Introduction
The global confirmed case count of Coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), surpassed 775 million as of December 2023 [1]. COVID-19 is now considered a multisystemic condition, which can lead to a broad spectrum of symptoms and long-term sequelae that may persist after remission of the infection up to one year and – in a proportion of cases – is still present even after two years [Reference Mehta, Bhandari, Raut, Kacimi and Huy2–Reference Ma, Deng, Liu, Du, Liu and Liu9]. Different terms and definitions have been proposed for this persisting condition, such as long-Covid or post-COVID-19 condition (PCC), which was described by a Delphi consensus as a condition occurring in subjects with a history of SARS-CoV-2 infection, usually 3 months from the infection, with symptoms that last for at least 2 months [Reference Soriano, Murthy, Marshall, Relan and Diaz10]. Long-term sequelae include hyposmia/anosmia, fatigue, dyspnea, and neuropsychiatric conditions, such as depression, anxiety, and cognitive impairment (CI) [Reference Coleman, Casiraghi, Blau, Chan, Haendel and Laraway11, Reference Stewart, Wasserman and Appelbaum12]. Among these, CI has been reported as one of the most prevalent and burdensome consequences of the infection, affecting over 20% of patients who contracted COVID-19 [Reference Fernandez-de-Las-Peñas, Notarte, Macasaet, Velasco, Catahay and Ver4].
Although deficits in different cognitive domains have been described, those more frequently reported involve executive functions, speed of processing, attention, and working memory, which may lead to substantial detriments to the quality of life and daily functioning of individuals [Reference Perrottelli, Sansone, Giordano, Caporusso, Giuliani and Melillo13–Reference Crivelli, Palmer, Calandri, Guekht, Beghi and Carroll15].
Despite the clinical relevance of post-COVID CI, several research questions remain open. First, its underlying pathophysiology is still unclear; several hypotheses have been proposed, including the direct infiltration of the virus in the central nervous system cells, or an indirect brain damage due to different mechanisms such as abnormal immunological response, inflammation, vasculopathy, or hypoxemia caused by the infection [Reference Leng, Shah, Ahmad, Premraj, Wildi and Li Bassi16, Reference García-Grimshaw, Chirino-Pérez, Flores-Silva, Valdés-Ferrer, Vargas-Martínez and Jiménez-Ávila17]. Second, given the novelty of this condition, its natural course and long-term outcome are still unknown; according to recent evidence, post-COVID CI can show different trajectories, probably related to different pathogenesis: some individuals may experience early short-term CI that resolves within weeks or months, while in others it may persist even for two years [Reference Fernandez-de-Las-Peñas, Notarte, Macasaet, Velasco, Catahay and Ver4, Reference Latifi and Flegr6, Reference Gonçalves, Aliberti, Bertola, Avelino-Silva, Dias and Apolinario18–Reference Liu, Chen, Wang, Wang, Jiang and Yang20]; in other cases, CI may emerge long after the acute infection has remitted, since the risk of developing post-COVID CI has been found still present even after 2 years from the acute infection [Reference Taquet, Sillett, Zhu, Mendel, Camplisson and Dercon21]. Third, unlike other common symptoms of PCC, such as dyspnea and depression, there are no established and effective treatments for post-COVID CI [Reference Quan, Wang, Gong, Wang, Li and Jia22].
The latest WHO recommendations [23] suggest the combination of multiple cognitive rehabilitation strategies including both restorative (e.g., repeated standardized cognitive exercises) and compensatory interventions (e.g., skills training on self-management strategies such as planning and prioritizing activities or simplifying large tasks into smaller components). However, the WHO guidelines identified no randomized or non-randomized controlled trials specifically testing the rehabilitation interventions for post-COVID CI; therefore, they were based on the evidence available for diverse populations, such as individuals with CI following traumatic brain injury and stroke-related CI [Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe and Langenbahn24]. Although findings collected in other populations affected by CI may still suggest valuable treatment strategies and clinical directions, there is a critical need for evidence-based therapeutic options specifically tailored to post-COVID CI. In fact, it is crucial to take into account the complexity of this condition, wherein patients may have CIs in one or more cognitive domains with a wide range of severity and duration [Reference Möller, Borg, Janson, Lerm, Normark and Niward25]. Furthermore, comorbidity with somatic or psychiatric conditions in patients affected by post-COVID CI may exacerbate these deficits, potentially influencing the extent and nature of the CI and the efficacy of treatments. For instance, some studies addressing post-COVID CI showed that patients with PCC are often easily fatigued, which might affect a patient’s tolerance to the cognitive rehabilitation training [Reference Vu, Fitzpatrick, Cope, Bertolli, Sotoodehnia and West26].
Previous reviews that focused on treatment strategies for PCC, even when including the management of CI, did not provide a comprehensive overview on the topic due to the following limitations: in some cases, they addressed the efficacy of a single type of intervention (e.g., noninvasive brain stimulation [NIBS] [Reference Linnhoff, Koehler, Haghikia and Zaehle27]), retrieved studies in which cognitive deficits were not assessed [Reference Fugazzaro, Contri, Esseroukh, Kaleci, Croci and Massari28, Reference Chandan, Brown, Simms-Williams, Bashir, Camaradou and Heining29], or included only study protocols that have not provided yet results on the efficacy of treatments [Reference Hawke, Nguyen, Ski, Thompson, Ma and Castle30, Reference Fawzy, Abou Shaar, Taha, Arabi, Sabbah and Alkodaymi31]. Two recent qualitative reviews that specifically addressed the current research evidence on management of post-COVID CI [Reference Quan, Wang, Gong, Wang, Li and Jia22, Reference Frontera, Guekht, Allegri, Ashraf, Baykan and Crivelli32] highlighted the scarcity of data, as they only retrieved, respectively, three and four clinical trials specifically carried out in subjects with post-COVID CI. The authors concluded that different treatment options, including lifestyle interventions (e.g., sleep management, physical activity, and dietary interventions), cognitive training programs, and possibly anti-dementia drugs, should further be considered and tested. However, it should be noticed that these reviews did not report a detailed characterization of the studies in terms of inclusion and exclusion criteria, clinical and demographical variables of the samples, and employed cognitive tests. In addition, they also included studies in which differences in cognitive functioning before and after treatments were not assessed through objective measures [Reference Fesharaki-Zadeh, Lowe and Arnsten33, Reference De Luca, Camaioni, Marra, Salzano, Carriere and Ricciardi34]; this is a crucial aspect given the discrepancy between subjective and objective cognitive assessments reported in patients with PCC [Reference Ceban, Ling, Lui, Lee, Gill and Teopiz14].
In light of these observations and of the existing gaps in the current literature, the aims of the present systematic review are 1) providing an updated overview of the existing evidence on the efficacy of treatments implemented to improve cognitive functioning, assessed with objective assessments, in individuals suffering from post-COVID CI; 2) identifying potential limitations of the current evidence and providing methodological recommendations for further studies on the topic.
Methods
The current systematic review was performed in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann and Mulrow35]. A systematic literature search was conducted for studies published up to December 31, 2023, using three databases (PubMed, Scopus, and ProQuest). The following combination of search terms was used: (COVID-19 OR SARS-COV-2 OR 2019-nCoV OR “long covid” OR “persistent covid” OR “post covid” OR “long-haul covid” OR “Post-covid brain fog”) AND (cognition OR neurocognition OR “cognitive deficit” OR “cognitive impairment”) AND (neuromodulation OR intervention OR training OR stimulation OR remediation OR management OR treatment OR therap* OR rehabilitation). Duplicates from the combination of the three databases were excluded. Three investigators independently screened all articles for eligibility based on titles and abstracts, and then the full text of the selected articles was reviewed. Discrepancies in the selection of suitable articles were discussed by all authors and were resolved through discussion and consensus.
Inclusion criteria for the articles were established prior to the article review based on PICOS framework (Table 1).
Inclusion criteria based on PICOS framework
Exclusion criteria were 1) articles published before the pandemic; 2) studies considering non-interventional trials, or trials on interventions aimed at preventing post-COVID CI; 3) studies addressing not relevant populations (e.g., individuals with no history of acute infection or with a diagnosis of cognitive deficit preceding the COVID-19 infection); 4) studies testing the efficacy of treatments based only on self-report measures, given the evidence that these assessments are not as reliable and standardized as objective measures [23]; 5) preclinical and nonhuman studies; and 6) articles with unavailable full text in English.
No article was excluded based on study design, sample size, demographic characteristics of the subjects, applied definition or duration of the PCC, or time elapsed since the acute infection.
For the articles meeting inclusion criteria, data extraction included authors; year of publication; design of the study; sample size; demographic and clinical characteristics of the patients (including age, gender, years of education, duration of acute illness, duration of the PCC, history of hospitalization, or intensive care unit management); inclusion criteria; applied definition of PCC; tests used for assessing cognitive functioning; type, duration, and description of the employed intervention; assessment of confounding factors such as depressive and anxiety symptoms; statistical analysis; and main findings in relation to cognitive outcomes.
The methodological quality of the included studies was assessed using the Joana Briggs Institute Critical Appraisal tools [36].
Results
Search results
The combined outcome of the three database results included 11468 records (PubMed: 7489; Scopus: 2240; ProQuest: 1739). Details on screening, eligibility assessment, and exclusion criteria are reported in Figure 1.
PRISMA flow diagram.
A total of 29 articles met our criteria and were included in the review. Nine studies were randomized controlled trials (RCTs) [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37–Reference Tanashyan, Morozova, Raskurazhev and Kuznetsova45], two non-RCTs [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46, Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47], nine noncontrolled trials [Reference Sasaki, Yamatoku, Tsuchida, Sato and Yamaguchi48–Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56], and nine were case reports [Reference Łuckoś, Cielebąk and Kamiński57–Reference Chang, Chen, Chen and Tsai61] or case series [Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62–Reference Robbins, Gonevski, Clark, Baitule, Sharma and Magar65].
In relation to the type of tested interventions, our search retrieved two articles on cognitive rehabilitation [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46, Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49], six on NIBS (one on transcranial alternating current stimulation [tACS] [Reference Sabel, Zhou, Huber, Schmidt, Sabel and Gonschorek59], two on transcranial direct current stimulation [tDCS] [Reference Oliver-Mas, Delgado-Alonso, Delgado-Álvarez, Díez-Cirarda, Cuevas and Fernández-Romero41, Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62], and three on transcranial magnetic stimulation [TMS] [Reference Sasaki, Yamatoku, Tsuchida, Sato and Yamaguchi48, Reference Chang, Chen, Chen and Tsai61, Reference Noda, Sato, Fujii, Nagano, Iwasa and Hirahata64]), seven on pharmacological or nutraceutical compounds [Reference Versace, Ortelli, Dezi, Ferrazzoli, Alibardi and Bonini40, Reference Momtazmanesh, Ansari, Izadi, Shobeiri, Vatankhah and Seifi42, Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43, Reference Tanashyan, Morozova, Raskurazhev and Kuznetsova45, Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50, Reference Tobinick, Spengler, Ignatowski, Wassel and Laborde60, Reference Zifko, Yacob, Braun and Dietz63], three on hyperbaric oxygen therapy [HBOT] [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37, Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58, Reference Robbins, Gonevski, Clark, Baitule, Sharma and Magar65], one on neurofeedback training (NFT) intervention [Reference Łuckoś, Cielebąk and Kamiński57], one on photobiomodulation [Reference Bowen and Arany38], one on meditation [Reference Hausswirth, Schmit, Rougier and Coste39], and eight on multidisciplinary rehabilitation interventions [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47, Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51–Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56].
Inclusion criteria and clinical and demographic features of the enrolled subjects
Inclusion/exclusion criteria, when reported (22 studies), were heterogeneous with respect to the clinical condition required to be included. In most studies, the requirement for inclusion was the presence of self-reported post-COVID CI or other post-COVID symptoms; only in four studies, inclusion criteria listed the presence of post-COVID CI assessed by means of objective tests; in other four studies, only a history of hospitalization due to COVID-19 infection was listed as an inclusion criterion. Also with respect to the minimal duration of the PCC, inclusion criteria were quite heterogeneous (in the majority of studies 12 weeks, in others a duration ranging from 1 week to 6 months, or no temporal criterion at all). A detailed description of inclusion/exclusion criteria for all included studies is provided in the Supplementary Material (results section, paragraph 1, and Supplementary Table S1).
Demographic and clinical features of the subjects included in all studies are reported in Table 2 and Supplementary Table S1.
Demographic and clinical features of retrieved samples
a Data reported in 4 studies only [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37, Reference Versace, Ortelli, Dezi, Ferrazzoli, Alibardi and Bonini40, Reference Oliver-Mas, Delgado-Alonso, Delgado-Álvarez, Díez-Cirarda, Cuevas and Fernández-Romero41, Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62].
b Data missing in 12 studies [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37–Reference Hausswirth, Schmit, Rougier and Coste39, Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43–Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46, Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49, Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50, Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53, Reference Amini, Vaezmousavi and Shirvani54, Reference Robbins, Gonevski, Clark, Baitule, Sharma and Magar65].
c Seven studies included exclusively subjects with a history of hospitalization [Reference Momtazmanesh, Ansari, Izadi, Shobeiri, Vatankhah and Seifi42, Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46, Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47, Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50, Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56, Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58, Reference Tobinick, Spengler, Ignatowski, Wassel and Laborde60].
d Information was missing in 9 studies [Reference Bowen and Arany38, Reference Hausswirth, Schmit, Rougier and Coste39, Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43–Reference Tanashyan, Morozova, Raskurazhev and Kuznetsova45, Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49, Reference Amini, Vaezmousavi and Shirvani54, Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62, Reference Robbins, Gonevski, Clark, Baitule, Sharma and Magar65].
CI assessment methodologies
High heterogeneity was observed among studies with regard to the tests and batteries employed for cognitive assessment. Details on the applied assessment tools in all included studies are reported in Table 3.
Cognitive impairment assessment methodologies
Abbreviations: BACS, Brief Assessment of Cognition in Schizophrenia scale; CAB, Cognitive Assessment Battery; CR, choice reaction time; MMSE, Mini Mental State Examination; MoCa, Montreal Cognitive Assessment; PC, Pattern Comparison Task; TAP, Test of Attentional Performance; TMT-A/B = Trail Making Test A and B; WAIS-IV, Wechsler Adult Intelligence Scale-IV; WMS-R: Wechsler Memory Scale-Revised.
Methodological quality of the included studies
Applying the Critical Appraisal tools of the Joana Briggs Institute [36], fourteen of the included studies were categorized as being of good methodological quality [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37, Reference Oliver-Mas, Delgado-Alonso, Delgado-Álvarez, Díez-Cirarda, Cuevas and Fernández-Romero41, Reference Momtazmanesh, Ansari, Izadi, Shobeiri, Vatankhah and Seifi42, Reference Tanashyan, Morozova, Raskurazhev and Kuznetsova45–Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50, Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52, Reference Łuckoś, Cielebąk and Kamiński57, Reference Tobinick, Spengler, Ignatowski, Wassel and Laborde60, Reference Chang, Chen, Chen and Tsai61, Reference Noda, Sato, Fujii, Nagano, Iwasa and Hirahata64], thirteen as being of average methodological quality [Reference Bowen and Arany38, Reference Versace, Ortelli, Dezi, Ferrazzoli, Alibardi and Bonini40, Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43, Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51, Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53–Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56, Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58, Reference Sabel, Zhou, Huber, Schmidt, Sabel and Gonschorek59, Reference Zifko, Yacob, Braun and Dietz63, Reference Robbins, Gonevski, Clark, Baitule, Sharma and Magar65], and two as being of poor methodological quality [Reference Hausswirth, Schmit, Rougier and Coste39, Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62]. The main factors affecting the methodological quality were the poor characterization of experimental samples, the inadequate sensitivity of assessment tools, and the lack of control for confounding variables. Data on the risk of bias are reported in the Supplementary Material file (results section, paragraph 2).
Results on the efficacy of the interventions
Cognitive remediation interventions
Cognitive remediation (CR) is defined as a behavioral training intervention targeting deficits in attention, memory, executive function, social cognition, or metacognition, using scientific principles of learning to improve cognitive skills and functional outcomes [Reference Bowie, Bell, Fiszdon, Johannesen, Lindenmayer and McGurk66].
A non-RCT [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46] tested in 15 patients the efficacy of the CogPack CR program consisting of 6 weekly sessions and tailored to the patients’ cognitive profiles, as assessed through the Brief Assessment of Cognition in Schizophrenia (BACS) scale. A significant improvement in global cognitive functioning was observed in the group of patients with respect to the control group [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46]. Depressive symptomatology did not affect the efficacy of the CR intervention on global cognitive improvement [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46].
A noncontrolled feasibility pilot study [Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49], testing an 8-week program of digital cognitive training, reported significant improvement in attention, memory, coordination, perception, and reasoning in a sample of 73 post-COVID individuals.
Noninvasive brain stimulation
NIBS refers to a range of techniques aimed at modulating brain electrical activity in targeted cortical areas, stimulating neuronal excitability, neural plasticity, and changes in connectivity patterns [Reference Elyamany, Leicht, Herrmann and Mulert67–Reference Siegert, Diedrich and Antal69]. In the last decades, a range of different NIBS techniques, such as tACS, TMS, and tDCS, have been tested for the treatment of CI in different psychiatric and neurological conditions, as well as in healthy cognitive aging [Reference Booth, Taylor, Brown and Pobric70–Reference Trifilio, Shortell, Olshan, O’Neal, Coyne and Lamb78].
Three studies tested the efficacy of TMS [Reference Sasaki, Yamatoku, Tsuchida, Sato and Yamaguchi48, Reference Chang, Chen, Chen and Tsai61, Reference Noda, Sato, Fujii, Nagano, Iwasa and Hirahata64]. The first was a pilot case series study in which 20 TMS sessions of intermittent theta-burst stimulation (iTBS) applied to the left dorsolateral prefrontal cortex (DLPFC) and to the right lateral orbitofrontal cortex (LOFC) led to an improvement in executive functions in a sample of 23 individuals [Reference Noda, Sato, Fujii, Nagano, Iwasa and Hirahata64]. A second noncontrolled study reported a significant post-intervention improvement in overall cognitive performance in 12 subjects receiving 10 TMS sessions applied to the frontal and occipital cerebral regions [Reference Sasaki, Yamatoku, Tsuchida, Sato and Yamaguchi48]. This study also reported that subjects showed an increase in blood flow in the frontal and occipital cortical areas, as compared to the pre-intervention assessment. A case report testing the effects of continuous accelerated theta-burst TMS applied to the right DLPFC, followed by intermittent accelerated theta-burst TMS applied to the left DLPFC, showed significant improvements in memory in a 30-year-old woman [Reference Chang, Chen, Chen and Tsai61].
The efficacy of tDCS was tested in two studies. One randomized sham-controlled trial reported no significant effects on a task assessing executive functioning and processing speed in 23 subjects receiving eight tDCS sessions [Reference Oliver-Mas, Delgado-Alonso, Delgado-Álvarez, Díez-Cirarda, Cuevas and Fernández-Romero41]. A case series on patients receiving tDCS combined with online cognitive training showed improvement in processing speed, verbal learning, and memory in the four included subjects [Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62].
As to tACS, one case report showed that 13 tACS sessions led to an improvement in attention, executive functions, verbal learning, and verbal memory, but not in working memory in a 40-year-old woman [Reference Sabel, Zhou, Huber, Schmidt, Sabel and Gonschorek59].
Pharmacological and nutraceutical interventions
Seven trials on potential pharmacological intervention for post-COVID CI were completed, while others, mainly testing the efficacy of anti-dementia drugs [Reference Quan, Wang, Gong, Wang, Li and Jia22, Reference Bonilla, Peluso, Rodgers, Aberg, Patterson and Tamburro79], are currently ongoing.
One RCT [Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43] tested the efficacy of donepezil chlorhydrate, a cholinesterase inhibitor which has shown neuroprotective and anti-inflammatory effects and is approved for the treatment of Alzheimer’s disease [Reference Kim, Kandiah, Hsu, Suthisisang, Udommongkol and Dash80, Reference Liu, Zhu, Fu and Yang81]. The study reported no differences between the intervention group (N = 15) and the control group (N = 10) on the overall score in memory performance, assessed through a test battery, after 4 and 12 weeks of treatment [Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43].
A second RCT tested the efficacy of the coordination complex between ethylmethylhydroxypyridine and trimethylhydrosinium propionate with succinate acid anion (CCSA), a newly marketed compound with potential neuroprotective effects, in 15 subjects as compared to 15 controls receiving placebo, and reported a significant improvement in cognitive performance only in the intervention group [Reference Tanashyan, Morozova, Raskurazhev and Kuznetsova45].
The efficacy of famotidine, a selective histamine H2 receptor antagonist, was tested in a RCT versus placebo, reporting that this drug significantly improved global cognition in the treatment group (N = 25), with no significant correlation with the improvements in depression and anxiety symptoms [Reference Momtazmanesh, Ansari, Izadi, Shobeiri, Vatankhah and Seifi42].
Another RCT tested the efficacy versus placebo of ultramicronized palmitoylethanolamide (PEA), an endocannabinoid drug which has been tested in several neurological and neurodegenerative conditions for its modulating role in neuroinflammation and synaptic neurotransmission [Reference Assogna, Di Lorenzo, Martorana and Koch82, Reference Cordaro, Cuzzocrea and Crupi83] combined with luteolin. The analysis showed no significant improvements in global cognitive functioning in both the PEA and control groups (n = 34) [Reference Versace, Ortelli, Dezi, Ferrazzoli, Alibardi and Bonini40].
A noncontrolled study investigating the effects of a nutraceutical including different compounds such as L-theanine, vitamin B6, vitamin D, biotin, folic acid, and vitamin B12 found an improvement of the global MoCA scores, particularly within the attention and executive functioning domains in a sample of 40 elderly patients [Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50].
A case series study testing the benefits of Ginkgo biloba extract EGb 761 reported improvements in global cognitive functioning in five patients [Reference Zifko, Yacob, Braun and Dietz63].
Finally, a case report of a patient found an improvement in global cognitive functioning and more specifically in the executive functions and verbal fluency domains, following the administration of perispinal etanercept, a tumor necrosis factor inhibitor [Reference Tobinick, Spengler, Ignatowski, Wassel and Laborde60].
Hyperbaric oxygen therapy
HBOT, the therapeutic administration of 100% oxygen at environmental pressures greater than one atmosphere, has been recently tested off-label for the treatment of CI associated with neurological disorders [Reference Gottfried, Schottlender and Ashery84, Reference Marcinkowska, Mankowska, Kot and Winklewski85], particularly in the case of traumatic brain injury [Reference Alashram, Padua, Romagnoli and Annino86] and vascular dementia [Reference You, Li, Xiong, Yan, Li and Yan87].
In relation to post-COVID CI, a randomized sham-controlled trial tested the efficacy of HBOT [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37] reporting significant improvements in global cognitive functioning, attention, and executive functions after HBOT in 37 subjects, as compared to the control group. Furthermore, the recovery of these cognitive domains was explained by neuroimaging data that suggested a restoration of functional connectivity between the frontoparietal, default mode and salience networks after the intervention [Reference Catalogna, Sasson, Hadanny, Parag, Zilberman-Itskovich and Efrati88].
A case series reported that global cognition, executive functions, attention, processing speed, and verbal fluency improved in a sample of 10 patients following 10 sessions of HBOT [Reference Robbins, Gonevski, Clark, Baitule, Sharma and Magar65].
Furthermore, a case report on a single patient showed that 20 sessions of HBOT significantly improved both pulmonary capacity and global cognitive functioning [Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58]. The improvements were associated with an increase in brain perfusion assessed through MRI [Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58].
Neurofeedback training
NFT is an electroencephalogram-based biofeedback technique aimed at training self-regulation of neurophysiological states to reach specifically targeted electroencephalography (EEG) signals [Reference Loriette, Ziane and Ben Hamed89, Reference Orendáčová and Kvašňák90].
One case report showed the effects of combining EEG-based NFT and goal-oriented cognitive training in the treatment of post-COVID CI [Reference Łuckoś, Cielebąk and Kamiński57]. The study involved a Sensory-Motor Rhythm and theta/beta training interventions showing improvements in attention, visual learning, memory, and executive functioning after 30 sessions during a period of 15 weeks [Reference Łuckoś, Cielebąk and Kamiński57].
Other interventions
A pilot study compared the efficacy of either transcranial or whole-body photo-biomodulation (PBM), a technique implementing ultraviolet rays with anti-inflammatory properties, reporting that both interventions were associated with significant improvement in cognitive performance in a sample of 14 individuals [Reference Bowen and Arany38].
Furthermore, a RCT tested the efficacy of a meditation program and found an improvement in processing speed in a sample of 17 individuals [Reference Hausswirth, Schmit, Rougier and Coste39].
Multidisciplinary interventions
Among studies included in this review, the highest number (N = 8) tested the efficacy of multidisciplinary interventions on CI. Of these, six [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51–Reference Amini, Vaezmousavi and Shirvani54, Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56] involved cognitive rehabilitation (five with restorative interventions [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52–Reference Amini, Vaezmousavi and Shirvani54, Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56] and one with compensatory interventions [Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51]).
A non-RCT testing the efficacy of in-person and supervised physical exercise (PE) together with dietary modules reported that the intervention group (N = 21) significantly improved in global cognitive functioning, as compared to the usual care group (N = 23) [Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47]. However, an RCT testing a remote-based digital multidisciplinary intervention including physical and cognitive exercise modules, as well as dietary and sleep hygiene recommendations, found no significant differences in cognitive performance improvements between subjects assigned to the multidisciplinary intervention (N = 52) and the control group (N = 48) [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44].
The remaining six studies were all noncontrolled.
One study investigated the efficacy of a multidisciplinary intervention that included individual and group-based cognitive behavioral therapy (CBT) [Reference Boschloo, Hieronymus and Cuijpers91–Reference Hedman-Lagerlöf, Carlbring, Svärdman, Riper, Cuijpers and Andersson95], individual and group-based cognitive training, and PE training during a 5-week stay in a rehabilitation facility in 80 subjects [Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52]. No significant improvements in memory and attention performance were observed at discharge [Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52].
Another study investigated the efficacy of an 8-week multidisciplinary intervention comprising both physical rehabilitation and digital CR interventions and reported improvements in verbal fluency, verbal learning, and memory, in a sample of 40 individuals [Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53].
The efficacy of a 30-day-long multidisciplinary intervention comprising both CR and physiotherapy was tested in a study that found significant improvements in global cognitive functioning, particularly in the domains of attention, abstract reasoning, memory, and visuospatial orientation in a sample of 64 individuals [Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56].
A cohort study on the efficacy of aerobic exercise, combined with educational sessions on fatigue, memory and concentration, and sleeping hygiene, reported a significant improvement in global cognitive functioning in a sample of 32 subjects [Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51].
A study testing the efficacy of a rehabilitation intervention consisting of physical and cognitive training reported improvements in attention, calculation, memory, and global functioning in a sample of 42 elderly men (65–80 years) recovering from hospitalization due to COVID-19 infection [Reference Amini, Vaezmousavi and Shirvani54].
Finally, a study testing a 3-month multidisciplinary intervention including physical training and psychological, dietary, and occupational assistance reported an improvement in global functioning scores in a sample of 22 subjects [Reference Everaerts, Heyns, Langer, Beyens, Hermans and Troosters55].
Discussion
Due to the dearth of evidence on the treatment of post-COVID CI, current guidelines are mainly based on the management of CI in other conditions [23, Reference Möller, Borg, Janson, Lerm, Normark and Niward25]. The present review retrieved 29 studies testing the efficacy of a broad range of interventions for post-COVID CI.
Evidence in six studies [Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47, Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51, Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53–Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56] included in the present review, of either good [Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47] or average [Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51, Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53–Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56] methodological quality and carried out in relatively large samples [Reference Asimakos, Spetsioti, Mavronasou, Gounopoulos, Siousioura and Dima47, Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53, Reference Amini, Vaezmousavi and Shirvani54, Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56], shows that multidisciplinary interventions encompassing different components, such as physical rehabilitation interventions, cognitive training, and support for lifestyle modifications (e.g., dietary recommendations and sleep hygiene), can significantly improve post-COVID CI. It could be hypothesized that multidisciplinary approaches target the different components of PCC, which, particularly in some cases (subjects experiencing also sleep disturbances, mood and anxiety symptoms, and physical fatigue), may impact cognitive performance, as well as overall functional outcomes and quality of life. However, two studies testing multidisciplinary interventions reported no significant benefits [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52]. In one of them [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44], which tested an app-based multidisciplinary intervention, the results might have been hindered by very low user compliance [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44]. The advantages of app- and home-based multidisciplinary interventions are the low costs and their easy dissemination; however, their acceptability and feasibility need to be investigated in further trials. In the second study [Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52], a 5-week multidisciplinary intervention in a rehabilitation clinic setting did not significantly improve attention and working memory, despite remission of depressive symptomatology. However, this study did not assess other cognitive domains or global functioning at the post-intervention visit, which hinders comparison with other studies [Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52]. Consistently with this finding, another study [Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43] found no significant improvement in either attention or working memory, evaluated with specific individual tests. Therefore, further studies should investigate the effectiveness of multidisciplinary interventions using comprehensive cognitive batteries.
It is worth noticing that all the multidisciplinary interventions included PE [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51–Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56], which is considered a cornerstone of the rehabilitation from COVID-19 infection; in one study [Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51], the sole combination of aerobic exercise and education on compensatory CR strategies (e.g., pacing strategies and management of daily activities) was found to improve global cognitive performance [Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51]. Besides targeting physical fatigue, physical activity may improve cognition through its effects on hormonal and cardiovascular systems, as well as through its modulatory effects on neuroplasticity and inflammatory cascades [Reference Farì, Lunetti, Pignatelli, Raele, Cera and Mintrone96].
Six of the studies on multidisciplinary interventions involved cognitive rehabilitation programs, either restorative – in the majority of cases – or compensatory [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51–Reference Amini, Vaezmousavi and Shirvani54, Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56]. In addition, the application of cognitive training interventions alone was also reported to effectively improve cognitive functioning in two other studies [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46, Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49]. Digital [Reference Albu, Rivas Zozaya, Murillo, García-Molina, Figueroa Chacón and Kumru53, Reference Torous, Myrick and Aguilera97] or virtual reality [Reference Groenveld, Achttien, Smits, de Vries, van Heerde and Staal98, Reference Ventura99] CR interventions can be home based and self-administered [Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49], thus providing a cost-effective intervention, which can be tailored to target specific cognitive domains based on patients’ characteristics [Reference Espie, Firth and Torous100–Reference Torous103]. Further RCT trials are needed in order to gain stronger evidence on the efficacy of these interventions and identify the most effective programs (e.g., CR exercises [Reference Samper-Pardo, León-Herrera, Oliván-Blázquez, Méndez-López, Domínguez-García and Sánchez-Recio44, Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46, Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49, Reference Kupferschmitt, Jöbges, Randerath, Hinterberger, Loew and Köllner52–Reference Amini, Vaezmousavi and Shirvani54, Reference Rabaiotti, Ciracì, Donelli, Oggioni, Rizzi and Savi56] or compensatory strategies [Reference Daynes, Gerlis, Chaplin, Gardiner and Singh51]) and the best methods of administration (e.g., home-based [Reference Duñabeitia, Mera, Baro, Jadad-Garcia and Jadad49] or supervised face-to-face interventions [Reference Palladini, Bravi, Colombo, Caselani, Di Pasquasio and D’Orsi46]).
Overall, our results also support the WHO recommendations on the management of post-COVID CI, which advise the use of a combination of restorative and compensatory cognitive rehabilitation interventions; the retrieved evidence, however, additionally suggests that these interventions may be more effective in the context of multidisciplinary approaches, when complemented by interventions such as PE, targeting the different symptoms that individuals may experience and that may contribute to cognitive and functional impairment.
In relation to NIBS techniques, their effects on neuronal responsiveness, long-term potentiation, and neurovascular modulation might improve the cortical hypometabolism described in the PCC [Reference Linnhoff, Koehler, Haghikia and Zaehle27, Reference Sabel, Zhou, Huber, Schmidt, Sabel and Gonschorek59]; intermittent theta-burst TMS stimulation, through its facilitatory role on the DLPFC and by enhancing theta-gamma coupling involved in cognitive functions, has been suggested to improve cognitive functioning and stimulate neuroplasticity [Reference Noda, Sato, Fujii, Nagano, Iwasa and Hirahata64]. The retrieved evidence, however, was limited to one pilot study, one case series, and case reports for TMS [Reference Sasaki, Yamatoku, Tsuchida, Sato and Yamaguchi48, Reference Chang, Chen, Chen and Tsai61, Reference Noda, Sato, Fujii, Nagano, Iwasa and Hirahata64], while only one case series and a single sham-controlled RCT were available for tDCS [Reference Oliver-Mas, Delgado-Alonso, Delgado-Álvarez, Díez-Cirarda, Cuevas and Fernández-Romero41, Reference Cavendish, Lima, Bertola, Charvet, Bikson and Brunoni62], and the latter did not find positive effects on cognition [Reference Oliver-Mas, Delgado-Alonso, Delgado-Álvarez, Díez-Cirarda, Cuevas and Fernández-Romero41]. This lack of efficacy might be explained by the fact that NIBS interventions have been reported to improve only specific cognitive domains, particularly working memory in the case of TMS, [Reference Booth, Taylor, Brown and Pobric70] and attention/vigilance and working memory for tDCS [Reference Nissim, McAfee, Edwards, Prato, Lin and Lu72, Reference Begemann, Brand, Ćurčić-Blake, Aleman and Sommer73]. Therefore, further studies using more comprehensive test batteries are needed to investigate their efficacy and whether specific post-COVID cognitive phenotypes can benefit from the use of these techniques.
Two case studies and one randomized sham-controlled trial reported evidence in favor of the efficacy of HBOT. Given its potential mechanisms on mitochondrial activity, neurogenesis, and angiogenesis [Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58], HBOT could be useful particularly in the case of patients who had suffered from hypoxia during acute COVID-19 illness and required treatment with high-flow oxygen [Reference Bhaiyat, Sasson, Wang, Khairy, Ginzarly and Qureshi58]. In the retrieved sham-controlled trial [Reference Zilberman-Itskovich, Catalogna, Sasson, Elman-Shina, Hadanny and Lang37], HBOT was associated with functional connectivity and white matter modifications, and the authors hypothesized that structural and functional connectivity analysis may represent both a treatment eligibility and response monitoring tool [Reference Catalogna, Sasson, Hadanny, Parag, Zilberman-Itskovich and Efrati88].
Evidence was sparse and limited to single studies for other interventions, including the pharmacological ones [Reference Versace, Ortelli, Dezi, Ferrazzoli, Alibardi and Bonini40, Reference Momtazmanesh, Ansari, Izadi, Shobeiri, Vatankhah and Seifi42, Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43, Reference Tanashyan, Morozova, Raskurazhev and Kuznetsova45, Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50, Reference Tobinick, Spengler, Ignatowski, Wassel and Laborde60, Reference Zifko, Yacob, Braun and Dietz63, Reference Lenze, Reiersen and Santosh104]. Given the increased risk of developing dementia associated with the PCC [Reference Olivera, Sáez, Carniglia, Caruso, Lasaga and Durand105, Reference Matsui, Mitsuma, Nagata, Matsushita and Asahi106] and the similarities between post-COVID CI and dementia [Reference Quan, Wang, Gong, Wang, Li and Jia22], the early start of anti-dementia drugs could be hypothesized for subsets of patients at a particularly higher risk of CI related to PCC, such as older subjects and those who experienced greater COVID-19 severity symptoms. However, the only retrieved trial testing the efficacy of an anti-dementia drug in individuals with PCC did not support its application [Reference Pooladgar, Sakhabakhsh, Soleiman-Meigooni, Taghva, Nasiri and Darazam43].
Overall, the evidence collected up to this date by the current review is quite hard to interpret and summarize due to several methodological factors. First, the included studies showed heterogeneity in relation to inclusion and exclusion criteria, particularly regarding the applied definition of PCC and the recruitment time range from acute illness. Such heterogeneity, which was found to be diffuse in the current literature [Reference Quan, Wang, Gong, Wang, Li and Jia22, Reference Cabrera Martimbianco, Pacheco, Bagattini  and Riera107–Reference Pan and Pareek109], is particularly crucial, given the current uncertainty over the longitudinal trajectory of post-COVID-19 CI and consequentially the identification of the optimal time window for treatment [Reference Soriano, Murthy, Marshall, Relan and Diaz10, Reference Möller, Borg, Janson, Lerm, Normark and Niward25, Reference Shah, Hillman, Playford and Hishmeh110].
Furthermore, the applied assessment tools for CI were also variable, with most of the studies including only screening tools, such as MoCA or MMSE, which might have inadequate sensitivity to assess improvements in cognitive functioning, as compared to neuropsychological test batteries that provide in-depth characterization of cognitive domains [Reference Bove, Fogacci, Quattrocchi, Veronesi and Cicero50, Reference Becker, Lin, Doernberg, Stone, Navis and Festa111]. Another factor to be taken into account is that many of the included studies recruited subjects based on self-reported CI, but employed objective assessment tools for the pre–post evaluation of treatment efficacy [Reference Miskowiak, Johnsen, Sattler, Nielsen, Kunalan and Rungby112–Reference Blackmon, Day, Powers, Bosch, Prabhakaran and Woolston115]. However, a meta-analysis focusing on COVID-19 patients showed that studies applying objective assessment tests reported significantly greater rates of individuals with CI in comparison with those employing self-reporting tools (36% versus 18%, respectively), which might suggest that the population who might benefit from treatments to improve CI might be much larger than the number of individuals with subjective complaints and may thus be not adequately represented in samples recruited through self-reported CI as the main inclusion criterion [Reference Ceban, Ling, Lui, Lee, Gill and Teopiz14].
In addition, all studies lacked a long-term longitudinal design to analyze the stability over time of the reported improvements [Reference Becker, Lin, Doernberg, Stone, Navis and Festa111]. Furthermore, the absence of control groups in many studies, together with the long duration of the intervention protocols, does not allow to control for placebo effects and for the potential spontaneous remission of this condition, respectively. Finally, experimental samples were heterogeneous and under-characterized in relation to several parameters related to potential confounding factors, such as the symptom severity of the acute infection, the occurrence of hospitalization, and the presence of comorbidities during the PCC, such as psychiatric conditions [Reference Wang, Wang, Davis, Volkow and Xu116–Reference Tzur Bitan, Givon-Lavi, Kridin, Kaliner, Krieger and Dov Cohen119].
With the end of the pandemic outbreak of COVID-19 and the emergence of new and milder variants of the disease, individuals are facing lower risks of both severe acute manifestations and sequelae; however, evidence shows that new variants of the COVID-19 virus are currently associated with similar risk of CI and overall neurological and psychiatric sequalae as compared to earlier variants [Reference Taquet, Sillett, Zhu, Mendel, Camplisson and Dercon21]. This evidence indicates that health services worldwide will continue to face high rates of post-COVID CI and PCC diagnoses [Reference Solomonov, Kanellopoulos, Grosenick, Wilkins, Goldman and Ritholtz120] and, together with the evidence described in the present paper, strongly suggests that further research is needed to address this largely unmet need. Future research will need to be grounded on well-established definitions of PCC and post-COVID CI, assess objectively CI through comprehensive cognitive batteries, and employ longitudinal evaluations and study designs that allow better stratification of the studied population and control for confounding factors.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1192/j.eurpsy.2024.1770.
Acknowledgments
The authors have no acknowledgment to make.
Financial support
This study was supported by the Italian Ministry of University and Research (PRIN 2020 project “Patterns, correlates, interrelationships and evolution of psychopathology and cognitive impairment after recovery from symptomatic COVID-19 infection”; grant no. 20202CEKKJ).
Competing interest
Giulia Maria Giordano reports consulting fees from Angelini. Armida Mucci reports consulting fees from Pierre Fabre, Rovi, and Boehringer Ingelheim; patents with Pierre Fabre; and participation on data safety monitoring boards with Angelini and Boehringer Ingelheim. Silvana Galderisi reports consulting fees from Gedeon Richter; honoraria from Angelini, Boehringer Ingelheim, Gedeon Richter, Janssen, Lundbeck, Otsuka, Recordati, and Rovi; and participation on data safety monitoring boards with Angelini, Boehringer Ingelheim, Janssen, and Rovi. No other competing interests were declared.
Open access
Comments
No Comments have been published for this article.