Increased Incidence of Gestational Diabetes in Women Receiving Prophylactic 17 -Hydroxyprogesterone Caproate for Prevention of R

Epidemiology/Health Services Research ORIGINAL ARTICLE

Increased Incidence of Gestational Diabetes in Women Receiving Prophylactic 17␣-Hydroxyprogesterone Caproate for Prevention of Recurrent Preterm Delivery

1 2 ANDREI REBARBER, MD DEBBIE J. RHEA, MPH term birth in the study population by 2 2 NIKI B. ISTWAN, RN GARY J. STANZIANO, MD 33% (3). Meta-analysis including both 3 1 KAREN RUSSO-STIEGLITZ, MD DANIEL H. SALTZMAN, MD 1 older and current studies has provided JANE CLEARY-GOLDMAN, MD further support for the use of 17P for preterm birth prevention (4), although the mechanism of action by which 17P pre- OBJECTIVE — Progesterone has a known diabetogenic effect. We sought to determine vents preterm birth remains poorly un- whether the incidence of gestational diabetes mellitus (GDM) is altered in women receiving ␣ derstood. weekly 17 -hydroxyprogesterone caproate (17P) prophylaxis for the prevention of recurrent The metabolic changes of normal preterm birth. pregnancy are essential to provide ade- RESEARCH DESIGN AND METHODS — Singleton gestations in women having a quate nutrients to the growing fetus. As history of preterm delivery were identified from a database containing prospectively collected pregnancy progresses, increased levels of information from women receiving outpatient nursing services related to a high-risk pregnancy. human chorionic sommatomam- Included were patients enrolled for outpatient management at Ͻ27 weeks’ gestation with doc- motropin, cortisol, prolactin, progester- umented pregnancy outcome and delivery at Ͼ28 weeks. Patients with preexisting diabetes were one, and estrogen lead to insulin excluded. The incidence of GDM was compared between patients who received prophylactic resistance. Studies in animal models dem- intramuscular 17P (250-mg weekly injection initiated between 16.0 and 20.9 weeks’ gestation) onstrate that progesterone plays an im- and those who did not. portant role in signaling insulin release RESULTS — Maternal BMI and age were similar. The incidence of GDM was 12.9% in the 17P and pancreatic function (5). The relatively group (n ϭ 557) compared with 4.9% in control subjects (n ϭ 1,524, P Ͻ 0.001; odds ratio 2.9 diabetogenic properties of progesterone, [95% CI 2.1–4.1]). which peak at 32 weeks’ gestation, have been described in humans (6). The Amer- CONCLUSIONS — The use of 17P for the prevention of recurrent preterm delivery is ican Diabetes Association (ADA) recom- associated with an increased risk of developing GDM. Early GDM screening is appropriate for mends screening all women at risk for women receiving 17P prophylaxis. gestational diabetes mellitus (GDM). The Diabetes Care 30:2277–2280, 2007 ADA considers women to be at risk for GDM unless they are aged Ͻ25 years, have normal body weight, have no first- reterm birth is the leading cause of egy after a recent study from the National degree relatives with diabetes, have no perinatal mortality and morbidity Institute of Child Health and Human De- history of glucose intolerance or poor ob- P for nonanomalous infants in the velopment Maternal-Fetal Medicine Units stetrical outcome, and are not a member Ͼ ϳ U.S. where 12% of infants, 480,000, Network (2,3). This study examined the of a high-risk ethnic group (7). A 2001 are born prematurely each year (1). Al- effectiveness of 17P in reducing the rate of Practice Bulletin of the American College though past studies of progestational preterm delivery in women with a single- of Obstetricians and Gynecologists rec- agents for the prevention of preterm de- ton gestation and a history of prior pre- ommended a similar risk-based ap- livery reported varied results, there has term delivery. 17P was administered via proach, but noted that because only a been renewed interest in the use of 17␣- weekly injections initiated between 16 small percentage of patients meet the cri- hydroxyprogesterone caproate (17P) as a and 20 weeks’ gestation and was shown to secondary preterm birth prevention strat- decrease the incidence of recurrent pre- teria for low risk, universal 50-g 1-h glu- ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● cose challenge test screening may be a more practical approach (8). Conversely, 1 From the Mount Sinai School of Medicine, Division of Maternal Fetal Medicine, New York, New York; the U.S. Preventive Services Task Force 2Matria Healthcare, Department of Clinical Research, Marietta, Georgia; and 3Valley Health System, Division of Maternal Fetal Medicine, Ridgewood, New Jersey. concluded that the evidence is insufficient Address correspondence and reprint requests to Andrei Rebarber, MD, 70 E. 90th St., New York, NY to recommend for or against routine 10029. E-mail: [email protected]. screening for GDM (9). Because glucose Received for publication 21 March 2007 and accepted in revised form 28 May 2007. intolerance increases during pregnancy, Published ahead of print at http://care.diabetesjournals.org on 11 June 2007. DOI: 10.2337/dc07-0564. Abbreviations: 17P, 17␣-hydroxyprogesterone caproate; ADA, American Diabetes Association; GDM, screening for GDM is most commonly gestational diabetes mellitus. conducted during the 24th–28th week of A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion gestation (9,10). The purpose of the factors for many substances. present study was to determine whether © 2007 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby there is an increased incidence of GDM in marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. women receiving supplemental proges-

DIABETES CARE, VOLUME 30, NUMBER 9, SEPTEMBER 2007 2277 Progesterone and GDM

Table 1—Maternal characteristics group received specialized education and counseling from a perinatal nurse, based 17P treatment Control P value on their clinical condition and the outpatient program in which they were en- n 557 1,524 rolled. All patients received scheduled Maternal age (years) 29 (16–44) 30 (16–45) 0.855 clinical assessment, which included eval- Ͼ37 years 53 (9.5) 125 (8.2) 0.376 uation of any patient-reported signs or Tobacco use 54 (9.7) 87 (5.7) 0.002 symptoms of preterm labor, and had Married 421 (75.6) 1,190 (78.1) 0.214 nursing support available via telephone History of preterm delivery 557 (100) 1,524 (100) — 24 h/day. Prescription of betamimetic Prepregnancy BMI 26.2 Ϯ 6.6 26.2 Ϯ 6.7 0.791 medications for tocolysis (in both groups) Obese BMI 140 (25.1) 340 (22.3) 0.178 was at the discretion of each patient’s in- Betamimetic tocolysis 101 (18.1) 375 (24.6) 0.002 dividual health care provider. Gestational age at start of out- 19.0 (16.0–26.9) 21.6 (4.7–25.9) Ͻ0.001 Data were analyzed using Student’s t, patient management (weeks) Mann-Whitney U, Pearson’s ␹2, and Fish- Data are median (range), means Ϯ SD, or n (%). er’s exact test statistics as appropriate on the basis of data distributions to compare differences between control and treat- terone by weekly 17P injection for the outpatient services, those who had a med- ment groups. Because maternal weight prevention of preterm birth. ical history of diabetes before the current and betamimetic medications are com- pregnancy, and those who had “Un- monly thought to influence development RESEARCH DESIGN AND known” designated for GDM in the ante- of GDM, data regarding prepregnancy METHODS — The study population partum outcome record. Also excluded BMI (Ͻ20, 20–24.9, 25–29.9, and Ն30 was identified retrospectively from a large were women experiencing recurrent pre- kg/m2) and use of betamimetics for toco- database containing information from term delivery before 28 weeks in the cur- lysis were also examined. The primary women who received outpatient perinatal rent pregnancy, as these women may not study outcome was the incidence of services for pregnancy-related conditions have yet received testing for GDM. Tim- GDM. through Matria Healthcare. Information ing of GDM testing was determined by stored in the database is collected pro- each patient’s physician and was not doc- RESULTS — Maternal characteristics spectively from the patient and her health umented in the outpatient record. For are presented in Table 1. As expected, be- care provider at the start of outpatient ser- most women, glucose screening is con- cause the recommended gestational age vices, as well as during the course of care. ducted between 24 and 28 weeks’ gesta- for initiation of 17P is 16–20 weeks, The data include medical and pregnancy tion unless they are known to have women receiving 17P started outpatient history, current pregnancy risk factors carbohydrate intolerance before the 24th services earlier than those in the control and diagnoses, biometric clinical data rel- week of gestation (9,10). group. The majority of women in the con- ative to services provided, medications re- Data were divided into treatment trol group (62.1%) received outpatient ceived, and maternal and neonatal (17P) and control (no 17P) groups. The preterm labor surveillance services (daily outcome data. All information is collected treatment group comprised 557 patients outpatient uterine contraction monitor- using standardized operating procedures, for whom weekly intramuscular injec- ing and nurse assessment), whereas those forms, and computer systems. Written tions of 250 mg of 17P were prescribed remaining received outpatient services for consent allowing the use of her anony- beginning at 16–20.9 weeks’ gestation conditions such as hyperemesis gravida- mous, de-identified data for research and and administered by Matria Healthcare rum or pregnancy-related hypertension. reporting purposes is obtained from each between April 2004 and January 2006, All patients in both groups had a history patient upon enrollment for services. whereas the control group comprised of at least one prior preterm delivery. Be- Each patient’s physician was responsible 1,524 patients at similar risk for recurrent tamimetic tocolytic medications were for all antepartum testing and treatment preterm delivery (history of prior preterm prescribed for almost 25% of control pa- decisions. delivery) who did not receive 17P through tients compared with 18.1% in the 17P For the present study, we used de- Matria Healthcare or any other source. 17P treatment group (P ϭ 0.002). Women in identified data from women with single- was compounded by a qualified com- the treatment group received 14.9 Ϯ 4.5 ton gestations who had a history of prior pounding pharmacy using an Insurance (mean Ϯ SD) injections of 250 mg of 17P. preterm delivery and who enrolled for Service Organization Class 5 Clean Room, Women receiving 17P had a signifi- outpatient services at Ͻ27 weeks’ gesta- with adequate quality control procedures cantly higher incidence of GDM than con- tion. Eligible for analysis were women and documentation to assure sterility and trol subjects (12.9% in the 17P group who had a documented height and potency of each vial. Unit-dose vials were compared with 4.9% in control subjects, prepregnancy weight and complete doc- delivered to the patient’s home for weekly P Ͻ 0.001; odds ratio [OR] 2.9 [95% CI umentation of pregnancy outcome (in- administration by a perinatal nurse. During 2.1–4.1]). Gestational age at delivery was cluding antepartum complications the weekly visits, patients were counseled similar between the groups (36.9 Ϯ 2.3 documented in discrete fields of “Yes, ” on the signs and symptoms of preterm la- weeks in the 17P group compared with “No, ” or “Unknown ” answers and GDM bor. Between weekly visits, nurses and 37.1 Ϯ 2.4 weeks in control subjects, P ϭ status) in the outpatient record. We ex- pharmacists were available at any time for 0.080). There were similar rates of spon- cluded patients reporting a preexisting di- patient questions and concerns through a taneous recurrent preterm delivery at agnosis of diabetes at admission for toll-free number. Patients in the control Ͻ35 weeks’ gestation between women

2278 DIABETES CARE, VOLUME 30, NUMBER 9, SEPTEMBER 2007 Rebarber and Associates

Table 2—Logistic model results that women receiving 17P prophylaxis be considered at high risk for GDM and that P value OR (95% CI) earlier screening in this population is warranted. Obese BMI (Ն30 kg/m2) Ͻ0.001 6.91 (2.93–16.28) The pregnant state is characterized by Overweight BMI (25.0–29.9 kg/m2) 0.004 3.70 (1.53–8.92) decreased insulin-stimulated tissue glu- 17P prophylaxis Ͻ0.001 3.09 (2.17–4.40) cose uptake and increased liver glucose Normal BMI (20.0–24.9 kg/m2) 0.192 1.80 (0.74–4.38) production. The adaptive response to this Betamimetic tocolysis 0.852 1.04 (0.67–1.64) increased insulin resistance is increased Gestational age at start of outpatient care 0.050 0.97 (0.933–1.000) production of insulin by the pancreatic Tobacco use 0.193 0.57 (0.24–1.33) ␤-cells. Progesterone is known to exhibit diabetogenic properties during pregnancy. Mechanisms proposed for this ef- treated with 17P and control subjects genic or glucocorticoid activity, no fect include enhancement of insulin (12.4 vs. 9.6%, respectively, P ϭ 0.062). virilizing effects on female fetuses, and no resistance through a reduction in GLUT4 Because of significant univariate dif- teratogenic effects. expressions or impairment of the normal ferences found in rates of maternal smok- The present study was not designed ␤-cell adaptive response of enhanced in- ing, betamimetic drug use, and to assess the efficacy of 17P for the pre- sulin secretion (20). Given the biological gestational age at the start of outpatient vention of preterm delivery. Gestational plausibility of progesterone-mediated management between the study groups age at delivery was similar between gestational hyperglycemia, we sought to (Table 1), plus the known association be- women who received 17P and those who define the actual clinical risk of GDM in tween maternal weight and GDM, a logis- did not, thus allowing a similar window of pregnant women treated with 17P. tic regression model was tested to assess opportunity for development of GDM. Al- To our knowledge, this is the first relative independent associations on the though all of the patients studied had a study to examine the impact of 17P pro- dependent outcome of GDM incidence history of prior preterm delivery, Ͼ35% phylaxis on the incidence of GDM. Limi- (Table 2). Although patients in the obese of patients in the control group received tations of the present study include the and overweight categories had the highest outpatient services unrelated to preterm inability to fully evaluate fetal outcomes risk of developing GDM, the use of 17P birth prevention (e.g., hyperemesis re- and the measurable metabolic impact on continued to impart a positive, indepen- lated services) and thus may have had an the maternal milieu, inability to stratify dent association with GDM incidence overall lesser risk for preterm delivery data by maternal race and ethnicity as (overall adjusted OR 3.09). than those in the study group. these data were not consistently available GDM, particularly if uncontrolled, is in the database, and lack of information CONCLUSIONS — It is estimated associated with an increased risk of peri- regarding timing of GDM testing and ma- that one in eight infants in the U.S. is born natal morbidity (18). The risks for shoul- ternal risk factors for GDM. Further stud- preterm, accounting for nearly 500,000 der dystocia, death, bone fracture, and ies are warranted to better elucidate the preterm births each year (1). Primary pre- nerve palsy can be increased without ap- association between 17P and GDM in vention of preterm birth is a public health propriate therapy (19). It is therefore pru- women with and without other risk fac- priority because of the short-term and dent to investigate the impact of weekly tors for development of GDM and to clar- long-term medical consequences and fi- supplemental 17P on the incidence of ify the quantifiable impact of 17P therapy nancial costs to the health care system. GDM, particularly given its new found on insulin resistance during pregnancy. There are Ͼ3 decades of data describing popularity. Initial concerns with this ther- In this early study, supplemental 17P in- the use of various progesterone com- apy involved the timing of the recom- jections of 250 mg given weekly for the pounds administered for the prolonga- mended onset of initiation (e.g.,16–20 prevention of preterm delivery appear to tion of pregnancy (11). The use of weeks’ gestation) in comparison with increase the incidence of GDM, a condi- progesterone supplementation as a pre- standard timing for screening for GDM tion that has now been clearly shown to ventative therapy for women with recur- (e.g., 24–28 weeks’ gestation). The ADA be associated with an adverse pregnancy rent preterm birth was recently evaluated recommends that risk assessment for outcome (19). Clinical implications of in two randomized, controlled trials GDM should be performed at the first pre- this finding may include screening (3,12). Both showed that progesterone natal visit. Women with clinical charac- women receiving 17P for GDM who oth- use substantially reduced the rate of pre- teristics consistent with a high risk for erwise would be considered to have a low term delivery. The American College of GDM (marked obesity, a personal history risk for the condition and earlier or more Obstetricians and Gynecologists’ Com- of GDM, glycosuria, or a strong family frequent GDM screening for those with a mittee on Obstetric Practice recommends history of diabetes) should undergo glu- moderate risk. that progesterone supplementation be cose testing as soon as feasible. If they are considered only for women with a history found not to have GDM at that initial of a previous spontaneous birth at Ͻ37 screening, they should be retested be- References weeks of gestation, pending the outcome tween 24 and 28 weeks’ gestation. 1. Martin JA, Hamilton BE, Sutton PD, Ven- tura SJ, Menacker F, Kirmeyer S: Births: of further investigations (13). The effects Women with an average risk should have final data for 2004, in National Vital Sta- of 17P on pregnancy in experimental an- testing performed at 24–28 weeks’ gesta- tistics Reports. vol. 55 no. 1. Hyattsville, imals has been studied in rats, rabbits, tion (7). If, in fact, exposure to supple- MD, National Center for Health Statistics, mice, and monkeys (14–17). These ear- mental 17P results in an increased risk for 2006 lier studies found no evidence of andro- the development of GDM, we postulate 2. Goldstein P, Berrier J, Rosen S, Sacks HS,

DIABETES CARE, VOLUME 30, NUMBER 9, SEPTEMBER 2007 2279 Progesterone and GDM

Chalmers TC: A meta-analysis of random- Gynecologists Committee on Practice 14. Johnstone EE, Franklin RR: Assay of pro- ized control trials of progestational agents Bulletins–Obstetrics: ACOG Practice Bul- gestins for fetal virilizing properties using in pregnancy. Br J Obstet Gynaecol 96: letin: clinical management guidelines for the mouse. Obstet Gynecol 23:359–362, 265–274, 1989 obstetrician-gynecologists. Number 30, 1964 3. Meis PJ, Klebanoff M, Thom E, Dom- September 2001. Obstet Gynecol 98:525– 15. Courtney KD, Valerio DA: Teratology in browski MP, Sibai B, Moawad AH, Sprong 538, 2001 the Macaca mulatta. Teratology 1:163– CY, Hauth JC, Miodovnik M, Varner MW, 9. U.S. Preventive Services Task Force: Screen- 172, 1968 Leveno KJ, Caritis SN, Iams JD, Wapner ing for Gestational Diabetes Mellitus: Recom- 16. Carbone JP, Brent RL: Genital and non- RJ, Conway D, O’Sullivan MJ, Carpenter mendations and Rationale. Rockville, MD, genital teratogenesis of prenatal progester- M, Mercer B, Ramin SM, Thorp JM, Peace- Agency for Healthcare Research and Qual- one therapy: the effects of 17␣-hydro- man AM, Gabbe S, National Institute of ity, 2003. Available from http://www.ahrq. xyprogesterone caproate on embryonic Child Health and Human Development gov/clinic/3rduspstf/gdm/gdmrr.htm. and fetal development and endochon- Maternal-Fetal Medicine Units Network: Accessed 14 March 2007. dral ossiﬁcation in the C57B1/6L mouse. Prevention of recurrent preterm delivery 10. Metzger BE, Coustan DR, Organizing Am J Obstet Gynecol 169:1292–1298, 1993 by 17-␣-hydroxyprogesterone caproate. Committee: Summary and recommenda- 17. Miller RE, Nelson GW, Johnson CK: Eval- N Engl J Med 348:2379–2385, 2003 tions of the 4th International Workshop uation of teratogenic potential of Delalu- 4. Sanchez-Ramos L, Kaunitz AM, Delke I: Conference on gestational diabetes. Dia- tin (17␣-hydroxyprogesterone caproate) Progestational agents to prevent preterm betes Care 21:B161–B167, 1998 in mice. Teratology 28:201–208, 1983 birth: a meta-analysis of randomized con- 11. Meis PJ: 17 Hydroxyprogesterone for the 18. Gonzalez-Quintero VH, Istwan NB, Rhea trolled trials. Obstet Gynecol 105:273– prevention of preterm delivery. Obstet Gy- DJ, Rodriguez LI, Cotter A, Carter J, Muel- 279, 2005 necol 105:1128–1135, 2005 ler A, Stanziano GJ: The impact of glyce- 5. Picard F, Wanatabe M, Schoonjans K, Ly- 12. da Fonseca EB, Bittar RE, Carvalho MHB, mic control on neonatal outcome in don J, O’Malley BW, Auwerz J: Progester- Zugaib M: Prophylactic administration of singleton pregnancies complicated by one receptor knockout mice have an progesterone by vaginal suppository to gestational diabetes. Diabetes Care 30: improved glucose homeostasis secondary reduce the incidence of spontaneous pre- 467–470, 2007 to ␤-cell proliferation. Proc Natl Acad Sci U term birth in women at increased risk: a 19. Crowther CA, Hiller JE, Moss JR, McPhee SA99:15644–15648, 2002 randomized placebo-controlled double- AJ, Jeffries WS, Robinson JS: Effect of 6. Carr DB, Gabbe S: Gestational diabetes: blind study, Am J Obstet Gynecol 188: treatment of gestational diabetes mellitus detection, management, and implica- 419–424, 2003 on pregnancy outcomes. N Engl J Med tions. Clin Diabetes 16:1–17, 1998 13. American College of Obstetricians and 352:2477–2486, 2005 7. American Diabetes Association: Diagnosis Gynecologists: ACOG Committee Opin- 20. Branisteanu D, Mathieu C: Progesterone and classiﬁcation of diabetes mellitus. Di- ion, Number 291: use of progesterone to in gestational diabetes mellitus: guilty or abetes Care 30:S42–S47, 2007 reduce preterm birth. Obstet Gynecol 102: not guilty? Trends Endocrinol Metab 14: 8. American College of Obstetricians and 1115–1116, 2003 54–56, 2003

2280 DIABETES CARE, VOLUME 30, NUMBER 9, SEPTEMBER 2007