This article was downloaded by: [University of California, San Francisco] On: 21 February 2015, At: 23:42 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Behavioral Medicine Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/vbmd20

Is Salivary pH a Marker of Depression among Older Spousal Caregivers for Cancer Patients? a

b

Rabia Khalaila , Miri Cohen & Jamal Zidan a

Zefat Academic College

b

University of Haifa

c

c

Ziv Medical Center Accepted author version posted online: 21 Nov 2013.Published online: 22 Apr 2014.

Click for updates To cite this article: Rabia Khalaila, Miri Cohen & Jamal Zidan (2014) Is Salivary pH a Marker of Depression among Older Spousal Caregivers for Cancer Patients?, Behavioral Medicine, 40:2, 71-80, DOI: 10.1080/08964289.2013.861794 To link to this article: http://dx.doi.org/10.1080/08964289.2013.861794

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

BEHAVIORAL MEDICINE, 40: 71–80, 2014 C Taylor & Francis Group, LLC Copyright  ISSN: 0896-4289 print/1940-4026 online DOI: 10.1080/08964289.2013.861794

Is Salivary pH a Marker of Depression among Older Spousal Caregivers for Cancer Patients? Rabia Khalaila Zefat Academic College

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

Miri Cohen University of Haifa

Jamal Zidan Ziv Medical Center

The pH in saliva, which decreases due to the activation of the sympathetic nervous system, may serve as a biomarker of psychological distress in caregivers but has rarely been studied in this context. The aims are to examine the levels of salivary pH as a possible biomarker of depression among caregivers and whether depression mediates the association between caregiving status (cancer caregivers vs. non-cancer caregivers) and pH levels. Cross-sectional data were collected from 68 consecutive-sampled spouses of cancer patients, and 42 agematched individuals. Lower levels of pH saliva were found among caregivers of cancer patients than in the comparison group. Being a caregiver, poor subjective health, higher depression, and lower mastery predicted lower pH levels. In addition, depression mediated the associations of mastery with pH levels. The study provides preliminary evidence that salivary pH may serve as an easily tested indicator of the stress of caregiving and its related depression. Keywords: biomarkers, mastery, mental health, pH saliva

INTRODUCTION The search for biomarkers has recently attracted considerable attention in the medical and psychological sciences.1 Biomarkers are defined as quantifiable biological indicators of a particular physiological process, whether normal or pathophysiologic, or of the risk of developing a disease. Several biomarkers are currently being used for routine screening, such as low density lipoprotein (LDL) as a biomarker for blood cholesterol,1 and some are being researched at present, such as the use of pro-inflammatory cytokine levels in serum for identifying individuals at high risk of sepsis.2 Biomarkers have the potential to improve the identification rate of individuals at risk of developing psychological symptoms (such as depression or posttraumatic stress syndrome) and predict the severity of these symptoms.1,3 Thus, they may lead to an improvement in the treatment of these at-risk individuals.1

However, the scientific search for biomarkers of psychological symptoms is just beginning.1 Efficient biomarkers should be stable and sensitive, but also easy to measure1 and the collection of biomarker specimens should be painless and non-invasive for patients. Therefore, special attention has been given to identifying physiological biomarkers of stress in a saliva test.4,5 For example, previous studies have shown that salivary cortisol is a significant indicator of stress4,6; however, it is not a very efficient measure due to being very costly and time-consuming. A simple potential low-priced biomarker, which has, however, not yet been studied thoroughly, is the test of salivary pH.7,8 The purpose of the current study is to examine the possibility of using the salivary pH level as a marker of caregiving depression among older spouses of cancer patients, by comparing levels of pH in caregivers of cancer patients with those of a matched comparison group. STRESS AND BIOMARKERS

Correspondence should be addressed to Rabia Khalaila, RN, PhD, Head of Nursing Department, Zefat Academic College, 11 Jerusalem St., P.O.B. 160, Zefat 13206, Israel. E-mail: [email protected]

The study of biomarkers for psychological outcomes of stressful encounters, such as caregiving for an ill family

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

72

KHALAILA ET AL.

member, draws from the rich empirical evidence of the effects of stress on bodily systems, for example, the immune or cardiovascular system.9,10 These effects of stress are exerted via the hypothalamic-pituitary-adrenocortical (HPA) axis and the sympathetic nervous system, resulting in the secretion of excessive levels of stress hormones, mainly cortisol and catecholamines.11,12 Chronic high concentrations of stress hormones can cause or accelerate processes of illness,10,13 especially in old age.14 Excessive levels of cortisol and catecholamines have an effect on a wide range of physiological processes throughout the body, such as the metabolism. Therefore, they may cause changes in the level of secretion of various molecules or substances that are expressed in blood or saliva, and may be assessed as biomarkers of the effects of stress. A large body of research demonstrated that proinflammatory cytokines in serum15 or cortisol levels in blood16 or saliva17 are associated with the degree of psychological symptoms, such as depression or posttraumatic syndrome, manifested in reaction to stress3,15–17 It has been suggested that several other biomarkers are associated with psychological measures. For example, levels of salivary amylase and protein18 or Plasminogen Activator Inhibitor-1 antigen,19 which has been implicated in the development of cardiovascular disease, were also found to be related to stress.

STRESS AND pH LEVELS IN SALIVA Using levels of salivary pH as a biomarker of stress or depression might be an easy, inexpensive and non-invasive measure; however, only two studies could be found in which the effect of stress on levels of pH was measured.7,8 In the first study, Sandin and Chorot7 examined changes in levels of skin, salivary, and urinary pH in reaction to academic oral examination stress in 32 women. The results showed that during the stress situation statistically significant reductions in the skin, salivary, and urinary pH were produced, and these were associated with levels of anxiety. In the second study, Morse et al.8 studied ten dental patients requiring nonsurgical endodontic therapy on upper anterior teeth who practiced simple word meditation in order to relax. Results showed that pre-test/post-test meditation reduced anxiety significantly as measured by responses to a questionnaire, increased salivary volume, and raised salivary pH. The physiological mechanism that may explain the stresssaliva pH association is based on previous findings that the salivary pH level reliably reflects blood pH (20). Salivary pH levels are found to change with the activation of the sympathetic nervous system under stress, which leads to a lowering of the rate of secretion from the saliva glands in the mouth.18,21 This lowered rate leads to a decrease in the bicarbonate secreted in saliva (alkaline), which causes a decrease in oral pH as the acidity rises.8,18,22 However, the use of salivary pH as a biomarker of stress must be examined

further. The present study examined this topic in the context of the stress of spousal caregiving for cancer patients.

THE STRESS OF CAREGIVING Caregiving stress has been found to have a negative effect on the psychological and physical health of individuals caring for ill spouses,9,23 particularly in those who care for cancer patients,24,25 and in older spousal caregivers.26 In most of the studies, perceived caregiving stress27 and depression24,27,28 appear as the most frequently used indicators for assessing the psychological symptoms of caregiving for relatives with a chronic disease,23 particularly among those caring for a spouse with cancer.28 However, very few studies have addressed the physiological effects of caregiving on older persons caring for partners who have cancer.28 This is of special importance given the decrease in the general health and changes in the functioning of bodily systems (eg, immune, endocrinological, and cardiovascular systems)29 of this population. In addition, older adults tend to express fewer negative feelings,30 and therefore depressed caregivers may not be identified by healthcare professionals. Thus, the need to search for feasible biomarkers is even greater with regard to this population. The physiological implications of caring for relatives ill with cancer were mostly studied in relation to immune parameters. For example, adult daughters of women with breast cancer were found to have lower level of Th1 cytokines (a group of cytokines responsible for regulation of cytotoxic function; eg, Interleukin [IL]-2 or IL-12) and lower natural killer cytotoxic activity.31,32 Several studies of spousal caregivers of Alzheimer patients showed lower cytotoxic functions of T and natural killer cells, lower secretion of Th1 cytokines, and a less efficient immune response to flu immunizations.14 However, pH in relation to caregiving stress has not yet been studied.

CAREGIVING STRESS, MASTERY, AND BIOMARKERS Personal resources, such as mastery or perceived control, were often found to be buffers of the impact of the stress of caregiving on mental and health measures,33,35 and on immune functions.31 Pearlin and Schooler36 suggested a theoretical model of the process of coping with the stress of caregiving, in which perceived mastery affects the psychological outcomes, such as depression, and physical health outcomes. They stated that individuals’ capacity for controlling their own life and aspects of the environment facilitates more efficient coping, thus minimizing the harmful effects of stress factors on their health. Other scholars found that stronger personal mastery may help reduce physiological reactions to caregiving stress, which in turn may have downstream

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

SALIVARY pH AND DEPRESSION IN OLDER SPOUSAL CAREGIVERS

health benefits.19,34 For example, Roepke and colleagues34 tested the association between personal mastery and sympathoadrenal medulary arousal in a sample of 69 spousal caregivers of Alzheimer’s patients. The results indicated that higher levels of personal mastery were associated with reduced norepinephrine reactivity to the stressor. In further studies, the effect of mastery on physiological parameters was found to be mediated through the psychological reaction to the stressor.37 In light of the lack of research on pH as a biomarker of psychological symptoms in reaction to stress, in this study, spousal caregiving for cancer patients was assessed in this respect as representative of chronic and intense stress and was compared to caregiving for non-cancer individuals. Depression, being a main psychological symptom of caregiving stress,33 was assessed as the psychological outcome, and as a mediator between personal mastery and pH levels. The study hypotheses were: Hypothesis 1: Higher levels of caregiving stress and depression and lower levels of saliva pH will be found among cancer caregivers as compared to age-matched non-cancer caregivers. Hypothesis 2: Caregiving stress and depression will be negatively associated with levels of salivary pH level: The higher the level of caregiving stress and depression, the lower the salivary pH, while mastery will be positively related to levels of pH. Hypothesis 3: Depression will mediate the presumed association between mastery and levels of pH.

METHODS Sample The sample comprised 68 spouses (all of whom were ≥ 50 years of age) of cancer patients who received ≥ 2 chemotherapy sessions at an oncology clinic in Safed, Israel. The comparison group consisted of a convenience sample recruited in urban and rural communities within the geographic area served by the hospital, comprising 42 agematched spouses who were caring for a partner in need of some assistance in the activities of daily living (ADL), but neither a cancer patient nor one receiving treatment, such as dialysis, for other serious health conditions. Inclusion criteria for both groups were: sufficient fluency in Hebrew or Arabic; being the primary caregiver; not suffering from any chronic diseases that might interfere with salivary pH (eg, diabetes mellitus, dialysis); and not taking medications that might affect the level of salivary pH (eg, proton pump inhibitors, histamine 2 inhibitors). The sample size was determined (using power and sample calculations software; PS, version 3.0.43) for two independent groups of subjects with a ratio of 0.5 subjects from

73

the comparison group per one subject in the cancer group. Assuming that the pH level within each group would be normally distributed with standard deviation 0.7, for effect size (difference in the cancer and non-cancer groups means) of 0.5, we would need to examine at least 54 cancer caregiver subjects and 27 comparison subjects in order to be able to reject the null hypothesis that the population means of the cancer and non-cancer groups are equal with probability (power) .85. The Type I error probability associated with t-test of this null hypothesis was .05. Measures

Dependent Variable Saliva pH Level. Saliva pH level indicates the level of acidity in the saliva. A saliva specimen given by participants prior to the interview was deposited into a special receptacle. The test was taken in the morning after a minimum rest of 10 minutes, after at least two hours had elapsed since the last meal, smoke, or snack (except for water). Each participant was asked by the tester (one of two graduate nursing students who were trained as research assistants) to fill his or her mouth with saliva, swallow it, repeat the action two more times consecutively, and then, on the third time, to spit into a small receptacle (a plastic container). A valid digital device was used, the CyberScan pH 501 made by El-Hamma Instruments Ltd., Israel, which comprises a kit of multi-purpose sensors. The measurement of the pH level was carried out immediately to minimize the environmental impact on the measurement. The tester dipped the sensor connected to the device into the collection receptacle with the saliva, until hearing the beep several seconds later, indicating that the measurement was completed. Then, the tester marked the salivary pH level displayed on the device screen on a questionnaire form. The optimum pH measurement for saliva is between 6.5 and 7.5 maintained all day (pH = 7), which means it is at a neutral level of balance between acid and alkaline in the body. A measurement of pH > 7 shows alkaline (base) saliva while pH < 7 indicates acidic saliva. Independent Variables Caregivers’ Background Details. Caregivers’ background details included age, education, ethnicity, caregiving status, and family relationship to the patient. Care Recipients’ Details. Care recipients’ details included age, education, number of symptoms, ADL and IADL status; and for the cancer patients they also included type of cancer, stage, and time elapsed since diagnosis. For the number of symptoms variable, the respondents in both groups indicated “yes” or “no” to each symptom on a list of 14 symptoms consisting of pain, nausea, vomiting, loss of appetite, sleep disturbances, fatigue, constipation, diarrhea, dry mouth, weight loss, confusion, difficulty swallowing,

74

KHALAILA ET AL.

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

difficulty breathing, and urination disorders.38 The sum of positive responses was then calculated. The ADL scale39 ranged from 0, independent; to 8, highly dependent; for the following areas: transfer, dressing, bathing, eating or feeding, and continence. A mean score was calculated. The scale has an acceptable internal consistency reliability; for the sample in the present study it yielded 0.86 (α = 0.86). The IADL inventory developed by Lawton and Brody40 consists of the activities of telephone use, shopping, housekeeping, laundry, transportation, managing medications, and managing finances. IADL scores ranged from 0, independent; to 8, highly dependent. Internal consistency reliability (Cronbach’s alpha) in the present study was 0.88.

Perceived Caregiving Stress. Perceived caregiving stress was assessed by the short version of the ZBI (Zarit Burden Interview) comprising 12 items.41 Similarly to the original 22-item scale,42 this short scale is composed of two subscales: personal strain (eg, “Do you feel that because of the time you spent with your relative you do not have enough time for yourself?”) and role strain (eg, “Do you feel you should be doing more for your relative?”). The scale was translated into Hebrew and validated by Bachner and Ayalon.43 All items are rated on a five-point Likert scale from 0, never; to 4, almost always; with a higher score representing a higher level of perceived caregiving stress in the current study (α = 0.87). Mastery. Mastery was assessed by the 7-item Pearlin and Schooler’s36 scale. The wording of five of the items is negative (eg, “ I have little control over the things that happen to me”), and of two items is positive (eg, “What happens to me in the future depends mostly on me”). Responses to the negative items were reverse-coded. Responses were rated on a scale from 1, strongly disagree; to 4, strongly agree. A mean score was calculated after reversing two items. A higher score indicates higher levels of mastery. The scale was translated into Hebrew and Arabic and then validated by Khalaila and Litwin.33 Internal consistency reliability (Cronbach’s alpha) in the present study was 0.76. Self-rated Health. Self-rated health was assessed by a single 4-point scale measure of general health used in cancer research,27 with items such as “In general, how would you rate your health today?” Respondents were asked to rate their health on the day of the interview from 1 to 4, with 1, poor; 2, fair; 3, good; and 4, excellent. Thus, a higher score reflects good self-perceived health. Level of Depression. The level of depression was measured by the Center for Epidemiologic Studies Depression Scale (CES-D),44 using the short ten-item version.45 Participants were asked to rate how often over the past week they experienced symptoms associated with depression, such as

restless sleep, poor appetite, and the like (eg, “I felt that everything I did was an effort” and “I felt depressed”). Responses were rated on a 4-point Likert scale ranging from 0, rarely or none of the time, defined as less than once a day; to 3, most or all of the time, defined as 5–7 days. The range of the composite score was 0–30, with higher scores representing greater depressive symptomatology. A cutoff score of ≤ 10 represents a risk of major depression in older adults.45 Internal consistency reliability (Cronbach’s alpha) in the present study was 0.84. Procedure A cross-sectional study with comparison group was conducted. The research assistants were stationed in the clinic waiting rooms and approached all eligible caregivers accompanying the patients for a chemotherapy session on two days a week during the research period. Five caregivers refused to participate (yielding a 93.2% participation rate). The comparison group was recruited by the two research assistants from urban and rural communities within the same geographic area served by the hospital, based on the inclusion criteria and using the snowball approach. A face-to-face structured interview was conducted for each respondent at the hospital oncology department for the cancer group, or at the respondent’s home for the comparison group. Each participant was asked to give a saliva test prior to the beginning of the approximately half-hour interview. The study instruments were administered in Arabic or Hebrew, according to the respondent’s preference. All instruments had been translated and validated in a pilot study of 25 spousal caregivers from this same population group. Back translation was applied to confirm the clarity of the statements. The data from the main survey were collected for one year from November 2010 to October 2011. Institutional review board approval was obtained from the Research Ethics Committee of the Safed Medical Center, Israel, prior to the study. All participants gave written informed consent, and received an explanation of their right to withdraw from the study at any time. Confidentiality was strictly maintained. None of the patients received any incentives for participation in study. Data Analyses Differences in means or distributions of the background and study variables were examined using t-tests, chi square, or ANOVA. To test Hypothesis 1 regarding differences between the cancer and non-cancer caregiver groups, t-tests were conducted. This was followed by calculating Cohen’s d for the magnitude of the effect (0.2 small effect, 0.5 medium effect, 0.8 large effect). In addition, the comparative assessment of pH levels of depressed caregivers (CES-D ≥ 10) and nondepressed caregivers (CES-D < 10) was conducted using a t-test. Pearson product-moment correlations were computed

SALIVARY pH AND DEPRESSION IN OLDER SPOUSAL CAREGIVERS

75

TABLE 1 Background Characteristic of Spousal Caregiver and Care-Recipients, by Group

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

Variables Care recipients characteristics Age, mean (SD), y Range Sex, n (%) Female Male Cancer site, n (%) Lung Prostate Colorectal Breast Other Stage at diagnosis, n (%) I II III IV Time since diagnosis, mo Total symptoms, mean (SD) Range ADL, mean (SD) IADL, mean (SD) No. of children, mean (SD) Caregivers characteristics Ethnicity Jewish Arabic Relationship Wife Husband Age, mean (SD), y Range Education, mean (SD), y Range Self-rated health, mean (SD) Range ∗p

Non-cancer group (n = 42)

Cancer group (n = 68)

Test statistic

63.7 (7.3) 53–80

63.6 (8.3) 50–87

t = 1.20

13 (31.0) 29 (69.0)

32 (47.1) 36 (52.9)

χ 2 = 2.70

— — — — —

11 (16.2) 10 (14.7) 17 (25.0) 16 (23.5) 14 (20.6)

— — — —

4 (7.0) 11 (19.3) 20 (35.1) 22 (38.6) 22.1 (24.0) 2–96 6.3 (2.9) 0–13 1.4 (1.9) 4.2 (2.3) 2.0 (2.2)

2.3 (2.0) 0–9 0.1 (0.3) 1.2 (1.4) 3.7 (2.5)

t = 8.40∗∗∗ t = 5.70∗∗∗ t = 8.50∗∗∗ t = –2.80∗∗ χ 2 = 0.20

26 (61.9) 16 (38.1)

45 (66.2) 23 (33.8)

29 (69.0) 13 (31.0) 60.4 (11.9) 45–79 11.8 (3.0) 6–17 3.1 (0.6) 2.0–4.0

36 (52.9) 32 (47.1) 62.7 (7.9) 50–79 11.2 (4.6) 0–26 2.6 (0.9) 1.0–4.0

χ 2 = 2.78 t = 2.70 t = –0.70 t = –4.80∗∗

< .05; ∗∗ p < .01; ∗∗∗ p < .001.

to clarify the associations between the study variables (Hypothesis 2). Bonferroni correction was conducted to control for multiple comparisons. To further assess Hypothesis 2 regarding the associations between study variables and pH levels, while controlling for confounders, hierarchical linear regressions were conducted. In the first step, the caregiving status (cancer caregivers vs. non cancer caregivers) variable was entered, and then background variables were entered in the second step. In the third and fourth step, depression and mastery were entered. Caregiving stress was not entered since a correlation coefficient of greater than 0.60 was used as an indicator of the potential for multicollinearity, as previously suggested.46 In addition, details of tolerance and variance inflation factor (VIF) were examined and indicated that multicollinearity does not exist among the independent variables entered to the regression model. To assess the third hypoth-

esis regarding mediation, we applied the Baron and Kenny47 method and Sobel test 48 for determining the significance of the mediation effect. pH levels were assessed for normality and found to be normally distributed.

RESULTS Table 1 shows the background characteristics of the care recipients and caregivers in the cancer and non-cancer caregiver groups. The background characteristics of the care recipients in the two groups were comparable on the variables of age and sex. About two-thirds of the comparison group, and slightly more than the half of the cancer group were men. The mean age of the care recipients in both groups was about 63 years, ranging from 50 to 87 years. However, the number

76

KHALAILA ET AL. TABLE 2 Means (SDs) and Differences of Study Variables, for Cancer and Non-Cancer Groups Non-cancer group (n = 42)

Variables Caregiving stress Mastery Depression (CES-D) pH level

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

∗p

Cancer group (n = 68)

Mean (SD)

Range

Mean (SD)

Range

t-test

Effect size (Cohen’s D)

0.6 (0.6) 2.8 (0.6) 5.5 (4.1) 6.7 (0.7)

0–3.3 1.6–4.0 0–14 5.2–8.5

1.3 (0.8) 2.5 (0.7) 12.4 (6.5) 6.1 (0.8)

0–3.3 1–4 0–23 4.5–7.9

−5.3∗∗ 2.43∗ −6.2∗∗ 3.5∗

0.99 0.46 1.27 0.80

< .01; ∗∗ p < .001.

of somatic symptoms was about three times greater in the cancer than in the comparison group. The care recipients’ ADL was low in both groups (possible range was 0–8 in both groups), while IADL (instrumental activities of daily living) was in the middle of the scale (possible range was 0–8 in both groups). However, both measures were significantly higher in the cancer than in the comparison group. Members of the cancer group had fewer children on average than members of the comparison group. The cancer and non-cancer caregiver groups were comparable in all the background characteristics: ethnicity; age; education; and self-rated health. About two-thirds of the caregivers in both groups were Jewish. Slightly more than the half of the cancer group and about two-thirds of the comparison group were wives. The mean age of caregivers in both groups was slightly more than 60 years (range was 45–79 years old). Finally, the average education of the caregivers in both groups was about 11 years, with a relatively wide range (0–26). Levels of self-rated health were in the middle of the range, but lower in the cancer group than in the comparison group. Table 2 shows the means (SDs) and the comparisons between the cancer caregivers and the non-cancer caregivers group on study variables. Levels of mastery were in the middle of the scale, but significantly (medium effect size) lower for the study group. In support of the first hypothesis, levels of perceived stress were low in both groups (possible ranges were 0–4), but the score in the cancer caregivers group was more than double of that in the comparison group. This effect was significant and with large effect size. In addition, in support of the first hypothesis, the mean depression score for caregivers in the cancer group was more than double of that in the comparison group, and the difference was statistically significant with a very large effect size. Moreover, the mean depression score of the cancer group was above the cut-off point of 10, indicative of a risk of clinical depression in older people,42 while in the non-cancer group the mean score was much below the cut-off point. Further analysis was conducted to compare the pH levels of depressed (CES-D ≥ 10) (N = 45) and non-depressed caregivers (CES-D < 10) (N = 45). The result showed that the pH level was lower (mean = 5.9, SD = 0.7) among depressed as compared to nondepressed caregivers (mean = 6.6, SD = 0.8), and this dif-

ference was also statistically significant, t(88) = 4.01, p < .001. Further in support of hypothesis 1, the means of the salivary pH levels were lower (ie, higher acidity levels) in the cancer group than in the comparison group. The difference was statistically significant with large effect size. Table 3 presents the bivariate correlation matrix between the study variables. Caregiving burden and depression were positively correlated, while each was negatively correlated with mastery. That is, the higher the levels of perceived burden and depression, the lower were the reported levels of mastery. The pH level was negatively associated with a higher level of caregiving burden and depressive symptoms, and positively associated with mastery. Thus, the higher the level of depression and perceived burden and the lower the level of mastery, the lower were the pH levels. After Bonferroni correction, except for the associations of mastery with caregiving stress, all correlations remained significant. Table 4 displays the results of the hierarchical regression analysis. The background variables (age, family status, number of children, economic status, ethnicity, care recipients’ ADL, IADL, and disease-related variables) were not associated with salivary pH levels, and therefore were not included in the model. In addition, caregiving status was highly associated with the depression score, CES-D (r = .66, p < .001), and excluded from the analysis due to concern for multicollinearity.46 In the first step, the caregiving status (being cancer or non-cancer caregiver) variable was entered. Being a cancer caregiver was negatively and strongly associated with pH levels, that is, cancer caregivers’ saliva had higher acidity levels. The caregiving status variable explained 12% TABLE 3 Pearson Correlations between Study Variables (N = 110) Variables 1. Caregiving stress 2. Mastery 3. Depression 4. pH level ∗p

< .01; ∗∗ p < .001.

1

2

4

5

1.00 −0.46∗ 0.66∗∗ −0.43∗∗

1.00 −0.52∗∗ 0.40∗∗

1.00 −0.53∗∗

1.00

77

SALIVARY pH AND DEPRESSION IN OLDER SPOUSAL CAREGIVERS TABLE 4 Summary of Hierarchical Regression Analysis for Predicting pH Saliva among Spousal Caregivers (N = 110) Model 1 Variable

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

Caregiving status (1 = cancer) Sex (1 = male) Education Self-rated health Mastery Depression (CES-D) R (Adjusted R2) F for change in R2

Model 2

Model 3

Model 4

B

SE

β

β

SE

β

B

SE

β

B

−.61

.18

−.35∗∗

−.29 −.08 −.01 .52

.16 .14 .02 .09

−.17 −.05 .06 .55∗∗∗

−.23 −.19 .03 .44 .34

.17 .14 −.02 .08 .11

−.02 −.12 −.12 .48∗∗∗ .30∗∗

−.05 −.21 −.03 .40 .20 −.31

.12 (.11) 12.34∗∗

.39 (.36) 13.37∗∗∗

0.46 (.43) 10.43∗∗

SE .16 .14 .02 .08 .11 .12 .50 (.46) 6.91∗

β −.03 −.12 −.16 .43∗∗∗ .18 −.29∗∗

Note. Reference category: Caregiving status (0 = non-cancer); sex (0 = female). < .05; ∗∗ p < .01; ∗∗∗ p < .001.

∗p

of the variance of salivary pH. Background variables of sex, education, and self-rated health were entered in the second step, which added 26% to the explained variance of pH, with only self-rated health appearing to predict salivary pH uniquely, and with the group variable no longer significantly associated with pH: the lower the level of self-rated health, the lower was the pH level (ie, the higher the acidity level). In the third step, mastery was added and contributed 7% to the explained variance of pH, with self-rated health and mastery being significantly associated with pH (in accordance with the second hypothesis). In the final step, CES-D was added and contributed 4% to the explained variance of pH, with self-rated health, CES-D being significantly associated with pH (in accordance with the second hypothesis). Mastery ceased to be significant when CES-D was added, due to the effect of depression as a possible mediator. The final model explained 48% of the pH level with an adjusted R2 of .44 (p < .001). The hypothesized mediation (Hypothesis 3) was further assessed, based on Baron and Kenny’s method (47) and the Sobel test for the assessment of the statistical significance of the mediation effect. The mediating role of depression on the relations of mastery and pH levels was assessed. The pH level was regressed on mastery (β = .40, p < .001), which resulted in a beta decreased to β = .15; (p < .05) when CES-D was added. The Sobel test showed z = –2.04 (p < .05). Thus, the third hypothesis was also confirmed, that is, CES-D was a partial mediator of the mastery and pH associations.

DISCUSSION The present study is innovative and provides a unique contribution to the literature in that it shows lower salivary pH levels among spouses caregiving for cancer patients, who also had higher caregiving stress (represented by perceived caregiving stress and depression scores) than the comparison group. The present results also demonstrate a mediation

effect of depression on the associations of caregiving status and mastery according to pH levels. The cancer caregivers perceived the caregiving role as more stressful and exhibited higher levels of depression than the age-matched non-cancer caregivers who tapped the lower levels of pH in saliva. These differences were found to have a large effect size. In addition, the current study supported the “known-group” validity of the pH measure by demonstrating differences in pH salivary levels between the depressed caregivers (CES-D ≥ 10) and non depressed caregivers (CES-D < 10) in the direction that was expected. Moreover, controlling for background variables, levels of salivary pH were related to levels of caregiving stress, depression, and mastery. Therefore, the present results, although preliminary, suggest that salivary pH may be a simple biomarker of the long-term outcomes of the stress of caregiving. Only two previous studies were found that assessed the association between any psychological outcomes of stress and salivary pH levels, and they support our findings.7,8 In addition, the current results are supported by the large amount of empirical knowledge that has been accumulating since Selye49 introduced his model of the HPA axis’s response to physiological stress,11 particularly, the immune functions9,10,31 or cardiovascular measures, such as heart rate and heart rate variability50 and the neuro-endocrine system, which regulates pH levels.4,51 In addition, several recent studies reported an association between salivary pH and an increased risk of diabetes and cardiovascular diseases,29 Moreover, the results are in accordance with previous studies that assessed different salivary biomarkers, such as cortisol4,6 or amylse,18 and found them to be associated with psychological symptoms in reaction to stress. However, further studies are required to clarify the association between salivary pH and other saliva markers using a control design and larger sample. The present study focused on caregivers’ perceived stress and depression, as they comprise the most common outcomes of caregiving stress.24,27,28 pH is probably not specific for these psychological states, but rather may be a

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

78

KHALAILA ET AL.

general indicator of psychological distress, including also other symptoms such as anxiety or posttraumatic stress disorder. Similarly, other physiological parameters related to stress, such as cortisol, catecholamines, or immune parameters, are general indicators of distress and were found associated with different psychological symptoms, such as anxiety7 or posttraumatic symptoms.3 However, the possible associations of pH with other psychological symptoms need to be studied. Biomarkers in general, including saliva pH, may be affected by various factors, such as sickness or other physical problems18 smoking,52 medications, 53 and dietary intake.53 In order to control for these possible confounders, we have excluded from the present study participants with severe chronic diseases or those using medications having direct effects on pH levels and the effect of self-rated health was controlled for. In addition, saliva specimens were taken two hours after partaking of any food or drink, or smoking in order to neutralize their effects.56 Nevertheless, future largescale and longitudinal studies are required in order to control for a range of possible confounders. The association between self-rated health and salivary pH also warrants attention. Self-rated health is considered to be a reliable indicator of actual health state,55 thus the possibility suggested by the present results that pH may be a predictor of self-rated health may have both theoretical and clinical significance that should be further assessed. It is important to note that the absolute difference between means of pH found between the caregivers for cancer patients and non-cancer caregivers, although having a large effect size, was relatively small. However, even for more established biomarkers, such as pro-inflammatory cytokines, there is no clear consensus on what size of difference in mean levels is clinically relevant. Studies using cytokines as biomarkers, with small,57,3 and large samples,58 also found small absolute differences in levels of cytokines or other inflammatory markers that were statistically significant. Although several studies link high levels of biomarkers to various health problems, such as inflammatory diseases, slower wound healing, and lower efficiency of vaccines,59 the question of what magnitude of elevated levels of cytokines is a risk factor for health problems is not yet clear,59 and especially not for saliva pH, which has only now been proposed as a biomarker. These issues should be addressed in future research. Coping theories allocate a central role for mastery in the process of coping with stressors36 due to its being a major personal resource, impacting the health outcomes of stressful encounters, especially caregiving stress,24,3 and related to physiological responses to stressors.19,34,37 Indeed, in the present study mastery was found to be strongly associated with levels of pH, and this association was partially mediated by the levels of depression. According to this observed mediation patterns it appears that level of mastery influences degree of depression, which in turn affects salivary pH level.

This effect of mastery on salivary pH through the levels of depression also accords to the coping theories,60 and previous studies showing that the association of mastery with physiological biomarkers was mediated by depression.37 This may have clinical implications, such as a possibility to identify caregivers at higher risk for depression and negative health implications and to offer them psycho-social support or help. It should be noted that both groups demonstrated relatively low levels of perceived caregiving stress, in contrast to relatively higher levels of depression. The low rating of stress might be due to the tendency of older individuals to conceal negative feelings through emotional control strategies (61), especially when they are directly asked about negative feelings. In contrast, the items of the depression scale probe on specific behavioral, physical and mental symptoms and therefore provide a more accurate picture on psychological symptoms. The current study has several limitations that must be taken into account. First, because it was a relatively small, cross-sectional, one-point study, causal associations cannot be assumed. An additional limitation is that it was a singlecenter study with a non-randomized sample, which limits the generalizability of the study results to other groups of cancer patients. However, the study is the first to indicate the possibility that salivary pH may serve as a biomarker of stress and its resulting impact on the physical and psychological health associated with being caregivers for cancer patients. Future studies with longitudinal designs are required to explore the associations of different psychological states, such as strain, anxiety, posttraumatic syndrome, life chaos, and the role of personal resources, such as mastery, self-efficacy, and coping strategies, with saliva pH levels. In addition, further studies are needed to elucidate the biological mechanisms and the clinical implications that underlie the changes in salivary pH in reaction to stress from a clinical point of view. CONCLUSIONS The results suggest that it is possible to identify at-risk caregivers and assist them to regain personal mastery and reduce distress by providing suitable tools to deal with the effects of stress and its ramifications on their physical and mental health. These tools could include psychosocial interventions such as support from the multi-professional oncology staff, support groups, cognitive-behavior therapy or relaxation training. The effect of these interventions on levels of pH should be studied in further randomized studies with longitudinal designs. REFERENCES [1] Macaluso M, Preskorn SH. How biomarkers will change psychiatry: from clinical trials to practice. Part I: introduction. J Psychiatr Pract. 2012;18:118–121.

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

SALIVARY pH AND DEPRESSION IN OLDER SPOUSAL CAREGIVERS [2] Stryjewski GR, Nylen ES, Bell MJ, et al. Interleukin-6, interleukin-8, and a rapid and sensitive assay for calcitonin precursors for the determination of bacterial sepsis in febrile neutropenic children. Pediatr Crit Care Med. 2005;6:129–135. [3] Cohen M, Meir T, Klein E, Volpin G, Assaf M, Pollack S. Cytokine levels as potential biomarkers for predicting the development of posttraumatic stress symptoms in casualties of accidents. Int J Psychiatr Med. 2011;42:117–131. [4] Hellhammer DH, Wus S, Kudielka BM. Salivary cortisol as a biomarker in stress research. Psychoneuroendocrinol. 2009;34: 163–171. [5] Nater UM, Rohleder N. Salivary alpha amylase as a non-invasive biomarker for the sympathetic nervous system: current state of research. Psychoneuroendocrinol. 2009;34:486–496. [6] De Vugt ME, Nicolson NA, Aalten P, Lousberg R, Jolle J, Verhey FRJ. Behavioral problems in dementia patients and salivary cortisol patterns in caregivers. J Neuropsychiatr Clin Neurosci. 2005;17:201–207. [7] Sandin B, Chorot P. Changes in skin, salivary, and urinary pH as indicators of anxiety level in humans. Psychophysiology. 1985;22:226–230. [8] Morse DR, Schacterle GR, Furst ML, Esposito JV, Zaydenburg M. Stress, relaxation and saliva: relationship to dental caries and its prevention with a literature review. Ann Dent. 1983;42:47–54. [9] Davis LL, Weaver M, Zammrini E, Stevens A, Kang DH, Parker CR. Biopsychological markers of distress in informal caregivers. Biol Res Nurs. 2004;6:90–99. [10] Karatsoreos IN, McEwen BS. Psychobiological allostasis: resistance, resilience and vulnerability. Trends Cogn Sci. 2011;15:576–584. [11] Brummett B, Boyel SH, Siegler IC, et al. HPA axis function in male caregivers: effect of the monoamine oxidase-A gene promoter. Biol Psychol. 2008;79:250–255. [12] Hugo FN, Hilgert JB, Corso S, et al. Association of chronic stress, depression symptoms and cortisol with low saliva flow in a sample of south-Brazilians aged 50 years and older. Gerontology. 2008;25:18–25. [13] Natelson BH. Stress, hormones and disease. Physiol Behav. 2004;82:139–143. [14] Kiecolt-Glaser J, Glaser R, Gravensrein S, Malarkey WB, Sheridan J. Chronic stress alters the immune response to influenza versus vaccine in older adults. Proc Natl Acad Sci USA. 1996;93:3043–3047. [15] Kiecolt-Glaser Jk, Glaser R. Depression and immune function central pathways to morbidity and mortality. J Psychosom Res. 2002;53:873–876. [16] Burk HM, Davis MC, Otte C, Moher DC. Depression and cortisol responses to psychological stress; a meta-analysis. Psychoneuroendoceinol. 2005;30:846–856. [17] Kunz-Ebrecht SR, Mohamed-Ali V, Feldman PJ, Kirschbaum C, Steptoe A. Cortisol responses to mild psychological stress are inversely associated with proinflammatory cytokines. Brain Behav Immun. 2003;17:373–383. [18] Wu KP, Ke JY, Chung CY, et al. Relationship between unstimulated salivary flow rate and saliva composition of healthy children in Taiwan. Chang Gung Med J. 2008;31:281–286. [19] Mausbach BT, von K¨anel R, Depp C, et al. The moderating effect of personal mastery on the relations between stress and Plasminogen Activator Inhibitor-1 (PAI-1) antigen. Health Psychol. 2008;27:S172–S179. [20] Hofman LF. Innovative non- or minimally-invasive technologies for monitoring health and nutritional status in mothers and young children. J Nutr. 2001;131:1621S–1625S. [21] Rojas-Morales T, Navas R, Viera N, Alvarez CJ, Chaparro N. pH and salivary sodium bicarbonate in cancer patients: correlation with seric concentration. Med Oral, Patol Oral. 2008;13:E456–E459. [22] Bardow A, Moe D, Nyvad B, Nauntofte B. The buffer capacity and buffer systems of human whole saliva measured without loss of CO2. Arch Oral Biol. 2000;45:1–12.

79

[23] Kiecolt-Glaser JK, Dura JR, Speicher CE, Trask J, Glaser R. Spousal caregivers of dementia victims: longitudinal changes in immunity and health. Psychosom Med. 1991;53:345–362. [24] Gaugler JE, Linder J, Given CW, Kataria R. Family cancer caregiving and negative outcomes: the direct and meditational effects of psychosocial resources. J Fam Nurs. 2009;15:417–444. [25] Kurtz ME, Kurtz JC, Given CW, Given B. Depression and physical health among family caregivers of geriatric patients with cancer: a longitudinal view. Med Sci Mon. 2004;10:CR447–CR456. [26] Haley WE, Lamonde LA, Han B, Burton AM, Schonwetter R. Predictors of depression and life satisfaction among spousal caregivers in hospice: application of a stress process model. J Palliat Med. 2003;6:215–224. [27] Fitzell A, Pakenham KI. Application of a stress and coping model to positive and negative adjustment outcomes in colorectal cancer caregiving. Psychooncology. 2010;19:1171–1178. [28] Kim Y, Duberstein PR, Sorensen S, Larson MR. Levels of depressive symptoms in spouses of people with lung cancer: effects of personality, social support, and caregiving burden. Psychosomatics. 2005;46:123–130. [29] Margaret F, Bevans MF,Sternberg EM. Caregiving burden, stress, and health effects among family caregivers of adult cancer patients. JAMA. 2012;307:398–403. [30] Moshe CE, Danoff-Burg S. A review of age differences in psychological adjustment to breast cancer. J Psychosoc Oncol. 2005;23: 101–114. [31] Cohen M, Klein E, Kuten A, Fried G, Zinder O, Pollack S. Increased emotional distress in daughters of breast cancer patients is associated with decreased natural cytotoxic activity, elevated levels of stress hormones and decreased secretion of Th1 cytokines. Int J Cancer. 2002;100:347–354. [32] Cohen M, Pollack S. Mothers with breast cancer and their adult daughters: the relationship between mothers’ reaction to breast cancer and their daughters’ emotional and neuroimmune status. Psychosom Med. 2005;67:64–71. [33] Khalaila R, Litwin H. Does filial piety decrease depression among family caregivers? Aging Ment Health. 2011;15:679–686. [34] Roepke SK, Mausbach BT, Aschbacher K, et al. Personal mastery is associated with reduced sympathetic arousal in stressed Alzheimer caregivers. Am J Geriatr Psychiatr. 2008;16:310–317. [35] Sherwood PR, Given CW, Given B, Schiffman RF, Murman DL, von Eye A, Lovely M, Rogers LR, Remer S. The influence of caregiver mastery on depressive symptoms. J Nur Scholarship. 2007;3:249–255. [36] Pearlin LI, Schooler C. The structure of coping. J Health Soc Behav. 1978;19:2–21. [37] Harmell AL, Chattillion EA, Roepke SK, Mausbach BT. A review of the psychobiology of dementia caregiving: a focus on resilience factors. Curr Psychiatr Rep. 2011;13:219–224. [38] Tranmer JE, Heyland D, Dudgeon D, Groll D, Squires-Graham M, Coulson K. Measuring the symptom experience of seriously ill cancer and noncancer hospitalized patients near the end of life with the memorial symptom assessment scale. J Pain Symptom Manage. 2003;25:420–429. [39] Katz S, Downs TD, Cash HR, Grotz RC. Progress in development of the index of ADL. Gerontologist. 1970;1:20–30. [40] Lawton MP, Brody EM. Assessment of older people: selfmaintaining and instrumental activities of daily living. Gerontologist. 1969;9:179–186. [41] Bedard M, Molloy W, Squire L, Dubois S, Lever JA, O’Donnell M. The Zarit burden interview: a new short version and screening version. Gerontologist. 2001;41:652–657. [42] Zarit SH, Todd P, Zarit JM. Subjective burden of husbands and wives as caregivers: a longitudinal study. Gerontologist. 1986;26:260–266. [43] Bachner YG, Ayalon L. Initial examination of the psychometric properties of the short Hebrew version of the Zarit burden interview. Aging Ment Health. 2010;14:725–730.

Downloaded by [University of California, San Francisco] at 23:42 21 February 2015

80

KHALAILA ET AL.

[44] Radloff LS. The CES-D: a self-report depression scale for research in the general population. Appl Psych Meas. 1977;1:384–401. [45] Andresen EM, Malmgren JA, Carter WB, Patrick DL. Screening for depression in well older adults: evaluation of a short form of the CESD (Center for Epidemiologic Studies Depression Scale). Am J Prev Med. 1994;10:77–84. [46] Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Erlbaum; 1988:567. [47] Baron RM, Kenny DA. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51: 1173–1182. [48] Preacher KJ, Hayes AF. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behav Res Methods Instrum Comput. 2004;36:717–731. [49] Selye H. The Stress of Life. New York, NY: McGraw-Hill; 1956. [50] De Vente W, Olff M, Van Amsterdam JGC, Kamphuis JH, Emmelkamp PMG. Physiological differences between burnout patients and healthy controls: blood pressure, heart rate, and cortisol responses. Occup Environ Med. 2003;60:i54–i61. [51] Bhagwagar Z, Hafizi S, Cowen PJ. Increase in concentration of waking salivary cortisol in recovered patients with depression. Am J Psychiat. 2003;160:1890–1891. [52] Avs¸ar A, Darka O, Topalo˘glu B, Bek Y. Association of passive smoking with caries and related salivary biomarkers in young children. Arch Oral Biol. 2008;53:969–974.

[53] Bardow A, Nyvad B, Nauntofte B. Relationships between medication intake, complaints of dry mouth, salivary flow rate and composition, and the rate of tooth demineralization in situ. Arch Oral Biol. 2001;46:413–423. [54] Ernest SL. Dietary intake, food preferences, stimulated salivary flow rate, and masticatory ability in older adults with complete dentitions. Spec Care Dentist. 1993;13:102–106. [55] Park NS, Jang Y, Lee BS, Chiriboga DA. Racial/ethnic differences in predictors of self-rated health: finding from the survey of older Floridians. Res Aging. 2012;19. doi:10.1177/0164027512440572. [56] Fenoll-Palomares C, Munoz JV, Sanchiz V, et al. Unstimulated salivary flow rate, pH and buffer capacity of saliva in healthy volunteers. Rev Esp Enferm Dig. 2004;96:773–783. [57] Brummett B, Boyle S, Ortel T, Becker R, Siegler I, Williams R. Associations of depressive symptoms, trait hostility, and gender with C-reactive protein and interleukin-6 response after emotion recall. Psychosom Med. 2010;72:333–339. [58] Morozink JA, Friedman EM, Coe CL, Ryff CD. Socioeconomic and psychosocial predictors of interleukin-6 in the MIDUS national sample. Health Psychol. 2010;29:626–635. [59] Kiecolt-Glaser JK. Psychoneuroimmunology: psychology’s gateway to the biomedical future. Perspect Psychol Sci. 2009;4:367–369. [60] Lazarus R, Folkman S. Stress, Appraisal and Coping. New York, NY: Springer; 1984. [61] Carstensen LL, Charles ST. Emotion in the second half of life. Curr Direct. 1998;7:144–149.