Stressful life events, perceived stress, and 12-month course of geriatric depression: Direct effects and moderation by the 5-HTTLPR and COMT Val158Met polymorphisms

Figures & data

Table I.  Demographic data of participants by remission status at 6 months.

Demographic and clinical variables (N = 216)	Depressed-remitted (N = 84)	Depressed-nonremitted (N = 132)	N (missing values)	Test statistic	P value
Age (SD)	69.94 (7.89)	69.42 (7.01)	216 (0)	t214 = 0.50	0.6160
Sex, N (%) of women	53 (63.10)	80 (60.61)	216 (0)	χ^2 = 0.13, 1 df	0.7139
Education, years (SD)	13.80 (2.73)	13.83 (3.03)	216 (0)	t214 = − 0.07	0.9449
CIRS score	4.23 (3.50)	4.30 (2.89)	199 (17)	t197 = − 0.15	0.8838
Baseline MADRS score (SD)	27.70 (7.67)	26.70 (7.09)	216 (0)	t214 = 0.98	0.3262
MMSE score	28.13 (2.25)	28.15 (2.24)	209 (7)	t207 = − 0.07	0.9445
5-HTTLPR genotype, N (%) of subjects			216 (0)	χ^2 = 0.08, 1df	0.7831
L/L	29 (34.52)	48 (36.36)
L/S or S/S	55 (65.48)	84 (63.64)
COMT Val158Met genotype, N (%) of subjects			201 (15)	χ^2 = 3.56, 2df	0.1689
Met/Met	24 (30.38)	23 (18.85)
Met/Val	39 (49.37)	70 (57.38)
Val/Val	16 (20.25)	29 (23.77)
Average stress (SD)	6.20 (2.12)	6.76 (2.01)	213 (3)	t211 = − 1.94	0.0542
Total stress (SD)	2.14 (1.52)	2.64 (1.77)	216 (0)	t214 = − 2.14	0.0338
Negative stress (SD)	1.20 (1.31)	1.61 (1.54)	216 (0)	t214 = − 2.03	0.0441
Positive stress (SD)	0.37 (0.67)	0.48 (0.73)	216 (0)	t214 = − 1.10	0.2729

Baseline average stress data were missing for three subjects. All participants had genetic data for the 5-HTTLPR polymorphism and 201 subjects had data for the COMT Val¹⁵⁸Met polymorphism. MADRS, Montgomery–Asberg depression rating scale; MMSE, mini-mental state exam; SD, standard deviation; df, degrees of freedom. The values in italics indicate the statistical significance achieved at alpha = 0.05.

Table II.  Demographic data of participants by remission status at 12 months.

Demographic and clinical variables (N = 212)	Depressed-remitted (N = 102)	Depressed-nonremitted (N = 110)	N (missing values)	Test statistic	P value
Age (SD)	69.32 (6.93)	69.39 (7.18)	212 (0)	t210 = − 0.07	0.9447
Sex, N (%) of women	59 (57.84)	68 (61.82)	212 (0)	χ^2 = 0.35, 1 df	0.5552
Education, years (SD)	13.92 (2.69)	13.73 (3.03)	212 (0)	t210 = 0.49	0.6232
CIRS score	3.85 (2.94)	4.33 (2.92)	194 (18)	t192 = − 1.15	0.2506
Baseline MADRS score (SD)	26.56 (7.00)	27.61 (7.68)	212 (0)	t210 = − 1.04	0.3005
MMSE score	28.29 (2.03)	28.13 (2.22)	201 (11)	t199 = 0.53	0.5959
5-HTTLPR genotype, N (%) of subjects			212 (0)	χ^2 = 0.00, 1df	0.9627
L/L	34 (33.33)	37 (33.64)
L/S or S/S	68 (66.67)	73 (66.36)
COMT Val158Met genotype,  N (%) of subjects			196 (16)	χ^2 = 6.44, 2df	0.0400
Met/Met	29 (31.18)	17 (16.50)
Met/Val	44 (47.31)	64 (62.14)
Val/Val	20 (21.51)	22 (21.36)
Average stress (SD)	6.11 (2.24)	6.87 (1.95)	210 (2)	t208 = − 2.62	0.0094
Total stress (SD)	2.49 (1.64)	2.42 (1.73)	212 (0)	t210 = 0.31	0.7566
Negative stress (SD)	1.42 (1.43)	1.50 (1.53)	212 (0)	t210 = − 0.38	0.7010
Positive stress (SD)	0.51 (0.74)	0.37 (0.69)	212 (0)	t210 = 1.40	0.1643

Baseline average stress data were missing for two subjects. All participants had genetic data for the 5-HTTLPR polymorphism and 196 subjects had data for the COMT Val¹⁵⁸Met polymorphism. MADRS, Montgomery–Asberg depression rating scale; MMSE, mini-mental state exam; SD, standard deviation; df, degrees of freedom. The values in italics indicate the statistical significance achieved at alpha = 0.05.

Table III.  Logistic regression models examining baseline stress measures and their interactions with genotypes as predictors of 6- and 12-month remission status.

Baseline stress measures and interactions with genotypes as predictors of diagnostic status at 6 and 12 months	Diagnostic status at 6 months			Diagnostic status at 12 months
	N (remitted/nonremitted)	Wald χ^2 value	P value	N (remitted/nonremitted)	Wald χ^2 value	P value
Average stress	213 (82/131)	4.76	0.0291	210 (101/109)	6.33	0.0119
Average stress × 5-HTTLPR	213 (82/131)	1.17	0.2798	210 (101/109)	0.00	0.9807
Average stress × COMT Val158Met	199 (78/121)	0.87	0.6481	195 (93/102)	1.54	0.4628
Total stress	216 (84/132)	4.55	0.0329	212 (102/110)	0.09	0.7669
Total stress × 5-HTTLPR	216 (84/132)	0.00	0.9561	212 (102/110)	1.53	0.2162
Total stress × COMT Val158Met	201 (79/122)	1.62	0.4450	196 (93/103)	1.07	0.5842
Negative stress	216 (84/132)	4.09	0.0432	212 (102/110)	0.17	0.6773
Negative stress × 5-HTTLPR	216 (132/84)	1.34	0.2479	212 (102/110)	5.84	0.0156
Negative stress × COMT Val158Met	201 (79/122)	0.29	0.8654	196 (93/103)	1.40	0.4977
Positive stress	216 (84/132)	1.07	0.3010	212 (102/110)	1.76	0.1845
Positive stress × 5-HTTLPR	216 (132/84)	0.24	0.6256	212 (102/110)	1.23	0.2666
Positive stress × COMT Val158Met	201 (79/122)	1.96	0.3756	196 (93/103)	0.31	0.8577

All models examined remission status as a dichotomous-dependent variable at the 6- and 12-month assessments. Initial models examined the four stress measures as independent variables (with 1 degrees of freedom (df)), while controlling for age, education, and baseline MADRS score. These models were subsequently used to test gene by stress interactions (with 2 df), while controlling for age, education, baseline MADRS score, and genotype. The values in italics indicate the statistical significance achieved at alpha = 0.05.

Figure 1.  Probability of remission at the 12-month assessment by number of negative SLEs and 5-HTTLPR genotype. The solid line represents S allele carriers (L/S and S/S genotypes). The dotted line represents L/L homozygous subjects; (A) examines probability of remission with baseline number of negative stressors and (B) examines probability of remission with change in number of negative stressors.

Table IV.  Logistic regression models examining changes in stress scores and their interactions with genotypes as predictors of 12-month remission status.

Delta-stress measure and their interactions with genotypes as predictors of diagnostic status at 12 months	Diagnostic status at 12 months
	N (remitted/nonremitted)	Wald χ^2 value	P value
Delta-average stress	210 (101/109)	23.91	< 0.0001
Delta-average stress by 5-HTTLPR	210 (101/109)	0.12	0.7297
Delta-average stress by COMT Val158Met	195 (93/102)	0.47	0.7924
Delta-total stress	212 (102/110)	2.98	0.0842
Delta-total stress by 5-HTTLPR	212 (102/110)	3.08	0.0795
Delta-total stress by COMT Val158Met	196 (93/102)	1.35	0.5094
Delta-negative stress	212 (102/110)	14.09	0.0002
Delta-negative stress by 5-HTTLPR	212 (102/110)	7.52	0.0061
Delta-negative stress by COMT Val158Met	196 (93/102)	4.95	0.0840
Delta-positive stress	212 (102/110)	0.27	0.6002
Delta-positive stress by 5-HTTLPR	212 (102/110)	0.36	0.5475
Delta-positive stress by COMT Val158Met	196 (93/102)	0.30	0.8622

Initial models examined the 12-month remission status as a dichotomous-dependent variable with the change in stress score (with 1 df) serving as an independent variable. Models also controlled for age, education, baseline MADRS score, and baseline stress score. These models were subsequently used to test for gene by environment interactions (with 2df) while controlling for age, education, baseline MADRS score, the respective baseline and change in stress score, and genotype. The values in italics indicate the statistical significance achieved at alpha = 0.05.