Subgroups with typical courses of depressive symptoms in an elderly population during 13 years of observation: Results from the Heinz Nixdorf Recall Study

Abstract

Background:

Depressive symptoms are volatile over time but empirical studies of intra-individual variations of depressive symptoms over longer periods are sparse.

Aims:

We aim to examine fluctuation patterns of depressive symptoms and to investigate the possible influence of age, sex and socioeconomic factors on fluctuation in a population-based sample over a period of 13 years.

Methods:

We used data of 4,251 participants (45–75 years; 51.0% women at baseline) of the Heinz Nixdorf Recall Study with at least two of nine possible measurements obtained in the period between 2000 and 2017. Depressive symptoms were assessed via the Center for Epidemiologic Studies Depression Scale (CES-D) short form. Based on the individual mean values and standard deviation from all measurements, we categorized participants as G1 ‘stable low’, G2 ‘stable high’, G3 ‘stable around cutoff’ and G4 ‘large fluctuations’.

Results:

Most participants (82.3%) showed stable low depressive symptoms (G1), whereas 2.3% performed stable high values (G2), 6.9% stable around the cutoff (G3) and 8.6% large fluctuations (G4).

Conclusion:

Our longitudinal results reveal that almost 18% (G2, G3 and G4) of the participants have an increased depression score or strong fluctuations at times. According to our classification, a higher proportion of the participants show anomalies with regard to depression compared to a simple classification into depressed and nondepressed, especially if this is based on a single measurement. Thus, longitudinal measurements of depression can prevent misclassification and provide valuable information about the course of depressive symptoms for a better understanding of the changes of depression.

Keywords

Depression epidemiology prospective cohort study Center for Epidemiologic Studies Depression Scale (CES-D)

Introduction

Depression is the leading cause of disability worldwide and contributes significantly to the global burden of disease (Kessler et al., 2009; Moussavi et al., 2007). More than 300 million people of all ages worldwide suffer from depression, equivalent to 4.4% of the world population (World Health Organization, 2017). According to results reported in the ‘Global Burden of Diseases, Injuries, and Risk Factors Study 2010’, mental and substance use disorders contributed 7.4% to the total global burden of disease (Whiteford et al., 2013). Because of its high prevalence and its association with physical health and quality of life, depression is considered a main public health problem (Ohrnberger et al., 2017).

In general, more women are affected by depression than men with a prevalence of depression twice as high (Grigoriadis & Robinson, 2007; Kuehner, 2003). Most notably, depression is highly prevalent in older people and is associated with considerable functional impairment, morbidity and mortality (Lenze et al., 2005) as well as negative impacts on their quality of life (Almeida, 2014; Forlani et al., 2014). Luppa et al. (2012) estimated in a meta-analysis a pooled prevalence of 17.1% for depressive disorders and 7.2% for major depressions based on community-based studies with an elderly population worldwide.

Depression is typically characterized by specific depressive symptoms and an episodic course; that is, these symptoms are temporary and often subside even without therapeutic measures. The main symptoms of depression encompass depressive mood, loss of appetite, psychomotor retardation, suicidal ideation, sleep disorders and feelings of guilt, worthlessness, hopelessness or helplessness (Rapaport et al., 2002). The course of depressive episodes varies widely in frequency, from a singular week-long episode to a lifelong disorder with recurring severe depressive episodes (Richards, 2011). Especially, over a long period and with moderate or severe intensity, depression can have serious consequences, including the increased burden of physical illness and risk of suicide or even the inability to work, learn or perform tasks of everyday life (Fiske et al., 2009).

Elevated depressive symptoms serve as criteria for the diagnosis of major depression or a depressive episode (Rodriguez et al., 2012) and are often measured as an indicator of depression in population-based epidemiological studies using paper and pencil questionnaires. Even persons who do not fulfill the criteria for major depression but tend to have an elevated depression score may already suffer from impaired physical health and quality of life.

Since most studies are assessing depressive symptoms only at one time point, the development of depressive symptoms as well as the heterogeneity between individuals is seldom considered. For a better understanding of the course and reasons for the certainty of depression even at the subthreshold level, it is essential to analyze at the course of symptoms from a long-term perspective, too (Colman & Ataullahjan, 2010). Musliner et al. concluded in their systematic review of heterogeneity in long-term trajectories of depressive symptoms that trajectories in the general population vary in terms of two characteristics: severity (low, medium and high) and stability (stable, increasing and decreasing) (Musliner et al., 2016). Most people follow a trajectory characterized by no depressive symptoms, but a remarkable proportion of 2.1%−7.2% experience persistently high depressive symptoms over long periods.

Of the 25 studies included in the review by Musliner et al., eight targeted on older adults with depressive symptoms measured more than four times (Andreescu et al., 2008; Byers et al., 2012; Hsu, 2012; Hybels et al., 2016; Kuchibhatla et al., 2012; Kuo et al., 2011; Liang et al., 2011; Montagnier et al., 2014). Concerning the long-run course of depressive symptoms, these studies considered, besides stability, exclusively a steady increase or decrease in changes of symptoms. However, there is a lack of studies that not only examine the stable or continuous development of depressive symptoms but also take into account fluctuations in the course of the depression. Therefore, it is important to investigate whether subgroups with different courses of depressive symptoms are existing and to describe specific patterns. Besides, it would also be worth knowing to what extent distinct subpopulations differ in their socioeconomic or disease-related characteristics. This research can help to develop a better understanding of the distribution of the burden of depression and the underlying causes and development of diseases concerning depressive symptoms.

Accordingly, the main objective of this study is to describe typical courses of depressive symptoms over a long time period and to investigate fluctuation patterns of depressive symptoms with regard to age, sex, socioeconomic factors and diseases in a large population-based cohort study of older adults over 13 years.

Methods

Study population

The Heinz Nixdorf Recall (HNR) Study is an ongoing prospective population-based cohort study. The study rationale and design have been described in detail elsewhere (Schmermund et al., 2002; Stang et al., 2005). Briefly, for baseline examination, starting in 2000, 4,814 women and men (49.8% men) aged 45 to 75 years were recruited from mandatory citizen registries of three large cities (Bochum, Essen and Mülheim/Ruhr, Germany). Two study components can be distinguished (Figure 1): (1) The participants were invited to the study center for examination three times in total, with an interval of 5 years (t0: 2000–2003, t5: 2006–2008, t10: 2011–2015). Data assessment at the study center included standardized interviews, clinical examination, comprehensive laboratory tests and self-administered questionnaires. (2) In addition, a yearly postal follow-up questionnaire was provided between these examinations. Figure 1 shows the corresponding timeline of the HNR Study, the sample size for each time point and CES-D assessment for depressive symptoms. Accordingly, the CES-D was assessed three times at the study center within 10 years. On the other survey dates, the CES-D was completed by mailed questionnaires since the seventh follow-up. The average follow-up time was 9.1 years.

Figure 1.

Timeline of the HNR study and Center for Epidemiologic Studies Depression Scale (CES-D) assessment.

The HNR study was approved by the local ethics committees, and all participants gave written informed consent before participation.

Assessment of depressive symptoms

While the gold standard for the diagnosis of depression is a structured clinical interview, self-reporting tools play an important role in screening for depression and are widely used in research studies (Dunstan et al., 2017). We assessed depressive symptoms using the Center for Epidemiologic Studies Depression Scale (CES-D) (Radloff, 1977). The CES-D is a validated tool often used in the general population. It is asking for the presence and frequency of symptoms and emotional states in the week before the interview including depressive mood, feelings of guilt or worthlessness, sleep disorders and self-doubt (Clark et al., 1981; Fountoulakis et al., 2007). The CES-D is considered an indicator of depression and is highly correlated with a clinical diagnosis of depression (Lewinsohn et al., 1997).

In the HNR Study, a short version of the CES-D with 15 questions was used. Answers are given on a 4-point Likert-type scale ranging from ‘less than 1 day’ (0 point) to ‘5–7 days’ (3 points). We calculated a sum score ranged from 0 to 45 points with a higher score indicating more and/or more frequent depressive symptoms. Positively formulated items were coded backward and an average value was calculated over all 15 items. For up to three missing answers, the item value was replaced by the mean value of the answered questions. In the HNR Study, a cutoff point of ⩾17 was assumed to indicate depression (Hautzinger & Bailer, 1993; Icks et al., 2013). The participants completed the CES-D self-administered questionnaire at the study center at each of the three examinations. The CES-D was further assessed from the seventh postal follow-up onward (t7, t8, t9, t11, t12, t13; Figure 1).

In a standardized computer-assisted face-to-face interview (CAPI), performed by trained personnel, participants were asked additionally if they (1) ever had depression and if so and (2) if the exact diagnosis is known and the depression is currently being treated or was previously. To assess medication intake, we used the brown bag method, asking the participants to bring all the medication they had been taking in the last 7 days with them to the study center. Antidepressants were categorized using the Anatomical Therapeutic Chemical (ATC) classification index. CAPI and medication intake were only assessed at the three visits at the study center (Figure 1).

In our study, three different ways of defining depression are used: (1) depressive symptoms as a continuous variable based on the CES-D, which means an increase in CES-D score indicates an increase in the depressive state; (2) depressive symptoms as a dichotomous variable with a CES-D cutoff value of ⩾17 at which depression occurs; and (3) a comprehensive definition with a CES-D value above the cutoff, the intake of antidepressants and/or a current medical care for depression.

Socioeconomic status, lifestyle and health-related factors and morbidity

Socioeconomic status, lifestyle and morbidity were assessed as part CAPI (see above). Education was classified according to the International Standard Classification of Education (ISCED-97) as total years of formal education, combining school and vocational training and was categorized into four groups (⩽10, 11–13, 14–17, ⩾18 years). Income was measured as the monthly household equivalent income, which was calculated by dividing the total household net income by a weighting factor for each household member. We also recorded if participants were cohabiting with a partner or not. Physical activity is defined as the regular practice of any type of sports activities during the past 4 weeks. Smoking status was grouped into three categories: (1) current, (2) former and (3) never smoker and pack-years were additionally calculated. Important life events (e.g. death or serious illness of a close person, momentous professional change, separation, relocation) in the last 6 months were asked, too. Anthropometry was measured to calculate body mass index (BMI) (weight in kilograms/(height in meters)²). Diabetes in HNR was defined according to Moebus et al. (2009). Coronary heart disease and stroke were assessed. Information on myocardial infarction, heart failure, stroke, emphysema, asthma, cancer, rheumatism, slipped disk or migraine at baseline was used to build a categorical variable indicating the number of comorbidities (0, 1 or ⩾2).

Statistical analysis

A total of 4,251 HNR participants with at least two measurements on CES-D were included. The intra-individual mean value of the CES-D score from all available measurements was calculated for each participant to generate a CES-D mean score over time. Their corresponding standard deviation was calculated as well to show the variation of the CES-D score over time for each participant. Based on these two values, the populations were divided into four subgroups (stable low, stable high, stable around cutoff for depressive symptoms, large fluctuations). See Table 1 for more detailed information regarding the classification into the different subgroups.

Table 1.

Criteria for classification into the four subgroups.

Four subgroups	Stable low(n = 3,498, 82.3%)	Stable high(n = 98, 2.3%)	Stable around cutoff(n = 291, 6.9%)	Large fluctuations(n = 364, 8.6%)
Individual standard deviation	<6.8	<6.8	<6.8	⩾6.8
Overall CES-D mean	<13.6	>20.4	⩾13.6; ⩽20.4	x

CES-D: Center for Epidemiologic Studies Depression Scale.

Descriptive statistics were used to outline the subgroups according to age, sex, socioeconomic status, lifestyle and health-related factors. Continuous parameters are represented as mean ± standard deviation (SD) or quartile distribution Q2 (25th percentile, Q1; 50th percentile, Q2; 75th percentile, Q3). Categorical variables are given as number and percentage (%). Pearson’s correlation was calculated for the different CES-D assessments.

All analyses were conducted using SAS statistical software, version 9.4.

Results

The baseline characteristics of the analyzed population (n = 4,251) are shown in Table 1 Supplement, for the entire population as well as for women (n = 2,167) and men (n = 2,084) separately. The participants (51.4% of women) were aged between 45 and 75 years (M = 59.3 ± 7.7). At baseline, the mean CES-D score was 7.9 (correspond to definition i) and 7.9% of the participants were above the cutoff (⩾17) (definition ii). A diagnosis for depression was reported by 8.6% of the participants and 5.0% were taking an antidepressant. Women showed higher CES-D mean score (women = 8.8, men = 7.0) and reported more frequently a diagnosis of depression (women = 11.6%, men = 5.4%) as well as intake of antidepressant medication (women = 8.9%, men = 2.7%).

Figure 2 depicts the longitudinal CES-D scores as a continuous variable (definition i) stratified by sex. Women revealed higher CES-D scores than men over the whole study period. Overall, a decrease in the CES-D score is observable over time in both men and women. The highest share of participants above the cutoff is observed at t7 with 8.8% dropping to 6.3% at t13 (definition ii) (Table 2). Stratified by sex, more women had a CES-D above the cutoff compared to men. Using the comprehensive definition of depression (definition iii: encompassing intake of antidepressants, current medical care for depression and/or being above the cutoff) reveals a similar course over time like definition ii. It is noticeable that, according to this definition, 19.7% of women in t10 suffer from depression.

Figure 2.

CES-D distribution by follow-up year and sex.

Table 2.

Prevalence of all investigated depression variables by time points.

	CES-D ⩾ 17			Intake of antidepressants			Medical care for depression currently			CES-D ⩾ 17 and/or antidepressants and/or medical care
	All	Men	Women	All	Men	Women	All	Men	Women	All	Men	Women
t0 (n = 4,152)	327 (7.9%)	98 (4.8%)	229 (10.9%)	199 (5.0%)	53 (2.7%)	146 (7.2%)	84 (2.0%)	27 (1.4%)	57 (2.7%)	530 (12.8%)	162 (7.9%)	368 (17.4%)
t5 (n = 4,170)	364 (8.7%)	121 (5.9%)	243 (11.5%)	260 (6.3%)	83 (4.1%)	177 (8.5%)	89 (2.2%)	26 (1.3%)	63 (3.0%)	614 (14.7%)	201 (9.8%)	413 (19.4%)
t7 (n = 2,118)	186 (8.8%)	79 (7.5%)	107 (10.0%)
t8 (n = 3,478)	303 (8.7%)	121 (7.1%)	182 (10.3%)
t9 (n = 3,223)	272 (8.4%)	99 (6.3%)	173 (10.5%)
t10 (n = 2,948)	202 (6.9%)	58 (4.0%)	144 (9.6%)	226 (7.3%)	67 (4.5%)	159 (10.1%)	117 (3.8%)	39 (2.6%)	78 (4.9%)	419 (14.2%)	124 (8.6%)	295 (19.7%)
t11 (n = 2,508)	195 (7.8%)	71 (5.8%)	124 (9.7%)
t12 (n = 2,091)	158 (7.6%)	69 (6.7%)	89 (8.4%)
t13 (n = 1,138)	72 (6.3%)	26 (4.7%)	46 (7.8%)

CES-D: Center for Epidemiologic Studies Depression Scale.

Of the participants with elevated depressive symptoms at baseline, 44.1% and 36.9% also had elevated depressive symptoms at the 5-year follow-up and 10-year follow-up as well (Table 3). Of the participants who are above the cutoff at baseline, 26.8% are still or again affected by increased symptoms at t13. It is shown that participants with elevated CES-D scores above the cutoff tended to have higher depressive symptoms during the following year, for example, from t7 to t8, 67.4%, or t10 to t11, 56.3%.

Table 3.

Proportion of participants with CES-D ⩾ 17 by time points.

	t0 ⩾ 17	t5 ⩾ 17	t7 ⩾ 17	t8 ⩾ 17	t9 ⩾ 17	t10 ⩾ 17	t11 ⩾ 17	t12 ⩾ 17	t13 ⩾ 17
t0 ⩾ 17 (n = 449)	x	44.1%	32.5%	35.3%	35.87%	36.9%	30.9%	34.0%	26.8%
t5 ⩾ 17 (n = 431)		x	43.8%	51.0%	46.2%	45.4%	46.2%	44.1%	36.7%
t7 ⩾ 17 (n = 215)			x	67.4%	55.0%	44.8%	49.5%	44.2%	35.3%
t8 ⩾ 17 (n = 346)				x	56.7%	49.8%	51.3%	48.0%	39.7%
t9 ⩾ 17 (n = 313)					x	50.9%	56.7%	53.4%	37.5%
t10 ⩾ 17 (n = 244						x	56.3%	59.6%	48.5%
t11 ⩾ 17 (n = 218)							x	56.2%	48.8%
t12 ⩾ 17 (n = 179)								x	50.7%

Pearson correlation coefficients were calculated to identify any associations between the CES-D score for the different assessments. The correlation coefficients of CES-D as reported at different time points during the study are depicted in the supplemental material (Table 2 Supplement). The coefficients mostly ranged between 0.46 and 0.73, indicating a moderate to high correlation. Measurements that are only 1 year apart correlate more strongly with each other, for example, from t7 to t8 or t11 to t12, compared with the lowest correlation from baseline to t13. The correlation coefficients for the four subgroups show similar correlations for ‘stable low’ as for the entire cohort (e.g. t11 and t12, r = 0.64) (Table 3 Supplement). The ‘stable high’ subgroup also indicates a positive correlation for measurements of the following years (e.g. t7 and t8, r = 0.49), but no uniform trend for larger time intervals (Table 4 Supplement). There is also no correlation trend seen for the groups ‘stable around cutoff’ and ‘large fluctuations’ (Tables 5 and 6 Supplement). For the last three subgroups, it should be noted that the number of measurements is partly very small (<200).

Definitions of subgroups

We defined four subgroups with distinct courses of change in CES-D levels (Figure 3). The largest group of participants (n = 3,498, 82.3%) shows a stable low CES-D score over time (G1). Interestingly, the mean CES-D score amounted to 6.5 at baseline and kept stable on this level for the CES-D score over time (6.0 ± 3.3). The subgroup ‘stable high’ (G2), which shows a CES-D score well above the cutoff for depression, comprised 98 participants (2.3%) with a mean baseline CES-D score of 23.6 and a constant high CES-D score during follow-up with an average mean of 25.0. The subgroup ‘stable around cutoff’ included 291 participants (6.9%) fluctuating during the study course around the cut-point ⩾17. This group has a mean baseline CES-D score of 15.5 and shows an average CES-D score of 16.1 during follow-up. These first groups have in common that the CES-D performs stable over time. The fourth subgroup ‘large fluctuations’ (n = 364, 8.6%) had a mean baseline CES-D score of 11.9 and an increasing CES-D score of 14.5 over time, however, always accompanied by a substantially higher standard deviation of 5.4 as compared to the more stable groups. Figure 4 depicts real examples (randomly selected from these with eight or nine CES-D measurements) of the four individual CES-D courses for each subgroup.

Figure 3.

Development of the CES-D means by depression subgroups and 95% confidence intervals.

Figure 4.

Examples of four individual CES-D courses for each subgroup.

The characterization of the subgroups is presented in Table 4. The ‘stable low’ subgroup has a slightly higher male proportion (51.5%), in comparison with the other groups with a lower proportion of men (G2, 32.2%; G3, 36.4%; G4, 39.3%). The stable high group is mainly characterized by women (67.3%). Men in this group are relatively young (M_age = 56.5 years). The average follow-up time in years is the longest for stable low participants (men: 10.1, women: 10.2). In G1 (men: 6.0, women: 7.0) and G4 (men: 10.7, women: 12.7), women have a higher CES-D score than men at baseline. A difference in the CES-D score over time between women and men is almost non-existent in G3 (men: 16.0, women: 16.2) and G4 (men: 14.5, women: 14.5). The greatest proportion of highly educated is in G1 (G1 men: 15.6%, women: 8.4%, G2 men: 6.3%, women: 1.5%, G3 men: 8.5%, women: 6.4%, G4 men: 8.4% women: 9.0%). The health-related risk factors show that stable high group is less physically active (men: 40.6%, women: 41.0%) compared to the other groups (G1 men: 56.2%, women: 58.6%, G3 men: 48.1%, women: 47.6%, G4 men: 48.3%, women: 46.6%). Women in the stable high group have a higher average BMI (G2 women: 30.2, G1 women: 27.3, G3 women: 28.6, G4 women: 27.9). On the other hand, men hardly show this difference in BMI between the groups.

Table 4.

Sex-stratified baseline characteristics by depression subgroups.

Characteristics	Stable low(n = 3,498, 82.3%)		Stable high(n = 98, 2.3%)		Stable around cutoff(n = 291, 6.9%)		Large fluctuations(n = 364, 8.6%)
Characteristics	Men(n = 1,803, 51.5%)	Women(n = 1,695, 48.5%)	Men(n = 32, 32.6%)	Women(n = 66, 67.3%)	Men(n = 106, 36.4%)	Women(n = 185, 63.6%)	Men(n = 143, 39.3%)	Women(n = 221, 60.7%)
Follow-up time mean (SD) (missings n = 0)	10.1 (2.2)	10.2 (2.3)	8.0 (3.1)	8.8 (2.6)	9.3 (2.5)	9.7 (2.2)	9.3 (2.3)	9.6 (2.3)
CES-D Baseline mean (SD) (missings n = 0)	6.0 (4.0)	7.0 (4.6)	24.2 (5.7)	23.2 (6.0)	15.8 (4.7)	15.8 (4.7)	10.7 (8.1)	12.7 (9.5)
CES-D over time mean (SD) (missings n = 0)	5.6 (3.1)	6.4 (3.3)	25.3 (3.6)	24.8 (3.9)	16.0 (1.8)	16.2 (1.9)	14.5 (5.4)	14.5 (5.4)
Age mean (SD) (missings n = 0)	59.3 (7.5)	59.2 (7.5)	56.5 (7.5)	60.0 (8.3)	61.1 (8.2)	61.1 (8.2)	59.1 (8.5)	58.0 (8.1)
Age group
45–54 years	565 (31.3%)	542 (32.0%)	16 (50.0%)	20 (30.3%)	33 (31.3%)	49 (26.5%)	52 (36.4%)	87 (39.4%)
55–64 years	763 (42.3%)	708 (41.8%)	10 (31.3%)	24 (36.4%)	34 (32.1%)	62 (33.5%)	48 (33.6%)	82 (37.1%)
65–75 years	475 (26.3%)	445 (26.2%)	6 (18.8%)	22 (33.3%)	39 (36.8%)	74 (40.0%)	43 (30.1%)	52 (23.5%)
Partnership^a (missings n = 3)	1,681 (93.3%)	1,320 (78.0%)	25 (78.1%)	42 (66.6%)	95 (90.5%)	132 (71.4%)	163 (94.4%)	163 (73.8%)
Education^b (missings n = 6)
⩽10 years	61 (3.4%)	248 (14.6%)	4 (12.5%)	17 (25.7%)	12 (11.3%)	40 (21.6%)	11 (7.7%)	37 (16.8%)
11–13 years	848 (47.1%)	1,094 (64.6%)	18 (56.3%)	46 (70.0%)	45 (42.5%)	117 (63.2%)	64 (44.8%)	146 (66.1%)
14–17 years	609 (33.9%)	208 (12.3%)	8 (25.0%)	2 (3.0%)	40 (37.7%)	16 (8.7%)	56 (39.2%)	18 (8.1%)
⩾18 years	281 (15.6%)	143 (8.4%)	2 (6.3%)	1 (1.5%)	9 (8.5%)	12 (6.4%)	12 (8.4%)	20 (9.0%)
Net income in Euro in P50 (P25; P75) (missings n = 262)	1,619 (1,175; 2,208)	1,406 (1,104; 1,875	1,414 (898; 1,875)	937 (766; 1,406)	1,406 (1,036; 1,036)	1,278 (937; 1,619)	1,489 (1,104; 1,875)	1,313 (937; 1,796)
Physical activity yes (missings n = 0)	1,013 (56.2%)	994 (58.6%)	13 (40.6%)	27 (41.0%)	51 (48.1%)	88 (47.6%)	69 (48.3%)	103 (46.6%)
Smoking (missings n = 2)
Non-smokers	528 (29.3%)	950 (56.0%)	8 (25.0%)	37 (56.0%)	38 (35.9%)	119 (64.3%)	29 (20.3%)	106 (48.0%)
Former smokers	855 (47.4%)	410 (24.2%)	14 (43.8%)	9 (13.7%)	43 (40.6%)	36 (19.5%)	63 (44.0%)	61 (27.7%)
Current smokers	419 (23.3%)	334 (19.7%)	10 (31.2%)	20 (30.3%)	25 (23.6%)	30 (16.2%)	51 (35.7%)	54 (24.4%)
Pack-years mean (SD) (missings n = 112)	31.1 (30.6)	20.4 (18.3)	41.2 (43.4)	25.7 (20.8)	33.5 (27.3)	19.7 (16.6)	34.8 (30.2)	23.3 (18.6)
BMI^c mean (SD) (missings n = 17)	28.1 (3.9)	27.3 (4.9)	28.1 (3.7)	30.2 (6.3)	28.3 (4.0)	28.6 (4.3)	28.6 (4.3)	27.9 (5.7)
BMI^c classes (missings n = 17)
<25 kg/m²	362 (20.2%)	614 (36.4%)	7 (21.9%)	12 (18.2%)	22 (20.8%)	56 (30.4%)	29 (20.3%)	77 (34.8%)
25–30 kg/m²	999 (55.6%)	659 (39.1%)	16 (50.0%)	22 (33.3%)	51 (48.1%)	62 (33.7%)	74 (51.8%)	74 (33.5%)
>30 kg/m²	435 (24.2%)	413 (24.5%)	9 (28.1%)	32 (48.5%)	33 (31.1%)	66 (35.9%)	40 (28.0%)	70 (31.7%)
Physician diagnosed depression (missings n = 742)	67 (3.8%)	128 (7.7%)	10 (32.3%)	26 (40.0%)	16 (15.7%)	48 (26.5%)	17 (12.1%)	46 (21.2%)
Intake of antidepressants medication (missings n = 266)	31 (1.9%)	75 (4.7%)	7 (23.3%)	18 (27.7%)	8 (8.1%)	21 (12.0%)	7 (5.1%)	32 (15.3%)
Cardiovascular disease (missings n = 0)	159 (8.8%)	39 (2.3%)	4 (15.5%)	2 (3.0%)	13 (12.3%)	4 (2.2%)	16 (11.2%)	8 (3.6%)
Stroke (missings n = 12)	46 (2.6%)	25 (1.5%)	2 (4.7%)	1 (3.3%)	5 (4.7%)	6 (3.3%)	3 (2.1%)	7 (3.2%)
Diabetes (missings n = 0)	288 (16.0%)	139 (8.2%)	3 (9.4%)	10 (15.1%)	14 (13.2%)	25 (13.5%)	28 (20.6%)	18 (8.1%)
Comorbidities (missings n = 0)
None	175 (39.7%)	654 (38.6%)	10 (31.3%)	14 (21.2%)	37 (34.9%)	53 (28.7%)	48 (33.6%)	84 (38.0%)
1	619 (34.3%)	549 (32.4%)	7 (21.9%)	20 (30.3%)	28 (26.4%)	51 (27.6%)	47 (43.9%)	53 (24.0%)
2 or more	469 (26.1%)	492 (29.0%)	15 (46.9%)	32 (48.5%)	41 (38.7%)	81 (43.8%)	48 (33.6%)	84 (38.0%)
Important events last 6 months (missings n = 17)	368 (20.5%)	415 (24.6%)	10 (33.3%)	28 (42.4%)	19 (17.9%)	61 (33.0%)	41 (29.3%)	76 (34.9%)

CES-D: Center for Epidemiologic Studies Depression Scale; BMI: body mass index.

Married or living together with a partner in the same household. ^bcombining school and vocational training. ^cBody mass index: Weight (kg)/height (m)².

Discussion

This study aimed to describe typical subgroups with a specific course of depressive symptoms over a long time period. Using up to nine CES-D measurements over 13 years we could identify four groups with specific courses: (1) ‘stable low’, (2) ‘stable high’, (3) ‘stable around cutoff’ and (4) ‘large fluctuations’. The largest group with 82% has a stable low CES-D score, followed by the groups ‘large fluctuations’ (8.6%), ‘stable around cutoff’ (6.9%) and ‘stable high’ (2.3%).

The baseline prevalence of depression in our study (7.9%) is at the lower limit of the prevalence range of 8%–16% reported by Blazer (2003). According to definition (iii), 12.3% of participants at baseline suffered from depression, which corresponds to the middle of the range by Blazer. Furthermore, our study replicates the well-known higher proportion of women with depression.

In their review on long-term trajectories of depressive symptoms, Musliner et al. report two main findings: severity (low, medium and high) and stability (stable, increasing and decreasing) of depressive symptoms (Musliner et al., 2016). We also classified our sample according to stability but subdivided only the groups displaying stable symptoms to the severity of depressive symptoms. We identified subgroups with high and low depressive symptoms. The group of ‘stable around cutoff’ can be understood as medium severity according to Musliner’s classification.

A comparison with other studies reveals that the classification into ‘low’ and ‘high’ is the most common. Kuchibhatla et al. (2012) identified four different classes: (1) stable low depressive symptoms (76.6%); (2) initially low, increasing to the subsyndromal level (10.0%); (3) stable high depressive symptoms (5.4%); and (4) high depressive symptomatology improving over 6 years before reverting somewhat (8.0%). With ‘low stable’ and ‘stable high’, this roughly corresponds to our distribution among the groups. Byers et al. (2012) also described four groups: (1) minimal depressive symptoms (27.8%), (2) persistently low depressive symptoms (54.0%), (3) increasing depressive symptoms (14.8%) and (4) persistently high depressive symptoms (3.4%). Summarizing the first two groups by Byers, this matches the ‘stable low’ group of our study with 81.8%. These two studies were conducted with older adults in the United States, with Byers examining only women. In our study, the group of ‘stable around cutoff’ can best be translated as moderate depressive symptoms, such as those found by Liang et al. (2011), but their group of ‘moderate and stable depressive symptoms’ is much larger with 29.2%. Kou et al. (2011) presented quite different results, identifying four classes: (1) persistent low, 41.8%; (2) persistent mild, 46.8%; (3) late peak, 4.2%; and (4) high-chronic, 7.2%. In comparison, they found more participants with high and much less with low depressive symptoms. This difference can be explained by the fact that the study population is Taiwanese and which makes a comparison with Europeans or Americans more difficult.

None of these studies that investigated the course of depressive symptoms explicitly considered fluctuations in both directions. If the studies identified groups taking into account fluctuations, then they refer only to the increase or decrease of depressive symptoms (Byers et al., 2012; Hsu, 2012).

Strengths and limitations

The major strength of the HNR Study is the fact that it is a large representative sample of middle and older age followed annually. The availability of multiple measurements of depressive symptoms allowed investigating patterns of depressive symptoms over 13 years. The number of participants lost to follow-up is very small. Another strength of the HNR Study is that it is carefully conducted and includes a large number of health-related control variables. Although no clinical diagnosis is used to determine depression, the CES-D is a widely utilized and well-established instrument. For the description of the distribution of depression, additional information on the intake of antidepressants and current medical care was used. The large sample provided an acceptable number of participants in each group.

The study also has a few limitations. Although the loss to follow-up is not high, it cannot be excluded that drop out is unbalanced due to depression or morbidity. In the entire study population, the CES-D score decreased during 13 years of observation. This can be interpreted either as a higher drop off in the number of depressed participants or as a decrease in depressive symptoms with increasing age. The fact that participants with a stable low score had longer follow-up times speaks in favor of the first explanation. Along with their higher social status, this group had lower morbidity. Looking at the CES-D scores of the subgroups over time, a decreasing trend is not observable. Luppa et al. (2012) report the growth of depressive symptoms with increasing age in their systematic review on age- and gender-specific prevalence of depression in the latest life. This increase can also be explained by the fact that physical disability, cognitive impairment or declining socioeconomic status also rise with advancing age. These are all factors strongly associated with depression.

In addition, not all participants have the same number of measurements due to loss to follow-up. A further limitation is that the interval between measurements is not always the same. The CES-D was not measured between t0 and t5 nor in t6, so no fluctuations in depressive symptoms that might have occurred in the meantime were detected. The questionnaire was always filled out at a different location during the study period (study center or at home) but was always self-administrated. Moreover, we could not identify systematic differences in the distributions of the scores between the different assessments. The definition of the subgroups based on the distribution of the CES-D is only one possibility of categorization and neither medication intake or known clinical diagnoses nor history of depression have been considered. In addition, the group with large fluctuations is in itself very heterogeneous.

Future research

After descriptively characterizing the study population and categorizing different temporal courses of depressive symptoms, further research can continue to work with these findings. In further research, it is necessary to investigate potential differences in the results if depressive symptoms are measured only once or if more measurements are available at different times. The categorization into the four subgroups not only allows determining whether a person is classified as depressed or not at a certain point in time but it also considers if a person is chronically depressed or fluctuates greatly in his or her emotional state. This classification can be taken into account for future work.

Finally, as the number of categories was determined by the distribution of depressive symptoms in our study, the results need to be carefully interpreted and confirmed by future work or by other classification methods.

Conclusion

This is one of the few studies that has so far attempted to describe and categorize the different courses of depressive symptoms in older people over time in the general population. Our study described four characteristic subgroups of different long-term courses of depressive symptoms characterized by socioeconomic status, morbidity and lifestyle factors. This classification can prevent misclassification and improve the accurate detection of depression, especially of a minor depression, as the participants are categorized precisely. Our study shows a higher proportion of anomalies in depression than simply distinguishing between depressed and nondepressed people.

This article presents a reasonable definition of the depression subgroups as well as a first overview and description of the characteristics of these subgroups. In future studies, these subgroups can be used for the analyses of an association between depression and different kinds of health- and environment-related exposures and outcomes. Overall, the different subgroups of depressive symptoms provide new insights into the heterogeneity underlying the average changes in depressive symptoms focusing on the fluctuations of symptoms in older adults.

Supplemental Material

Supplement_Material – Supplemental material for Subgroups with typical courses of depressive symptoms in an elderly population during 13 years of observation: Results from the Heinz Nixdorf Recall Study

Supplemental material, Supplement_Material for Subgroups with typical courses of depressive symptoms in an elderly population during 13 years of observation: Results from the Heinz Nixdorf Recall Study by Miriam Engel, Karl-Heinz Jöckel, Nico Dragano, Miriam Engels and Susanne Moebus in International Journal of Social Psychiatry

Footnotes

Acknowledgements

The authors thank all study participants, the personnel of the HNR study center, the investigative group and all former employees of the HNR study.

Author contributions

M.E. and S.M. contributed to conceptualization; M.E., S.M. and K-H.J. to methodology; M.E. to formal analysis and investigation and writing – original draft preparation; M.E., K-H.J., N.D., M.Es. and S.M. to writing – review and editing; K-H.J., N.D. and S.M. to funding acquisition; S.M. to supervision.

Conflict of interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethical standards

The HNR study was approved by the local ethics committees of the Medical Faculty of the University of Duisburg-Essen, and all participants gave written, informed consent prior to participation.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: The authors thank the Heinz Nixdorf Foundation (Chairman: Martin Nixdorf; Past Chairman: Dr jur. Gerhard Schmidt (†)) for their generous support of this study. The study was also supported by the German Ministry of Education and Research (BMBF project: 01EG0401, 01GI0856, 01GI0860, 01GS0820_WB2C, 01ER1001D, 01GI0205); the German Research Council (DFG) (DFG project: SI 236/8-1, SI 236/9-1, SI 236/10-1, JO 170/8-1, ER155/6-1, KN885/3-1, HO 3314/2-1, ER155/6-2, HO 3314/2-3, EI 969/2-3, INST 58219/32-1, PE 2309/2-1); the Ministry of Innovation, Science, Research and Technology (MIWFT-NRW); North Rhine-Westphalia and Deutsche Gesetzliche Unfallversicherung (DGUV project: FF-FP0295).

ORCID iD

Miriam Engel

Supplemental material

Supplemental material for this article is available online.

References

Almeida

O. P.

(2014). Prevention of depression in older age. Maturitas, 79(2), 136–141. https://doi.org/10.1016/j.maturitas.2014.03.005

Andreescu

Chang

C.-C. H.

Mulsant

B. H.

Ganguli

(2008). Twelve-year depressive symptom trajectories and their predictors in a community sample of older adults. International Psychogeriatrics, 20(2), 221–236. https://doi.org/10.1017/S1041610207006667

Blazer

D. G.

(2003). Depression in late life: Review and commentary. Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 58(3), M249–M265. https://doi.org/10.1093/gerona/58.3.M249

Byers

A. L.

Vittinghoff

Lui

L. Y.

Hoang

Blazer

D. G.

Covinsky

K. E.

…Yaffe

(2012). Twenty-year depressive trajectories among older women. Archives of General Psychiatry, 69(10), 1073–1079. https://doi.org/10.1001/archgenpsychiatry.2012.43

Clark

V. A.

Aneshensel

C. S.

Frerichs

R. R.

Morgan

T. M.

(1981). Analysis of effects of sex and age in response to items on the CES-D scale. Psychiatry Research, 5(2), 171–181.

Colman

Ataullahjan

(2010). Life course perspectives on the epidemiology of depression. The Canadian Journal of Psychiatry, 55(10), 622–632. https://doi.org/10.1177/070674371005501002

Dunstan

D. A.

Scott

Todd

A. K.

(2017). Screening for anxiety and depression: Reassessing the utility of the Zung scales. BMC Psychiatry, 17(1), Article 329. https://doi.org/10.1186/s12888-017-1489-6

Fiske

Wetherell

J. L.

Gatz

(2009). Depression in older adults. Annual Review of Clinical Psychology, 5(1), 363–389. https://doi.org/10.1146/annurev.clinpsy.032408.153621

Forlani

Morri

Ferrari

Dalmonte

Menchetti

De Ronchi

Atti

A. R.

(2014). Prevalence and gender differences in late-life depression: A population-based study. American Journal of Geriatric Psychiatry, 22(4), 370–380. https://doi.org/10.1016/j.jagp.2012.08.015

10.

Fountoulakis

K. N.

Bech

Panagiotidis

Siamouli

Kantartzis

Papadopoulou

… St. Kaprinis

(2007). Comparison of depressive indices: Reliability, validity, relationship to anxiety and personality and the role of age and life events. Journal of Affective Disorders, 97(1), 187–195. https://doi.org/10.1016/j.jad.2006.06.015

11.

Grigoriadis

Robinson

G. E.

(2007). Gender issues in depression. Annals of Clinical Psychiatry, 19(4), 247–255. https://doi.org/10.1080/10401230701653294

12.

Hautzinger

Bailer

(1993). Allgemeine Depressions Skala. Manual [General Depression Scale. A Manual]. Beltz-Test-GmbH.

13.

Hsu

H. C.

(2012). Group-based trajectories of depressive symptoms and the predictors in the older population. International Journal of Geriatric Psychiatry, 27(8), 854–862. https://doi.org/10.1002/gps.2796

14.

Hybels

C. F.

Bennett

J. M.

Landerman

L. R.

Liang

Plassman

B. L.

(2016). Trajectories of depressive symptoms and oral health outcomes in a community sample of older adults. International Journal of Geriatric Psychiatry, 31(1), 83–91. https://doi.org/10.1002/gps.4292

15.

Icks

Albers

Haastert

Pechlivanis

Pundt

Slomiany

…Moebus

(2013). Risk for high depressive symptoms in diagnosed and previously undetected diabetes: 5-year follow-up results of the Heinz Nixdorf Recall Study. PLOS ONE, 8(2), Article e56300. https://doi.org/10.1371/journal.pone.0056300

16.

Kessler

R. C.

Aguilar-Gaxiola

Alonso

Chatterji

Lee

Ormel

…Wang

P. S.

(2009). The global burden of mental disorders: An update from the WHO World Mental Health (WMH) Surveys. Epidemiology and Psychiatric Sciences, 18(1), 23–33.

17.

Kuchibhatla

M. N.

Fillenbaum

G. G.

Hybels

C. F.

Blazer

D. G.

(2012). Trajectory classes of depressive symptoms in a community sample of older adults. Acta Psychiatrica Scandinavica, 125(6), 492–501. https://doi.org/10.1111/j.1600-0447.2011.01801.x

18.

Kuehner

(2003). Gender differences in unipolar depression: An update of epidemiological findings and possible explanations. Acta Psychiatrica Scandinavica, 108(3), 163–174. https://doi.org/10.1034/j.1600-0447.2003.00204.x

19.

Kuo

S. Y.

Lin

K. M.

Chen

C. Y.

Chuang

Y. L.

Chen

W. J.

(2011). Depression trajectories and obesity among the elderly in Taiwan. Psychological Medicine, 41(8), 1665–1676. https://doi.org/10.1017/S0033291710002473

20.

Lenze

E. J.

Schulz

Martire

L. M.

Zdaniuk

Glass

Kop

W. J.

…Reynolds

C. F.

III . (2005). The course of functional decline in older people with persistently elevated depressive symptoms: Longitudinal findings from the Cardiovascular Health Study. Journal of the American Geriatrics Society, 53(4), 569–575. https://doi.org/10.1111/j.1532-5415.2005.53202.x

21.

Lewinsohn

P. M.

Seeley

J. R.

Roberts

R. E.

Allen

N. B.

(1997). Center for Epidemiologic Studies Depression Scale (CES-D) as a screening instrument for depression among community-residing older adults. Psychology and Aging, 12(2), 277–287.

22.

Liang

Quinones

A. R.

Bennett

J. M.

(2011). Multiple trajectories of depressive symptoms in middle and late life: Racial/ethnic variations. Psychology and Aging, 26(4), 761–777. https://doi.org/10.1037/a0023945

23.

Luppa

Sikorski

Luck

Ehreke

Konnopka

Wiese

…Riedel-Heller

S. G.

(2012). Age- and gender-specific prevalence of depression in latest-life – Systematic review and meta-analysis. Journal of Affective Disorders, 136(3), 212–221. https://doi.org/10.1016/j.jad.2010.11.033

24.

Moebus

Stang

Mohlenkamp

Dragano

Schmermund

Slomiany

…Jöckel

K. H.

(2009). Association of impaired fasting glucose and coronary artery calcification as a marker of subclinical atherosclerosis in a population-based cohort – Results of the Heinz Nixdorf Recall Study. Diabetologia, 52(1), 81–89. https://doi.org/10.1007/s00125-008-1173-y

25.

Montagnier

Dartigues

J.-F.

Rouillon

Pérès

Falissard

Onen

(2014). Ageing and trajectories of depressive symptoms in community-dwelling men and women. International Journal of Geriatric Psychiatry, 29(7), 720–729. https://doi.org/https://doi.org/10.1002/gps.4054

26.

Moussavi

Chatterji

Verdes

Tandon

Patel

Ustun

(2007). Depression, chronic diseases, and decrements in health: Results from the World Health Surveys. The Lancet, 370(9590), 851–858. https://doi.org/10.1016/s0140-6736(07)61415-9

27.

Musliner

K. L.

Munk-Olsen

Eaton

W. W.

Zandi

P. P.

(2016). Heterogeneity in long-term trajectories of depressive symptoms: Patterns, predictors and outcomes. Journal of Affective Disorders, 192, 199–211. https://doi.org/10.1016/j.jad.2015.12.030

28.

Ohrnberger

Fichera

Sutton

(2017). The dynamics of physical and mental health in the older population. Journal of the Economics of Ageing, 9, 52–62. https://doi.org/10.1016/j.jeoa.2016.07.002

29.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1(3), 385–401. https://doi.org/10.1177/014662167700100306

30.

Rapaport

M. H.

Judd

L. L.

Schettler

P. J.

Yonkers

K. A.

Thase

M. E.

Kupfer

D. J.

…Rush

A. J.

(2002). A descriptive analysis of minor depression. American Journal of Psychiatry, 159(4), 637–643. https://doi.org/10.1176/appi.ajp.159.4.637

31.

Richards

(2011). Prevalence and clinical course of depression: A review. Clinical Psychology Review, 31(7), 1117–1125. https://doi.org/10.1016/j.cpr.2011.07.004

32.

Rodriguez

M. R.

Nuevo

Chatterji

Ayuso-Mateos

J. L.

(2012). Definitions and factors associated with subthreshold depressive conditions: A systematic review. BMC Psychiatry, 12, Article 181. https://doi.org/10.1186/1471-244x-12-181

33.

Schmermund

Mohlenkamp

Stang

Gronemeyer

Seibel

Hirche

…Erbel

(2002). Assessment of clinically silent atherosclerotic disease and established and novel risk factors for predicting myocardial infarction and cardiac death in healthy middle-aged subjects: Rationale and design of the Heinz Nixdorf RECALL Study. Risk Factors, Evaluation of Coronary Calcium and Lifestyle. American Heart Journal, 144(2), 212–218.

34.

Stang

Moebus

Dragano

Beck

E. M.

Mohlenkamp

Schmermund

…Jöckel

K. H.

(2005). Baseline recruitment and analyses of nonresponse of the Heinz Nixdorf Recall Study: Identifiability of phone numbers as the major determinant of response. European Journal of Epidemiology, 20(6), 489–496.

35.

Whiteford

H. A.

Degenhardt

Rehm

Baxter

A. J.

Ferrari

A. J.

Erskine

H. E.

…Vos

(2013). Global burden of disease attributable to mental and substance use disorders: Findings from the Global Burden of Disease Study 2010. The Lancet, 382(9904), 1575–1586. https://doi.org/10.1016/s0140-6736(13)61611-6

36.

World Health Organization. (2017). Depression and other common mental disorders: Global health estimates.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB