PFAPA flares observed during COVID outbreak: can emotional stress trigger PFAPA attacks? A multicenter cohort study

Periodic fever, aphthous stomatitis, pharyngitis and cervical adenitis (PFAPA) is one of the most common autoinflammatory syndrome. It was reported in 1987 by Marshall et al., who described recurrent episodes of fever along with pharyngitis, cervical lymph nodes, and mouth ulcers mimicking cyclic neutropenia, but with a normal neutrophil count [1]. The exact prevalence of this syndrome is unknown, but PFAPA appears to be more common than other periodic fever diseases [2].

A typical PFAPA attack usually lasts 3-7 days, with intervals of 2-8 weeks [3]. At certain stages of the disease episodes may occur at regular intervals, with predictable attacks [4]. Medications may reduce the frequency of attacks, especially glucocorticoids therapy where a single dose usually halts the attack within hours, but in about half of patients frequency of attacks increases [5]. Our group recently found that a Mediterranean ancestry may play a role in the development of the disease [6]; however, other factors that affect the development or severity of the disease remain unknown.

We and others have clinically observed that emotional stress may play a significant role in the development of PFAPA attacks. However, periodic emotional stress as a contributor to an immunologic burst is hard to prove using scientific tools, especially in toddlers, although the association between psychological factors and attacks in patients with familial Mediterranean fever (FMF) is well-established [7].

In March 2020, during COVID-19 pandemic outbreak, a general lockdown was announced in Israel, closing all educational institutions and work places for approximately 2 months. Subsequently and on short notice (~2 weeks), preschools and schools were reopened. We assumed that familial and personal stress due to COVID-19 outbreak consequences, along with returning to school (considered a significant stressor in children [8]) on short notice may have served as a stressful event to the patients, and therefore as a trigger for induction of flare.

The objectives of this study were to examine whether emotional stress triggers PFAPA attacks. A secondary goal was to examine whether parents of children with PFAPA identify specific environmental triggers for their children's flares.

Of 279 potentially eligible patients, 187 were reached and 81 excluded due to their child being in remission for at least 6 months prior; 106 patients' parents were enrolled and filled the first questionnaire; and 99 parents filled out the second questionnaire. The enrollment flow chart is depicted in Fig. 1.

Table 1 summarizes the sociodemographic and clinical information regarding the paediatric patients in the study: 58/99 (58.5%) were male, average age was 7.28 ± 2.7 years, and 60 (60.6%) attended the early years of elementary school at time of enrollment. Most (91, 91.9%) paediatric patients were treated with single dose of a corticosteroid during attacks, and the mean duration of attacks when treated with corticosteroids was 5.2 ±11.5 hours.

On the paediatric emotional distress scale, during the stressful period the mean paediatric emotional distress scale score was statistically significant higher than in the subsequent less stressful period (35.6 ± 8.1 vs. 32.1 ± 7.7, respectively, P <0.047), as shown in Table 2. This served to validate that the timing of the first and second phone calls corresponded with our assumption of stressful and less stressful periods. The distribution of paediatric emotional distress scale among the patients in the 2 different periods is illustrated in supplementary figure S2.

More PFAPA attacks were recorded during the stressful compared to the less stressful period, with 41 (38.7%) of parents reporting at least one PFAPA attack in the preceding two weeks to the first call, compared to only 24 (22.6%) prior to the second call (p = 0.017). Moreover, a statistically significant number of patients reported more than one attack during the stressful period, as compared to the less stressful period (14, 13.2% vs. 5, 4.7%, respectively, P = 0.04).

Finally, the majority 62 (58.5%) of parents reported triggers for PFAPA attacks, 30 (28.3%) reported no triggers, and 14 did not answer the question. Among parents who reported triggers, the most common (45, 72.6%) were emotional triggers, with 14 parents identifying other triggers: physical activity -5, seasonal - 4, changes in temperature - 2, contact with water - 1, type of food - 1, and exposure to the sun - 1.

After achieving validation of, and comparing stressful and less stressful periods, we were able to show that a higher percentage of children experienced at least one PFAPA attack during the first stressful period, as compared to the second less stressful period (control). Furthermore, the number of attacks was significantly higher during the stressful period. In addition, the majority of parents reported emotional factors as triggers for attacks.

Although development of PFAPA attack in response to emotional stress is a frequent complaint from the parents in our clinic (e.g. 1^st day of school, birthday celebration, familial vacation, anxiety due to fight with a friend or a parent), it was never evaluated. The COVID-19 outbreak served as an opportunity to assess this clinical observation due to widespread emotional and economic crisis at the national level, therefore standardizing the events and their timing, and thus minimizing potential bias.

We believe these findings demonstrate, with a high degree of probability, that mental factors such as stress or excitement are triggers for PFAPA attacks.

One of the strengths of the work is the use of relatively objective tools to test the hypothesis. Instead of relying only on a questionnaire based on retrospective self-reporting, which may be biased by previous beliefs or different perceptions of the disease by parents, we chose to contact the parents at a time which we assumed was mentally stressful for the children, and initially asked a simple and objective question ("Have there been attacks in the last two weeks"). Only later, to reinforce the findings, we presented the open question regarding possible triggers and submitted the PEDS questionnaire.

Recent studies have shown that the proportion of children presenting with psychological problems during the COVID-19 pandemic was high and it is likely that the pandemic is worsening the mental health of youth [14]. Leeb et al. reported that the proportion of pediatric ED visits due to mental health issues during the "first wave" of the COVID-19 pandemic increased by 24% for children 5-11 years compared to the same periods in the past [15]. This period is parallel with our study period. During the pandemic there was an increase in hospital mental health related admissions among children and adolescents [16, 17].

The information on triggers for PFAPA attacks is scarce. Although PFAPA attacks tend to occur on a regular cycle, in the first report on the syndrome Marshal et al. commented that intervals between attacks may vary between and among patients, although they found no seasonal variations or association with infections [1]. Many parents report on irregular frequency, for example, to intermittently skip an episode of fever, [5] and the treatment with glucocorticoids may also varies the intervals between attacks [5]. Mean age of patients in our study is higher than typical PFAPA age, which may explain majority of the patients had the disease for a while and it is possible that in later years in the course of the disease regular periodicity is not seen. Environmental factors such as lack of breastfeeding or maternal smoking were found to be associated with the onset of the disease in a small study, but were not examined as triggers of attacks [18]. Padgett reported menstruation as a trigger for PFAPA attacks in an adolescent girl [19]. In addition, a claim was raised regarding the role of the microbes of the tonsils as a trigger for attacks [20], but the issue has not yet been unequivocally proven [21].

The fact that mental state is a trigger for an attack of periodic fever is known in other diseases, especially FMF. In that disease, the frequency of attacks is less predictable than for PFAPA⁴, and several attempts have been made to identify triggers for flares. Karadag et al. conducted a questionnaire study that showed 49.8% of patients (mostly adults) identified psychological stress as a trigger for attacks [22]. Yenokyan et al. found a positive association between FMF attacks and stress events in the two days preceding them [23]. For children with FMF, it was found that psychosocial conditions may explain 27% of the variability of attacks [24]. Another study found that the mean depression and anxiety scores of FMF paediatric patients were positively correlated with the number of attacks [25].

The biological explanation for the effect of stress on FMF attacks may be related to activation of the sympathetic system, initiating the release of catecholamines (epinephrine and norepinephrine), which leads to the secretion of Il-1β, activation of the inflammasome via activation of cAMP signaling pathway protein kinase A (PKA), and an increase in gene expression of REDD1 (regulated in development and DNA damage responses) [7]. The pathogenesis of PFAPA involves the activation of IL-1β as well as inflammasome [26], and therefore it is possible that the effect of emotional triggers is similar to what was found with FMF.

This study reports PFAPA flares observed during COVID outbreak in Israel. These results show that emotional distress is associated with higher frequency of PFAPA attacks, thus may be considers as environmental trigger for PFAPA attack. Future research is needed to confirm these findings and characterize additional triggers.