Nasopharyngeal carcinoma (NPC) is rare in most regions of the world, with an incidence of less than 1 per 100,000 person-years [1]. However, in some regions of South China and Southeast Asia, the incidence of NPC approaches 20–50 per 100,000 person-years [2, 3]. The pathogenesis is not fully understood; however, accumulated evidence indicates that genetic susceptibility, Epstein–Barr virus (EBV) infection [4], and environmental factors, such as salted-fish consumption and cigarette smoking, increase the risk of NPC development [5, 6]. NPC easily spreads to the lymph nodes and distant organs, and most patients are diagnosed only when the tumor has reached an advanced stage (stages III and IV) [7].

More than 95% of adults in some NPC endemic areas are infected with EBV. Infection typically occurs in childhood, and the virus remains in a latent phase within resting memory B cells under the strict monitoring of the immune system [8, 9]. However, the virus may be periodically reactivated in response to environmental stress [5]. Several studies have reported that smoking [5], Cantonese-style salted fish [10], and Chinese herbs [11], all of which contain EBV-inducing substances, may drive EBV into the lytic cycle [12].

Given the unclear etiology of NPC, primary prevention remains difficult, and the main strategy for reducing NPC mortality is to identify the disease at an early stage [13]. In the preclinical phase, serum antibodies against EBV-related antigens remain increased for an average of 38 months [14–16], and the serological tests for these markers are simple and inexpensive [15]. Therefore, these antibodies have been used as screening markers for NPC in endemic areas since the 1970s. The immunoglobulin A (IgA) antibodies against viral capsid antigen (VCA-IgA) and early antigen (EA-IgA) are the most popular screening markers. VCA is the structural protein expressed by EBV during the late stage of the lytic cycle following viral DNA synthesis, whereas EA is a nonstructural protein expressed during the switch from the latent phase to lytic replication before viral DNA synthesis. The two markers have distinct features for NPC screening: VCA-IgA has a higher sensitivity than EA-IgA (79.6% vs. 31.6%); however, its specificity is lower than EA-IgA (92.0% vs. 99.7%) [13]. Thus, in practice, these two markers are typically used in combination for NPC screening [15, 17].

Several prospective screening studies have demonstrated early detection rates for NPC of greater than 70% [15, 18, 19]. Nevertheless, EBV infection is ubiquitous; thus, routine use as a screening tool may have relatively low specificity, which may lead to potentially costly and invasive follow-up diagnostics [19–21]. Therefore, an investigation of the natural variations of EBV antibodies in non-NPC participants may identify particular serostatus patterns for improved NPC screening.

In this study, we used EBV VCA-IgA and EA-IgA data from a prospective, population-based screening study with no clinical evidence of NPC during a median 15.9-year follow-up in South China [17] to determine the distribution of baseline EBV VCA-IgA between sexes and among age groups as well as the long-term titer fluctuations of the VCA-IgA and EA-IgA serostatus in non-NPC participants. The aim was to identify additional factors useful for EBV antibody-based NPC screening strategies in endemic areas.

The current findings indicated that the baseline EBV VCA-IgA-positive rate was 7.6% in males and 6.6% in females, and the positive rate tended to increase with age from 6.0% in the 30–39 year age group to 8.4% in the 50–59 year age group. These findings are consistent with the epidemiological characteristics of areas with a high NPC risk, with increased incidence in males and peak incidence among the population aged 50–59 years [3]. Therefore, an age- and sex-dependent high-risk cutoff value for VCA-IgA should be considered. The currently recommended referral criteria for nasopharyngeal endoscopy and biopsy (VCA-IgA ≥1:40 and/or VCA-IgA and EA-IgA double positive) may be appropriate for men of 30–39 years old; however, these criteria become less specific with age and may thus increase the possibility of unnecessary diagnostic tests in participants of 50–59 years old. Larger randomized clinical trials should be conducted to further address this concern.

Titer fluctuations in the EBV antibody serostatus were common in this non-NPC population. In a previous NPC screening study in an endemic area, many non-NPC participants were advised to undergo expensive and invasive nasopharyngeal endoscopy and biopsy examinations as a result of increased serum EBV antibodies [14]. Our findings demonstrated that 55.5% of the participants with an initial VCA-IgA-positive status met the high-risk criteria, whereas 20.6% of the participants with an initial VCA-IgA-negative status met the high-risk criteria at least once during the first 5-year study period. One potential strategy for reducing unnecessary follow-up examinations is to repeatedly test for serum EBV antibodies several weeks before referral for endoscopy and nasopharyngeal biopsy to not only preclude random measurement errors but also eliminate participants who exhibit a short-term EBV antibody increase caused by environmental factors. If the repeated tests indicate that the antibody levels remain increased, a fiberscopic examination or biopsy is recommended. Otherwise, these expensive and invasive examinations may be avoided.

Approximately 85.2% of the participants with an initial VCA-IgA-positive status converted to negative within 5 years. However, we determined that the seroconversion rate of NPC participants from positive to negative was substantially lower (15.4%) in our original study (P < 0.001) [17] than the seroconversion rate of non-NPC participants. This finding suggested that serum EBV antibody titer fluctuation may be used to determine the optimal screening interval. Similar to several other NPC screening studies in South China [14, 15], the preset screening interval was 1 year for participants with both VCA-IgA and EA-IgA positive, and the positive predictive value was less than 1% [13, 15]. We speculated that the use of results from only one test to determine the clinical screening interval does not take into account transient increases induced by environmental stressors, which may thus result in many false-positive evaluations. Thus, a rational screening interval for participants with EBV seropositive at baseline may be adjusted by retest results. If the repeated EBV antibody tests remain positive, then a retest in 1 year should be recommended. In contrast, the screening interval may be prolonged in negative cases.

These frequent seroconversions of EBV antibodies may arise from environmental inducers that change throughout an individual’s lifetime. The virus is typically restricted to resting memory B cells by the host immune system; however, it may be periodically reactivated. Following EBV reactivation, it switches from a latent phase to a lytic phase with concomitant expression of viral antigens. The host immune system produces a series of antibodies against viral antigens, especially IgA antibodies against lytic phase antigens, including VCA-IgA and EA-IgA. The living environment may frequently change. When these inducers are eliminated, the virus may return to the latent state in memory B cells, and the serum antibody levels gradually decrease, which results in seroconversion to a negative status.

We determined that the seroconversion rate of EA-IgA from positive to negative was significantly higher than that of VCA-IgA in the first 5-year period (100.0% vs. 85.2%, P = 0.001). In contrast, the seroconversion rate of EA-IgA from negative to positive was lower than that of VCA-IgA (19.6% vs. 100.0%, P < 0.001). A high-specificity screening marker must yield a low false-positive rate throughout the screening course for the non-disease population. Our previous study [13] and other studies [16, 21] have reported that the baseline EA-IgA-positive status is lower than the VCA-IgA-positive status in the non-NPC population. Furthermore, VCA-IgA exhibited a higher two-way seroconversion rate (from negative to positive and from positive to negative) in the non-NPC population, whereas the EA-IgA conversion was only one-way (from positive to negative). This pattern results in a higher false-positive rate for VCA-IgA and a higher specificity for EA-IgA during follow-up. Thus, EA-IgA may be more specific for NPC screening than VCA-IgA.

This study had several limitations. First, the study was conducted in Sihui County, China, which has a high NPC incidence worldwide. Therefore, the results may not be translated to other populations in medium- or low-incidence areas. However, NPC screening is more likely to succeed if conducted in a high-risk population; thus, these findings are useful for developing screening strategies for this population. Second, the participants recruited for NPC screening may not represent the general population. For example, screening participants may be healthier or more health-conscious than the general population. However, no report to date has demonstrated that the serum EBV antibody status in the non-NPC population is affected by compliance for NPC screening.

In summary, serum EBV antibody levels in a non-NPC population from an endemic area were increased in males and older participants. Thus, an age- and sex-specific cutoff for the anti-EBV antibody level should be considered for NPC screening. Repeated testing indicated that EBV seroconversions were common in the non-NPC population. This finding may contribute to the design of more efficient NPC screening strategies with a rational adjustment of rest-retest intervals.