Hostname: page-component-77f85d65b8-g4pgd Total loading time: 0 Render date: 2026-04-17T16:25:54.444Z Has data issue: false hasContentIssue false
  • English
  • Français

Bipolar Disorder Symptoms in Old Age: Gender and Age Differences Network Analysis

Published online by Cambridge University Press:  26 November 2025

Berta Ausín ,
Miguel Ángel Castellanos  and
Manuel Muñoz
Berta Ausín*
Affiliation:
Universidad Complutense de Madrid , Spain
Miguel Ángel Castellanos
Affiliation:
Universidad Complutense de Madrid , Spain
Manuel Muñoz
Affiliation:
Universidad Complutense de Madrid , Spain
*
Corresponding author: Berta Ausín; Email: bertaausin@psi.ucm.es
Rights & Permissions [Opens in a new window]

Abstract

The clinical presentation and course of illness of older-age bipolar disorder (OABD) are highly variable. In addition, the presentation and course of bipolar disorder (BD) differ between females and males. This study aims to carry out a network analysis of older people with symptoms compatible with BD. Using a sample from the MentDis_ICF65+ study, a symptom network analysis was conducted according to gender and age in 555 people over 65 in the Community of Madrid (Spain). The network was estimated using the InsingFit package that implements a procedure called eLasso. These results reveal differences in the strength, closeness, and betweenness of the networks according to gender and for the age groups 65–74 and 75–84. Females present a network that is much more sparse, with a lower density, and consisting of two sub-networks: one composed of TALK (more talkative than usual) and RACIN (a flight of ideas, racing thoughts) and the other of PAINF (activities with painful consequences), SLEEP (the decreased need for sleep), GRAND (inflated self-esteem), and AGIT (psychomotor agitation). In the case of men, a denser network is obtained, with greater connections between all the symptoms, being the edge with greater weight than the one integrated by RACIN and GRAND. In relation to age, it is possible to observe changes in the model between the two age groups. These network differences support viewing OABD dimensionally and emphasize considering gender and age to improve understanding and personalize treatments for older adults with bipolar disorder symptoms.

Keywords

agingbipolar disordergender differencesnetwork analysisolder adults

Information

Type
Research Article
Information
The Spanish Journal of Psychology , Volume 28 , 2025 , e23
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of Universidad Complutense de Madrid and Colegio Oficial de la Psicología de Madrid

Introduction

Bipolar disorders (BDs) are under-researched in the older population. Volkert et al. (Reference Volkert, Schulz, Härter, Wlodarcyzk and Andreas2013) conducted a meta-analysis assessing the prevalence of mental disorders in older people in Europe and North America. This study found a prevalence of 0.53% for current BD and 1.10% for lifetime BD, which is similar to the prevalence found in the population under 65 years of age. A European epidemiological study of people over 65 years of age reports a lifetime prevalence of BD of 4.4%, and a 12-month prevalence of 2.5% (Andreas et al., Reference Andreas, Schulz, Volkert, Dehoust, Sehner, Suling, Ausín, Canuto, Crawford, Da Ronch, Grassi, Hershkovitz, Muñoz, Quirk, Rotenstein, Santos-Olmo, Shalev, Strehle, Weber and Härter2017). In the Spanish population, the lifetime prevalence of BD is 6.61%, the 12-month prevalence is 4.68%, and the current prevalence is 0.18% (Ausín et al., Reference Ausín, Muñoz, Santos-Olmo, Pérez-Santos and Castellanos2017). This study shows that among older people, the prevalence of BD decreases significantly with age. Females are affected to a greater extent by this disorder. Moreover, the presentation and course of BD differ between females and males (Arnold, Reference Arnold2003; Dias et al., Reference Dias, Kerr-Corrêa, Torresan and Santos2006; Parial, Reference Parial2015). Arnold (Reference Arnold2003) notes that the onset of BD tends to occur later in females than males, and females more often have a seasonal pattern of mood disturbance. Furthermore, females experience depressive episodes, mixed mania, and rapid cycling more often than males. Bipolar II disorder, which is predominated by depressive episodes, also appears to be more common in females than males. In addition, males more often debut with a manic episode, whereas females seem more likely to present with depression (Dell’Osso et al., Reference Dell’Osso, Cafaro and Ketter2021) and experience depressive episodes more frequently than males (Altshuler et al., Reference Altshuler, Kupka, Hellemann, Frye, Sugar, McElroy, Nolen, Grunze, Leverich, Keck, Zermeno, Post and Suppes2010). Finally, comorbidity of medical and psychiatric disorders is more common in females than males. On the other hand, regarding gender differences in BD treatments, females are more likely to be in contact with specialist services for BD (Cunningham et al., Reference Cunningham, Crowe, Stanley, Haitana, Pitama, Porter, Baxter, Huria, Mulder, Clark and Lacey2020). These authors found in their study that females were more likely to receive outpatient treatment only and have recorded comorbid anxiety, whereas more males had recorded substance use disorder, were convicted of crimes when unwell, received compulsory treatment, and received inpatient treatment.

Furthermore, in the Global Aging and Geriatric Experiments in Bipolar Disorder study (Blanken et al., Reference Blanken, Oudega, Almeida, Schouws, Orhan, Beunders, Klumpers, Sonnenberg, Blumberg, Eyler, Forester, Forlenza, Gildengers, Mulsant, Rajji, Rej, Sarna, Sutherland, Yala and Dols2024), significant clinical differences between genders were observed. Older males showed higher rates of substance use disorders. These study results suggest that older females tend to have a more severe course of BD compared with older males, experiencing higher rates of psychiatric hospitalization along with increased symptoms of anxiety and hypochondriasis.

Sajatovic et al. (Reference Sajatovic, Strejilevich, Gildengers, Dols, Al Jurdi, Forester, Kessing, Beyer, Manes, Rej, Rosa, Schouws, Tsai, Young and Shulman2015) note that older-age bipolar disorder (OABD; BD in individuals aged ≥60 years) represents as much as 25% of the population with BD and represents a heterogeneous group, including those with early-onset BD as well as late-onset BD, with potentially different pathogenesis, clinical course, and care needs. Furthermore, the clinical presentation and course of illness of OABD are highly variable, often characterized by mood episode recurrence, medical comorbidity, cognitive deficits, and impaired functioning (Chen et al., Reference Chen, Dols, Rej and Sajatovic2017). These authors conclude that treating OABD is challenging due to medical complexity, comorbidity, diminished tolerance to treatment, and a limited evidence base. In this line, Dols and Beekman (Reference Dols and Beekman2018) indicate that further understanding of OABD may lead to more specific recommendations for treatment adjusted to the specific characteristics and needs caused by age-related somatic and cognitive changes.

Furthermore, data characterizing BD in older people are scarce, particularly on functional status. Different studies report a negative impact on the functioning and quality of life of symptoms of BD in older people (Cotrena et al., Reference Cotrena, Branco, Kochhann, Shansis and Fonseca2016; De la Fuente-Tomás et al., Reference De la Fuente-Tomás, Sierra, Sanchez-Autet, García-Blanco, Safont, Arranz and García-Portilla2018; Depp et al., Reference Depp, Davis, Mittal, Patterson and Jeste2006; Sajatovic et al., Reference Sajatovic, Strejilevich, Gildengers, Dols, Al Jurdi, Forester, Kessing, Beyer, Manes, Rej, Rosa, Schouws, Tsai, Young and Shulman2015; Tatay-Manteiga et al., Reference Tatay-Manteiga, Cauli, Tabarés-Seisdedos, Michalak, Kapczinski and Balanzá-Martínez2019).

The Diagnostic and Statistical Manual of Mental Disorders (DSM-5; APA, 2014) points out seven symptoms for the BD: inflated self-esteem or grandiosity; decreased need for sleep; increased talkativeness; racing thoughts; distracted easily; increase in goal-directed activity or psychomotor agitation, and engaging in activities that hold the potential for painful consequences, and states, a-theoretically, that they manifest together. This view has the potential to obscure important differences between specific symptoms and does not indicate a significant relationship between these symptoms, which may establish different relationships according to age and gender variables.

Borsboom (Reference Borsboom2017) proposes an alternative vision to these categorical classifications, which is the perspective of symptom networks in psychopathology.

Borsboom and Cramer (Reference Borsboom and Cramer2013) assume that in-network approaches to psychopathology disorders result from the causal interplay between symptoms (e.g., worry causes insomnia and insomnia causes fatigue). This technique provides a visual description of the partial correlations between symptoms, which can be interpreted as complex associations. Borsboom and Cramer (Reference Borsboom and Cramer2013) note that networks consist of nodes and edges. Nodes represent the objects of study, and edges represent the connections between them.

A review of published studies on BD symptom network analysis or older people with symptoms compatible with BD network analysis shows a lack of studies on this subject. An advanced Web of Science search on articles on network analysis and BD in the last 5 years, using the following query ((TS = (network NEAR/0 analysis) AND TS = (bipolar) AND SU = (Psychology OR Psychiatry) NOT TS = (neur* OR gene OR genes OR “brain network” OR “social network” OR “social media analysis” OR “text mining” OR “data mining” OR “content analysis” OR “semantic network análisis” OR “thematic network analysis”), reveals that only five studies have been published (Corponi et al., Reference Corponi, Anmella, Verdolini, Pacchiarotti, Samalin, Popovic, Azorin, Angst, Bowden, Mosolov, Young, Perugi, Vieta and Murru2020; Curtiss et al., Reference Curtiss, Fulford, Hofmann and Gershon2019; Peralta et al., Reference Peralta, Gil-Berrozpe, Sánchez-Torres and Cuesta2020; Strauss et al., Reference Strauss, Esfahlani, Kirkpatrick, Allen, Gold, Visser and Sayama2019; Weintraub et al., Reference Weintraub, Schneck and Miklowitz2020).

Peralta et al. (Reference Peralta, Gil-Berrozpe, Sánchez-Torres and Cuesta2020) network analysis identified four dimensions in subjects with BD (depression, mania, positive and negative), and the most central symptom in BD was delusions. Corponi et al. (Reference Corponi, Anmella, Verdolini, Pacchiarotti, Samalin, Popovic, Azorin, Angst, Bowden, Mosolov, Young, Perugi, Vieta and Murru2020) found that mixed manic-depressive symptoms were the most central and highly interconnected nodes in the network, particularly agitation and irritability. These authors note that, besides mixed symptoms, appetite gain and hypersomnia were significantly endorsed in BD patients.

Weintraub et al. (Reference Weintraub, Schneck and Miklowitz2020) conducted a network analysis of mood symptoms in adolescents with BD or at risk for bipolar-spectrum disorders and found that symptoms within the depressive and manic mood poles were more related to each other than to symptoms of the opposing mood pole. Furthermore, symptoms related to activity/energy levels (i.e., fatigue or hyperactivity) and depressed mood were the most prominent mood symptoms among youth with bipolar spectrum disorders.

Strauss et al. (Reference Strauss, Esfahlani, Kirkpatrick, Allen, Gold, Visser and Sayama2019) examined how densely interconnected individual negative symptom domains are in BD, whether some domains are more central than others, and whether gender influenced network structure. These authors found that anhedonia was the most central symptom in BD. They also reported gender differences in centrality, which seems to indicate that the search for pathophysiological mechanisms and targeted treatment development should be focused on different sets of symptoms in males and females. As noted by Blanken et al. (Reference Blanken, Oudega, Almeida, Schouws, Orhan, Beunders, Klumpers, Sonnenberg, Blumberg, Eyler, Forester, Forlenza, Gildengers, Mulsant, Rajji, Rej, Sarna, Sutherland, Yala and Dols2024), there is an urgent need for a deeper understanding of gender differences throughout the bipolar lifespan to create targeted treatments and enhance therapeutic outcomes.

To the best of our knowledge, however, no symptom network study examined symptoms compatible with BD in the population over 65, despite the clinical importance of these symptoms in older adults.

We hypothesize that it is possible to construct different networks based on gender and age by examining symptoms compatible with BD among people over 65. Specifically, since the present study evaluates symptoms compatible with BD type I, which are manic symptoms and males more often debut with a manic episode, it is expected to find the frequencies of occurrence for most symptoms to be higher for males than for females. This study has two aims: (1) to examine the relationship between symptoms compatible with BD among older subjects from the general population, using network analysis, and (2) to estimate the network structure among symptoms compatible with BD and analyze gender and age-related differences in a sample of undiagnosed people over 65 years old in the Community of Madrid (Spain).

Materials and Method

Design

This study uses national data from the Health and Well-Being of People Aged 65–84 in Europe (MentDis_ICF65+ study), being a European study in which the authors of this article participated on behalf of the Spanish team. The sample was drawn from the MentDis_ICF65+ study, which focuses on the health and well-being of older adults across Europe (Andreas et al., Reference Andreas, Härter, Volkert, Hausberg, Sehner, Wegscheider, Rabung, Ausín, Canuto, Da Ronch, Grassi, Hershkovitz, Lelliott, Muñoz, Quirk, Rotenstein, Santos-Olmo, Shalev, Siegert and Schulz2013). The primary goal of the MentDis65+ project is to estimate the prevalence of mental health disorders in the European population aged 65 and above (Andreas et al., Reference Andreas, Härter, Volkert, Hausberg, Sehner, Wegscheider, Rabung, Ausín, Canuto, Da Ronch, Grassi, Hershkovitz, Lelliott, Muñoz, Quirk, Rotenstein, Santos-Olmo, Shalev, Siegert and Schulz2013, Reference Andreas, Schulz, Volkert, Dehoust, Sehner, Suling, Ausín, Canuto, Crawford, Da Ronch, Grassi, Hershkovitz, Muñoz, Quirk, Rotenstein, Santos-Olmo, Shalev, Strehle, Weber and Härter2017). The sampling strategy specifically targets individuals aged 65 to 85 years, using diagnostic tools and measurement instruments tailored to the needs and characteristics of older adults. A random, stratified sample was selected based on age groups (65–74 and 75–85) and gender, from individuals aged 65–85 living in the Community of Madrid, including all 21 districts of the city.

Participants

A total of 555 participants, who met the inclusion criteria, were interviewed and included in the network analysis.

The inclusion criteria for the sample were as follows:

  1. i) residency in Madrid;

  2. ii) age between 65 and 84 years; and

  3. iii) ability to provide informed consent to participate.

The exclusion criteria were as follows:

  1. i) individuals over 85 years old;

  2. ii) nursing home residents;

  3. iii) severe cognitive impairment, as assessed by a Mini-Mental State Examination (Mini-mental; Folstein et al., Reference Folstein, Folstein and McHugh1975) with a cutoff score greater than 18; and

  4. iv) language barriers preventing the interview from taking place.

Informed consent was obtained from each participant. The study was conducted in line with the Declaration of Helsinki, and the research protocol received approval from the Deontological Commission of the Faculty of Psychology at the Complutense University of Madrid (Reference No. 2203201), as well as from the European Commission.

Variables and Instruments

The following variables were assessed using the corresponding instruments:

Sociodemographic variables: Data on age, gender, marital status, education level, medical diagnoses, economic situation (subjective perception ranging from very poor to very good), and the significance of religious beliefs were collected through custom-made questions.

Assessment of symptoms compatible with BD: To evaluate and diagnose symptoms compatible with BD type I, the Composite International Diagnostic Interview for people over 65 years (CIDI65+) was used (Wittchen et al., Reference Wittchen, Strehle, Gerschler, Volkert, Dehoust, Sehner, Wegscheider, Ausìn, Canuto, Crawford, Da Ronch, Grassi, Hershkovitz, Muñoz, Quirk, Rotenstein, Santos-Olmo, Shalev, Weber and Andreas2014). This structured diagnostic interview was employed to gather 12-month prevalence data of mental disorders in older adults. It assesses the main conditions listed in the DSM-IV-TR, providing diagnoses based on the criteria of the DSM-IV-TR classification (APA, 2000). There are seven DSM-5 symptoms (APA, 2014) for BD, as in the DSM-IV-TR: (1) inflated self-esteem or grandiosity (GRAND); (2) decreased need for sleep (SLEEP); (3) increased talkativeness (TALK); (4) racing thoughts (RACING); (5) distracted easily (DISTR); (6) increase in goal-directed activity or psychomotor agitation (AGIT); and (7) engaging in activities that hold the potential for painful consequences (PAINF).

Statistical Analyses

The frequencies and percentages of the items in the CIDI65+ interview were calculated as descriptive statistics. A test for two proportions (Z-test) was calculated to compare the proportions of each symptom in each variable under study: gender (female and male) and age groups (65–74 and 75–84 years).

The analyses have been performed with R statistical software (v3.5.6) using several packages. The network was estimated using the InsingFit package created by van Borkulo (van Borkulo et al., Reference van Borkulo, Boschloo, Borsboom, Penninx, Waldorp and Schoevers2015); this package implements a procedure called eLasso. This procedure is an extension of the lasso procedure that is widely used for continuous data and that imposes an L1 penalty on the estimation of the inverse covariance matrix. eLasso has been shown to work best when the data are binary.

To guarantee the sparsity of the network, an L1 penalty is imposed on these estimated coefficients. This contraction is controlled by a penalty parameter that, instead of being arbitrarily chosen, in this model, is chosen based on the extended Bayesian information criterion, which has been shown (van Borkulo et al., Reference van Borkulo, Boschloo, Borsboom, Penninx, Waldorp and Schoevers2015) to have good metric properties (converges with increasing sample size and has a low false-positive rate). The network visualizations were created using the qgraph package (Epskamp et al., Reference Epskamp, Cramer, Waldorp, Schmittmann and Borsboom2012) (γ = .25). Each node represents an item, and each edge represents the relationship between each pair of items; bolder lines indicate stronger relationships between items.

The centrality of each symptom is represented in a plot with its standardized values for the statistics of strength, closeness, and betweenness, which are calculated as z-scores.

Estimates of the networks were made independently according to gender (male vs. female) and age (65–74 vs. 75–84). The estimation procedure was the same for all, but to facilitate the visual comparability of the results, the nodes of each group were forced to coincide positively using the layout found for the network with all subjects.

The stability of the network obtained for all subjects was estimated with the botnet package by calculating the stability of the correlations between the edges against the loss of subjects using bootstrapping subsamples. The stability index CS(cor = 0.7) (Epskamp et al., Reference Epskamp, Borsboom and Fried2018), used by Epskamp, reports the percentage of subsamples that have found a correlation between the original edge and those of the samples equal or higher than 0.7 and shows reasonable stability for values equal or higher than 0.2.

Results

Characteristics of the Sample

The sample included 555 males and females between 65 and 84 years of age. The mean age was 73.5 years. The lifetime prevalence of BD in the sample was 6.6%, the 12-month prevalence was 4.7%, and the current prevalence was 0.2%. Table 1 shows the sociodemographic characteristics of the sample. As Table 1 shows, the distribution of males and females is consistent with census reports (48% males and 52% females; INE, 2020).

Table 1. Sociodemographic characteristics of the sample (N = 555)

Symptomatology and Differences in Age and Gender

Descriptive statistics with frequencies and percentages of occurrence of each BD symptom for gender (female and male) and age groups (65–74 and 75–84 years) are found in Table 2.

Table 2. Frequencies and percentages of symptoms for gender and age variables in bipolar disorder symptoms

Note: GRAND = inflated self-esteem; SLEEP = the decreased need for sleep; TALK = more talkative than usual; RACIN = a flight of ideas, racing thoughts; DISTR = distractibility; AGIT = psychomotor agitation; PAINF = activities with painful consequences.

* p < .05.

** p < .05.

*** p < .001.

Considering all the participants, the most frequent symptoms were GRAND and AGIT (25% and 20%, respectively), while the remaining symptoms presented much lower values: 10% for PAINF, 6% for SLEEP, 5% for DISTR, and, finally, 4% for TALK and RACIN.

The gender variable presented a very similar frequency distribution for males and females (but not the relationships between symptoms, as will be seen later). For most symptoms, the frequencies of occurrence are slightly higher for males than for females, although these differences are not statistically significant. The only symptom with a significant difference (Z = 4.27, p < .05) was DISTR, with a frequency of 3% for females and 8% for males.

In the age variable, statistically significant differences are found between the two groups in all symptoms except RACIN and PAINF. The greatest presence of symptoms is found in the younger age group (65–74). The symptoms that show the greatest difference in frequencies between the groups (greater value of Z) are SLEEP (Z = 12.66), TALK (Z = .54), and AGIT (Z = 7.02).

Gender-Based Networks

To facilitate the interpretation of the results, a single network with all subjects was estimated, and its layout was used as the layout for the calculated subgroup networks. The stability statistics were calculated using these networks.

The networks for female and male groups can be seen in Figure 1. There is a noticeable difference between the two networks. The network for females is much sparser, with lower density, consisting of two sub-networks: one composed of TALK and RACIN and the other of PAINF, SLEEP, GRAND, and AGIT. The DISTR symptom remains completely isolated from the rest. The network for males turns out to be denser, with greater connections between all the symptoms, being the edge with the greater weight the one integrated by RACIN and GRAND, whereas this edge is nonexistent in the group of females.

Figure 1. Manic symptom networks by gender groups: female (left) and male (right). Each node represents an item, and each link represents a relationship between each pair of items (bolder lines indicate stronger relations). Centrality indices are represented with blue (female) and red (male). Note: GRAND = inflated self-esteem; SLEEP = decreased need for sleep; TALK = more talkative than usual; RACIN = flight of ideas, racing thoughts; DISTR = distractibility; AGIT = psychomotor agitation; PAINF = activities with painful consequences. Centrality indices are represented with blue (female) and red (male). Network stability for edges with all subjects: CS(cor = 0.7) = 0.28.

The symptoms with greater centrality for both groups are GRAND and RACIN, which appear strongly connected in the male group but not in the female group. The greatest difference in centrality is found in the DISTR symptom, which is practically irrelevant in the case of females.

Age-Based Networks

The network for the age groups (65–74 and 75–84) can be found in Figure 2. The network for the younger group has a pattern based on two subnetworks: one composed of the connection between TALK and RACIN (similar to that found in the female group) and the other consisting of the DISTR, AGIT, and GRAND symptoms. The rest of the edges turned out to be weaker. In the network calculated for the older group (75–84), the result was very sparse, reflecting almost only the relationship between PAINF, GRAND, and AGIT. Despite the clear differences in the structure of both networks, centrality is reasonably similar concerning strength (except for PAINF, which for the group of 65–74 has a very low value); however, the other values of centrality do reflect these structural differences, as can be seen in the values of betweenness.

Figure 2. Manic symptom networks by age groups: 65–74 (left) and 75–84 (right). Each node represents an item, and each link represents a relationship between each pair of items (bolder lines indicate stronger relations). Centrality indices are represented with blue (65–74) and red (75–84). Note: GRAND = inflated self-esteem; SLEEP = decreased need for sleep; TALK = more talkative than usual; RACIN = flight of ideas, racing thoughts; DISTR = distractibility; AGIT = psychomotor agitation; PAINF = activities with painful consequences. Centrality indices are represented with blue (65–74) and red (75–84). Network stability for edges with all subjects: CS(cor = 0.7) = 0.28.

Network Stability

The stability of the network estimated with all subjects and calculated as the CS(cor = 0.7) index (percentage of subsamples that have found a correlation between the original edge and those of the samples equal or higher than 0.7), proposed by Epskamp, presents a reasonable value for the edges (0.28); however, it presents low values (lower than 0.05) for the stability of the intercepts of the model and the estimated values of strength.

Discussion

This is the first study in the scientific literature to report the network analysis of DSM-IV-TR BD symptoms in the older adult population, analyzing age and gender differences. The results confirm our hypothesis that it is possible to construct different networks based on gender and age by examining symptoms of BD among people over 65. The results reveal differences in the strength, closeness, and betweenness of the networks according to gender and age, thus supporting the idea of BD as a complex dynamic system, with unique characteristics among people and not as a prototypical classification with an underlying mental disorder.

The gender variable presented a very similar frequency distribution for males and females. For most symptoms, the frequencies of occurrence are slightly higher for males than for females; the only symptom with a significant difference was DISTR with a frequency of 3% for females and 8% for males. These results can be explained by the fact that this study evaluated symptoms compatible with BD type I, which are manic symptoms (which are the data available in the MentDis_ICF65+ study), and it is known that males more often debut with a manic episode, whereas females seem more likely to present with depression (Dell’Osso et al., Reference Dell’Osso, Cafaro and Ketter2021). Therefore, it is to be expected that when studying only symptoms of mania, it is men who score higher on this symptomatology. If we had studied major depressive symptoms compatible with BD type II, we would have expected to find more of this type of symptomatology in women.

Concerning the networks developed in terms of gender, females present a network that is much more sparse, with a lower density and consisting of two sub-networks: one composed of TALK and RACIN, and the other of PAINF, SLEEP, GRAND, and AGIT. The DISTR symptom remains completely isolated from the rest. On the contrary, in the case of males, a denser network is obtained, with greater connections between all the symptoms, being the edge with greater weight the one integrated by RACIN and GRAND, whereas this edge is nonexistent in the group of females. The symptoms with greater centrality for both groups are GRAND and RACIN, which appear strongly connected in the male group but not in the female group. The greatest difference in centrality is found in the DISTR symptom, which is practically irrelevant in the case of females.

To the best of our knowledge, there are no precedent studies on network analysis in BD and its differences by gender that would allow us to compare our findings. Concerning gender differences in BD, it is known that BD in females is unique in its presentation, and it is characterized by a later age of onset, seasonality, atypical presentation, and a higher degree of mixed episodes (Parial, Reference Parial2015). Perhaps differences in BD between males and females develop with exposure to life events and experiences, being greater in adulthood when an individual’s full development has already occurred, and differences in experiences may be more determined by gender (Castellanos et al., Reference Castellanos, Ausín, Bestea, González-Sanguino and Muñoz2020).

These gender differences could be critical in making a more complete evaluation of the BD and an early and more accurate diagnosis, being information that is not provided through traditional categorical classifications. In this sense, one repercussion in clinical practice related to the intervention of BD in old age could address primarily the symptoms of GRAND (inflated self-esteem) and AGIT (psychomotor agitation). In females with BD symptoms, it would presumably be more effective to carry out interventions targeting two distinct symptom sub-networks: one composed of TALK (more talkative than usual) and RACIN (a flight of ideas, racing thoughts), and the other of PAINF (activities with painful consequences), SLEEP (the decreased need for sleep), GRAND (inflated self-esteem), and AGIT (psychomotor agitation), while in males, perhaps more useful interventions would focus on addressing RACIN (a flight of ideas, racing thoughts) and GRAND (inflated self-esteem) symptomatology. Similarly, the clinician should pay particular attention to certain differential BD symptoms in males and females in the assessment.

Regarding age, it is possible to observe changes in the model between the two age groups. In people aged 65–74, the network obtained has a pattern based on two subnetworks: one composed of the connection between TALK and RACIN (similar to that found with the female group), and the other consisting of the DISTR, AGIT, and GRAND symptoms. Nevertheless, the network in people aged 75–84 was very sparse, reflecting almost only the relationship between PAINF, GRAND, and AGIT.

The results obtained in the different networks developed show that the symptomatology of BD behaves differently in females and males, and also in different age groups in people over 65. As clinical repercussions relate to the intervention of BD, it is also of interest to consider the clinical case formulation approach, in which treatments are personalized, focusing on the symptoms that appear most frequently in each of the networks found, taking age and gender into account. In addition, focusing on symptoms and the relationship between symptoms aligns well with clinical practice, such as cognitive behavioral therapy (Borsboom et al., Reference Borsboom, Cramer and Kalis2018). Furthermore, the network approach allows for individual heterogeneity, because every individual can have their own network structure (Bringmann, Reference Bringmann2021). The use of individual symptom networks has important clinical implications. On the one hand, it could be used to uncover the person-specific risk factors for developing mental disorders (Borsboom & Cramer, Reference Borsboom and Cramer2013). On the other hand, knowing that a specific symptom of a person is a central symptom in the network structure of a patient can pave the way for idiographic targeted treatment (Bringmann, Reference Bringmann2021). Furthermore, the usefulness of using the clinical formulation or a network analysis of a person’s symptoms goes beyond the overall explanation of the case and extends to organizing complex and sometimes contradictory information; to guiding treatment planning and direction; to coordinating the intervention team; to therapist-client and interprofessional communication; to using it as a basis for assessing change; as a support for the therapist’s understanding of the patient; and, finally, as a re-organizing element of the patient’s experience, narratives, and identity (Muñoz et al., Reference Muñoz, Ausín and Panadero2019).

Although this study highlights the possibilities of psychopathological symptom combinations, however, only two studies have focused on the analysis of symptom networks in the over-65 population and their clinical applications (Belvederi Murri et al., Reference Belvederi Murri, Amore, Respino and Alexopoulos2020; Castellanos et al., Reference Castellanos, Ausín, Bestea, González-Sanguino and Muñoz2020), and they have done so by studying major depressive disorder. Castellanos et al. (Reference Castellanos, Ausín, Bestea, González-Sanguino and Muñoz2020) suggest that in females with major depressive disorder, it would presumably be more effective to carry out interventions at the emotional level, along with behavioral and sleep activation guidelines, while in males, perhaps more useful interventions would focus on addressing somatic-type problems. Based on the results of the symptom network analysis for major depression in older people, Belvederi Murri et al. (Reference Belvederi Murri, Amore, Respino and Alexopoulos2020) found that death wishes, depressed mood, loss of interest, and pessimism had the highest values of centrality in the network structure of depressive symptoms in late-life and suggest that these symptoms may be used as targets for novel, focused interventions and in studies investigating neurobiological processes central to late-life depression.

On the other hand, McNally (Reference McNally2016) notes that, even though no single method in the field of psychopathology is likely to provide answers to all the questions about the origins and treatment of psychological disorders, network analysis holds promise as both a scientific and practical approach to conceptualizing and guiding the treatment of these conditions. Despite the important repercussions at a clinical level that adopting a dimensional perspective of psychopathology can have, Contreras et al. (Reference Contreras, Nieto, Valiente, Espinosa and Vazquez2019) highlight that the clinical utility of network analysis is still uncertain, as there are important limitations on the types of data included, the analytic procedures, and the psychometric and clinical validity of the results. For these authors, it would be premature to defend that central variables found in specific studies should automatically become new intervention targets.

One of the main limitations of the study that we found was related to the representativeness of the sample. In the present study, exclusion criteria were applied for various technical reasons, barring people with a severe cognitive deficit or who could not be interviewed due to some sort of cognitive deficit, or being nursing home residents, homeless, non-Spanish speakers, or people over 85 years old, which is a limitation to the study. On the other hand, the results are based on DSM-IV-TR criteria for BD, and using DSM-5 diagnoses could have led to different results. In addition, it should be noted that it would have been interesting to have data on current symptomatology and not only for the last 12 months, and on other interesting variables such as medication or comorbidity. In addition, the results of this study are limited because this study evaluates symptoms compatible with BD type I, but does not include symptoms compatible with BD type II.

In conclusion, network differences found in this study seem to support the conceptualization of OABD from a dimensional point of view. These results highlight the importance of considering gender and age when studying older people with symptoms compatible with BD, which could facilitate the personalization of treatments offered to this population in healthcare services.

Nevertheless, future studies would have to advance to demonstrate these differential networks between males and females and to propose differentiated assessments and treatments based on these BD symptom networks. This study provides a starting point to question the uniformity of interventions in BD and the need for further research in this area in the future.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author (B.A.). The data are not publicly available due to restrictions, for example, their containing information that could compromise the privacy of research participants.

Acknowledgements

The authors would like to thank Professor Martino Belvederi Murri (Department of Clinical and Experimental Medicine, University of Ferrara, Italy) for his comments on a draft of this manuscript. They gratefully thank all participants and all interviewers in their study.

Author contribution

Conceptualization: B.A., M.M.; data curation: M.A.C.; formal analysis: M.A.C.; funding acquisition: B.A., M.M.; investigation: B.A.; methodology: B.A., M.M.; project administration: M.M.; resources: B.A., M.M.; supervision: B.A., M.M.; visualization: M.A.C.; writing—original draft: B.A., M.A.C.; writing—review and editing: all authors. All authors have read and agreed to the published version of the manuscript.

Funding statement

This work was supported by the European Commission under Grant No. 223105.

Competing interests

The authors declare no competing interests.

References

Altshuler, L. L., Kupka, R. W., Hellemann, G., Frye, M. A., Sugar, C. A., McElroy, S. L., Nolen, W. A., Grunze, H., Leverich, G. S., Keck, P. E., Zermeno, M., Post, R. M., & Suppes, T. (2010). Gender and depressive symptoms in 711 patients with bipolar disorder evaluated prospectively in the Stanley Foundation bipolar treatment outcome network. The American Journal of Psychiatry, 167(6), 708–715. https://doi.org/10.1176/appi.ajp.2009.09010105.CrossRefGoogle ScholarPubMed
American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders. American Psychiatric Association.Google Scholar
American Psychiatric Association (2014). Diagnostic and statistical manual of mental disorders DSM-5 (5. ed.). Madrid: Editorial Médica Panamericana.Google Scholar
Andreas, S., Härter, M., Volkert, J., Hausberg, M., Sehner, S., Wegscheider, K., Rabung, S., Ausín, B., Canuto, A., Da Ronch, C., Grassi, L., Hershkovitz, Y., Lelliott, P., Muñoz, M., Quirk, A., Rotenstein, O., Santos-Olmo, A. B., Shalev, A., Siegert, J.,… Schulz, H. (2013). The MentDis_ICF65+ study protocol: Prevalence, 1-year incidence and symptom severity of mental disorders in the elderly and their relationship to impairment, functioning (ICF) and service utilisation. BMC Psychiatry, 13(1), 1–10. https://doi.org/10.1186/1471-244X-13-62CrossRefGoogle ScholarPubMed
Andreas, S., Schulz, H., Volkert, J., Dehoust, M., Sehner, S., Suling, A., Ausín, B., Canuto, A., Crawford, M., Da Ronch, C., Grassi, L., Hershkovitz, Y., Muñoz, M., Quirk, A., Rotenstein, O., Santos-Olmo, A. B., Shalev, A., Strehle, J., Weber, K., … Härter, M. (2017). Prevalence of mental disorders in elderly people: The European MentDis_ICF65+ study. The British Journal of Psychiatric, 210(2) 125–131 https://doi.org/10.1192/bjp.bp.115.180463.CrossRefGoogle ScholarPubMed
Arnold, L. M. (2003). Gender differences in bipolar disorder. The Psychiatric Clinics of North America, 26(3), 595–620. https://doi.org/10.1016/s0193-953x(03)00036-4.CrossRefGoogle ScholarPubMed
Ausín, B., Muñoz, M., Santos-Olmo, A. B., Pérez-Santos, E, & Castellanos, M. A. (2017). Prevalence of mental disorders in the elderly in the Community of Madrid: Results of the Mentdis_ICF65+ study. The Spanish Journal of Psychology, 20, Article e6. https://doi.org/10.1017/sjp.2017.3.CrossRefGoogle Scholar
Belvederi Murri, M., Amore, M., Respino, M., & Alexopoulos, G. S. (2020). The symptom network structure of depressive symptoms in late-life: Results from a European population study. Molecular Psychiatry, 25(7), 1447–1456. https://doi.org/10.1038/s41380-018-0232-0.CrossRefGoogle ScholarPubMed
Blanken, M. A. J. T., Oudega, M. L., Almeida, O. P., Schouws, S. N. T. M., Orhan, M., Beunders, A. J. M., Klumpers, U. M. H., Sonnenberg, C., Blumberg, H. P., Eyler, L. T., Forester, B. P., Forlenza, O. V., Gildengers, A., Mulsant, B. H., Rajji, T., Rej, S., Sarna, K., Sutherland, A., Yala, J.,… Dols, A. (2024). Sex differences among older adults with bipolar disorder: Results from the Global Aging & Geriatric Experiments in Bipolar Disorder (GAGE-BD) project The American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry, 32(3), 326–338. https://doi.org/10.1016/j.jagp.2023.10.008.CrossRefGoogle ScholarPubMed
Borsboom, D. (2017). A network theory of mental disorders. World Psychiatry: Official Journal of the World Psychiatric Association, 16(1), 5–13. https://doi.org/10.1002/wps.20375.CrossRefGoogle ScholarPubMed
Borsboom, D., & Cramer, A. O. (2013). Network analysis: An integrative approach to the structure of psychopathology. Annual Review of Clinical Psychology, 9, 91–121. https://doi.org/10.1146/annurev-clinpsy-050212-185608.CrossRefGoogle Scholar
Borsboom, D., Cramer, A. O. J., & Kalis, A. (2018). Brain disorders? Not really: Why network structures block reductionism in psychopathology research. The Behavioral and Brain Sciences, 42, Article e2. https://doi.org/10.1017/S0140525X17002266.CrossRefGoogle ScholarPubMed
Bringmann, L. F. (2021). Person-specific networks in psychopathology: Past, present, and future. Current Opinion in Psychology, 41, 59–64. https://doi.org/10.1016/j.copsyc.2021.03.004.CrossRefGoogle ScholarPubMed
Castellanos, M. A., Ausín, B., Bestea, S., González-Sanguino, C., & Muñoz, M (2020). A network analysis of major depressive disorder symptoms and age and gender-related differences in people over 65 in a Madrid Community sample (Spain). International Journal of Environmental Research and Public Health, 17(23), Article 8934. https://doi.org/10.3390/ijerph17238934.CrossRefGoogle Scholar
Chen, P., Dols, A., Rej, S., & Sajatovic, M. (2017). Update on the epidemiology, diagnosis, and treatment of mania in older-age bipolar disorder. Current Psychiatry Reports, 19(8), Article 46. https://doi.org/10.1007/s11920-017-0804-8.CrossRefGoogle ScholarPubMed
Contreras, A., Nieto, I., Valiente, C., Espinosa, R., & Vazquez, C. (2019). The study of psychopathology from the network analysis perspective: A systematic review. Psychotherapy and Psychosomatics, 88(2), 71–83. https://doi.org/10.1159/000497425.CrossRefGoogle Scholar
Corponi, F., Anmella, G., Verdolini, N., Pacchiarotti, I., Samalin, L., Popovic, D., Azorin, J. M., Angst, J., Bowden, C. L., Mosolov, S., Young, A. H., Perugi, G., Vieta, E., & Murru, A. (2020). Symptom networks in acute depression across bipolar and major depressive disorders: A network analysis on a large, international, observational study. European Neuropsychopharmacology: The Journal of the European College of Neuropsychopharmacology, 35, 49–60. https://doi.org/10.1016/j.euroneuro.2020.03.017.CrossRefGoogle ScholarPubMed
Cotrena, C., Branco, L. D., Kochhann, R., Shansis, F. M., & Fonseca, R. P. (2016). Quality of life, functioning and cognition in bipolar disorder and major depression: A latent profile analysis. Psychiatry Research, 241, 289–296. https://doi.org/10.1016/j.psychres.2016.04.102.CrossRefGoogle ScholarPubMed
Cunningham, R., Crowe, M., Stanley, J., Haitana, T., Pitama, S., Porter, R., Baxter, J., Huria, T., Mulder, R., Clark, M., & Lacey, C. (2020). Gender and mental health service use in bipolar disorder: National cohort study. BJPsych Open, 6(6), e138. https://doi.org/10.1192/bjo.2020.117.CrossRefGoogle ScholarPubMed
Curtiss, J., Fulford, D., Hofmann, S. G., & Gershon, A. (2019). Network dynamics of positive and negative affect in bipolar disorder. Journal of Affective Disorders, 249, 270–277. https://doi.org/10.1016/j.jad.2019.02.017.CrossRefGoogle ScholarPubMed
De la Fuente-Tomás, L., Sierra, P., Sanchez-Autet, M., García-Blanco, A., Safont, G., Arranz, B., & García-Portilla, M. P. (2018). Sleep disturbances, functioning, and quality of life in euthymic patients with bipolar disorder. Psychiatry Research, 269, 501–507. https://doi.org/10.1016/j.psychres.2018.08.104.CrossRefGoogle ScholarPubMed
Dell’Osso, B., Cafaro, R., & Ketter, T. A. (2021). Has bipolar disorder become a predominantly female gender related condition? Analysis of recently published large sample studies. International Journal of Bipolar Disorders, 9(1), 3. https://doi.org/10.1186/s40345-020-00207-z.CrossRefGoogle ScholarPubMed
Depp, C. A., Davis, C. E., Mittal, D., Patterson, T. L., & Jeste, D. V. (2006). Health-related quality of life and functioning of middle-aged and elderly adults with bipolar disorder. The Journal of Clinical Psychiatry, 67(2), 215–221. https://doi.org/10.4088/jcp.v67n0207.CrossRefGoogle ScholarPubMed
Dias, R. D. S., Kerr-Corrêa, F., Torresan, R. C., & Santos, C. H. R. D. (2006). Transtorno bipolar do humor e gênero [Bipolar mood disorder and gender]. Archives of Clinical Psychiatry (São Paulo), 33, 80–91.CrossRefGoogle Scholar
Dols, A., & Beekman, A. (2018). Older age bipolar disorder. The Psychiatric Clinics of North America, 41(1), 95–110. https://doi.org/10.1016/j.psc.2017.10.008.CrossRefGoogle ScholarPubMed
Epskamp, S., Borsboom, D., & Fried, E. I. (2018). Estimating psychological networks and their accuracy: A tutorial paper. Behaviour Research Methods, 50(1), 195–212. https://doi.org/10.3758/s13428-017-0862-1CrossRefGoogle ScholarPubMed
Epskamp, S., Cramer, A. O. J., Waldorp, L. J., Schmittmann, V. D., & Borsboom, D. (2012). Qgraph: Network visualizations of relationships in psychometric data. Journal of Statistical Software, 48(4), 1–18. https://doi.org/10.18637/jss.v048.i04CrossRefGoogle Scholar
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. https://doi.org/10.1016/0022-3956(75)90026-6CrossRefGoogle Scholar
Instituto Nacional de Estadística. (2020). Censo de poblaciones y viviendas (Census of population and housing). https://www.ine.es/dyngs/INEbase/es/operacion.htm?c=Estadistica_C&cid=1254736176951&menu=ultiDatos&idp=1254735572981Google Scholar
McNally, R. J. (2016). Can network analysis transform psychopathology? Behaviour Research and Therapy, 86, 95–104. https://doi.org/10.1016/j.brat.2016.06.006CrossRefGoogle ScholarPubMed
Muñoz, M., Ausín, B., & Panadero, S. (2019). Manual práctico de evaluación psicológica clínica. Síntesis.Google Scholar
Parial, S. (2015). Bipolar disorder in women. Indian Journal of Psychiatry, 57(Suppl 2), S252–S263. https://doi.org/10.4103/0019-5545.161488.CrossRefGoogle ScholarPubMed
Peralta, V., Gil-Berrozpe, G. J., Sánchez-Torres, A., & Cuesta, M. J. (2020). The network and dimensionality structure of affective psychoses: An exploratory graph analysis approach. Journal of Affective Disorders, 277, 182–191. https://doi.org/10.1016/j.jad.2020.08.008.CrossRefGoogle Scholar
Sajatovic, M., Strejilevich, S. A., Gildengers, A. G., Dols, A., Al Jurdi, R. K., Forester, B. P., Kessing, L. V., Beyer, J., Manes, F., Rej, S., Rosa, A. R., Schouws, S. N., Tsai, S. Y., Young, R. C., & Shulman, K. I. (2015). A report on older-age bipolar disorder from the International Society for Bipolar Disorders Task Force. Bipolar Disorders, 17(7), 689–704. https://doi.org/10.1111/bdi.12331.CrossRefGoogle ScholarPubMed
Strauss, G. P., Esfahlani, F. Z., Kirkpatrick, B., Allen, D. N., Gold, J. M., Visser, K. F., & Sayama, H. (2019). Network analysis reveals which negative symptom domains are most central in schizophrenia vs bipolar disorder. Schizophrenia Bulletin, 45(6), 1319–1330. https://doi.org/10.1093/schbul/sby168.CrossRefGoogle ScholarPubMed
Tatay-Manteiga, A., Cauli, O., Tabarés-Seisdedos, R., Michalak, E. E., Kapczinski, F., & Balanzá-Martínez, V. (2019). Subjective neurocognition and quality of life in patients with bipolar disorder and siblings. Journal of Affective Disorders, 245, 283–288. https://doi.org/10.1016/j.jad.2018.11.012.CrossRefGoogle ScholarPubMed
van Borkulo, C., Boschloo, L., Borsboom, D., Penninx, B. W., Waldorp, L. J., & Schoevers, R. A. (2015). Association of symptom network structure with the course of depression. JAMA Psychiatry, 72(12), 1219–1226. https://doi.org/10.1001/jamapsychiatry.2015.2079CrossRefGoogle Scholar
Volkert, J., Schulz, H., Härter, M., Wlodarcyzk, O., & Andreas, S. (2013). The prevalence of mental disorders in older people in Western countries—a meta-analysis. Ageing Research Reviews, 12, 339–353 https://doi.org/10.1016/j.arr.2012.09.004.CrossRefGoogle ScholarPubMed
Weintraub, M. J., Schneck, C. D., & Miklowitz, D. J. (2020). Network analysis of mood symptoms in adolescents with or at high risk for bipolar disorder. Bipolar Disorders, 22(2), 128–138. https://doi.org/10.1111/bdi.12870.CrossRefGoogle ScholarPubMed
Wittchen, H. U., Strehle, J., Gerschler, A., Volkert, J., Dehoust, M., Sehner, S., Wegscheider, B., Ausìn, B., Canuto, A., Crawford, M., Da Ronch, C., Grassi, L., Hershkovitz, Y., Muñoz, M., Quirk, A., Rotenstein, O., Santos-Olmo, A.B., Shalev, A., Weber, K., … Andreas, S. (2014). Measuring symptoms and diagnosing mental disorders in the elderly community: The test-retest reliability of the CIDI65+. International Journal of Methods in Psychiatric Research, 24(2), 116–129. https://doi.org/10.1002/mpr.1455CrossRefGoogle Scholar
Figure 0

Table 1. Sociodemographic characteristics of the sample (N = 555)

Figure 1

Table 2. Frequencies and percentages of symptoms for gender and age variables in bipolar disorder symptoms

Figure 2

Figure 1. Manic symptom networks by gender groups: female (left) and male (right). Each node represents an item, and each link represents a relationship between each pair of items (bolder lines indicate stronger relations). Centrality indices are represented with blue (female) and red (male). Note: GRAND = inflated self-esteem; SLEEP = decreased need for sleep; TALK = more talkative than usual; RACIN = flight of ideas, racing thoughts; DISTR = distractibility; AGIT = psychomotor agitation; PAINF = activities with painful consequences. Centrality indices are represented with blue (female) and red (male). Network stability for edges with all subjects: CS(cor = 0.7) = 0.28.

Figure 3

Figure 2. Manic symptom networks by age groups: 65–74 (left) and 75–84 (right). Each node represents an item, and each link represents a relationship between each pair of items (bolder lines indicate stronger relations). Centrality indices are represented with blue (65–74) and red (75–84). Note: GRAND = inflated self-esteem; SLEEP = decreased need for sleep; TALK = more talkative than usual; RACIN = flight of ideas, racing thoughts; DISTR = distractibility; AGIT = psychomotor agitation; PAINF = activities with painful consequences. Centrality indices are represented with blue (65–74) and red (75–84). Network stability for edges with all subjects: CS(cor = 0.7) = 0.28.