A/B DNA, Epstein-Barr virus and cytomegalovirus was performed by PCR: HHV-6A/B DNA was detected, Ct value (predictor X4) — 31. When examining blood by ELISA for specific antibodies of the IgM class, IgG to herpesviruses, the following results were obtained: IgG to cytomegalovirus (CMV) detected. IgM antibodies to CMV, IgG antibodies to the nuclear and capsid antigens of EBV, HHV-6A/B not detected. Given the positive result of PCR on HHV-6A/B DNA, final clinical diagnosis was interpreted as persistent HHV-6A/B infection in the activation stage. The child was prescribed immunotropic therapy with antiviral action consisting of meglumine acridonacetate according to the manufacturer’s instructions in a course of 23 days.

The calculation was performed retrospectively using the formulas:

LDF1 = 42.7×36.5 – 49.2×0 – 6.2×3.35 + 5.2×31 – 865 = 834.98

LDF2 = 45.1×36.5 – 53×0 – 6.7×3.35+4.7×31 – 940 = 829.41

Since LDF1 > LDF2, a 94.4% probability of a latent form of HHV-6А/В infection was pronounced.

This case demonstrates that, according to the proposed discriminant model, the probability of an active form of HHV-6A/B infection was extremely low (5.6%); on this basis, it was possible to avoid a long course of immunotropic therapy.

The existing problem of overdiagnosis of the active form of HHV-6A/B infection is related to the interpretation of the results of high-quality PCR of whole blood without taking into account the possible latency of the virus in mononuclear cells. Despite their unsuitability in most cases, the detection of HHV-6A/B DNA generally leads to the prescription of long-term repeated courses of immunotropic therapy. Thus, the developed prognostic model using discriminant analysis based on a comprehensive assessment of clinical and laboratory parameters solves the problem of DD of active and latent forms of HHV-6A/B infection in children with lymphoproliferative and respiratory syndromes.

Utkin et al. proposed a method for identifying infectious mononucleosis associated with human betaherpesvirus 6A/B. The invention takes into account such biomarkers as AVEN mRNA, CHUK transcript 2, CIRBP transcript 2, TRAF3 transcript 2 and IRAK4 transcript 10. Although the method is extremely interesting from the point of view of studying the molecular mechanisms of pathogenesis, its disadvantages include the assessment of infection activity in relation to only one nosological form (infectious mononucleosis), inaccessibility of implementation in practical healthcare with a large flow of patients, lack of technical equipment of laboratories and trained personnel necessary for the study of genes and transcripts, as well as high financial costs and difficulty in interpreting the obtained results [22].

As a solution of this problem, Melekhina et al. developed a method for identifying the indications for antiherpetic therapy of HHV-6A/B infection in children with acute respiratory diseases. The active form of infection, which requires the prescription of antiviral therapy, is established when more than 100 copies of the of HHV-6A/B DNA/105 cells are detected in the blood by quantitative PCR. The proposed method is characterized by its speed of execution and adequate approach to the obtained results taking into account different levels of viral load, as well as its convenience in the presence of quantitative PCR — which, however, is not an affordable diagnostic method at all levels of medical care [23].

It is possible to use reverse transcription PCR to differentiate the active and latent forms of HHV-6A/B infection. This method is characterized by its higher sensitivity and specificity than culture-based studies. The determination of the form of infection depends on the detected matrix RNA of viral proteins: the open reading frames U42, U22, U38, U100 indicate an active form of infection, while U94, conversely, indicates a latent infection [20]. Thus, the method is promising, but currently unavailable in practice.

Our described model offers additional advantages such as simplicity, accessibility, speed of execution, as well as the possibility of providing DD of active and latent forms of HHV-6A/B infection without involving additional financial costs for expensive laboratory tests. The uniqueness of the method is confirmed by the patent for the invention.

The developed prognostic model can be applied in clinical practice for the diagnosis of active and latent forms of HHV-6A/B infection in children with catarrhal, lymphoproliferative syndromes when HHV-6A/B DNA is detected in whole blood, as well as to justify the appointment of immunotropic therapy with antiviral action. When establishing the latent form of HHV-6A/B infection, it is necessary to conduct a further diagnostic search to exclude autoinflammatory diseases accompanied by an appropriate symptom complex or the active form of infectious mononucleosis of mono– and combined herpesvirus etiology along with lesions of organs and systems associated with viral infection.