Dear Arnold, you have received several helpful suggestions from colleagues. We have utilized the collection of salivary cortisol to answer a number of questions related to the diurnal regulation (rhythmicity), response to stress (responsivity) and moderating factors in children with autism spectrum disorders. While you may not be able to control all potential confounding factors, among the more critical factors that we control for include: food/drink intake (nothing an hour before salivary collection), medications (many can impact the level of cortisol), sleep (use a sleep diary), time of day (utilize a Trackcap devise) and sampling four times per day (immediate waking, 30-min post waking, afternoon and evening 30-min before bedtime). We also employ several statistical approaches to take into consideration intraindividual variability and between-group variablity). For example, we use linear mixed model approaches. If you need more detailed information, I can direct you to specific papers from our lab and others that describe the aforementioned procedures.

For other confounding factors it may be important to consider diet and activity. These can be relatively easy to control prior to collection which makes the analysis later easier.

While I'm sure your patients are unlikely to be very active it is worth ensuring patients are rested (i.e. no physical activity prior to sample) and also fasted with particular care to avoid caffeine.

Thanks Maurice. Drinking, eating and physical activity will be prohibited prior to the saliva collection.

Dear Arnold,

I think you should add in this list of yours the following factors: religion, psychiatric medication, presence of chronic non-communicable diseases. I honestly do not understand the reason you choose to evaluate salivary cortisol levels after exercise. In my humble opinion, you should use the minor protocol, which includes 3 salivettes in one day. 1st: Exactly after the the morning arousal 2nd: 20-30' after the arousal 3rd: 12 h after the first one. In this way, you may evaluate adequately the CAR (Cortisol awakening response).

Hi Chariklela, Thanks for your advice. I intend to investigate the relation between salivary cortisol concentrations and clinical improvements because cortisol level may provide an objective measure for stress level. After doing some literature search, it is apparent that salivary cortisol level may not be a good indicator of stress as cortisol levels can be confounded by psychological, behavioural, cognitive and biological factors. Please correct me if I am wrong.

You are right my friend. But please, keep in mind that it is not invasive. Also, you should consider hair cortisol evaluation which gives you the opportunity to safely assess stress levels 5 months prior to evaluation and thus evaluating chronic stress.

Consider using hair cortisol instead. It will provide you with a more stable biomarker of stress exposure during a much longer period of time. So you can examine total cortisol exposure during the 3 months prior to your treatment to cortisol exposure during the 3 months after treatment. A must-read reference is "Hair as a retrospective calendar of cortisol production—Increased cortisol incorporation into hair in the third trimester of pregnancy" by Kirschbaum et al., (2009) PNE, 34(1).

Arnold, you have received great answers from other researchers in this area. We have a point of care test for salivary cortisol, with our main clients in the field of sports and the corporate world. This type of test maybe useful for you(especially for pilot work), but I do urge caution, because keeping close tabs on the compliance to sample collection to genuinely investigate CAR is challenging at best. We achieve this easily in training camp environments where subjects are with us all of the time, but find it far more challenging in the corporate world. The problem is, people do not like to admit that they did not follow the correct protocol (forgot a sample, or did it 15 minutes late etc) because they do not like to et you down. We have seen some very interesting case studies of elevated cortisol levels just (two or three days) before subsequent injury in sport. However, we routinely measure these athletes twice per week through the season, so have a good feel for what is "normal" for each individual and thus look for deviations from individual profiles that we gather. In sport we do about 25,000 tests per year and do see some nice associations, however, we cannot publish these as our clients want to keep their data confidential in an attempt to gain a competitive edge (the irony to us is that they are all doing pretty much the same thing anyway). We also see interesting patterns in cortisol when athletes (or corporate clients for that matter) become more chronically fatigued. Intuitively one looks for elevations in "stress hormones" for such situations, but I can assure you that in these two fields, the people we are most worried about are those with suppressed cortisol levels. Cortisol is a nice marker, if it has been well controlled in your methodology, but we also use a whole host of other markers in our settings. I don't really want to use this forum as a sales pitch, because that is not what researchgate is for, but if you are interested in discussion, feel free to contact me direct. On the positive side there are at least two English Premier League soccer clubs who will let us publish the data that we have been gathering and this will come into the public domain later in 2013 and 2014. These will be significant contributions to the literature related to the use of biomarkers in the field of elite sport. Good luck with your project.

Dear Arnold, you have received several helpful suggestions from colleagues. We have utilized the collection of salivary cortisol to answer a number of questions related to the diurnal regulation (rhythmicity), response to stress (responsivity) and moderating factors in children with autism spectrum disorders. While you may not be able to control all potential confounding factors, among the more critical factors that we control for include: food/drink intake (nothing an hour before salivary collection), medications (many can impact the level of cortisol), sleep (use a sleep diary), time of day (utilize a Trackcap devise) and sampling four times per day (immediate waking, 30-min post waking, afternoon and evening 30-min before bedtime). We also employ several statistical approaches to take into consideration intraindividual variability and between-group variablity). For example, we use linear mixed model approaches. If you need more detailed information, I can direct you to specific papers from our lab and others that describe the aforementioned procedures.

Hi Arnold:

Please be aware the HPA Axis dysregulation is a feature of many psychiatric disorders (e.g., PTSD and depression) which makes it difficult to interpret cortisol levels (i.e., the idea that a high cortisol level is an indicator of high stress does not apply to persons with PTSD). I know you hinted at this issue in your initial comments but it really deserves very careful consideration. Lastly, there is a well known spike in cortisol that occurs upon awakening so you might want to give further consideration to the timing of your measurement(s).

On a completely different note, you might want to consider using salivary alpha amylase as a measure of sympathetic activation. Also, you might want to consider using a provocation task such as the cold pressor so you can look at stress reactivity and recovery in your sample. It may be more telling than just measuring basal levels of these markers.

Regards,

Mike

I suggest you evaluate serum cortisol before and after the procedure (treatment). On the question of the presence of depressive symptoms, really, cortisol can be altered due to impairment of the hypothalamic-pituitary-adrenal axis. I suggest you make the appropriate adjustments in statistical when analyzing results.

First you have to know that the cortisol hormone has a rhythmic production, it means its level in the morning is different and higher than nocturne levels. So during the day we are already submited a different levels of cortisol. After that, any emotional or physical stimulus during the day are able to increased cortisol levels. So for study this hormone you have to design a very specific protocol and have to pay attention in your population routine.

I suspect that quality of sleep will be a confounding factor as well. I've not studied it, but I suspect that people with LBP will not sleep as well as someone without it. You might want to be careful about immune responses as well. Cytokines will likely be involved in both pain perception and cort levels.

Hi Arnold,

you got some fantastic answers, but I would like to add that Blythe's suggestion of a fourth sample (before bedtime) is quite important, because most studies show that the main difference in cortisol salivary or plasma levels between chronically stressed individuals and healthy ones is revealed in the nighttime sample. Obviously pre- and post-treatment samples are particularly important for your purpose, but as a piece of advice, try to assay the pre-treatment samples before you distribute your patients into the treatment groups, so you will not be catch into the surprise of having pre-treatment differences among your groups (believe me, that can happen). Good luck with you project!

Hi Arnold,

as some some people here have already mentioned, basal cortisol is not a good measure for your purpose. Our research on 400 people showed that basal cortisol is connected to some personality traits. It was not affected by PTSD nor depression.

Thank you everyone for your advice. Given that salivary cortisol may not be a good indicator for monitoring sympathetic activation, I look into the possibility of using salivary alpha amylase (as suggested by Prof Saladin). I find that alpha amylase levels may respond to the changes in stress more quickly than cortisol. Although salivary alpha amylase is affected by the flow rate, diurnal cycle, age, it is not affected by the menstrual cycle. It seems to me that alpha amylase is a better choice. What are your opinions? Given the concern of budget and small sample size (32 subjects), will you recommend me to use a cold pressor test to examine the stress reactivity and recovery?

As Joe Dunbar, I suggest CAR measure. For PTSD (chronic stress) is considered a reliable measure for the acute reagibility of the HPA axis (for a great review: http://p113367.typo3server.info/uploads/media/lit0905.pdf).

Dear Arnold,

The best way is to analyse alpha amylase too, and see the reactivity and recovery of cold test or another stressor test together with the basal concentrations of both hormones.

Dear Leah, Rui and Helena, thanks for your suggestion. I think alpha amylase is a valid biomarker for measuring stress. I will use a standardized saliva collection procedure to control for the environmental confounders. Salivary alpha amylase is good because it won't be affected by the menstrual cycle or female hormones. In short, I don't need to adjust for too many confounders in my statistical analysis. I plan to use the State-Trait Anxiety Inventory (STAI) as an anxiety and stress self-reported measure. Will you suggest other scales?

A colleague of mine has conducted a PhD study of the effect of music therapy on salivary cortisol levels. Its text and references may be useful:

Bolette Daniels Beck (2012): Guided Imagery and Music (GIM) with adults on sick leave suffering from work-related stress – a mixed methods experimental study

Download abstract

Download thesis

Download from : http://www.mt-phd.aau.dk/phd-theses/

Best wishes,

Erik Christensen, Aalborg University, Denmark

Hi Arnold:

I think you have the benefit of many views on how to proceed with your work... in fact, it may be difficult trying to sort through it all... on a final note, depending on your budget, you may want to consider examining salivary cortisol and AA (Salimetrics has commercial assay kits to do both)... additionally, you could consider obtaining basal measures and do a challenge task like the cold pressor task... that would cover a lot of ground and increase the measurement density and rigor of your protocol.

I have attached some relevant publications on AA for your consideration... best of luck Arnold.

Warmest regards,

Mike

Hi Prof Saladin, Thank you for the article and suggestions. I am also grateful for everybody's advice. I have a clear picture for my experiment now. Arnold :)

Dear Arnold, I want to go back to your first topic lower back pain and stress level.

In depression the major marker is interleukine-6 which is always elevated and which is causing higher Cortisol levels.The conclusion here is if cytokines are involved like

non specific IL-1beta, INF gamma and specific IL-6 depression must be an immune disease. One of the known factors which are keeping an immune system at a low

function is the lack of sunshine and physical activity or more precise the lack of

vitamin D3. Some years ago a friend told me he had been suffering from lower back pain for 2 years and all he got from his doctors had been pain pills. After following an advice from a friend he added 4000 IU Vitamin D3 daily for three month and lower back pain vanished and no more pain pills needed. He posted in a forum "Why did MY doctors did not tell me about vitamin D3?" . If you look up results of vitamin D3 testing in general population you will find that a quarter or a third of western populations are deficient in vitamin D3 which is an door opener to many diseases

like infections of the upper respiratory tract which is 1.7 fold elevated in depressed people. In my opinion it would enhance the quality of your stress study of LBP

patients if you measure vitamin D3 levels in addition and look for connections

to stress level and pain felt.

I am also using salivary cortisol as a biomarker in my research study for dissertation. I am doing a pilot study soon and will be taking an early morning specimen and an afternoon specimen to look at diurnal rhythm. My confounders are age, sex, pubertal status, smoking status, time of awakening, pregnancy, any medications that would change the concentration of cortisol, Addison's Disease, Cushings Disease, Sexual gender identity and time of last eating or drinking anything except water. Are you familiar with Salimetrics? Where are you sending your samples? Is your study for dissertation or other research? Higher stress levels may not always mean higher cortisol. Have you read about blunting of cortisol? Bruce McEwen has written a ton on Allostatic Load. Take a look at his work. Another one to follow is Doug Granger at Salimetrics.

Susan, in our opinion Salimetrics make the best commercial ELISA kits on the market for cortisol and this was pretty much confirmed in Psychneuroendocrinology early this year. We have our own in-house assays, as well as our Point of Care test, which is the most cost effective option if measuring small sample numbers, or rapid results are requred. Otherwise go with Salimetrics for research

Thanks, that is the plan!

It is frequent or chronic activation of the HPA axis that is more likely associated with subjective stress, since it has been linked to negative health outcomes such as depression, cognitive decline, and cardiovascular disease (see Adam & Kumari, 2009). Thus, a better physiological indicator of stress is one that reflects overloading of the HPA axis. The degree of change in cortisol levels in response to a stressor can reveal such overloading, and has in fact been shown to be a more accurate marker of subjective stress (Campbell & Ehlert, 2012; Vedhara et al., 2003).

The Trier Social Stress Test (TSST) is a laboratory procedure that has been used to reliably induce changes in salivary cortisol (Kirschbau et al, 1993; Weerth et al., 2007). One advantage of the TSST over ecological momentary assessment of cortisol is that the type of stressor and timing of the presentation of a stressor are carefully controlled in a laboratory setting. In addition, the TSST permits for a prolonged reading of a person’s response to a stressor, assessing total salivary cortisol production up to sixty minutes after its introduction.

By the way, if you are looking for a good measure to assess the stress level of your patients, I recently published one:

Description: The Stress Overload Scale (SOS) is comprised of 30 items and is designed to measure “stress overload”, a state described in stress theories as occurring when demands overwhelm resources. The respondent uses a 5‐point Likert scale (1= not at all, 5= a lot) to indicate subjective feelings and thoughts experienced over the prior week. There are two factors underlying overload: Personal Vulnerability and Event Load, which are measured by two distinct but correlated subscales. Even numbered items on the SOS comprise the Event Load, and the odd numbered items (item 5 is reversed scored) comprise the Personal Vulnerability scale; there are also 6 filler items used to discourage negative response sets that are not scored. The scales can be summed to obtain a continuous total score, with higher scores indicating higher levels of stress overload. Or, the subscales can be split at their means to form a four-category diagnostic matrix; those scoring in the High EL-High PV category have been shown to be at the greatest risk for subsequent pathology.

Unique Features: The SOS is the only stress scale that was wholly empirical constructed, through a sequenced series of factor analytic and psychometric studies, all using community samples matched to US Census demographic proportions. It differs from other measures of stress in that it is (1) psychometrically strong--especially in terms of validity; (2) appropriate to community research, due to its brevity and fit to a broad demographic spectrum; (3) unique in its ability to cross-section individuals into risk categories.

Reliability and Validity: The internal consistency of the SOS is excellent (with Cronbach’s alphas > .94 for both subscales and the measure as a whole). Test-retest reliability is good (with coefficients averaging .75 over one week). Construct validity has been demonstrated in significant correlations with other measures of stress and illness; Criterion validity has been shown in the SOS’ ability to predict illness following a stressful event.

Reference: Amirkhan (2012). Stress Overload: A new approach to the assessment of stress. American Journal of Community Psychology, 49, 55-71.

Hi Arnold, if you plan to measure 30 minutes after waking, it sounds like you are going for the Cortisol Awakening Response (CAR). There are a couple of great review papers about what factors influence this response, particularly:

Fries, E., Dettenborn, L., Kirschbaum, C., 2009. The cortisol awakening response (CAR): Facts and future directions. Int. J. Psychophysiol., 72 (1), 67-73.

From this review and my knowledge, menstrual cycle stage will not affect the CAR much, if at all. Oral contraceptive use may or may not make a difference. There are also mixed findings for whether or not age matters. Smoking definitely will raise cortisol levels in general. Also, the CAR will be higher on week-days (work days) vs. weekend/days off. So you will want to standardize which days you get samples from your participants.

Also see:

Chida, Y., Steptoe, A., 2009. Cortisol awakening response and psychosocial factors: A systematic review and meta-analysis. Biol. Psychol., 80 (3), 265-278.

Keep in mind, you should get a sample right after waking as well as 30 minutes later to assess the magnitude of the response. Otherwise, you will not know whether your 30 min measurement represents that typical peak/burst of cortisol people have after waking, or just their baseline morning level.

I agree with your colleague who recommended collecting samples for at least 3 days to get a more stable, "trait" measurement in each person. See, for example:

Hellhammer, J., Fries, E., Schweisthal, O.W., Schlotz, W., Stone, A.A., Hagemann, D., 2007. Several daily measurements are necessary to reliably assess the cortisol rise after awakening: State- and trait components. Psychoneuroendocrinology, 32 (1), 80-86.

Finally, it seems to me that with only 32 participants, you might not have the power to control for all these factors. So your best bet would be to restrict your sample. Perhaps you should focus on a particular age group, stick to women not on (or on) oral contraceptives, and exclude for clinical depression or depression scores above a well-established cut-off.

Hope this helps!

I would also suggest doing a Pilot study. This has helped me tremendously in knowing the logistics of data collection with a population. My pilot so far as 26 participants that we collected all in one day. I am trying to get about 6 more to be able to see if the study is logistically feasible and has some effect that can be determined statistically.

I would like to echo those who mentioned Salimetrics being a good company for salivary cortisol research. Recently, I collected saliva from athletes across a season and performed 22 cortisol assays using Salimetrics collection equipment (SOS & SST) and EIAs. Everything was easy, fast, and low variability. I'm not in any way affiliated with the company, but I like their stuff. Also, thanks everyone for your insights. I didn't ask the original question, but your answers helped me, too.

Hi Arnold,

Have you thought of cortisol in scalp hair measurement? Salviary cortisol is a great way to look at the acute response but has its downsides. Cortisol in hair measurement is a novel tool to measure long term cortisol exposure, which might be relevant for your research question.

regards Erica

In regards to Michele's response, I just reviewed a paper that showed cortisol levels to indeed be affected by oral contraception.

It has also been reported that salivary cortisol levels can be affected by a number of factors, including eating, drinking, smoking, medications, time of saliva collection, and gestational age (Hellhamme et al., 2009).

However, it seems that salivary cortisol levels are most stable later in the day (Kirschbaum et al., 1993), so that would seem the best time for collection.

What a fantastic thread. This is exactly what ResearchGate is good for!

I wonder if anyone has considered the impact of infection upon salivary (or hair) cortisol measurements?

That is an excellent idea. Infection is certainly a form of stress and would show up in saliva and the cortisol levels. Controlling for that may be difficult unless you have a very detailed screening tool to screen out any one with infections.

Does anyone have a clue whether allergies influence cortisol levels too? Do any studies exist?

Thank you for your suggestions. I have modified my experiment to exclude saliva collection because of the financial constraint. I love all the discussion. I will consider this option again when I have funding for my future experiment.

Sorry to hear that. It is expensive. I just paid almost 1,000 for 60 samples to be analyzed. My next study will have 300 samples. The cost will be around 4,000 or more. I was fortunate to receive a grant for the first pilot study and hope to receive one for the next study. Have you been looking for funding?

Since I have to complete my PhD studies on time, I just leave this part behind. I will save it for my future projects.

Perhaps a wise choice Arnold. Best of success in the future.

Dear Arnold, you have to pay attention in which time you will make the saliva collection. It is interesting that your patients has chronic pain and high cortisol levels, we found hypocortisolism related to chronic pain.

It will be interesting you pay attention how long your patients presents the symtoms.

Success in your research.

I agree that you should use multiple samples if at all possible.  Awakening, 30 minutes post-awakening, noon, early evening, and before bedtime.  This can be expensive though unless you are doing the analysis yourself.  Sending them to a lab costs me 14.00 per sample to be analyzed in duplicate.  I would definitely exclude those on birth control and those on any meds that would interfere with cortisol such as asthma meds.  Contacting Dr. Douglas Granger would be a great one to discuss your study with, even if just over email.  He is a very nice man and spoke with me multiple times about my dissertation study.  Do you have a grant to support the cost of the cortisol analysis?  What clinical treatments are you measuring, and what clinical outcomes are you expecting?  Attaching articles that might be helpful.

Been in the lookout for factors affecting cortisol release in animals and guess i just found amazing ideas on this thread. On a differenet note altogether, i would appreciate some clarification, i think fecal material could be assessed of corticosterone hormones to measure stress in animals held in captivity. Is this right? Thank you.

Simon

MSc Wildlife Conservation

University of Chester