Population demographic characteristics are presented in table 1. The results are in line with those obtained in our previous results on this cohort (28, 29).
ApoE4 statistics
Table 2 shows the genotyping frequency in each group (no statistical analysis was performed because some genotypes such as E2/E2 are absent from some groups), the percentage of E4 carriers and the allelic frequency (E2, E3, E4) in each diagnostic group. Statistical analyses show that there are significantly more E4 carriers (DLB vs AD vs AD/DLB vs CS: χ2=32.22, P<0.0001) in the AD and AD/DLB groups than in the CS (CS vs AD: P=0.0012; CS vs AD/DLB: P<0.0001), but also in the DLB (DLB vs AD: P=0.0073; DLB vs AD/DLB: P<0.0001) groups. There was no significant difference between AD/DLB and the AD group (AD vs AD/DLB: P=0.1274). Allelic frequency analysis confirms these results (DLB vs AD vs AD/DLB vs CS: χ2=33.43, P<0.0001), there are significantly more E4 carriers in AD patients (AD and AD/DLB) than in the other patients (CS and DLB) (CS vs AD: χ2=15.33, P=0.0005; CS vs AD/DLB: χ2=16.33, P=0.0003; DLB vs AD: χ2=11.71, P=0.0029; DLB vs AD/DLB: χ2=17.53, P=0.0002). No significant difference was found between CS and pure DLB patients both in terms of E4 carriers and allele frequency (E4 carriers P=0.1911; allele frequency χ2=4.62, P=0.10).
Socio-educational level and age of disease onset according to ApoE4 genotyping
The socio-educational level (SEL) was assessed in the cohort. No differences were observed between the different groups, all groups of patients had on average the same SEL (Table 1). In a second step, we studied the correlation between the SEL and the reported age at onset of the diseases (DLB, AD, AD/DLB). Note that for a number of patients, the number of years before the onset of symptoms is not known, we decided to exclude these patients from the analysis rather than using the age at the time of the first consultation (DLB N=76; AD N=24; AD/DLB N=17). In AD patients, there is a positive correlation between the age at onset and the SEL (r = 0.425, p = 0.039). For DLB patients, although the correlation is not significant, there is a relatively strong positive trend between age at onset and SEL (r = 0.22, p = 0.056). For patients with AD/DLB comorbidity, there is no correlation between age at onset and SEL (r = -0.0269, p = 0.918) (fig. 2).
In Figure 2, ApoE4 patients were marked with a red circle. Looking at the results of the DLB patients, we wanted to further analyse the SEL according to the presence or not of ApoE4. For this purpose we calculated the median of the SEL in the DLB group in order to separate the high SEL from the low SEL in a somewhat equitable way (median Education = 11 years, i.e. one year before bachelor). Thus, we separated patients with a SEL higher or equal to 11 and patients with a SEL strictly lower than 11 (Table 3). We observe that the group with a high SEL have significantly more ApoE4 patients than the group with a low SEL (Fisher's exact test p = 0.03). Interestingly, DLB patients with a high SEL have a strong tendency to have more ApoE4 patients than controls (p = 0.064) and have a strong tendency to have fewer than AD patients (p = 0.057).
Age at onset according to ApoE4 presence
It is interesting to determine whether the presence of ApoE4 in patients leads to an earlier onset of the disease. For each of the diseases, ApoE4 carriers do not have an earlier onset of the disease. It can be noted that DLB ApoE4 patients start their disease significantly earlier than AD ApoE4 patients (Fig. 3; p = 0.044).
Alzheimer biomarkers and ApoE4
On a part of the cohort (141 patients), we have the results of the Alzheimer biomarkers (flowchart fig.1 and Table 4). These biomarkers have been analysed previously (29, 38), but without including the ApoE4 results. Figure 4 shows that the presence or absence of ApoE4 does not seem to significantly influence the levels of the different biomarkers. Note that the results of the Aβ42/Aβ40 ratio were not presented for the AD patients because we only had the results of one ApoE4 carrier and 4 non-carriers, which was not sufficient to perform statistical analysis.
We decided to take the analysis a step further. Indeed, we had previously shown an Aβ42 decrease between the prodromal and demented stages of DLB patients (29, 38). We therefore analysed Aβ42 levels between prodromal and demented DLB patients according to the presence or absence of ApoE4. Aβ42 levels are significantly lower in demented ApoE4 than in prodromal non-ApoE4 (fig. 5; p = 0.015). The ANOVA one-way did not reveal any significant difference between the demented and prodromal ApoE4 groups (p = 0.151), however a t-test between these two groups shows a significant difference (p = 0.016) and a t-test between non-ApoE4 demented versus ApoE4 demented indicates a strong trend (p = 0.068). A larger group would probably be needed to confirm these variations. Overall, we can conclude that the Aβ42 decrease at the demented stage seems to be more prominent in ApoE4 patients.
As the ApoE4 patients in the DLB group are mainly patients with a high SEL, we wanted to check whether the Aβ42 levels were lower in this group compared to the low SEL and compared to the non-ApoE4. We found that this was not the case, we found no significant difference (ApoE4 high SEL; Non-ApoE4 high SEL; ApoE4 low SEL; Non-ApoE4 low SEL; ANOVA, p=0.461; data not shown).