COVID-19 Pandemic and Eating Disorders: What Can We Learn About Psychopathology and Treatment? A Systematic Review

The PRISMA guidelines were followed to select and assess published articles [25].

In order to perform a systematic review of the literature, the following search keys were used in PubMed: “(COVID) AND (((eating) AND (disord*)) OR (anorexia) OR (bulimi*) OR (bing*))”. Bibliographies from relevant papers were inspected to identify studies not yielded by the initial search.

Articles were selected according to the following inclusion criteria: the paper (1) was a peer-reviewed research article published in English; (2) included samples of people with a current or lifetime diagnosis of any ED; and (3) was published between January 1st, 2020, and April 30th, 2021. Review papers, meta-analyses, commentary, study protocols, and case reports were excluded.

The literature search identified 696 papers, which were screened against the inclusion criteria. Fifty-two full-texts were assessed. Thirty studies were excluded because they did not meet the eligibility criteria: fifteen were editorial/commentary/letter, six were interview of healthcare providers or caregivers, three were study protocols, three were case series, and three were conducted on general population. This resulted in the inclusion of 22 studies in the qualitative synthesis. Figure 1 reports the flow diagram of study inclusion.

All the studies were conducted during the first wave of SarsCov-2 pandemic.

Most studies (14 of 22) were quantitative studies, 4 showed a quantitative–qualitative design, and 4 were qualitative studies. The main characteristics (diagnosis, sample size of each patients’ group, and diagnosis), the assessed outcomes, and the main findings of quantitative studies are reported in Table 1. The main qualitative findings studies are reported in Table 2.

All the selected studies, except those by Schlegl et al. [26•, 27•], Leenaerts et al. [28], and Frayn et al. [29], were conducted in mixed ED samples. Fifteen studies were conducted in patients with a current ED, 4 were in mixed samples with a current or a past ED, and 3 studies were conducted in samples with a self-reported diagnosis of an ED (1 current diagnosis, 2 current or past diagnosis). Five studies conducted a longitudinal assessment, although four of these [28, 30••, 31, 32] compared levels of symptomatology during the pandemic with those in the pre-pandemic period. The remaining studies adopted a cross-sectional design (e.g., asking participants if their symptoms had changed during the pandemic period), with the exception of 2 studies [33•, 34•] which conducted a retrospective evaluation of psychopathology. Only two studies [30••, 35] compared symptomatology levels between patients and healthy controls. Twelve studies reported that patients with EDs were in treatment, 5 did not report this data, and 3 studies were conducted in both treated and untreated patients and 2 in recently discharged patients. Three studies included outpatients, and 3 study included both inpatients and outpatients.

Seven studies reported the prevalence of SarsCov-2 infection among patients with EDs ranging from 0 to 5%.

Most studies [14, 26•, 27•, 30••, 31, 33•, 36, 37•, 38•] identified a significant impairment in ED core symptoms (i.e., food restriction, binge-purging behaviors, and physical exercise). Considering the studies adopting a descriptive procedure, we identified a worsening of ED symptomatology occurring in a range from 38 [14] to 83% [38•] of the assessed samples. However, no change in the severity of symptomatology was found in other two studies adopting a longitudinal design [28, 32], while an improvement in eating symptoms was observed by Fernandez-Aranda et al. [39]. Worsening in the severity of symptomatology did not differ between patients with a current ED diagnosis and those with a lifetime diagnosis in two studies [37•, 40] but not in Branley-Bell and Talbot [41] who reported greater impairment in those currently ill. The studies [26•, 27•] conducted in people with a single diagnosis (AN or BN) found that almost 50% of the recruited samples reported ED symptom worsening. These studies [26•, 27•] also highlighted that when ED-related cognitions were evaluated, the impairment was even more common than that of behavioral ED symptoms, occurring in 70%, 80%, and 87% of samples with AN [27•], with BN [26•], or with mixed ED diagnoses [41], respectively. A general worsening of ED-related cognitions was also found in other studies [33•, 35]. Across ED-related aberrant behaviors, physical exercise is worth of a specific mention. Indeed, the possibility to do physical activity was reduced as result of pandemic restrictions: this promoted a widespread increase of anxiety related to inactivity effects [30••, 33•, 37•, 42] with high variability [41] in the amount of physical exercise performed by the patients.

When differences between the main ED diagnoses were investigated, a greater concern about food restriction was found in AN individuals, while more frequent binge eating was detected in the BN ones, suggesting that differences between the ED diagnoses are consistent with diagnostic characteristics [37•]. Differences between people with AN and those with BN were identified also by Castellini et al. [30••] who found that the latter group was more vulnerable to the pandemic restrictions because of their interference with the recovery process. On the other hand, three different research groups [33•, 36, 38•] failed to identify an effect of the diagnosis on the ED symptom trajectory during and after the pandemic lockdown, although the comparisons were conducted between AN individuals and mixed ED groups.

Only two studies [30••, 35] compared ED symptom impairment between people with EDs and healthy controls by employing a longitudinal approach. Castellini et al. [30••] found that the intensity of symptom (i.e., objective binge eating and physical exercise) worsening was significantly greater in patients than in controls. Nisticò et al. [35] found that the severity of ED symptom decreased in the re-opening period following the first lockdown (March to May 2020). This was consistent with the results of another study [33•] adopting a retrospective design and highlighting that in the re-opening period, the ED symptoms returned to the levels seen before the lockdown.

Limitations of these studies need to be acknowledged. First, except for Schlegl et al.’s study [27•], differences between adults and adolescents were not assessed: this precludes the possibility to predict age-related vulnerability to EDs during the COVID-19 pandemic. Second, only two studies adopted a prospective design and included a comparison group, and a few studies included patients with a clinically defined diagnosis. Third, most of the studies did not assess differences across the main ED diagnoses: this limits the possibility to draw transdiagnostic conclusions.

Changes in general psychopathology during the lockdown were assessed in 11 studies. Three of them focused on specific psychopathology variables and revealed an increase in anxiety [14, 37•] and post-traumatic stress symptoms [30••] during the lockdown period. A more comprehensive evaluation of several internalizing symptoms was conducted in the remaining studies [26•, 27•, 28, 29, 33•, 35, 42]. Overall, these studies agreed that people with EDs experienced heightened anxious and depressive symptoms during the lockdown. Schlegl et al. [26•, 27•] identified loneliness, sadness, and inner restlessness as the most pronounced general symptoms in AN and BN people with 70–75% of the assessed patients reporting a deterioration of these symptoms. Remarkably, a longitudinal design was employed in three of these studies [28, 30••, 35]. Furthermore, Monteleone et al. [33•] and Nisticò et al. [35] found that the worsening of internalizing symptoms persisted in the re-opening period that followed the first lockdown in Italy. Furthermore, an increased rate of comorbidity, affective disorders, and suicide risk was observed in children and adolescents recovered for their ED in the first months of the 2020 in comparison to those hospitalized in the same period of the previous year [31]. However, it is worth to outline that only two studies [30••, 35] adopted a prospective design and a comparison with a control group, while only Monteleone et al. [33•] included a large sample of people with EDs.

The quality of life perception was evaluated in three studies through a quantitative assessment [26•, 27•, 32]. Reduced satisfaction was observed in 62% of BN individuals and in 50% of AN people discharged from previous hospital admission [26•, 27•], while no significant change was reported by Machado et al. [32] who evaluated the ED-induced clinical impairment.

Predictors of symptom change during the COVID-19 lockdown period were evaluated in three studies adopting a quantitative design [30••, 34•, 43]. Two of these studies [30••, 43] pointed to low self-directedness, childhood traumatic experiences, and insecure attachment as predictors of the COVID-19-related ED symptoms deterioration [43] and post-traumatic stress symptoms onset [30]. In a large population with mixed ED diagnoses, the path analysis [34•] showed that heightened isolation and fear of contagion predicted ED and general symptom worsening as well as reduced satisfaction with family and with friends’ relationships and reduced perceived social support were associated with ED and general symptoms deterioration, respectively. The quality of the therapeutic relationship was a resilient factor for people with EDs [34•].

The factors related to the COVID-19 psychopathology worsening were assessed in 8 qualitative studies [29, 31, 37•, 38•, 40‐42, 44]. Social restrictions, negative emotions, changes in routine, and thin-related social media messages were described as possible factors contributing to mental health deterioration in most of those studies. Heightened social isolation was reported in all the qualitative studies. Negative emotions included heightened rumination and anxiety [29, 38•, 40]; changes in routine activities encompassed disruption in living situation, which promoted hiding their ED from others and increased pressure from relatives to eat more [40, 41, 44], more free time with boredom and lack of distraction [38•, 40, 41], reduced opportunities to exercise [38•, 41], change in food availability at home [37•, 38•, 41], and increased intentionality and responsibility in planning their own actions [44]. These studies pointed to perceived uncertainty and lack of control as the common mechanisms by which the disruption in routine activities promoted psychopathology deterioration in ED people during the COVID-19 lockdown. However, routine changes [29, 44] and social isolation [29] were sometimes associated with symptom improvement. In this line, useful strategies helping patients to face with COVID-19-related distress were detected and can be divided in two groups: heightened self-care and reduced pressure to engage in social activities or reduced social/work pressure [29, 38•, 40, 41, 44]. The former included increased focus and responsibility for recovery [37•, 44], creating boundaries to look after self [38•], time spent in enjoyable activities/hobbies, or mild physical exercise [26•, 27•, 38•, 42].

The main limitation of the qualitative studies is their small sample sizes. Furthermore, a few studies have evaluated the effects of personality-related characteristics and of theoretically suggested variables (i.e., early abuse) that may contribute to explain the observed variation in psychopathology trajectories.

The main COVID-19-induced treatment change was a reduced access to in-person treatment [26•, 27•, 41, 42, 45]. Schlegl et al. [26•] found that the rate of BN patients receiving face-to-face treatment decreased from 82 to 36% during the lockdown. The parallel increase of online treatment was often perceived as characterized by impairment in the quality of the therapy [26•, 27•, 37•, 38•, 41, 42, 44‐46]. In this line, Lewis et al. [45] also reported that 54% of the ED sample would not recommend the online treatment and 68% would not choose to continue the online therapy. Positive predictors of a good perception of the online therapy were longer illness duration, higher COVID-19-related anxiety, and stronger therapeutic relationship [45]. Fernández-Aranda et al. [39] found that the patients with AN were those reporting lower satisfaction with the online transition. On the other hand, a positive perception of the online therapy was reported in some other studies [29, 44, 46], and there is evidence that patients who interrupted all kinds of treatment were those showing the highest symptom worsening during the lockdown [37•]. In this line, the online treatment allowed patients to maintain a strong and safe therapeutic relationship [38•] and made treatment more accessible for some patients [29, 38•, 46]. In another study, no effect of the treatment delivery strategy (i.e., direct access or telehealth) was found on the psychopathology worsening experienced during the lockdown [34•]. The main barriers identified by the patients regarding the online treatment were perceiving a detached connection with the therapist [38•, 39, 46]; technological difficulties (e.g., low quality of Internet connection or lack of private space) [38•, 46]; and concerning about self-monitoring due to reduction of the therapist’s pressure that patients need to resist the demands of the illness [41, 46]. Overall, the online treatment was described as the best alternative when face-to-face therapy was not available [38•, 41, 46]. Finally, a few studies found that individuals with EDs described their need for mental care as less important than that for physical care related to the COVID-19 infection and perceived themselves as an unjustified burden on the health system [38•, 41, 44, 46]. It is worth mentioning that the comparison between face-to-face and telehealth therapies as well as the treatment successful rates during the pandemic has been not sufficiently explored. Although previously recommended [13•], no evaluation of self-help treatment effectiveness has been provided.