Primary infection by Epstein–Barr virus (EBV) typically presents with all the main features of mononucleosis: fever, asthenia, enlarged lymph nodes, exudative pharyngitis, hepatosplenomegaly and exanthema following administration of penicillin. But in some cases it presents with few or atypical symptoms, making diagnosis more challenging. In younger patients, primary infection is usually asymptomatic or oligosymptomatic, and it is in adolescence that primary infection starts to present with the full picture of mononucleosis.

The typical age of primary infection by EBV varies based on cultural and socioeconomic factors. In developing countries and rural areas, most children have been infected by EBV between ages 3 and 6 years. In contrast, in developed countries and urban areas primary infection tends to occur at later ages and is more likely to produce more symptoms or a mononucleosis syndrome.1–3

The diagnosis of infectious mononucleosis is based on the clinical presentation supported by abnormal findings of laboratory tests (elevated white blood cell count with lymphocytosis and monocytosis, atypical lymphocytes and elevation of liver enzymes). The presence of heterophile antibodies is highly specific but has a low sensitivity, and their production increases with age.4–6 Oligosymptomatic forms of primary infection by EBV are usually confirmed by serologic testing.

Specific EBV viral capsid antigen (VCA) IgG and IgM antibody tests, EBV early antigen tests and EBV nuclear antigen tests are the serologic tests of choice to diagnose acute infection in immunocompetent hosts and to monitor the course of the infection.5–7 New molecular diagnostic techniques, such as detection of viral DNA by polymerase chain reaction, could help interpret uncertain results of antibody tests, although the published evidence on this approach is still scarce, especially when it comes to healthy individuals and the paediatric population, and further studies are needed to standardise diagnostic viral load values.8–10 These tests are not used to diagnose mononucleosis in immunocompetent patients.11

Few studies have investigated the seroprevalence of EBV and the ages at which primary infection occurs in the Spanish population. One of the studies conducted in Spain, by Pariente et al., found a peak between ages 2 and 4 years and another between ages 14 and 18 years.12

Most of the studies in the literature focus on hospitalised patients with symptoms of infectious mononucleosis,13–15 so few data are available on other forms of primary infection in children without a previous diagnosis of immune disease.

The aim of our study was to establish the epidemiological, clinical and laboratory characteristics of primary infection by EBV in healthy children in our catchment population and to assess their association with the clinical presentation of primary infection.

In our study, we found 2 age peaks in the incidence of primary infection by EBV before adolescence: 2–4 years and 7 years. Historically, the literature has described a higher incidence in adolescence.2 In our sample, up to 64% of patients with a serologic diagnosis of primary infection by EBV were aged less than 10 years, probably because older children have a more florid presentation, which in many cases leads to a clinical diagnosis without the ordering of confirmatory laboratory tests that was the basis of our study.14–16

Despite the higher frequency in younger children, our study found a predominance of the typical presentation (65 cases). In the group of patients with typical features, 13 only had 2 symptoms and 52 more than 2 symptoms. However, in many instances the third symptom belonged to the composite variables “other symptoms”, which encompassed clinical manifestations that are not specifically part of mononucleosis syndromes. This approach to the classification of clinical features may have led to an overestimation of the frequency of typical presentations in our sample.

The most frequent clinical manifestations at the time of primary infection by EBV were fever, cervical lymphadenopathy and exudative tonsillitis, which was consistent with most other published case series,3,13–15,17 although with different proportions. These differences are probably due to our sample including children that were not hospitalised, unlike most other studies in the reviewed literature.

One salient finding of our study was that coinfection by other viruses was detected in 22% of the sample. In some of these cases it is difficult to identify the aetiological agent based solely on the clinical presentation. After studying the outcomes in these children, we considered that in the cases that a follow-up evaluation was requested and patients had positive results of the EBV antinuclear antigen IgG test, the patients may have had coinfection in addition to primary infection with EBV, although we were unable to confirm this with certainty with the available data. There is a greater certainty that cases with typical presentations and positive for heterophile antibodies are cases of coinfection. There are few references in the literature about coinfection with other viruses, although there have been published reports of cases with serologic tests positive for both EBV and CMV.13,18 There are also 2 published cases with severe thrombocytopenia and coinfection by herpes simplex 6 and CMV,18,19 similar to patient 46 in our sample.

We considered the cases of patients 46, 76 and 93 possible false positives for EBV, although we were unable to confirm this, as DNA testing for detection of EBV was not performed at the time of diagnosis of primary infection and some individuals may not mount an immune response following primary infection.

The cases with indeterminate results of EBV-VCA IgM tests are also difficult to interpret, especially in the context of an atypical presentation. Of the patients for who we could obtain follow-up data, 5 did not seroconvert (Table 3). Patients 46 and 76 had positive IgM tests for other viruses (herpes simplex 6), and taking into account their presentation and the negative results of the heterophile antibody tests, we believe that these cases were probably false positives for EBV. In contrast, in patients 34 and 39, the EBV-VCA IgM test became positive after the initial test, and no other data were recorded suggesting coinfection, so we cannot tell whether they represented false positives for EBV or just had not developed immunological memory against the virus. None had complications and there was no evidence supporting the possibility of an underlying immune disorder at the time, although patient 39 had a history of lymphoblastic leukaemia diagnosed 6 years before that was currently in remission without treatment. It is important not to rule out the diagnosis of primary infection in patients with indeterminate results of IgM tests, as this is a possible occurrence in the early stage of infection. It may also be helpful to rule out coinfection by herpes viruses and repeat the serologic test later on in cases with indeterminate EBV-VCA IgM test results to see if there had been changes in serologic status.

The percentage of patients that tested positive for heterophile antibodies (33%) was lower in both groups compared to the current literature. Authors such as Sumaya and Ench found a prevalence of heterophile antibodies of up to 83% in children aged more than 4 years; although this proportion dropped to 50% in children aged 2 and 3 years.4,6 Other European authors have reported a somewhat lower prevalence (42%), in addition to a frequent proportion of negative tests in the early and late stages of primary infection.20 A subsequent review, however, has reported positive rates of 25–50% in children aged less than 12 years,21 which is more consistent with the prevalence found in our study. It is also worth noting that the different techniques used to detect heterophile antibodies do not have the same sensitivity and specificity.22 The test used in our laboratory (Monolatex, Biokit) has a sensitivity of 76% and a specificity of 98% according to the manufacturer. The low production of heterophile antibodies in our sample maybe due to 64% of the patients being less than 10 years of age. Another possible explanation is that tests may have been ordered in the early or the late stage of primary infection, especially in the 22% that were ordered in primary care clinics. Other Spanish authors have also found a low prevalence (32%) of heterophile antibodies in children aged more than 4 years without a clear explanation.13

Patients with oligosymptomatic forms of disease tended to be younger, as has been described in the literature.1–3 We did not find significant differences in the rest of the epidemiological and laboratory characteristics between both groups, although the number of patients for which we had data for some variables was small. We did not find descriptive studies in the literature on the different clinical presentations of primary infection. Most published studies have focused on hospitalised patients and patients with a presentation of infectious mononucleosis.13,14,17

Our study has the limitations inherent in retrospective designs, as we could not obtain follow-up data that could have helped interpret test results in some cases. We are aware of the potential bias involved in not being able to include the entire population of children with EBV infection, as we could only identify those that presented at least 1 symptom and in whom serologic tests were ordered for confirmation. Furthermore, the non-typical presentation group included patients that could have exhibited divergent patterns. On one hand, there were oligosymptomatic patients in whom one would expect a benign course of disease. On the other, there are patients with atypical manifestations, in whom we cannot rule out that the atypical presentation was due to some impairment in the immune response. At any rate, none of the patients with non-typical manifestations developed complications, and all had favourable outcomes.

On the other hand, we have managed to exclude the source of bias involved in having a sample consisting solely of hospitalised patients, and included patients with oligosymptomatic and atypical presentations.

In our catchment population, primary infection by EBV mainly occurs in younger children, with a predominance of oligosymptomatic presentations. The percentage of patients that tested positive for heterophile antibodies was very low in both the typical and the non-typical presentation groups, with positive results being more frequent in patients with a typical presentation. Cases with indeterminate results of the EBV-VCA IgM test were more frequent in the non-typical presentation group, in both oligosymptomatic patients and patients with atypical features.

Coinfection by other viruses is common, most frequently by CMV and parvovirus B19; in these cases, it is advisable to rule out a false positive.

The authors have no conflicts of interest to declare.