Pioglitazone and metformin for increased small low-density lipoprotein in polycystic ovary syndrome: Counterpoint

Rizzo and colleagues, in this issue of the Journal, reemphasize that small low-density lipoprotein (LDL) particles1, 2, 3 are part of an atherogenic dyslipidemia in women with polycystic ovary syndrome (PCOS).4 They suggest because women with PCOS cluster cardiovascular risk factors, consideration should be given to using insulin-sensitizing agent combinations. They point out that diet and exercise often fail for obese women. To prevent vascular consequences LDL particles should be normalized.

The authors are to be congratulated for bringing this issue to attention. We are facing an epidemic of obesity in most industrialized countries. Women with PCOS are likely to develop a metabolic syndrome (MBS) with atherogenic dyslipidemia.5 Most believe origins in utero6 and genetic endowment and excessive caloric intake lead to childhood obesity.7 Gynecologists often see patients for menstrual dysfunction, hyperandrogenism, and/or infertility. We are in an ideal position to diagnose these problems and to help in prevention.

Atherogenic dyslipidemia is a summary statement for low high-density lipoprotein (HDL) cholesterol (less than 50 mg/dL in women), high fasting triglycerides (greater than 150 mg/dL), and LDL cholesterol levels often not markedly abnormal. Excess LDL particles arise from impaired triglyceride metabolism often associated with insulin resistance. Pattern B or small LDL particle increase can happen with normal LDL levels.8 Mechanistically, this pattern confers atherogenic activity in conjunction with a host of risk factors including hypertension, inactivity, smoking, altered carbohydrate metabolism, central obesity, etc.

We agree that this problem needs diagnosis and prevention. Many issues need to be worked out, however, before we can suggest with confidence the best steps for treatment. Ideally our interventions are based on solid evidence-based recommendations most often driven by large quality, randomized, clinical trials with hard endpoints (like cardiovascular disease death or stroke). This entire area is hampered by a paucity of clinical trial outcome data for optimum medications for lipid control for women in general,9 let alone for young women who have an atherogenic lipid profile. They may or may not have impaired glucose tolerance (IGT) or diabetes. They may not yet demonstrate significant structural vascular disease.

Vascular lesions begin in childhood.10 Most likely the majority of us have lesions by adult age. These lesions are hard to visualize without expensive technology. Most are not associated with a high likelihood of an immediate cardiovascular event. Data regarding vascular disease burden either as a surrogate marker (coronary and/or carotid intima change and/or plaque by ultrasound or abdominal aortic change by ultrasound) or disease diagnosed by more expensive or more invasive procedures are quite meager in women with PCOS in general.11, 12

Many women with IGT will progress to frank diabetes.13, 14 Even though hard outcome data for the benefit of lipid lowering for any person under age 40 years are sparse, many of the guidelines for hypertension treatment,15 cholesterol control,16 or diabetes therapy and/or prevention17 were based on data from a prior generation. Our current patients tend to eat more and work more at a younger age and have less leisure time necessary to stay fit.

How should we proceed? The history of medicine is littered with recommendations based on mechanistic inference later proven to be incorrect. What best evidence helps with this issue?

The Adult Treatment Panel National Cholesterol Education Program is an excellent resource to help. Current recommendations18 are to determine a Framingham risk score in all persons before deciding on targets for therapy. First, we should target LDL cholesterol and use a statin in persons with known diabetes (because diabetes is a risk equivalent for a very high Framingham score or for known disease like prior heart attack or significant carotid disease or equivalent). Clinical trial evidence is available. Reduction in events occurs in diabetics, regardless of the LDL level. Angiotensin-converting enzyme inhibitors and/or statins are contraindicated in pregnancy. Primary considerations should be given to lifestyle intervention and secondary consideration to Fib rate or niacin when the non HDL cholesterol (total HDL cholesterol) is not a goal. Therapy is dictated in part by the presence or absence of IGT. Niacin can aggravate insulin resistance. Each of these agents is not without cost or potential side effects. Each needs appropriate monitoring and follow up. Target goals are determined by risk status (non-HDL goal is 30 points higher than the LDL target).

Although abnormal LDL particles are common in PCOS patients, this test is often not available, it is expensive, and it is often not reimbursable. There is a move to substitute apolipoprotein B or use the non-HDL cholesterol value or the HDL to triglyceride ratio (greater than 3.5) to help determine whether pattern B or LDL particle excess is likely. Nonfasting triglyceride concentrations are being evaluated as predictors as well. No lipid screening test is perfect.

Not all patients with PCOS have the MBS; not all have LDL particle excess. Although criteria for the MBS differ slightly between professional organizations, 1 practical definition includes: (1) increased waist measurement circumference greater than 35 inches in Caucasians or lower in certain high-risk groups; (2) HDL less than 50 ng/mL, and (3) triglycerides greater than 150 mg/dL, blood pressure greater than 130/85 mm Hg, or a fasting blood sugar greater than 100 mg/dL fasting or, equally, treatment for hypertension, or diabetes. Should we be worried only about all patients with PCOS who also meet criteria for metabolic syndrome?

Before we start committing our patients to treatment with a thiazolidine, cost-benefit issues and how to best decide when exercise and/or diet have failed and medication is indicated need clarity. At the time of this writing, troglitazone was taken off the market for liver problems, and rosiglitazone is under scrutiny for potential heart problems and risk for bone fracture and edema.19 The entire class of drugs is under scrutiny to determine whether they really prevent diabetes. In contrast, diet and exercise and metformin prevent diabetes in patients with IGT.20

We advocate looking at the vital signs, including pulse, blood pressure (calibrated large cuff when needed), and waist circumference in addition to height and weight. The risk for all-cause death (mortality) rises when the body mass index exceeds 26-27 kg/m2; risk for coronary heart disease and diabetes rises when the waist circumference exceeds approximately 75 cm for both sexes.21 Central obesity, as simply measured by the waist circumference, predicts risk. A waist of 36 inches in Europids and Africa Americans, 32 inches in Indian Asians, and 34 inches in Southeast Asians portends the likelihood for abnormal low HDL and triglyceride elevation.

Knowing that the culprit is accumulation of central body fat, we should focus our efforts on intensive lifestyle alterations as proven by evidence-based trials. A pedometer is often helpful to assure 10,000 steps per day to maintain weight and to encourage 15,000 steps per day to lose weight or to maintain after weight loss (equivalent of 30 minutes or 60 minutes of brisk walking to maintain weight loss or lose weight, respectively). Providing adequate nutrition is easier now. In almost all cites, a credentialed nutritionist is easily accessed through the web. Our first goal should be to avoid weight gain. Losing 10% of current weight is associated with a 50% chance of returning ovulation.22 This amount of weight loss is realistic and tends to improve almost all of the metabolic parameters.23

It is time for evidence-based research to understand old and new medications to control blood pressure, dyslipidemia, and the glucose levels for reproductive-age females with the MBS. A multitarget approach can accomplish much more than any single monotherapy or single-target combination therapy for cardiometabolic risk.

Many women have a cardiometabolic syndrome. Many women with cardiometabolic syndrome have PCOS. Not all women with PCOS have cardiometabolic syndrome. In our opinion the use of metformin plus a thiazolidine, although mechanistically attractive, has the potential to do harm. When we must step outside the clinical trial evidence box, we should use the best evidence available from other populations and best clinical judgment to extrapolate.

The future also needs to address optimum tools to assess burden of disease. Carotid ultrasound to detect sublicinical disease,24 because of cost, ease of performance, wide availability (carotid intima-thickening and/or plaque), is under scrutiny. Reliability and observer variation need to be optimized. Likely the need for both drug intervention and monitoring will be by assessing burden of disease. We must intensify our efforts to get our patients to adhere to our instruction. We need to acknowledge they are trying to lose weight and teach a skill set to help achieve a realistic goal. Medication should be reserved for: (1) diagnosed IGT when lifestyle alteration fails; (2) high blood pressure (international and national guidelines are easily available); (3) proven vascular disease; (4) a high Framingham risk score; or (5) recalcitrant lipid abnormalities including low HDL cholesterol, high triglycerides, and the presence of excess LDL cholesterol particles if this information is available.