Longitudinal cognitive biomarkers predicting symptom onset in presymptomatic frontotemporal dementia

Longitudinal cognitive biomarkers predicting symptom onset in presymptomatic frontotemporal dementia Introduction We performed 4-year follow-up neuropsychological assessment to investigate cognitive decline and the prog- nostic abilities from presymptomatic to symptomatic familial frontotemporal dementia (FTD). Methods Presymptomatic MAPT (n = 15) and GRN mutation carriers (n = 31), and healthy controls (n = 39) underwent neuropsychological assessment every 2 years. Eight mutation carriers (5 MAPT, 3 GRN) became symptomatic. We investi- gated cognitive decline with multilevel regression modeling; the prognostic performance was assessed with ROC analyses and stepwise logistic regression. Results MAPT converters declined on language, attention, executive function, social cognition, and memory, and GRN converters declined on attention and executive function (p < 0.05). Cognitive decline in ScreeLing phonology (p = 0.046) and letter fluency ( p = 0.046) were predictive for conversion to non-fluent variant PPA, and decline on categorical fluency (p = 0.025) for an underlying MAPT mutation. Discussion Using longitudinal neuropsychological assessment, we detected a mutation-specific pattern of cognitive decline, potentially suggesting prognostic value of neuropsychological trajectories in conversion to symptomatic FTD. Keywords Presymptomatic · Frontotemporal dementia · Familial · Biomarkers · Cognition · Neuropsychological assessment · Longitudinal Introduction characterized by deficits in executive function, social cog - nition and language, whereas memory and visuoconstruction Frontotemporal dementia (FTD) is a presenile neurodegen- are initially spared [3–5]. Non-fluent variant PPA (nfvPPA) erative disorder, leading to a heterogeneous clinical pres- patients show agrammatism and speech sound distortions, entation, involving behavioural (behavioural variant FTD; while semantic variant PPA (svPPA) patients experience def- bvFTD) and/or language deterioration (primary progressive icits in confrontation naming and word comprehension [6]. aphasia; PPA) [1]. FTD has an autosomal dominant pattern GRN mutations often lead to a clinical diagnosis of bvFTD, of inheritance in 30 percent of cases, with mutations in the nfvPPA or parkinsonism. In MAPT mutations, bvFTD is the progranulin (GRN) and microtubule-associated protein tau main phenotype, and semantic and memory impairments can (MAPT) genes as its two main causes [2]. The cognitive be prominent neuropsychological symptoms [7]. profile of FTD varies depending on the clinical phenotype Research in familial FTD has demonstrated the presence and the underlying genotype. Patients with bvFTD are of a presymptomatic stage in which subtle cognitive changes have been identified [ 8–12]. More specifically, cognitive decline can start as early as 8 years prior to estimated symp- Electronic supplementary material The online version of this tom onset and shows mutation-specific patterns, with GRN article (https ://doi.org/10.1007/s0041 5-018-8850-7) contains supplementary material, which is available to authorized users. mutation carriers declining in memory, and MAPT mutation carriers declining in language, social cognition and memory * Janne M. Papma [8, 10]. This suggests that cognitive measures could function j.papma@erasmusmc.nl as disease-tracking biomarkers in the presymptomatic stage. Extended author information available on the last page of the article Vol.:(0123456789) 1 3 1382 Journal of Neurology (2018) 265:1381–1392 However, it is currently unknown what the long-term cogni- tive profiles of presymptomatic FTD mutations are, whether neuropsychological assessment can be used to track disease progression to the symptomatic stage, and what the prognos- tic value is of cognitive trajectories in the presymptomatic and early symptomatic stage of FTD. In this study, we investigated longitudinal cognitive decline on neuropsychological assessment in presympto- matic mutation carriers (MAPT or GRN) and controls from the same families within our longitudinal presymptomatic Dutch familial FTD Risk Cohort (FTD-RisC). Second, we assessed the difference in cognitive course between convert- ers’ genotypes (i.e. MAPT vs. GRN) and phenotypes (i.e. bvFTD vs. nfvPPA) versus non-converters. Lastly, we inves- tigated the prognostic value of neuropsychological trajecto- ries in predicting symptom onset within 2–4 years. Methods Participants In FTD-RisC, we follow healthy 50% at-risk family mem- bers from genetic FTD families on a 2-year basis. In the Fig. 1 Participant in- and exclusion and sample size per time point. Two controls were excluded as they had multiple cognitive disorders current study, we included 87 participants from MAPT or (≤ 2 SD below reference mean) on neuropsychological testing. Eight GRN families with study entries between December 2009 mutation carriers converted to clinical FTD within the study window. and January 2013 [8, 9, 13]. The follow-up period was Their data were restructured, so that there were three time points: 4 years, in which we acquired neuropsychological assess- 4  years before symptom onset, 2  years before symptom onset and symptom onset. Four years before symptom onset, only data of six ments at study entry, follow-up after 2 years and follow-up converters were available, as two mutation carriers converted between after 4 years. DNA genotyping (see “Procedure”) assigned baseline and first follow-up. The data of converters were compared participants either to the presymptomatic mutation carrier to, respectively, baseline, follow-up after 2  years and follow-up after (n = 46; 31 GRN, 15 MAPT), or control group (n = 39; 29 4 years in non-converters and healthy controls GRN, 10 MAPT family members). We excluded two controls as they had cognitive disorders (≥ 2 SD below mean) on neurological examination, and MR imaging of the brain. multiple domains, ultimately including 85 participants (46 mutation carriers, 37 controls; Fig. 1). DNA sequencing was performed at study entry. All par- ticipants were asymptomatic according to established Standard protocol approvals, registrations, diagnostic criteria for bvFTD [3] or PPA [6] at baseline. Knowledgeable informants were asked about cognitive and/ and patient consents or behavioural deterioration at each study visit by means of a structured interview and a well-validated questionnaire Clinical investigators were blind for participants’ genetic sta- tus if they had not undergone predictive testing. In case of (Neuropsychiatric Inventory; NPI) [14]. conversion to clinical FTD, we offered the patient and family members genetic counselling and unblinding of genetic sta- Converters tus, to confirm the presence of the pathogenic mutation. At study entry, all participants gave written informed consent. Eight mutation carriers became symptomatic within the study time window (“converters”). Symptom onset was The study was approved by the Medical and Ethical Review Committee of the Erasmus Medical Center. determined by means of the above mentioned assessment (anamnesis, MR imaging of the brain, neuropsychological Procedure assessment, heteroanamnestic information and unblinding of genetic status). Conversion was determined in a multi- Every 2  years, participants underwent a standardized disciplinary consensus meeting of the Erasmus MC FTD Expertise Centre, involving neurologists (LDK, JCvsS), assessment consisting of a neuropsychological test battery, 1 3 Journal of Neurology (2018) 265:1381–1392 1383 neuropsychologists (LCJ, JLP, SF, EvdB, JMP), medical 2 years, follow-up after 4 years) into the following three doctors (LHM, ELvdE), as well as neuroradiologists, geri- new time points (Fig. 1): atricians, a clinical geneticist (RvM), and a care consultant. Six converters (5 MAPT, 1 GRN) presented with progres- 4 years before symptom onset: we used the data of the sive behaviour deterioration, functional decline, and frontal study visit 4 years before diagnosis. Analyses could were and/or temporal lobe atrophy on MRI, fulfilling the interna- performed in six converters, as two (1 GRN, 1 MAPT—2 tional diagnostic consensus criteria of Rascovsky et al. [3] bvFTD) developed symptoms between baseline and first for bvFTD with definite FTLD pathology. Two converters follow-up (i.e. at 2 years follow-up), and therefore no (both GRN) presented with isolated language difficulties data 4 years prior to symptom onset were available. and no impairments in daily living activities, thereby fulfill- 2 years before symptom onset: we used the data of the ing the diagnostic criteria for PPA of Gorno-Tempini et al. study visit 2 years before diagnosis. Analyses included [6]. Both developed nfvPPA, as they showed a non-fluent, all eight converters. halting speech, with sound errors and agrammatism. See After symptom onset: we used the data of the diagnosis Supplementary Table 1 for demographic, clinical and neu- visit. Analyses included all eight converters. ropsychological data of the converters. We defined mutation carriers remaining without FTD symptoms as non-convert- In non-converters and controls, we used the original time ers (n = 38; 28 GRN, 10 MAPT). points: baseline (data were compared to “4  years before symptom onset” data of converters), follow-up after 2 years Neuropsychological assessment (data were compared to “2 years before symptom onset data of converters) and follow-up after 4 years (data were com- We screened global cognitive functioning by means of the pared to “after symptom onset data of converters). Mini-Mental State Examination [15] (MMSE) and frontal assessment battery [16] (FAB). Experienced neuropsycholo- Statistical analysis gists (LCJ, JLP, SF) administered neuropsychological tests within six cognitive domains: language, attention and mental Statistical analyses were performed using SPSS Statistics processing speed, executive functioning, social cognition, 21.0 (IBM Corp., Armonk, NY) and GraphPad Prism 7 memory, and visuoconstruction. We rated language with the (La Jolla, California, USA), with the significance level at 60-item Boston Naming Test (BNT) [17], verbal Semantic p < 0.05 (two-tailed) across all comparisons. We compared Association Test (SAT) [18], ScreeLing phonology [19], and demographic data between MAPT mutation carriers, GRN categorical fluency [20]. We assessed attention and mental mutation carriers and controls, and between converters, non- processing speed by means of Trail making Test (TMT)-A converters and controls by means of one-way ANOVAs. We [21], Stroop Color-Word Test I and II [22], Wechsler Adult performed Pearson Χ tests to investigate differences in sex. Intelligence Scale III (WAIS-III) Digit Span forwards [23], Longitudinal comparisons of clinical data were performed and Letter Digit Substitution Test (LDST) [24]. Execu- with repeated measures ANOVAs. We standardized all raw tive functioning was evaluated using TMT-B [21], Stroop neuropsychological test scores by converting them into Color-Word Test III [22], WAIS-III Digit Span backwards z-scores (i.e. individual test score minus the baseline mean [23], modified Wisconsin Card Sorting Test (WCST) con- of the controls, divided by the baseline SD of the controls) cepts [25], letter fluency [ 20], and WAIS-III Similarities per time point, after which we calculated composite z-scores [23]. Happé cartoons [26] and Ekman Faces [27] measured for the respective six cognitive domains by averaging the social cognition. We assessed memory using the Dutch Rey z-scores of the individual tests per domain. For the longi- Auditory Verbal Learning Test (RAVLT) [28] and Visual tudinal comparisons we used multilevel linear regression Association Test (VAT) [29]. We evaluated visuoconstruc- modeling. This analysis corrects for bias when data absence tion by means of clock drawing [30] and WAIS-III Block is dependent on characteristics present in the model, and Design [23]. Alternate forms were used at follow-up visits, can therefore efficiently handle missing and unbalanced time when applicable (letter fluency, RAVLT, VAT). Depressive points. There were two levels in the models: the partici- symptoms were rated with the Beck’s Depression Inventory pants constituted the upper level; their repeated measures the (BDI) [31]. lower level. We ran two analyses to assess cognitive decline per mutation (1) and clinical status (2): Study design 1. We entered mutation status (MAPT mutation carrier, In converters, we restructured the three original time points GRN mutation carrier or control), time (4 years before within our study window (i.e. baseline, follow-up after symptom onset, 2 years before symptom onset, and after symptom onset), and first-order interactions, with age, 1 3 1384 Journal of Neurology (2018) 265:1381–1392 gender and educational level as covariates. We reran the in MAPT converters than GRN converters (p = 0.043) and analyses excluding the converters to exclude convert- non-converters (p = 0.001; Table 1). Both MAPT and GRN ers driving the cognitive decline in the mutation carrier converters declined significantly with respect to MMSE groups; score (p < 0.001) and they developed more neuropsychi- 2. We split the converter group according to genotype atric symptoms in the form of higher BDI (p = 0.001) and (MAPT or GRN) and phenotype (bvFTD or nfvPPA) NPI (p = 0.021) scores in comparison to non-converters and to investigate specific profiles of cognitive decline over controls. FAB scores did not significantly change over time time. We then entered clinical status (converter, non- (p > 0.05). converter or control), time, and first-order interactions, with age, gender and educational level as covariates. Longitudinal cognitive decline in MAPT and GRN mutation carriers Third, to investigate the prognostic abilities of cognitive decline in discriminating between converters and non-con- The whole group of MAPT mutation carriers declined sig- verters, we determined the area under the curve (AUC) by nificantly within the domains language, social cognition and receiver operating characteristic (ROC) analyses on the neu- memory compared with controls (Table 2; Fig. 1). This was ropsychological trajectories between visits. For this, we cal- reflected in lower scores on the BNT and categorical fluency, culated deltas between test scores; one between 4 and 2 years Happé cartoons, VAT and RAVLT delayed recall (Table 2). before symptom onset and one between 2 years before symp- In the whole group of GRN mutation carriers, no longitu- tom onset and symptom onset. Optimal cut-off levels were dinal decline was found in comparison to controls. In com- given by the highest Youden’s index [32]. Again, we split the parison to GRN mutation carriers, MAPT mutation carriers converter group according to genotype (MAPT or GRN) and declined significantly on the domains language (β = − 0.015, phenotype (bvFTD or nfvPPA). Next, we performed logistic p < 0.001) and memory (β = −  0.016, p = 0.008), reflected regression analyses, taking group (converter vs. non-con- in lower BNT (β = −  0.085, p = 0.01), SAT (β = −  0.027, verter) as the dependent variable and the deltas (tests with p = 0.015), category fluency (β = −  0.107, p = 0.002), and significant diagnostic performance in abovementioned ROC RAVLT delayed recall (β = − 0.047, p = 0.001) scores. There analyses) as the independent variables. The models were were no cognitive domains or tests on which GRN muta- selected with a forward stepwise method according to the tion carriers declined more than MAPT mutation carriers likelihood ratio test and applying the standard p values for (Table  2). By excluding the five MAPT converters from variable inclusion (0.05) and exclusion (0.10), with age, sex the analyses, none of the domain scores in MAPT mutation and education as covariates. Goodness of fit was evaluated carriers continued to show significant decline over time in 2 2 with the HL Χ test. Nagelkerke R is reported as measure of comparison to controls. Regarding individual tests, however, effect size. We checked predictor variables for multicollin- the decline on the RAVLT delayed recall remained signifi- earity. All models were corrected for multiple comparisons cant (β = − 0.032, p = 0.023). The results did not change by (Bonferroni). excluding the three GRN converters from the analyses. In comparison to GRN, MAPT mutation carriers still declined more on language (β = − 0.010, p = 0.004), reflected in lower Results ScreeLing phonology (β = − 0.008, p = 0.024) and category fluency (β = −  0.007, p = 0.041). There was no cognitive Demographics decline in controls—but significant improvement was found on social cognition (Happé non-ToM and Ekman Faces) and MAPT mutation carriers were significantly younger than memory (RAVLT immediate and delayed recall) (Table 2). GRN mutation carriers (p = 0.012; Table  1). The mean The raw neuropsychological test scores per time point can familial symptom onset age was lower in MAPT than in be found in Supplementary Table 2. GRN mutation carriers and controls (both p < 0.001). There were no significant differences between groups regarding Longitudinal cognitive decline in converters estimated years to symptom onset (p > 0.05). Longitudinal and non‑converters analyses demonstrated that MAPT mutation carriers declined significantly more than GRN mutation carriers and controls Converters with a MAPT mutation deteriorated significantly with regards to the MMSE (p = 0.014), and also developed on all domains but visuoconstruction (Fig. 2a–d, f; Table 3). more depressive symptoms (p = 0.028). FAB and NPI scores Within these domains, performances declined on BNT did not significantly change over time (p > 0.05). Convert- (p < 0.001), LDST (p = 0.035), Stroop I, II and III (I: p = 0.017; ers, non-converters and controls did not differ regarding II: p < 0.001; III: p = 0.021), categorical fluency (p = 0.001), demographic variables, apart from a younger family onset WAIS similarities (p < 0.001), Happé ToM (p = 0.011), and 1 3 Journal of Neurology (2018) 265:1381–1392 1385 Table 1 Demographics and clinical data Demographics HC (n = 39) MAPT carriers GRN carriers p value* MAPT GRN converters Non-converters p value** (n = 15) (n = 31) converters (n = 3) (n = 38) (n = 5) Age at study 49.1 ± 12.2 41.9 ± 10.0 52.1 ± 8.2 0.012 45.3 ± 8.5 54.9 ± 9.0 48.8 ± 10.3 0.704 entry, years Sex, female (%) 20 (56%) 7 (47%) 20 (65%) 0.506 1 (20%) 3 (100%) 23 (60.5%) 0.154 Education 5.2 ± 1.0 5.1 ± 1.6 5.7 ± 0.9 0.102 6.0 ± 0.7 5.7 ± 0.6 5.4 ± 1.3 0.409 (Verhage) a,b c,d Onset age fam- 59.0 ± 5.8 51.3 ± 6.7 61.0 ± 2.4 < 0.001 48.0 ± 4.7 59.7 ± 0.0 58.8 ± 6.1 0.002 ily, years Estimated years − 10.2 ± 11.2 − 7.7 ± 9.6 − 9.4 ± 7.9 0.690 − 2.7 ± 4.0 − 4.8 ± 9.0 − 10.0 ± 8.5 0.335 to onset, years Clinical Years HC (n = 39) MAPT carriers GRN carriers p value* MAPT GRN converters Non-converters p value** data to (n = 15) (n = 31) converters (n = 3) (n = 38) onset (n = 5) MMSE 4 29.1 ± 1.3 29.6 ± 0.5 29.1 ± 1.6 0.451 29.5 ± 0.6 29.0 ± 1.4 29.2 ± 1.4 0.924 2 29.2 ± 1.3 28.7 ± 2.2 28.9 ± 1.6 0.513 29.8 ± 0.4 28.0 ± 1.0 27.7 ± 1.5 0.271 c,d 0 29.2 ± 1.0 28.4 ± 1.5 29.2 ± 1.4 0.099 27.2 ± 1.6 27.7 ± 1.5 29.3 ± 1.2 0.001 FAB 4 – – – – – – – – 2 17.4 ± 0.9 17.4 ± 0.8 17.5 ± 0.9 0.883 17.3 ± 1.0 17.5 ± 0.7 17.5 ± 0.9 0.929 0 16.7 ± 1.7 16.5 ± 1.6 17.0 ± 1.1 0.639 15.4 ± 1.5 16.3 ± 1.5 17.1 ± 1.1 0.120 BDI 4 4.1 ± 4.5 4.0 ± 6.3 3.2 ± 3.9 0.693 1.3 ± 1.0 2.0 ± 2.8 3.7 ± 5.0 0.645 2 3.7 ± 3.9 4.5 ± 5.0 3.2 ± 4.0 0.638 5.0 ± 4.7 2.7 ± 3.8 3.5 ± 4.4 0.866 c,d 0 3.5 ± 4.3 7.6 ± 9.5 3.0 ± 6.7 0.108 11.6 ± 13.0 6.3 ± 5.1 3.1 ± 6.5 0.042 c–e NPI 4 0.1 ± 0.5 4.6 ± 11.2 1.4 ± 3.4 0.180 0.0 ± 0.0 – 3.0 ± 7.5 0.006 2 0.6 ± 1.2 6.4 ± 20.7 0.3 ± 0.7 0.095 0.2 ± 0.4 0.0 ± 0.0 2.9 ± 13.3 0.767 a,b d 0 0.8 ± 1.5 12.3 ± 18.7 2.1 ± 6.6 0.001 15.6 ± 16.3 10.7 ± 15.9 3.4 ± 11.4 0.009 Values indicate: mean ± standard deviation. Significant comparisons are displayed in bold GRN progranulin, HC healthy control, MMSE Mini-Mental State Examination, FAB frontal assessment battery, BDI Beck’s depression inven- tory, NPI neuropsychiatric inventory *p value represents result of overall ANOVA between MAPT mutation carriers, GRN mutation carriers and healthy controls **p value represents result of overall ANOVA between MAPT converters, GRN converters, non-converters and HC Significant post hoc test between MAPT and GRN mutation carriers Significant post hoc test between MAPT mutation carriers and healthy controls Significant post hoc test between converters and non-converters Significant post hoc test between converters and healthy controls Only data of MAPT converters available, therefore the p value represents the comparison between MAPT converters, non-converters and HC Dutch educational system categorized into levels from 1 = less than 6 years of primary education to 7 = academic schooling (Verhage, 1964) Data only available on follow-up visits RAVLT immediate (p = 0.004) and delayed recall (p = 0.030). as GRN converters, with lower scores on attention and execu- Converters with a GRN mutation deteriorated significantly on tive function (Table 3). There were no differences in decline attention and mental processing speed, and executive function between converters with bvFTD and nfvPPA (Table 3). The (Fig. 2b, c; Table 3). Within these domains, performances on raw neuropsychological test scores per time point can be found TMT-B (p < 0.001), Stroop III (p < 0.001), WCST (p = 0.005), in Supplementary Table 3. letter fluency (p = 0.012) and WAIS similarities (p < 0.001) deteriorated significantly over time. Converters with bvFTD Classification between converters had a similar pattern of cognitive decline as MAPT convert- and non‑converters ers, with lower scores on social cognition, memory, language, attention and executive function (Table 3). Comparably, con- Between 4 and 2  years before symptom onset, the delta verters with nfvPPA had a similar pattern of cognitive decline domain and individual neuropsychological test scores 1 3 1386 Journal of Neurology (2018) 265:1381–1392 Table 2 Cognitive trajectories in mutation carriers (converters, non-converters) and healthy controls Domain test Healthy controls (n = 39) MAPT mutation carriers (n = 15) GRN mutation carriers (n = 31) Baseline β p Baseline β p Baseline β p Language 0.0 ± 0.6 0.000 0.931 0.2 ± 0.6 − 0.010 0.002 0.1 ± 0.7 0.004 0.121 BNT 53.4 ± 4.5 0.026 0.105 52.6 ± 5.3 − 0.080 0.005 55.1 ± 3.7 0.006 0.786 SAT 27.8 ± 1.1 − 0.003 0.604 27.9 ± 1.5 − 0.008 0.604 27.5 ± 2.0 0.019 0.033 ScreeLing phonology 23.5 ± 0.8 0.001 0.733 23.9 ± 0.3 − 0.005 0.190 23.8 ± 0.5 − 0.001 0.863 Categorical fluency 23.9 ± 4.9 0.026 0.141 26.5 ± 6.6 − 0.087 0.006 23.4 ± 5.7 0.021 0.424 Attention and processing speed 0.0 ± 0.8 − 0.001 0.084 0.3 ± 0.6 − 0.003 0.096 0.1 ± 0.9 − 0.003 0.075 TMT part A 31.8 ± 15.0 − 0.022 0.416 26.1 ± 9.7 0.065 0.192 31.4 ± 12.2 0.060 0.145 Stroop card I 47.1 ± 8.0 0.039 0.011 43.2 ± 8.8 − 0.017 0.529 45.0 ± 8.4 − 0.001 0.951 Stroop card II 58.5 ± 10.6 0.012 0.539 54.9 ± 8.5 0.027 0.470 60.2 ± 13.2 0.001 0.969 Digit Span forwards 8.7 ± 1.9 0.001 0.871 9.0 ± 2.6 − 0.010 0.294 9.4 ± 2.4 − 0.016 0.055 LDST 34.5 ± 6.8 0.001 0.894 34.2 ± 4.7 − 0.636 0.699 33.2 ± 7.4 0.005 0.798 Executive function 0.0 ± 0.7 0.001 0.505 0.3 ± 0.6 − 0.005 0.065 0.2 ± 0.8 − 0.004 0.052 TMT part B 67.8 ± 29.3 0.052 0.494 61.0 ± 28.5 0.079 0.570 72.2 ± 42.7 − 0.099 0.390 Stroop card III 93.7 ± 22.6 − 0.087 0.021 83.8 ± 14.7 0.141 0.042 96.6 ± 26.2 0.013 0.815 Digit span backwards 6.1 ± 2.0 0.008 0.194 6.6 ± 1.8 0.002 0.877 6.6 ± 2.1 − 0.011 0.222 WCST concepts 5.5 ± 0.9 0.002 0.592 5.6 ± 1.1 − 0.009 0.296 5.80 ± 0.6 − 0.010 0.144 Letter fluency 32.1 ± 9.9 0.134 < 0.001 36.1 ± 14.3 − 0.108 0.049 38.9 ± 12.0 − 0.062 0.173 Similarities 24.8 ± 4.7 0.006 0.645 25.5 ± 4.7 − 0.034 0.122 26.2 ± 5.0 − 0.011 0.556 Social cognition 0.0 ± 0.8 0.000 0.878 0.2 ± 0.7 − 0.009 0.007 0.3 ± 0.7 − 0.003 0.332 Happé ToM 11.8 ± 3.4 0.013 0.172 12.6 ± 3.7 − 0.044 0.011 12.9 ± 2.9 − 0.005 0.707 Happé non-Tom 11.7 ± 2.9 0.020 0.013 12.4 ± 2.8 − 0.036 0.017 13.0 ± 2.6 − 0.012 0.331 Ekman faces 45.7 ± 6.4 0.038 0.009 47.0 ± 5.5 − 0.028 0.293 47.10 ± 5.5 − 0.013 0.548 Memory 0.0 ± 0.7 0.000 0.848 0.1 ± 1.3 − 0.017 < 0.001 0.1 ± 0.9 − 0.001 0.745 VAT 11.8 ± 0.6 0.001 0.740 11.4 ± 1.6 − 0.012 0.019 11.5 ± 0.9 0.000 0.926 RAVLT imm. recall 42.6 ± 9.8 0.157 < 0.001 47.5 ± 9.7 − 0.076 0.090 46.3 ± 10.6 − 0.015 0.686 b a,b RAVLT del. recall 8.4 ± 3.2 0.050 < 0.001 9.7 ± 3.9 − 0.048 < 0.001 9.4 ± 3.3 − 0.000 0.983 RAVLT recognition 28.6 ± 2.1 0.014 0.127 29.0 ± 2.0 − 0.022 0.176 29.2 ± 1.2 − 0.009 0.505 Visuoconstruction 0.0 ± 0.8 − 0.001 0.656 − 0.2 ± 0.7 − 0.005 0.266 0.0 ± 1.0 0.000 0.963 Block design 36.5 ± 14.0 0.034 0.305 35.5 ± 20.8 − 0.006 0.917 39.3 ± 18.5 − 1.164 0.246 Clock drawing 12.6 ± 1.4 0.003 0.453 12.2 ± 1.3 − 0.009 0.284 12.4 ± 1.8 0.005 0.475 Values indicate: mean ± standard deviation; β represents estimate of change over time. Composite domain scores are z-scores, individual test scores are raw scores. Composite domain scores are expressed as z-scores, the individual test scores are raw scores. p values represent compari- sons to healthy controls. Significant comparisons are displayed in bold MAPT microtubule-associated protein tau, GRN progranulin, BNT Boston Naming Test, SAT semantic association test, TMT Trail making Test, WAIS Wechsler Adult Intelligence Scale, LDST letter digit substitution test, WCST Wisconsin card sorting test, ToM theory of mind, VAT visual association test, RAVLT Rey Auditory Verbal Learning Test, imm immediate, del delayed Remained significant after excluding converters from the analyses Survived Bonferroni correction for multiple comparisons Higher scores and β weights indicate worse performance failed to distinguish significantly between converters and Discussion non-converters. Between 2 years before symptom onset and symptom onset decline on categorical fluency was predic- This study examined a large cohort of at-risk participants tive of an underlying MAPT mutation (p = 0.025; Table 4). from GRN and MAPT FTD families by means of neuropsy- Decline on ScreeLing phonology (p = 0.046) and letter flu- chological assessment during a 4-year follow-up. Within the ency (p = 0.046) was predictive of conversion to nfvPPA study time window, eight mutation carriers became symp- (Table 4). tomatic. Converters with a MAPT and GRN mutation had mutual as well as gene-specific profiles of cognitive decline. 1 3 Journal of Neurology (2018) 265:1381–1392 1387 Fig. 2 Multilevel linear regression model displaying longitudi- tion and mental processing speed, c executive functioning, d memory, nal decline (4  years, 2  years and after symptom onset) in composite e visuoconstruction, and f language. NB: the healthy controls have a domain z-score in the total group of converters (light green), MAPT mean z-score of zero by default as the z-scores of mutation carriers converters (light blue dotted line), GRN converters (dark blue dotted were based on that (raw score minus mean score of healthy controls, line), non-converters (dark green) and healthy controls (black). Mod- divided by the standard deviation of healthy controls). MAPT micro- els are displayed per cognitive domain: a social cognition, b atten- tubule-associated protein tau, GRN progranulin Cognitive decline on categorical fluency between 2 years neuropsychological test scores remain static while mutation before conversion and symptom onset was predictive for carriers are presymptomatic, and cognitive decline starts an underlying MAPT mutation, and decline on ScreeLing only near or at symptom onset [34–36], suggesting an explo- phonology and letter fluency was predictive for conversion sive rather than gradual start of the symptomatic disease to nfvPPA. These results suggest that neuropsychological stage. Alternatively, we might be unable to pick up subtle assessment could provide sensitive clinical biomarkers to cognitive changes in presymptomatic mutation carriers due identify and track FTD mutation carriers at-risk of convert- to lack of power. Also, although well-validated, most of our ing to the symptomatic stage. These findings hold potential neuropsychological tests were not developed for repeated for improving early clinical diagnosis by identifying the administration in a preclinical population [37]. We there- most sensitive neuropsychological tests for conversion, and fore cannot rule out that familiarity and/or practice effects use in upcoming disease-modifying clinical trials. are obscuring subtle cognitive decline, a notion that seems Following the MAPT mutation carriers over a 4-year to be underwritten by improvement in social cognition and period, we found significant decline in language, social cog- memory in controls, but not mutation carriers. nition and memory. This is consistent with findings from In our exploratory analyses in converters, we discovered previous presymptomatic familial FTD studies, in which both common as well as mutation-specific profiles of cogni- both cross-sectional [9–11, 33] and longitudinal [8] decline tive decline in MAPT and GRN. In both mutations, decline was found. Specifically, in our first follow-up study [8 ], we in attention, mental processing speed and executive function demonstrated decline in the domains language, social cog- was found—while only converters with a MAPT mutation nition and memory 5–8 years before estimated symptom demonstrated decline on language, memory and social cogni- onset. It should be taken into account that this study made tion. Previous studies in familial FTD also point to distinct use of estimated onset as a proxy, instead of actual symp- profiles for MAPT and GRN [8, 10–12], and are largely con- tom onset as in the present study—but the similar profile sistent with our present findings. Another important aspect is of decline confirms the presence of early changes in these the longitudinal tracking of the different clinical phenotypes. three domains. As in our previous study, the present results The similar patterns of cognitive decline in bvFTD as MAPT, are largely driven by the converters. This could suggest that and nfvPPA as GRN are related to the dominant genotype in 1 3 1388 Journal of Neurology (2018) 265:1381–1392 1 3 Table 3 Cognitive trajectories in MAPT, GRN, bvFTD and nfvPPA converters, and non-converters Domain test MAPT converters (n = 5) GRN converters (n = 3) bvFTD converters (n = 6) nfvPPA converters (n = 2) Non-converters (= 38) Baseline β p Baseline β p Baseline β p Baseline β p Baseline β p a a Language 0.1 ± 0.7 − 0.028 < 0.001 0.6 ± 0.2 − 0.007 0.299 0.1 ± 0.7 − 0.025 < 0.001 0.6 ± 0.2 − 0.014 0.061 0.1 ± 0.6 0.002 0.408 a a BNT 54.3 ± 6.9 − 0.239 < 0.001 57.5 ± 2.1 − 0.019 0.604 54.3 ± 6.9 − 0.224 < 0.001 57.5 ± 2.1 − 0.033 0.396 54.2 ± 4.2 − 0.001 0.960 SAT 27.0 ± 1.4 − 0.040 0.034 28.0 ± 1.4 0.006 0.805 27.0 ± 1.4 − 0.036 0.052 28.0 ± 1.4 0.000 0.993 27.7 ± 2.0 0.013 0.127 ScreeLing phonology 24.0 ± 0.0 0.002 0.617 24.0 ± 0.0 − 0.011 0.114 24.0 ± 0.0 0.004 0.358 24.0 ± 0.0 − 0.017 0.018 23.8 ± 0.4 − 0.002 0.551 a a Categorical fluency 25.8 ± 4.6 − 0.250 < 0.001 28.0 ± 2.8 − 0.149 0.022 25.8 ± 4.6 − 0.237 < 0.001 28.0 ± 2.8 − 0.170 0.015 24.0 ± 6.3 0.014 0.546 Attention and mental 0.3 ± 0.6 − 0.010 0.006 0.2 ± 0.3 − 0.013 0.005 0.3 ± 0.6 − 0.010 0.004 0.2 ± 0.3 − 0.013 0.006 0.1 ± 0.8 − 0.001 0.321 processing speed TMT part A 20.0 ± 6.3 0.067 0.448 25.0 ± 8.5 0.073 0.539 20.0 ± 6.3 0.065 0.449 25.0 ± 8.5 0.090 0.483 31.1 ± 11.8 0.051 0.181 Stroop card I 44.0 ± 5.2 0.101 0.030 46.5 ± 6.4 0.058 0.349 44.0 ± 5.2 0.106 0.020 46.5 ± 6.4 0.044 0.503 44.4 ± 8.9 − 0.020 0.345 b a a Stroop card II 58.5 ± 7.6 0.331 < 0.001 56.5 ± 0.7 0.186 0.006 57.5 ± 7.6 0.319 < 0.001 56.5 ± 0.7 0.194 0.008 58.8 ± 12.9 − 0.032 0.217 Digit Span forwards 9.5 ± 1.7 0.010 0.609 9.0 ± 0.0 − 0.038 0.146 9.5 ± 1.7 0.010 0.601 9.0 ± 0.0 − 0.043 0.119 9.3 ± 2.6 − 0.013 0.088 LDST 34.8 ± 6.7 − 0.100 0.012 35.0 ± 0.0 − 0.061 0.235 34.8 ± 6.7 − 0.098 0.011 35.0 ± 0.0 − 0.061 0.270 33.3 ± 6.9 0.004 0.809 a a Executive function 0.6 ± 0.4 − 0.018 < 0.001 0.6 ± 0.1 − 0.032 <0.001 0.6 ± 0.4 − 0.020 < 0.001 0.6 ± 0.1 − 0.029 < 0.001 0.2 ± 0.8 − 0.001 0.515 b a TMT part B 57.0 ± 27.0 0.472 0.038 48.0 ± 32.5 1.448 <0.001 57.0 ± 27.0 0.684 0.006 48.0 ± 32.5 0.921 0.010 71.2 ± 40.4 − 0.132 0.195 b a a a a Stroop card III 87.5 ± 23.4 0.468 < 0.001 86.5 ± 7.8 0.734 <0.001 87.5 ± 23.4 0.449 < 0.001 86.5 ± 7.8 0.815 < 0.001 93.7 ± 24.8 − 0.026 0.577 Digit span backwards 8.0 ± 1.4 − 0.018 0.284 5.5 ± 0.7 − 0.039 0.082 8.0 ± 1.4 − 0.022 0.186 5.5 ± 0.7 − 0.033 0.172 6.5 ± 2.0 − 0.003 0.721 WCST concepts 6.0 ± 0.0 − 0.015 0.193 6.0 ± 0.0 − 0.040 0.007 6.0 ± 0.0 − 0.021 0.073 6.0 ± 0.0 − 0.032 0.035 5.7 ± 0.8 − 0.006 0.323 Letter fluency 35.8 ± 7.9 − 0.143 0.101 45.5 ± 17.7 − 0.328 0.010 35.8 ± 7.9 − 0.156 0.066 45.5 ± 17.7 − 0.339 0.013 37.9 ± 13.0 − 0.048 0.245 a a a a Similarities 29.0 ± 1.2 − 0.151 < 0.001 29.0 ± 1.4 − 0.175 <0.001 29.0 ± 1.2 − 0.155 < 0.001 29.0 ± 1.4 − 0.175 < 0.001 25.5 ± 4.0 0.004 0.775 a a Social cognition 0.0 ± 1.0 − 0.022 < 0.001 0.8 ± 0.1 − 0.012 0.127 0.0 ± 1.0 − 0.021 < 0.001 0.8 ± 0.1 − 0.016 0.071 0.3 ± 0.7 − 0.002 0.336 Happé ToM 12.3 ± 5.1 − 0.096 0.002 13.5 ± 2.1 0.017 0.672 12.3 ± 5.1 − 0.078 0.012 13.5 ± 2.1 − 0.019 0.669 12.8 ± 3.0 − 0.012 0.380 Happé non-Tom 12.3 ± 2.4 − 0.067 0.010 15.5 ± 0.7 − 0.041 0.215 12.3 ± 2.4 − 0.060 0.016 15.5 ± 0.7 − 0.062 0.080 12.8 ± 2.7 − 0.012 0.267 Ekman faces 43.5 ± 6.1 − 0.089 0.023 50.0 ± 0.0 − 0.175 0.001 43.5 ± 6.1 − 0.118 0.003 50.0 ± 0.0 − 0.127 0.024 47.3 ± 5.4 − 0.001 0.965 a a Memory − 1.0 ± 2.0 − 0.050 < 0.001 0.7 ± 0.8 0.002 0.751 − 1.0 ± 2.0 − 0.044 < 0.001 0.7 ± 0.8 − 0.005 0.525 0.2 ± 0.8 − 0.002 0.473 VAT 10.0 ± 2.4 − 0.030 0.005 12.0 ± 0.0 0.004 0.675 10.0 ± 2.4 − 0.027 0.011 12.0 ± 0.0 0.000 0.983 11.6 ± 0.8 − 0.002 0.705 RAVLT imm. recall 42.5 ± 9.1 − 0.241 0.001 54.5 ± 19.1 − 0.111 0.226 42.5 ± 9.1 − 0.210 0.003 54.5 ± 19.1 − 0.177 0.067 46.7 ± 10.0 − 0.009 0.797 a a RAVLT del. recall 7.5 ± 5.5 − 0.085 < 0.001 10.5 ± 5.0 0.002 0.951 7.5 ± 5.5 − 0.080 < 0.001 10.5 ± 5.0 − 0.002 0.954 9.7 ± 3.2 − 0.009 0.359 RAVLT recognition 27.3 ± 3.1 − 0.037 0.005 30.0 ± 0.0 − 0.014 0.266 27.3 ± 3.1 − 0.036 0.004 30.0 ± 0.0 − 0.014 0.308 29.3 ± 1.1 − 0.009 0.461 Visuoconstruction 0.2 ± 0.8 − 0.009 0.217 0.2 ± 0.2 − 0.010 0.312 0.2 ± 0.8 − 0.008 0.250 0.2 ± 0.2 − 0.013 0.237 − 0.1 ± 1.0 0.000 0.895 Block design 51.0 ± 27.1 − 0.222 0.064 32.0 ± 1.4 − 0.148 0.333 51.0 ± 27.1 − 0.235 0.042 32.0 ± 1.4 − 0.109 0.503 37.1 ± 18.5 − 0.006 0.898 Clock drawing 11.8 ± 2.1 − 0.002 0.876 13.5 ± 0.7 − 0.014 0.459 11.8 ± 2.1 − 0.001 0.966 13.5 ± 0.7 − 0.023 0.281 12.3 ± 1.6 0.001 0.888 Values indicate: mean ± standard deviation; β represents estimate of change over time. Composite domain scores are z-scores, individual test scores are raw scores. Composite domain scores are expressed as z-scores, the individual test scores are raw scores. p values represent comparisons to non-converters. Significant comparisons are displayed in bold MAPT microtubule-associated protein tau, GRN progranulin, bvFTD behavioural variant frontotemporal dementia, nfvPPA non-fluent variant primary progressive aphasia, BNT Boston Naming Test, SAT semantic association test, TMT Trail making Test, WAIS Wechsler Adult Intelligence Scale, LDST letter digit, substitution test, WCST Wisconsin card sorting test, ToM theory of mind, VAT visual association test, RAVLT Rey Auditory Verbal Learning Test, imm immediate, del delayed Survived Bonferroni correction for multiple comparisons Higher scores and β weights indicate worse performance Journal of Neurology (2018) 265:1381–1392 1389 1 3 Table 4 ROC analyses on neuropsychological decline between 2 years before conversion and symptom onset in converters Domain and individual neuropsychological bvFTD vs. nfvPPA converters MAPT vs. GRN converters tests a b AUC 95% CI p Optimal Δ Sensitivity (%) Specificity (%) AUC 95% CI p Optimal Δ Sensitivity (%) Specificity (%) Language 0.667 0.29–1.00 0.505 – – – 0.867 0.51–1.00 0.101 – – – BNT 0.708 0.34–1.00 0.405 – – – 0.90 0.67–1.00 0.074 – – – SAT 0.625 0.24–1.00 0.617 – – – 0.833 0.54–1.00 0.136 – – – ScreeLing phonology 1.000 1.00–1.00 0.046 − 0.5 100 100 0.700 0.21–1.00 0.371 – – – Categorical fluency 0.833 0.53–1.00 0.182 – – – 1.000 1.00–1.00 0.025 − 6.5 100 100 Attention and mental processing speed 0.750 0.41–1.00 0.317 – – – 0.600 0.19–1.00 0.655 – – – TMT part A 0.542 0.00–1.00 0.868 – – – 0.50 0.05–0.95 1.000 – – – Stroop card I 0.583 0.19–0.97 0.739 – – – 0.600 0.17–1.00 0.655 – – – Stroop card II 0.583 0.12–1.00 0.739 – – – 0.667 0.22–1.00 0.456 – – – Digit Span forwards WAIS-III 0.750 0.40–1.00 0.317 – – – 0.633 0.23–1.00 0.551 – – – LDST 0.625 0.23–1.00 0.617 – – – 0.633 0.22–1.00 0.551 – – – Executive function 0.583 0.19–0.98 0.739 – – – 0.733 0.36–1.00 0.297 – – – TMT part B 0.667 0.29–1.00 0.617 – – – 0.900 0.64–1.00 0.121 – – – Stroop card III 0.833 0.51–1.00 0.182 – – – 0.600 0.15–1.00 0.655 – – – Digit span backwards WAIS-III 0.542 0.09–1.00 0.868 – – – 0.567 0.14–0.99 0.766 – – – WCST concepts 0.500 0.10–0.90 1.000 – – – 0.700 0.32–1.00 0.371 – – – Letter fluency 1.000 1.00–1.00 0.046 − 16 100 100 0.767 0.36–1.00 0.233 – – – Similarities WAIS-III 0.625 0.14–1.00 0.617 – – – 0.567 0.13–1.00 0.766 – – – Social cognition 0.500 0.00–1.00 1.000 – – – 0.667 0.13–1.00 0.456 – – – Happé ToM 0.458 0.00–1.00 0.868 – – – 0.700 0.21–1.00 0.371 – – – Happé non-Tom 0.500 0.00–1.00 1.000 – – – 0.667 0.22–1.00 0.456 – – – Ekman faces 0.667 0.15–1.00 0.505 – – – 0.567 0.07–1.00 0.766 – – – Memory 0.750 0.41–1.00 0.317 – – – 0.933 0.75–1.00 0.053 – – – VAT 0.792 0.45–1.00 0.243 – – – 0.933 0.75–1.00 0.053 – – – RAVLT immediate recall 0.667 0.15–1.00 0.505 – – – 0.600 0.09–1.00 0.655 – – – RAVLT delayed recall 0.667 0.27–1.00 0.505 – – – 0.867 0.58–1.00 0.101 – – – RAVLT recognition 0.750 0.37–1.00 0.317 – – – 0.900 0.65–1.00 0.074 – – – Visuoconstruction 0.583 0.19–0.98 0.739 – – – 0.600 0.19–1.00 0.655 – – – Block design WAIS-III 0.808 0.35–1.00 0.405 – – – 0.500 0.07–0.93 1.000 – – – Clock drawing 0.667 0.29–1.00 0.505 – – – 0.600 0.16–1.00 0.655 – – – AUC area under the curve, CI confidence interval, bvFTD behavioural variant frontotemporal dementia, nfvPPA non-fluent variant frontotemporal dementia, MAPT microtubule-associated pro- tein tau, GRN progranulin, BNT Boston Naming Test, SAT semantic association test, TMT Trail making Test, WAIS Wechsler Adult Intelligence Scale, LDST letter digit substitution test, WCST Wisconsin Card Sorting Test, ToM theory of mind, VAT visual association test, RAVLT Rey Auditory Verbal Learning Test Negative delta represents decline in test performance in nfvPPA vs. bvFTD (i.e. when a converter declines on this particular task, he/she is more likely to develop nfvPPA Negative delta represents decline in test performance in MAPT vs GRN (i.e. when a converter declines on this particular task, he/she is more likely to have a underlying MAPT mutation 1390 Journal of Neurology (2018) 265:1381–1392 each group (e.g. all nfvPPA converters have a GRN muta- are affected in both presymptomatic [8 , 11] and symptomatic tion). These findings suggest that neuropsychological assess- FTD [49, 50]. Future research could additionally investigate the ment can be used to track the different mutations and phe- use of qualitative assessment of verbal fluency (e.g. clustering, notypes from the presymptomatic to the symptomatic stage, switching between clusters), as recent research [49] points to which is advantageous considering the need for good clinical die ff rences between FTD and PPA subtypes—making this a endpoints in future disease-modifying trials. promising application of verbal fluency for a precise clinical Extending the findings from our first follow-up study [8 ], differentiation in presymptomatic and early stage FTD. we demonstrated significant decline on the RAVLT recall in Key strengths of our study constitute our longitudinal design, presymptomatic MAPT mutation carriers. The additional finding spanning a 4-year follow-up of at-risk participants from both MAPT that lower memory scores over time were also found in MAPT, and GRN families. Although our group of converters is currently and not GRN converters—suggesting a mutation-specific small, this is the first study tracking FTD mutation carriers from aetiology—corroborate this. Although memory loss has been the presymptomatic to symptomatic disease stage. Being aware of described in GRN [38, 39], this is usually a later symptom, while the caveats of small sample sizes and administering a large amount episodic memory impairment has been found as the presenting of neuropsychological tests with respect to statistical power, our and most prominent symptom in MAPT [7, 40, 41]. Interest- results warrant replication in our cohort as well as larger interna- ingly, the Genetic Frontotemporal dementia Initiative (GENFI) tional cohorts such as GENFI [10], in which with the passing of consortium revealed hippocampal atrophy in presymptomatic time more mutation carriers will approach symptom onset and/ MAPT from 15 years before estimated symptom onset [10], and or convert to clinical FTD. The dropout rate is very low, creating as this medial temporal structure is critical for episodic memory balanced datasets across the three time points. Additionally, use of processing [42] this offers a good explanation for our findings. multilevel linear modeling further handles ec ffi iently with miss - In line with earlier studies [42, 43], we did n fi d dec fi its in verbal ing data. Directions for future research entail the development of recall but not visual associative memory. Semantically loaded neuropsychological tasks more suited to administer in the presymp- tasks such as the RAVLT can be particularly more difficult than tomatic phase (robust to ceiling effects) and repeated administra- visual memory tasks like the VAT, as a result of the prominent tion (robust to practice and able to measure small changes). More semantic impairments seen early in MAPT-associated FTD [44]. extensive quantification tools of behavioural functioning are also Our results contribute to the present thinking that memory defi- needed to capture the entire clinical spectrum of (presymptomatic) cits can be an integral part of the clinical spectrum [42], and FTD, as well as assessment methods that rely less on the accuracy comprehensive memory tasks should therefore be incorporated of informant report [37]. A disadvantage of the study is the fact that in the standard diagnostic work-up. the neuropsychological assessment was part of the clinical assess- Knowing the cognitive profile of decline indicative for con- ment with which we determined conversion to the symptomatic version is important to get more insight into the timing of clini- stage. This has likely led to a circular reasoning, as we demonstrated cal changes in the earliest disease stage. We found that conver- that converters declined over time, while cognitive decline was con- sion can be predicted based on cognitive decline in the 2 years sidered a prerequisite for conversion. Ideally, the tests assessed in prior to symptom onset, but not earlier. As the cognitive decline our study should not have been used in the diagnosis of conver- was part of the diagnostic process of determining conversion, sion. However, in our multidisciplinary meeting, we followed the this is not a surprising finding. However, it does suggest a more international consensus criteria for bvFTD [3] and PPA [6], using explosive disease development with cognitive decline acceler- all available clinical information—e.g. MR imaging of the brain, ating rapidly in proximity of symptom onset, which is in line anamnestic and heteroanamnestic information, behavioural and with evidence from a large familial Alzheimer’s disease cohort neuropsychiatric questionnaires, unblinding of genetic status—so [45]. By selectively choosing tests within the domains that that symptom onset did not solely depend on the neuropsychologi- have prognostic abilities, the neuropsychological battery can cal assessment. Furthermore, as the multilevel model assumes a be shortened, which would benefit patient burden and helps linear relationship between genetic status and cognitive decline over cutting healthcare expenses. Especially fluency tasks seem to time, we could have missed non-linear effects over time. Lastly, the be promising candidates, as they were able to distinguish accu- analyses on the non-converters and controls were performed using rately between future phenotype and underlying genotype. The the original baseline and follow-up visits, regardless of, e.g. age and latter is essential for patient stratification in future clinical trials time to estimated symptom onset. It is possible that these analyses targeting specific pathologies, and ideally these interventions therefore lost some sensitivity to detect cognitive decline over time. should be applied in the presymptomatic stage [46]. Reliable However, as between-group analyses on age and estimated years to phenotypic prediction furthermore optimizes the diagnostic symptom onset in converters, non-converters, and controls did not process by shortening the current diagnostic delay [47], and show significant differences (respectively, p = 0.99 and p = 0.19), is helpful for the patient, caregiver and clinician in knowing we believe this effect is minimal. what disease presentation and course to expect. Verbal fluency Our study investigates longitudinal neuropsychological per- tests are widely used in dementia diagnosis setting [48], and formance in a large cohort of at-risk individuals from genetic 1 3 Journal of Neurology (2018) 265:1381–1392 1391 5. Adenzato M, Cavallo M, Enrici I (2010) Theory of mind ability in FTD families. We provide evidence of mutation-specific cog- the behavioural variant of frontotemporal dementia: an analysis of nitive decline when moving from the presymptomatic into the neural, cognitive, and social levels. Neuropsychologia 48:2–12 symptomatic stage, and of neuropsychological trajectories 6. Gorno-Tempini ML, Hillis AE, Weintraub S et al (2011) Classifi- predicting symptom onset. These results suggest the potential cation of primary progressive aphasia and its variants. Neurology 76(11):1006–1014 biomarker value of neuropsychological assessment in both 7. Rohrer JD, Warren JD (2011) Phenotypic signatures of genetic disease-monitoring and predicting conversion to clinical FTD. frontotemporal dementia. Curr Opin Neurol 24(6):542–549 8. Jiskoot LC, Dopper EGP, den Heijer T et al (2016) Presympto- Acknowledgements We would like to thank all the participants and matic cognitive decline in familial frontotemporal dementia: a their families for taking part in our study. This work was supported by longitudinal study. Neurology 87:384–391 Dioraphte Foundation Grant 09-02-03-00, the Association for Fronto- 9. Dopper EG, Rombouts SA, Jiskoot LC et al (2014) Structural and temporal Dementias Research Grant 2009, Alzheimer Nederland and functional brain connectivity in presymptomatic familial fronto- Memorabel ZonMw Grant 733050102 (Deltaplan Dementie). temporal dementia. Neurology 83:e19–e26 10. Rohrer JD, Nicholas JM, Cash DM et al (2015) Presymptomatic Author contributions LCJ contributed to the conception and design cognitive and neuroanatomical changes in genetic frontotemporal of the study, acquired and analysed data, and drafted the manuscript, dementia in the genetic frontotemporal dementia initiative (GENFI) figures and tables. JLP acquired data. LvA acquired and analysed data. study: a cross-sectional analysis. Lancet Neurol 14:253–262 SF acquired data. LHM acquired data and contributed to the design 11. Geschwind DH, Robidoux J, Alarcón M et al (2001) Dementia of the figures. LDK acquired data. ELvdE acquired data. EGPD con- and neurodevelopmental predisposition: cognitive dysfunction in tributed to the conception of the study and acquired data. RT contrib- presymptomatic subjects precedes dementia by decades in fron- uted to the design of the study and data analysis. RvM is the genetic totemporal dementia. Ann Neurol 50:741–746 guardian of the study. JvS contributed to the conception and design 12. Hallam BJ, Jacova C, Hsiung GYR et al (2014) Early neuropsy- of the study and is PI of the project. EvdB contributed to the design chological characteristics of progranulin mutation carriers. J Int and data interpretation of the study. JMP contributed to the design of Neuropsychol Soc 20:694–703 the study, and drafting the manuscript, figures and tables. All authors 13. Dopper EG, Chalos V, Ghariq E et al (2016) Cerebral blood flow were involved in copyediting and approval of the final draft of the in presymptomatic MAPT and GRN mutation carriers: a longitu- manuscript. dinal arterial spin labeling study. Neuroimage Clin 12:460–465 14. Kaufer DI, Cummings JL, Ketchel P et al (2000) Validation of the NPI-Q, a brief clinical form of the neuropsychatric inventory. J Compliance with ethical standards Neuropsychiatry Clin Neurosci 12(2):233–239 15. Folstein MF, Folstein SE, McHugh PR (1975) “Mini-mental Conflicts of interest LCJ, JLP, LvA, LHM, LDK, ELvdE, EGPD, RT, state”. A practical method for grading the cognitive state of RvM, JvS, EvdB, JMP report no conflicts of interest. patients for the clinician. J Psychiatr Res 12(3):189–198 16. Dubois B, Slachevsky A, Litvan I, Pillon B (2000) The Ethical standard All procedures performed in studies involving human FAB: a frontal assessment battery at bedside. Neurology participants were in accordance with the ethical standards of the insti- 55(11):1621–1626 tutional and national research committee, and with the 1964 Helsinki 17. Kaplan E, Goodglass H, Weintraub S (1978) The Boston Naming declaration and its later amendments or comparable ethical standards. Test. Lea & Febiger, Philadelphia 18. Visch-Brink E, Stronks D, Denes G (2005) Semantische Associa- Informed consent Informed consent was obtained from all individual tie test. Swets & Zeitlinger, Lisse participants included in the study. 19. Doesborgh SJ, van de Sandt-Koenderman WM, Dippel DW et al (2003) Linguistic deficits in the acute phase of stroke. J Neurol 250(8):977–982 Open Access This article is distributed under the terms of the Crea- 20. Thurstone LLT, Thurstone TG (1962) Primary mental abilities. tive Commons Attribution 4.0 International License (http://creat iveco Science Research Associates, Chicago mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- 21. Battery Army Individual Test (1994) Manual of directions and scor- tion, and reproduction in any medium, provided you give appropriate ing. War Department, Adjutant General’s office, Washington, DC credit to the original author(s) and the source, provide a link to the 22. Stroop JR (1935) Studies of interference in serial verbal reactions. Creative Commons license, and indicate if changes were made. J Exp Psychol 18:643–662 23. Wechsler D (2005) WAIS-III nederlandse bewerking, technische handleiding. Harcourt Test Publishers, Lisse 24. Jolles J, Houx PJ, van Boxtel MPJ, Ponds RWHM (1995) Maas- References tricht aging study: determinants of cognitive aging. Neuropsych Publishers, Maastricht 1. Snowden JS, Pickering-Brown SM, Mackenzie IR et al (2006) Pro- 25. Nelson HE (1976) A modie fi d card sorting test sensitive to frontal granulin gene mutations associated with frontotemporal dementia lobe defects. Cortex 12:313–324 and progressive non-fluent aphasia. Brain 129:3091–3102 26. Happe F, Brownell H, Winner E (1999) Acquired ‘theory of mind’ 2. Warren JD, Rohrer JD, Rossor MN (2013) Clinical review. Fron- impairments following stroke. Cognition 70(3):211–240 totemporal dementia. BMJ 347:f4827 27. Ekman P, Friesen WV (1976) Pictures of facial affect. Consulting 3. Rascovsky K, Hodges JR, Knopman D et al (2011) Sensitivity of Psychologists Press, Palo Alto revised diagnostic criteria for the behavioural variant of fronto- 28. Rey A (1958) L’examen clinique en psychologie. Presses Univer- temporal dementia. Brain 134:2456–2477 sitaires de France, Paris 4. Seelaar H, Rohrer JD, Pijnenburg Y et al (2011) Clinical, genetic 29. Lindeboom J, Schmand B, Tulner L et al (2002) Visual associa- and pathological heterogeneity of frontotemporal dementia: a tion test to detect early dementia of the Alzheimer type. J Neurol review. J Neurol Neurosurg Psychiatry 82:476–486 Neurosurg Psychiatry 73(2):126–133 1 3 1392 Journal of Neurology (2018) 265:1381–1392 30. Royall DR, Cordes JA, Polk M (1998) CLOX: an executive clock 41. Tolboom N, Koedam ELGE, Schott JM et al (2010) Dementia drawing task. J Neurol Neurosurg Psychiatry 64:588–594 mimicking Alzheimer’s disease owing to a tau mutation: CSF and 31. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) PET findings. Alzheimers Dis Assoc Disord 24:303–307 An inventory for measuring depression. Arch Gen Psychiatry 42. Hornberger M, Piguet O (2012) Episodic memory in frontotem- 4:561–571 poral dementia: a critical review. Brain 135:678–692 32. Youden WJ (1950) Index for rating diagnostic tests. Cancer 3:32–35 43. Spina S, Schonhaut D, Boeve BF et  al (2017) Frontotempo- 33. Barandiaran M, Estanga A, Moreno F et al (2012) Neuropsycho- ral dementia with the V337M MAPT mutation. Neurology logical features of asymptomatic c.709-1G > A progranulin muta- 88(8):758–766 tion carriers. JINS 18:1086–1090 44. Snowden JS, Adams J, Harris J et al (2015) Distinct clinical and 34. Janssen JC, Schott JM, Cipolotti L et al (2005) Mapping the onset pathological phenotypes in frontotemporal dementia associated and progression of atrophy in familial frontotemporal lobar degen- with MAPT, PGRN and C9orf72 mutations. Amyotroph Lateral eration. JNNP 76(2):162–168 Scler Frontotemp Degener 16:497–505 35. Rohrer JD, Warren JD, Barnes J et al (2008) Mapping the progres- 45. Hassenstab J, Aschenbrenner AJ, Balota DA et al (2016) Cog- sion of progranulin-associated frontotemporal lobar degeneration. nitive trajectories in DIAN: relationships with symptom onset, Nat Clin Pract Neurol 4(8):455–460 mutation types and clinical status. Alzheimers Dement 12(7):368 36. Ferman TJ, McRae CA, Arvanitakis Z et al (2003) Early and 46. Meeter LH, Donker Kaat L, Rohrer JD, van Swieten JC (2017) pre-symptomatic neuropsychological dysfunction in the PPND Imaging and fluid biomarkers in frontotemporal dementia. Nat family with the N279K tau mutation. Parkinsonism Relat Disord Rev Neurol 13(7):406–419 9(5):265–270 47. Rosness TA, Engedal K, Chemali Z (2016) Frontotemporal 37. Miller JB, Banks SJ, Léger GC, Cummings JL (2014) Randomized dementia: an updated clinician’s guide. J Geriatr Psychiatry controlled trials in frontotemporal dementia: cognitive and behav- 29(5):271–280 ioral outcomes. Transl Neurodegener 3:12 48. Pakhomov SVS, Hemmy LS (2014) A computational linguistic 38. Whitwell JL, Jack CR, Boeve BF et al (2009) Voxel-based mor- measure of clustering behavior on semantic verbal fluency task phometry patterns of atrophy in FTLD with mutations in MAPT predicts risk of future dementia in the Nun study. Cortex 55:97–106 or PGRN. Neurology 72:813–820 49. van den Berg E, Jiskoot LC, Grosveld MJH et al (2017) Qualita- 39. Rohrer JD, Ridgway GR, Modat M et al (2010) Distinct profiles tive assessment of verbal fluency performance in frontotemporal of brain atrophy in frontotemporal lobar degeneration caused by dementia. Dement Geriatr Cogn Disord 44:35–44 progranulin and tau mutations. Neuroimage 53:1070–1076 50. Laisney M, Matuszewski V, Mézenge F et al (2009) The under- 40. Smith R, Puschmann A, Schöll M et  al (2017) 18F-AV-1451 lying mechanisms of verbal fluency deficit in frontotemporal tau PET imaging correlates strongly with tau neuropathology in dementia and semantic dementia. J Neurol 256:1083 MAPT mutation carriers. Brain 139(9):2372–2379 Affiliations 1,2 1,2 1 1 1 Lize C. Jiskoot  · Jessica L. Panman  · Lauren van Asseldonk  · Sanne Franzen  · Lieke H. H. Meeter  · 1,3 1 1 4 5 Laura Donker Kaat  · Emma L. van der Ende  · Elise G. P. Dopper  · Reinier Timman  · Rick van Minkelen  · 1,6 1 1 John C. van Swieten  · Esther van den Berg  · Janne M. Papma Lize C. Jiskoot John C. van Swieten l.c.jiskoot@erasmusmc.nl j.c.vanswieten@erasmusmc.nl Jessica L. Panman Esther van den Berg j.panman@erasmusmc.nl e.vandenberg@erasmusmc.nl Lauren van Asseldonk Department of Neurology, Erasmus Medical Center laurenvanasseldonk@gmail.com Rotterdam, Room Ee2240, ’s-Gravendijkwal 230, Sanne Franzen 3015 CE Rotterdam, The Netherlands s.franzen@erasmusmc.nl Department of Radiology, Leiden University Medical Center, Lieke H. H. Meeter Leiden, The Netherlands h.meeter@erasmusmc.nl Department of Clinical Genetics, Leiden University Medical Laura Donker Kaat Center, Leiden, The Netherlands l.donkerkaat@erasmusmc.nl Department of Psychiatry, Section of Medical Psychology Emma L. van der Ende and Psychotherapy, Erasmus Medical Center, Rotterdam, e.vanderende@erasmusmc.nl The Netherlands Elise G. P. Dopper Department of Clinical Genetics, Erasmus Medical Center, e.dopper@erasmusmc.nl Rotterdam, The Netherlands Reinier Timman Department of Clinical Genetics, VU Medical Center, r.timman@erasmusmc.nl Amsterdam, The Netherlands Rick van Minkelen r.vanminkelen@erasmusmc.nl 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Neurology Springer Journals

Longitudinal cognitive biomarkers predicting symptom onset in presymptomatic frontotemporal dementia

Free
12 pages
Loading next page...
 
/lp/springer_journal/longitudinal-cognitive-biomarkers-predicting-symptom-onset-in-AUIWBZ8pYP
Publisher
Springer Berlin Heidelberg
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Neurology; Neurosciences; Neuroradiology
ISSN
0340-5354
eISSN
1432-1459
D.O.I.
10.1007/s00415-018-8850-7
Publisher site
See Article on Publisher Site

Abstract

Introduction We performed 4-year follow-up neuropsychological assessment to investigate cognitive decline and the prog- nostic abilities from presymptomatic to symptomatic familial frontotemporal dementia (FTD). Methods Presymptomatic MAPT (n = 15) and GRN mutation carriers (n = 31), and healthy controls (n = 39) underwent neuropsychological assessment every 2 years. Eight mutation carriers (5 MAPT, 3 GRN) became symptomatic. We investi- gated cognitive decline with multilevel regression modeling; the prognostic performance was assessed with ROC analyses and stepwise logistic regression. Results MAPT converters declined on language, attention, executive function, social cognition, and memory, and GRN converters declined on attention and executive function (p < 0.05). Cognitive decline in ScreeLing phonology (p = 0.046) and letter fluency ( p = 0.046) were predictive for conversion to non-fluent variant PPA, and decline on categorical fluency (p = 0.025) for an underlying MAPT mutation. Discussion Using longitudinal neuropsychological assessment, we detected a mutation-specific pattern of cognitive decline, potentially suggesting prognostic value of neuropsychological trajectories in conversion to symptomatic FTD. Keywords Presymptomatic · Frontotemporal dementia · Familial · Biomarkers · Cognition · Neuropsychological assessment · Longitudinal Introduction characterized by deficits in executive function, social cog - nition and language, whereas memory and visuoconstruction Frontotemporal dementia (FTD) is a presenile neurodegen- are initially spared [3–5]. Non-fluent variant PPA (nfvPPA) erative disorder, leading to a heterogeneous clinical pres- patients show agrammatism and speech sound distortions, entation, involving behavioural (behavioural variant FTD; while semantic variant PPA (svPPA) patients experience def- bvFTD) and/or language deterioration (primary progressive icits in confrontation naming and word comprehension [6]. aphasia; PPA) [1]. FTD has an autosomal dominant pattern GRN mutations often lead to a clinical diagnosis of bvFTD, of inheritance in 30 percent of cases, with mutations in the nfvPPA or parkinsonism. In MAPT mutations, bvFTD is the progranulin (GRN) and microtubule-associated protein tau main phenotype, and semantic and memory impairments can (MAPT) genes as its two main causes [2]. The cognitive be prominent neuropsychological symptoms [7]. profile of FTD varies depending on the clinical phenotype Research in familial FTD has demonstrated the presence and the underlying genotype. Patients with bvFTD are of a presymptomatic stage in which subtle cognitive changes have been identified [ 8–12]. More specifically, cognitive decline can start as early as 8 years prior to estimated symp- Electronic supplementary material The online version of this tom onset and shows mutation-specific patterns, with GRN article (https ://doi.org/10.1007/s0041 5-018-8850-7) contains supplementary material, which is available to authorized users. mutation carriers declining in memory, and MAPT mutation carriers declining in language, social cognition and memory * Janne M. Papma [8, 10]. This suggests that cognitive measures could function j.papma@erasmusmc.nl as disease-tracking biomarkers in the presymptomatic stage. Extended author information available on the last page of the article Vol.:(0123456789) 1 3 1382 Journal of Neurology (2018) 265:1381–1392 However, it is currently unknown what the long-term cogni- tive profiles of presymptomatic FTD mutations are, whether neuropsychological assessment can be used to track disease progression to the symptomatic stage, and what the prognos- tic value is of cognitive trajectories in the presymptomatic and early symptomatic stage of FTD. In this study, we investigated longitudinal cognitive decline on neuropsychological assessment in presympto- matic mutation carriers (MAPT or GRN) and controls from the same families within our longitudinal presymptomatic Dutch familial FTD Risk Cohort (FTD-RisC). Second, we assessed the difference in cognitive course between convert- ers’ genotypes (i.e. MAPT vs. GRN) and phenotypes (i.e. bvFTD vs. nfvPPA) versus non-converters. Lastly, we inves- tigated the prognostic value of neuropsychological trajecto- ries in predicting symptom onset within 2–4 years. Methods Participants In FTD-RisC, we follow healthy 50% at-risk family mem- bers from genetic FTD families on a 2-year basis. In the Fig. 1 Participant in- and exclusion and sample size per time point. Two controls were excluded as they had multiple cognitive disorders current study, we included 87 participants from MAPT or (≤ 2 SD below reference mean) on neuropsychological testing. Eight GRN families with study entries between December 2009 mutation carriers converted to clinical FTD within the study window. and January 2013 [8, 9, 13]. The follow-up period was Their data were restructured, so that there were three time points: 4 years, in which we acquired neuropsychological assess- 4  years before symptom onset, 2  years before symptom onset and symptom onset. Four years before symptom onset, only data of six ments at study entry, follow-up after 2 years and follow-up converters were available, as two mutation carriers converted between after 4 years. DNA genotyping (see “Procedure”) assigned baseline and first follow-up. The data of converters were compared participants either to the presymptomatic mutation carrier to, respectively, baseline, follow-up after 2  years and follow-up after (n = 46; 31 GRN, 15 MAPT), or control group (n = 39; 29 4 years in non-converters and healthy controls GRN, 10 MAPT family members). We excluded two controls as they had cognitive disorders (≥ 2 SD below mean) on neurological examination, and MR imaging of the brain. multiple domains, ultimately including 85 participants (46 mutation carriers, 37 controls; Fig. 1). DNA sequencing was performed at study entry. All par- ticipants were asymptomatic according to established Standard protocol approvals, registrations, diagnostic criteria for bvFTD [3] or PPA [6] at baseline. Knowledgeable informants were asked about cognitive and/ and patient consents or behavioural deterioration at each study visit by means of a structured interview and a well-validated questionnaire Clinical investigators were blind for participants’ genetic sta- tus if they had not undergone predictive testing. In case of (Neuropsychiatric Inventory; NPI) [14]. conversion to clinical FTD, we offered the patient and family members genetic counselling and unblinding of genetic sta- Converters tus, to confirm the presence of the pathogenic mutation. At study entry, all participants gave written informed consent. Eight mutation carriers became symptomatic within the study time window (“converters”). Symptom onset was The study was approved by the Medical and Ethical Review Committee of the Erasmus Medical Center. determined by means of the above mentioned assessment (anamnesis, MR imaging of the brain, neuropsychological Procedure assessment, heteroanamnestic information and unblinding of genetic status). Conversion was determined in a multi- Every 2  years, participants underwent a standardized disciplinary consensus meeting of the Erasmus MC FTD Expertise Centre, involving neurologists (LDK, JCvsS), assessment consisting of a neuropsychological test battery, 1 3 Journal of Neurology (2018) 265:1381–1392 1383 neuropsychologists (LCJ, JLP, SF, EvdB, JMP), medical 2 years, follow-up after 4 years) into the following three doctors (LHM, ELvdE), as well as neuroradiologists, geri- new time points (Fig. 1): atricians, a clinical geneticist (RvM), and a care consultant. Six converters (5 MAPT, 1 GRN) presented with progres- 4 years before symptom onset: we used the data of the sive behaviour deterioration, functional decline, and frontal study visit 4 years before diagnosis. Analyses could were and/or temporal lobe atrophy on MRI, fulfilling the interna- performed in six converters, as two (1 GRN, 1 MAPT—2 tional diagnostic consensus criteria of Rascovsky et al. [3] bvFTD) developed symptoms between baseline and first for bvFTD with definite FTLD pathology. Two converters follow-up (i.e. at 2 years follow-up), and therefore no (both GRN) presented with isolated language difficulties data 4 years prior to symptom onset were available. and no impairments in daily living activities, thereby fulfill- 2 years before symptom onset: we used the data of the ing the diagnostic criteria for PPA of Gorno-Tempini et al. study visit 2 years before diagnosis. Analyses included [6]. Both developed nfvPPA, as they showed a non-fluent, all eight converters. halting speech, with sound errors and agrammatism. See After symptom onset: we used the data of the diagnosis Supplementary Table 1 for demographic, clinical and neu- visit. Analyses included all eight converters. ropsychological data of the converters. We defined mutation carriers remaining without FTD symptoms as non-convert- In non-converters and controls, we used the original time ers (n = 38; 28 GRN, 10 MAPT). points: baseline (data were compared to “4  years before symptom onset” data of converters), follow-up after 2 years Neuropsychological assessment (data were compared to “2 years before symptom onset data of converters) and follow-up after 4 years (data were com- We screened global cognitive functioning by means of the pared to “after symptom onset data of converters). Mini-Mental State Examination [15] (MMSE) and frontal assessment battery [16] (FAB). Experienced neuropsycholo- Statistical analysis gists (LCJ, JLP, SF) administered neuropsychological tests within six cognitive domains: language, attention and mental Statistical analyses were performed using SPSS Statistics processing speed, executive functioning, social cognition, 21.0 (IBM Corp., Armonk, NY) and GraphPad Prism 7 memory, and visuoconstruction. We rated language with the (La Jolla, California, USA), with the significance level at 60-item Boston Naming Test (BNT) [17], verbal Semantic p < 0.05 (two-tailed) across all comparisons. We compared Association Test (SAT) [18], ScreeLing phonology [19], and demographic data between MAPT mutation carriers, GRN categorical fluency [20]. We assessed attention and mental mutation carriers and controls, and between converters, non- processing speed by means of Trail making Test (TMT)-A converters and controls by means of one-way ANOVAs. We [21], Stroop Color-Word Test I and II [22], Wechsler Adult performed Pearson Χ tests to investigate differences in sex. Intelligence Scale III (WAIS-III) Digit Span forwards [23], Longitudinal comparisons of clinical data were performed and Letter Digit Substitution Test (LDST) [24]. Execu- with repeated measures ANOVAs. We standardized all raw tive functioning was evaluated using TMT-B [21], Stroop neuropsychological test scores by converting them into Color-Word Test III [22], WAIS-III Digit Span backwards z-scores (i.e. individual test score minus the baseline mean [23], modified Wisconsin Card Sorting Test (WCST) con- of the controls, divided by the baseline SD of the controls) cepts [25], letter fluency [ 20], and WAIS-III Similarities per time point, after which we calculated composite z-scores [23]. Happé cartoons [26] and Ekman Faces [27] measured for the respective six cognitive domains by averaging the social cognition. We assessed memory using the Dutch Rey z-scores of the individual tests per domain. For the longi- Auditory Verbal Learning Test (RAVLT) [28] and Visual tudinal comparisons we used multilevel linear regression Association Test (VAT) [29]. We evaluated visuoconstruc- modeling. This analysis corrects for bias when data absence tion by means of clock drawing [30] and WAIS-III Block is dependent on characteristics present in the model, and Design [23]. Alternate forms were used at follow-up visits, can therefore efficiently handle missing and unbalanced time when applicable (letter fluency, RAVLT, VAT). Depressive points. There were two levels in the models: the partici- symptoms were rated with the Beck’s Depression Inventory pants constituted the upper level; their repeated measures the (BDI) [31]. lower level. We ran two analyses to assess cognitive decline per mutation (1) and clinical status (2): Study design 1. We entered mutation status (MAPT mutation carrier, In converters, we restructured the three original time points GRN mutation carrier or control), time (4 years before within our study window (i.e. baseline, follow-up after symptom onset, 2 years before symptom onset, and after symptom onset), and first-order interactions, with age, 1 3 1384 Journal of Neurology (2018) 265:1381–1392 gender and educational level as covariates. We reran the in MAPT converters than GRN converters (p = 0.043) and analyses excluding the converters to exclude convert- non-converters (p = 0.001; Table 1). Both MAPT and GRN ers driving the cognitive decline in the mutation carrier converters declined significantly with respect to MMSE groups; score (p < 0.001) and they developed more neuropsychi- 2. We split the converter group according to genotype atric symptoms in the form of higher BDI (p = 0.001) and (MAPT or GRN) and phenotype (bvFTD or nfvPPA) NPI (p = 0.021) scores in comparison to non-converters and to investigate specific profiles of cognitive decline over controls. FAB scores did not significantly change over time time. We then entered clinical status (converter, non- (p > 0.05). converter or control), time, and first-order interactions, with age, gender and educational level as covariates. Longitudinal cognitive decline in MAPT and GRN mutation carriers Third, to investigate the prognostic abilities of cognitive decline in discriminating between converters and non-con- The whole group of MAPT mutation carriers declined sig- verters, we determined the area under the curve (AUC) by nificantly within the domains language, social cognition and receiver operating characteristic (ROC) analyses on the neu- memory compared with controls (Table 2; Fig. 1). This was ropsychological trajectories between visits. For this, we cal- reflected in lower scores on the BNT and categorical fluency, culated deltas between test scores; one between 4 and 2 years Happé cartoons, VAT and RAVLT delayed recall (Table 2). before symptom onset and one between 2 years before symp- In the whole group of GRN mutation carriers, no longitu- tom onset and symptom onset. Optimal cut-off levels were dinal decline was found in comparison to controls. In com- given by the highest Youden’s index [32]. Again, we split the parison to GRN mutation carriers, MAPT mutation carriers converter group according to genotype (MAPT or GRN) and declined significantly on the domains language (β = − 0.015, phenotype (bvFTD or nfvPPA). Next, we performed logistic p < 0.001) and memory (β = −  0.016, p = 0.008), reflected regression analyses, taking group (converter vs. non-con- in lower BNT (β = −  0.085, p = 0.01), SAT (β = −  0.027, verter) as the dependent variable and the deltas (tests with p = 0.015), category fluency (β = −  0.107, p = 0.002), and significant diagnostic performance in abovementioned ROC RAVLT delayed recall (β = − 0.047, p = 0.001) scores. There analyses) as the independent variables. The models were were no cognitive domains or tests on which GRN muta- selected with a forward stepwise method according to the tion carriers declined more than MAPT mutation carriers likelihood ratio test and applying the standard p values for (Table  2). By excluding the five MAPT converters from variable inclusion (0.05) and exclusion (0.10), with age, sex the analyses, none of the domain scores in MAPT mutation and education as covariates. Goodness of fit was evaluated carriers continued to show significant decline over time in 2 2 with the HL Χ test. Nagelkerke R is reported as measure of comparison to controls. Regarding individual tests, however, effect size. We checked predictor variables for multicollin- the decline on the RAVLT delayed recall remained signifi- earity. All models were corrected for multiple comparisons cant (β = − 0.032, p = 0.023). The results did not change by (Bonferroni). excluding the three GRN converters from the analyses. In comparison to GRN, MAPT mutation carriers still declined more on language (β = − 0.010, p = 0.004), reflected in lower Results ScreeLing phonology (β = − 0.008, p = 0.024) and category fluency (β = −  0.007, p = 0.041). There was no cognitive Demographics decline in controls—but significant improvement was found on social cognition (Happé non-ToM and Ekman Faces) and MAPT mutation carriers were significantly younger than memory (RAVLT immediate and delayed recall) (Table 2). GRN mutation carriers (p = 0.012; Table  1). The mean The raw neuropsychological test scores per time point can familial symptom onset age was lower in MAPT than in be found in Supplementary Table 2. GRN mutation carriers and controls (both p < 0.001). There were no significant differences between groups regarding Longitudinal cognitive decline in converters estimated years to symptom onset (p > 0.05). Longitudinal and non‑converters analyses demonstrated that MAPT mutation carriers declined significantly more than GRN mutation carriers and controls Converters with a MAPT mutation deteriorated significantly with regards to the MMSE (p = 0.014), and also developed on all domains but visuoconstruction (Fig. 2a–d, f; Table 3). more depressive symptoms (p = 0.028). FAB and NPI scores Within these domains, performances declined on BNT did not significantly change over time (p > 0.05). Convert- (p < 0.001), LDST (p = 0.035), Stroop I, II and III (I: p = 0.017; ers, non-converters and controls did not differ regarding II: p < 0.001; III: p = 0.021), categorical fluency (p = 0.001), demographic variables, apart from a younger family onset WAIS similarities (p < 0.001), Happé ToM (p = 0.011), and 1 3 Journal of Neurology (2018) 265:1381–1392 1385 Table 1 Demographics and clinical data Demographics HC (n = 39) MAPT carriers GRN carriers p value* MAPT GRN converters Non-converters p value** (n = 15) (n = 31) converters (n = 3) (n = 38) (n = 5) Age at study 49.1 ± 12.2 41.9 ± 10.0 52.1 ± 8.2 0.012 45.3 ± 8.5 54.9 ± 9.0 48.8 ± 10.3 0.704 entry, years Sex, female (%) 20 (56%) 7 (47%) 20 (65%) 0.506 1 (20%) 3 (100%) 23 (60.5%) 0.154 Education 5.2 ± 1.0 5.1 ± 1.6 5.7 ± 0.9 0.102 6.0 ± 0.7 5.7 ± 0.6 5.4 ± 1.3 0.409 (Verhage) a,b c,d Onset age fam- 59.0 ± 5.8 51.3 ± 6.7 61.0 ± 2.4 < 0.001 48.0 ± 4.7 59.7 ± 0.0 58.8 ± 6.1 0.002 ily, years Estimated years − 10.2 ± 11.2 − 7.7 ± 9.6 − 9.4 ± 7.9 0.690 − 2.7 ± 4.0 − 4.8 ± 9.0 − 10.0 ± 8.5 0.335 to onset, years Clinical Years HC (n = 39) MAPT carriers GRN carriers p value* MAPT GRN converters Non-converters p value** data to (n = 15) (n = 31) converters (n = 3) (n = 38) onset (n = 5) MMSE 4 29.1 ± 1.3 29.6 ± 0.5 29.1 ± 1.6 0.451 29.5 ± 0.6 29.0 ± 1.4 29.2 ± 1.4 0.924 2 29.2 ± 1.3 28.7 ± 2.2 28.9 ± 1.6 0.513 29.8 ± 0.4 28.0 ± 1.0 27.7 ± 1.5 0.271 c,d 0 29.2 ± 1.0 28.4 ± 1.5 29.2 ± 1.4 0.099 27.2 ± 1.6 27.7 ± 1.5 29.3 ± 1.2 0.001 FAB 4 – – – – – – – – 2 17.4 ± 0.9 17.4 ± 0.8 17.5 ± 0.9 0.883 17.3 ± 1.0 17.5 ± 0.7 17.5 ± 0.9 0.929 0 16.7 ± 1.7 16.5 ± 1.6 17.0 ± 1.1 0.639 15.4 ± 1.5 16.3 ± 1.5 17.1 ± 1.1 0.120 BDI 4 4.1 ± 4.5 4.0 ± 6.3 3.2 ± 3.9 0.693 1.3 ± 1.0 2.0 ± 2.8 3.7 ± 5.0 0.645 2 3.7 ± 3.9 4.5 ± 5.0 3.2 ± 4.0 0.638 5.0 ± 4.7 2.7 ± 3.8 3.5 ± 4.4 0.866 c,d 0 3.5 ± 4.3 7.6 ± 9.5 3.0 ± 6.7 0.108 11.6 ± 13.0 6.3 ± 5.1 3.1 ± 6.5 0.042 c–e NPI 4 0.1 ± 0.5 4.6 ± 11.2 1.4 ± 3.4 0.180 0.0 ± 0.0 – 3.0 ± 7.5 0.006 2 0.6 ± 1.2 6.4 ± 20.7 0.3 ± 0.7 0.095 0.2 ± 0.4 0.0 ± 0.0 2.9 ± 13.3 0.767 a,b d 0 0.8 ± 1.5 12.3 ± 18.7 2.1 ± 6.6 0.001 15.6 ± 16.3 10.7 ± 15.9 3.4 ± 11.4 0.009 Values indicate: mean ± standard deviation. Significant comparisons are displayed in bold GRN progranulin, HC healthy control, MMSE Mini-Mental State Examination, FAB frontal assessment battery, BDI Beck’s depression inven- tory, NPI neuropsychiatric inventory *p value represents result of overall ANOVA between MAPT mutation carriers, GRN mutation carriers and healthy controls **p value represents result of overall ANOVA between MAPT converters, GRN converters, non-converters and HC Significant post hoc test between MAPT and GRN mutation carriers Significant post hoc test between MAPT mutation carriers and healthy controls Significant post hoc test between converters and non-converters Significant post hoc test between converters and healthy controls Only data of MAPT converters available, therefore the p value represents the comparison between MAPT converters, non-converters and HC Dutch educational system categorized into levels from 1 = less than 6 years of primary education to 7 = academic schooling (Verhage, 1964) Data only available on follow-up visits RAVLT immediate (p = 0.004) and delayed recall (p = 0.030). as GRN converters, with lower scores on attention and execu- Converters with a GRN mutation deteriorated significantly on tive function (Table 3). There were no differences in decline attention and mental processing speed, and executive function between converters with bvFTD and nfvPPA (Table 3). The (Fig. 2b, c; Table 3). Within these domains, performances on raw neuropsychological test scores per time point can be found TMT-B (p < 0.001), Stroop III (p < 0.001), WCST (p = 0.005), in Supplementary Table 3. letter fluency (p = 0.012) and WAIS similarities (p < 0.001) deteriorated significantly over time. Converters with bvFTD Classification between converters had a similar pattern of cognitive decline as MAPT convert- and non‑converters ers, with lower scores on social cognition, memory, language, attention and executive function (Table 3). Comparably, con- Between 4 and 2  years before symptom onset, the delta verters with nfvPPA had a similar pattern of cognitive decline domain and individual neuropsychological test scores 1 3 1386 Journal of Neurology (2018) 265:1381–1392 Table 2 Cognitive trajectories in mutation carriers (converters, non-converters) and healthy controls Domain test Healthy controls (n = 39) MAPT mutation carriers (n = 15) GRN mutation carriers (n = 31) Baseline β p Baseline β p Baseline β p Language 0.0 ± 0.6 0.000 0.931 0.2 ± 0.6 − 0.010 0.002 0.1 ± 0.7 0.004 0.121 BNT 53.4 ± 4.5 0.026 0.105 52.6 ± 5.3 − 0.080 0.005 55.1 ± 3.7 0.006 0.786 SAT 27.8 ± 1.1 − 0.003 0.604 27.9 ± 1.5 − 0.008 0.604 27.5 ± 2.0 0.019 0.033 ScreeLing phonology 23.5 ± 0.8 0.001 0.733 23.9 ± 0.3 − 0.005 0.190 23.8 ± 0.5 − 0.001 0.863 Categorical fluency 23.9 ± 4.9 0.026 0.141 26.5 ± 6.6 − 0.087 0.006 23.4 ± 5.7 0.021 0.424 Attention and processing speed 0.0 ± 0.8 − 0.001 0.084 0.3 ± 0.6 − 0.003 0.096 0.1 ± 0.9 − 0.003 0.075 TMT part A 31.8 ± 15.0 − 0.022 0.416 26.1 ± 9.7 0.065 0.192 31.4 ± 12.2 0.060 0.145 Stroop card I 47.1 ± 8.0 0.039 0.011 43.2 ± 8.8 − 0.017 0.529 45.0 ± 8.4 − 0.001 0.951 Stroop card II 58.5 ± 10.6 0.012 0.539 54.9 ± 8.5 0.027 0.470 60.2 ± 13.2 0.001 0.969 Digit Span forwards 8.7 ± 1.9 0.001 0.871 9.0 ± 2.6 − 0.010 0.294 9.4 ± 2.4 − 0.016 0.055 LDST 34.5 ± 6.8 0.001 0.894 34.2 ± 4.7 − 0.636 0.699 33.2 ± 7.4 0.005 0.798 Executive function 0.0 ± 0.7 0.001 0.505 0.3 ± 0.6 − 0.005 0.065 0.2 ± 0.8 − 0.004 0.052 TMT part B 67.8 ± 29.3 0.052 0.494 61.0 ± 28.5 0.079 0.570 72.2 ± 42.7 − 0.099 0.390 Stroop card III 93.7 ± 22.6 − 0.087 0.021 83.8 ± 14.7 0.141 0.042 96.6 ± 26.2 0.013 0.815 Digit span backwards 6.1 ± 2.0 0.008 0.194 6.6 ± 1.8 0.002 0.877 6.6 ± 2.1 − 0.011 0.222 WCST concepts 5.5 ± 0.9 0.002 0.592 5.6 ± 1.1 − 0.009 0.296 5.80 ± 0.6 − 0.010 0.144 Letter fluency 32.1 ± 9.9 0.134 < 0.001 36.1 ± 14.3 − 0.108 0.049 38.9 ± 12.0 − 0.062 0.173 Similarities 24.8 ± 4.7 0.006 0.645 25.5 ± 4.7 − 0.034 0.122 26.2 ± 5.0 − 0.011 0.556 Social cognition 0.0 ± 0.8 0.000 0.878 0.2 ± 0.7 − 0.009 0.007 0.3 ± 0.7 − 0.003 0.332 Happé ToM 11.8 ± 3.4 0.013 0.172 12.6 ± 3.7 − 0.044 0.011 12.9 ± 2.9 − 0.005 0.707 Happé non-Tom 11.7 ± 2.9 0.020 0.013 12.4 ± 2.8 − 0.036 0.017 13.0 ± 2.6 − 0.012 0.331 Ekman faces 45.7 ± 6.4 0.038 0.009 47.0 ± 5.5 − 0.028 0.293 47.10 ± 5.5 − 0.013 0.548 Memory 0.0 ± 0.7 0.000 0.848 0.1 ± 1.3 − 0.017 < 0.001 0.1 ± 0.9 − 0.001 0.745 VAT 11.8 ± 0.6 0.001 0.740 11.4 ± 1.6 − 0.012 0.019 11.5 ± 0.9 0.000 0.926 RAVLT imm. recall 42.6 ± 9.8 0.157 < 0.001 47.5 ± 9.7 − 0.076 0.090 46.3 ± 10.6 − 0.015 0.686 b a,b RAVLT del. recall 8.4 ± 3.2 0.050 < 0.001 9.7 ± 3.9 − 0.048 < 0.001 9.4 ± 3.3 − 0.000 0.983 RAVLT recognition 28.6 ± 2.1 0.014 0.127 29.0 ± 2.0 − 0.022 0.176 29.2 ± 1.2 − 0.009 0.505 Visuoconstruction 0.0 ± 0.8 − 0.001 0.656 − 0.2 ± 0.7 − 0.005 0.266 0.0 ± 1.0 0.000 0.963 Block design 36.5 ± 14.0 0.034 0.305 35.5 ± 20.8 − 0.006 0.917 39.3 ± 18.5 − 1.164 0.246 Clock drawing 12.6 ± 1.4 0.003 0.453 12.2 ± 1.3 − 0.009 0.284 12.4 ± 1.8 0.005 0.475 Values indicate: mean ± standard deviation; β represents estimate of change over time. Composite domain scores are z-scores, individual test scores are raw scores. Composite domain scores are expressed as z-scores, the individual test scores are raw scores. p values represent compari- sons to healthy controls. Significant comparisons are displayed in bold MAPT microtubule-associated protein tau, GRN progranulin, BNT Boston Naming Test, SAT semantic association test, TMT Trail making Test, WAIS Wechsler Adult Intelligence Scale, LDST letter digit substitution test, WCST Wisconsin card sorting test, ToM theory of mind, VAT visual association test, RAVLT Rey Auditory Verbal Learning Test, imm immediate, del delayed Remained significant after excluding converters from the analyses Survived Bonferroni correction for multiple comparisons Higher scores and β weights indicate worse performance failed to distinguish significantly between converters and Discussion non-converters. Between 2 years before symptom onset and symptom onset decline on categorical fluency was predic- This study examined a large cohort of at-risk participants tive of an underlying MAPT mutation (p = 0.025; Table 4). from GRN and MAPT FTD families by means of neuropsy- Decline on ScreeLing phonology (p = 0.046) and letter flu- chological assessment during a 4-year follow-up. Within the ency (p = 0.046) was predictive of conversion to nfvPPA study time window, eight mutation carriers became symp- (Table 4). tomatic. Converters with a MAPT and GRN mutation had mutual as well as gene-specific profiles of cognitive decline. 1 3 Journal of Neurology (2018) 265:1381–1392 1387 Fig. 2 Multilevel linear regression model displaying longitudi- tion and mental processing speed, c executive functioning, d memory, nal decline (4  years, 2  years and after symptom onset) in composite e visuoconstruction, and f language. NB: the healthy controls have a domain z-score in the total group of converters (light green), MAPT mean z-score of zero by default as the z-scores of mutation carriers converters (light blue dotted line), GRN converters (dark blue dotted were based on that (raw score minus mean score of healthy controls, line), non-converters (dark green) and healthy controls (black). Mod- divided by the standard deviation of healthy controls). MAPT micro- els are displayed per cognitive domain: a social cognition, b atten- tubule-associated protein tau, GRN progranulin Cognitive decline on categorical fluency between 2 years neuropsychological test scores remain static while mutation before conversion and symptom onset was predictive for carriers are presymptomatic, and cognitive decline starts an underlying MAPT mutation, and decline on ScreeLing only near or at symptom onset [34–36], suggesting an explo- phonology and letter fluency was predictive for conversion sive rather than gradual start of the symptomatic disease to nfvPPA. These results suggest that neuropsychological stage. Alternatively, we might be unable to pick up subtle assessment could provide sensitive clinical biomarkers to cognitive changes in presymptomatic mutation carriers due identify and track FTD mutation carriers at-risk of convert- to lack of power. Also, although well-validated, most of our ing to the symptomatic stage. These findings hold potential neuropsychological tests were not developed for repeated for improving early clinical diagnosis by identifying the administration in a preclinical population [37]. We there- most sensitive neuropsychological tests for conversion, and fore cannot rule out that familiarity and/or practice effects use in upcoming disease-modifying clinical trials. are obscuring subtle cognitive decline, a notion that seems Following the MAPT mutation carriers over a 4-year to be underwritten by improvement in social cognition and period, we found significant decline in language, social cog- memory in controls, but not mutation carriers. nition and memory. This is consistent with findings from In our exploratory analyses in converters, we discovered previous presymptomatic familial FTD studies, in which both common as well as mutation-specific profiles of cogni- both cross-sectional [9–11, 33] and longitudinal [8] decline tive decline in MAPT and GRN. In both mutations, decline was found. Specifically, in our first follow-up study [8 ], we in attention, mental processing speed and executive function demonstrated decline in the domains language, social cog- was found—while only converters with a MAPT mutation nition and memory 5–8 years before estimated symptom demonstrated decline on language, memory and social cogni- onset. It should be taken into account that this study made tion. Previous studies in familial FTD also point to distinct use of estimated onset as a proxy, instead of actual symp- profiles for MAPT and GRN [8, 10–12], and are largely con- tom onset as in the present study—but the similar profile sistent with our present findings. Another important aspect is of decline confirms the presence of early changes in these the longitudinal tracking of the different clinical phenotypes. three domains. As in our previous study, the present results The similar patterns of cognitive decline in bvFTD as MAPT, are largely driven by the converters. This could suggest that and nfvPPA as GRN are related to the dominant genotype in 1 3 1388 Journal of Neurology (2018) 265:1381–1392 1 3 Table 3 Cognitive trajectories in MAPT, GRN, bvFTD and nfvPPA converters, and non-converters Domain test MAPT converters (n = 5) GRN converters (n = 3) bvFTD converters (n = 6) nfvPPA converters (n = 2) Non-converters (= 38) Baseline β p Baseline β p Baseline β p Baseline β p Baseline β p a a Language 0.1 ± 0.7 − 0.028 < 0.001 0.6 ± 0.2 − 0.007 0.299 0.1 ± 0.7 − 0.025 < 0.001 0.6 ± 0.2 − 0.014 0.061 0.1 ± 0.6 0.002 0.408 a a BNT 54.3 ± 6.9 − 0.239 < 0.001 57.5 ± 2.1 − 0.019 0.604 54.3 ± 6.9 − 0.224 < 0.001 57.5 ± 2.1 − 0.033 0.396 54.2 ± 4.2 − 0.001 0.960 SAT 27.0 ± 1.4 − 0.040 0.034 28.0 ± 1.4 0.006 0.805 27.0 ± 1.4 − 0.036 0.052 28.0 ± 1.4 0.000 0.993 27.7 ± 2.0 0.013 0.127 ScreeLing phonology 24.0 ± 0.0 0.002 0.617 24.0 ± 0.0 − 0.011 0.114 24.0 ± 0.0 0.004 0.358 24.0 ± 0.0 − 0.017 0.018 23.8 ± 0.4 − 0.002 0.551 a a Categorical fluency 25.8 ± 4.6 − 0.250 < 0.001 28.0 ± 2.8 − 0.149 0.022 25.8 ± 4.6 − 0.237 < 0.001 28.0 ± 2.8 − 0.170 0.015 24.0 ± 6.3 0.014 0.546 Attention and mental 0.3 ± 0.6 − 0.010 0.006 0.2 ± 0.3 − 0.013 0.005 0.3 ± 0.6 − 0.010 0.004 0.2 ± 0.3 − 0.013 0.006 0.1 ± 0.8 − 0.001 0.321 processing speed TMT part A 20.0 ± 6.3 0.067 0.448 25.0 ± 8.5 0.073 0.539 20.0 ± 6.3 0.065 0.449 25.0 ± 8.5 0.090 0.483 31.1 ± 11.8 0.051 0.181 Stroop card I 44.0 ± 5.2 0.101 0.030 46.5 ± 6.4 0.058 0.349 44.0 ± 5.2 0.106 0.020 46.5 ± 6.4 0.044 0.503 44.4 ± 8.9 − 0.020 0.345 b a a Stroop card II 58.5 ± 7.6 0.331 < 0.001 56.5 ± 0.7 0.186 0.006 57.5 ± 7.6 0.319 < 0.001 56.5 ± 0.7 0.194 0.008 58.8 ± 12.9 − 0.032 0.217 Digit Span forwards 9.5 ± 1.7 0.010 0.609 9.0 ± 0.0 − 0.038 0.146 9.5 ± 1.7 0.010 0.601 9.0 ± 0.0 − 0.043 0.119 9.3 ± 2.6 − 0.013 0.088 LDST 34.8 ± 6.7 − 0.100 0.012 35.0 ± 0.0 − 0.061 0.235 34.8 ± 6.7 − 0.098 0.011 35.0 ± 0.0 − 0.061 0.270 33.3 ± 6.9 0.004 0.809 a a Executive function 0.6 ± 0.4 − 0.018 < 0.001 0.6 ± 0.1 − 0.032 <0.001 0.6 ± 0.4 − 0.020 < 0.001 0.6 ± 0.1 − 0.029 < 0.001 0.2 ± 0.8 − 0.001 0.515 b a TMT part B 57.0 ± 27.0 0.472 0.038 48.0 ± 32.5 1.448 <0.001 57.0 ± 27.0 0.684 0.006 48.0 ± 32.5 0.921 0.010 71.2 ± 40.4 − 0.132 0.195 b a a a a Stroop card III 87.5 ± 23.4 0.468 < 0.001 86.5 ± 7.8 0.734 <0.001 87.5 ± 23.4 0.449 < 0.001 86.5 ± 7.8 0.815 < 0.001 93.7 ± 24.8 − 0.026 0.577 Digit span backwards 8.0 ± 1.4 − 0.018 0.284 5.5 ± 0.7 − 0.039 0.082 8.0 ± 1.4 − 0.022 0.186 5.5 ± 0.7 − 0.033 0.172 6.5 ± 2.0 − 0.003 0.721 WCST concepts 6.0 ± 0.0 − 0.015 0.193 6.0 ± 0.0 − 0.040 0.007 6.0 ± 0.0 − 0.021 0.073 6.0 ± 0.0 − 0.032 0.035 5.7 ± 0.8 − 0.006 0.323 Letter fluency 35.8 ± 7.9 − 0.143 0.101 45.5 ± 17.7 − 0.328 0.010 35.8 ± 7.9 − 0.156 0.066 45.5 ± 17.7 − 0.339 0.013 37.9 ± 13.0 − 0.048 0.245 a a a a Similarities 29.0 ± 1.2 − 0.151 < 0.001 29.0 ± 1.4 − 0.175 <0.001 29.0 ± 1.2 − 0.155 < 0.001 29.0 ± 1.4 − 0.175 < 0.001 25.5 ± 4.0 0.004 0.775 a a Social cognition 0.0 ± 1.0 − 0.022 < 0.001 0.8 ± 0.1 − 0.012 0.127 0.0 ± 1.0 − 0.021 < 0.001 0.8 ± 0.1 − 0.016 0.071 0.3 ± 0.7 − 0.002 0.336 Happé ToM 12.3 ± 5.1 − 0.096 0.002 13.5 ± 2.1 0.017 0.672 12.3 ± 5.1 − 0.078 0.012 13.5 ± 2.1 − 0.019 0.669 12.8 ± 3.0 − 0.012 0.380 Happé non-Tom 12.3 ± 2.4 − 0.067 0.010 15.5 ± 0.7 − 0.041 0.215 12.3 ± 2.4 − 0.060 0.016 15.5 ± 0.7 − 0.062 0.080 12.8 ± 2.7 − 0.012 0.267 Ekman faces 43.5 ± 6.1 − 0.089 0.023 50.0 ± 0.0 − 0.175 0.001 43.5 ± 6.1 − 0.118 0.003 50.0 ± 0.0 − 0.127 0.024 47.3 ± 5.4 − 0.001 0.965 a a Memory − 1.0 ± 2.0 − 0.050 < 0.001 0.7 ± 0.8 0.002 0.751 − 1.0 ± 2.0 − 0.044 < 0.001 0.7 ± 0.8 − 0.005 0.525 0.2 ± 0.8 − 0.002 0.473 VAT 10.0 ± 2.4 − 0.030 0.005 12.0 ± 0.0 0.004 0.675 10.0 ± 2.4 − 0.027 0.011 12.0 ± 0.0 0.000 0.983 11.6 ± 0.8 − 0.002 0.705 RAVLT imm. recall 42.5 ± 9.1 − 0.241 0.001 54.5 ± 19.1 − 0.111 0.226 42.5 ± 9.1 − 0.210 0.003 54.5 ± 19.1 − 0.177 0.067 46.7 ± 10.0 − 0.009 0.797 a a RAVLT del. recall 7.5 ± 5.5 − 0.085 < 0.001 10.5 ± 5.0 0.002 0.951 7.5 ± 5.5 − 0.080 < 0.001 10.5 ± 5.0 − 0.002 0.954 9.7 ± 3.2 − 0.009 0.359 RAVLT recognition 27.3 ± 3.1 − 0.037 0.005 30.0 ± 0.0 − 0.014 0.266 27.3 ± 3.1 − 0.036 0.004 30.0 ± 0.0 − 0.014 0.308 29.3 ± 1.1 − 0.009 0.461 Visuoconstruction 0.2 ± 0.8 − 0.009 0.217 0.2 ± 0.2 − 0.010 0.312 0.2 ± 0.8 − 0.008 0.250 0.2 ± 0.2 − 0.013 0.237 − 0.1 ± 1.0 0.000 0.895 Block design 51.0 ± 27.1 − 0.222 0.064 32.0 ± 1.4 − 0.148 0.333 51.0 ± 27.1 − 0.235 0.042 32.0 ± 1.4 − 0.109 0.503 37.1 ± 18.5 − 0.006 0.898 Clock drawing 11.8 ± 2.1 − 0.002 0.876 13.5 ± 0.7 − 0.014 0.459 11.8 ± 2.1 − 0.001 0.966 13.5 ± 0.7 − 0.023 0.281 12.3 ± 1.6 0.001 0.888 Values indicate: mean ± standard deviation; β represents estimate of change over time. Composite domain scores are z-scores, individual test scores are raw scores. Composite domain scores are expressed as z-scores, the individual test scores are raw scores. p values represent comparisons to non-converters. Significant comparisons are displayed in bold MAPT microtubule-associated protein tau, GRN progranulin, bvFTD behavioural variant frontotemporal dementia, nfvPPA non-fluent variant primary progressive aphasia, BNT Boston Naming Test, SAT semantic association test, TMT Trail making Test, WAIS Wechsler Adult Intelligence Scale, LDST letter digit, substitution test, WCST Wisconsin card sorting test, ToM theory of mind, VAT visual association test, RAVLT Rey Auditory Verbal Learning Test, imm immediate, del delayed Survived Bonferroni correction for multiple comparisons Higher scores and β weights indicate worse performance Journal of Neurology (2018) 265:1381–1392 1389 1 3 Table 4 ROC analyses on neuropsychological decline between 2 years before conversion and symptom onset in converters Domain and individual neuropsychological bvFTD vs. nfvPPA converters MAPT vs. GRN converters tests a b AUC 95% CI p Optimal Δ Sensitivity (%) Specificity (%) AUC 95% CI p Optimal Δ Sensitivity (%) Specificity (%) Language 0.667 0.29–1.00 0.505 – – – 0.867 0.51–1.00 0.101 – – – BNT 0.708 0.34–1.00 0.405 – – – 0.90 0.67–1.00 0.074 – – – SAT 0.625 0.24–1.00 0.617 – – – 0.833 0.54–1.00 0.136 – – – ScreeLing phonology 1.000 1.00–1.00 0.046 − 0.5 100 100 0.700 0.21–1.00 0.371 – – – Categorical fluency 0.833 0.53–1.00 0.182 – – – 1.000 1.00–1.00 0.025 − 6.5 100 100 Attention and mental processing speed 0.750 0.41–1.00 0.317 – – – 0.600 0.19–1.00 0.655 – – – TMT part A 0.542 0.00–1.00 0.868 – – – 0.50 0.05–0.95 1.000 – – – Stroop card I 0.583 0.19–0.97 0.739 – – – 0.600 0.17–1.00 0.655 – – – Stroop card II 0.583 0.12–1.00 0.739 – – – 0.667 0.22–1.00 0.456 – – – Digit Span forwards WAIS-III 0.750 0.40–1.00 0.317 – – – 0.633 0.23–1.00 0.551 – – – LDST 0.625 0.23–1.00 0.617 – – – 0.633 0.22–1.00 0.551 – – – Executive function 0.583 0.19–0.98 0.739 – – – 0.733 0.36–1.00 0.297 – – – TMT part B 0.667 0.29–1.00 0.617 – – – 0.900 0.64–1.00 0.121 – – – Stroop card III 0.833 0.51–1.00 0.182 – – – 0.600 0.15–1.00 0.655 – – – Digit span backwards WAIS-III 0.542 0.09–1.00 0.868 – – – 0.567 0.14–0.99 0.766 – – – WCST concepts 0.500 0.10–0.90 1.000 – – – 0.700 0.32–1.00 0.371 – – – Letter fluency 1.000 1.00–1.00 0.046 − 16 100 100 0.767 0.36–1.00 0.233 – – – Similarities WAIS-III 0.625 0.14–1.00 0.617 – – – 0.567 0.13–1.00 0.766 – – – Social cognition 0.500 0.00–1.00 1.000 – – – 0.667 0.13–1.00 0.456 – – – Happé ToM 0.458 0.00–1.00 0.868 – – – 0.700 0.21–1.00 0.371 – – – Happé non-Tom 0.500 0.00–1.00 1.000 – – – 0.667 0.22–1.00 0.456 – – – Ekman faces 0.667 0.15–1.00 0.505 – – – 0.567 0.07–1.00 0.766 – – – Memory 0.750 0.41–1.00 0.317 – – – 0.933 0.75–1.00 0.053 – – – VAT 0.792 0.45–1.00 0.243 – – – 0.933 0.75–1.00 0.053 – – – RAVLT immediate recall 0.667 0.15–1.00 0.505 – – – 0.600 0.09–1.00 0.655 – – – RAVLT delayed recall 0.667 0.27–1.00 0.505 – – – 0.867 0.58–1.00 0.101 – – – RAVLT recognition 0.750 0.37–1.00 0.317 – – – 0.900 0.65–1.00 0.074 – – – Visuoconstruction 0.583 0.19–0.98 0.739 – – – 0.600 0.19–1.00 0.655 – – – Block design WAIS-III 0.808 0.35–1.00 0.405 – – – 0.500 0.07–0.93 1.000 – – – Clock drawing 0.667 0.29–1.00 0.505 – – – 0.600 0.16–1.00 0.655 – – – AUC area under the curve, CI confidence interval, bvFTD behavioural variant frontotemporal dementia, nfvPPA non-fluent variant frontotemporal dementia, MAPT microtubule-associated pro- tein tau, GRN progranulin, BNT Boston Naming Test, SAT semantic association test, TMT Trail making Test, WAIS Wechsler Adult Intelligence Scale, LDST letter digit substitution test, WCST Wisconsin Card Sorting Test, ToM theory of mind, VAT visual association test, RAVLT Rey Auditory Verbal Learning Test Negative delta represents decline in test performance in nfvPPA vs. bvFTD (i.e. when a converter declines on this particular task, he/she is more likely to develop nfvPPA Negative delta represents decline in test performance in MAPT vs GRN (i.e. when a converter declines on this particular task, he/she is more likely to have a underlying MAPT mutation 1390 Journal of Neurology (2018) 265:1381–1392 each group (e.g. all nfvPPA converters have a GRN muta- are affected in both presymptomatic [8 , 11] and symptomatic tion). These findings suggest that neuropsychological assess- FTD [49, 50]. Future research could additionally investigate the ment can be used to track the different mutations and phe- use of qualitative assessment of verbal fluency (e.g. clustering, notypes from the presymptomatic to the symptomatic stage, switching between clusters), as recent research [49] points to which is advantageous considering the need for good clinical die ff rences between FTD and PPA subtypes—making this a endpoints in future disease-modifying trials. promising application of verbal fluency for a precise clinical Extending the findings from our first follow-up study [8 ], differentiation in presymptomatic and early stage FTD. we demonstrated significant decline on the RAVLT recall in Key strengths of our study constitute our longitudinal design, presymptomatic MAPT mutation carriers. The additional finding spanning a 4-year follow-up of at-risk participants from both MAPT that lower memory scores over time were also found in MAPT, and GRN families. Although our group of converters is currently and not GRN converters—suggesting a mutation-specific small, this is the first study tracking FTD mutation carriers from aetiology—corroborate this. Although memory loss has been the presymptomatic to symptomatic disease stage. Being aware of described in GRN [38, 39], this is usually a later symptom, while the caveats of small sample sizes and administering a large amount episodic memory impairment has been found as the presenting of neuropsychological tests with respect to statistical power, our and most prominent symptom in MAPT [7, 40, 41]. Interest- results warrant replication in our cohort as well as larger interna- ingly, the Genetic Frontotemporal dementia Initiative (GENFI) tional cohorts such as GENFI [10], in which with the passing of consortium revealed hippocampal atrophy in presymptomatic time more mutation carriers will approach symptom onset and/ MAPT from 15 years before estimated symptom onset [10], and or convert to clinical FTD. The dropout rate is very low, creating as this medial temporal structure is critical for episodic memory balanced datasets across the three time points. Additionally, use of processing [42] this offers a good explanation for our findings. multilevel linear modeling further handles ec ffi iently with miss - In line with earlier studies [42, 43], we did n fi d dec fi its in verbal ing data. Directions for future research entail the development of recall but not visual associative memory. Semantically loaded neuropsychological tasks more suited to administer in the presymp- tasks such as the RAVLT can be particularly more difficult than tomatic phase (robust to ceiling effects) and repeated administra- visual memory tasks like the VAT, as a result of the prominent tion (robust to practice and able to measure small changes). More semantic impairments seen early in MAPT-associated FTD [44]. extensive quantification tools of behavioural functioning are also Our results contribute to the present thinking that memory defi- needed to capture the entire clinical spectrum of (presymptomatic) cits can be an integral part of the clinical spectrum [42], and FTD, as well as assessment methods that rely less on the accuracy comprehensive memory tasks should therefore be incorporated of informant report [37]. A disadvantage of the study is the fact that in the standard diagnostic work-up. the neuropsychological assessment was part of the clinical assess- Knowing the cognitive profile of decline indicative for con- ment with which we determined conversion to the symptomatic version is important to get more insight into the timing of clini- stage. This has likely led to a circular reasoning, as we demonstrated cal changes in the earliest disease stage. We found that conver- that converters declined over time, while cognitive decline was con- sion can be predicted based on cognitive decline in the 2 years sidered a prerequisite for conversion. Ideally, the tests assessed in prior to symptom onset, but not earlier. As the cognitive decline our study should not have been used in the diagnosis of conver- was part of the diagnostic process of determining conversion, sion. However, in our multidisciplinary meeting, we followed the this is not a surprising finding. However, it does suggest a more international consensus criteria for bvFTD [3] and PPA [6], using explosive disease development with cognitive decline acceler- all available clinical information—e.g. MR imaging of the brain, ating rapidly in proximity of symptom onset, which is in line anamnestic and heteroanamnestic information, behavioural and with evidence from a large familial Alzheimer’s disease cohort neuropsychiatric questionnaires, unblinding of genetic status—so [45]. By selectively choosing tests within the domains that that symptom onset did not solely depend on the neuropsychologi- have prognostic abilities, the neuropsychological battery can cal assessment. Furthermore, as the multilevel model assumes a be shortened, which would benefit patient burden and helps linear relationship between genetic status and cognitive decline over cutting healthcare expenses. Especially fluency tasks seem to time, we could have missed non-linear effects over time. Lastly, the be promising candidates, as they were able to distinguish accu- analyses on the non-converters and controls were performed using rately between future phenotype and underlying genotype. The the original baseline and follow-up visits, regardless of, e.g. age and latter is essential for patient stratification in future clinical trials time to estimated symptom onset. It is possible that these analyses targeting specific pathologies, and ideally these interventions therefore lost some sensitivity to detect cognitive decline over time. should be applied in the presymptomatic stage [46]. Reliable However, as between-group analyses on age and estimated years to phenotypic prediction furthermore optimizes the diagnostic symptom onset in converters, non-converters, and controls did not process by shortening the current diagnostic delay [47], and show significant differences (respectively, p = 0.99 and p = 0.19), is helpful for the patient, caregiver and clinician in knowing we believe this effect is minimal. what disease presentation and course to expect. Verbal fluency Our study investigates longitudinal neuropsychological per- tests are widely used in dementia diagnosis setting [48], and formance in a large cohort of at-risk individuals from genetic 1 3 Journal of Neurology (2018) 265:1381–1392 1391 5. Adenzato M, Cavallo M, Enrici I (2010) Theory of mind ability in FTD families. We provide evidence of mutation-specific cog- the behavioural variant of frontotemporal dementia: an analysis of nitive decline when moving from the presymptomatic into the neural, cognitive, and social levels. Neuropsychologia 48:2–12 symptomatic stage, and of neuropsychological trajectories 6. Gorno-Tempini ML, Hillis AE, Weintraub S et al (2011) Classifi- predicting symptom onset. These results suggest the potential cation of primary progressive aphasia and its variants. Neurology 76(11):1006–1014 biomarker value of neuropsychological assessment in both 7. Rohrer JD, Warren JD (2011) Phenotypic signatures of genetic disease-monitoring and predicting conversion to clinical FTD. frontotemporal dementia. Curr Opin Neurol 24(6):542–549 8. Jiskoot LC, Dopper EGP, den Heijer T et al (2016) Presympto- Acknowledgements We would like to thank all the participants and matic cognitive decline in familial frontotemporal dementia: a their families for taking part in our study. This work was supported by longitudinal study. Neurology 87:384–391 Dioraphte Foundation Grant 09-02-03-00, the Association for Fronto- 9. Dopper EG, Rombouts SA, Jiskoot LC et al (2014) Structural and temporal Dementias Research Grant 2009, Alzheimer Nederland and functional brain connectivity in presymptomatic familial fronto- Memorabel ZonMw Grant 733050102 (Deltaplan Dementie). temporal dementia. Neurology 83:e19–e26 10. Rohrer JD, Nicholas JM, Cash DM et al (2015) Presymptomatic Author contributions LCJ contributed to the conception and design cognitive and neuroanatomical changes in genetic frontotemporal of the study, acquired and analysed data, and drafted the manuscript, dementia in the genetic frontotemporal dementia initiative (GENFI) figures and tables. JLP acquired data. LvA acquired and analysed data. study: a cross-sectional analysis. Lancet Neurol 14:253–262 SF acquired data. LHM acquired data and contributed to the design 11. Geschwind DH, Robidoux J, Alarcón M et al (2001) Dementia of the figures. LDK acquired data. ELvdE acquired data. EGPD con- and neurodevelopmental predisposition: cognitive dysfunction in tributed to the conception of the study and acquired data. RT contrib- presymptomatic subjects precedes dementia by decades in fron- uted to the design of the study and data analysis. RvM is the genetic totemporal dementia. Ann Neurol 50:741–746 guardian of the study. JvS contributed to the conception and design 12. Hallam BJ, Jacova C, Hsiung GYR et al (2014) Early neuropsy- of the study and is PI of the project. EvdB contributed to the design chological characteristics of progranulin mutation carriers. J Int and data interpretation of the study. JMP contributed to the design of Neuropsychol Soc 20:694–703 the study, and drafting the manuscript, figures and tables. All authors 13. Dopper EG, Chalos V, Ghariq E et al (2016) Cerebral blood flow were involved in copyediting and approval of the final draft of the in presymptomatic MAPT and GRN mutation carriers: a longitu- manuscript. dinal arterial spin labeling study. Neuroimage Clin 12:460–465 14. Kaufer DI, Cummings JL, Ketchel P et al (2000) Validation of the NPI-Q, a brief clinical form of the neuropsychatric inventory. J Compliance with ethical standards Neuropsychiatry Clin Neurosci 12(2):233–239 15. Folstein MF, Folstein SE, McHugh PR (1975) “Mini-mental Conflicts of interest LCJ, JLP, LvA, LHM, LDK, ELvdE, EGPD, RT, state”. A practical method for grading the cognitive state of RvM, JvS, EvdB, JMP report no conflicts of interest. patients for the clinician. J Psychiatr Res 12(3):189–198 16. Dubois B, Slachevsky A, Litvan I, Pillon B (2000) The Ethical standard All procedures performed in studies involving human FAB: a frontal assessment battery at bedside. Neurology participants were in accordance with the ethical standards of the insti- 55(11):1621–1626 tutional and national research committee, and with the 1964 Helsinki 17. Kaplan E, Goodglass H, Weintraub S (1978) The Boston Naming declaration and its later amendments or comparable ethical standards. Test. Lea & Febiger, Philadelphia 18. Visch-Brink E, Stronks D, Denes G (2005) Semantische Associa- Informed consent Informed consent was obtained from all individual tie test. Swets & Zeitlinger, Lisse participants included in the study. 19. Doesborgh SJ, van de Sandt-Koenderman WM, Dippel DW et al (2003) Linguistic deficits in the acute phase of stroke. J Neurol 250(8):977–982 Open Access This article is distributed under the terms of the Crea- 20. Thurstone LLT, Thurstone TG (1962) Primary mental abilities. tive Commons Attribution 4.0 International License (http://creat iveco Science Research Associates, Chicago mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- 21. Battery Army Individual Test (1994) Manual of directions and scor- tion, and reproduction in any medium, provided you give appropriate ing. War Department, Adjutant General’s office, Washington, DC credit to the original author(s) and the source, provide a link to the 22. Stroop JR (1935) Studies of interference in serial verbal reactions. Creative Commons license, and indicate if changes were made. J Exp Psychol 18:643–662 23. Wechsler D (2005) WAIS-III nederlandse bewerking, technische handleiding. Harcourt Test Publishers, Lisse 24. Jolles J, Houx PJ, van Boxtel MPJ, Ponds RWHM (1995) Maas- References tricht aging study: determinants of cognitive aging. Neuropsych Publishers, Maastricht 1. Snowden JS, Pickering-Brown SM, Mackenzie IR et al (2006) Pro- 25. Nelson HE (1976) A modie fi d card sorting test sensitive to frontal granulin gene mutations associated with frontotemporal dementia lobe defects. Cortex 12:313–324 and progressive non-fluent aphasia. Brain 129:3091–3102 26. Happe F, Brownell H, Winner E (1999) Acquired ‘theory of mind’ 2. Warren JD, Rohrer JD, Rossor MN (2013) Clinical review. Fron- impairments following stroke. Cognition 70(3):211–240 totemporal dementia. BMJ 347:f4827 27. Ekman P, Friesen WV (1976) Pictures of facial affect. Consulting 3. Rascovsky K, Hodges JR, Knopman D et al (2011) Sensitivity of Psychologists Press, Palo Alto revised diagnostic criteria for the behavioural variant of fronto- 28. Rey A (1958) L’examen clinique en psychologie. Presses Univer- temporal dementia. Brain 134:2456–2477 sitaires de France, Paris 4. Seelaar H, Rohrer JD, Pijnenburg Y et al (2011) Clinical, genetic 29. Lindeboom J, Schmand B, Tulner L et al (2002) Visual associa- and pathological heterogeneity of frontotemporal dementia: a tion test to detect early dementia of the Alzheimer type. J Neurol review. J Neurol Neurosurg Psychiatry 82:476–486 Neurosurg Psychiatry 73(2):126–133 1 3 1392 Journal of Neurology (2018) 265:1381–1392 30. Royall DR, Cordes JA, Polk M (1998) CLOX: an executive clock 41. Tolboom N, Koedam ELGE, Schott JM et al (2010) Dementia drawing task. J Neurol Neurosurg Psychiatry 64:588–594 mimicking Alzheimer’s disease owing to a tau mutation: CSF and 31. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) PET findings. Alzheimers Dis Assoc Disord 24:303–307 An inventory for measuring depression. Arch Gen Psychiatry 42. Hornberger M, Piguet O (2012) Episodic memory in frontotem- 4:561–571 poral dementia: a critical review. Brain 135:678–692 32. Youden WJ (1950) Index for rating diagnostic tests. Cancer 3:32–35 43. Spina S, Schonhaut D, Boeve BF et  al (2017) Frontotempo- 33. Barandiaran M, Estanga A, Moreno F et al (2012) Neuropsycho- ral dementia with the V337M MAPT mutation. Neurology logical features of asymptomatic c.709-1G > A progranulin muta- 88(8):758–766 tion carriers. JINS 18:1086–1090 44. Snowden JS, Adams J, Harris J et al (2015) Distinct clinical and 34. Janssen JC, Schott JM, Cipolotti L et al (2005) Mapping the onset pathological phenotypes in frontotemporal dementia associated and progression of atrophy in familial frontotemporal lobar degen- with MAPT, PGRN and C9orf72 mutations. Amyotroph Lateral eration. JNNP 76(2):162–168 Scler Frontotemp Degener 16:497–505 35. Rohrer JD, Warren JD, Barnes J et al (2008) Mapping the progres- 45. Hassenstab J, Aschenbrenner AJ, Balota DA et al (2016) Cog- sion of progranulin-associated frontotemporal lobar degeneration. nitive trajectories in DIAN: relationships with symptom onset, Nat Clin Pract Neurol 4(8):455–460 mutation types and clinical status. Alzheimers Dement 12(7):368 36. Ferman TJ, McRae CA, Arvanitakis Z et al (2003) Early and 46. Meeter LH, Donker Kaat L, Rohrer JD, van Swieten JC (2017) pre-symptomatic neuropsychological dysfunction in the PPND Imaging and fluid biomarkers in frontotemporal dementia. Nat family with the N279K tau mutation. Parkinsonism Relat Disord Rev Neurol 13(7):406–419 9(5):265–270 47. Rosness TA, Engedal K, Chemali Z (2016) Frontotemporal 37. Miller JB, Banks SJ, Léger GC, Cummings JL (2014) Randomized dementia: an updated clinician’s guide. J Geriatr Psychiatry controlled trials in frontotemporal dementia: cognitive and behav- 29(5):271–280 ioral outcomes. Transl Neurodegener 3:12 48. Pakhomov SVS, Hemmy LS (2014) A computational linguistic 38. Whitwell JL, Jack CR, Boeve BF et al (2009) Voxel-based mor- measure of clustering behavior on semantic verbal fluency task phometry patterns of atrophy in FTLD with mutations in MAPT predicts risk of future dementia in the Nun study. Cortex 55:97–106 or PGRN. Neurology 72:813–820 49. van den Berg E, Jiskoot LC, Grosveld MJH et al (2017) Qualita- 39. Rohrer JD, Ridgway GR, Modat M et al (2010) Distinct profiles tive assessment of verbal fluency performance in frontotemporal of brain atrophy in frontotemporal lobar degeneration caused by dementia. Dement Geriatr Cogn Disord 44:35–44 progranulin and tau mutations. Neuroimage 53:1070–1076 50. Laisney M, Matuszewski V, Mézenge F et al (2009) The under- 40. Smith R, Puschmann A, Schöll M et  al (2017) 18F-AV-1451 lying mechanisms of verbal fluency deficit in frontotemporal tau PET imaging correlates strongly with tau neuropathology in dementia and semantic dementia. J Neurol 256:1083 MAPT mutation carriers. Brain 139(9):2372–2379 Affiliations 1,2 1,2 1 1 1 Lize C. Jiskoot  · Jessica L. Panman  · Lauren van Asseldonk  · Sanne Franzen  · Lieke H. H. Meeter  · 1,3 1 1 4 5 Laura Donker Kaat  · Emma L. van der Ende  · Elise G. P. Dopper  · Reinier Timman  · Rick van Minkelen  · 1,6 1 1 John C. van Swieten  · Esther van den Berg  · Janne M. Papma Lize C. Jiskoot John C. van Swieten l.c.jiskoot@erasmusmc.nl j.c.vanswieten@erasmusmc.nl Jessica L. Panman Esther van den Berg j.panman@erasmusmc.nl e.vandenberg@erasmusmc.nl Lauren van Asseldonk Department of Neurology, Erasmus Medical Center laurenvanasseldonk@gmail.com Rotterdam, Room Ee2240, ’s-Gravendijkwal 230, Sanne Franzen 3015 CE Rotterdam, The Netherlands s.franzen@erasmusmc.nl Department of Radiology, Leiden University Medical Center, Lieke H. H. Meeter Leiden, The Netherlands h.meeter@erasmusmc.nl Department of Clinical Genetics, Leiden University Medical Laura Donker Kaat Center, Leiden, The Netherlands l.donkerkaat@erasmusmc.nl Department of Psychiatry, Section of Medical Psychology Emma L. van der Ende and Psychotherapy, Erasmus Medical Center, Rotterdam, e.vanderende@erasmusmc.nl The Netherlands Elise G. P. Dopper Department of Clinical Genetics, Erasmus Medical Center, e.dopper@erasmusmc.nl Rotterdam, The Netherlands Reinier Timman Department of Clinical Genetics, VU Medical Center, r.timman@erasmusmc.nl Amsterdam, The Netherlands Rick van Minkelen r.vanminkelen@erasmusmc.nl 1 3

Journal

Journal of NeurologySpringer Journals

Published: Apr 7, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off