Human Reproduction Vol.21, No.9 pp. 2252–2256, 2006
Advance Access publication June 19, 2006.
2252 © The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: firstname.lastname@example.org
Metformin treatment is effective in obese teenage girls
Vincenzo De Leo1, M.C.Musacchio, G.Morgante, P.Piomboni and F.Petraglia
Department of Pediatrics, Obstetrics and Reproductive Medicine, Institute of Obstetrics and Gynecology, University of Siena, Siena, Italy
1To whom correspondence should be addressed at: Institute of Obstetrics and Gynecology, University of Siena, Policlinico Le Scotte,
Via Bracci, 53100 Siena, Italy. E-mail: email@example.com
BACKGROUND: Polycystic ovary syndrome (PCOS) is the most frequent cause of menstrual disorders in teenage
girls. Little information is available about the effects of metformin in adolescent girls with PCOS and its dose and its
efficacy in regulating menstrual cyclicity and hyperandrogenic symptoms. We evaluated the effects of metformin
treatment on ovulatory function, hirsutism, acne, hormonal patterns and body weight in adolescent girls with PCOS.
METHODS: Eighteen girls, ranging in age from 15 to 18 years, were enrolled in the study. Clinical diagnosis of
PCOS was based on the consensus criteria for PCOS accepted in May 2003 at Rotterdam. All subjects received 1700
mg/day metformin as tablets continuously for 6 months. They were then followed up for 6 months. RESULTS: Two
patients complained of side effects for >2 weeks and interrupted treatment; they were not evaluated. All the others
showed an improvement in menstrual cyclicity. Menstrual periods were ovulatory, with progesterone levels up to
6 ng/ml in luteal phase and a significant reduction in testosterone, androstenedione and free testosterone. BMI was
restored within normal limits in all girls between 21 and 24 kg/m2. Six months after the end of metformin treatment,
menstrual cycles continued to be regular and ovulatory with normal BMI. Side effects were slight. CONCLUSIONS:
The present results confirm the positive effects of metformin on menstrual periods and show that the drug can be
administered to young women to improve ovulation and hyperandrogenic symptoms such as hirsutism, acne and
Key words: adolescents/androgens/hirsutism/metformin/obese/PCOS
Polycystic ovary syndrome (PCOS) is the most frequent cause
of menstrual disorders in teenage girls. After menarche, young
women may show many initially minor or moderate symptoms
that aggravate with time. These symptoms include menstrual
irregularity, hirsutism, acne and obesity. Hormonal profile and
ultrasound examination indicate PCOS. Insulin-resistance tests
[homeostasis model assessment (HOMA) or oral glucose toler-
ance test (OGTT)] are now performed as well. Insulin resist-
ance and consequent elevated plasma concentrations of insulin
are reported to be associated with high androgen concentra-
tions and related symptoms such as anovulation, oligomenor-
rhoea, hirsutism, acne and seborrhoea in adolescent girls with
PCOS (Adams et al., 1986; Gilling-Smith and Franks, 1993).
More than 50% of girls with this syndrome are obese, and
many develop type 2 diabetes in later life (Chang et al., 1983;
Jalal et al., 1987). The clinical disorders and infertility have
been ascribed to high levels of insulin and LH, high LH/FSH
ratio and increased ovarian androgen production (Wajchenberg
et al., 1986; Waldestreicher et al., 1988; De Leo et al., 2003).
Hyperinsulinaemic insulin resistance and increased ovarian
cytochrome P-450 and 17,20-lyase enzymatic activity, two fea-
tures of PCOS, are pathogenically linked (La Marca et al., 2000).
Insulin has been shown to stimulate the proliferation of thecal
cells, increase LH-stimulated androgen secretion, increase
P450c17mRNA levels, up-regulate LH receptors and up-regulate
ovarian insulin-like growth factor-I (IGF-I) receptors (De Leo
et al., 2003). In PCOS patients, high local androgen concentra-
tions are responsible for anovulation by a direct effect on the
ovary (Ehrmann et al., 1995).
In the last 10 years, insulin-lowering drugs have become
widely used in the treatment of PCOS, particularly for the
induction of ovulation. Many studies have demonstrated the
efficacy of metformin, a biguanide normally used to treat non-
insulin-dependent diabetes, in inducing ovulation in PCOS
patients with insulin resistance (Nestler et al., 1998; De Leo
et al., 1999). Metformin is the most thoroughly investigated
insulin-lowering agent used to treat PCOS; it enhances insulin
sensitivity in the liver, where it inhibits hepatic glucose pro-
duction, and in muscle, where it improves glucose uptake and
use (De Leo et al., 2003).
One of the main reasons PCOS patients seek medical advice
is menstrual irregularity. Oligomenorrhoea and amenorrhoea
are usually linked to the absence of ovulation. Available
by guest on November 15, 2015
Metformin in adolescent girls with PCOS
studies indicate that 6 months of metformin therapy induces an
improvement in menstrual regularity in at least 50% of patients
(De Leo et al., 2003). Although confirmation is needed, it
seems that responders have high insulin levels, low androgen
levels and less severe menstrual abnormalities before therapy.
This suggests that the best candidates for metformin therapy
have oligomenorrhoea rather than amenorrhoea and are over-
weight rather than obese.
The relationship between PCOS, oligomenorrhoea, hyperan-
drogenism and hyperinsulinism has been well documented in
adults, but less in young women (Rosenfield et al., 2000; Van
Hooff et al., 2000, 2004). Hyperinsulinaemia has been docu-
mented in a small group of oligomenorrhoeic adolescents with
obesity (Apter et al., 1995) and in a group of lean anovulatory
or oligo-ovulatory hyperandrogenaemic adolescents (Ibanez
et al., 2001). Regarding medical treatment, little data are yet
available about the effects of metformin in young women with
PCOS, especially about its dose and ability to regulate men-
strual cyclicity and hyperandrogenic symptoms.
In this study, we evaluated the effect of metformin treatment
on ovulatory function, hirsutism, acne, hormone patterns and
body weight in 18 girls with PCOS.
Materials and methods
Eighteen obese girls with PCOS, ranging in age from 15 to 18 years,
were enrolled in the study. The study was approved by the ethical
committee of the University Medical School, and informed consent
was obtained from each patient. The clinical diagnosis of PCOS was
defined according to the consensus criteria for PCOS with the pres-
ence of clinical and biochemical signs of hyperandrogenism, chronic ano-
vulation and/or oligomenorrhoea and polycystic ovaries (The Rotterdam
ESHRE/ASRM-sponsored PCOS consensus workshop group, 2004).
Hirsutism was evaluated by the modified Ferriman–Gallwey method
(Hatch et al., 1981), and the mean score was 12 ± 2.5 (mean ± SD).
Menstrual bleeding occurred every 45–60 days. All women were
obese, with a BMI between 25.5 and 27 kg/m2.
Insulin resistance was diagnosed by OGTT. The criteria for hyper-
insulinemia were insulin basal levels >15 pmol/l and response >80
pmol/l at 90 min. All women had fasting insulin concentrations >15
pmol/l and cumulative insulin concentrations >48 000 pmol/min [area
under the curve (AUC) insulin during a 2-h 75 g oral glucose test].
Basal hormone concentrations [LH, FSH, androstenedione, testo-
sterone, free testosterone, estradiol (E2), 17-hydroxyprogesterone (17-
OHP), progesterone and sex hormone-binding globulin (SHBG)] were
evaluated every 4 days from days 7 to 45 of the cycle, the month
before metformin therapy was started and in the last month of therapy.
Before treatment, basal hormone levels indicated anovulatory cycles,
and androgens were at the upper limits of the normal range. All
women were normoprolactinaemic and had normal thyroid function.
None of the patients had virilization or congenital adrenal hyperplasia
(on the basis of normal levels of 17-OHP).
A baseline ultrasound scan of uterus and ovaries was performed.
The ultrasonographic diagnosis of PCOS was based on the presence of
≥10 follicles (2–10 mm in diameter) in one or both ovaries.
The patients entered the study on days 6–8 of induced menstrual
bleeding. All received 1700 mg/day metformin (Metforal; Guidotti,
Pisa, Italy) as tablets uninterruptedly for 6 months. Hirsutism was
evaluated by a modified form of the Ferriman–Gallwey method that
has been used in our clinic for many years. A score was assigned to
patients before and after 6 months of metformin therapy. A score >8 is
indicative of hirsutism. Patients were followed up for 6 months with
plasma progesterone determinations on day 22 of the menstrual cycle.
Plasma LH, FSH, E2, testosterone, androstenedione, free testosterone,
17-OHP, progesterone and SHBG were measured by double-antibody
radioimmunoassay using Radim Kits (Rome, Italy) for LH and FSH,
Sorin Kits (Saluggia, Italy) for androstenedione, progesterone and tes-
tosterone, DPC Kits (Los Angeles, CA, USA) for SHBG, free testo-
sterone and 17-OHP and Biodata Kits (Rome, Italy) for E2. Free
androgen index (FAI) was calculated as testosterone/SHBG × 100.
Samples were assayed in duplicate at two dilutions. All samples from
each subject were assayed together. Quality-control pools at low,
medium and high hormone concentrations were included in each assay.
The assay detection limits were 0.2 IU/l for LH, 0.18 IU/l for FSH,
18 pmol/l for E2, 0.16 nmol/l for progesterone, 0.52 pmol/l for free tes-
tosterone, 277 pmol/l for testosterone, 104 pmol/l for androstenedione,
0.21 nmol/l for 17-OHP and 0.2 nmol/l for SHBG. The intra- and inter-
assay variations were 7.8 and 8.2% for LH, 6.2 and 6.5% for FSH, 4.2
and 4.9% for E2, 8.5 and 10.8% for progesterone, 3.4 and 4.6% for
testosterone, 3.2 and 3.4% for free testosterone, 5.6 and 6.4% for
androstenedione, 4 and 4.8% for 17-OHP and 6.9 and 13% for SHBG.
Analytical methods were highly specific for each hormone with an
extremely low cross-reactivity (<0.05%) to other naturally occurring
hormones or therapeutic drugs that may be present in samples.
To assess responses to metformin therapy, we compared basal hormone
concentrations, maximum increments (maximum rise above baseline)
and AUC (cumulative rise above baseline) using the non-parametric
Wilcoxon test. Differences were considered significant for P < 0.05.
The drug was well tolerated by all patients. Two patients who
complained of side effects for >2 weeks and interrupted treat-
ment were not evaluated. The others showed an improvement
in menstrual cyclicity. After the first month of treatment, men-
strual bleeding returned to once a month within 30 days of the
start of the second month of therapy. Menstrual periods were
ovulatory, with progesterone levels up to 6 ng/ml in luteal
phase and a significant reduction in testosterone, androstenedione
and free testosterone. The Ferriman–Gallwey score for hirsutism
decreased significantly from 12 ± 2.5 to 7 ± 1.2 (mean ± SE) after
6 months of metformin treatment (Figure 1). In many girls,
acne and/or seborrhoea were greatly alleviated after 6 months
of treatment; only in 4 patients did mild acne persist, especially
before menstrual bleeding. BMI was restored within normal
limits in all girls between 21 and 24 kg/m2. Six months after
the end of metformin treatment, menstrual cycles continued to
be regular and ovulatory with normal BMI (Table I).
Table II summarizes hormone concentrations during follicu-
lar phase before metformin therapy and at the sixth month of
by guest on November 15, 2015
V.De Leo et al.
treatment. Significant reduction (P < 0.01) in basal levels of
LH, androstenedione, testosterone, free testosterone and LH/
FSH ratio was observed. SHBG concentrations showed a signi-
ficant increase in all women.
Figure 2 shows the patterns of plasma concentrations (medians
and ranges) of LH and FSH measured every 4 days for 5 weeks
before therapy and during the sixth month of treatment. Before
metformin treatment, basal LH levels were high and without a
peak, having a median value of 12.1 mIU/ml (range 16.2–7.5) for
all cycles. During the sixth month of therapy, LH showed a reduc-
tion in basal levels and an ovulatory peak at days 13–15 in all
women. FSH also showed a small increment in follicular phase
during treatment with a small peak on the same days as LH.
Before metformin treatment, the LH/FSH ratio was 2.23 (range
5.2–0.9). During therapy, LH showed a reduction in basal levels,
and the LH/FSH ratio decreased.
Before metformin therapy, E2 and progesterone showed con-
stant patterns. During metformin therapy, E2 and progesterone
patterns reflected ovulatory cycles, with median luteal-phase
progesterone levels of 6100 ± 1400 pg/ml (Figure 3).
Plasma androgen levels were high before treatment. During
metformin treatment, testosterone, free testosterone and
androstenedione showed significant reductions in all women.
The mean reduction was 22% for testosterone, 30% for free
testosterone, 22% for androstenedione and 35% for FAI.
Fasting insulin levels and AUC insulin were low during met-
formin treatment with basal plasma levels <5 mIU/l.
Progesterone levels during follow-up (day 22 of the cycle)
showed maintenance of ovulation with progesterone concen-
trations >5000 pg/ml in all women.
The results of this study confirm the beneficial effects of met-
formin on ovulation in young women with PCOS and provide
new information about the drug’s hormonal and clinical effects
in this disorder.
Figure 1. A graph showing the Ferriman–Gallwey score in patients
with polycystic ovary syndrome (PCOS) before and after 6 months of
metformin treatment. *P < 0.01.
before metforminafter metformin
Ferriman and Gallwey score
Table I. Clinical parameters of the 16 women evaluated before and after 6
months of metformin treatment
After sixth month
Days of cycles (range)
BMI (kg/m2) (range)
Ferriman–Gallwey score (mean ± SE)
12 ± 2.5
7 ± 1.2
Table II. Hormone concentrations (medians and ranges) in blood samples of
16 women drawn on day 8 of the menstrual cycle before metformin therapy
and on day 8 of the cycle in the sixth month of treatment showing significant
reductions (P < 0.01) in LH, androstenedione, testosterone, free testosterone
and the LH/FSH ratio
Sex hormone-binding globulin (SHBG) concentrations showed a significant
increase in these women.
Sixth month of treatment
Free testosterone (pg/ml)
Free androgen index (FAI)
Estradiol (E2) (pg/ml)
Figure 2. Plasma LH and FSH levels in polycystic ovary syndrome
(PCOS) patients before metformin therapy and during the sixth month
48 12 162024 283236 40
48 12 1620 24 2832 3640
by guest on November 15, 2015
Metformin in adolescent girls with PCOS
Insulin resistance has long been recognized as a feature of
PCOS (Chang et al., 1983; De Leo et al., 2003). Insulin res-
istance and elevated plasma levels of insulin are associated
with high androgen concentrations in PCOS patients. There is
evidence that insulin stimulates ovarian androgens in vitro
(Poretsky et al., 1999). Insulin inhibits liver secretion of
SHBG, increasing the availability of androgens (Preziosi et al.,
1993). Another mechanism of action could be the inhibition of
liver synthesis of IGF-I in the ovary, directly influencing ovar-
ian steroidogenesis through a specific receptor (Poretsky et al.,
After menarche, adolescent women show a transient state of
anovulation. This period is considered physiological within the
ontogeny of fertility (Metcalf et al., 1983). However, in certain
young women, regular menstrual cycles fail to occur within
2–3 years of menarche and the anovulatory condition persists.
The prevalence of PCOS in women aged 18–25 years is 8%
(Michelmore et al., 1999). However, studies suggest that
PCOS begins in early puberty and evolves through adolescence
into adulthood (Van Hooff et al., 2004).
The most common treatment in these patients worldwide is
the contraceptive pill, especially those with anti-androgenic
progestins (Lemay and Poulin, 2002; Wiegratz and Kuhl,
2002). The clinical outcome is good during therapy, but after
termination, the symptoms recur and the young women
undergo an increase in body weight.
Many articles have shown that metformin induces a signific-
ant improvement in insulin levels and menstrual cycles in
oligo-amenorrhoeic teenagers with PCOS (Ibanez et al., 2001).
Six months of treatment was followed by significant reductions
in serum lipids and hirsutism (Ibanez et al., 2000), indicating
that insulin lowering has a normalizing effect on multiple aber-
rations in the endocrine metabolic profile of adolescents with
ovarian hyperandrogenism. Metformin treatment of 15 obese
adolescents with impaired glucose tolerance and PCOS for
6 months was found to be beneficial for glucose tolerance,
insulin sensitivity, insulinaemia and elevated androgen levels
(Arslanian et al., 2002).
Our present study confirms the positive effects of metformin
on menstrual periods and shows that the drug can be adminis-
tered in young women to improve hyperandrogenic symptoms
such as hirsutism and acne, as well as ovulation. The beneficial
effect of metformin on acne and hirsutism in adolescents could
therefore be due to the restoration of ovulation and the normal-
ization of estrogens.
It is recognized that high insulin levels exert anabolic effects
and modify fat distribution: our results showed a significant
reduction in BMI to <25 kg/m2. This weight reduction may be
due to the normalization of plasma insulin resulting in reduced
appetite. Our young women were therefore prescribed a low-
carbohydrate high-protein diet to reduce BMI, restore ovula-
tion and reduce acne and hirsutism. It has been shown that
weight loss, accompanied by an increase in insulin sensitivity,
can improve metabolic and hormonal abnormalities character-
istic of the PCOS (Pasquali et al., 1989; Anderson et al., 1995).
Finally, early pharmacological intervention with metformin
could possibly prevent manifestation of the complete spectrum
of PCOS in young overweight girls. The persistence of regular
ovulatory menstrual cycles in the 6 months after the end of
treatment demonstrates that metformin treatment provides last-
ing benefits. All girls maintained a BMI <25 kg/m2, and this
can play a role in normal ovulation menstrual cycles. These
patients are under follow-up for weight changes and to see how
long normal menstrual cycles last.
This study shows the efficacy of metformin in adolescent
women with anovulation and moderate obesity, confirming
previous trials and contributing important information about
the preventive effect of the drug on full manifestation of PCOS
in young women. In these patients, metformin could well be a
more appropriate treatment than estroprogestin pills.
Adams J, Polson DW and Franks S (1986) Prevalence of polycystic ovaries in
women with anovulation and idiopathic hirsutism. Br Med J 293,355–359.
Anderson P, Seljeflot I, Abdelnoor M, Arnesen H, Dale PO, Lovik A and Bir-
keland K (1995) Increased insulin sensitivity and fibrinolytic capacity after
dietary intervention in obese women with polycystic ovary syndrome.
Apter D, Bützow T, Laughlin A and Yen SS (1995) Metabolic features of
polycystic ovary syndrome are found in adolescent girls with hyperandro-
genism. J Clin Endocrinol Metab 80,2966–2973.
Arslanian SA, Lewy V, Danadian K and Saad R (2002) Metformin therapy in
obese adolescents with polycystic ovary syndrome and impaired glucose tol-
erance: amelioration of exaggerated adrenal response to adrenocorticotropin
with reduction of insulinemia/insulin resistance. J Clin Endocrinol Metab
Chang JR, Nakamura RM, Judd HL and Kaplan SA (1983) Insulin resistance
in nonobese patients with polycystic ovarian disease. J Clin Endocrinol
De Leo V, La Marca A, Ditto A, Morgante G and Cianci A (1999) Effects of
metformin on gonadotropin-induced ovulation in women with polycystic
ovary syndrome. Fertil Steril 72,282–285.
Figure 3. Plasma estradiol (E2) and progesterone levels in polycystic
ovary syndrome (PCOS) patients before metformin therapy and dur-
ing the sixth month of treatment.
48 12 162024 28 32 3640
4812 1620242832 36 40
by guest on November 15, 2015
V.De Leo et al. Download full-text
De Leo V, La Marca A and Petraglia F (2003) Insulin-lowering agents in the
management of polycystic ovary syndrome. Endocr Rev 24,633–667.
Ehrmann DA, Barnes RB and Rosenfield RL (1995) Polycystic ovary syn-
drome as a form of functional ovarian hyperandrogenism due to dysregula-
tion of androgen secretion. Endocr Rev 16,322–353.
Gilling-Smith C and Franks S (1993) Polycystic ovary syndrome. Reprod Med
Hatch R, Rosenfield R, Kim MH and Tredway D (1981) Hirsutism: implica-
tions, etiology, and management. Am J Obstet Gynecol 140,815–826.
Ibanez L, Valls C, Potau N, Marcos MV and De Zegher F (2000) Sensitization
to insulin in adolescent girls to normalize hirsutism, hyperandrogenism, oli-
gomenorrhea, dyslipidemia, and hyperinsulinism after precocious pubarche.
J Clin Endocrinol Metab 85,3526–3530.
Ibanez L, Valls C, Ferrer A, Marcos MV, Rodriguez-Hierro F and de
Zegher F (2001) Sensitization to insulin induces ovulation in nonobese
adolescents with anovulatory hyperandrogenism. J Clin Endocrinol
Jalal I, Naiker P, Reddi K, Moodley J and Joubert SM (1987) Evidence for
insulin resistance in nonobese patients with polycystic ovarian disease after
inhibition of ovarian steroid secretion. Fertil Steril 45,327–333.
La Marca A, Egbe TO, Morgante G, Paglia T, Cianci A and De Leo V (2000)
Metformin treatment reduces ovarian cytochrome P-450c17α response to
human chorionic gonadotrophin in women with insulin resistance-related
polycystic ovary syndrome. Hum Reprod 15,21–23.
Lemay A and Poulin Y (2002) Oral contraceptives as anti-androgenic treat-
ment of acne. J Obstet Gynaecol Can 24,559–567.
Metcalf MG, Skidmore DS, Lowry GF and Mackenzie JA (1983) Incidence of
ovulation in the years after the menarche. J Endocrinol 97,213–219.
Michelmore KF, Balen AH, Dunger DB and Vessey MP (1999) Polycystic
ovaries and associated clinical and biochemical features in young women.
Clin Endocrinol 51,779–786.
Nestler JE, Jakubowicz DJ, Evans WS and Pasquali R (1998) Effects of met-
formin on spontaneous and clomiphene-induced ovulation in the polycystic
ovary syndrome. N Engl J Med 338,1876–1880.
Pasquali R, Antenucci D, Casimirri F, Venturoli S, Paradisi R, Fabbri F,
Balestra V, Melchionda N and Barbara L (1989) Clinical and hormonal
characteristics of obese amenorrheic hyperandrogenic women before and
after weight loss. J Clin Endocrinol Metab 68(1),173–179.
Poretsky L, Cataldo NA, Rosenwaks Z and Giudice LC (1999) The insulin-related
ovarian regulatory system in health and disease. Endocr Rev 20,535–582.
Preziosi P, Barrett-Connor E, Papoz L, Roger M, Saint-Paul M, Nahoul K and
Simon D (1993) Interrelation between plasma sex hormone-binding globu-
lin and plasma insulin in healthy adult women: the telecom study. J Clin
Endocrinol Metab 76,283–287.
Rosenfield RL, Ghai K, Ehrmann DA and Barnes RB (2000) Diagnosis of the
polycystic ovary syndrome in adolescence: comparison of adolescent and
adult hyperandrogenism. J Pediatr Endocrinol Metab 13,1285–1289.
The Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group
(2004) Revised consensus on diagnostic criteria and long-term health risks
related to polycystic ovary syndrome (PCOS). Hum Reprod 19,41–47.
Van Hooff MH, Voorhorst FJ, Kaptein MB, Hirasing RA, Koppenaal C and
Schoemaker J (2000) Polycystic ovaries in adolescents and the relationship
with menstrual cycle patterns, luteinizing hormone, androgens, and insulin.
Fertil Steril 74,49–58.
Van Hooff MH, Voorhorst FJ, Kaptein MB, Hirasing RA, Koppenaal C and
Schoemaker J (2004) Predictive value of menstrual cycle pattern, body mass
index, hormone levels and polycystic ovaries at age 15 years for oligo-
amenorrhoea at age 18 years. Hum Reprod 19,383–392.
Wajchenberg BL, Achando SS, Okada H, Czeresnia CE, Peixoto S, Lima SS
and Goldman J (1986) Determination of the sources of androgen overpro-
duction in hirsutism associated with polycystic ovary syndrome by simulta-
neous adrenal and ovarian venous catheterization: comparison with the
dexamethasone suppression test. J Clin Endocrinol Metab 63,1204–1210.
Waldestreicher J, Santoro NF, Hall JE, Filicori M and Crowley WF Jr (1988)
Hyperfunction of the hypothalamic-pituitary axis in women with polycystic
ovary disease: indirect evidence for partial gonadotroph desensitization. J
Clin Endocrinol Metab 66,165–172.
Wiegratz I and Kuhl H (2002) Managing cutaneous manifestations of hyperandro-
genic disorders: the role of oral contraceptives. Treat Endocrinol 1,372–386.
Submitted on February 2, 2006; resubmitted on April 4, 2006, April 27, 2006;
accepted on May 3, 2006
by guest on November 15, 2015