Introduction
White matter hyperintensities (WMH) are a very common magnetic resonance imaging (MRI) finding in older individuals, appearing with increased signal on T2-weighted and fluid-attenuated inversion recovery sequences. While they are most often regarded as a feature of small vessel disease, multiple pathophysiological mechanisms have been incriminated such as blood-brain barrier leakage, neuroinflammation and neurodegeneration (Fernando et al., Reference Fernando, Simpson, Matthews, Brayne, Lewis, Barber, Kalaria, Forster, Esteves, Wharton, Shaw, O’Brien and Ince2006; Shim et al., Reference Shim, Yang, Roe, Coats, Benzinger, Xiong, Galvin, Cairns and Morris2015). Demyelination, axonal loss and gliosis with variable pathological severity are also among the nonspecific neuropathological substrates of WMH (Gouw et al., Reference Gouw, Seewann, van der Flier, Barkhof, Rozemuller, Scheltens and Geurts2011). WMH prevalence increases with age and vascular risk burden (Ryu et al., Reference Ryu, Woo, Schellingerhout, Chung, Kim, Jang, Park, Hong, Jeong, Na, Cho, Kim, Kim, Han, Lee, Cha, Kim, Lee, Ko, Cho, Lee, Yu, Oh, Park, Kang, Lee, Park, Lee, Choi, Lee, Bae and Kim2014). WMH have been related to cognitive decline, incident stroke and dementia, as well as fatigue, physical (e.g., imbalance, gait abnormalities), and neuropsychiatric symptoms (NPS) (Clancy et al., Reference Clancy, Gilmartin, Jochems, Knox, Doubal and Wardlaw2021; Debette and Markus, Reference Debette and Markus2010).
NPS (such as depression, anxiety, delusions, and apathy) are non-cognitive disturbances that are quite prevalent in individuals with MCI and almost universal in those with dementia (Lyketsos et al., Reference Lyketsos, Carrillo, Ryan, Khachaturian, Trzepacz, Amatniek, Cedarbaum, Brashear and Miller2011, Reference Lyketsos, Lopez, Jones, Fitzpatrick, Breitner and DeKosky2002). Among cognitively unimpaired (CU) older adults, NPS have been linked to worse cognitive test performance (Liampas et al., Reference Liampas, Siokas, Lyketsos and Dardiotis2022b), more precipitous cognitive decline (Krell-Roesch et al., Reference Krell‐Roesch, Syrjanen, Machulda, Christianson, Kremers, Mielke, Knopman, Petersen, Vassilaki and Geda2021) and an elevated hazard of Alzheimer’s (AD) or non-AD dementia (Liew, Reference Liew2020). In individuals with MCI, NPS have been linked to steeper cognitive trajectories (Roberto et al., Reference Roberto, Portella, Marquié, Alegret, Hernández, Mauleón, Rosende-Roca, Abdelnour, Esteban de Antonio, Tartari, Vargas, López-Cuevas, Bojaryn, Espinosa, Ortega, Pérez-Cordón, Sanabria, Orellana, de Rojas, Moreno-Grau, Montrreal, Alarcón-Martín, Ruíz, Tárraga, Boada and Valero2021) and inflated risk of future dementia (Liew, Reference Liew2019), while in those with dementia, NPS are a harbinger of more abrupt cognitive decline (Defrancesco et al., Reference Defrancesco, Marksteiner, Kemmler, Dal-Bianco, Ransmayr, Benke, Mosbacher, Höller and Schmidt2020), among other unfavorable outcomes (Bränsvik et al., Reference Bränsvik, Granvik, Minthon, Nordström and Nägga2021). Therefore, the presence of NPS in older adults should be regarded as a precursor of cognitive decline throughout the normal cognitive aging to dementia continuum.
Of note, in the continuum of healthy aging – dementia, Taragano and colleagues introduced the construct of mild behavioral impairment (MBI) – the neuropsychiatric equivalent of MCI –, as a transitional stage between normal aging and dementia which confers greater risk of incident dementia than MCI (Taragano et al., Reference Taragano, Allegri, Krupitzki, Sarasola, Serrano, Loñ and Lyketsos2009). Although an affinity towards non-AD dementias is apparent (frontotemporal dementia [FTD] and Lewy body dementia [LBD]), many individuals with MBI may convert to AD as well, owing to its considerably larger prevalence (Taragano et al., Reference Taragano, Allegri, Heisecke, Martelli, Feldman, Sánchez, García, Tufro, Castro, Leguizamón, Guelar, Ruotolo, Zegarra, Dillon and Lanctôt2018). The predominant hypothesis suggests that the association between NPS and cognitive decline probably reflects the relationship of NPS with undergoing neuropathological alterations (Peters and Lyketsos, Reference Peters and Lyketsos2015). Different NPS have been related to different neurodegenerative processes and by extension to heterogeneous cognitive trajectories and progression to different neurocognitive entities. For instance, psychosis has been linked to neuritic plaques, neurofibrillary tangles and Lewy body disease (an in turn, incident AD and Lewy body dementia [LBD]), agitation and aggression have been associated with TDP-43 pathology (a common substrate of FTD), whereas results on the involvement of vascular lesions in MBI have been inconsistent (Devanand et al., Reference Devanand, Lee, Huey and Goldberg2022; Matsuoka et al., Reference Matsuoka, Imai and Narumoto2023).
To date, few have investigated the relationship between WMH and NPS in individuals without dementia (MCI or normal cognition). These studies had several limitations such as small samples, not accounting for the confounding of neurocognitive status with vascular risk, non-distinctions between CU and MCI individuals, use of composite NPS scores over individual NPS (and then focus on depression and anxiety). It is not surprising that they have reported contradictory results (Chan et al., Reference Chan, Pettigrew, Soldan, Zhu, Wang, Albert, Rosenberg and Ismail2022; Clancy et al., Reference Clancy, Gilmartin, Jochems, Knox, Doubal and Wardlaw2021; Miao et al., Reference Miao, Chen, Robert, Smith and Ismail2021; Misquitta et al., Reference Misquitta, Dadar, Louis Collins and Tartaglia2020; Staekenborg et al., Reference Staekenborg, Gillissen, Romkes, Pijnenburg, Barkhof, Scheltens and van der Flier2008; Tumati et al., Reference Tumati, Boyd, Ismail, Mah, Seitz, Herrmann and Lanctôt2023; Yang et al., Reference Yang, Shu, Yan, Yang, Xu and Wei2022). The aim in this was to estimate the associations between WMH and individual NPS in older adults without dementia. Specifically, we hypothesized that the odds of having NPS would be to the presence of WMH in both CU and MCI.
Methods and materials
This cross-sectional analysis capitalized on data from the ongoing Uniform Data Set (UDS). UDS is a central repository of multidisciplinary, longitudinally collected data by National Institute on Aging / National Institute of Health - funded Alzheimer’s Disease Research Centers (ADRCs) across the United States (Beekly et al., Reference Beekly, Ramos, Lee, Deitrich, Jacka, Wu, Hubbard, Koepsell, Morris and Kukull2007; Morris et al., Reference Morris, Weintraub, Chui, Cummings, DeCarli, Ferris, Foster, Galasko, Graff-Radford, Peskind, Beekly, Ramos and Kukull2006; Siokas et al., Reference Siokas, Liampas, Lyketsos and Dardiotis2022; Weintraub et al., Reference Weintraub, Salmon, Mercaldo, Ferris, Graff-Radford, Chui, Cummings, DeCarli, Foster, Galasko, Peskind, Dietrich, Beekly, Kukull and Morris2009). UDS was initiated in 2005 and since has been stewarded by the National Alzheimer’s Coordinating Center (NACC). Clinician-, self- and family-referred volunteers, or actively recruited individuals with a cognitive status ranging from normal cognition to dementia are enrolled according to each ADRC’s discrete protocol. Standardized evaluations take place on an approximately annual basis. Participants or surrogates provide informed consent before participation. All procedures are overseen by local Institutional Review Board(s) and performed in accordance with the ethical standards of the declaration of Helsinki and its later amendments. For further information on access to the NACC database, please contact NACC at https://naccdata.org/.
Eligibility criteria and diagnostic procedures
The current analysis was based on UDS data from the December 2022 data freeze, collected from a total of 46 ADRCs. Older (≥50 years) participants with available data on WMH (only the 1st visit with available data was considered for eligibility) and a concurrent diagnosis of MCI or CU, were eligible (those with dementia or cognitive impairment not MCI were excluded). Cognitive diagnoses were established by either expert consensus panels (in the majority of cases) or single physicians (i.e., those who conducted the examination), according to each ADRC’s discrete protocol. CU was defined by the absence of a diagnosis of dementia, MCI or cognitive impairment not MCI. MCI [subjective and objective (based on the typical threshold of 1.5 standard deviations) cognitive disorder in the absence of repercussions on daily life] and dementia were diagnosed using standard clinical criteria (McKeith et al., Reference McKeith, Boeve, Dickson, Halliday, Taylor, Weintraub, Aarsland, Galvin, Attems, Ballard, Bayston, Beach, Blanc, Bohnen, Bonanni, Bras, Brundin, Burn, Chen-Plotkin, Duda, El-Agnaf, Feldman, Ferman, ffytche, Fujishiro, Galasko, Goldman, Gomperts, Graff-Radford, Honig, Iranzo, Kantarci, Kaufer, Kukull, Lee, Leverenz, Lewis, Lippa, Lunde, Masellis, Masliah, McLean, Mollenhauer, Montine, Moreno, Mori, Murray, O'Brien, Orimo, Postuma, Ramaswamy, Ross, Salmon, Singleton, Taylor, Thomas, Tiraboschi, Toledo, Trojanowski, Tsuang, Walker, Yamada and Kosaka2017; McKhann et al., Reference McKhann, Drachman, Folstein, Katzman, Price and Stadlan1984; Neary et al., Reference Neary, Snowden, Gustafson, Passant, Stuss, Black, Freedman, Kertesz, Robert, Albert, Boone, Miller, Cummings and Benson1998; Petersen et al., Reference Petersen, Smith, Waring, Ivnik, Tangalos and Kokmen1999; Roman et al., Reference Román, Tatemichi, Erkinjuntti, Cummings, Masdeu, Garcia, Amaducci, Orgogozo, Brun, Hofman, Moody, O’Brien, Yamaguchi, Grafman, Drayer, Bennett, Fisher, Ogata, Kokmen, Bermejo, Wolf, Gorelick, Bick, Pajeau, Bell, DeCarli, Culebras, Korczyn, Bogousslavsky, Hartmann and Scheinberg1993; Winblad et al., Reference Winblad, Palmer, Kivipelto, Jelic, Fratiglioni, Wahlund, Nordberg, Bäckman, Albert, Almkvist, Arai, Basun, Blennow, De Leon, DeCarli, Erkinjuntti, Giacobini, Graff, Hardy, Jack, Jorm, Ritchie, Van Duijn, Visser and Petersen2004). Participants with cognitive impairment who did not clearly fit into these categories were diagnosed with cognitive impairment – not MCI.
Measurement of NPS
The Neuropsychiatric Inventory Questionnaire (NPI-Q) is an informant administered, widely used tool for the evaluation of NPS in dementia research (Kaufer et al., Reference Kaufer, Cummings, Ketchel, Smith, MacMillan, Shelley, Lopez and DeKosky2000; Liampas et al., Reference Liampas, Siokas, Zoupa, Lyketsos and Dardiotis2024). NPI-Q evaluates 12 domains: delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability, aberrant motor behavior, night-time behaviors, and eating behaviors. Informants initially report the presence or absence of cardinal symptomatology for each domain in the month preceding the examination and subsequently rate the severity of any symptoms according to a 3-point severity scale: mild (noticeable, but not a significant change); moderate (significant, but not a dramatic change); or severe (very marked or prominent; a dramatic change) (Liampas et al., Reference Liampas, Siokas, Lyketsos and Dardiotis2022c). For the purposes of the current analysis, participants were dichotomized for presence of each NPS (0: absent; 1: present). Delusions and hallucinations were grouped together (psychotic symptoms) owing to their very low prevalence. Two additional composite NPS indices were analyzed: total number of NPS (0–11) and total NPS severity score (0–22). For the latter, absence of NPS was scored with 0, mild symptomatology conferred 1 point and moderate to severe symptomatology conferred 2 points.
Measurement of WMH
The Cardiovascular Health Study (CHS) score is an ordinal quantification scale that uses visual inspection of a MRI to WMH burden ranging from 0 to 8 (Manolio et al., Reference Manolio, Kronmal, Burke, Poirier, O'Leary, Gardin, Fried, Steinberg and Bryan1994). Periventricular and subcortical volumes of WMH are assessed on spin density-weighted axial images and scored between 0 (no WMH) and 8 (extensive, confluent WMH). Areas of large vessel infarction or small vessel lacunar strokes are excluded from the scoring. In the NACC database, based on WMH status, participants are coded as follows: with no to mild WMH (CHS score: 0–4), with moderate WMH (CHS score: 5–6) and with extensive WMH (CHS score: 7–8).
Covariates considered
Chronological age upon entry to the study and years of formal education were treated as scale variables. Sex, race (Caucasian, African American, other) and the following comorbidities were treated as categorical (yes/no) variables: cerebrovascular disease (CEVD), atrial fibrillation (AF), hypertension, diabetes mellitus (DM) and dyslipidaemia. CEVD was defined as a positive history of stroke and/or transient ischemic attack. Comorbidities were primarily evaluated based on participant or co-participant reporting. However, to limit the amount of missing data, clinician reporting was utilized when necessary.
Regarding cognition, the five MCI subtypes in the NACC database were treated as dichotomous variables, i.e., MCI memory: yes/no; MCI language: yes/no; MCI executive function: yes/no; MCI visuospatial: yes/no; MCI attention: yes/no. On the other hand, Mini-Mental State Examination (MMSE) scores were treated as scale variables. Instead of MMSE, Montreal Cognitive Assessment (MoCA) was utilized in the last (3rd) version of UDS. To limit the amount of missing data, MoCA values were converted to MMSE scores according to conversion tables provided by a NACC crosswalk study (Liampas et al., Reference Liampas, Siokas, Lyketsos and Dardiotis2023; Monsell et al., Reference Monsell, Dodge, Zhou, Bu, Besser, Mock, Hawes, Kukull and Weintraub2016).
Statistical analysis
Individuals with unimpaired cognition or with MCI were analyzed separately using the same approach. Demographic and other characteristics of the three groups defined by WMH status were compared using analysis of variance (ANOVA; scale variables) and Pearson’s chi-squared test (categorical variables). The frequencies of different NPS subtypes were also compared between the three WMH groups using Pearson’s chi-squared test.
The unadjusted and adjusted odds, and 95% confidence intervals, of having each NPS by WMH severity were estimated. Binary logistic regression models were performed to estimate the adjusted odds of reporting each NPS by WMH status. Models were adjusted for age, years of education, sex, race, history of CeVD, AF, hypertension, DM and dyslipidaemia. To account for cognitive status, analyses involving CU participants were additionally adjusted for MMSE scores, whereas analyses involving MCI participants were additionally adjusted for MCI subtypes. Finally, composite NPS measures (scale variables: total number of NPS or total NPS severity) were sequentially inserted into univariate generalized linear models (GLMs) as dependent variables (robust to violations of normality). Again, both measures were separately analyzed in CU and MCI individuals. GLMs were adjusted for the same covariates as described before.
Statistical analyses were performed using the IBM SPSS Statistics Software Version 26 (Chicago, IL, USA). Despite performing multiple comparisons, the conventional threshold of α = 0.05 was implemented for the revelation of statistical significance in order to retain a fair statistical power for our analyses, considering the low frequency of certain NPS, and the low prevalence of extensive WMH.
Results
Participant characteristics – CU group
In total, 4617 CU participants with available WMH assessments were eligible for the analysis. Of these 376 had moderate and 54 had extensive WMH. Baseline differences by WMH status are in Table 1. The sample comprised predominantly older, well-educated, Caucasian individuals. CU individuals with greater WMH burden were older, more often of female sex and African American, and performed worse on MMSE. Vascular comorbidities were more common in the presence of more severe WMH. Participants with normal cognition and extensive WMH had greater average severity of NPS. Anxiety and elation were the only NPS that were more prevalent in the high severity WMH group.
Table 1. Baseline comparison of older, cognitively unimpaired individuals by white matter hyperintensities status
Bold denotes statistically significant differences between the two groups; CHS, cardiovascular health study.
Associations between WMH status and NPS – CU group
After adjusting for age, education, MMSE scores, sex, race and vascular risk factors, the average number of NPS was 1.28 (95% CI: 1.08–1.48) for those with no to mild WMH, 1.38 (1.13–1.63) for those with moderate and 1.99 (1.50–2.49) for those with severe. Post hoc comparisons (pooled p = 0.007) found that the total number of NPS was greater in CU individuals with extensive WMH compared to those with no to mild WMH [Mean difference (MD) = 0.71 (0.24–1.18), p = 0.003] or compared to those with moderate WMH [MD = 0.61 (0.12–1.10), p = 0.014]. Average NPS severity by WMH status was 1.71 (1.41–2.00) for those with no to mild WMH, 1.83 (1.46–2.19) for those with moderate and 2.76 (2.03–3.49) for those with severe. Post hoc comparisons (pooled p = 0.008) revealed that total NPS severity was greater in the group of CU individuals with extensive WMH compared to those with no to mild WMH [MD = 1.05 (0.37–1.74), p = 0.003], and compared to those with moderate WMH [MD = 0.93 (0.21–1.65), p = 0.011].
The odds of having each NPS by WMH status in CU individuals are in Table 2. The odds of elation (∼9.9 times), disinhibition (∼4.4 times), agitation (∼3.5 times) and anxiety (∼2.7 times) were significantly elevated in CU older adults with extensive WMH compared to those with no or mild WMH (Figure 1). On the other hand, compared to those with no or mild WMH,moderate WMH increased only the odds of having disinhibition by about 1.9 times. Of note, the older adults with extensive WMH had higher odds (between 1.7 and 2.1 times) of having every remaining NPS, apart from psychotic symptoms, but findings were not significant.
Figure 1. Associations between extensive white matter hyperintensities and neuropsychiatric symptoms in cognitively unimpaired older adults. Odds ratios and 95% confidence intervals (95% CIs) are provided.
Table 2. Odds of having neuropsychiatric symptoms by white matter hyperintensities (WMH) status among older, cognitively unimpaired adults. The group with the lowest cardiovascular health study (CHS) score (<5) was used as reference
Bold denotes statistically significant differences between the two groups; between group differences were considered significant only if among group differences were significant as well – p-value for among group differences in provided by each NPS; analyses were adjusted for age, education, MMSE scores, sex, race, and vascular risk factors; OR, odds ratio; CI, confidence interval; NA, non-applicable – no participant with extensive WMH had psychotic symptoms.
Participant characteristics – MCI group
In total, 3170 participants with MCI had available WMH assessments and were eligible for the analysis. Among them, 471 had moderate and 88 had extensive WMH. Baseline differences by WMH status are in Table 3. The MCI group primarily consisted of older, well-educated, Caucasian individuals. Older adults with MCI and greater WMH burden were older, more often African American and performed worse on MMSE. Vascular comorbidities were more prevalent in the presence of more severe WMH. The number of total NPS, total NPS severity, as well as the presence of individual NPS did not differ by extent of WMH changes.
Table 3. Baseline comparison of older individuals with mild cognitive impairment (MCI) by white matter hyperintensities status
Bold denotes statistically significant differences between the two groups; CHS, cardiovascular health study.
Associations between WMH status and NPS – MCI group
After adjusting for age, education, MCI subtypes, sex, race and vascular risk factors, the average total number of NPS was 2.32 (2.03–2.61) for those with no to mild, 2.50 (2.17–2.83) for those with moderate and 1.89 (1.34–2.45) for those with severe WMH. Post hoc comparisons showed that the total number of NPS did not differ between groups. Average NPS severity was 3.14 (2.69–3.59) for those with no to mild, 3.41 (2.90–3.92) for those with moderate and 2.56 (1.71–3.41) for those with severe WMH. Post hoc comparisons found no NPS differences by WMH group. The odds of having each NPS by WMH status in older adults with MCI are in Table 4. Those with moderate WMH had approximately 34% more odds of having anxiety. However, considering the small effect size and inconsistency (no other significant associations were found between WMH and NPS in the MCI group – an opposite trend was revealed) this finding is likely the result of chance.
Table 4. Odds of having neuropsychiatric symptoms by white matter hyperintensities(WMH) status among older adults with mild cognitive impairment (MCI). The group with the lowest cardiovascular health study (CHS) score (<5) was used as reference
Bold denotes statistically significant differences between the two groups; between group differences were considered significant only if among group differences were significant as well – p-value for among group differences in provided by each NPS; analyses were adjusted for age, education, MCI subtypes, sex, race and vascular risk factors; OR, odds ratio; CI, confidence interval; NA, non-applicable – no participant with extensive WMH had elation.
Discussion
We report that CU older adults with extensive WMH in MRI studies had higher odds of having NPS compared to individuals with no to mild WMH, in particular, elation, disinhibition, agitation, and anxiety. In the CU group with moderate WMH only disinhibition was associated with WMH in comparison with the reference group (no to mild WMH). Our estimates accounted for important demographic and vascular-related factors, as well as cognitive testing. These findings align with the hypothesis that severe WMH contribute to the overall neuropsychiatric burden of CU individuals, independent of cognitive impairment or vascular risk burden. Moderate WMH do not appear to appreciably increase NPS risk overall in this group.
Older individuals with MCI and extensive or moderate WMH did not have greater odds of NPS compared to those with no to mild WMH. This lack of association may stem from the fact that co-existing neuropathologic alterations in older adults with MCI are almost universal: AD pathology predominates with vascular and Lewy body (LB) pathologies not infrequent (Dugger et al., Reference Dugger, Davis, Malek-Ahmadi, Hentz, Sandhu, Beach, Adler, Caselli, Johnson, Serrano, Shill, Belden, Driver-Dunckley, Caviness, Sue, Jacobson, Powell and Sabbagh2015; Schneider et al., Reference Schneider, Arvanitakis, Leurgans and Bennett2009). Therefore, those with no to mild (as well as moderate) WMH may have more co-existing neurodegenerative alterations that account for cognitive impairment. Considering the well-established, strong relationships between amyloid or LB pathology and several NPS it is likely that any association of WMH with NPS was overwhelmed by that of the other pathologies (Gibson et al., Reference Gibson, Grinberg, ffytche, Leite, Rodriguez, Ferretti‐Rebustini, Pasqualucci, Nitrini, Jacob‐Filho, Aarsland and Suemoto2023; Goukasian et al., Reference Goukasian, Hwang, Romero, Grotts, Do, Groh, Bateman and Apostolova2019; Krell-Roesch et al., Reference Krell-Roesch, Vassilaki, Mielke, Kremers, Lowe, Vemuri, Machulda, Christianson, Syrjanen, Stokin, Butler, Traber, Jack, Knopman, Roberts, Petersen and Geda2019). These findings may provide a potential explanation for the incongruous results of previous publications that did not consistently reveal a relationship between WMH and NPS in individuals without dementia: analyzing individuals without dementia, especially small groups, with and without MCI, with MCI owing to heterogeneous neuropathologies or with different levels of MCI and NPS severity may modulate true associations between WMH and NPS.
WMH have been associated with disrupted brain network dynamics (Tuladhar et al., Reference Tuladhar, Reid, Shumskaya, de Laat, van Norden, van Dijk, Norris and de Leeuw2015). Impaired transferring of information between interconnected cerebral areas is theorized to underlie the relationship between WMH and cognitive impairment (Yang et al., Reference Yang, Huang, Luo, Li, Qin, Ma, Shao, Xu, Zhang, Xu and Zhang2020). The same mechanism could be crucial in the occurrence of neuropsychiatric symptoms in individuals with WMH (Desmarais et al., Reference Desmarais, Gao, Lanctôt, Rogaeva, Ramirez, Herrmann, Stuss, Black, Keith and Masellis2021). The functional connectivity of the brain is compromised across multiple psychiatric conditions with younger ages of onset (e.g., autism, attention deficit hyperactivity disorder, bipolar disorder, schizophrenia, and so on) and may be similarly undermined in older adults with late-onset NPS. Therefore, the correlation of greater WMH load with reduced functional connectivity may provide a potential explanation for the prominent associations between extensive WMH and NPS (Crockett et al., Reference Crockett, Hsu, Dao, Tam, Eng, Handy and Liu-Ambrose2021; Quandt et al., Reference Quandt, Fischer, Schröder, Heinze, Lettow, Frey, Kessner, Schulz, Higgen, Cheng, Gerloff and Thomalla2020). Future research conducting mediation analyses could explore whether disrupted brain connectivity assumes a pivotal role in NPS among older individuals, whether different mechanisms are implicated or whether WMH are just epiphenomena of neurodegeneration with variable severity across the spectrum of heterogeneous neuropathologies.
Of note, WMH may not only contribute directly to the epidemiology of NPS but also via their interference with cognition. WMH burden has been correlated with global but also domain-specific cognitive impairment (Prins and Scheltens, Reference Prins and Scheltens2015). More specifically, frontal operations such as executive function and attention appear to be correlated to the volume of periventricular and subcortical WMH (Puzo et al., Reference Puzo, Labriola, Sugarman, Tripodis, Martin, Palmisano, Steinberg, Stein, Kowall, McKee, Mez, Killiany, Stern and Alosco2019; Sudo et al., Reference Sudo, Alves, Alves, Ericeira-Valente, Tiel, Moreira, Laks and Engelhardt2012). Again, disrupted brain (mainly fronto-parietal) networks seem to mediate these cognitive associations (Li et al., Reference Li, Liang, Chen, Zhang, Wei, Chen, Shu, Reiman and Zhang2015). Of interest, specific cognitive deficits have been more strongly linked to particular NPS: impairments in principally frontally mediated functions have been specifically associated with anxiety and lability symptoms such as disinhibition, agitation, irritability and elation (Liampas et al., Reference Liampas, Siokas, Lyketsos and Dardiotis2022b; Rosenberg et al., Reference Rosenberg, Mielke, Appleby, Oh, Leoutsakos and Lyketsos2011). Of course, some spatial specificity is to be expected; WMH in strategic brain regions may be more or less related to different cognitive and NPS manifestations (Brugulat-Serrat et al., Reference Brugulat-Serrat, Salvadó, Sudre, Grau-Rivera, Suárez-Calvet, Falcon, Sánchez-Benavides, Gramunt, Fauria, Cardoso, Barkhof, Molinuevo and Gispert2020; Lampe et al., Reference Lampe, Kharabian-Masouleh, Kynast, Arelin, Steele, Löffler, Witte, Schroeter, Villringer and Bazin2019). Based on the above, our findings may at least partially be driven by the cognitive associations of WMH.
Overall, our findings enhance current knowledge on the associations between neurodegenerative alterations and NPS in older adults without dementia. Previous positron emission tomography (PET) imaging studies of individuals without dementia has indicated that higher Aβ42 deposition is related to higher NPI-Q scores (Ng et al., Reference Ng, Chiew, Rosa-Neto, Kandiah, Ismail and Gauthier2021). Apathy, anxiety, depression and psychotic symptoms were most consistently associated with amyloid pathology – while agitation, disinhibition or elation present weaker to marginal associations (Gibson et al., Reference Gibson, Grinberg, ffytche, Leite, Rodriguez, Ferretti‐Rebustini, Pasqualucci, Nitrini, Jacob‐Filho, Aarsland and Suemoto2023; Goukasian et al., Reference Goukasian, Hwang, Romero, Grotts, Do, Groh, Bateman and Apostolova2019; Krell-Roesch et al., Reference Krell-Roesch, Vassilaki, Mielke, Kremers, Lowe, Vemuri, Machulda, Christianson, Syrjanen, Stokin, Butler, Traber, Jack, Knopman, Roberts, Petersen and Geda2019; Ng et al., Reference Ng, Chiew, Rosa-Neto, Kandiah, Ismail and Gauthier2021). On the other hand, elation and disinhibition appear to be linked to the presence of frontotemporal lobar degeneration, whereas psychotic symptoms and impulse control disorders are very prevalent with LB and PD neuropathology, respectively (Cajanus et al., Reference Cajanus, Solje, Koikkalainen, Lötjönen, Suhonen, Hallikainen, Vanninen, Hartikainen, de Marco, Venneri, Soininen, Remes and Hall2019; Cotta Ramusino et al., Reference Cotta Ramusino, Perini, Vaghi, Dal Fabbro, Capelli, Picascia, Franciotta, Farina, Ballante and Costa2021; Sokołowski et al., Reference Sokołowski, Roy, Goh, Hardy, Datta, Cobigo, Brown, Spina, Grinberg, Kramer, Rankin, Seeley, Sturm, Rosen, Miller and Perry2023). Although significant overlap is to be expected, based on the neuropsychiatric manifestations of an individual -especially as part of a comprehensive examination- clinicians and researcher can make some inferences about potential ongoing neurodegenerative alterations and select more sophisticated laboratory means to establish more accurate diagnoses.
Strengths and limitations
Our study has several strengths including the large sample of individuals with available WMH assessments with an adequate number of individuals with extensive WMH. The NPI-Q was uniformly used to assess the presence of NPS. The neurocognitive status of the participants (along with important demographic and vascular-related confounders) was accounted for in the analytical part of the article.
This analysis has several weaknesses, as well. First, the number of certain NPS (especially psychotic and motor symptoms and secondarily apathy) was very small, underpowering several aspects of our analysis. Therefore, it is not surprising that we failed to reproduce some previously established associations, such as between WMH and apathy (Manca et al., Reference Manca, Jones and Venneri2022). This is reflected in the broad confidence intervals and may have obscured several non-trivial associations. Second, although several crucial factors and covariates were taken into account, our findings may have been driven by residual confounding (it is not be possible to capture the effect of every potential confounder) or the non-trivial proportion of missing data (Liampas et al., Reference Liampas, Hatzimanolis, Siokas, Yannakoulia, Kosmidis, Sakka, Hadjigeorgiou, Scarmeas, Dardiotis and O’Caoimh2022a; Samara et al., Reference Samara, Liampas, Dadouli, Siokas, Zintzaras, Stefanidis, Daponte, Sotiriou and Dardiotis2022). Third, the presence or absence of WMH was not uniformly assessed by a central, blinded evaluator (or group of evaluators). Some variability is expected among different assessors in the quantification of WMH. In addition, we did not correct our findings for multiple comparisons to retain a fair statistical power despite the low frequency of certain NPS and the low prevalence of extensive WMH. Nevertheless, in view of the sizeable and consistent associations in the CU group we are confident that our results reflect true associations. Moreover, we did not include additional imaging (or not) biomarkers, such as global or parietal atrophy, hippocampal volumes, and so on. Finally, another limitation is the observational nature of our study. Hence, it is not possible to make etiologic inferences about NPS and WMH.
Conclusions
Extensive WMH were associated with the overall neuropsychiatric burden of CU individuals, independent of cognitive impairment or vascular risk burden. The odds of elation, disinhibition, agitation and anxiety were particularly elevated. On the other hand, WMH were not related to the neuropsychiatric burden of individuals with MCI. Considering that alternative neuropathologic alterations in older adults with MCI (especially AD, vascular and LB pathology) may account for cognitive impairment (instead of extensive WMH), the very strong established relationships between these pathologies and NPS is likely overwhelming the association of WMH with NPS. Therefore, it would be interesting if future research looked into the same associations using larger samples of individuals with different MCI subtypes and by extension different underlying pathologies.
Data availability and materials
For further information on access to the NACC database, please contact NACC (contact details can be found at https://naccdata.org/).
Conflict of interest
The authors declare that they have no conflict of interest.
Funding
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Description of author(s)’ roles
IL: original draft preparation, data curation, formal analysis, design of the study, interpretation of data, and review & editing of manuscript; VS, EZ, PS: validation, review & editing of manuscript; AP, ZT, VT: data curation review & editing of manuscript; CGL, ED: conceptualization, formulation of research question, design of the study, supervision, review & editing.
Ethical standards
Participants or surrogates provide informed consent before participation. All procedures are overseen by Institutional Review Boards at each ADRC. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Acknowledgements
The NACC database is funded by NIA/NIH Grant U24 AG072122. NACC data are contributed by the NIA-funded ADRCs: P30 AG062429 (PI James Brewer, MD, PhD), P30 AG066468 (PI Oscar Lopez, MD), P30 AG062421 (PI Bradley Hyman, MD, PhD), P30 AG066509 (PI Thomas Grabowski, MD), P30 AG066514 (PI Mary Sano, PhD), P30 AG066530 (PI Helena Chui, MD), P30 AG066507 (PI Marilyn Albert, PhD), P30 AG066444 (PI John Morris, MD), P30 AG066518 (PI Jeffrey Kaye, MD), P30 AG066512 (PI Thomas Wisniewski, MD), P30 AG066462 (PI Scott Small, MD), P30 AG072979 (PI David Wolk, MD), P30 AG072972 (PI Charles DeCarli, MD), P30 AG072976 (PI Andrew Saykin, PsyD), P30 AG072975 (PI David Bennett, MD), P30 AG072978 (PI Neil Kowall, MD), P30 AG072977 (PI Robert Vassar, PhD), P30 AG066519 (PI Frank LaFerla, PhD), P30 AG062677 (PI Ronald Petersen, MD, PhD), P30 AG079280 (PI Eric Reiman, MD), P30 AG062422 (PI Gil Rabinovici, MD), P30 AG066511 (PI Allan Levey, MD, PhD), P30 AG072946 (PI Linda Van Eldik, PhD), P30 AG062715 (PI Sanjay Asthana, MD, FRCP), P30 AG072973 (PI Russell Swerdlow, MD), P30 AG066506 (PI Todd Golde, MD, PhD), P30 AG066508 (PI Stephen Strittmatter, MD, PhD), P30 AG066515 (PI Victor Henderson, MD, MS), P30 AG072947 (PI Suzanne Craft, PhD), P30 AG072931 (PI Henry Paulson, MD, PhD), P30 AG066546 (PI Sudha Seshadri, MD), P20 AG068024 (PI Erik Roberson, MD, PhD), P20 AG068053 (PI Justin Miller, PhD), P20 AG068077 (PI Gary Rosenberg, MD), P20 AG068082 (PI Angela Jefferson, PhD), P30 AG072958 (PI Heather Whitson, MD), P30 AG072959 (PI James Leverenz, MD).
Open access
