Evaluation of chronic stress indicators in geriatric and oncologic caregivers: a cross-sectional study

Abstract

Caregiving induces chronic stress with physical and psychological impact on informal caregivers health. Therefore, subjective and objective indicators are needed for the early diagnosis of pathologic stress to prevent the risk of developing stress-related diseases in caregivers. Our aim was to assess the self-perceived stress, that is, how and how much the stressor affects the individual, through endocrine, metabolic, and immunologic biomarkers levels in geriatric and oncologic informal caregivers. Informal caregivers and non-caregivers were invited to participate in a cross-sectional study at the Clinic Hospital of Barcelona. Demographic and lifestyle characteristics, self-perceived stress (Perceived Stress Scale, State-Trait Anxiety Inventory and Stress Visual Analogue Scale), and biomarkers (copeptin, glucose, glycated hemoglobin, low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), cholesterol, triglycerides, α-amylase, cortisol, tumor necrosis factor (TNF-α), and Interleukins (IL-6 and IL-10)) were evaluated. Descriptive and non-parametric statistical data analysis were performed. Fifty-six subjects (19 non-caregivers, 17 geriatric caregivers, and 20 oncologic caregivers) participated. Median age (IQR) was 57 years (47–66) and 71.46% were women. Self-perceived stress was higher in oncologic caregivers than geriatric caregivers in all psychometric test analyzed (Wilcoxon Rank Sum test, p value < .05). Glucose concentrations and glycated hemoglobin levels differed statistically among groups (Kruskal–Wallis test (K–W tests), p value < .05), even though the median levels were not clinically relevant. Levels of other biomarkers did not differ significantly (K–W tests, p value > .05). These findings suggest that perceived stress is not homogeneous in the caregivers community and thus these two groups could be differentiated. These results provide the baseline information to initiate social actions addressed to each group of caregivers to increase their wellbeing.

Keywords:

• Oncologic caregivers
• geriatric caregivers
• perceived stress
• chronic stress
• hair cortisol
• stress biomarkers

Introduction

Informal caregiver community represents a great value for any National Health System, but they have many neglected health needs (Oliva-Moreno, Peña-Longobardo, & Vilaplana-Prieto, 2015). Informal caregiver is defined as the nonprofessional and non-remunerated person, usually member of the family setting, in charge of the care of a person with limited autonomy due to physical or cognitive impairment.

Daily caregiving leads to a burden escalation with impact on the caregiver health (Schulz & Sherwood, 2008) from the emotional reaction and the objective burden related to the daily tasks and funding of the care expenses (Oliva-Moreno, Trapero-Bertran, Peña-Longobardo, & del Pozo-Rubio, 2017; Pinquart & Sorensen, 2003). In Spain, the economic impact value was estimated in 1.7–4.9% of Spanish Gross Domestic Product (GDP) of 2008 (Oliva-Moreno et al., 2015, 2017). Common physical and psychological problems are sleeping disturbance, fatigue, pain, depression, anxiety, and stress, collectively referred to as caregiver stress syndrome (Carmeli, 2014; Pinquart & Sorensen, 2003; Stenberg, Ruland, & Miaskowski, 2010; Vitaliano, Zhang, & Scanlan, 2003). Altogether contributes to the caregiver’s health decline and higher risk for developing cardiovascular disorders and diabetes mellitus (Roepke et al., 2012; Schulz & Beach, 1999; Siddiqui, Madhu, Sharma, & Desai, 2015; Von Känel et al., 2011).

Global life expectancy has increased by five years between 2000–2015. For instance; in Spain, healthy life expectancy is around 70 years and life expectancy exceeds 82 years (WHO, 2016). In the interval, the risk of developing one or more chronic diseases, such as certain cancers or dementia, increases in the population rising the demands for care (Global Alliance for Chronic Diseases, GACD annual report for 2015–16). Demand for caregivers of the elderly in Spain by 2060 was estimated at 2.5 million (Geerts, Willemé, & Mot, 2012). However, this estimation would increase if other non-geriatric users were considered.

Chronic stress is considered pathology (García, 2010), nevertheless, there is not unanimous consensus to define the stress condition or objective measure of signs that allow the early diagnosis of patients (Aguiló et al., 2015). Selye (1950) proposed that anything that causes stress endangers life, unless it is met by adequate adaptive responses and conversely, anything that endangers life causes stress and adaptive responses. Another definition presented by Lazarus (1993) is that stress state arises when a specific event threatens the endocrine, physiological, or psychological homeostatic stability of the individual.

Psychometric tests are common screening tools to assess the perceived stress. However, their multidimensional nature and the lack of standardization make difficult to identify the most appropriate test and to measure the stress level objectively (Whalen & Buchholz, 2009). Some studies show that chronic stressors may contribute to metabolic, endocrine and immunologic deregulation (Hjortskov, Garde, Ørbaek, & Hansen, 2004; Siddiqui et al., 2015; Vitaliano et al., 2003). These biomarkers may apply as potential candidates of stress indicators.

Our research is oriented toward the search for measurable variables that systematically, objectively, and quantitatively act as indicators of stress in individuals subjected to different stressors (Aguiló et al., 2015). The present study aims to assess the self-perceived stress and endocrine, metabolic, and immunologic biomarkers levels in geriatric and oncologic informal caregivers.

Methods

Participants

Caregivers of elder chronic patients (geriatric caregivers) and caregivers of oncologic patients (oncologic caregivers) attending the Integrated Care Unit and the Oncology Service of the Hospital Clínic of Barcelona (Barcelona, Spain), respectively, were invited to participate in a cross-sectional study to investigate signs of stress by using psychometric tests and analytical measures of biomarkers. Non-caregiver volunteers unrelated with the patients were invited to participate in this study as a comparison group. The sample was drawn from a larger, ongoing study of chronic stress biomarkers and although the sample size was modest, it was comparable to other studies examining biomarkers in stressed individuals (Siegenthaler, Walti, Urwyler, Schuetz, & Christ-Crain, 2014; Urwyler, Schuetz, Sailer, & Christ-Crain, 2015). Nevertheless, the results of our assay should be interpreted with caution.

The study was performed in the Hospital Clínic of Barcelona between January 2015 and June 2016. The study was approved by the institutional Committee of Ethics and in accordance with the ethical principles of the World Medical Association and the Helsinki Declaration. All participants were informed about the purpose and characteristics of the study and their rights to withdraw of the study at any time. All participants signed an informed consent.

Participants were consecutively included in the study, always in accordance to the inclusion and exclusion criteria for each group.

Inclusion criteria for caregivers were: to be older than 18 years of age, to be in charge of either an oncological patient with progressive incurable cancer under oncospecific or palliative treatment, or a geriatric patient (>65 years of age) diagnosed with a chronic disease (chronic obstructive pulmonary disease, heart failure, or neurodegenerative disease, and Barthel Index below 60); to be unpaid; to live with the patient in the same household; to be emotionally involved; and to have undertaken the caregiving activity for, at least, the last six months.

Inclusion criteria for non-caregivers were: to be older than 18 years of age; not to be subject to any recognized stressor; and not to be in charge of any adult family member requiring care.

Exclusion criteria for all participants were: failure to meet any of the inclusion criteria; habitual intake of psychotropic or corticoids; to be diagnosed with alteration of the hypothalamic pituitary adrenal axis; or to have a body mass index (BMI) greater than 35.

Endpoints

Socio-demographic and lifestyle variables analyzed were: age, gender, BMI, habits (smoking status, coffee and alcohol intake, drugs consume, and sport practice), health condition (diagnosed chronic diseases, medication, and treatment) and other variables (cohabitation condition and job status). Detailed characteristics of these variables are shown in Table 1.

Table 1. Demographic and lifestyle characteristics.

	Non-caregivers	Geriatric caregivers	Oncologic caregivers	Total
Participants, N (%)	19 (33.93)	17 (30.36)	20 (35.71)	56
Age. Median, years (IQR)	53.00 (45.00–65.00)	67.50 (57.50–71.00)	55.50 (47.00-62.00)	57.00 (47.00–66.00)	Kruskal-Wallis Test, (H2 = 9.75, p = .0076)*
Gender, N (%)
Men	6 (31.58)	3 (17.65)	7 (35.00)	16 (28.57)	Fisher’s Exact Test, p = .4961
Women	13 (68.42)	14 (82.35)	13 (65.00)	40 (71.43)
BMI Median (IQR)	25.61 (20.83–28.39)	26.67 (25.24–29.09)	25.40 (22.24–28.04)	26.23 (22.04–28.76)	Kruskal-Wallis Test, (H2= 1.69; p = .4290)
Health condition, N	19	16^a	20	55
Chronic disease, N (%)
Non	15 (78.95)	7 (43.75)	8 (40.00)	30 (54.55)	Fisher’s Exact Test, p = .0312*
Yes	4 (21.05)	9 (56.25)	12 (60.00)	25 (45.45)
Medication, N (%)
Non	11 (57.89)	6 (37.50)	9 (45.00)	26 (47.29)	Fisher’s Exact Test, p = .4634
Yes	8 (42.11)	10 (62.50)	11 (55.00)	29 (52.73)
Psychotropic, N (%)
Non	17 (89.47)	16 (100.00)	14 (70.00)	47 (85.45)	Fisher’s Exact Test, p = .0382*
Occasionally ^b	2 (10.53)	0 (0.00)	6 (30.00)	8 (14.55)
Habits, N	19	16^a	20	55
Smoker. N (%)
Non	18 (94.74)	14 (87.50)	12 (60.00)	44 (80.00)	Fisher’s Exact Test, p = .0239*
Yes	1 (5.26)	2 (12.50)	8 (40.00)	11 (20.00)
Coffee, N (%)
None	1 (5.26)	6 (37.50)	1 (5.00)	8 (14.55)	Fisher’s Exact Test, p = .0091*
≥1 cup	18 (94.74)	10 (62.50)	19 (95.00)	47 (85.45)
Alcohol daily consume, N (%)
None	2 (10.53)	12 (75.00)	9 (45.00)	23 (41.82)	Fisher’s Exact Test, p = .0004*
≥1 glass	17 (89.47)	4 (25.00)	11 (55.00)	32 (58.18)
Sport practice, N (%)
Non	3 (15.79)	14 (87.50)	15 (75.00)	32 (58.18)	Fisher’s Exact Test, p < .0001*
Yes^c	16 (84.21)	2 (12.50)	5 (25.00)	23 (41.82)
Others, N	19	16^a	20	55
Married, N (%)
Non	8 (42.11)	5 (31.25)	5 (25.00)	18 (32.73)	Fisher’s Exact Test, p = .5153
Yes	11 (57.89)	11 (68.75)	15 (75.00)	37 (67.27)
Cohabitation, N (%)
Family	6 (31.58)	3 (18.75)	8 (40.00)	17 (30.91)	Fisher’s Exact Test, p = .4890
Partner	8 (42.11)	11 (68.75)	9 (45.00)	28 (50.91)
Alone	5 (26.32)	2 (12.50)	3 (15.00)	10 (18.18)
Job status, N (%)
Inactive	4 (21.05)	9 (56.25)	10 (50.00)	23 (41.82)	Fisher’s Exact Test p = .0671
Active	15 (78.95)	7 (43.75)	10 (50.00)	32 (58.18)

^a Data from one geriatric caregiver participant were not provided and therefore, discarded for these analyses.

^b Less than twice/week and any during the last 24 h.

^c Either occasionally or currently.

* Group comparison statistically significant (p < .05).

Primary endpoints were the perceived stress scores measured with the Spanish version of the following psychometric tests: Perceived Stress Scale (PSS) (Cohen, Kamarck, & Mermelstein, 1983), State-Trait Anxiety Inventory (STAI) (Spielberg, Gorsuch, & Lushene, 2008), and Stress Visual Analogue Scale (VASS) (Monk, 1989).

PSS provides an overall measure of perceived stress in a situation appraised as stressful considering the feelings and thoughts that have occurred during the last month. The test consists of a 10-question survey with Likert scaling responses (never, almost never, sometimes, fairly often, and very often). STAI measures a subjective emotional state. The test consists of 40 questions with Likert scaling responses. Twenty of these 40 questions measure the trait anxiety (almost never, sometimes, often, and almost always) and the remaining 20 measure the state anxiety (not at all, somewhat, moderately so, and very much so). VASS consists of a simple visual 100-point scale (0, not at all; 100, absolutely stressed), in which participants rate their present perceived stress. Completing all the tests takes less than 20 minutes depending on the patient. No special education or training is required to administer them.

Secondary endpoints were endocrine, metabolic, and inflammatory markers: plasma copeptin (pmol/L), copeptin osmolarity (mOsm), glucose (mg/dL), glycated hemoglobin (HbA1c; IFCC and NGSP/DCCT units) (%), low-density lipoprotein cholesterol (LDL; mg/dL), high-density lipoprotein cholesterol (HDL; mg/dL), cholesterol (mg/dL), triglycerides (mg/dL), salivary α-amylase (U/mL), salivary cortisol (μg/dL), hair-cortisol (pg/mg), tumor necrosis factor (TNF-α; pg/mL), and Interleukins (IL-6) and IL-10) (pg/mL).

Participants abstained from eating, drinking stimulants (such as tea, coffee or alcohol), brushing their teeth, or smoking during the two hour period before arriving at the hospital. The experimental procedure was performed between 11:00 am and 14:00 pm, in order to control hormonal variations attributable to the circadian rhythm or eating habits. Each session lasted approximately for 45 min.

Plasma samples were obtained from 10 mL of blood collected in serum tubes for the biochemical and immunological analysis and ethylenediaminetetraacetic acid (EDTA) tubes for the HbA1c and copeptin analysis. All analyses were carried out by the Centre de Diagnòstic Biomèdic (CDB) of the Hospital Clínic of Barcelona, except for the immunology factors that were analyzed at the Centro de Investigaciones Biológicas (CIB) in Madrid (Spain).

Salivary samples were collected using salivettes (Sarstedt, Granollers, Spain) at two time points separated by 15 min from each other. Subjects removed the swab from the salivettes, placed it in the mouth chewing for about 60 seconds to stimulate salivation, returned the swab with the absorbed saliva to the salivette and finally they were stored at −20 °C until the analysis day. Then, salivettes were thawed at room temperature and centrifuged at 4 °C for 10 min at 1000 g. Salivary α-amylase was analyzed using Salimetrics^® Salivary Alpha-Amylase Assay (Salimetrics^®, Carlsbad, CA), per duplicate using the same kit batch and by following the instructions of the manufacturer. The amount of α-amylase activity is directly proportional to the 405 nm absorbance. Salivary cortisol was analyzed using Salimetrics^® Cortisol Enzyme Immunoassay (Salimetrics^®, Carlsbad, CA), per duplicate using the same corresponding kit batch and by following the instructions of the manufacturer. The limit of quantification of cortisol assay is 0.012 µg/dL. The means of both samples separated by 15 min were then used for the statistical analysis.

Hair-specific characteristics were assessed including hair color and treatment (none, highlights, and dying). Hair samples were stored at room temperature in propylene tubes protected from light. The accumulated cortisol was extracted according to a previously published procedure with minor modifications (Scorrano et al., 2015). Briefly, the 3 cm of hair closest to the scalp were obtained and washed two times for 1 min in 10 ml of isopropanol and completely air-dried at room temperature. For the extraction of cortisol, 40 mg of hair was mixed with 1.6 mL of methanol overnight by continuous rotation (20 rpm/min). After centrifugation (10,000 g, 10 min, 4 °C), the methanol was recovered in a new clean glass tube and the procedure was repeated. Total recovered methanol was pooled and dried up under a nitrogen stream. Cortisol extracts were reconstituted in 200 µL of sodium phosphate buffer and analyzed by using Salimetrics^® Cortisol Enzyme Immunoassay. Salivary and hair samples were assessed in the Endocrinology and Radioimmunoanalysis Department (SER) at the Universitat Autònoma of Barcelona (Spain).

Data analysis

Descriptive analysis of data was performed with SAS^® software version 9.4 (SAS Inc., Cary, NC). For categorical variables, percentages of events were calculated and the comparisons of groups were analyzed by Fisher’s exact test. For quantitative variables, the mean and standard deviation (SD) or median and the inter-quartile range (IQR) were calculated depending on the data. In the frame of cross-sectional exploratory study, comparisons among groups were analyzed by parametric or non-parametric analysis according to the sample size and criteria of normality (Shapiro–Wilk test). Comparisons among the three groups were analyzed by either ANOVA or Kruskal–Wallis test. Comparisons between two groups were analyzed by either Student t-test or Wilcoxon Rank Sum test (Mann–Whitney U-test). Finally, the median and the IQR were used due to the limited sample, although it might be less powerful. All tests were two-tailed, and p values <.05 were considered statistically significant.

Results

A total of 56 subjects participated in the study (19 non-caregivers, 17 geriatric caregivers, and 20 oncologic caregivers). The median age (IQR) of participants was 57 years (47–66), 71.4% (40/56) were women and the median (IQR) of BMI of all participants was 26.23 (22.04–28.76). Non-significant differences were found among groups in terms of gender and BMI (p value >.05). Geriatric caregivers resulted to be slightly older (median age, 67.5 years; IQR: 57.50–71.00) compared to oncologic caregivers and non-caregivers (median age, 55.50 years; IQR: 47.00–62.00, and median age, 53.00 years; IQR: 45.00–65.00 respectively, p value <.05) (Table 1).

The three groups were not homogeneous in health condition, habits, and job status (Table 1). The highest percentage of participants with chronic diseases was found in the oncologic caregivers (60.0%, 12/20 total oncologic caregivers) (p value = .0312). 14.5% (8/55) of total participants took occasionally psychotropic; the highest percentage was found in the oncologic caregivers group (30.0%, 6/20 total oncologic caregivers) (p value = .0382). A 20.0% (11/55) of total participants were smokers; the highest percentage of smokers was found in the oncologic caregivers group (40.0%, 8/20 total oncologic caregivers) (p value = .0239). An 85.5% (47/55) of total participants drank coffee regularly; the lowest percentage of coffee drinkers was found in the geriatric caregivers group (62.5%, 10/16 total geriatric caregivers) (p value = .0091). 58.2% (32/55) of total participants did not practice sport, mainly caregivers (80.55%, p value <.0001). Lastly, 58.2% (32/55) of total participants were daily job active, mainly non-caregivers (78%, p value = .067), but differences were not statistically significant (p value = .0671). Non-significant differences were found among the three groups in taking other prescribed medication, married status, nor household cohabitation (p value > .05; Table 1).

Perceived stress scores were analyzed in 55 out of 56 total participants because the psychometric tests responses obtained from one geriatric caregiver participant was missing.

We found evidence that the level of perceived stress in non-caregivers, geriatric caregivers, and oncologic caregivers differs significantly overall (p value <.05), being higher in caregivers. Oncologic caregivers presented the highest score distribution in all psychometric tests: PSS test (median: 24.50; IQR: 23.00–27.50), STAI-s test (median: 47.50; IQR: 35.00–53.00), STAI-t test (median: 28.50; IQR: 20.00–40.00), and in VASS rate (median: 80.00; IQR: 65.00–82.50). Non-caregivers showed the lowest score distribution in all psychometric tests, except for PSS whose score distribution was similar to that found in the geriatric caregivers group (Table 2).

Table 2. Primary endpoints. Psychometric tests.

	Non-caregivers median (IQR) N = 19	Geriatric caregivers median (IQR) N^a = 16	Oncologic caregivers Median (IQR) N = 20
PSS	22.00 (18.00–23.00)	20.50 (17.50–23.00)	24.50 (23.00–27.50)	K–W test (H2 = 13.10, p = .0014*) WRS test (S = 209.00; p = .0045*)
STAI-s	11.00 (6.00–16.00)	26.00 (22.00–43.50)	47.50 (35.00–53.00)	K–W test (H2 = 28.53; p < .0001*) WRS test (S = 228.50; p = .0398*)
STAI-t	13.00 (9.00–21.00)	17.00 (12.50–41.00)	28.50 (20.00–40.00)	K–W test (H2 = 8.89; p = .0117*) WRS test (S = 259.00; p = .2528)
VASS	40.00 (23.00–50.00)	50.00 (50.00–77.50)	80.00 (65.00–82.50)	K–W test (H2 = 22.90; p < .0001*) WRS test (S = 228.00; p = .0287*
Summary score^b	99.00 (67.00–116.00)	149.00 (107.00–227.50)	218.50 (190.00–249.00)	K–W test (H2 = 2.7123; p < .0001*) WRS test (S = 235.00; p = .0621)

^a Psychometric tests answers from one geriatric caregiver participant were missing and therefore discarded for these analyses.

^b Summary score obtained from the sum of psychometric test scores of each participant for a general overall analysis.

* Group comparison statistically significant (p < .05).

K–W test: Kruskal-Wallis test; WRS: Wilcoxon Rank Sum test of geriatric caregivers vs. oncologic caregivers comparison group.

Furthermore, we found that the oncologic caregivers group had median (IQR) score of PSS test, STAI-s test, and median (IQR) VASS rate significantly higher than the geriatric caregivers group (p value <.05) (Table 2). However, median (IQR) score of STAI-t was not significantly different between both, oncologic and geriatric caregivers group (S = 259.00; p value = .2528).

Regarding biomarkers, the levels of glucose concentrations and glycated hemoglobin in non-caregivers, geriatric caregivers, and oncologic caregivers differed slightly but were statistically significant overall (glucose: Kruskal-Wallis tests, p value = .0494; glycated hemoglobin: Kruskal–Wallis tests, p value = .0031). However, the median level was not clinically relevant in any group (normal values: glucose <125 mg/dL and HbA1c < 6.5%) (Table 3). We observed abnormal concentrations of glucose typed as pre-diabetes in some oncologic caregivers.

Table 3. Secondary endpoints. Endocrine, metabolic, and inflammatory biomarkers.

	N^a	Non-caregivers median (IQR)	N^a	Geriatric caregivers median (IQR)	N^a	Oncologic caregivers median (IQR)
Copeptin (pmol/L)	12	5.45 (2.25–11.50)	16	6.25 (4.35–13.85)	19	6.20 (3.50–8.60)	K–W test (H2 = 0.87; p = .6466)
Copeptin osmolarity (mOsm)	12	300.00 (296.00–304.50)	16	303.50 (299.50–307.50)	19	303.00 (298.00–307.00)	K–W test (H2 = 1.14; p = .5665)
Glucose (mg/dL)	19	93.00 (89.00–101.00)	17	100.00 (91.00–105.00)	19	101.00 (90.00–132.00)	K–W test (H2 = 6.02; p = .0494*)
HbA1c (IFCC) (%)	18	5.00 (4.90–5.20)	17	5.50 (5.40–5.70)	19	5.20 (4.70–5.60)	K–W test (H2 = 11.56; p = .0031*)
HbA1c (NGSP/DCCT) (%)	18	5.40 (5.30–5.60)	17	5.90 (5.80–6.10)	19	5.60 (5.10–6.00)	K–W test (H2 = 11.56; p = .0031*)
LDL (mg/dL)	18	109.70 (100.00–135.40)	17	131.80 (98.60–163.60)	15	127.60 (104.60–154.00)	K–W test (H2 = 1.40; p = .4970)
HDL (mg/dL)	19	58.00 (51.00-67.00)	17	53.00 (43.00-62.00)	19	58.00 (53.00-71.00)	K–W test (H2 = 3.16; p = .2062)
Cholesterol (mg/dL)	19	199.00 (172.00–214.00)	17	207.00 (166.00–238.00)	19	208.00 (181.00–244.00)	K–W test (H2 = 2.22; p = .3297)
Triglycerides (mg/dL)	19	81.00 (58.00–96.00)	17	105.00 (82.00–132.00)	19	77.00 (57.00–156.00)	K–W test (H2 = 3.08; p = .2144)
Salivary α-amylase (U/mL)	15	198.65 (98.15–217.25)	0	NA	20	260.78 (135.90–382.05)	WRS test (S = 223.00; p = .1222)
Salivary cortisol (μg/dL)	15	0.16 (0.12–0.21)	0	NA	20	0.14 (0.08–0.21)	WRS test (S = 297.00; p = .3819)
Hair-cortisol (pg/mg)	19	11.52 (7.76–16.51)	17	11.24 (9.81–12.23)	20	9.46 (6.70–18.88)	K–W test (H2 = 0.3534; p = .8380)
TNFα (pg/mL)	19	0.73 (0.00–1.56)	15	0.10 (0.00–1.15)	20	0.42 (0.00–1.56)	K–W test (H2 = 0.98; p = .6111)
IL-6 (pg/mL)	19	1.15 (0.00–3.46)	15	0.58 (0.00–2.12)	20	2.40 (0.00–5.38)	K–W test (H2 = 2.15; p = .3413)
IL-10 (pg/mL)	19	3.00 (2.27–3.74)	15	3.44 (2.85–4.32)	20	3.15 (1.53–3.74)	K–W test (H2 = 2.68; p = .2620)

^a Missing samples due to either the unsuitable preservation or improper laboratory analysis.

* Group comparison statistically significant (p < .05).

WRS: Wilcoxon Rank Sum test; K–W test: Kruskal–Wallis test.

Distribution values of concentration of copeptin, copeptin osmolarity, LDL, HDL, cholesterol, triglycerides, salivary α-amylase, salivary cortisol, TNFα, IL-6, and IL-10 did not differ significantly among non-caregivers, geriatric caregivers, and oncologic caregivers groups (p value > .05) (Table 3). Finally, no significant group differences were observed, neither for the distribution of hair characteristics nor for cortisol concentration (p value >.05).

Discussion

In this study, we assessed the perceived stress and the concentration of endocrine, metabolic, and inflammatory biomarkers in geriatric and oncologic caregivers in comparison with a non-caregivers group. Our results demonstrate that perceived stress, as well as glucose and glycated hemoglobin levels are not homogeneous in the caregivers community under chronic stress.

From a sociodemographic point of view, non-caregivers enjoy a better health condition and a more active life, i.e. they participate in sports and have a job, compared to the caregivers group. However, since this study is a cross-sectional analysis, this difference cannot be directly attributed to stress.

Oncologic caregivers presented higher overall perceived stress, higher current state of anxiety, and higher perceived self-reported psychological stress rate than caregivers of geriatric patients with chronic diseases. Trait anxiety (STAI-t) did not differ between both groups.

Cancer diagnosis and progress have an important impact on oncologic patients and their relatives in terms of stress. Frequent clinical visits are required by the patients who also have to be submitted in drug-treatments with devastating toxic effects, while health deterioration becomes more and more evident. Due to the frequently unexpected evolution of the patient, emotional burden produced when taking care of an oncologic patient differs a lot from that suffered when taking care of an elder. PSS scores obtained in our oncologic caregivers group were consistent with the scores published in similar studies (Kessler et al., 2014). Our study design did not allow us to assess the effect of age difference nor the effect of lifestyle differences between geriatric and oncologic caregivers (chronic diseases, occasionally intake of psychotropic, smoking, coffee drinking, and job status). However, we highlight that PSS measures the response to a high-impact life event within a particular timespan and its validity criterion is not affected by age or gender (Cohen et al., 1983).

Several studies have shown that psychological stressors may have physiological impact in the organism (Hjortskov et al., 2004; Schulz & Beach, 1999; Siegenthaler et al., 2014; Urwyler et al., 2015). To evaluate this impact, our study assessed the levels of copeptin, glucose, glycated hemoglobin, LDL, HDL, cholesterol, triglycerides, α-amylase, cortisol, TNFα, IL-6, and IL-10 in geriatric caregivers, oncologic caregivers, and non-caregivers during a unique session.

We found evidence of higher levels of glucose in oncologic caregivers and higher levels of glycate hemoglobin in geriatric caregivers. The geriatric results could seem controversial considering that the levels of glucose were found lower in geriatric than oncologic caregivers. In any case, the medians were not clinically alarming. We think that the differences seen between glucose and glycated hemoglobin in both groups could have been produced by bias caused by the session setting as for the oncologic caregivers and/or by the age of geriatric caregivers. Oncologic caregivers were appointed at the hospital during a programed visit of the patient to his doctor, while the geriatric caregivers were invited to participate at home during a routine elder's home visit. On the other hand, metabolic dysfunction is often observed in elderly. The influence of these factors should be addressed in future studies.

Siddiqui et al. (2015) found elevated levels of salivary cortisol and salivary α-amylase in type 2 diabetes mellitus patients as a response to stress. In our study, some oncologic caregivers presented abnormal concentrations of glucose, at almost pre-diabetes levels, but salivary and hair-cortisol levels were within the normal range. This discrepancy may be attributed to methodological details. Cortisol concentration increases in the organism in an acute stress situation. In a chronic stress scenario, the association between self-reported mental stress and the cortisol response is controversial because the diversity of stressors, the mental stress measures, and the intra-individual variability of diurnal cortisol fluctuation (Hjortskov et al., 2004; Schulz, Kirschbaum, Prüßner, & Hellhammer, 1998).

Other approaches suggest hair sampling to estimate cortisol concentrations. Hair samples obtained 5 cm away from the scalp are not affected by diurnal fluctuation or changes in cortisol binding globulin, but they are not exempt of sampling challenges (Russell, Koren, Rieder, & Van Uum, 2012). Studies performed in animals and humans showed that hair cortisol concentration increases under stress situations although cosmetic treatments could decrease its detectable value. Therefore, although cortisol in hair could be a potential biomarker of chronic stress, clinical results might be biased and hence not valid, unless these artifacts could be eliminated. A reliable correlation between hair cortisol levels and psychological tests and the clinical relevant range have not been demonstrated yet (Russell et al., 2012).

In our study, the levels of remaining endocrine, metabolic, and immunologic biomarkers did not differ significantly among non-, geriatric, and oncologic caregivers groups, respectively.

Beyond defining concrete values of biomarkers, which would help us asses the level of stress and its impact on the health of specific groups, like the caregivers, it would be interesting to monitor caregivers populations in long-term: the incidence of illnesses and health expenditure compared with a similar population of non-caregivers.

Our study presents several strengths and limitations. The main strengths are: (1) for the first time our study compares perceived stress scores obtained from geriatric, oncologic and non-caregivers all together and (2) we have demonstrated that the perceived stress level is different between oncologic and geriatric caregivers, whereas propensity to be anxious did not differ among groups. Thus, our study may provide the baseline information to promote specific social actions and empowerment programs customized to different types of caregivers who suffer of chronic stress. The main limitations are: (1) the small size of the sample enrolled in each group which limits the power of the statistical analysis and the prediction capacity of our set of biomarkers, so results should be validated in further studies and (2) despite cortisol levels were also analyzed in hair samples to reduce cortisol diurnal fluctuations bias, hair-samples might not be exempt of artifacts produced by cosmetic hair treatments. Our results ought to be consolidated in prospective cohort studies with larger groups of caregivers with the ultimate goal to define appropriate medical and social actions that could prevent comorbidities.

Conclusions

Caregivers of chronic patients are an emerging population given the sociodemographic characteristics in which we live and the continuous medical advances especially since the last 20 years. The value of stress perceived is not homogeneous in the caregivers community summited to a chronic situation of stress. Caregivers of oncologic patients perceive higher stress and have higher levels of glucose in plasma than caregivers of geriatric patients with chronic diseases.

These results provide the baseline information to initiate social actions addressed to increase the empowerment targeting different types of caregivers with chronic stress.

Ethical disclosures

The authors declare that the procedures followed were in accordance with the regulations of the responsible Clinical Research Ethics Committee and in accordance with those of the World Medical Association and the Helsinki Declaration. The authors declare that no patient data appears in thiarticle. The authors have obtained the informed consent of the patients and/or subjects mentioned in the article. The author for correspondence is in possession of this document.

Acknowledgements

We thank A. Lobo, C. de la Cámara, and R. López-Antón from the ZaraDemP Department of the Lozano Blesa Hospital in Zaragoza for their contribution to the psychological and psychiatric side of the study and in the definition of the psychometric tests; B. López-Barbeito from the Critical Care Unit of the Clinic Hospital of Barcelona for his contribution in the clinical protocol preparation; A. Armario from the Autonomous University of Barcelona: and to Angel Corbí from the Myeloid Cell Laboratory of the Centro de Investigaciones Biológicas, CSIC for their contribution to the study concerning biochemical variables acquisition, analysis, and interpretation and to Cristina Gil Roda for providing medical writing support, which was funded by CIBER-BBN, Madrid (Spain) in accordance with Good Publication Practice (GPP3) guidelines (http://www.ismpp.org/gpp3).

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was funded by the MINECO (Spanish Ministry of Economy and Competitiveness) Project FIS - PI12/00514 and by the Centro de Investigación Biomédica en Red sobre Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN) of the Instituto de Salud Carlos III de España. There was no involvement in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication from part of the sponsor.