Insulin Resistance in Polycystic Ovary Syndrome (PCOS): Evaluation and Important Associations

Abstract

PCOS is common, and as understanding evolves to recognize many diverse presentations, prevalence of PCOS continues to increase. There are clear associations of tremendous interest to the clinician between PCOS and an array of common patient complaints, and insulin resistance seems to play a key role in most such presentations. The gold-standard tool for assessing insulin sensitivity, the euglycemic-hyperinsulinemic clamp technique is reserved for research purposes and is not broadly available as a diagnostic test. This review will showcase associations between PCOS, insulin resistance, and several important common patient concerns. The review will also attempt to determine if an adequate alternative to the euglycemic-hyperinsulinemic clamp technique exists for accurate determination of insulin resistance in outpatient practice.

Introduction

The prevalence of polycystic ovarian syndrome (PCOS) is high. Insulin resistance is present in PCOS at a higher rate than in the general population, therefore, it benefits the clinician to understand insulin resistance in the context of PCOS. This article will review the prevalence of insulin resistance, the clinical outcomes associated with it, and the measurement of insulin resistance specifically within patients who meet the diagnostic criteria for PCOS. Given the ability of insulin resistance to predict many negative outcomes or comorbidities in patients with PCOS, in-office tools to accurately determine insulin resistance would be very valuable.

Prevalence and Severity of Insulin Resistance in PCOS

Insulin resistance is present in most cases of PCOS. A 2016 meta-analysis, which looked exclusively at results of euglycemic-hyperinsulinemic clamp studies including 1224 cases and 741 controls, found that insulin sensitivity was 27% lower in women with PCOS. They did not report on the rate of insulin resistance among PCOS cases, nor distinguish the severity of insulin resistance at different BMIs. Elevated BMI had a greater effect on insulin sensitivity in cases than in controls. BMI had a greater impact on insulin sensitivity than did PCOS (Cassar et al 2016).

Another meta-analysis conducted in 2017 found that 74.9% of those meeting Rotterdam criteria for PCOS were insulin resistant as determined by euglycemic-hyperinsulinemic clamp. The percentage of insulin resistant subjects was 59.3% in normal-weight, 77.5% in overweight, and 93.9% in obese patients(Rezvanian et al 2009).

A study published after these meta-analyses, using euglycemic-hyperinsulinemic clamp, which had 448 cases and 40 controls, found a rate of insulin resistance of 56.3% in women with PCOS. Insulin resistance was present in 18.8% of lean women with PCOS. This study included Chinese women only(Li et al 2019).

A 2019 meta-analysis found that in nonobese persons with PCOS, the odds ratio of insulin resistance was 5.70 (1.46-22.32). However, the study did not make clear how insulin resistance was measured. Only three studies met their criteria for analysis of insulin resistance(Zhu et al 2019).

The quality of the information about prevalence and severity of insulin resistance in PCOS is low relative to the worldwide incidence of the syndrome. From the available studies, we can conclude that most people with PCOS are insulin resistant.

Androgen Levels and Insulin Sensitivity

A small amount of evidence links severity of insulin resistance with severity of hyperandrogenism.

In a small study of 27 people with PCOS, testosterone levels were associated with the Matsuda index of insulin sensitivity, independent of BMI(Luotola et al 2018).

A larger study of 378 PCOS subjects found that free androgen index (FAI) was negatively related to HOMA-IR r = -0.413 p <0.001. Across tertiles of FAI, insulin sensitivity decreased(Zhang et al 2018).

Similarly, in a study of 1000 PCOS cases, FAI was significantly correlated with fasting insulin (FIN) and HOMA-IR (r = 0.240, P < 0.001 and r = 0.191, P < 0.001, respectively). Free testosterone was significantly correlated with FIN after adjusting for the influence of age(Zhang et al 2019).

In a study of 448 cases and 40 controls, insulin resistance as determined by euglycemic-hyperinsulinemic clamp was associated with testosterone levels, but insulin resistance as determined by HOMA-IR was not (Li et al 2019).

To summarize; when PCOS gets worse, IR gets worse.

This small body of evidence does not explain the nature of the relationship between insulin resistance and hyperandrogenism, which remains an important nut to crack in understanding PCOS.

Insulin Resistance in PCOS and Outcomes of Interest

Some of the body of evidence exploring PCOS has examined the relationship between outcomes of interest to patients and insulin resistance. These studies all use surrogate markers of insulin resistance, as opposed to euglycemic-hyperinsulinemic clamp studies, which means they are likely underestimating prevalence and severity of insulin resistance, as will be discussed below. Studies looking at incidence of metabolic syndrome in PCOS have not been included here, as they fail to report specifically on insulin sensitivity of their subjects.

Fertility

In 1000 PCOS cases, the odds ratio of live birth was 1.81 in the group with HOMA-IR <2.69 compared to HOMA-IR >/= 2.69 (1.26, 2.55, p= 0.001). Live birth is the ultimate outcome for fertility. However, this study also found statistically significant correlations between fasting insulin and ovulation per ovulation induction cycle, conception, and pregnancy. No associations were found between fasting insulin or HOMA-IR and miscarriage (Zhang et al 2019).

Depression 

PCOS patients are at a three to four-times greater risk of depression and five to six-times greater risk of anxiety than those without PCOS (Cooney et al 2017, Greenwood et al 2018).In one study, there was a significant association between HOMA-IR and depression, but no association between HOMA-IR and anxiety. In a study of 738 PCOS cases, HOMA-IR >2.2 was associated with a 2.3-fold increased risk of depression (Greenwood et al 2018).

Sleep Apnea

A systematic review including six studies found that subjects with PCOS and obstructive sleep apnea (OSA) had higher HOMA-IR than those with PCOS but without OSA(Kahal et al 2018).

Hirsutism

A controlled trial of laser hair removal compared to laser hair removal plus metformin 500mg tid found that in the metformin group, HOMA-IR decreased from 2.83 to 1.35, and that the metformin group had superior response to laser hair removal(Rezvanian et al 2009).

Bone Mineral Density

Bone mineral density appears to be an outlier; worse insulin sensitivity is associated with higher bone mineral density. However, in the single study examining insulin sensitivity in PCOS, the controls had a very high mean HOMA-IR of 3.6 (Pereira-Eshraghi et al 2019).

Measuring Insulin Resistance in Practice

The high prevalence of insulin resistance in PCOS, as well as its association with hyperandrogenism and with outcomes of clinical significance, demonstrates the desirability of being able to measure insulin resistance in clinical practice. To this end, we will review the reliability of surrogate markers of insulin resistance in the context of PCOS. We will additionally attempt to find any useful cutoffs for those markers.

Seeing average HOMA-IR scores in PCOS compared to healthy controls can help orient us. A meta-analysis compared HOMA-IR in 3037 subjects divided into four groups: obese/PCOS, non-obese/PCOS, obese/non-PCOS and non-obese/non PCOS. The pooled mean of HOMA IR in each group was 4.38 (3.84, 4.92), 2.68 (2.16, 3.20), 2.44 (2.06, 2.82) and 1.34 (1.06, 1.63) respectively. The difference between non-obese/PCOS and obese/non-PCOS was not statistically significant(Behboudi-Gandevani et al 2016).

HOMA-IR is relatively simple for a clinician to calculate. Fasting insulin (mU/L) is multiplied by fasting glucose (mg/dL). A value between 0.5 and 1.4 denotes good insulin sensitivity. Higher scores denote increasingly poor insulin sensitivity (The Blood Code 2022). This simple calculation, which can be easily performed in-office, provides a decent marker of insulin sensitivity, yet unfortunately, as reviewed below, significantly underestimates insulin resistance relative to the euglycemic-hyperinsulinemic clamp technique.

A 2019 study of 448 cases and 40 controls comparing the euglycemic-hyperinsulinemic clamp to HOMA-IR found the Kappa value to be 0.069 for lean PCOS subjects and 0.139 for obese PCOS subjects, where a Kappa value of zero indicates no relationship. The prevalence of IR estimated by HOMA-IR was lower than that determined by clamp technique. By clamp technique 34.5% of lean PCOS subjects and 80.2% of obese PCOS subjects were insulin resistant. Using HOMA-IR, 15.6% of lean PCOS subjects and 41.6% of obese PCOS subjects were insulin resistant(Li et al 2019).

Another 2019 study with 537 subjects found a much more promising correlation coefficient of r= -0.6828 when comparing the euglycemic-hyperinsulinemic clamp and HOMA-IR results(Long et al 2019).

The 2017 Tosi et al study that found a high rate of 74.9% insulin resistance in PCOS subjects also looked at the euglycemic-hyperinsulinemic clamp compared to surrogate indices. The percentage of PCOS subjects with IR was lower as determined by HOMA-IR than by clamp, at 41.1%. The correlation coefficient of HOMA-IR and the euglycemic-hyperinsulinemic clamp was r= 0.622 p<0.001. Going a step further, the authors reported positive and negative predictive values, concluding that for all surrogate markers, specificity for detecting insulin resistance was fair but sensitivity was low(Tosi et al 2017).

There are a couple of smaller studies as well, both specifically in adolescents.

A 2018 study of 28 adolescents with PCOS found the following correlation coefficients with comparison to the euglycemic-hyperinsulinemic clamp: HOMA-IR r= -0.78, p<0.001, e-IS r= 0.70, p<0.001, Matsuda r= 0.533, p<0.001 (Cree-Green et al 2018).

A 2020 study comparing the euglycemic-hyperinsulinemic clamp to various surrogate indices in 26 adolescents with PCOS found the following correlation coefficients: oral minimal model estimation of insulin sensitivity r=0.64, p<0.0001, oral minimal model estimation of dynamic insulin sensitivity r=0.73 p <0.0001, Matsuda index r=0.59 p<0.003(Carreau et al 2020).

Beauty is in the eye of the beholder; a desirable correlation coefficient in this case lies with the judgment of the clinician. There is considerable variability. The worst result was in lean PCOS subjects. It is probably safe to say that surrogate markers should not be used to rule out insulin resistance, since all surrogate markers appear to underestimate the rate of insulin resistance.

Other Possibilities

Fortunately, the search for a solution for the clinician looking to understand the level of insulin sensitivity in a particular patient continues.

Inferring insulin resistance based on sex hormone binding globulin levels may prove to be an acceptable workaround. In a study of 100 PCOS patients and 61 controls, in the non-overweight group, a SHBG level of < 38.4 nmol/L predicted HOMA-IR >2.29 with 92.86% sensitivity (95% CI: 82.7–98.0%) and 74.39% specificity (95% CI: 63.6–83.4%) (Chen et al 2021).

Many diverse molecules are being investigated for their utility in identifying insulin resistance, although none are sufficiently well-established for clinical use yet (Polak et al 2017).

Conclusion

Insulin resistance is present in most patients with PCOS, and strongly predicts the presence and severity of common comorbidities associated with PCOS. It would be beneficial if the clinician could accurately assess insulin resistance in individual patients in order to treat only those who require it, and to gauge improved risk of various undesirable outcomes. HOMA-IR, although widely used in studies looking at outcomes of importance to PCOS patients, correlates moderately with insulin resistance determined by the euglycemic-hyperinsulinemic clamp, but does not correlate perfectly. It certainly underestimates insulin resistance, and may underestimate it drastically. Other surrogate markers perform similarly and are less frequently used in studies of clinical outcomes. Researchers are working diligently to provide a better solution.

References

Behboudi-Gandevani S, Ramezani Tehrani F, Rostami Dovom M, Farahmand M, Bahri Khomami M, Noroozzadeh M, Kabir A, Azizi F. Insulin resistance in obesity and polycystic ovary syndrome: systematic review and meta-analysis of observational studies. Gynecol Endocrinol. 2016;32(5):343-353.

Carreau AM, Xie D, Garcia-Reyes Y, Rahat H, Bartlette K, Behn CD, Pyle L, Nadeau KJ, Cree-Green M. Good agreement between hyperinsulinemic-euglycemic clamp and 2 hours oral minimal model assessed insulin sensitivity in adolescents. Pediatr Diabetes. 2020 Nov;21(7):1159-1168.

Cassar S, Misso ML, Hopkins WG, Shaw CS, Teede HJ, Stepto NK. Insulin resistance in polycystic ovary syndrome: a systematic review and meta-analysis of euglycaemic-hyperinsulinaemic clamp studies. Hum Reprod. 2016 Nov;31(11):2619-2631.

Chen F, Liao Y, Chen M, Yin H, Chen G, Huang Q, Chen L, Yang X, Zhang W, Wang P, Yin G. Evaluation of the Efficacy of Sex Hormone-Binding Globulin in Insulin Resistance Assessment Based on HOMA-IR in Patients with PCOS. Reprod Sci. 2021 Sep;28(9):2504-2513.

Cooney LG, Lee I, Sammel MD, Dokras A. High prevalence of moderate and severe depressive and anxiety symptoms in polycystic ovary syndrome: a systematic review and meta-analysis. Hum Reprod. 2017 May 1;32(5):1075-1091.

Cree-Green M, Cai N, Thurston JE, Coe GV, Newnes L, Garcia-Reyes Y, Baumgartner AD, Pyle L, Nadeau KJ. Using simple clinical measures to predict insulin resistance or hyperglycemia in girls with polycystic ovarian syndrome. Pediatr Diabetes. 2018 Dec;19(8):1370-1378.

Greenwood EA, Pasch LA, Cedars MI, Legro RS, Eisenberg E, Huddleston HG; Eunice Kennedy Shriver National Institute of Child Health and Human Development Reproductive Medicine Network. Insulin resistance is associated with depression risk in polycystic ovary syndrome. Fertil Steril. 2018 Jul 1;110(1):27-34.

Kahal H, Kyrou I, Uthman O, Brown A, Johnson S, Wall P, Metcalfe A, Tahrani AA, Randeva HS. The association between obstructive sleep apnea and metabolic abnormalities in women with polycystic ovary syndrome: a systematic review and meta-analysis. Sleep. 2018 Jul 1;41(7).

Li W, Chen Q, Xie Y, Hu J, Yang S, Lin M. Prevalence and degree of insulin resistance in Chinese Han women with PCOS: Results from euglycemic-hyperinsulinemic clamps. Clin Endocrinol (Oxf). 2019 Jan;90(1):138-144.

Long J, Li L, Wang F, Yang G, Cheng W, Wei J, Chen M, Liu D. Screening for a Simple and Effective Indicator of Insulin Resistance in Chinese Reproductive-Aged Women, with the Insulin Clamp Technique as a Reference. Metab Syndr Relat Disord. 2019 Oct;17(8):423-429.

Luotola K, Piltonen TT, Puurunen J, Morin-Papunen LC, Tapanainen JS. Testosterone is associated with insulin resistance index independently of adiposity in women with polycystic ovary syndrome. Gynecol Endocrinol. 2018 Jan;34(1):40-44.

Pereira-Eshraghi CF, Chiuzan C, Zhang Y, Tao RH, McCann M, Neugut YD, Printz A, Fennoy I, Cree-Green M, Oberfield SE, Sopher AB. Obesity and Insulin Resistance, Not Polycystic Ovary Syndrome, Are Independent Predictors of Bone Mineral Density in Adolescents and Young Women. Horm Res Paediatr. 2019;92(6):365-371.

Polak K, Czyzyk A, Simoncini T, Meczekalski B. New markers of insulin resistance in polycystic ovary syndrome. J Endocrinol Invest. 2017;40(1):1-8.

Rezvanian H, Adibi N, Siavash M, Kachuei A, Shojaee-Moradie F, Asilian A. Increased insulin sensitivity by metformin enhances intense-pulsed-light-assisted hair removal in patients with polycystic ovary syndrome. Dermatology. 2009;218(3):231-236.

The Blood Code. https://thebloodcode.com/homa-ir-calculator/. Accessed February 2022. 

Tosi F, Bonora E, Moghetti P. Insulin resistance in a large cohort of women with polycystic ovary syndrome: a comparison between euglycaemic-hyperinsulinaemic clamp and surrogate indexes. Hum Reprod. 2017 Dec 1;32(12):2515-2521.

Zhang B, Wang J, Shen S, Liu J, Sun J, Gu T, Ye X, Zhu D, Bi Y. Association of Androgen Excess with Glucose Intolerance in Women with Polycystic Ovary Syndrome. Biomed Res Int. 2018 Mar 8;2018:6869705.

Zhang D, Yang X, Li J, Yu J, Wu X. Effect of hyperinsulinaemia and insulin resistance on endocrine, metabolic and fertility outcomes in women with polycystic ovary syndrome undergoing ovulation induction. Clin Endocrinol (Oxf). 2019 Sep;91(3):440-448.