Gender Disparities in Neurological Symptoms of Long COVID: A Systematic Review and Meta-Analysis

Abstract

Background

Female gender is a known risk factor for long COVID. With the increasing number of COVID-19 cases, the corresponding number of survivors is also expected to rise. To the best of our knowledge, no systematic review has specifically addressed the gender differences in neurological symptoms of long COVID.

Methods

We included studies on female individuals who presented with specific neurological symptoms at least 12 weeks after confirmed COVID-19 diagnosis from PubMed, Central, Scopus, and Web of Science. The search limit was put for after January 2020 until June 15, 2024. We excluded studies that did not provide sex-specific outcome data, those not in English, case reports, case series, and review articles

Results

A total of 5,632 eligible articles were identified. This article provides relevant information from 12 studies involving 6,849 patients, of which 3,414 were female. The sample size ranged from 70 to 2,856, with a maximum follow-up period of 18 months. The earliest publication date was September 16, 2021, while the latest was June 11, 2024. The following neurological symptoms had a significant difference in the risk ratio (RR) for female gender: fatigue RR 1.40 (95% confidence interval [CI]: 1.22–1.60, p < 0.001), headache RR 1.37 (95% CI: 1.12–1.67, p = 0.002), brain-fog RR 1.38 (95% CI 1.08–1.76, p = 0.011) depression RR 1.49 (95% CI: 1.2–1.86, p < 0.001), and anosmia RR 1.61 (95% CI: 1.36–1.90, p < 0.001). High heterogenicity was found for fatigue, brain fog, and anxiety due to the diverse methodologies employed in the studies.

Conclusion

Our findings suggest that women are at a higher risk for long-COVID neurological symptoms, including fatigue, headaches, brain fog, depression, and anosmia, compared to men. The prevalence of these symptoms decreases after 1 year, based on limited data from the small number of studies available beyond this period.

Introduction

Global reports from the World Health Organization (WHO) indicate that over 773,449,299 cases of coronavirus disease 2019 (COVID-19) have been documented. As the world navigates through the aftermath of the pandemic, COVID-19 lingers in diverse variants and subvariants [1]. With the ongoing surge in COVID-19 cases, the population of survivors is also anticipated to grow. A systematic review indicated that 45% of these survivors, regardless of their hospitalization status, continued to experience at least one lingering symptom at the 4-month mark [2]. Long COVID, otherwise referred to as postacute sequelae of COVID-19, has been defined by the WHO as the persistence or emergence of new symptoms 12 weeks following initial infection with the SARS-CoV-2 virus, with these symptoms lasting for a minimum of 2 months and lacking any other identifiable cause [3]. While the male gender has been identified as a risk factor for COVID-19 mortality [4], the female gender is recognized as a risk factor for long COVID [5–7]. Some studies posit that this might be attributed to women’s heightened self-awareness of such symptoms [8], potentially leading to more diagnoses of long COVID among females and subsequently increasing the estimated prevalence. This assertion finds support in a study indicating a predominance of females among patients seeking care for long COVID [9]. However, studies indicate that this alone cannot account for the elevated prevalence of prolonged COVID-19 among females compared to males [8]. Studies propose hypotheses indicating potential specific pathogenesis contributing to this gender disparity in long COVID [8].

Roughly 30% of long-COVID sufferers, aged 30–59 years, report that neurological issues severely hinder their work [10], highlighting the need for healthcare providers to understand common long-COVID symptoms in women. Neurological symptoms like fatigue and anosmia are common in long COVID [11], and while the female gender has been identified as a risk factor for long COVID [12], research on its link to specific neurological symptoms is sparse. This review seeks to investigate the discrepancies in neurological symptoms between genders in long-COVID cases.

Methods

This study used the following methodological framework, in conjunction with the extended Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist for systemic reviews [13]. The study protocol was preregistered on the International Prospective Register Reviews (PROSPERO; CRD42024512266).

Search Strategy and Selection Criteria

This systematic review included studies on individuals with long COVID, as defined by NICE guidelines, sourced from databases including PubMed, Central, Scopus, and Web of Sciences to establish an extensive pool of helpful information regarding the neurological effects of long COVID. We deliberately placed a search limit after January 2020 to June 15, 2024. This time frame was chosen because COVID-19 cases began to emerge in December 2019, marking the onset of the pandemic. We used a search strategy focused on identifying post-COVID-19 neurological symptoms. The following search strategy was used: (covid OR coronavirus OR “COVID-19” OR “COVID 2019” OR “COVID-2019” OR “SARS-CoV-2” OR “corona virus disease 2019” OR “severe acute respiratory syndrome coronavirus 2” OR “2019 pandemic” OR “Post-Acute COVID-19 Syndrome”) AND (female OR woman) AND (post OR long) AND (neurological OR neuro* OR “Neurologic Manifestations” OR “Post Acute COVID-19 Syndrome” OR “Post COVID-19” OR “Post COVID 19”) AND (headache OR fatigue OR depression OR anosmia OR anxiety OR “brain fog” OR “Mental Fatigue” OR “long COVID brain fog” OR “Post-traumatic stress disorder” OR ptsd OR “Stress Disorders, Post-Traumatic”).

Inclusion and Exclusion Criteria

This review investigated neurological symptoms in female patients of all ages following COVID-19 infection. The studies encompassed patients who exhibited particular neurological symptoms following a confirmed COVID-19 diagnosis, persisting for at least 12 weeks as per the NICE guideline [14]. We excluded studies that reported generalized neurological diagnoses without detailed symptom descriptions. Additionally, studies lacking sex-disaggregated data or failing to clearly differentiate male and female data were excluded. Non-English language studies, case reports, case series, systematic reviews, and meta-analyses were also excluded.

Data Extraction and Quality Assessment

Screening and data extraction were meticulously conducted by four independent reviewers (A.G., T.L., S.S., and L.L.). Each reviewer independently screened the studies for inclusion. In cases where discrepancies arose during the screening phase, they were discussed among all four reviewers to reach a consensus. If a unanimous decision could not be achieved, the issue was referred to a senior reviewer (Y.S.) for final resolution. For studies reporting on both control and patient groups, only data pertaining to the patient groups were extracted. A standardized extraction tool was utilized to gather essential information, including the first author’s name, study location, publication date, sample size, the number of long-COVID patients, the number of patients exhibiting neurological symptoms, sex, and specific neurological manifestations. Four authors (A.G., T.L., S.S., and L.L.) independently extracted data from the full texts of the 11 studies that passed the initial screening. Any inconsistencies in the extracted data were discussed collectively, and if necessary, a senior reviewer (Y.S.) was consulted to resolve persistent conflicts. During the quality assessment phase, two authors (A.G. and T.L.) independently evaluated the studies based on several criteria: the diagnostic criteria for COVID-19, the duration of neurological symptoms attributed to long COVID, and the objectives related to evaluating the impact of neurological symptoms on long COVID. In instances where disagreements occurred between the reviewers during this phase, the senior reviewer (Y.S.) provided the decisive judgment to resolve these issues.

Data Tabulation and Statistical Analyses

The effect size to be analyzed was the risk ratio (RR) and 95% confidence intervals (95% CIs) that were calculated for each neurological symptom. A Mantel-Haenszel-type estimator in a random effect model was used to estimate RR. Pooled estimates and 95% CIs were calculated from the individual RR. The results were pooled when the number of studies was ≥3 and a forest plot was used to plot the data. I² was used to study the heterogeneity between the studies. High heterogeneity is considered as i² >50%. Sources of heterogeneity were investigated in subgroup analyses. Publication bias was investigated by using a funnel plot that tests the hypothesis that the regression intercept is zero. Duval and Tweedie’s “trim and fill” method was applied to correct the point estimates for publication bias. Funnel plots of the standard error by log-RR are presented for the main analyses. Values of p < 0.05 were considered statistically significant. The data analyses were conducted using the R Foundation Statistical software (version 4.3.2).

Risk of Bias Assessment

Risk of bias assessment was performed using the Newcastle-Ottawa Scale (NOS) to critically appraise the literature included in the systematic review using common variables. Methodological quality and risk of bias were assessed by independent reviewers according to the NOS, which is valid for use in cohort studies [15] and the adapted version [16] for cross-sectional studies. The scale consists of eight items with three quality parameters: (i) selection, (ii) comparability, and (iii) outcome. The quality of the studies (poor, fair, and good) was scored by allocating stars to each domain as follows: a poor-quality score was allocated 0 or 1 star(s) in the selection, 0 stars in comparability, and 0 or 1 star(s) in the outcome domain; a fair quality score was awarded, two stars in the selection, one or two stars in comparability, and two or three stars in outcomes. A good quality score was awarded, with three or four stars in selection, one or two in comparability, and two or three stars in outcomes [15].

Results

In total, 5,632 articles were reviewed after the removal of duplicates, 582 articles were screened for titles and abstracts, and 81 articles met the criteria. These studies presented data on the long-COVID symptoms experienced by females. Afterward, the full texts of these articles were evaluated following the inclusion and exclusion criteria that were issued above, and after quality assessment, 12 studies [9, 17–27] were included in this review. Figure 1 shows the flow diagram detailing the review process and study selection based on the PRISMA flowchart.

Fig. 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.

Characteristics of the Results

Table 1 presents the characteristics of the studies included. The sample size ranged from 70 to 2,856, with a maximum follow-up period of 18 months and a minimum follow-up of 12 weeks. The earliest publication date was February 11, 2021, while the latest was June 11, 2024. Eleven countries were included in the review: the USA, Norway, the UK, Spain, Italy, Poland, Switzerland, and Korea. The majority of the studies (n = 8) originated from Europe, with Asia (n = 2) and North America (n = 2) contributing fewer studies. Our review discovered a total of 6,849 COVID-19 patients, among whom 3,414 were female. The predominant neurological symptoms that were different between the genders were fatigue (38.37% female prevalence, 26.84% male prevalence), headache (26.43% female prevalence, 19.81% male prevalence), anosmia (12.2% female prevalence, 6.66% male prevalence), anxiety (13.23% female prevalence, 10.57% male prevalence), depression (14.74% female prevalence, 8.93% male prevalence), brain-fog symptoms (16.9% female prevalence, 11.09% male prevalence), posttraumatic stress disorder (PTSD) (30.65% female prevalence, 17.9% male prevalence), and anosmia (12.2% female prevalence, 6.66% male prevalence). Table 2 illustrates the prevalence of symptoms, as outlined in the review.

Table 1.

Characteristics of the included studies

References	Country	Published time	Study type	No. of patients with past COVID-19 infection	Female
Mazurkiewicz et al. [17]	Poland	Nov 30, 2022	Retrospective observational	303	191 (63.03%)
Kim et al. [18]	Korea	Jan 27, 2022	Prospective cohort	241	164 (68.04%)
Bai et al. [19]	Italy	Apr 28, 2022	Prospective cohort	377	133 (35.27%)
Fernandez-de-las-Peñas et al. [20]	Spain	May 26, 2022	Retrospective cohort	1,969	915 (46.47%)
Michelutti et al. [21]	Italy	Oct 15, 2022	Retrospective observational	213	151 (70.89%)
Fjelltveit et al. [22]	Norway	Feb 1, 2023	Prospective case-control study	233	124 (53.21%)
Weinstock et al. [23]	USA	May 1, 2022	Retrospective cross-sectional	136	122 (89.7%)
Aparisi et al. [24]	Spain	Mar 26, 2022	Prospective case-control study	70	45 (64.28%)
Sykes et al. [25]	UK	Feb 11, 2021	Retrospective cohort	134	46 (34.32%)
Pelà et al. [26]	Italy	May 16, 2022	Retrospective cohort	223	89 (39.91%)
Ganesh et al. [9]	USA	Feb 5, 2022	Prospective cohort	108	58 (53.7%)
Gebhard et al. [27]	Switzerland	Jun 11, 2024	Prospective observational	2,856	1,307 (45.7%)

Table 2.

Analysis of female prevalence comparison to male prevalence

Symptom	Number of studies	Female sample size	Male sample size	Female prevalence	Male prevalence	Risk Ratio (95% CI)	p value
Fatigue	10	3,247	3,397	1,246 (38.37%)	912 (26.84%)	1.40 (1.22–1.6)	<0.001
Headache	7	2,856	2,883	755 (26.43%)	551 (19.81%)	1.37 (1.12–1.67)	0.002
Anxeity	5	2,569	3,008	340 (13.23%)	318 (10.57%)	1.32 (0.91–1.92)	0.13
Depression	4	2,523	2,920	372 (14.74%)	261 (8.93%)	1.49 (1.2–1.86)	<0.001
Brain-fog symptoms	7	2,863	3,173	486 (16.9%)	352 (11.09%)	1.38 (1.08–1.76)	0.011
Post-traumic stress disorder	1	137	240	42 (30.65%)	43 (17.90%)	NE	NE
Anosmia	5	2,563	2,993	313 (12.2%)	200 (6.66%)	1.61 (1.36–1.90)	<0.001
Paresthesia	2	1,458	1,611	43 (2.9%)	40 (2.4%)	NE	NE
Dysgeusia	2	1,444	1,789	141 (9.76%)	102 (5.7%)	NE	NE
Vertigo	2	1,458	1,611	64 (4.3%)	60 (3.72%)	NE	NE
Visual Disturbances	2	1,458	1,611	49 (3.3%)	49 (3%)	NE	NE

CI, Confidence interval; NE, non estimable.

Bold entries in the table idicate a p value of ≤0.05.

Risk of Bias in Included Studies

Table 3 summarizes our assessments of the risk of bias in the included studies. Among the observational studies two studies [20, 23], we rated the overall risk of bias as “fair” due to identified limitations in methodological reliability and insufficient follow-up. We rated the overall low risk of bias for the rest of the observational studies as “good.”

Table 3.

Risk of bias quality assessment

	Study type	Selection				Comparability		Exposure/Outcome			Sub total assessment			Conclusion	Total
		1	2	3	4	1a	1b	1	2	3	S	C	E
Mazurkiewicz et al. [17]	cross-sectional	*	no	*	**	*	*	*	*	*	good	good	good	GOOD	7
Kim et al. [18]	cohort	*	*	*	no	*	*	*	*	no	good	good	good	GOOD	7
Bai et al. [19]	cohort	*	*	*	no	*	*	*	no	no	good	good	fair	GOOD	6
Fernandez-de-las-Peñas et al. [20]	cohort	*	*	no	no	*	*	*	no	no	fair	good	fair	FAIR	5
Michelutti et al. [21]	cross-sectional	*	*	*	*	*	*	*	*	*	good	good	good	GOOD	8
Fjelltveit et al. [22]	case-control-	*	*	*	no	*	*	*	*	no	good	good	good	GOOD	7
Weinstock et al. [23]	cross-sectional	*	no	no	*	*	*	*	no	*	fair	good	good	FAIR	5
Aparisi et al. [24]	case-control	*	*	*	no	*	*	*	*	no	good	good	good	GOOD	7
Sykes et al. [25]	cohort	*	*	*	no	*	*	*	*	no	good	good	good	GOOD	7
Pelà et al. [26]	cohort	*	no	*	*	*	no	*	*	*	good	fair	good	GOOD	7
Ganesh et al. [9]	cohort	no	*	*	*	*	*	*	*	*	good	good	good	GOOD	8
Gebhard et al. [27]	cohort	*	*	*	*	*	*	*	*	*	good	good	good	GOOD	9

Pooled Analysis

Fatigue

Ten studies reported fatigue. The female estimated the prevalence of 38.37%. Fatigue was the most frequently reported symptom within the neurological symptoms. For fatigue, the pooled RR of females experiencing fatigue compared to males was 1.40 (95% CI: 1.22–1.6, p < 0.001). A forest plot for fatigue is shown in Figure 2.

Fig. 2.

The forest plot compares the risk of fatigue (upper panel) and headache (lower panel) between females and males across various studies. Each square represents the RR for a study, with square size indicating the study’s weight in the analysis. Horizontal lines through squares denote the 95% CI. The diamond at the bottom shows the overall pooled RR and its 95% CI, summarizing the combined effect size. A vertical line at RR = 1 indicates no difference in risk between groups. Heterogeneity measures assess variability among study results.

Headache

Seven studies reported headaches. The female estimated prevalence is 26.43%. For headache, the pooled RR of females experiencing headache compared to males was 1.37 (95% CI: 1.12–1.67, p = 0.002). A forest plot for headache is shown in Figure 2.

Brain-Fog Symptoms

Seven studies mentioned brain-fog symptoms. The female estimated prevalence is 16.9%. For brain fog, the pooled RR of females experiencing brain fog compared to males was 1.38 (95% CI: 1.08–1.76, p = 0.011). A forest plot for brain fog is shown in Figure 3a.

Fig. 3.

This forest plot compares the risk of brain fog (a), anxiety (b), depression (c), and anosmia (d) between females and males across various studies. Each square represents the RR for an individual study, with the square size proportional to the study’s weight. Horizontal lines through squares denote the 95% CI. The diamond at the bottom shows the overall pooled RR and its 95% CI, indicating the combined effect size. A vertical line at RR = 1 serves as a reference for no difference in risk between groups. Heterogeneity measures assess variability among the study results.

Anxiety

Three studies reported anxiety. The female anxiety estimated prevalence is 13.23%. For anxiety, the pooled RR of females experiencing anxiety compared to males was 1.32 (95% CI: 0.91–1.92, p = 0.13). A forest plot for anxiety is shown in Figure 3b.

Depression

Three studies reported depression. The depression estimated prevalence is 14.74%. For depression, the pooled RR of females experiencing depression compared to males was 1.49 (95% CI: 1.2–1.86, p = 016). A forest plot for fatigue is shown in Figure 3c.

Anosmia

Four studies mentioned anosmia. The female estimated prevalence is 12.2%. For anosmia, the pooled RR of females experiencing anosmia compared to males was 1.61 (95% CI: 1.36–1.9, p < 0.001). A forest plot for brain fog is shown in Figure 3d.

Restless Leg Syndrome

One study by Weinstock et al. [23] identifies 134 patients through a cross-sectional survey. He found 13 long-COVID patients with new restless leg syndrome, and all 13 were female. He found that the baseline prevalence of restless legs syndrome in females with long COVID was similar to that in the general population of control females. However, the prevalence of restless legs syndrome was higher in the long-COVID state [23].

Posttraumatic Stress Disorder

One study by Bai et al. [19] identifies 377 patients in a single center at San Paolo Hospital in Milan. The authors identified 42 female patients with PTSD and 43 male patients. The estimated prevalence for female is 30.65%.

Publication Bias

The funnel plot analysis of fatigue indicates a possibility of publication bias or heterogeneity, given the asymmetric distribution of studies, particularly with clustering to the right. This suggests that smaller studies might exhibit more pronounced effects. To evaluate the presence of publication bias, Egger’s test was employed. The test, applied to 10 studies focusing on fatigue symptoms, yielded a p value of 0.21, suggesting a lack of significant publication bias. Figure 4 illustrates the funnel plot about fatigue.

Fig. 4.

This funnel plot evaluates potential publication bias in studies comparing the risk of fatigue between females and males. Each red dot represents a study included in the meta-analysis. The horizontal axis shows the RR from each study, while the vertical axis represents the standard error (SE) of the RR. Symmetry around the vertical line indicates no publication bias. The dashed lines form a funnel shape outlining the expected range of study results. Asymmetry or clustering outside the funnel may suggest potential publication bias.

Subgroup Analysis – Prospective Versus Retrospective Studies

A subgroup analysis was performed to evaluate the differences between retrospective and prospective studies. The analysis found no significant differences between retrospective and prospective studies for all the neurological symptoms included in the meta-analysis. This suggests that both retrospective and prospective studies provide comparable results in our investigation. Figure 5 provides a comparative analysis of the study designs specifically for the fatigue symptom. It shows the comparison between retrospective and prospective studies in terms of their results on fatigue. Further details on the comparative analysis for other neurological symptoms are available in online suppl. Figure 1 (for all online suppl. material, see https://doi.org/10.1159/000540919).

Fig. 5.

Forest plot comparing the risk of fatigue in females (experimental group) versus males (control group) across 10 studies, divided into retrospective and prospective subgroups. Each square represents the study’s odds ratio (OR) and its 95% CI, with the square size indicating the study’s weight in the meta-analysis. Diamonds show the pooled OR and 95% CI for each subgroup and overall. The vertical line at OR = 1 indicates no difference between groups. Heterogeneity is moderate to high in retrospective studies (I² = 74%) and low in prospective studies (I² = 28%).

Discussion

This meta-analysis aimed to explore the difference in the prevalence of neurological symptoms between females and males with long COVID. Our results indicate that female patients are at a significant higher risk compared to males for developing long-COVID symptoms fatigue, headaches, depression, and anosmia. Additionally, while not significant, the prevalence of PTSD, anxiety, dysgeusia, and vertigo is greater in the female gender. This observation is consistent with a recent meta-analysis by Natarajan et al. [28], which included 36 studies and a total sample of 11,598 patients with long COVID. The meta-analysis reported estimated prevalence rates for neurological symptoms in a forest plot as follows: fatigue 29.2%, headache 9.49%, memory deficit 18.43%, attention/concentration deficit 20.22%, smell dysfunction 14.38%, and PTSD 22.44%. Comparing these figures with our results, it is evident that the estimated prevalence of fatigue, headache, and PTSD was higher in the female population than in the general long-COVID population. This could be attributed to the increased susceptibility of the female gender to autoimmune responses triggered by COVID-19. Our review demonstrated high heterogeneity (i² >50%) in the following symptoms: fatigue, brain fog, and anxiety. This variability might stem from several factors, one being the varied methodologies across the studies: while some conducted physical follow-ups to assess symptoms, others utilized surveys [18]. Additionally, there was diversity in patient selection, with some research focusing on hospitalized COVID-19 patients [19, 20, 25] and others including nonhospitalized or a mix of patient groups. Another factor contributing to the heterogeneity could be our inclusion criteria necessitated a long-COVID diagnosis without imposing limits on the follow-up duration. This absence of temporal constraints might have contributed to the review’s heterogeneity, potentially amplifying it beyond what might have been observed had the follow-up period been restricted to 1 year. This notion is supported by a study conducted by Mizrahi et al. [29], which analyzed health outcomes in a cohort of 1,913,234 individuals, comparing unvaccinated individuals with COVID-19 to those without the infection. The findings suggested that individuals with mild COVID-19 cases were generally at a lower risk for adverse health outcomes, with most symptoms resolving within a year [29]. This is consistent with the studies by Ye et al. [30], which were conducted 2 years after hospital discharge and reported the lowest prevalence among those reviewed (fatigue, 5.22%; headache, 0.66%; anxiety, 6.11%; depression, 8.67%; anosmia, 0.11%). However, although the prevalence might be affected, it is noteworthy that we observed persistent symptoms beyond a year [15, 19], and it appears that the occurrence of symptoms does not decrease significantly [19]. Furthermore, other studies indicate that these symptoms can impact the quality of life even 2 years postinfection [31, 32]. Based on these studies, it is indicated that the prevalence of long-COVID neurological symptoms generally decreases over time, although some patients continue to experience persistent symptoms.

The underlying mechanism for females being a risk factor to developed long-COVID neurological symptoms remains incompletely elucidated; nonetheless, one hypothesis suggests that females exhibit heightened susceptibility to COVID-19-induced autoimmunity. Evidence indicates that female COVID-19 patients manifest a greater propensity to generate antibody responses against the COVID-19 spike protein antigen than their male counterparts [33]. Additionally, distinct microbiome compositions between sexes may play a role, with females harboring a greater abundance of bacteria that modulate immune function [34]. Moreover, studies have suggested that sex hormones may contribute to the development of an autoimmune state, thereby playing a potential role in pathogenesis [35]. Estrogen has a greater propensity to enhance humoral immune responses and antibody generation than androgenic hormones, such as testosterone [36]. Another possible mechanism is that the angiotensin-converting enzyme 2 (ACE2) gene is located on the X chromosome, leading to ACE2 overexpression. This overexpression of ACE2 offers protection against severe disease, but at the cost of prolonging disease manifestations, leading to symptoms that can be found in long COVID [37]. This finding regarding sex hormones is consistent with studies, showing that premenopausal women are more affected than postmenopausal women with long-COVID symptoms [38]. The potential mechanisms underlying female-specific risks of long-COVID symptoms are shown in Figure 6.

Fig. 6.

Potential mechanisms underlying female-specific risks of long-COVID symptoms. Estrogen and factors related to the X chromosome and female gut microbiome can modulate immune responses. Specifically, these factors can influence the humoral response. This may lead to less severe COVID-19 cases; however, it can result in the persistence of COVID-19 remnants and subsequent long-COVID symptoms. The presence of an additional X chromosome in females might elevate the expression of ACE2, further contributing to the sustained presence of COVID-19 in the system. Moreover, COVID-19 can disrupt immune tolerance mechanisms, leading to the generation of autoreactive immune cells and antibodies. This phenomenon may occur through processes, such as molecular mimicry. When COVID-19 gains access to the central nervous system (CNS), it utilizes various pathways, including the direct invasion of CNS cells, retrograde axonal transport, and penetration through the endothelial cells of the blood-brain barrier. Once inside the CNS, COVID-19 can prompt microglial cells to release proinflammatory agents, leading to mitochondrial dysfunction and oxidative stress. This cascade of events leads to neuroinflammation, demyelination, and neurodegeneration. Collectively, these processes may increase the vulnerability of females to long-COVID neurological complications.

In our study, chronic fatigue emerged as the symptom with the highest prevalence, which is consistent with the findings of other studies [31, 39, 40]. COVID-19 is not the only entity associated with post-viral fatigue. Research indicates that severe acute respiratory syndrome coronavirus 1 and the Middle East respiratory syndrome have also been linked to such fatigue [41]. However, compared to other post-viral fatigue, our review indicates that in long-COVID patients, being female is a risk factor for developing fatigue. Moreover, in a study by Ye et al. [30], fatigue was correlated with depression, rather than anxiety [21]. This is consistent with our findings, indicating that depression exhibited a significant difference between females and males, whereas anxiety showed a similar prevalence between the genders in long-COVID patients. This association may be attributed to the shared pathophysiological mechanisms between fatigue and depression, such as immune-mediated injury and neuroinflammation. Indeed, some studies have identified a causal relationship between these factors [42]. Ganesh et al. [9] found in their study that female patients tend to have higher IL-6 levels than males and predominant fatigue is associated with higher IL-6 levels compared to those with dyspnea and chest pain. They suggested a potentially shared immune dysregulation pathway between long-COVID fatigue and central sensitization (CS) phenotypes, including conditions such as chronic fatigue syndrome, fibromyalgia, and postural tachycardia syndrome. This association with the CS phenotype is linked to increased IL-6 levels [43]. The mechanism underlying CS involves heightened sensitivity of the brain and spinal cord to stimuli, thereby lowering the threshold for stimulation and amplifying the existing stimuli [44]. Although the reasons for CS are multifaceted and intricate, IL-6 plays a role in its development [45]. Komaroff and Lipkin [46] highlighted the similarities between long-COVID fatigue and myalgic encephalomyelitis/chronic fatigue syndrome, noting numerous parallels in the underlying biological mechanisms of these two conditions.

The expression “brain-fog symptoms” refers to a wide spectrum of neurocognitive deficits, including difficulties with concentration and multitasking, as well as both short-term and long-term memory impairments. Studies indicate that these symptoms significantly diminish the quality of life of the affected individuals [47]. Moreover, patients reporting such symptoms often express feelings of guilt juxtaposed with gratitude for their survival [48]. Our review found similar findings regarding the reduction in quality of life due to these symptoms. The mechanisms underlying brain fog in the context of COVID-19 are not clearly understood. However, several theories have been proposed. One prevailing hypothesis suggests that COVID-19 infects the CNS cells via ACE2 receptors, particularly astrocytes, which are the predominant cells in the CNS. When astrocytes become infected, they may alter their metabolic pathways. This disruption could potentially harm neighboring neurons, as astrocytes play a supportive role in neurons. Such damage could explain the symptoms observed in individuals experiencing brain fog [49]. Another theory posits that microglial cells may be activated and triggered by an entry point from the hypothalamus. The activation of these microglial cells can result in the release of proinflammatory molecules [50]. Additionally, COVID-19 may exacerbate oxidative stress and induce mitochondrial dysfunction in microglial cells [51]. Such neuroinflammatory responses and impaired redox processes are believed to play a significant role in the progression of neurological effects associated with long COVID [52].

Headache was found to be the second most common symptoms for the female population with long COVID. A long-COVID headache is defined as a form of a new headache that starts during the acute infection of COVID-19 or after a delay [53]. The prevalence of long-COVID headaches is 18%, which is higher in middle-aged women [53]. The reason for the higher prevalence in female gender could be due to the common reason that headaches are more common in females than in the general population [54]. However, specific headache patterns might be related to long-COVID syndrome, and some studies suggest that bilateral and pressing quality and the phenotype of tension-type headaches are more common than migraine headaches [55].

The mechanism behind long-COVID headaches is not widely accepted, but one hypothesis suggests that these headaches occur due to the CNS invasion by COVID-19 [56]. This invasion may occur through multiple pathways. COVID-19 can bind to ACE-2 receptors, facilitating the cellular invasion of CNS cells, including neurons and glia in structures such as the olfactory groove and trigeminal ganglia [57]. Another significant pathway involves retrograde axonal transport, where the virus infects peripheral neurons in the nasal epithelium and travels backward along the olfactory nerve to the olfactory bulb in the brain [58]. This process uses the neuron’s transport machinery to move viral particles from the nasal cavity directly into the CNS, effectively bypassing the blood-brain barrier. Additionally, COVID-19 can enter the bloodstream (viremia) and infect endothelial cells of the blood-brain barrier or be transported by infected immune cells, such as monocytes or macrophages [58, 59]. These infected cells can cross the BBB and release the virus into the CNS.

Our subanalysis comparing prospective and retrospective studies on long-COVID neurological symptoms found no significant differences between these study types. This suggests that retrospective studies, despite their inherent limitations, produce results consistent with prospective studies. This finding validates the inclusion of retrospective data in our study and supports its use in future research. Consequently, this insight allows future studies and clinical guidelines to potentially rely on both retrospective and prospective data when evaluating neurological outcomes in long COVID patients. This strengthens the robustness of our findings across different study designs. The strength of our study was that we performed a comprehensive search of a wide number of electronic databases. The limitations of this review are the limited number of studies included, high heterogeneity in the estimation of the prevalence of some neurological manifestations, and the inclusion of studies with very small sample sizes.

Conclusion

Our review found that female is a risk factor for long-COVID fatigue, headache, depression, and anosmia, and the prevalence of these symptoms decreases after 1 year, based on limited data from the small number of studies available beyond this period. Given the higher association of long COVID with females, it is essential for physicians to understand the predominant neurological symptoms. Specifically, recognizing the most prevalent symptoms and linking their presentation to long-COVID symptoms in the female population are crucial. Hence, ongoing research on this topic remains paramount, especially among populations at a higher risk of the disease. We hope that future studies will focus on sex differences in long-term symptoms and try to explain the reason for the increased prevalence.

Statement of Ethics

The study approval was not required, since this was a meta-analysis of published studies. Written informed consent was not required, since this was a meta-analysis of published studies.

Conflict of Interest Statement

The authors declare that they have no competing interests.

Funding Sources

This work did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

Author Contributions

Conceptualization: A.G., T.L., and Y.S.; methodology and data curation: A.G., S.S., L.L., and T.L.; formal analysis: A.G. and Y.S.; writing – original draft preparation: A.G.; writing – review and editing; A.G., Y.S., and L.L.; and supervision: Y.S.

Funding Statement

This work did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

Data Availability Statement

The data are available upon reasonable request.