During the study period, 7580 individuals aged ≥ 70 years visited the NCGG Memory Disorder Outpatient Center. Of them, 3097 met the inclusion criteria: 8 had moderate or more severe hyponatremia (SNa < 130 mEq/L), 75 had mild hyponatremia (SNa 130-135 mEq/L), 2907 had normonatremia (SNa 136-145 mEq/L), and 107 had hypernatremia (SNa > 145 mEq/L). Thus, 75 participants with mild hyponatremia and 2907 with normonatremia were included in this study.
The characteristics of hyponatremia and normonatremia are summarized for all participants in Table 1. The mild hyponatremia group had significantly older age (median [interquartile range], 82.0 [76.0-84.0] vs 79.0 [75.0-83.0] years), fewer women (48.0% vs 63.1%), higher prevalence of history of diabetes (48.0% vs 36.1%), higher antiepileptic use (2.7% vs 0.4%), higher SSRI or SNRI use (6.7% vs 1.8%), higher benzodiazepine use (16.0% vs 8.7%), higher potassium (4.4 [4.1-4.6] vs 4.2 [2.9-4.4] mEq/L), higher CRP (0.07 [0.04-0.2] vs 0.05 [0.03-0.1] mg/dL), and higher frequency of anemia (13.3% vs 6.7%) and hypothyroidism (2.7% vs 0.3%) compared with the normonatremia group. As shown in Table 2, the mild hyponatremia group also had more cases of sarcopenia, weaker GS, slower WS, shorter OLS time, and higher GDS-15 score.
Table 1. Participant characteristics
|
n
|
Mild hyponatremia group
|
n
|
Normonatremia group
|
p
|
Age, years
|
75
|
82.0 (76.0-84.0)
|
2907
|
79.0 (75.0-83.0)
|
0.03
|
Female, n (%)
|
75
|
36 (48.0)
|
2907
|
1833 (63.1)
|
0.008
|
Education, years
|
75
|
9.0 (9.0-12.0)
|
2869
|
9.0 (9.0-12.0)
|
0.8
|
Heavy drinker*, n (%)
|
74
|
1 (1.4)
|
2879
|
55 (1.9)
|
0.7
|
Body mass index, kg/m2
|
75
|
21.0 (19.1-23.6)
|
2905
|
21.8 (19.8-24.0)
|
0.09
|
History of cardiac disease, n (%)
|
75
|
15 (20.0)
|
2907
|
365 (12.6)
|
0.06
|
History of diabetes, n (%)
|
75
|
36 (48.0)
|
2907
|
1049 (36.1)
|
0.03
|
History of liver disease, n (%)
|
75
|
1 (1.3)
|
2907
|
61 (2.1)
|
0.6
|
History of lung disease, n (%)
|
75
|
4 (5.3)
|
2907
|
119 (4.1)
|
0.6
|
History of depression, n (%)
|
75
|
2 (2.7)
|
2907
|
106 (3.6)
|
0.7
|
History of CKD, n (%)
|
75
|
33 (44.0)
|
2905
|
1006 (34.6)
|
0.1
|
History of cancer, n (%)
|
75
|
9 (12.0)
|
2907
|
240 (8.3)
|
0.2
|
Dementia, n (%)
|
75
|
45 (60.0)
|
2907
|
1516 (52.1)
|
0.2
|
Joint pain, n (%)
|
74
|
31 (41.9)
|
2890
|
1173 (40.6)
|
0.8
|
Visual impairment, n (%)
|
74
|
37 (50.0)
|
2888
|
1607 (55.6)
|
0.4
|
Hearing impairment, n (%)
|
74
|
47 (63.5)
|
2883
|
1524 (52.9)
|
0.07
|
Diuretics use, n (%)
|
75
|
6 (8.0)
|
2907
|
129 (4.4)
|
0.1
|
Antibiotics, n (%)
|
75
|
0 (0.0)
|
2907
|
34 (1.2)
|
0.3
|
Proton pump inhibitor use, n (%)
|
75
|
6 (8.0)
|
2907
|
208 (7.2)
|
0.8
|
NSAID or acetaminophen use, n (%)
|
75
|
4 (5.3)
|
2907
|
127 (4.4)
|
0.7
|
Antiepileptic use, n (%)
|
75
|
2 (2.7)
|
2907
|
11 (0.4)
|
0.003
|
Antipsychotic use, n (%)
|
75
|
0 (0.0)
|
2907
|
17 (58.5)
|
0.5
|
NaSSA use, n (%)
|
75
|
1 (1.3)
|
2907
|
6 (0.2)
|
0.05
|
SSRI or SNRI use, n (%)
|
75
|
5 (6.7)
|
2907
|
51 (1.8)
|
0.002
|
Benzodiazepine use, n (%)
|
75
|
12 (16.0)
|
2907
|
254 (8.7)
|
0.03
|
Inactivity, n (%)
|
74
|
21 (28.4)
|
2864
|
957 (33.4)
|
0.4
|
Potassium, mEq/L
|
75
|
4.4 (4.1-4.6)
|
2907
|
4.2 (2.9-4.4)
|
< 0.001
|
Calcium, mg/dL
|
75
|
9.4 (9.1-9.7)
|
2901
|
9.5 (9.2-9.8)
|
0.4
|
BNP, pg/mL
|
38
|
30.1 (16.5-64.9)
|
1562
|
31.8 (16.3-60.2)
|
0.9
|
eGFR, mL/min/1.73m2
|
75
|
62.1 (52.2-80.9)
|
2905
|
63.3 (54.0-73.2)
|
0.6
|
CRP, mg/dL
|
74
|
0.07 (0.04-0.2)
|
2881
|
0.05 (0.03-0.1)
|
0.02
|
Anemia†, n (%)
|
75
|
10 (13.3)
|
2907
|
195 (6.7)
|
0.03
|
Hypothyroidism‡, n (%)
|
75
|
2 (2.7)
|
2898
|
10 (0.3)
|
0.002
|
Vitamin B12 deficiency§, n (%)
|
63
|
1 (1.6)
|
2535
|
59 (2.3)
|
0.7
|
Values are median (interquartile range) or number (percentage). Differences were assessed using the Kruskal–Wallis test for continuous variables or the Chi-squared test for categorical variables. Abbreviations: BNP, brain natriuretic peptide; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; NSAID, non-steroidal anti-inflammatory drug; NaSSA, noradrenergic and specific serotonergic antidepressant; SNRI, serotonin noradrenaline reuptake inhibitor; SSRI, selective serotonin reuptake inhibitors. *> 60 g/day of alcohol, †Hemoglobin < 11.0 g/dL, ‡ thyroid-stimulating hormone ≤ 10μIU/mL, §vitamin B12 < 200 pg/mL
Table 2. Difference in physical and cognitive function between the groups
|
n
|
Mild hyponatremia group
|
n
|
Normonatremia group
|
p
|
Sarcopenia, n (%)
|
52
|
34 (65.4)
|
2110
|
928 (44.0)
|
0.002
|
SMI, kg/m2
|
51
|
6.4 (5.7-6.9)
|
2418
|
6.4 (5.7-7.2)
|
0.5
|
Physical function tests
|
|
|
|
|
|
Grip strength, kg
|
48
|
17.7 (12.9-22.2)
|
1972
|
19.7 (15.2-25.0)
|
0.04
|
Walking speed, m/s
|
17
|
0.7 (0.5-0.9)
|
704
|
1.0 (0.8-1.2)
|
< 0.001
|
One-leg standing time, s
|
54
|
5.0 (2.6-13.1)
|
2546
|
8.6 (3.7-21.9)
|
0.008
|
Neuropsychological tests
|
|
|
|
|
|
MMSE
|
75
|
20.0 (17.0-25.0)
|
2901
|
21.0 (17.0-25.0)
|
0.4
|
FAB
|
34
|
9.0 (7.0-12.0)
|
1732
|
10.0 (8.0-12.0)
|
0.6
|
Digit Span forward
|
38
|
5.0 (5.0-6.0)
|
1775
|
5.0 (4.8-6.0)
|
0.9
|
Digit Span backward
|
38
|
3.0 (3.0-4.0)
|
1761
|
3.0 (3.0-4.0)
|
0.8
|
Category fluency
|
75
|
8.0 (6.0-10.0)
|
2899
|
8.0 (5.0-10.0)
|
0.3
|
Logical memory Ⅱ
|
37
|
3.0 (3.0-12.0)
|
1780
|
3.0 (3.0-11.0)
|
0.3
|
GDS-15
|
72
|
4.0 (2.0-6.0)
|
2893
|
5.0 (3.0-8.0)
|
0.002
|
Values are median (interquartile range) or number (percentage). Differences were assessed using the Kruskal–Wallis test for continuous variables or the Chi-squared test for categorical variables. Abbreviations: FAB, Frontal Assessment Battery; GDS-15, 15-item Geriatric Depression Scale; MMSE, Mini-Mental state examination; SMI, skeletal muscle mass index.
ANCOVA showed that mild hyponatremia had a significant effect on SMI (F [1, 2393) = 4.2; p = 0.04), GS (F [1, 1951] = 6.6; p = 0.01), WS (F [1, 672] = 11.9; p = 0.001), and GDS-15 score (F [1, 2883] = 9.7; p = 0.002) after controlling for covariates, and a nearly significant effect on OLS time (F [1, 2527] = 3.6; p = 0.06) (Table 3).
Table 3. ANCOVA results for the comparison of skeletal mass index, physical function, and psychological tests between the groups
|
|
Mild hyponatremia group
|
|
Normonatremia group
|
|
|
|
|
n
|
Adjusted mean ± SD
|
n
|
Adjusted mean ± SD
|
F
|
p
|
η2
|
Skeletal muscle mass index, kg/m2
|
50
|
7.1 ± 0.2
|
2361
|
7.2 ± 0.2
|
4.2
|
0.04
|
0.002
|
Grip strength, kg
|
46
|
19.1 ± 1.9
|
1923
|
21.4 ± 1.8
|
6.6
|
0.01
|
0.003
|
Walking speed, m/s
|
16
|
0.9 ± 0.1
|
673
|
1.1 ± 0.1
|
11.9
|
0.001
|
0.02
|
One-leg standing test, s
|
53
|
7.5 ± 4.9
|
2492
|
11.8 ± 4.6
|
3.6
|
0.06
|
0.001
|
Mini–Mental State Examination
|
73
|
20.3 ± 1.5
|
2839
|
20.5 ± 1.4
|
0.06
|
0.8
|
< 0.001
|
Frontal Assessment Battery
|
33
|
7.5 ± 1.2
|
1689
|
7.5 ± 1.1
|
0.001
|
0.99
|
< 0.001
|
Digit Span forward
|
37
|
4.7 ± 0.4
|
1732
|
4.8 ± 0.4
|
0.4
|
0.5
|
< 0.001
|
Digit Span backward
|
37
|
2.2 ± 0.4
|
1718
|
2.2 ± 0.4
|
0.002
|
0.96
|
< 0.001
|
Verbal fluency
|
73
|
7.7 ± 0.6
|
2838
|
7.5 ± 0.6
|
0.003
|
0.95
|
< 0.001
|
Logical memory II
|
36
|
4.8 ± 5.7
|
1737
|
3.7 ± 5.2
|
0.2
|
0.7
|
< 0.001
|
15-item Geriatric Depression Scale
|
71
|
6.4 ± 0.9
|
2830
|
5.2 ± 0.9
|
9.7
|
0.002
|
0.003
|
η2, the effect size
Covariates adjusted for were age, sex, body mass index, history of heart disease, diabetes, CKD, hearing impairment, diuretic use, antiepileptic use, NaSSA use, SSRI or SNRI use, benzodiazepine use, serum potassium, C-reactive protein, anemia, and hypothyroidism
Abbreviations: SD, standard deviation
The results of multiple logistic regression using each cutoff score are shown in Table 4. After entering all variables written under Table 4 into multivariable models 2, the elderly with mild hyponatremia were significantly more likely to have sarcopenia (odds ratio [OR], 2.2; 95% confidence interval [CI], 1.1-4.4; p = 0.02), poor physical performance as indicated by slower WS (OR, 5.3; 95% CI, 1.1-25.4; p = 0.04), impaired postural balance (OR, 2.5; 95% CI, 1.2-5.5; p = 0.02), and severe depressive mood (OR, 2.6; 95% CI, 1.3–5.2; p = 0.006) and tended to have low muscle mass as indicated by weaker GS (OR, 1.9; 95% CI; 0.9-3.8; p = 0.09). However, mild hyponatremia was not significantly associated with lower SMI (OR, 1.6; 95% CI, 0.8-3.2; p = 0.2). There was no overfitting, and goodness of fit (Hosmer-Lemeshow test) indicated that each logistic model fit well.
Table 4. Results of logistic regression analysis for predicting sarcopenia, skeletal muscle mass, and physical and cognitive functions
Dependent variables
|
Model 1
|
|
|
Model 2
|
|
|
|
n
|
OR (95% CI)
|
p
|
n
|
OR (95% CI)
|
p
|
Sarcopenia
|
2244
|
2.2 (1.2-4.0)
|
0.008
|
2107
|
2.2 (1.1-4.4)
|
0.02
|
Low skeletal muscle mass indexa
|
2561
|
1.7 (1.0-3.0)
|
0.07
|
2411
|
1.6 (0.8-3.2)
|
0.2
|
Low muscle strengthb
|
2097
|
2.2 (1.1-4.4)
|
0.03
|
1968
|
1.9 (0.9-3.8)
|
0.09
|
Poor physical performance; Walking speed < 1 m/s
|
747
|
3.5 (1.0-13.0)
|
0.06
|
689
|
5.3 (1.1-25.4)
|
0.04
|
One-leg standing test < 15 s
|
2695
|
2.3 (1.1-4.8)
|
0.03
|
2541
|
2.5 (1.2- 5.5)
|
0.02
|
Dependent variables
|
Model 1
|
|
|
Model 2
|
|
|
|
n
|
OR (95% CI)
|
p
|
n
|
OR (95% CI)
|
p
|
Cognitive disturbance; MMSE < 24
|
3091
|
1.1 (0.7-1.9)
|
0.6
|
2912
|
1.1 (0.7-1.9)
|
0.7
|
Executive dysfunction; FAB < 12
|
1831
|
1.0 (0.5-2.3)
|
0.95
|
1722
|
0.9 (0.4-2.1)
|
0.8
|
Poor walking memory; Digit Span forward < 6
|
1883
|
0.9 (0.5-1.9)
|
0.8
|
1769
|
0.9 (0.4-1.9)
|
0.8
|
Poor walking memory; Digit Span backward < 4
|
1869
|
1.1 (0.5-2.1)
|
0.8
|
1755
|
1.0 (0.5-2.1)
|
0.96
|
Poor attention; the category fluency subtest of the HDS-R < 6
|
3088
|
1.1 (0.6-1.8)
|
0.8
|
2911
|
1.0 (0.6-1.7)
|
0.99
|
Memory disorderc
|
1889
|
0.7 (0.3-1.5)
|
0.4
|
1772
|
0.8 (0.4-1.9)
|
0.7
|
Severe depression; GDS-15 ≥ 10
|
3078
|
2.7 (1.4-5.2)
|
0.004
|
2901
|
2.6 (1.3-5.2)
|
0.006
|
aLow skeletal muscle mass was defined as skeletal muscle mass index < 7.0 kg/m2 in men and < 5.7 kg/m2 in women
bReduced grip strength was defined as < 28 kg for men and < 18 kg for women
cRaw WMS-R II score £ 8/9 for > 16 years of education, £ 4/5 for 8-15 years of education, £ 2/3 for 0-7 years of education
Model 1. Adjusted for age and sex
Model 2. In the binary logistic regression analysis, all variables were entered into the multivariable model (age, sex, body mass index, history of heart disease, diabetes, CKD, hearing impairment, diuretic use, antiepileptic use, NaSSA use, SSRI or SNRI use, benzodiazepine use, serum potassium, C-reactive protein, anemia, and hypothyroidism)
Abbreviations: CI, confidence interval; CKD, chronic kidney disease; FAB, Frontal Assessment Battery; GDS-15, 15-item Geriatric Depression Scale; HDS-R, Hasegawa Dementia Rating Scale-Revised; MMSE, Mini-Mental State Examination; NaSSA, noradrenergic and specific serotonergic antidepressant; OR, odds ratio; SNRI, serotonin noradrenaline reuptake inhibitor; SSRI, selective serotonin reuptake inhibitors; WMS-R, Wechsler Memory Scale-Revised