Assessing insulin sensitivity. (Controlling PCOS, part 1).(polycystic ovary disease)
Author/s: Robert P. Kauffman
While experts continue to debate the exact causes of PCOS, there’s little doubt that insulin resistance plays an important role in its pathophysiology. In the first installment of our series on PCOS, two authorities offer a detailed yet practical discussion on how to evaluate insulin sensitivity.
Most first-year medical students know that the inability of a patients tissues to respond to insulin is the hallmark of Type 2 diabetes, a fact that came to light in 1938, and was later confirmed with development of an insulin immunoassay in 1960. It wasn’t until 1980, however, that the relationship between insulin resistance and polycystic ovary syndrome (PCOS) was clearly established. (1-3)
The sensitivity of cells to insulin is determined by the response of circulating glucose to endogenous or exogenous insulin. Insulin resistance is characterized by a diminished glucose response to the metabolic actions of the hormone, or expressed in another way, it’s the inverse of insulin sensitivity. (4) And although other metabolic and physiologic parameters like lipid metabolism, vascular endothelial function, and gene expression are partially regulated by insulin action, these parameters are not considered when assessing insullin sensitivity. (4)
Of course, clinicians should keep in mind that a patient who has hyperinsulinemia is not necessarily insulin resistant. Serum insulin levels are primarily dependent on the amount of hormone released by the pancreatic [beta]-cell and the rate at which the hormone is removed from plasma–the plasma clearance rate. Other factors that come into play in this equation are obesity and fasting glucose concentrations, which both affect basal insulin secretion. (5)
In the United States, PCOS is usually defined by (1) the presence of hyperandrogenism, (2) six or fewer menstrual cycles a year, and (3) the absence of other endocrine disorders that could contribute to ovulatory dysfunction, including thyroid disease, hyperprolactinemia, Cushing’s disease, and nonclassic congenital adrenal hyperplasia. (6)
Although the presence of “polycystic ovaries” on ultrasound is commonly included by European investigators in the definition of PCOS, US clinicians have not widely embraced this diagnostic criteria for two reasons: Approximately 20% of women who ovulate normally have several subcortical ovarian follicles on ultrasonography; and about 20% of PCOS patients do not have multiple small ovarian cysts on their ultrasound scans. (7) Without an international consensus on the definition of PCOS, it is difficult to compare epidemiologic data provided in the medical literature.
About 50% to 70% of all women with PCOS have some degree of insulin resistance, or insulin insensitivity, and research has shown that insulin resistance probably contributes to the hyperandrogenism that is responsible for the symptoms of PCOS. The precise incidence of the syndrome, however, is probably dependent on how the syndrome is defined and the method used to detect and quantify insulin resistance. (8)
Despite the uncertain statistics, some investigators have suggested that physicians consider all women with PCOS as insulin resistant, which would make screening for insulin resistance unnecessary. (9) Unfortunately, this recommendation does not take into consideration the large differences in insulin sensitivity among women with PCOS. Nor does it take into account the fact that some patients with PCOS do not respond to metformin, which improves insulin sensitivity. (10) And finally, the suggestion to label all women with PCOS as insulin resistant fails to factor in ethnicity. Studies indicate, for example, that Caribbean-Hispanic, Mexican American, Indian, Mori, Pacific Island, and South Asian patients have greater degrees of insulin resistance than matched Caucasian populations. (11-14)
The value of measuring insulin sensitivity
The concept of insulin resistance is relatively easy to understand, but determining precisely who is insulin resistant is more complicated. The relationship between glucose and insulin is quite complex and involves the interaction of many metabolic and regulatory factors. (15) Normal insulin sensitivity varies widely and is influenced by age, ethnicity and obesity Simply put, not all people with impaired insulin sensitivity are necessarily suffering from a disorder, and pregnancy is a perfect example of this. (16-18
A World Health Organization consensus group recently concluded that the insulin sensitivity index ([S.sub.I]) of the lowest 25% of a general population can be considered insulin resistant. (19) The European Group for the Study of Insulin Resistance took a more restricted view, defining insulin resistance as the [S.sub.I] of the lowest 10% of a non-obese, nondiabetic, normotensive Caucasian population. (20) Richard Legro and his associates also used the [S.sub.I] of the lowest 10% of an obese, non-PCOS population to define insulin resistance. (21) Ideally we should be deriving the normal [S.sub.I] range from a population of women who are not obese, have regular menstrual cycles, are not suffering from hirsutism, and have normal circulating androgen levels.
Considering the important role that insulin evidently plays in PCOS, obstetrician/gynecologists may find it useful to evaluate insulin sensitivity quantitatively Having those data in hand will allow you to select the most appropriate candidates for insulin sensitizing drugs. It will also help in efforts to assess a patient’s clinical response to weight loss, insulin sensitizers, or other treatments. And finally, identifying insulin-resistant patients may assist in planning interventions for patients at high risk for Type 2 diabetes, hypertension, dyslipidemia, and cardiovascular disease, all of which have been linked to insulin resistance.
Choosing the best assessment technique
The hyperinsulinemic-euglycelmic clamp technique is the most scientifically sound technique for measuring insulin sensitivity, and it’s against this standard that all other tests are usually compared. Because this and similar “clamp” techniques are expensive, time consuming, and labor intensive, they are not very practical in an office setting.
To overcome these obstacles, alternative tests have been developed, including the frequently sampled IV glucose tolerance test (FSlVGTT), insulin tolerance test (ITT), insulin sensitivity test (1ST), and continuous infusion of glucose with model assessment (CIGMA) (Table 1). Unfortunately, all of these methods require IV access and multiple venipunctures, making them relatively impractical for office assessment. The oral glucose tolerance test (OGTT) does not require IV access but does involve several venipunctures and 2 to 4 hours of patient and technician time. Each of these tests has been shown to correlate reasonably well with dynamic clamp techniques. (8,17)
In the quest for a noninvasive measurement of insulin sensitivity, several fasting or “homeostatic” models have been proposed as well, and each has also correlated reasonable well with clamp techniques. (21-23) The fasting insulin level ([l.sub.o]), homeostatic model assessment (HOMA), fasting glucose/insulin ratio (G/I ratio), and quantitative insulin sensitivity check index (QUICKI) have been the most frequently used techniques in clinical investigations. The fact that these tests only require a single venipuncture in the fasting state and don’t call for concomitant IV access makes them particularly attractive to patients and clinicians alike.
When considering these approaches, keep in mind the units of measure used to report data. Insulin can be reported in [micro]/mL by conventional units (C.U or pmol/L by the system internationale (S.I.). On the other hand, serum glucose is usually expressed in mg/dL (C.U.) or mmol/L (S.I.). To convert insulin readings in [micro]U/mL to pmol/L, multiply by 7.175. To convert in the other direction, multiply by 0.139. To convert glucose values in mg/dL to mmol/L, multiply by 0.05551. To convert in the other direction, multiply by 18.
Clamp techniques and insulin infusion tests
Hyperinsulinemic-euglycemic clamp. The gold standard for evaluating insulin sensitivity, this “clamp” technique requires a steady IV infusion of insulin to be administered in one arm. The serum glucose level is “clamped” at a normal fasting concentration by administering a variable IV glucose infusion in the other arm. (8,16,21,24) Numerous blood samplings are then taken to monitor serum glucose so that a steady “fasting” level can be maintained. (In theory the IV insulin infusion should completely suppress hepatic glucose production and not interfere with the test’s ability to determine how sensitive target tissues are to the hormone.) The degree of insulin resistance should be inversely proportional to the glucose uptake by target tissues during the procedure. In other words, the less glucose that’s taken up by tissues during the procedure, the more insulin resistant a patient is.
A variation of this technique, the hyperinsulinemic-hyperglycemic clamp provides a better measurement of pancreatic beta cell function but is less physiologic than the euglycemic technique. (17)
Insulin sensitivity test (IST). IST involves IV infusion of a defined glucose load and a fixed-rate infusion of insulin over approximately 3 hours. Somatostatin may be infused simultaneously to prevent insulin secretion, inhibit hepatic gluconeogenesis, and delay secretion of counter-regulatory hormones–particularly glucagon, growth hormone, cortisol, and catecholamines. Fewer blood samples are required for this test, compared to clamp techniques. The mean plasma glucose concentration over the last 30 minutes of the test reflects insulin sensitivity Although lengthy, IST is less labor intensive than clamp techniques and the FSIVGTT. (17)
Insulin tolerance test (ITT). A simplified version of 1ST, ITT measures the decline in serum glucose after an IV bolus of regular insulin (0.1-0.5 U/kg) is administered. Several insulin and glucose levels are sampled over the following 15 minutes (depending on the protocol used). The ITT primarily measures insulin-stimulated uptake of glucose into skeletal muscle. Because this test is so brief, there’s very little danger of counter-regulatory hormones interfering with its results. (17,25)
IV access should be established for insulin injection, blood sampling, and for rapid administration of [D.sub.50] W should severe hypoglycemia occur. Normal values for women with PCOS have not been published to date, but normal ranges for insulin sensitivity in a general population have been published for persons with a body mass index below 30 kg/[m.sup.2] and for obese subjects (BMI >30 kg/[m.sup.2]) at 0.026 to 0.085 mmol/L * [minute.-1] and 0.0 12 to 0.0 17 mmol/L * [minute.sup.-1] respectively. (25) These values reflect the rate of decline of log transformed glucose values.
Taking the minimalist approach
“Minimal” models require IV or oral administration of glucose only, unlike studies we discussed previously, which require IV insulin. Frequently sampled IV glucose tolerance tests (FSIVGTT). This method is less labor intensive than clamp techniques yet still requires as many as 25 blood samples over a 3-hour period, and a computer-assisted mathematical analysis. (17,21) Several variations of the FSIVGTT have been published. One recently published study infused 0.3 g/kg of glucose over 1 minute, followed by tolbutamide 500 mg IV at 20 minutes into the glucose infusion. The [S.sub.1] was calculated by a computer-based program. (21) Tolbutamide administration can also be used during FSIVGTT to augment endogenous insulin secretion and is particularly useful in women with diabetes. (26)
Continuous infusion of glucose with model assessment (CIGMA).
Like ITT, CIGMA requires fewer venipunctures and is less laborious than clamp techniques. A constant IV glucose infusion is administered, and samples for glucose and insulin are drawn at 50, 55, and 60 minutes. A mathematical model is then used to calculate [S.sub.1] The results are reasonably compatible with clamp techniques; however, few laboratories have used CIGMA for insulin sensitivity testing in diabetic patients and there is no substantive data using the CIGMA technique in women with PCOS. (17)
Oral glucose tolerance test (OGTT). OGTT, a mainstay in the diagnosis of impaired glucose tolerance (IGT) and diabetes mellitus in pregnant and nonpregnant women, may be used to assess insulin sensitivity as well. Because no IV access is needed, OGTT is better suited for assessment of large populations than the other techniques we outlined. A modified OGTT that uses a 75- or 100-g glucose load and measures glucose and insulin at various intervals over 2 to 4 hours has been used in clinical studies. (12,24,27-29)
Like other minimal approaches to diagnosis, OGTT provides information on beta cell secretion and peripheral insulin action, and various mathematical equations have been used to provide an [S.sub.1] value. (24,27,30-31) Insulin sensitivity has been assessed by calculating insulin area under the curve ([AUC.sub.insulin]), [AUC.sub.glucose])/[AUC.sub.insulin], and by an insulin sensitivity index ([ISI.sub.0,120]) that applies only the glucose and insulin values from 0 and 120 minutes into a complex mathematical formula. (21,24,27,31)
A 2-hour G/I ratio has also been reported to give a reasonably accurate [S.sub.1], with a ratio of [less than or equal to]1.0 suggestive of insulin resistance in a PCOS population. (21) Insulin resistance has also been assessed qualitatively if one or more insulin values exceed an upper limit of normal at appropriate intervals. (12,32)
Researchers have compared various methods for assessing insulin sensitivity in type 2 diabetics using the OGTT and found good correlations between [AUC.sub.insulin], insulin level at 120 minutes ([I.sub.120]), and the steady state plasma glucose concentrations derived from a modified ITT. (27) OGTT provides a better assessment for impaired glucose tolerance and diabetes mellitus than fasting techniques because these patients may have normal fasting glucose values despite abnormal 2-hour levels. (4)
Fasting methods for assessing insulin sensitivity
As we mentioned before, the search for uncomplicated and inexpensive quantitative tools to evaluate insulin sensitivity has led to development of fasting state (homeostatic) assessments. These tests are based on fasting glucose and fasting insulin, and use straightforward mathematical calculations to assess insulin sensitivity and beta cell function. Several homeostatic approaches have been developed in recent years, each with its merits and deficiencies. One of the weaknesses of these models is that they assume the relationship between glucose and insulin is linear when in fact it’s parabolic. (8,15-17,23)
Fasting insulin ([I.sub.0]). Fasting serum insulin is an inexpensive assay, and does not require any mathematical calculations. A fasting insulin [greater than or equal to]20 [micro]U/mL in Caucasian women (12,21) and [greater than or equal to]23 [micro]U/mL in MexicanAmerican women (12) probably indicates insulin resistance in women with PCOS. At least one researcher has advocated averaging two or three readings to account for day-to-day variability. (32)
Although [I.sub.0] is less variable than other fasting procedures in normoglycemic patients, clinicians must still interpret results cautiously. (22,34) A fasting level of 30 [micro]U/mL indicates greater insulin resistance in a diabetic individual than in a normoglycemic patient since a similar basal insulin level would not proportionately suppress glucose in a diabetic patient as well as in the normal subject. (23,28) Remember that insulin sensitivity is the ability of the hormone to reduce serum glucose. If fasting glucose is high–for example, in a patient with impaired glucose tolerance–that may indicate a diminished effect from circulating insulin (or in severe cases of insulin resistance, diminished quantity of the hormone). Hence [I.sub.0] should not be used in glucose-intolerant or diabetic patients.
Glucose/insulin ratio (G/I ratio).
The G/I ratio has become very popular since its first description in 1998 as an accurate index of insulin sensitivity in women with PCOS. (21) The ratio of glucose to insulin is easily calculated, with lower values depicting higher degrees of insulin resistance (Table 2). A G/I ratio of less than 4.5 has been shown to be sensitive (95%) and specific (84%) for insulin resistance in a group of women with PCOS, when compared to a control group. (21)
In this study, “insulin resistance” was determined by performance of FSIVGTT on normal and PCOS women. The subjects in the study were obese, non-Hispanic, Caucasian individuals with PCOS living in southern Pennsylvania. In the final paragraph of this paper, the authors comment, “Further studies will be needed to validate this measure in other populations.” (21) Indeed, other studies comparing fasting G/I ratio and other homeostatic indices between Caucasian women and other regional or ethnic populations with PCOS have confirmed the suspicion that these screening values should be population specific. (12,14-15)
In a group of women living in western Texas, for instance, Kauffman and his associates found that a G/I ratio of [less than or equal to] 7.2 suggested insulin resistance in Caucasian women with PCOS while [less than or equal to] 4.0 was more appropriate for Mexican-American women. (12) The apparent differences among Caucasian women of Pennsylvania and Texas may be due to physical and genetic differences between the populations studied, differences in the procedures used to identify insulin resistance, and variations in criteria used to define PCOS. Homeostatic model assessment (HOMA). HOMA has been widely employed in clinical research to assess insulin sensitivity. (15,16,18,27,28) Rather than using fasting insulin or a G/I ratio, the product of the fasting values of glucose (expressed as mg/dL) and insulin (expressed as [micro]U/mL) is divided by a constant:
[I.sub.0] x [G.sub.0]/405
The constant 405 should be replaced by 22.5 if glucose is expressed in S.I. units. Unlike [I.sub.0] and the G/I ratio, the HOMA calculation compensates for fasting hyperglycemia. (23) Also keep in mind that HOMA and 10 values increase in the insulin-resistant patient while the G/I ratio decreases. The HOMA value correlates well with clamp techniques and has been frequently used to assess changes in insulin sensitivity after treatment. (16) HOMA has also been used to study insulin resistance among PCOS patients of differing ethnic origins. (12,14)
Quantitative insulin sensitivity check index (QUICKI). Like HOMA, QUICKI can be applied to normoglycemic and hyperglycemic patients. (23) It is derived by calculating the inverse of the sum of logarithmically expressed values of fasting glucose and insulin (34):
1/[log([I.sub.0]) + log([G.sub.0])]
Many investigators believe that QUICKI is superior to HOMA as a way of determining insulin sensitivity, (34,35) although the two values correlate well. (34) As the [S.sub.1] decreases, QUICKI values increase.
Puffing the research to practical use
Although the concept of insulin resistance is easy to understand, quantitative assessment of insulin sensitivity and the ability to determine exactly who is insulin resistant is more difficult. Ob/gyns obviously want a simple, reliable, and inexpensive approach to measuring insulin sensitivity in women with PCOS, and such a test would enable them to pin point which patients would most likely benefit from treatment. It would also enhance their ability to monitor treatment response.
At present, homeostatic procedures like fasting insulin, G/I ratio, HOMA, and QUICKI are easily obtained and appear reasonably accurate and economically feasible. However, because normal cutoff values have only been published for fasting insulin and G/I ratio, and since only the G/I ratio screens for hyperglycemia, we believe that fasting G/I ratio is currently the most clinically useful screen for office-based assessments of insulin resistance in PCOS patients. As previously noted, however, one should keep in mind that hyperglycemia (fasting glucose [greater than ore equal to] 100 mg/dL), essentially negates the values of G/I ratio and [I.sub.0] (Table 2).
HOMA and QUICKI may prove more useful across the full range of glycemia, particularly when screening values become available for large populations. Other serum markers for insulin resistance seem likely to be found in the future.
While we can’t yet be certain, it’s possible that early detection and treatment of insulin resistance may ultimately reduce the incidence or severity of diabetes mellitus, dyslipidemia, hypertension, and cardiovascular disease. But even if that proves to be the case, there are still several problems with our current approach to insulin sensitivity assessment in PCOS:
* the apparent lack of consensus on what defines PCOS and “normal” insulin sensitivity,
* ethnic and genetic variability,
* the presence of other factors contributing to insulin resistance such as obesity, stress, and aging, and
* concern about whether simplified models of insulin sensitivity have the precision to predict treatment needs, responses, and future morbidity.
Insulin-sensitizing agents such as metformin have been advocated for all women with PCOS, (9) but it seems logical that “one size may not fit all.” Do all women with PCOS benefit from the same dose of metformin, or should dosing be tailored to the individual? Is universal use of metformin cost effective? Might universal use of metformin be counterproductive in some women? Which endpoint is optimal to follow (ovulation, homeostatic [S.sub.1], testosterone)? Would some women benefit from a specific insulin sensitizer (such as metformin, rosiglitazone, or pioglitazone) over another? Do some with PCOS have a genetic variant that would not benefit from these drugs? Do insulin sensitizers improve IVF outcomes? Future research should hold answers to these questions, and easily obtained, reliable methods to assess insulin sensitivity should assist in that quest.
* About 50% to 70% of all women with polycystic ovary syndrome have some degree of insulin resistance, and this hormone insensitivity probably contributes to the hyperandrogenism that’s responsible for the signs and symptoms of PCOS.
* While the hyperinsulinemic-euglycemic clamp technique is the gold standard for measuring insulin sensitivity, it’s too expensive, time-consuming, and labor intensive to be of practical use in an office setting.
* Fasting glucose/insulin ratio is currently the most clinically useful screen for office-based assessment of the PCOS patient.
An overview of insulin senstivity indicies
Test Type technique
Hyperinsulinemiceuglycemic Dynamic Gold
clamp clamp Standard
Insulin tolerance Insulin Good
test (ITT) infusion
Insulin sensitivity Insulin Good
test (IST) infusion
Continuous infusion Minimal Good
of glucose with
Frequently sampled Minimal Good
IV glucose tolerance
Oral glucose Minimal Good
Fasting insulin ([I.sub.0]) Homeostatic Good
Fasting glucose/ Homeostatic Good
insulin ratio (G/I ratio)
Homeostatic model Homeostatic Good
Quantitative insulin Homeostatic Good
sensitivity check index
Insulin tolerance None
Insulin sensitivity None
Continuous infusion None
of glucose with
Frequently sampled None
IV glucose tolerance
Oral glucose [G.sub.120]/[I.sub.120] +
Fasting insulin ([I.sub.0]) [I.sub.0]
Fasting glucose/ [G.sub.0]/[I.sub.0]
insulin ratio (G/I ratio)
Homeostatic model [I.sub.0] * [G.sub.0]/405
Quantitative insulin 1/[log([I.sub.0]) + log([G.sub.0])]
sensitivity check index
IV access insulin resistance
Test needed in PCOS *
Hyperinsulinemiceuglycemic Yes None
Insulin tolerance Yes None
Insulin sensitivity Yes None
Continuous infusion Yes None
of glucose with
Frequently sampled Yes None
IV glucose tolerance
Oral glucose No [less than or equal to] 1.0 (21)
Fasting insulin ([I.sub.0]) No [greater than or equal to] 20
Caycasian (12, 21)
[greater than or equal to] 23
Fasting glucose/ No [less than or equal to] 4.5
insulin ratio (G/I ratio) Caucasian (21)
[less than or equal to] 7.2
[less than or equal to] 4.0
Homeostatic model No [greater than or equal to] 3.8
assessment (HOMA) Caucasian (12)
[greater than or equal to] 4.5
Quantitative insulin No None
sensitivity check index
Hyperinsulinemiceuglycemic Most labor intensive,
offers better assessment
of [beta]-cell function.
Protocol variations **
Insulin tolerance Short interval avoids
test (ITT) effect of counter-
Protocol variations **
Insulin sensitivity Protocol variations **
Continuous infusion Protocol variations **
of glucose with
Frequently sampled Protocol variations **
IV glucose tolerance
Oral glucose Different methods to
tolerance test assess insulin
(OGTT) sensitivity have been
[I.sub.120] ratio, [ISI.sub.0.120],
interval insulin value exceeding
upper limit of normal)
Protocol variations **
Fasting insulin ([I.sub.0]) Loss of accuracy with
1/[I.sub.0] also used in some
Fasting glucose/ Loss of accurancy with
insulin ratio (G/I ratio) hyperglycemia.
Different values may be
obtained in different
Homeostatic model Use 22.5 instead of
assessment (HOMA) 405 when glucose expressed
as mmol/L. Applicable to
Quantitative insulin May have best
sensitivity check index correlation with clamp
(QUICKI) among homeostatic
models. Applicable to
* All values are expressed in conventional units
+ Other OGTT models involve more complex mathematical equations (not
** The protocols used to perform this test vary widely from institution
Homeostatic models under different metabolic conditions
Patient Fasting serum values [I.sub.0]
A [I.sub.0] 20 ([micro]U/mL) 20
[G.sub.0] 100 (mg/dL)
B [I.sub.0] 30 30
C [I.sub.0] 30 30
Patient [G.sub.0]/[I.sub.0] ratio HOMA QUICKI
A 5.0 4.94 0.303
B 3.33 7.41 0.287
C 5.0 11.11 0.273
Fasting insulin ([I.sub.0]) increases and fasting glucose/insulin ratio
([G.sub.0]/[I.sub.0]) decreases with increasing insulin resistance in
normoglycemic patients (A and B). However, in the case of diabetes
mellitus (patient C), neither [I.sub.0] nor [G.sub.0]/[I.sub.0]
accurately reflect severity of insulin resistance. In contrast, both
HOMA and QUICKI detect changes in insulin sensitivity in normoglycemic
and diabetic patients. Values for insulin and glucose are expressed in
(1.) Hinsworth HP. Insulin-sensitive and insulin-insensitive types of diabetes mellitus. Clin Sci. 1938;4:119-152.
(2.) Yalow RS, Berson SA. Immunoassay of endogenous plasma insulin in man. J Clin Invest. 1960;39:1157-1175.
(3.) Burghen GA, Givens JR, Kitabchi AE. Correlation of hyperandrogenism with hyperinselinism in polycystic ovarian disease. J Clin Endocrinol Metab. 1980;50:113-116.
(4.) Consensus development conference on Insulin Resistance, 5-6 November 1997. American Diabetes Association. Diabetes Care. 1998;21:310-314.
(5.) Ferrannini E, Balkau B. Insulin: in search of a syndrome. Diabet Med. 2002;19:724-729.
(6.) Zawadzki JK, Dunaif A. Diagnostic criteria for polycystic syndrome: towards a rational approach. In: Dunaif A, Givens JR, Haseltine FP, et al, eds. Polycystic Ovary Syndrome. Boston, Mass: Blackwell Scientific: 1992;337-384.
(7.) Polson DW, Adams J, Wadsworth J, et al. Polycystic ovaries–a common finding in normal women. Lancet. 1988;1:870-872.
(8.) Ovalle F, Azziz R. Insulin resistance, polycystic ovary syndrome, and type 2 diabetes mellitus. Fertil Steril. 2002;77:1095-1105.
(9.) Nestler JE, Stovall D, Akhter N, et al. Strategies for the use of insulin-sensitizing drugs to treat infertility in women with polycystic ovary syndrome. Fertil Steril. 2002;77:209-215.
(10.) Ehrmann DA, Cavaghan MK, Imperial J, et al. Effects of metformin on insulin secretion, insulin action, and ovarian steroidogenesis in women with polycystic ovary syndrome. J Clin Endocrinol Metab. 1997;82:524-530.
(11.) Dunaif A, Sorbara L, Delson R, et al. Ethnicity and polycystic ovary syndrome are associated with independent and additive decreases in insulin action in Caribbean-Hispanic women. Diabetes. 1993;42:1462-1468.
(12.) Kauffman RP, Baker VM, DiMarino P, et al. Polycystic ovarian syndrome and insulin resistance in white and Mexican American women: a comparison of two distinct populations. Am J Obstet Gynecol. 2002;187:1362-1369.
(13.) Williamson K, Gunn AJ, Johnson N, et al. The impact of ethnicity on the presentation of polycystic ovarian syndrome. Aust N Z J Obstet Gynecol. 2001;41:202-206.
(14.) Wijeyaratne CN, Balen AH, Barth JH, et al. Clinical manifestations and insulin resistance (IR) in polycystic ovary syndrome (PCOS) among South Asians and Caucasians: is there a difference? Clin Endocrinol. 2002;57:343-350.
(15.) Radziuk J. Insulin sensitivity and its measurement: structural commonalities among the methods. J Clin Endocrinol Metab. 2000;85:4426-4433.
(16.) Mather KJ, Hunt AE, Steinberg HO, et al. Repeatability characteristics of simple indices of insulin resistance: implications for research applications. J Clin Endocrinol Metab. 2001;86:5457-5464.
(17.) Wallace TM, Matthew DR. The assessment of insulin resistance in man. Diabet Med. 2002;19:527-534.
(18.) Jayagopal V, Kilpatrick ES, Holding S, et al. The biological variation of insulin resistance in polycystic ovarian syndrome. J Clin Endocrinol Metab. 2002;87:1560-1562.
(19.) Alberti KG, Zimmet PZ. Definition, diagnosis, and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med. 1998;15:539-553.
(20.) Ferrannini E, Natali A, Bell P, et al. Insulin resistance and hypersecretion in obesity. European Group for the Study of Insulin Resistance (EGIR). J Clin Invest. 1997;100:1166-1173.
(21.) Legro RS, Finegood D, Dunaif A. A fasting glucose to insulin ratio is a useful measure of insulin sensitivity in women with polycystic ovary syndrome. J Clin Endocrinol Metab. 1998;83:2694-2698.
(22.) Laakso M. How good a marker is insulin level for insulin resistance? Am J Epidemiol. 1993;137:959-965.
(23.) Quon MJ. Limitations of the fasting glucose to insulin ratio as an index of insulin sensitivity. J Clin Endocrinol Metab. 2001;85:4615-4617.
(24.) Yildiz BO, Gedik O. Insulin resistance in polycystic ovary syndrome: hyperandrogenemia versus normoandrogenemia. Eur J Obstet Gynecol Reprod Biol. 2001;100:62-66.
(25.) Akinmokun A, Selby PL, Ramaiya K, et al. The short insulin tolerance test for determination of insulin sensitivity: a comparison with the euglycemic clamp. Diabet Med. 1992;9:432-437.
(26.) Yang YJ, Youn JH, Bergman RN. Modified protocols improve insulin sensitivity estimation using the minimal model. Am J Physiol. 1987;253:E595-602.
(27.) Yeni-Komshian H, Carantoni M, Abbasi F, et al. Relationship between several surrogate estimates of insulin resistance and quantification of insulin-mediated glucose disposal in 490 healthy nondiabetic volunteers. Diabetes Care. 2000;23:171-175.
(28.) Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic Insulin clamp. Diabetes Care. 1999;22:1462-1470.
(29.) Ganda OP, Day JL, Soeldner JS, et al. Reproducibility and comparative analysis of repeated intravenous and oral glucose tolerance tests. Diabetes. 1978;27:715-725.
(30.) Mari A, Pacini G, Murphy E, et al. A model-based method for assessing insulin sensitivity from the oral glucose tolerance test. Diabetes Care. 2001;24:539-548.
(31.) Gutt M, Davis CL, Spitzer SB, et al. Validation of the insulin sensitivity index (ISI(0,120)): comparison with other measures. Diabetes Res Clin Pract. 2000;47:177-184.
(32.) Kidson W. Polycystic ovary syndrome: a new direction in treatment. Med J Aust. 1998;169:537-540.
(33.) McAuley KA, Williams SM, Mann JI, et al. Diagnosing insulin resistance in the general population. Diabetes Care. 2001;24:460-464.
(34.) Katz A, Nambi SS, Mather K, et al. Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab. 2000;85:2402-2410.
(35.) Hrebicek J, Janout V, Malincikova J, et al. Detection of insulin resistance by simple quantitative insulin sensitivity check index QUICKI for epidemiological assessment and prevention. J Clin Endocrinol Metab. 2002;87:144-147.
Dr. Kauffman is Assistant Professor, Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center, Amarillo, Tex:
Dr. Castracane is Professor, Department of Obstetrics and Gynecology, Texas Tech University Health Sciences Center, Amarillo, Tex.
Series editor: Richard Legro, MD, is Associate Professor, Department of Obstetrics and Gynecology, Hershey Medical Center, Hershey, Pa.