The impact of Hormones

When I first started using a continuous glucose monitor (CGM), my blood sugar control looked excellent for the first couple of weeks. Then one day, when I was competing in a K9 Nosework Trial with my dog, my blood sugar was soaring all day long. I assumed it was a stress response from the nerves I always feel when trialing. But the next day, my blood sugar was still spiking easily and staying high for an extended period of time.

Confused, I turned to Google. Realizing that I was in the phase when I usually start experiencing PMS, I googled “PMS insulin resistance,” and sure enough, it was a thing! (Note, this applies to women who are cycling naturally, without hormonal birth control. There’s a whole literature exploring hormonal birth control and its impacts on insulin resistance, and that is a post for another day.)

I’ve noticed this pattern of worsened blood sugar control in myself every single cycle while wearing a CGM. There is some variation in when it starts though. Sometimes it’s as early as 10 days before my period arrives, and sometimes only a week before. There is also variation in just how high the blood sugar spikes might be in response to a certain food. Various factors like stress, sleep, and nutrient status can impact our blood sugar control, so a single reading on a single day is a result of many different factors, not just reproductive hormones.

So what does the literature say?

Much of the literature seems to have been inspired by anecdotal evidence from women with type 1 diabetes who observe more incidents of hyperglycemia (high blood sugar) during the luteal phase of their cycles (To refresh on cycle phases, check out this post and the figure below.). When I dug into the literature, I was honestly a bit shocked and disappointed at how the earlier studies were designed. Scientists studied up to 13 women per study across a single cycle with a single measurement of insulin resistance in the follicular phase and a single measurement in the luteal phase (Bingley et al. 2008, Blum et al. 2005, Escalante Pulido and Alpizar Salazar 1999, Scott et al. 1990, Valdes and Elkind-Hirsch 1990).

For those without a statistics background, when we test to see if averages differ across two groups of people, two cycle phases, etc., we use something called a t-test. The bigger the t-test, the more confident we are that a true difference exists between the groups, and there is a threshold t-test size above which we say there is a “statistically significant” difference. This usually means we are 90 or 95% confident (or more) that there is a difference. T-tests are smaller when there is:

1) A small difference in observed averages
2) A smaller sample size
3) More variation in measurements across observations

So, if we have a small sample size for a measurement with a lot of variation in it both across people and across the same person from day to day, our chances of getting a big t-test, and thus, a “statistically significant” difference is really small. In my field of economics, we would consider this kind of study to be “under powered.” You haven’t designed a study with enough statistical power to be able to observe the difference you are testing for, amongst all the noise in the data.

So unsurprisingly, the differences observed across cycle phases were not statistically significant in these studies. And as happens all too often with women’s health, we assume the women must be crazy. Or it must just be some sort of coincidence. (Check out Doing Harm by Maya Dusenbery if you want to read about how prevalent this is.)

But is this actually the case?

What does the high quality literature say?

Two studies were designed in such a way to actually answer our question! (Thank goodness!)

One study (Yeung et al. 2010) recruited 259 premenopausal women. They attempted to follow all of them for two cycles, and were able to do so for 250 of them. They collected measurements on 8 different days (menses, mid-follicular, late-follicular, LH/FSH surge, ovulation, early-luteal, mid-luteal, and late-luteal) across the menstrual cycle. On each of these days, they collected fasting insulin and glucose levels to calculate the homeostasis model of insulin resistance (HOMA-IR). The higher HOMA-IR is, the more insulin resistant you are. They also measured estradiol (a form of estrogen), progesterone, sex hormone binding globulin (SHBG), luteinizing hormone, and follicle stimulating hormone.

As we can see in the graph, their measure of insulin resistance is, indeed, higher, in the luteal phase. This difference is small, but, due to their larger sample size, it is statistically significant; that is to say that we can be pretty confident that a difference exists, on average, because we were able to observe it across a large number of women.

Their study focused only on healthy women, and excluded women with diabetes, PCOS, or other chronic health conditions. If healthy women experience a shift in insulin sensitivity, it seems plausible that women with some degree of insulin impairment might as well, and this shift could potentially be magnified by any existing insulin resistance.

Additionally, their study looked at correlations between hormones and insulin sensitivity. They find that both estradiol and progesterone are positively correlated with insulin resistance; higher levels of these hormones coincided with higher levels of insulin resistance. While progesterone has been hypothesized as the culprit behind luteal phase insulin resistance, there is conflicting evidence on estrogen (more on this in a later post). Conversely, lower levels of SHBG were associated with higher levels of insulin resistance. SHBG tends to be low in women with PCOS.

A second study sheds even more light on this topic (Brown et al. 2015). This study only recruited 12 subjects, but the subjects wore continuous glucose monitors for three complete cycles to collect ample observations for each participant. This study included only women with type 1 diabetes, and it tracked rates of high and low blood sugar as well as a measure of insulin sensitivity (note, the reverse of HOMA-IR discussed above).

Like the previous study, they found that insulin resistance increased, and insulin sensitivity decreased, on average, during the luteal phase of the women’s cycles. As observed by previous women with type 1 diabetes, their subjects had higher probabilities of hyperglycemia events during the luteal phase, and this difference was statistically significant. Interestingly, in their data, the highest risk of hyperglycemia occurred in the early luteal phase, whereas my issues start in the mid-luteal phase and continue through the late luteal phase.

Accordingly, this study considers individual variation as well as just looking at sample averages and trends. 9 out of their 12 subjects experienced decreased insulin sensitivity during the luteal phase of at least 2 of the 3 cycles. However, 3 women experienced different patterns and some of the 9 had one anomalous cycle that didn’t fit the pattern of their other two cycles. This is not entirely unexpected given the myriad other factors that influence blood sugar regulation as well as the complexity of female hormone dynamics across the menstrual cycle. I appreciate that the study authors provided a discussion of the observed variations instead of clinging to their overarching “statistically significant” findings.

So what does all of this tell us??

Menstrual phase and insulin resistance- Should we care?

On average, there is reason to believe that at least some women may be less insulin sensitive/more insulin resistant during their luteal phases. This means that a snack or meal that would not cause a big spike in blood sugar at other times in your cycle might do so during your luteal phase. For a lot of women, this occasional spike is not a big deal.

For women with PCOS, there are two factors that make this more concerning.

First, studies have demonstrated that women with PCOS take longer to clear insulin from their bodies (Peppard et al. 2001). A big spike in blood sugar, followed by a big spike in insulin, will leave large amounts of insulin in the bodies of women with PCOS for longer, further contributing to insulin resistance.

Second, there is preliminary evidence that the ovaries of women with PCOS are more sensitive to the presence of insulin (Nestler et al. 1998). In all women, insulin stimulates the release of androgens like testosterone from the ovaries; in women with PCOS, the ovaries may release higher levels of androgens. In my own experience, reducing carbohydrate intake in my late luteal phase has reduced the overt signs of excess androgens, suggesting that managing my blood sugar and resulting insulin better during this phase is decreasing the release of these excess androgens.

Now I know your next questions are:

How do I know what is happening across my cycle??

And of course, what should I do about this??

Part III of this series is going to dig into all of the ways to improve your insulin sensitivity, so sit tight!

As for the first question, that’s a little tricky. For starters, I am not a physician, so you should talk to your doctor if you have major concerns about insulin resistance. If you want to experiment on your own in the lowest cost way, you can play around with the amount of carbohydrates you consume in the luteal phase and see if that makes a difference in any of your symptoms or how you feel. I don’t recommend going as extreme as following a ketogenic diet to start. Much of the literature on the impact of keto on insulin resistance only considers men and/or postmenopausal women. Instead, you might experiment with cutting out a high density carb source or two (potatoes, grain, bread) and replace it with lower density carbs (any vegetable you can think of!) when you suspect you might have lowered insulin sensitivity. For me, the telltale signs of PMS are about when my insulin resistance kicks in, so that might be a good place to start. If you notice improvements in how you feel or your symptoms, stick with it. If you don’t notice any changes or if you feel tired and sluggish, that probably wasn’t the right adjustment for you. Add those carbs back in!

If you really want detailed answers, continuous glucose monitors are available. I love the Signos program and app. This will let you determine exactly what works for your body and how this might vary with your cycle (as well as with your stress level, sleep duration, etc.). It’s an investment but it’s been a game-changer in the management of PCOS for me. I plan on using CGMs with my future clients if they are willing to wear one. If you want to chat more about them, please reach out!

References

Bingley CA, Gitau R, and Lovegrove JA. 2008. Impact of Menstrual Cycle Phase on Insulin Sensitivity Measures and Fasting Lipids. Hormone and Metabolic Research 40(12):901-906.

Blum CA, Mueller B, Huber P, Kraenzlin M, Schindler C, de Geyter C, Keller U, and Puder JJ. 2005. Low-grade Inflammation and Estimates of Insulin Resistance during the Menstrual Cycle in Lean and Overweight Women. The Journal of Clinical Endocrinology & Metabolism 90(6):3230-3235.

Brown SA, Jiang B, McElwee-Malloy M, Wakerman C, and Breton MD. 2015. Fluctuations of Hyperglycemia and Insulin Sensitivity Are Linked to Menstrual Cycle Phases in Women with T1D. Journal of Diabetes, Science and Technology 9(6):1192-1199.

Nestler JE, Jakubowicz DJ, Falcon de Vargas A, Brik C, Quintero N, and Medina F. 1998. Insulin Stimulates Testosterone Biosynthesis by Human Thecal Cells from Women with Polycystic Ovary Syndrome by Activating Its Own Receptor and Using Inositolglycan Mediators as the Signal Transduction System. The Journal of Clinical Endocrinology & Metabolism 83(6):2001-2005.

Peppard HR, Marfori J, Iuorno MJ, and Nestler JE. 2001. Prevalence of Polycystic Ovary Syndrome among Premenopausal Women with Type 2 Diabetes. Diabetes Care 24:1050–1052.

Escalante Pulido JM and Alpizar Salazar M. 1999. Changes in Insulin Sensitivity, Secretion and Glucose Effectiveness During Menstrual Cycle. Archives of Medical Research 30(1):19-22.

Scott AR, Macdonald IA, Bowman CA, and Jeffcoate WJ. 1990. Effect of Phase of Menstrual Cycle on Insulin Sensitivity, Peripheral Blood Flow and Cardiovascular Responses to Hyperinsulinaemiea in Young Women with Type 1 Diabetes. Diabetes Medicine 7(1):57-62.

Valdes CT and Elkind-Hirsch KE. 1990. Intravenous Glucose Tolerance Test-Derived Insulin Sensitivity Changes during the Menstrual Cycle. The Journal of Clinical Endocrinology & Metabolism 72(3):642-646.

Yeung EH, Zhang C, Mumform SL, Ye A, Trevisan M, Chen L, Browne RW, Wactawski-Wende J, and Schistermen EF. 2010. Longitudinal Study of Insulin Resistance and Sex Hormones over the Menstrual Cycle. The BioCycle Study. The Journal of Clinical Endocrinology & Metabolism 95(12):5435-5442.

Disclaimer: This post is not intended to diagnose or treat any medical issues. It is intended for informational purposes only. I am not a medical practitioner. Always consult a trusted healthcare provider with any questions you may have about a medical condition or treatment and before starting any new health care regimen.