Thermoanalytical study of sweetener myo-inositol: α and β polymorphs

Food Chemistry 237 (2017) 1149–1154

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Thermoanalytical study of sweetener myo-inositol: a and b polymorphs Rafael Turra Alarcon a, Caroline Gaglieri a, Flávio Junior Caires a, Aroldo Geraldo Magdalena a, Ricardo António Esteves de Castro b, Gilbert Bannach a,⇑ a b

São Paulo State University (UNESP), School of Sciences, Chemistry Department, 17033-260 Bauru, SP, Brazil Coimbra University, College of Pharmacy, 3000-548 Coimbra, Portugal

a r t i c l e

i n f o

Article history: Received 24 April 2017 Received in revised form 6 June 2017 Accepted 9 June 2017 Available online 10 June 2017 Keywords: Inositol Polymorphism Thermal studies DSC-Photovisual Polarized Light Thermomicroscopy Food storage Melting point

a b s t r a c t This work investigates the thermal behavior of a and b myo-inositol polymorphs. The inositol is a natural compound widely used in the food industry due to its presence in carbohydrate metabolism and its sweet taste. The occurrence of polymorphism could change some physico-chemical properties, such as melting and sublimation temperatures, and solubility. Therefore, the thermal study of polymorphism is important to ensure better conditions for synthesis, storage, and transportation of food that contains the myo-inositol. Simultaneous Termogravimetry-Differential Thermal Analysis, Photovisual Differential Scanning Calorimetry, Polarized Light Thermomicroscopy, and Powder X-ray Diffraction were used in investigation. The data show a new thermal event associated to b myo-inositol melting at 221.43 °C, suggesting that the solid-solid transition at 185.68 °C was incomplete. The kinetics data made it possible to determine the transition lifetime of myo-inositol to occur 5% of solid-solid transition at 20 °C and 37 °C: 126 and 8 years, respectively. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Myo-Inositol (1,2,3,4,5,6-Hexahydroxycyclohexane, MI) or mesoInositol, is a natural compound that can be found in microorganisms, animals, and plants due to its metabolism role (Holub, 1986; Loewus & Murthy, 2000). MI is related to complex B vitamins, has a sweet taste, is soluble in water, and insoluble in organic solvents (O’Neil et al., 2006). Some researchers evince that MI could be a substitute to Metformin in diabetes treatment (Arendrup, Gregersen, Hawley, & Hawthorne, 1989; Corrado et al., 2011; D’Anna et al., 2012), used by women with Polycystic Ovarian Syndrome (Govindarajan et al., 2015; Turan et al., 2015; Unfer, Carlomagno, Dante, & Facchinetti, 2012), and used against some diseases such as cancer (Hecht et al., 1999) and Acute Respiratory Distress Syndrome (Howlett, Ohlsson, & Prakkal, 2015). Due to its presence in carbohydrate metabolism, MI is mostly used in foods for dietary treatments and supplementary foods (Önay-Uçar et al., 2014). Notwithstanding, it is also found in energy drinks (Babu, Church, & Lewander, 2008), honey (De la Fuente, Sanz, Martínez-Castro, Sanz, & Ruiz-Matute, 2007; Sanz, Sanz, & Martínez-Castro, 2004a),

fruit juices (Sanz, Villamiel, & Martínez-Castro, 2004b), olive fruit (Marsilio, Campestre, Lanza, & De Angelis, 2001), and Infant Food Formulation such as milk powder by FDA approval (Flores, Moreno, Frenich, & Vidal, 2011; Food & Drug Administration, 1985; Indyk, Lawrence, & Broda, 1993; Sabater, Prodanov, Olano, Corzo, & Montilla, 2016; Thompkinson & Kharb, 2007). Thus, it is important to know the thermal behavior of MI, to optimize the conditions for storage and transportation of foods that contain MI in their formulation, which is necessary to ensure the physico-chemical quality. Polymorphs could change physicochemical properties of foods such as their stability, solubility, melting point, and crystallization point (Braga, Grepioni, Maini, & Polito, 2009). Thermal techniques including ThermogravimetryDifferential Thermal Analyses (TG-DTA) and Differential Scanning Calorimetry (DSC) analyze the thermal behavior of polymorphs (Giron, 2002). One research found inositol b polymorphs, which are a result of crystallization of MI (Izutsu et al., 2014). This paper proposes a deep thermal investigation of a and b MI polymorphs. Hence, a new thermal event of b Myo-Inositol was discovered. 2. Materials and methods

⇑ Corresponding author. E-mail addresses: [email protected] (R.T. Alarcon), carolinegaglieri@ gmail.com (C. Gaglieri), [email protected] (F.J. Caires), [email protected]. br (A.G. Magdalena), [email protected] (R.A.E. de Castro), [email protected] (G. Bannach). http://dx.doi.org/10.1016/j.foodchem.2017.06.072 0308-8146/Ó 2017 Elsevier Ltd. All rights reserved.

2.1. Materials Myo-Inositol analytical grade
99% (PubChem CID: 892) was purchased from Aldrich.

1150

R.T. Alarcon et al. / Food Chemistry 237 (2017) 1149–1154

2.2. Polarized Light Thermomicroscopy (PLTM) The Polarized Light Thermomicroscopy (PLTM) studies were performed in a Linkam hot stage system, Model DSC600. For optical observation a Leica DMRB microscope, a Sony CCD-IRIS/RGB video camera, and a Sony HR Triniton monitor were used. Real Time Video Measurement System software by Linkam was used for image analysis. A small amount of the myo-inositol was dispersed throughout the glass crucibles, which were covered with glass lid. Using the samples as loose powder allows study of the thermal behavior of individual particles, which is very useful for organic compounds, which can present heterogeneous thermal behavior. The images were obtained by means of combined use of polarized light and wave compensators, resulting in a nonblack monochromatic background. 2.3. Thermal Analysis: Thermogravimetry (TG), Differential Thermal Analysis (DTA), and Differential Scanning Calorimetry (DSC) Simultaneous TG-DTA curves for each polymer were obtained using the thermal analysis system from Netzsch, model STA 449 F3. Approximately 5 mg of sample were weighted and placed in a 70-mL a-alumina open crucible. The parameters were set at a

heating rate of 10.0 °C min1, and a flow rate of 50 mL min1 in a dry air atmosphere. The temperature range was from 30.0 °C to 600.0 °C. The DSC analyses and images were obtained on a MettlerToledo equipment, model DSC 1 Stare System with digital camera SC30 3.3 megapixel and magnification of 6.5x. Approximately 3 mg of sample were placed in a 40 mL closed aluminum crucible with perforated lid. The heating rate was 10 °C min1, and the flow rate was 50 mL min1. The environment used was dry air atmosphere. The cycle of heating/cooling began at room temperature (25.0 °C). Then, the samples were heated to 250.0 °C, and cooled once again to 25.0 °C. 2.4. Powder X-Ray Diffraction (PXRD) The powder X-ray Diffraction were performed on a Rikagu D-TEX Ultra using Cu Ka radiation (k = 1.5403 Å) and settings of 40 kV and 15 mA, in the 2h range of 5–50° and step of 0.04°. 2.5. Kinetic parameters The solid-solid transition of a-inositol was studied using Nonisothermal method by DSC. The same equipment and parameters

Fig. 1. DSC cyclic curve: complete cycle curve with all thermal events (a), maximize solid-solid thermal event (b), and maximize b-inositol melting (c).

1151

R.T. Alarcon et al. / Food Chemistry 237 (2017) 1149–1154

used before were employed, except for the heating rate, which was reset to 5.0, 10, 15, 20, and 25 °C min1. All the data were processed using the ThermoKinetics 3.1 software by Netzsch (2017). 3. Results and discussion 3.1. Thermal analysis: Thermogravimetry (TG), Differential thermal Analysis (DTA), and Differential Scanning Calorimetry (DSC) The TG/DTG-DTA curves of the composites are presented in Supplementary material (Fig. S1). The First Derivative of TG curve (DTG) was applied to obtain the thermal steps with more accuracy. All samples exhibited two mass loss steps of thermal degradation (both related with exothermic events in DTA) and both were anhydrous, due to exhibited no step mass loss below 100.0 °C. The nonrecrystallized myo-inositol (MI) had thermal stability at 271.6 °C. The first mass loss occurred in the range of 271.6–405.1 °C (Dm = 86.19%), which is associated with an exothermic peak at 382.9 in the DTA curve. The second mass loss occurred at the 405.1–526.7 °C (Dm = 13.81%), related with exothermic event (426.8–547.0 °C) in DTA curve. The recrystallized myo-inositol (RMI) was stable up to 272.0 °C. Thus, the first mass loss occurred in the range of 272.0–411.2 °C (Dm = 88.45%), which is related with an exothermic peak at 398.6 °C in DTA; the second event of mass loss occurred at 411.2–535.6 °C (Dm = 11.55%), related with exothermic event (431.5–549.5 °C) in DTA curve. Hence, both samples had the same thermal stability and mass loss steps, and they both showed the endothermic peak at 222.2 °C in DTA. The cyclic DSC curve of inositol can be viewed in Fig. 1. In the first heating, an endothermic peak was observed at 227.9 °C (DHa-melt = 260.7 J g1), which was associated with the inositol melting, and is in agreement with the TG-DTA curves (Fig. S1). In the first cooling cycle, the inositol was recrystallized at 183.8 °C (Exothermic peak, DHcryst = 190.9 J g1). During the second heating cycle, an exothermic peak was located at 185.7 °C (DHsolid-solid = 31.1 J g1), which was recently related as a solid-solid transition to b-inositol (Izutsu et al., 2014). The transition is followed by two endothermic peaks at 221.4 °C (DHb-melt = 0.26 J g1) and 228.0 °C (DHa-melt = 260.3 J g1). The first endothermic peak has

not yet been described in literature, and it could be associated with b-inositol melting. The second is associated to a-inositol melting, which occurs in first heating too. This b-phase melting is in agreement with the PLTM analysis, which showed both melting processes (Fig. 2). The thermal behavior of inositol repeats in all cycles, and all the Enthalpies of each thermal event can be viewed in Table 1. In DSC Photovisual (Fig. 3) and PLTM (Fig. 4) analyses, the solid-solid transition was evident; this transition changes the sample structure during the heating-cooling cycle. In 183.8 °C, the change is abrupt. For better visualization, a video of all the cycle can be view in Supplementary material. The a-inositol diffractogram at 25 °C (Fig. 5a) had the relative intensity and the peak positions in agreement with the literature (Izutsu et al., 2014). The RMI diffractogram is shown in Fig. 5b, in which the peaks are very similar in comparison with the ainositol, except for some peaks at 2h = 12.6°, 17.9°, 18.12°, 25.4°, and 27.2°, indicating the presence of b phase (Izutsu et al., 2014). The third diffractogram refers to inositol after the solid-solid transition and shows the presence of characteristic peaks of phase: a (2h = 22.5°, 26.8°, 29.5°) and b (2h = 25.4°, 27.3°), which prove the presence of both phases. Thus, to understand better these changes in diffractograms, the Table S1 (Supplementary material) shows all the peaks of the three diffractograms, which appears in a and/or b phases. Associating the data of DSC curve and diffractograms made it possible to propose the thermal behavior of myo-inositol: During the first cooling, the inositol was recrystallized in both phases (a + b) at 183.8 °C. Then in the second heating, part of b phase was transformed to a phase at 185.68 °C, explaining the exothermic event; thus, the residual b phase melted at 221.43 °C and then a phase melt occurred at 228.04 °C. 3.2. Related kinetics studies for solid-solid transition of a-inositol The nonisothemal kinetic study is related to the reaction rate and is described by Eq. (1) (Hayama, Takahashi, Kikutake, Yokota, & Nemoto, 1999; Flynn & Wall, 1966):

da ¼ kðTÞf ðaÞ dt

ð1Þ

where da/dt is the reaction rate, a is the extent of the reaction, k(T) is the rate constant, t is time, T is temperature, and f(a) is the reaction model. Using the expression proposed by Ozawa-Flynn-Wall, it was possible to define the activation energy for each extent of reaction, Eq. (2) (Flynn & Wall, 1966; Ozawa, 1965; Aboulkas & El Harfi, 2008):

ln½b ¼ ln½A  5:331  1:052

Fig. 2. PLTM images: (a) a and b MI at 217.0 °C (b), melting of b MI at 218.6 °C (c), b MI melted and a MI solid at 221.8 °C (d), and Melting of a MI at 223.4 °C.

Ea RT a;i

ð2Þ

For the nonisothermal kinetic study, DSC curves at differents heating rates were made, as shown in Fig. S2. Increasing the heating rate dislocated the thermal events to higher temperatures and reduced the time of thermal event. This was expected because of the heat diffusing in the sample, notable for non-heat conducting materials, such as organic materials. The extent of reaction rates were adjusted in Ozawa-Flynn-Wall Method; Eq. (2) and a graph lnðbÞ vs 1000/T were plotted (Fig. S3). The relationship was observed between activation energy, Ea, and extent of reactions, a. The relationship between the two variables is shown in Fig. S4. Note that the activation energy assumed a constant value, Ea = 148.3 kJ mol1, until a = 0.7, when the activation energy decreased to 121.2 kJ mol1. The constant value of activation energy shows the occurrence of one thermal event, and the decrease of this Ea value in the final of the thermal event, could be explained by the reduction of b phase at system.

1152

R.T. Alarcon et al. / Food Chemistry 237 (2017) 1149–1154

Fig. 3. Photovisual images: MI at 25.0 °C (a); MI initial melting at 226.6 °C (b); MI total melting at 228.2 °C (c); recrystallization of MI in the cooling cycle at 183.0 °C (d); RMI before solid-solid transition in the second heating cycle at 170.6 °C (e); solid-solid transition at 185.0 °C (f); melting and evaporation of b-phase at 226.2 °C (g); and total melting of a-phase at 233.0 °C (h).

Fig. 4. PLTM images: RMI at 50.1 °C (a), RMI at 180.5 °C (b), RMI at 189.1 °C (c), solid-solid transition at 195.3 °C (d), Initial melting at 224.4 °C (e), and total melting at 226.0 °C (f).

Table 1

a-Melting Temperature Peak (Ta-pmelt), a-melt Enthalpy (DHa-melt), Crystallization Temperature Peak (Tpcryst), Crystallization Enthalpy (DHcrist), Solid-solid transition

Temperature peak (Tpsolid-solid), Solid-solid transition Enthalpy (DHsolid-solid), b-melting Temperature peak (Tb-pmelt) and b-melt Enthalpy observed in cycle DSC. Cycle step

Ta-pmelt/°C

DHa-melt/J g1

Tpcryst/°C

DHcrist/J g1

Tpsolid-solid/°C

DHsolid-solid/J g1

Tb-pmelt/°C

DHb-melt/J g1

Firts heating First Cooling Second Heating Second Cooling Third Heating Third Cooling

227.94 – 227.94 – 227.94 –

260.70 – 260.30 – 260.30 –

– 183.80 – 183.80 – 183.80

– 190.88 – 190.88 – 190.88

– – 185.68 – 185.68 –

– – 31.08 – 31.08 –

– – 221.43 – 221.43 –

– – 0.26 – 0.26 –

R.T. Alarcon et al. / Food Chemistry 237 (2017) 1149–1154

1153

important to the food industry and food formulations. All the thermal events were confirmed by DSC-Photovisual, PLTM, and PXRD, which found the presence of both phases in crystallization process and solid-solid transition. The kinetic date show that the choice of temperature that the inositol is submitted is very important due to the decrease of solid-solid transition lifetime with the increase of temperature. At 20 °C, the lifetime of inositol is 126 years and at 37 °C is 8 years for 5% solid-solid transition to occur. Acknowledgements The authors wish to thank CAPES (proc. 024/2012 Proequipment), POSMAT/UNESP, FAPESP (processes: 2012/21450-1 and 2013/09022-7) for financial support and PhD Rita de Cássia da Silva and Netzsch-Brazil to provided kinetic computational program (Thermokinetics 3.1).

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foodchem.2017. 06.072.

References Fig. 5. PXRD diffractograms: MI (a), RMI (b), and RMI after solid-solid transition (c).

To determine the kinetic model to understand better the experimental data, ThermoKinetics 3.1 software was used as previously mentioned. The best model found was the autocatalytic reaction. This result is in agreement with the images obtained in DSC photovisual, which show the quick change at 183.8 °C, suggesting that the formation of a phase increases the rate of solid-solid transition. All values of activation energy, standards deviations, and correlation coefficient obtained by program are shown in Table S2. The Activation Energy was calculated using the ASTM E1641 (1999), and the value obtained was 121.63 kJ mol1. Therefore, it was possible to calculate the inositol lifetime using the Eq. (3), suggested by ASTM 1877 (1999). The resulting graph is shown in Fig. S5.

log t f ¼

  Ea Ea a þ log 2:303RT f Rb

ð3Þ

The inositol solid-solid transition lifetime for 5% of b-phase convert to a-phase at 20.0 °C is about 126 years, whereas the lifetime decreases to 8 years, if the drug is submitted to 37.0 °C. This difference shows the importance of knowing the thermal behavior of inositol in order to choose better conditions for storage and transportation and avoid the formation of polymorphs. Consequently, the polymorphism can change several physico-chemical proprieties, such as density, hygroscopicity, vapor pressure, solubility, dissolution rate, melting and sublimation temperatures and stability. Hence, these factors are important in food formulations and storage of raw myo-inositol (Brittain, 2009, Chap. 1). 4. Conclusions The TG/DTG-DTA curves showed that the MI and RMI had thermal stability at 272.0 °C and both decomposed below 600.0 °C in two steps, with no material remaining. The DSC curves provided important data about a-Inositol melting, crystallization, solidsolid transition (185.68 °C), and a new thermal event associated to b-Inositol melting (221.43 °C). Hence, this information is

Aboulkas, A., & El Harfi, K. (2008). Study of the kinetics and mechanisms of thermal decomposition of Moroccan Tarfaya oil shale and its kerogen. Oil Shale, 25, 426–443. American Society for Testing and Materials – ASTM (1999). ASTM-E1641: Standard test method for decomposition kinetics by thermogravimetry. American Society for Testing and Materials – ASTM (1999). ASTM-E1877: Standard practice for calculating thermal endurance of materials from thermogravimetric decomposition data. Arendrup, K., Gregersen, G., Hawley, J., & Hawthorne, J. N. (1989). High-dose dietary myo-inositol supplementation does not alter the ischaemia phenomenon in human diabetics. Acta Neurologica Scandinavica, 80, 99–102. Babu, K. M., Church, R. J., & Lewander, W. (2008). Energy drinks: The new eyeopener for adolescents. Clinical Pediatric Emergency Medicine, 9, 35–42. Braga, D., Grepioni, F., Maini, L., & Polito, M. (2009). Crystal polymorphism and multiple crystal forms. Structure and Bonding, 132, 25–50. Brittain, H. G. (2009). Polymorphism in pharmaceutical solids (2th ed.). New York: CRC Press. Corrado, F., D’Anna, R., Di Vieste, G., Giordano, B., Pintaudi, B., Santamaria, A., & Di Benedetto, A. (2011). The effect of myo-inositol supplementation on insulin resistance in patients with gestational diabetes. Diabetic Medicine, 28, 972–975. D’Anna, R., Di Benedetto, V., Rizzo, P., Raffone, E., Interdonato, M. L., Corrado, F., & Di Benedetto, A. (2012). Myo-inositol may prevents gestational diabetes in PCOS women. Gynecology Endocrinology and Reproductive Medicine, 6, 440–442. De la Fuente, E., Sanz, M. L., Martínez-Castro, I., Sanz, J., & Ruiz-Matute, A. I. (2007). Volatile and carbohydrate composition of rare unifloral honeys from Spain. Food Chemistry, 105, 84–93. Flores, M. I. A., Moreno, J. L. F., Frenich, A. G., & Vidal, J. L. M. (2011). Fast determination of myo-inositol in milk powder by ultra high performance liquid chromatography coupled to tandem mass spectrometry. Food Chemistry, 129, 1281–1289. Flynn, J. H., & Wall, L. A. (1966). A quick, direct method for the determination of activation energy from thermogravimetric data. Polymers Letter, 4, 323–328. Food and Drug Administration (1985). Nutrient requirements for infant formulas. Federal Register, 50, 45106–45108. Giron, D. (2002). Applications of thermal analysis and coupled techniques in pharmaceutical industry. Journal of Thermal Analysis and Calorimetry, 68, 335–357. Govindarajan, C., Pitchaipillai, R., Shanmugasundaram, B., Thangam, S., Arokiasamy, J., & Pillai, M. S. (2015). Myoinositol: A review of its use in patients with polycystic ovary syndrome. World Journal of Pharmacy and Pharmaceutical Sciences, 4, 137–155. Hayama, T., Takahashi, K., Kikutake, K., Yokota, I., & Nemoto, K. (1999). Analysis of polymerization behavior of dental dimethacrylate monomers by differential scanning calorimetry. Journal of Oral Science, 41, 9–13. Hecht, S. S., Kenney, P. M. J., Wang, M., Trushin, N., Agarwal, S., Rao, A. V., & Upadhyaya, P. (1999). Evaluation of butylated hydroxuanisole, myo-inositol, curcumin, esculetin, resveratrol, and lycopene as inhibitors of benzo[a]pyrene plus 4-(methylnitrosamino)-1-butanone induced lung tumorigenesis in A/J mice. Cancer Letters, 137, 123–130. Holub, B. J. (1986). Metabolism and function of myo-inositol and inositol phospholipids. Annual Review of Nutrition, 6, 563–597.

1154

R.T. Alarcon et al. / Food Chemistry 237 (2017) 1149–1154

Howlett, A., Ohlsson, A., & Prakkal, N. (2015). Inositol in preterm infants at risk for or having respiratory distress syndrome. Cochrane Database of Systematic Reviews, 2, 1–35. Indyk, H. E., Lawrence, R., & Broda, D. (1993). The micronutrient content of bovine whole milk powder: Influence of pasture feeding and season. Food Chemistry, 46, 389–396. Izutsu, K. I., Yomota, C., Okuda, H., Kawanishi, T., Yamaki, T., Ohdate, R., ... Terada, K. (2014). Effects of formulation and process factors on the crystal structure of freezedried myo-inositol. Pharmaceuticals, Drug Delivery and Pharmaceutical Technology. Loewus, F. A., & Murthy, P. P. N. (2000). Myo-Inositol metabolism in plants. Plant Science, 150, 1–19. Marsilio, V., Campestre, C., Lanza, B., & De Angelis, M. (2001). Sugar and polyol compositions of some European olive fruit varieties (Olea europea L.) suitable for table olive purposes. Food Chemistry, 72, 485–490. Netzsch – Thermokinetics (2017). Accessed 01.31.17 www.netzsch-thermalanalysis.com/us/products-solutions/software/netzsch-advanced-software/ thermokinetics. Önay-Uçar, E., Arda, N., Pekmez, M., Yilmaz, A. I., Böke-Sarikahya, N., Kirmizigül, S., & Yalçin, A. S. (2014). Comparison of antioxidant capacity, protein profile and carbohydrate content of whey protein fractions. Food Chemistry, 150, 34–40.

O’Neil, M. J., Heckelman, P. E., Koch, C. B., Roman, K. J., Kenny, C. M., & D’Arecca, M. R. (2006). The merck index (14th ed.). New York: Rahway. Ozawa, T. (1965). A new method of analyzing thermogravimetric data. Bulletin of the Chemical Society of Japan, 38, 1881–1886. Sabater, C., Prodanov, M., Olano, A., Corzo, N., & Montilla, A. (2016). Quantification of prebiotics in comercial infant formulas. Food Chemistry, 194, 6–11. Sanz, M. L., Sanz, J., & Martínez-Castro, I. (2004a). Presence of some cycliotols in honey. Food Chemistry, 84, 133–135. Sanz, M. L., Villamiel, M., & Martínez-Castro, I. (2004b). Inositol and carbohydrates in different fresh fruit juices. Food Chemistry, 87, 325–328. Thompkinson, D. K., & Kharb, S. (2007). Aspects of infant food formulation. Comprehensive Reviews in Food Science and Food Safety, 6, 79–102. Turan, G. A., Eskicioglu, F., Sivrikoz, O. N., Cengiz, H., Adakan, S., Gur, E. B., ... Yilmaz, O. (2015). Myo-Inositol is a promising treatment for prevention of ovarian hyperstimulation syndrome (OHSS): An animal study. Gynecology Endocrinology and Reproductive Medicine. Unfer, V., Carlomagno, G., Dante, G., & Facchinetti, F. (2012). Effects of Myo-inositol in women with PCOS: A systematic review of randomized controlled trials. Gynecological Endocrinology, 28, 509–515.