Persistence with oral contraceptive pills versus metformin in women with polycystic ovary syndrome

JOURNAL OF WOMEN’S HEALTHVolume 21, Number 6, 2012ª Mary Ann Liebert, Inc.
DOI: 10.1089/jwh.2011.3116 Persistence with Oral Contraceptive Pills Versus Metformin Nicole W. Karjane, M.D.,1 Kai I. Cheang, PharmD., M.S.,2 Gabriela A. Mandolesi, M.D.,3 and Dale W. Stovall, M.D.4 Objective: We studied patient persistence with oral contraceptive pills (OCPs) compared to metformin fortreatment of polycystic ovary syndrome (PCOS) in an urban university clinic population.
Methods: We conducted a retrospective cohort study of women with PCOS who were treated in our specialtyclinic between 2004 and 2006. All women with the diagnosis of PCOS, defined as oligomenorrhea or amenorrheain conjunction with clinical or biochemical evidence of hyperandrogenism, with exclusion of other causes, wereincluded in the study. We abstracted data on demographic characteristics, medical history, anthropometricalmeasures, desire for pregnancy, prescribed treatment, and patient report of persistence with treatment at 3, 6,and 12 months. The primary outcome measure was persistence with prescribed treatment.
Results: One hundred nineteen subjects were included in the study. Demographic and anthropometrical char-acteristics were similar between the groups. At 3 months, 57.1% were persistent with OCPs, and 57.8% werepersistent with metformin ( p = 0.93). At 6 months, the percentages dropped to 38.1% with OCPs and 43.9% withmetformin ( p = 0.46). At 12 months, only 21.7% continued with OCPs compared to 31.2% with metformin( p = 0.19). Subjects were significantly more likely to be persistent with either OCPs or metformin at 3 monthscompared to either 6 or 12 months ( p < 0.01).
Conclusions: Women with PCOS showed similar persistence rates with OCPs compared to metformin. Persis-tence with either treatment precipitously decreases over time and is modest at 12 months.
compliance or persistence with either treatment in this groupof women. Compliance, also known as adherence, is defined Polycysticovarysyndrome(PCOS)isthemostcommon as the extent to which a patient acts in accordance with the endocrinopathy in women and affects approximately prescribed dose and timing of medication, whereas persis- 5%–7% of reproductive-aged women in the United States.1 tence is defined as the accumulation of time from initiation to Clinical manifestations of this syndrome include hyperan- discontinuation of therapy. It is known that patients who are drogenemia, hirsutism, anovulation, menstrual irregularities, undergoing medical therapy for chronic diseases commonly obesity, and infertility. In addition, women with PCOS are at have high rates of both nonadherence and nonpersistence.7–13 increased risk for significant medical problems, including Because of its potential to improve biochemical parameters type 2 diabetes, cardiovascular disease (CVD), and endome- of PCOS and decrease progression to type 2 diabetes, met- formin has been emerging as a treatment of choice for women Women with PCOS are commonly prescribed both oral with PCOS, especially those with metabolic syndrome.1,14 contraceptive pills (OCPs) and metformin in an effort to treat Metformin has common side effects, however, such as nausea, many of the physical and metabolic manifestations associated vomiting, diarrhea, bloating, and abdominal pain, which may with this syndrome. Although there has been extensive re- cause a reduction in compliance or persistence or both.15 search into the outcomes of treatment using OCPs and met- Furthermore, because the potential benefits of metformin are formin in women with PCOS, little is known about patient not always clinically apparent to the patient, especially in the 1Departments of Obstetrics and Gynecology and Pediatrics, Virginia Commonwealth University School of Medicine, Richmond, Virginia.
2Department of Pharmacotherapy & Outcomes Science, Virginia Commonwealth University, Richmond, Virginia.
3Obstetrical and Gynecological Associates, Sterling, Virginia.
4Departments of Obstetrics and Gynecology and Internal Medicine, University of Virginia School of Medicine, Charlottesville, Virginia.
short term,16 she may view the medication as ineffective and of PCOS were included in the study. PCOS was defined ac- cording to the National Institute of Health (NIH) consensus Before the discovery of the association between insulin re- criteria as oligomenorrhea or amenorrhea in conjunction with sistance and PCOS, women with PCOS were primarily treated clinical or biochemical evidence of hyperandrogenism and with combination OCPs alone. Oral contraceptives are known exclusion of appropriate diseases. Ultrasound criteria were to be effective in improving menstrual irregularities and re- not used to define PCOS in this study, nor were they used to ducing the clinical signs of hyperandrogenemia, such as acne exclude a patient from the study. Subjects were excluded if and hirsutism. In general, OCPs are well tolerated; however, they did not meet diagnostic criteria for PCOS or if they were they can cause breakthrough bleeding and exacerbate head- not given a prescription for either OCPs, metformin, or both.
aches. OCPs provide the added benefit of protection against Once patients were prescribed either OCPs or metformin pregnancy, whereas metformin may induce ovulation; therapy, they were evaluated in an intention to treat meth- therefore, metformin is not a good single agent therapy for odology. A member of the research team (G.A.M.) reviewed women who do not want to conceive. Although compliance all the patient charts and abstracted data on demographic with OCPs in women seeking contraception alone is subop- characteristics, indigent status, prescription drug coverage, timal,13,17,18,19 one might speculate that the additional favor- medical history, anthropometrical measures, desire for future able effects of OCPs on the clinical signs and symptoms of pregnancy, prescribed treatment, and subject report of per- PCOS may improve both compliance and persistence in this sistence to treatment at 3, 6, and 12 months. The primary outcome measure was persistence with treatment.
This project is a pilot study designed to explore patient Categorical baseline characteristics and reasons for dis- persistence with OCPs compared to metformin for the treat- continuation of therapy among patients prescribed OCPs or ment of PCOS in a university-based clinic population. We metformin were compared with chi-square or Fishers’ exact hypothesize that women are more likely to continue using tests. Baseline characteristics in continuous variables were OCPs compared to metformin given the more favorable side compared between the groups by Student’s t tests. Equalities effect profile, more predictable improvement of menstrual of variances were tested with the Brown-Forsythe test. For irregularities, and contraceptive benefit of OCPs. By better comparisons with unequal variances, p values of Welch understanding the persistence to prescribed medical regimens analysis of variance (ANOVA) tests were reported, as indi- for PCOS in our population, we can anticipate potential rea- cated. Time trends of compliance among patients prescribed sons for nonpersistence and adapt our clinical approach to OCP and metformin were evaluated by logistic regression.
patients to improve treatment in the future.
We also evaluated whether the time trends of compliancewere different between OCP and metformin users within thelogic regression model.
To assess patients’ adherence to therapy, OCP and met- After obtaining approval from the Institutional Review formin prescriptions were used as the unit of analysis, with Board, we conducted a retrospective cohort study of women persistence determined by patient report of taking the pre- with PCOS who were treated in the PCOS clinic of the Wo- scribed medication. We chose to use each prescription, instead men’s Health Center at Virginia Commonwealth University of the individual patient, as the unit of analysis because we (VCU) in Richmond, Virginia. This clinic was specifically were interested in the persistence to the individual medica- designed for the assessment, counseling, and treatment of tions. For those who reported nonpersistence with treatment, women with PCOS, and it implemented diagnostic, treat- the reason for discontinuation of therapy, if documented, was ment, and follow-up protocols. These protocols were devel- analyzed. For patients who were noncompliant with their oped to increase uniformity, continuity, and quality of care for follow-up appointments and did not call for a refill of their this population of women. All patients were offered therapy prescription, it was assumed that they had discontinued with metformin, OCPs, or both, and the dose and dosing therapy after their last prescription had ended. Persistence regimen for metformin and OCPs, as well as the counseling rates were compared between metformin and OCPs by the and scheduled follow-up visits, were consistent from patient chi-square test. Univariate analyses of the relationship be- to patient. Current medical literature supported the use of tween persistence and social/economic factors, concomitant both metformin and OCPs for treatment of women with medical problems and therapy, and baseline demographic PCOS; therefore, patients were offered both treatments, and variables were determined using chi-square or Fishers’ exact each patient was given the clinical rationale, risks, and pos- tests for categorical variables. Association between time since sible benefits of each therapy. Because limited data suggested therapy initiation and persistence to either metformin or OCP that combination therapy with both OCPs and metformin was analyzed using repeated measures analyses. Given that might be as or more beneficial than either alone and was this was a pilot study with a relatively small sample, we did without known synergistic risks, combination therapy was not attempt multiple regression analyses, nor did we attempt recommended to all patients, particularly those who desired to adjust for multiple comparisons. Results were reported as contraception. In women who declined combination therapy means with 95% confidence intervals (CI). p values of < 0.05 and requested monotherapy, metformin was favored over (2-tailed) were considered significant. All analyses were per- OCPs because of the high obesity rates in our population and formed using JMP 8.0 software (SAS Institute, Cary, NC).
the prevailing notion that the hallmark of PCOS is insulinresistance.
All the charts of women attending the PCOS clinic from 2004 to 2006 were assessed for inclusion into the study. Wo- A total of 119 subjects were identified as meeting inclusion men with available charts and a well-documented diagnosis criteria and were included in the study. Baseline characteristics Table 1. Baseline Characteristics of Women with indigent status, drug coverage, employment status, race, ed- ucation, presence of hirsutism, diabetes, hypertension, whe- ther a patient desired pregnancy in the future, the addition ofa second therapy, age, baseline body mass index (BMI), and Patients with diabetes were more likely to continue to use their medication than those without diabetes at 3 and 6 44.23 (42.66-45.80) 44.14 (42.27-46.01) 0.94 months ( p = 0.02 and p = 0.005, respectively). Patients desiring future pregnancy had a lower persistence rate than those not desiring future pregnancy [44.4% (12 of 27) vs. 62.5% (55 of 88), respectively, p = 0.098] at 3 months. At 6 months, this re- lationship became significant, with only 6 of 27 women de- siring pregnancy being persistent with OCP, whereas 39 of 87 women not desiring pregnancy were persistent ( p = 0.03). This relationship remained significant at 12 months ( p = 0.02).
With regard to BMI, women who reported continued use of OCPs at 3 months had a lower BMI [38.1 (36.0–40.1) vs. 41.2 (38.8–43.6) kg/m2, p = 0.05] and were younger [25.0 (23.7– 26.4) vs. 27.8 (26.2–29.4) years, p = 0.01] than those who were not persistent. Younger age remained a predictor of persis- tence at 6 months but not at 12 months. BMI was no longer a predictor of persistence at 6 or 12 months.
27.17 (26.05-28.30) 26.47 (25.10-27.94) 0.45 Determinants of persistence with metformin Patients with hirsutism were more persistent with metfor- min, although this relationship did not reach statistical sig- nificance. At 3 months, 65 of 109 patients (59.6%) with 39.67 (37.97-41.38) 39.23 (37.01-41.45) 0.76 hirsutism were persistent vs. 23 of 52 without hirsutism (44.23%, p = 0.07). At 6 months, this relationship continued ( p = 0.09) but not at 12 months. The addition of OCPs de- creased persistence to metformin somewhat. In this regard, 54 of 108 patients on combination therapy were persistent at 3 months, compared to 36 of 55 patients who were on metfor- Each initial prescription serves as the unit of analysis.
min monotherapy ( p = 0.07). This relationship, however, was Data in n (%), or mean (95% confidence interval).
not significant and did not continue at 6 and 12 months. In- BMI, body mass index; OCP, oral contraceptive pill.
surance drug coverage positively predicted persistence tometformin therapy at 12 months ( p = 0.05).
of subjects were normally distributed and are described in Table 1. The two groups were similar except with regard todesire for future pregnancy, with women prescribed OCPs The reasons for discontinuing therapy are listed in Table 3.
being less likely to desire pregnancy in the future than those Loss to follow-up was the most common reason for non- prescribed metformin. Persistence to prescribed medication is persistence in both groups, and other reasons for non- shown in Table 2 and did not differ significantly between the persistence were similar between the groups, with the two groups. The percentage of women taking metformin and exception of desire for pregnancy. For OCPs, the next leading OCPs fell significantly over the 12-month follow-up period.
reason for discontinuing therapy was accessibility (defined as For each agent, we further analyzed whether socioeco- patients reporting they were unable to obtain the medication, nomic and disease state factors affect the persistence rate for usually due to insufficient insurance drug coverage), followed either metformin or OCPs. The factors we assessed included by desire for pregnancy, then gastrointestinal side effects. For aTime trends of compliance were not different between metformin and OCP users ( p = 0.2421).
Table 3. Reasons for Discontinuing Therapy analyses and lack of controlling for these in this pilot study,the findings may also be a result of chance.
The most common reason for stopping either therapy was loss to follow-up. Assuming nonpersistence in those patientsnot complying with follow-up visits (and not calling in for prescription refill) may have falsely increased the non- persistence rate and is a limitation of our study. Given the high proportion of indigent subjects seen in our clinic, most of whom do not have access to healthcare outside our facility, we thought that this was the most accurate way to categorize patients who were lost to follow-up. It is possible that these patients sought care elsewhere; however, our university clinic population tends to be quite stable overall, and our review revealed no requests for medical records (including labora- tory tests) that would have indicated transfer of care. Not surprisingly, the most commonly cited reasons for dis- bEducation means there was some misunderstanding of what continuing metformin were gastrointestinal side effects and patients were told vs. what they understood.
accessibility, and reasons cited for stopping OCPs were ac-cessibility and desire for pregnancy.
Other limitations to our study include the fact that patients metformin, the next most common reason for discontinuing self-reported their persistence with therapy. A better meth- therapy was gastrointestinal side effects, followed by other odology for assessment of persistence may have been the use of blister packs or pill counts or use of pharmacy claims da-tabases; however, these methods can also produce inaccurateresults. Another limitation is the accuracy of the documenta- tion of clinical information, which could overestimate per- Both adherence and persistence to treatment are essential to sistence, as discontinuation may not necessarily have been obtain a positive clinical result with any long-term medical documented by the providers. Additionally, not every patient intervention. Persistence with prescribed medications has received the exact same OCP because we respected the pa- been shown to reduce the risk of recurrent stroke or acute tient’s preference for one pill over another. One could spec- myocardial infarction in patients with a history of either dis- ulate that allowing the patient to have a choice in the OCP ease.20,21 Furthermore, higher adherence to oral hypoglyce- prescribed might increase her persistence with therapy. Al- mic agents has been shown to lower both diabetes-related and ternatively, if a patient requests a brand name OCP, the higher total healthcare costs in a Medicaid population with diabe- cost may subsequently lead to shorter duration of persistence tes.22 Prior to this study, little had been published in regard to with therapy. Previously reported data have shown that the persistence with medical therapies for PCOS.
prescribing generic or preferred medications can improve In our study, persistence with therapy with both OCPs and adherence to chronic medication therapy, including OCPs,25 metformin in women with PCOS was poor. More specifically, and that higher patient copayments have been associated persistence with either therapy fell from just over 50% at 3 with shorter duration of persistence with medications used to months to about 40% at 6 months and approximately 25% by 1 treat hypertension.26 Lastly, new index prescriptions, instead year. These rates of persistence are slightly lower than re- of individual patients, were used as the unit of analysis. It is ported persistence rates for hypertension therapy,9 diabetes possible that patients prescribed combination metformin and therapy,10 and lipid-lowering therapy,8,23 but similar to per- OCP therapy may decide to stop both therapies at the same sistence rates for antidepressants.12,24 Although we antici- time, and hence lack of persistence with one medication may pated that persistence with OCPs might exceed that of contaminate the persistence rate of the other. It is also plau- metformin, there were no differences in regard to persistence sible that the combination therapy may have led to com- pounding of perceived side effects and further contributed to Factors associated with persistence with OCPs included nonpersistence. Future studies with larger sample size may diabetes, younger age, and a lower BMI. The reason for discern whether persistence with monotherapy with either these relationships is unclear, but it is possible that these agent is different from that with combination therapy.
women may have a more vested interest in not becoming In summary, women with PCOS show similar persistence pregnant. Women with desire for future pregnancy were rates for treatment with OCPs compared to metformin. Per- less likely to be persistent with therapy, particularly at 6 sistence with either therapy decreases significantly with time, and 12 months. Although some of these women dis- which is consistent with findings for long-term treatment of continued OCPs in order to achieve pregnancy, this only other chronic conditions.7,8,12,27 As PCOS is a chronic condi- accounted for 18.4% of those stopping therapy. The com- tion with numerous potential long-term sequelae, it is essen- mon misperception that long-term OCP use causes infertil- tial to educate patients about the importance of long-term ity may have contributed to this finding. In regard to persistence to therapy. Strategies that have been shown to metformin, patients with hirsutism were more likely to be improve patient adherence to other medical therapies include persistent, as were patients with insurance drug coverage.
simplifying the dosage regimen, more thorough patient Again, reasons for this relationship are unclear and may counseling and instructions on medication use, and remind- benefit from further study. Furthermore, given the multiple ers, including telephone follow-up.28 Further investigation into strategies to improve medication persistence is war- open, long-term clinical evaluation. J Clin Endocrinol Metab 15. Strack T. Metformin: A review. Drugs Today (Barc) 2008; 16. Essah PA, Apridonidze T, Iuorno MJ, Nestler JE. Effects of The authors have no conflicts of interest to report.
short-term and long-term metformin treatment on menstrualcyclicity in women with polycystic ovary syndrome. FertilSteril 2006;86:230–232.
17. Rosenberg MJ, Burnhill MS, Waugh MS, Grimes DA, Hillard PJ. Compliance and oral contraceptives: A review. Contra- 1. American College of Obstetricians and Gynecologists Com- mittee on Practice Bulletins—Gynecology. ACOG practice 18. Cramer JA. Compliance with contraceptives and other bulletin No. 108: Polycystic ovary syndrome. Obstet Gynecol treatments. Obstet Gynecol 1996;88:4S–12S.
19. Nelson AL, Westhoff C, Schnare SM. Real-world patterns of 2. Castelo-Branco C, Steinvarcel F, Osorio A, Ros C, Balasch J.
prescription refills for branded hormonal contraceptives: A Atherogenic metabolic profile in PCOS patients: Role of reflection of contraceptive discontinuation. Obstet Gynecol obesity and hyperandrogenism. Gynecol Endocrinol 2010; 20. Shaya FT, El Khoury AC, Mullins CD, et al. Drug therapy 3. Karaer A, Cavkaytar S, Mert I, Buyukkagnici U, Batioglu S.
persistence and stroke recurrence. Am J Manag Care 2006; Cardiovascular risk factors in polycystic ovary syndrome. J 21. Shaya FT, Gu A, Yan X. Effect of persistence with drug 4. Lambrinoudaki I. Cardiovascular risk in postmenopausal therapy on the risk of myocardial re-infarction. P T 2008; women with the polycystic ovary syndrome. Maturitas 2011; 22. Shenolikar RA, Balkrishnan R, Camacho FT, Whitmire JT, 5. Teede H, Deeks A, Moran L. Polycystic ovary syndrome: A Anderson RT. Comparison of medication adherence and complex condition with psychological, reproductive and associated health care costs after introduction of pioglita- metabolic manifestations that impacts on health across the zone treatment in African Americans versus all other races in patients with type 2 diabetes mellitus: A retrospective 6. Tomlinson J, Millward A, Stenhouse E, Pinkney J. Type 2 data analysis. Clin Ther 2006;28:1199–1207.
diabetes and cardiovascular disease in polycystic ovary 23. Avorn J, Monette J, Lacour A, et al. Persistence of use of syndrome: What are the risks and can they be reduced? lipid-lowering medications: A cross-national study. JAMA 7. Osterberg L, Blaschke T. Adherence to medication. N Engl J 24. Sawada N, Uchida H, Suzuki T, et al. Persistence and compliance to antidepressant treatment in patients with 8. Abughosh SM, Kogut SJ, Andrade SE, Larrat P, Gurwitz JH.
depression: A chart review. BMC Psychiatry 2009;9:38.
Persistence with lipid-lowering therapy: Influence of the 25. Shrank WH, Hoang T, Ettner SL, et al. The implications of type of lipid-lowering agent and drug benefit plan option in choice: Prescribing generic or preferred pharmaceuticals elderly patients. J Manag Care Pharm 2004;10:404–411.
improves medication adherence for chronic conditions. Arch 9. Bautista LE. Predictors of persistence with antihypertensive therapy: Results from the NHANES. Am J Hypertens 2008; 26. Zhang D, Carlson AM, Gleason PP, et al. Relationship of the magnitude of member cost-share and medication persistence 10. Boccuzzi SJ, Wogen J, Fox J, Sung JC, Shah AB, Kim J. Uti- with newly initiated renin angiotensin system blockers. J lization of oral hypoglycemic agents in a drug-insured U.S.
population. Diabetes Care 2001;24:1411–1415.
27. Hasford J, Schroder-Bernhardi D, Rottenkolber M, Kostev K, 11. El-Serag HB, Fitzgerald S, Richardson P. The extent and Dietlein G. Persistence with antihypertensive treatments: determinants of prescribing and adherence with acid- Results of a 3-year follow-up cohort study. Eur J Clin Phar- reducing medications: A national claims database study. Am J Gastroenterol 2009;104:2161–2167.
28. Haynes RB, Ackloo E, Sahota N, McDonald HP, Yao X. In- 12. Esposito D, Wahl P, Daniel G, Stoto MA, Erder MH, Croghan terventions for enhancing medication adherence. Cochrane TW. Results of a retrospective claims database analysis of differences in antidepressant treatment persistence associatedwith escitalopram and other selective serotonin reuptake in- hibitors in the United States. Clin Ther 2009;31:644–656.
13. Kane SV, Brixner D, Rubin DT, Sewitch MJ. The challenge of compliance and persistence: Focus on ulcerative colitis. J Virginia Commonwealth University School of Medicine 14. Moghetti P, Castello R, Negri C, et al. Metformin effects on clinical features, endocrine and metabolic profiles, and insulin sensitivity in polycystic ovary syndrome: A randomized,double-blind, placebo-controlled 6-month trial, followed by

Source: ftp://ftp.lib.sumdu.edu.ua/ebooks/articles/Nicole%20W.%20Karjane.pdf