7/26/2019 1-s2.0-S0002937814006899-main

1/5

Sacral nerve stimulation reduces elevated

urinary nerve growth factor levels in womenwith symptomatic detrusor overactivityDara Faye Shalom, MD; Nirmala Pillalamarri, MD; Xiangying Xue, MD; Nina Kohn, MBA, MA;Lawrence Russell Lind, MD; Harvey Allen Winkler, MD; Christine Noel Metz, PhD

OBJECTIVES:To investigate changes in urinary nerve growth factor

(uNGF) in women with symptomatic detrusor overactivity (DO) following

peripheral nerve evaluation (PNE) for sacral neuromodulation vs

controls.

STUDY DESIGN: There were 23 subjects with overactive bladder

symptoms and DO who failed management with anticholinergics and

22 controls consented to participate in this prospective pilot study.

Urine specimens were collected from controls at baseline for evalu-

ation of uNGF and creatinine. Subjects were evaluated at baseline and5 days after a trial of sacral nerve stimulation referred to as a PNE.

Each visit included urine collection for uNGF and, Incontinence Quality

of Life Questionnaire, Urinary Distress Inventory Questionnaire, post-

void residual volume, and a 3-day voiding diary. uNGF levels were

measured by enzyme-linked immunosorbent assay and expressed as

uNGF pg/creatinine mg.

RESULTS: Subjects with DO had significantly higher baseline

uNGF levels (corrected for creatinine) compared with controls

(19.82 pg/mg vs 7.88 pg/mg, P< .002). Seventeen DO subjects

underwent PNE and were evaluated at the end of the testing period.

There was a significant improvement in quality of life scores for

subjects after PNE compared with baseline (Urinary Distress In-

ventory Questionnaire: 7.0 vs 13.7, P< .001; Incontinence Quality

of Life Questionnaire: 87.3 vs 52.8, P < .0001). Concordantly,

uNGF levels significantly decreased from 17.23 pg/mg to 9.24 pg/mg

(P< .02) after PNE.

CONCLUSION: uNGF levels decrease with symptomatic response in DOsubjects undergoing PNE. DO subjects had significantly higher uNGF at

baseline vs controls, and uNGF levels significantly decreased after only

5 days of sacral nerve stimulation. These findings support a larger

study to validate the use of uNGF as an objective tool to assess

therapeutic outcome in patients undergoing PNE for sacral

neuromodulation.

Key words:detrusor overactivity (DO), overactive bladder (OAB),

sacral neuromodulation, urinary nerve growth factor (uNGF)

Cite this article as: Shalom DF, Pillalamarri N, Xue X, et al. Sacral nerve stimulation reduces elevated urinary nerve growth factor levels in women with symptomatic

detrusor overactivity. Am J Obstet Gynecol 2014;211:561.e1-5.

O veractive bladder (OAB) is a con-dition often characterized bydisabling symptoms of urinary urgencyand frequency with or without urge in-

continence. Detrusor overactivity (DO)

can be a cause of OAB and is diagnosedwhen involuntary detrusor contractionsare noted during the lling phase of

urodynamic testing. OAB is common,

with an estimated prevalence of 16% in

women in the United States, and is

known to adversely affect quality oflife.1,2 Treatment options for patients

with OAB have expanded throughout

the past decade. In patients who are re-fractory to rst-line treatment withbehavioral modications and anticho-

linergics, sacral neuromodulation can

be offered.This therapy is unique in that

a trial phase, referred to as a peripheral

nerve evaluation (PNE), is performed inorder to determine whether patients

will respond to therapy. During PNE,

patients receive 3-5 days of sacral nervestimulation via a temporary test stimu-lation lead placed in the S3 foramina.

The decision to proceed with perma-

nent implantation of the sacral neuro-modulation device, the implantable

pulse generator, is based primarily on

voiding diary results. A patient meetscriteria for implantation if they experi-

ence a 50% decrease in the number ofvoids and/or leakage episodes during

the testing phase.

Nerve growth factor (NGF) is a smallsecreted protein that promotes differ-

entiation and survival of target neurons.

NGF is produced in the bladder and

is reported to sensitize afferent nervesand induce bladder overactivity.3 This

From the Division of Female Pelvic Medicine and Reconstructive Surgery, Department of Obstetricsand Gynecology (Drs Shalom, Pillalamarri, Lind, and Winkler), North ShoreeLong Island JewishHealthSystem, Great Neck, andthe Centerfor Immunology andInammation, The Feinstein Institutefor Medical Research (Drs Xue and Metz), and Biostatistics Unit, Feinstein Institute for MedicalResearch (Ms Kohn and Dr Metz), Manhasset, NY.

Received March 10, 2014; revised June 6, 2014; accepted June 23, 2014.

L.R.L. is a consultant for Boston Scientic. The remaining authors report no conict of interest.

This research was supported by a grant from Medtronic Inc, Minneapolis, MN (D.F.S.).

Presented in oral format at the 40th Annual Scientic Meeting of the Society of GynecologicSurgeons, Scottsdale, AZ, March 23-26, 2014.

Corresponding author: Nirmala Pillalamarri, MD.npillalama@nshs.edu

0002-9378/$36.00 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2014.07.007

NOVEMBER 2014 American Journal of Obstetrics &Gynecology 561.e1

SGS Papers ajog.org

mailto:npillalama@nshs.eduhttp://dx.doi.org/10.1016/j.ajog.2014.07.007http://ajog.org/http://ajog.org/http://ajog.org/http://dx.doi.org/10.1016/j.ajog.2014.07.007mailto:npillalama@nshs.edu

7/26/2019 1-s2.0-S0002937814006899-main

2/5

bladder overactivity is linked to me-chanical stretch and reex bladder

muscle activity.3 NGF has been identi-

ed in the bladder urothelium, bladdersmooth muscle, and urine of patients

with OAB.3-5 Recent studies have dem-

onstrated that levels of urinary NGF(uNGF) correlate with patient reported

bladder symptoms3-6 and maydecreasefollowing successful treatment.3,4,6,7

The cause of OAB remains unclear,and currently disease severity and re-

sponse to treatment is assessed through

patient report. The lack of a standardizednoninvasive objective test to evaluate

disease progression and response totreatment makes the evaluation of pa-

tients with OAB exceedingly difcult. It

has been established that NGF is elevatedin the urine of patients with OAB and

DO, and has been shown to decrease

in patients who have a symptomaticresponse to therapy.3-7 We aim to con-

rm this nding in our patient popula-tion and to investigate the use of NGF as

a potential biomarker for evaluating

symptomatic response in patients un-dergoing PNE for sacral neuromod-

ulation. Our hypothesis is that OAB

patients with DO will have elevated NGF

levelsas compared with controls and thatthe level of NGF will decrease in patientswho have symptomatic improvement

following PNE.

MATERIALS AND METHODSSubjects and study design

This prospective caseecontrol study was

approved by the institutional review

board of the North ShoreeLong IslandJewish Health System (#11-060A) before

its initiation. This is a pilot study con-

ducted to assess the change in uNGFlevels in DO patients after sacral nerve

stimulation. All subjects (women) gavewritten informed consent before study

procedures. Subjects (cases and controls)

were recruited from the urogynecologypractice of the North ShoreeLong IslandJewish Health System between Aug. 11,

2011, until March 13, 2013. Cases were

enrolled in the study if they had OABsymptoms including urinary frequency,

urgency or urge incontinence for

greater than 3 months, a urodynamicdiagnosis of DO, and experienced no

improvement following treatment withanticholinergics and behavioral modi-

cations. Controls were age-matched

(5 years), and denied symptoms ofurinary frequency, urgency, or incon-

tinence. Exclusion criteria for both

cases and controls included: presence ofacute cystitis (conrmed by positive

urine culture), urinary tract tumorsor stones, bladder outlet obstruction,

postvoid residual (PVR) volume >100mL, history of urinary tract operation

(including urogynecologic procedures

such as slings) within 6 months, historyof intravesical botox usage within 1

year, interstitial cystitis, neurologicdisorder, and use of anticholinergics

within the past 21 days.

At baseline, a clean-catch urine spec-imen was collected from all controls for

determination of uNGF and creatinine

(Cr) levels. All cases completed a 3-dayvoiding diary, Incontinence Quality of

Life Questionnaire (I-QOL) and theUrinary Distress Inventory Question-

naire (UDI-6) at baseline. The I-QOL is a

validated 22-item quality of life instru-ment specic to urinary incontinence.8

Higher scores indicate a better quality

of life. The UDI-6 is a validated 6-item

questionnaire speci

c to incontinencein which higher scores indicate worsesymptoms.9 A clean-catch midstream

urine specimen was collected for deter-mination of uNGF and Cr levels, and a

PVR volume was measured with a

bladder ultrasound. Cases then under-went a 5-day PNE whereby an electrode

(Model 3057 Test Stimulation Lead;

Medtronic Inc, Minneapolis, MN) wasplaced into a unilateral S3 foramen un-

der uoroscopic guidance. The tech-

nique of the PNE has been describedpreviously.10 Proper S3 lead placement

was conrmed byuoroscopy as well aspatient sensation of stimulation and

direct observation of plantar exion of

the great toe using the external teststimulator (InterStim, Medtronic Inc).The test stimulation lead was then placed

into the test stimulation cable and test

stimulator (Model 3625; MedtronicInc). The external handheld test stimu-

lator was adjusted to achieve an opti-

mum level of sensation, and instructionson using the test stimulator were given.

Cases then began a 5-day trial of sacralnerve stimulation. A voiding diary was

completed each day of the PNE trial.

After 5 days, cases returned to the ofcefor follow-up and lead removal. The

follow-up visit included collection of a

clean-catch urine specimen followed bydetermination of PVR (as described

above), and completion of the I-QOLand UDI-6.

Urine processing and uNGF and

serum creatinine determinations

Urine samples were centrifuged within3 hours of collection at 3000 g for 10

minutes. The liquid supernatant was

separated into 1.5 mL aliquots andstored at 80 C. uNGF (performed in

triplicate) levels were measured using aspecic and highly sensitive enzyme-

linked immunosorbent assay method(Emax; Promega, Madison, WI), ac-

cording to the manufacturers sugges-

tions. Prior studies on uNGF reportedresults using the same enzyme-linked

immunosorbent assay kit for reproduc-

ibility.11 Sample concentrations weredetermined by extrapolating from the

uNGF standard curve. Urinary creati-nine levels were determined using the

enzymatic creatinine assay (DiazymeLaboratories, Poway, CA), according tothe manufacturers directions. Total

uNGF levels (pg/mL) were normalized

by urinary creatinine to overcome dif-fering dilutions of urine between sub-

jects. Data are presented as: mean

(uNGF [pg]/Cr [mg]) SD.

Statistical analyses

Comparisons between groups (cases vs

controls) for uNGF/Cr levels weremade using the ManneWhitney test.

Change in number of leaks and voids

per day, uNGF/Cr levels, and quality oflife measures before and after PNE werecompared in DO subjects using the

Wilcoxon signed-rank test. P values

< .05 were considered signicant.

RESULTSCharacteristics of the studypopulation

A total of 23 female subjects with OAB

symptoms and urodynamically provenDO met inclusion criteria and were

SGS Papers ajog.org

561.e2 American Journal of Obstetrics &Gynecology NOVEMBER 2014

http://ajog.org/http://ajog.org/

7/26/2019 1-s2.0-S0002937814006899-main

3/5

included in the analyses. Twenty-two fe-

male controls were selected and matchedto subjects based on age5 years.Baseline

demographics are listed inTable 1. The

mean age for subjects was 65.3 years(range, 24e86years)and the meanage forcontrols was 51.1 years (range, 29e73

years). There was 1 subject who was not

matched for age by a control.At baseline, subjects with DO had

signicantly higher mean uNGF/Cr levels

compared with controls (19.82 pg/mg

21 pg/mg vs 7.88 pg/mg 7.7 pg/mg,

P< .002). These ndings are shown in

theFigure.All DO subjects underwent PNE and

17 completed follow-up after 5 days of

test stimulation. Six subjects did notcomplete the testing phase. Of the 6subjects, 3 had lead displacement and 3

discontinued the test because of dis-

comfort from nerve stimulation. Leaddisplacement was diagnosed in patients

who perceived stimulation in areas other

than the vagina, anus, and perineum,and was conrmed by the clinician who

noted improper location of a lead. All 3patients with lead displacement reported

accidentally pulling the lead and noticing

a signicant change in sensation. Datawere collected on the 17 remaining

subjects who underwent the complete

5-day testing phase. All subjects had a>50% decrease in the number of dailyvoids and/or leakage episodes following

PNE. Descriptive data for number of

leaks and voids (per 24-hour period) onday 5 compared with baseline are shown

inTable 2. There was not only a signi-

cant reduction in symptom distress asmeasured by the UDI-6 (7.0 5.8 vs

13.7 4.4,P< .001), but also a statis-tically and clinically signicant rise in

health-related quality of life as deter-

mined by the I-QOL (87.3

23.5 vs 52.8 21.2, P< .0001) for subjects after PNE

compared with baseline. Concordantly,

mean uNGF/Cr levels signicantly de-creased from 17.23 pg/mg 20.7 pg/mg

to 9.24 pg/mg 7.1 pg/mg (P < .02)after PNE. In addition, mean PVR

volume signicantly decreased from

17.9 mL 19.1 ml at baseline to 2.1 mL 5.9 ml after PNE (P< .002). Results

are presented inTable 3.

COMMENT

In this pilot study, uNGF levels were

measured in subjects before and aftertreatment with sacral neuromodulation.We demonstrated that uNGF levels

decrease with therapeutic response in

patients with OAB and DO undergoingPNE. Mean uNGF levels were more than

twice as high in DO subjects at baseline

compared with those observed in nor-mal controls, and uNGF levels signi-

cantly decreased after only 5 days of

sacral nerve stimulation.Several previous studies report

similar uNGF ndings in DO patientsundergoing alternative therapies. Levels

of uNGF have been shown to decrease

following successful treatment with pa-tients who respond to therapy is unclear.Giannantoni et al12 reported that botu-

linum toxin-A reduces NGF levels in

subjects with DO by decreasing acetyl-choline release at the presynaptic level

and thus decreasing detrusor contrac-

tility and production of NGF. In patientswho respond to anticholinergics, the

TABLE 1

Baseline demographics

FactorMean (SD) subjects,n[ 23

Mean (SD) controls,n[ 22 Pvalue

Age, y 65.3 (17.3) 51.1 (15.2) .005

BMI (kilograms/meter2

) 32.9 (6.9) 26.9 (7.9) .033

Gravity 2.7 (2.2) 1.3 (1.1) .106

Parity 2.5 (1.8) 1.3 (1.1) .106

BMI. body mass index; SD, standard deviation.

Shalom. Sacral nerve stimulation reduces urinary nerve growth factor in detrusor overactivity. Am J Obstet Gynecol2014.

FIGURE

Baseline uNGF/Cr levels in subjects compared with controls

Cr, creatine; uNGF, urinary nerve growth factor.

Shalom. Sacral nerve stimulation reduces urinary nerve growth factor in detrusor overactivity. Am J Obstet Gynecol 2014 .

ajog.org SGS Papers

NOVEMBER 2014 American Journal of Obstetrics &Gynecology 561.e3

http://ajog.org/http://ajog.org/

7/26/2019 1-s2.0-S0002937814006899-main

4/5

authors postulate that the reduction inuNGF may be associated with suppres-sion of bladder smooth muscle stretch-

ing as a consequence of improvement in

DO symptoms.13 Although the precisemechanism of action of sacral neuro-

modulation is unknown, it is generallyaccepted that it works via the stimula-

tion of afferent rather than efferent

nerve bers.14 The proposed pathwayfor the role of NGF in DO is that NGF

sensitizes afferent C-bers leading to

sensory urgency and detrusor over-activity. It has also been suggested thatNGF induces bladder overactivity by

altering the expression of sodium or

potassium channels used by the bladderafferent bers.15 Therefore, it is plau-

sible that sacral neuromodulation, a

treatment designed to stimulate afferentnerves in the overactive bladder, would

alter uNGF levels.

The success rate of sacral nerve stimu-lation is reported to be in the range of 55%to 80%.14 This success is limited by the fact

that no predictive variables of outcome

have been identied. Currently, theresponse to sacral neuromodulation is

predicted based on patient reportedsymptoms and voiding diaries during the

PNE trial. In this study, we introduced

uNGF as an additional tool to evaluatetreatment response to PNE. All DO sub-

jects in our study population experienced

symptomatic improvement during the 5-day PNE. PVR was found to decrease aswell; however, this is not clinically relevant

as volumes of

7/26/2019 1-s2.0-S0002937814006899-main

5/5

Another limitation was our use ofPNE as the only testing method for

sacral neuromodulation. Six subjects

did not complete the PNE because oflead displacement or removal second-

ary to discomfort. At the time this

study was performed, PNE was thepreferred method of testing for sacral

neuromodulation in our urogynecol-ogy practice. This was secondary to the

ability to perform PNE in the ofcesetting as well as patient convenience.

As a result, this was the only method of

testing included in our study. Becausethat time we have started to perform

the stage 1 procedure more frequentlyas the risk of lead migration is minimal

with implantation of the tined lead.

Our small sample size was an addi-tional limitation. We are currently

conducting a larger study with long-

term follow-up to assess the changesin uNGF levels in DO patients after 12

months of sacral nerve stimulation.That data should provide additional

information on uNGF variability

within an individual. In addition, weare investigating the use of uNGF in the

testing phase (PNE) as a predictor of

short- and long-term treatment re-

sponse to sacral neuromodulation.Results from this pilot study areconsistent with previous research that has

demonstrated that uNGF is elevated inpatients withOAB/DO anddecreases with

symptomatic response to treatment. Over

the past few years evidence has accumu-lated from both animal and human

studies supporting the use of uNGF as a

potential biomarker for OAB and DO.Although these results are promising,

questions remain regarding the speci-

city, sensitivity, and cost-effectiveness ofuNGF as a tool for diagnosis and evalua-

tion of treatment response. Further

investigation is necessary before the

consideration of uNGF as a biomarker forthis condition. -

REFERENCES

1. Basra R, Kelleher C. Disease burden ofoveractive bladder: quality-of-life data assessedusing ICI-recommended instruments. Pharma-coeconomics 2007;25:129-42.2. Cisternas MG, Foreman AJ, Marshall TS,

Runken MC, Kobashi KC, SeifeldinR. Estimatingthe prevalence and economic burden of over-active bladder among Medicare beneciariesprior to Medicare Part D coverage. Curr MedRes Opin 2009;25:911-9.3. Liu HT, Chen CY, Kuo HC. Urinary nervegrowth factor levels in overactive bladder syn-drome and lower urinary tract disorders.J Formos Med Assoc 2010;109:862-78.4. Kuo HC, Liu HT, Chancellor MB. Can urinarynerve growth factor be a biomarker for over-active bladder? Rev Urol 2010;12:e69-77.5. Liu HT, Chen CY, Kuo HC. Urinary nervegrowth factor in women with overactive bladder

syndrome. BJU Int 2011;107:799-803.6. Liu HT, Chancellor MB, Kuo HC. Decrease ofurinary nerve growth factor levels after anti-muscarinic therapy in patients with overactivebladder. BJU Int 2009;103:1668-72.7. Liu HT, Chancellor MB, Kuo HC. Urinarynerve growth factor levels are elevated in

patients with detrusor overactivity anddecreased in responders to detrusor botulinumtoxin-A injection. Eur Urology 2009;56:700-7.8. Wagner TH, Patrick DL, Bavendam TG,Martin ML, Buesching DP. Quality of life of per-sons with urinary incontinence: development ofa new measure. Urology 1996;47:67-71.9. Uebersax JS, Wyman JF, Shumaker SA,

McClish D, Fantl JA. Short forms to assess lifequality and symptom distress for urinary incon-tinence in women: the Incontinence ImpactQuestionnaire and the Urogenital Distress In-ventory. Continence Program for WomenResearch Group. Neurourol Urodyn 1995;14:131-9.10. Janknegt RA, Weil EH, Eerdmans PH.Improving neuromodulation technique for re-fractory voiding dysfunctions: two-stageimplant. Urology 1997;49:358-62.

11. Seth JH, Sahai A, Khan MS, et al. Nervegrowth factor (NGF): a potential urinarybiomarker for overactive bladder syndrome

(OAB)? BJU Int 2013;111:372-80.12. Giannantoni A, Di Stasi SM, Nardicchi V,et al. Botulinum-A toxin injections into thedetrusor muscle decrease nerve growth factorbladder tissue levels in patients with neurogenicdetrusor overactivity. J Urol 2006;175:2341-4.13. Cho KJ, Kim HS, Koh JS, Kim JC. Changesin urinary nerve growth factor and prostaglandinE2 in women with overactive bladder afteranticholinergics. Int Urogynecol J 2013;24:325-30.14. Groen J, Bosch JLH, Mastrigt RV. Sacralneuromodulation in women with idiopathicdetrusor overactivity incontinence: decreasedoveractivity but unchanged bladder contractionstrength and urethral resistance during voiding.J Urol 2006;175:1005-9.15. de Groat WC, Yoshimura N. Afferent nerveregulation of bladder function in health anddisease. Handb Exp Pharmacol 2009;194:91-138.

ajog.org SGS Papers

NOVEMBER 2014 American Journal of Obstetrics &Gynecology 561.e5

http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref4http://refhub.elsevier.com/S0002-9378(14)00689-9/sref4http://refhub.elsevier.com/S0002-9378(14)00689-9/sref4http://refhub.elsevier.com/S0002-9378(14)00689-9/sref5http://refhub.elsevier.com/S0002-9378(14)00689-9/sref5http://refhub.elsevier.com/S0002-9378(14)00689-9/sref5http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://ajog.org/http://ajog.org/http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref15http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref14http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref13http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref12http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref11http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref10http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref9http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref8http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref7http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref6http://refhub.elsevier.com/S0002-9378(14)00689-9/sref5http://refhub.elsevier.com/S0002-9378(14)00689-9/sref5http://refhub.elsevier.com/S0002-9378(14)00689-9/sref5http://refhub.elsevier.com/S0002-9378(14)00689-9/sref4http://refhub.elsevier.com/S0002-9378(14)00689-9/sref4http://refhub.elsevier.com/S0002-9378(14)00689-9/sref4http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref3http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref2http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1http://refhub.elsevier.com/S0002-9378(14)00689-9/sref1