cervix and
TRANSCRIPT
Curren/ Women's Health Reviews, 2006, 2, 207-221
Uterine EMG and Cervical LIF - Promising Technologies in Obstetrics
Robert E. Garfield*, William L. Mader, Leili Shi, Shao Q. Shi and George R. Saade
University of Texas Medical Branch, Department of Obstetrics and Gynecology, Division of Reproductive Sciences, 301University, Route 1062, Galveston, Texas 77555, USA
Abstract: No current objective measures exist to determine uterine or cervical function during pregnancy, or to accuratelypredict term or preterit) labor and delivery. Uterine electromyography (EMG) and cervical light-induced fluorescence(LIF) studies were performed in term and pre-term animals and humans to evaluate the state and function of the uterusand cervix, and to determine the predictive capability of measurements made using these new technologies. Both uterineEMG and cervical LIF produced high positive and negative predictive values, compared to other currently-used methods.Uterine EMG and cervical LIF show great promise as diagnostic tests during pregnancy and for patient monitoring inobstetrics.
Keywords: Uterus, EMG, cervix, LIF, labor, parturition.
207
INTRODUCTION
Perhaps the most important health problem in obstetricstoday is pre-term birth and its associated complications.Preterm birth is the primary reason for infant morbidity andmortality, accounting for around 35% of all health carespending on infants [1]. Preterm birth affects more than 12%of births in the USA [2] and of these, 40-50% have beenrelated to preteens labor [2-4]. Pre-term neonates, withbirthweights less than 2500 g, represent about 10% of' thetotal number of babies born each year. The complications ofpre-term birth include significant neurological, mental,behavioral and pulmonary problems in later life. Among thepre-term survivors, the rate of neurological impairmentvaries from 10% to 20% and growth restriction occurs inapproximately 20% of the surviving infants. The develop-ment of effective methods to prevent or reduce the incidenceof' pre-term birth depends upon the understanding of themechanisms that initiate labor. The greatest burden ofpreterm birth is in developing countries, where much of thefatality and morbidity is secondary to infectious diseases [5].Increased preterm labor risks in undeveloped countriesinclude: malaria and hyperpyrexia, HIV suppression of theimmune system leading to additional infection with tubercu-losis, syphilis and intestinal parasites, and bacterialvaginosis.
PHYSIOLOGY
Labor has been viewed as the transition from an inactiveto an active muscle either by the addition of a uterotonin orwithdrawal of tonic progesterone inhibition [6]. Althoughpast models recognized the importance of progesterone incontrolling uterine quiescence, they neither defined preciselythe uterine stages of labor nor identified the mechanism ofaction of the hormones involved. In addition, the models of
*Address correspondence to this author at the University of Texas MedicalBranch, Department of Obstetrics and Gynecology, Division ofReproductive Sciences, 301 University, Route [062, Galveston, Texas77555, USA; Tei: (409) 772-7590; Fax: (409) 772-226 I;E-mail: rgarfielii;ulmb.edu
1573-40006 550.00+.00
parturition did not consider the changes in the cervix as animportant component of parturition. The results of experi-mental and clinical studies with progesterone and its antago-nists indicate that parturition (and uterine contractility) iscomposed of several major steps, with a relatively shortconditioning phase followed by active labor [7, 8]. Theconditioning step leading to the softening of the cervix takesplace in a different and longer time frame from the condi-tioning step of the myometrium, indicating that the myome-trium and cervix are regulated by independent mechanisms.
Our studies, as well as those of others, have demonstratedthat myometrial cells are coupled together electrically by gapjunctions composed of connexin proteins [9], The groupingof connexins provides channels of low electrical resistancebetween cells that facilitate pathways for the efficientconduction of action potentials. Throughout most of preg-nancy, and in all species studied, these cell-to-cell channelsor contacts are few, indicating poor coupling and decreasedelectrical conductance. This condition favors quiescence ofthe myometrium and the maintenance of pregnancy. At term,however, the cell junctions increase and form an electricalsyncytium required for effective contractions. The presenceof the contacts seems to be controlled by changing estrogenand progesterone levels in the uterus. As action potentialspropagate over the surface of a myometrial cell, the depolari-zation causes voltage-dependent Ca,} channels (VDCC) toopen. When this occurs, Ca t- enters the muscle cell down itselectro-chemical gradient to activate the myolilaments andprovoke a contraction. We have also demonstrated byreverse transcriptase polymerase chain reaction that theexpression of VDCC subunits in the rat myometriumincreases during term and pre-term labor [10]. The increasedexpression, which appears to be controlled by progesteronewithdrawal, may facilitate uterine contractility during laborby increasing portals for Ca z' entry.
The composition of the cervix is smooth muscle (10%)and a large component of connective tissue (90%) consistingof collagen, elastin and macromolecular components, whichmake up the extracellular matrix f I I ]. Many biochemical and
© 2006 Bentham Science Publishers Ltd.
208 Current Women's Health Reviews, 2006, Mal. 2, No. 3
Garfield at ul.
functional changes occur in cervical connective tissuetowards term [II-13]. This process of cervical ripeningresults in softening, effacement and finally dilation of thecervix. Ripening is required for the normal progression oflabor and delivery of the fetus. The exact mechanismscontrolling the cervical ripening process are largelyunknown.
The processes governing changes in the myometrium andcervix ultimately become irreversible and lead to active laborand delivery. However, there may be a point at which thecervix and uterus are electrochemically and physicallyprepared for delivery, but during which time there are noteffective contractions, nor perceptible dilation. We believe itis during this critical interim phase that there exists the finalopportunity to effectively treat the uterus or cervix withtocolytics (in order to prevent pre-term labor) by reversingthe process. After the interim phase, which may beexceedingly short, it could be that the uterus is hormonallystimulated to contract (or alternatively freed from inhibitionto contract) and the cervix will dilate as a result. Once trueactive labor has started, delivery might not be delayed formore than a few days in humans because the changes, whichbegin in the preparatory phase, and culminate at the end ofthe interim phase, have by this time become well establishedand cannot be undone, even with currently availabletocolytics (Fig. 1).
In our opinion, the key to understanding parturition, andto developing suitable treatment methods, is to understandthe processes by which the myometrium and the cervixundergo these conditioning or conversion stages.
Unfortunately, currently available methods, includingthose that are based solely on monitoring contractions or oncervical examination, cannot conclusively detect whether apatient has entered the conditioning step, because changes inthese variables may be independent of the preparatory stage,
or may not become detectable by these methods until arelatively late and irreversible stage has been reached,
MONITORING DURING PREGNANCY (ANDDIAGNOSIS OF LABOR)
Diagnosis of labor is one of the most difficult andimportant tasks facing medical practitioners in maternity caretoday, Knowing that true labor (which will lead to delivery)has begun, as well as predicting when it will start, isimportant for both normal and aberrant pregnancies.Prediction of labor in normal pregnancies is important forminimizing unnecessary hospitalizations, interventions andexpenses. On the other hand, accurate prediction anddiagnosis of spontaneous pre-term labor will also allowclinicians to start treatment early in women with true laborand avert unnecessary treatment and hospitalization inwomen who are simply having pre-term contractions, butwho are not in true labor. Even noticeable dynamic cervicalchange may not be an accurate indicator of true labor, as alarge percentage of women with established cervical changedo not deliver pre-term when not treated with tocolytics [14].To date, the most important key to preventing pre-term laborhas been constant contact and care from health carepractitioners [15].
The state of labor monitoring, with currently-usedclinical methods, can be summarized as follows:
1. current methods are subjective and/or inaccurate;
2. intrauterine pressure catheters are limited by invasive-ness and need for ruptured membranes;
3. uterine contraction monitors (e.g. tocodynamometer -TOCO) are uncomfortable, inaccurate and/or subjective;
4. no method has been successful at predicting pre-termlabor;
UTERUS
CERVIX
Gap junctions increaseMyametrial Ion channels changeMyometrial receptorsleffectorschangeContinued cervicalsoftening
Not reversible.
r '
Can be slowed:
^I
tocolysis
Stimulation-induced
Not reversibleCan be slowed:
'analysisStrong, effectiveuterine contractionsCervix dilated
True LaborPhase
Duration = days
Duration = hoursDuration = months
Interimphase
Interimphase
Very long preparatory phase for labor
True LaborPhase
Uterus preparedfor Labor contractionsCervix ripeDilation may begin
Hormonal changesGrowth of uterusSlow cervical softening
Notreversible
---ai
1/
Fig. (1). Model of parturition depicting the chronic changes of the cervix and the acute changes in the myometrium. Some changes arereversible while others are not. Still others may be slowed or prevented by using treatments.
I..
Uterine L))1G and Cervical LIF - Promising Technologies in Obstetrics'
Current l1'onren'sHealth Reviews, 2006, Vol, 2, \'a. 3 209
3
)
5. no method has lead to effective treatment of pre-termlabor; and
6. no method makes an objective measurement of bothfunction and state of the uterus or the cervix duringpregnancy.
Few current assessment methods can identify signs ofimpending labor, and none of them offer objective data thataccurately predict labor. Uterine electromyography (EMG)and cervical light-induced fluorescence (LIF) are twoemerging alternative technologies that may prove to bevaluable contributions to the solution of this problem (Figs.2A and 213).
In the same way that modern electrocardiography (ECG)has clearly benefited heart patients [16-21], uterine EMGwhich, similarly, amounts to the acquisition of electricalsignals taken non-invasively from surface electrodes, wouldlikely benefit obstetric patients when adopted by physicians,and when utilized as an everyday tool in the antenatal andlabor wards and clinics. Like the function of the ECG forcardiac monitoring, uterine EMG can be used to monitoruterine electrical recordings in normal pregnancies in orderto diagnose or even predict abnormal conditions such as pre-term labor, insufficient labor progress, or dystocia [22]. Thishas been done by analyzing several different types ofelectrical parameters derived from the raw recorded uterine
00
C0mc0U
ao6J3
10050
0-50
-100
0 0w '2W
100
r5
50
253 0 0
4 0 0
5 0 0
6 0 0
700T I M E ( S e c o n d s)
Fig. (2A). The lypical uterine EMG setup consists of electrodes placed near the navel close to the midline (which has been determined toyield, generally, the best conduction path from the myometriuln to the surface, likely because subcutaneous tissue is thinnest here, especiallyin late pregnancy). The electrode sites are first prepared by removing excess oils with alcohol, and then by applying a conductive gel. Theelectrodes are connected to lead wires and cables which pass the uterine EMG signals on to an amplifier/filter. The uterine EMG signals arethen displayed on a monitor and stored in a computer for later analysis. Real-time analysis routines are currently being developed to yield apredictive measure while the patient is still being monitored. The tracings at the bottom show the correspondence of uterine EMG burstevents to contraction events recorded by tocodynamometer (TOCO) from a laboring patient. The action-potentials, as measured by uterineEMG, are actually responsible for the contractions of the uterus, and govern such characteristics as contraction strength, duration, andfrequency. The TOCO provides to clinicians only the number of contraction events per unit time and a crude, inaccurate measure ofcontraction force. Recent studies indicate that the TOCO possesses no objective predictive capability. By contrast, in addition to the numberof contraction events per unit time, the uterine EMG signals give critical information about the firing rate and number of action potentialsinvolved during a contraction. As such, the uterine EMG gives a direct measure of the state of myometrial development and preparedness forlabor and delivery.
210 Current Wamen'.s Health Reviews, 2006, Vol. 2, No. 3
Garfield of al.
340
400
460
520
580
Fig. (2B). Schematic of the Collascope, a device built specifically to measure cervical light-induced fluorescence, or LIF. The excitationlight source provides light of the proper frequency to elicit Fluorescence from cervical collagen pyridinium cross-links, a component whichdecreases as gestational age increases. The amount of fluorescence produced is proportional to the amount of collagen in the cervix, therebygiving a direct measure of cervical ripeness. The excitation light is applied to the cervix, and fluorescence light is collected from the cervix,via the hand-held probe. The fluorescent light is then separated into frequency components by a spectrometer, and intensity is measured. Theresults are displayed on a computer screen and stored for later analysis. The trace at the bottom shows the fluorescence spectrum of samplesof soluble and insoluble collagen, as measured by the collascope. The LIF-ratio value, of the cervix for example, is calculated by taking thepeak fluorescent value (which occurs at approximately 390nm wavelength), and dividing by the peak reference value (which occurs atapproximately 343nm wavelength). The LW-ratio has been found to be highly predictive of labor in various species, including humans.
electrical signals, including the power density spectrum(PDS), which decomposes electrical signals into theirindividual frequency components. On the other hand,fluorescence spectroscopy, a widely used research tool inbiosciences, can reveal important information with respect tomolecular and physical states [231. Fluorescence spectraprovide important details on the structure and dynamics ofmacromolecules and their location at microscopic levels,Fluorescence spectroscopy has been used to examine thecollagen content of a variety of tissues, including somecancers [211. We have used this LIE methodology tosuccessfully evaluate the cervix in a number of studies. Thiswas accomplished by calculating the amount of fluorescentlight emitted from the cervix at the specific fluorescencefrequency ofcollagen.
Once calculated, uterine [MG and cervical LIF valuescan be compared to gestational age, delivery outcome, ormeasurement-to-delivery intervals using receiver operatorcharacteristic (ROC) curves. This gives a measure of
predictability by determining the number of true and falsepositive and negative cases, which in turn is used to Findpositive and negative predictive values (PPV and NPV,respectively). The uterine EMG and cervical LIF monitoringmethods, therefore, could permit effective classification ofwomen, could enable much better treatment and manage-ment of' those women than the currently available tools, andcould also allow for labor prediction. Such forecasting oflabor in normal (term-delivery) pregnancies is important forminimizing unnecessary hospitalizations, interventions andexpense, while prediction and diagnosis of spontaneouspreterm labor will allow clinicians to start any necessarytreatment early in women with true labor, and avertunnecessary treatment and hospitalization in women who arenot in true labor. These two translational devices, used aloneor in conjunction, would therefore aid in patient monitoring.In the studies described below, we outline human and animalexperiments designed to test the use bf uterine EMG andcervical LIF.
L7erine E,l1G oral Cervical LIF- Prwnising Technologies in Obstetrics
Current Women '5 Health Reviews, 2005, t'ol. 2, .\'a. 3 211
UTERINE EMG STUDIES IN HUMANS
Prediction of Labor
We performed a human study wherein standard medicalelectrodes were used to non-invasively acquire uterine EMGdata from the abdominal surface. A technique known aspo^^er spectrum analysis was then performed on the uterineelectrical data from term and pre-term women who presentedwith signs and symptoms of labor, but in whom differen-tiation between false and true labor could not be madeclinically [25j. The average PDS peak frequency wasdetermined for each woman and plotted against the woman'smeasurement-to-delivery interval (Figs. 3A and 3B).
B 0.65N= 0.60
2 0.55a)3- 0.50
0.45 '
Q0.40•
0 0.35 •
a 0.3060
40
20
0Measurement-to-delivery
interval (days)
Fig. (3). A: An acute increase occurs in the PPS peak frequencyjust prior to delivery (within approximately 24 hours from delivery)in patients delivering spontaneously at term (?37 weeks gestation).Note that for clarity, individual data points are indicated For 40+hours. B: A similar increase occurs in the PDS peak frequency priorto delivery (within approximately 4 days of delivery) in patientsdelivering spontaneously prior to term (<37 weeks gestation). Notethat for clarity, individual data points are indicated for 10+ days.
Comparison of PDS peak frequency showed asignificantly higher value for term patients within 24 hoursof tabor and for pre-term patients within 4 days of labor, ascompared to non-labor controls (Figs, 4A and 4B). Our
results are supported by recent work from other groups [26 -29] that show either quantification of basal uterine activity orchanges in one or more uterine EMG parameters duringlabor, although it must be emphasized that sampling rates,electric/magnetic filtering characteristics, and analysistechniques in these other studies varied.
ROC curves were generated for endpoints of 8, 12, 24and 48 hours. Tables IA and 1B show the results of the ROCanalysis, including the corresponding PPV and NPV, respec-tively. From the tables, one sees that the PPV and NPV arehigh at the time-from-measurement-to-delivery endpointsinvestigated.
We concluded that labor (and subsequent delivery) couldbe predicted successfully using non-invasive uterine EMGby observing the change to higher frequencies in theelectrical signals. We believe that these higher frequenciesare attributable to a lowering of myometrial cellular thre-shold potentials, an increasing of myometrial cellular restingpotentials, increases in size and number of ion channels,and/or increases in gap junctions, any one or all of theseoccurring as the uterus becomes electrochemically preparedfor labor.
Longitudinal Changes
Patient uterine-burst EMG amplitude was shown toremain low and steady as gestation increases in longitu-dinally-measured patients (i.e. multiple measurements madeserially on the same patient as gestation increased), andbecoming maximal during labor. However, amplitude alonewas not successful in determining the onset of labor. Weperformed experiments to evaluate other EMG variables,such as PDS peak frequency and uterine burst electricalenergy for differentiating patient categories, forecastingparturition, and for monitoring patient progress longitudi-nally more effectively than traditional methods such astocodynamometry. The PDS peak frequency was seen to besignificantly lower for longitudinal patients early ingestation, and sharply increased for those who delivered atterm (Fig, 5).
We concluded that there is an increase in electricalfrequency (thought to be due to an increase in excitabilityand conduction), which occurs in the myometrium during itspreparation for labor.
Induction Effect
In one study, we compared the effects on the uterineEMG for the induction agents misoprostil and prepidil. Non-invasive surface electrodes were used to acquire the uterineEMG signals trans-abdominally, as with all of our humanstudies. The patients were grouped according to theinduction agent used. Uterine EMG activity from eachpatient was recorded for 30 minutes just prior to treatmentwith the induction agent, and was recorded for another 30minutes, 4 hours after treatment with the induction agent.Either misoprostol (50 µg, 6 patients) or prepidil (200 ltg, 5patients) was applied for cervical ripening. The meanelectrical activity of each of the 30-minute periods wascalculated and the increase or decrease in the mean activitywas found for each patient. The average increase in EMGactivity was calculated for each group. Misoprostil caused an
A 0.65N2 0.60
2 0.55 •a>'B. 0.50 .
0.45se
a 0.40•
0 0.35 .
0.30 .300 250 0200 150 100 50
Measurement-to-deliveryinterval (hours)
212 Current Women's Health Reviews, 2006, Vol. 2. No. 3
Garfield et ul.
B
0.70.7 P < .01 P < .05
*= 0.60.6
* u 0.50.5 r cw-r 0.40.4 g.
0.3 *-
0.3
0.2 N = 15 N=42 a 0.2 N = 31 N =11
0.1 O 0.1a
0.0 0.0> 24 hours
5 24 hours
>4Jays
54days
Fig. (4A,B). For patients delivering spontaneously at term, a significantly higher PDS peak frequency was seen for patients delivering within24 hours of EMG measurement as compared to those delivering more than 24 hours from EMG measurement (Fig. A). For patientsdelivering spontaneously preterm, a significantly higher PDS peak frequency was seen for patients delivering within 4 days of EMGmeasurement as compared to those delivering more than 4 days from EMG measurement (Fig. B).
Table IA. Predictive Capability of Uterine EMG (Term Patients)
Time to delivery PPV (%) NPV (%) Sensitivity Specificity P
2 Days 93.8 55.6 .918 .625 <.005
I Day 85.4 88.9 .976 .533 <.010
I2 flours 75.D 100,0 1.000 .542 <.001e.e
8 Hours 54.5 100.0 1.000 .394 <.005
Table IB. Predictive Capability of Uterine EMG (Pre-Term Patients)
Time to delivery PPV (%) NPV (%) Sensitivity Specificity P
6 Days 81.8 90.3 .750 .933 <.001
4 Days 85.7 B8.6 .600 .969 <.001
2 Days 71.4 88.6 .556 .939 <.001
1 Day 75.0 86.8 .375 .971 <.001
A
increase in uterine EMG activity for a greater percentage ofpatients compared to prepidil (Fig. 6A), and the averageincrease was only significant when using misoprostol (Fig.6B).
From this, we concluded that it is possible to use trans-abdominal EMG for quantitatively measuring uterine activityduring treatments.
UTERINE EMG STUDIES IN ANIMALS
The aim of this particular work was to analyze records ofuterine electrical activity made from the abdominal surfaceof pregnant and laboring rats to examine whether similarquantitative information can be extracted as from direct
recording from the uterine muscle [30]. Electrical activityduring pregnancy (non-labor, days 18 to 22), term labor (day22), and pre-term labor (onapristone injected on day 18,delivery on day 19) was measured with use of electrodesattached to the uterine wall and to the abdominal surface,The fast Fourier transform and wavelet transforms wereobtained for representative el ectromyographic bursts. Powerspectra were generated, Intrauterine pressure was alsomeasured. As a result of this investigation, several parame-ters were identified for use in following the progressiveincrease in uterine activity that occurs in preparation for andduring labor. Analyses of amplitude, frequency, and percenttime active were found to be effective methods for objec-
Uterine EMC and Cervical LIF - Promising Technologies in Obstetrics
Current Women's Health Reviews, 2006, Yni. 2, No. 3 213
0.8N
>, 0.7Ca.)10.6
U-0.5
03
CO 0.4O.dCO 0.3
25 Gestation30
al Age (a
(Weeks)Fig. (5). Antepartum patients who ultimately delivered spontaneously at term (measured at several time-points) tended to maintain relativelylow PDS peak frequency values, while labor patients delivering spontaneously at term had relatively high PDS peak frequency values.
A
B
U
n -Labor• -Antepartum
n n
40
MISOPROSTOL
PREPIDIL
i-?
N 0.12
0 0.10
0.08
0.060
W 0.04
Z 0.02
2 0.00
1009080TO60 .5040 •30 -2010 .0
Fig. (6A,B). A higher percentage of patients showed an increase in uterine EMG activity after treatment with Misoprostol than alter
treatment with Prepidil (Fig. A). The average increase in EMC activity was significantly higher (paired t-test) after Misoprostol treatment,but not after Prepidil treatment. When considering all treatments, there was a significant increase in uterine EMG activity after treatment ascompared to before treatment (Fig. B).
tively determining the efficiency of uterine contractions.Most of the changes in these parameters appeared in the last24 hours before delivery (Figs. 7 and 8).
Other similar unpublished studies also showed dramaticincreases in uterine EMG frequency and magnitude near andduring labor and delivery at term (Fig. 9a) and preterm (Pig.9b). Other groups have performed animal uterine EMGstudies, and their work also shows that uterine EMGparameters can quantify uterine contractile activity, and/orcan characterize changes therein [31-33].
We concluded that recording of uterine electromyo-graphic activity from the abdominal surface would be usefulin following the progression of pregnancy and in predictingand diagnosing labor by observing various parameters,including a change to higher frequencies in the electricalsignals. As with the human data, we believe that these higherfrequencies are attributable to a lowering of myometrialcellular threshold potentials, an increasing of myometrial
cellular resting potentials, increases in size and number ofion channels, and/or increases in gap junctions, any one orall of these occurring as the uterus becomes electro-chemically prepared for labor.
CERVICAL LIE STUDIES IN HUMANS
Prediction of Labor
Thirty-one (31) women, with inconclusive signs of labor,were assessed in the third trimester. Cervical LIF wasobtained non-invasively using an instrument specificallydesigned for this purpose (Collascope) [34]. LIF wasmeasured from the 12:00 position of' the cervix threesuccessive times and the results were averaged. In womenwith clinically inconclusive signs of labor, those whodelivered within 24 hours of measurement had significantlylower LIF than those who delivered more than 24 hours frommeasurement (mean [SD]: 0.571 [0.413] vs. 0.894 [0.228];P< 0.05; Fig. 10).
214 Current Women's Health Reviews, 2006, Vol. 2, No. 3
Garfield e1 r
A
Day 19
B
Day 21
500Ut
gV- 500
AS250
}tV C- 250
IUP60
C
Day 22 TL
D
500
}Al
-500
mmHg
ti
min
2 min
Fig. (7). Electromyographic recordings from uterus (Ut, upper trace), abdominal surface (AS, middle trace), and intrauterine pressure (IUP,bottom trace) from pregnant rats on day 19 of gestation, day 21 of gestation, labor at term (Day 22, TL), and preterm labor (Day 19, PL), afteronapristone administration. A, On day 19 of gestation, uterine electromyographic activity is irregular and of low amplitude with nodiscernible signal at the abdominal surface. There is low-amplitude change in intrauterine pressure that is not clearly related to electricalactivity. B, Record from rat on day 21, showing appearance of small bursts. C, Typical reca ,l from rat on day 22 of gestation during activedelivery. Uterine electromyographic activity was characterized by regular, high-amplitude, long-duration bursts of action potentials. D,
Preterm labor was induced on day 19 of gestation with onapristone. Note that the intensity of uterine activity produced is similar to thatoccurring during spontaneous term labor, C.
Day 19
Day 22 TL
Day 19 PL
A
Ut
B
ut
C
Ut
Ut
gV500
AS250 r
RV
250 LIUP
60
mmHg 0
mmHg
t_
0
2 minu2 min
Day 19 PL
µv
1iVs
Hz Hz
E
AS
F
AS
Hz
Hz
Hz
Fig. (8). Representative uterine and surface bursts of LMG activit) along with their power densit y spectra (power density in microvolt•secand frequency in hertz). A and D, On day 19 of gestation, power density spectra analysis showed low values with relatively flat power density
Uterine Eli/G and Cervical LIF- Promising Technologies . iu Obstetrics
Current Women '.v Health Ravines, 2096, Vol. 2, A'n. 3 215
(Legend ofFfg. 8) cunld....
spectra on both uterine (Ut) and abdominal surface (As) sites. B and E, Term labor (TL) was characterized by significant increases in uterineand surface electrical p-'ower. Uterine and surface EMG activity had similar power spectra, although surface power was lower. C and F,Preterm labor (PL) was also characterized by increases in electrical power of both uterine and surface activity. Peak elect r ical power andfrequency domain of EMG activity were similar for both uterus and surface.
A
B
IUP INTENSITY vs. EMGENERGY
NORMAL PREGNANCY GROU1800 R = .871 3500
P < .05r m
1200
z 2500 t7W m~ Ez 800 N
ee 1500400
0 .Delivery
50017 18 19 20 21 22 23
GESTATION (days)
(UP INTENSITY vs. EMGENERGY
PRE-TERM DELIVERY GROUP " 30001000-
R = .542 • 2500
P < .05
m2000 _
is
C mEGOO
i^r3
€
1500yel zm
}Hip N
800-
Fig. (9A,B). Dramatic and significant increases in the uterine EMC energy and intrauterine pressure are seen in term-delivering rats just priorto the onset of labor (on approximately day 22 of gestation). The two parameters were shown to correlate. Dramatic and significant increasesin the uterine EMG energy and intrauterine pressure are seen in preterm-delivering rats (treated with an induction agent) just prior to the onsetof labor (on approximately day 18 of gestation). The two parameters were shown to correlate.
1.4
1.2
1.00
0.8
0.6Ire
0.4
0.2
0.0>24 hrs
<24 hrsTime to Delivery
Fig. (10). A significantly lower average cervical LIF ratio wasobserved in those patients delivering within 24 hours of LIFmeasurement as compared to those delivering more than 24 hoursfrom measurement.
ROC analysis showed that LIF was predictive of deliverywithin 24 hours (area under ROC curve: 0.73; P< 0.01;sensitivity = 59%, specificity = 100%, PPV = 78.9% andNPV = 80.0%, at an L1F cutoff of 0.57).
We concluded from this study that cervical L1F valuesdecrease significantly as gestational age increases and arepredictive of delivery within 24 hours.
Longitudinal Changes
Another experiment involving cervical collagen wasconducted to estimate whether LIF correlates with the time-to-delivery interval [34]. Twenty-one healthy gravidae
without signs of labor had LIF measured approximatelyweekly during the last trimester. Each cervical collagenfluorescence measurement was accomplished in 10 seconds.For each patient and time point, the mean of threemeasurements from different spots around the 12 o'clockposition of the exocervix was found. In all, 98 LIF measure-ments (median, 4 measurements; range, 2-11 measurementsper patient) were taken. Overall, 13 of tile women hadalready delivered by the time analysis was complete, andthey were included in a correlation analysis for time todelivery. LIF measurements negatively correlated withgestational age (R= _ 0.340; P< 0.05) and with time todelivery (R= 0.370; P< 0.05; n= 13; Fig. 11).
We concluded that cervical collagen decrease, asmeasured non-invasively by LIF, occurs gradually over time,reaching a critical value just before labor and delivery.
Induction Effect
We compared cervical collagen changes after prostag-landin treatment in 41 patients [35]. Several cervical LIE'measurements were taken prior to treatment with inductionagent for each patient and averaged. Several measurementswere then taken and averaged from the same patients 4 hoursafter treatment with the induction agent. This study showedsignificant decreases (P<0.05) in L1F values of' the cervixfour hours after treatment (0.885-1-/-0.037) in patients withhigh values before treatment (0.982+1-0.040), but no changein patients with low values before treatment (Figs. 12A and128).
Paired Student's t-test, Wilcoxon Signed Rank test,Linear Regression, Spearman correlation and Fisher Exact
Garfield et r11.216 Crrrrenf f omen's Health Reviews, 2006, Vol, 2, No. 3
2.2
2.0 -
1.8 -
1.6 -
1 .4 -
LI- 1.2 -
22
24
26
28
30
32
34
36
38
Gestational age (weeks)
Fig. (11). A significant negative correlation was found between the LIF ratio and the gestational age.
42
p=O.f1013
p .6556
1.6
16
0.4
a 0sU
04
t 021L
0A
IA
12
i.0
09
46
02
an
A
before
after
PG-Apprenbort
before
B
after
PG-AppGration
Fig. (I2A,B). For those patients needing cervical induction treatment (i.e. cervical collagen content too high), a significant decrease was seenin the cervical collagen, as measured by LIE, when comparing average LIF values before and after treatment with prostaglandins (Fig, A).For those patients with ripened cervices (i.e. collagen content reduced), no change was seen in the cervical collagen, as measured by LIF,when comparing average LIF values before and after treatment with prostaglandins (Fig. B).
test were used as appropriate. The decrease in LIF correlatedwith the initial LIF before PG-application (r=0.61). TheBishop-Score increased in all 41 patients. No correlation wasfound between LIF versus Bishop-Score. Note that once thecervix is ripe, the LIF value is at its lowest, and no furtherdecrease will be seen (Fig. 12B, again).
From these studies, we concluded that there is value inthe quantification of LIF for identifying patients requiringinduction treatment.
CERVICAL LIF STUDIES IN ANIMALS
The aim of one experiment was to characterize changesin cervical resistance and collagen content during pregnancyin rats [36]. To determine the change in cervical resistance,non-pregnant and timed pregnant Sprague Dawley rats weresacrificed at various times. Their cervices were isolated and
suspended in organ baths connected to a cervimeter formeasurement of' the stretch-tension relationship. In adifferent group of animals, cervical collagen content wasmeasured using LIF in non-pregnant and, longitudinally, inpregnant rats, Cervical resistance and collagen contentdecreased progressively during pregnancy (Figs. 13A and13B).
The changes in cervical resistance mirrored those incervical collagen content and the nadir in both occurredabout two days prior to the onset of labor.
Our conclusion was that cervical preparation for deliverydoes not occur acutely at the time of labor and that cervicalcollagen content determines cervical resistance.
In another experiment, LIF was used to investigate in-vivo changes of cervical collagen in guinea pigs during
Uterine &VG and Cervical LIF- Promising Technologies in Obstetrics
Current 14'muen's Health Reviews, 2006, Vol. 2, No. 3 217
Nonpregnantmo+ During gestation15:2I Postpartum
$ 0.6 -o
-0.4 -
0.2 -
0.0 a-
NonpregnantDuring gestation
® Postpartum
5-
4-
3-
2-
0
1
B
O J J
N CO. LC) Qa
o a
ca aDays of Pregnancy and Postpartum
Fig. (134,B). Cervical resistance to stretch in non-pregnant (NP, empty bar), pregnant (day 12-22, black bars), and postpartum (PP, day 1-2,hatched bars) rats. Cervical resistance was represented by the slope of a straight line fitted through the linear part of the force vs.displacement curve. Data presented as mean ± SEM (a = 4-9 at each time point). Fluorescence intensity of rat cervical collagen in non-pregnant (NP, empty bar), pregnant (day 13-21), day 22 non-labor (NL), or during delivery (L), in black bars, and postpartum (PP day 1-3, 5,10, 12, 17, hatched bars) rats. Data presented as mean ± SEM (n = 6 at each time point).
gestation and following sodium nitroprusside treatment [371.LIF of cervical collagen was measured from the surface ofthe exocervix in anesthetized non-pregnant and timed-pregnant guinea pigs at different times of gestation.Measurements were also performed in guinea pigs at mid-gestation before, and 8 hours after, intra-cervical treatmentwith sodium nitroprusside. Collagen fluorescence decreasedsignificantly as pregnancy progressed, reached lowest valuesat delivery, and increased gradually postpartum. Treatmentwith sodium nitroprusside, but not with the vehicle, caused asignificant decrease in LIE.
We concluded that LIF changes in the cervix reflect thegradual cervical softening (ripening) during pregnancy andthe return to the rigid state of the cervix postpartum, and thatcervical softening during pregnancy, as well as after sodiumnitroprusside treatment, is associated with a decrease incervical collagen cross-links.
FORECASTING LABOR - CURRENT CLINICALMETHODS VS. EMG AND LIF
From our own studies, as well as those of others, it isclear that for forecasting labor and delivery, the effectivenessof the technologies now generally accepted in clinicalpractice are limited, especially in relation to sensitivity and
positive predictive value, as compared with EMG and LIF,as shown in Table 2 [38].
Infra-uterine pressure catheters cannot be used to predictlabor. The invasive nature of the procedure involved canincrease the risk of infection or cause serious complications.Such infections could be a risk factor for pre-term labor [39].No real predictive capability exists for intrauterine pressurecatheter devices because they are mainly used in cases wherelabor has already been diagnosed clinically, which is whythey do not appear in the Table above.
Tocodynamometer devices are used in over 90% of' allhospital births. Physicians have been quick to adopt thesemachines because they supply uterine contraction data withlittle risk. Ironically, most physicians also agree that thesebulky force-transducer devices provide limited informationon labor. The instruments are largely dependent upon theskill of the clinician, but have not changed treatments orimproved outcomes following pre-term labor. The tocodyna-mometer generally cannot be used to predict true pre-termlabor, as a recent study has shown [401. That is becausetocodynamometer devices are inaccurate. Many differentvariables affect the uterine external force measurement, suchas instrument placement, amount of fat, and uterine wall
218 Current Women's Heintz Reviews, 2006, Vol. 2, No. 3
Garfield et al.
Table 2.
Predictive Capability of Various Obstetric Technologies
Sens (Sensitivity)Test Sens Spec PPV NPV
MPTLS 50.0 63.5 21.4 86.4
Ctx >_ 4/hr 6.7 92.3 25.0 84.7Spec (Specificity)
BS
4 32.0 91.4 42.1 87.4
Cx < 25mm 40.8 89.5 42.6 88.8PPV {Positive predictive value)
CFFN + 18.0 95.3 42.9
i
85.6
EMG 75.0 93.3 88.0
93.0NPV (Negative predictive value)
L1F 59.0 100.0
i 78.9
80.0
Multiple pre-teen labor symptoms
- Maximal uterine contractions (tocodynamometer - TOCO)
- Bishop Score
Cervical fetal tibroncctin test
Cervical length (measured with ultrasound)
Noninvasive transabdominal uterine electrontyogram
Cervical Light-Induced Fluorescence
pressure. Tocodynamometry of uterine contractions isessentially equivalent to measuring cardiac contractions andheart function with a contraction transducer placed on thechest. Clearly, information about normal and abnormalmuscle contraction can never be achieved with these devices.Home uterine activity monitoring (HUAM), based onexternal tocodynamometer recordings, has been used in aneffort to predict spontaneous pre-term labor and decrease thefrequency of pre-term birth [41]. However, HUAM has beenshown to be no better in lowering the frequency of pre-termbirth than weekly contact with a nurse [42]. The lack ofsignificant clinical usefulness of HUAM is directly related tothe inability of the tocodynamometer and its user todifferentiate between false and true contractions (Fig. 14).
Non-labor
LaborvA%d-A.-q.)'''NN,11A
0 1 2 3 4 5 6Time (minutes)
Fig. (14). The tocodynamometer is ineffective at discerningbetween true and false labor, since contractions, as measured bythis device, can be large or small in amplitude, long or short induration, and can occur with the same frequency, regardless ofwhether the patient is in labor or not.
Contractions measured with the tocodynamometer can belarge or small, and can occur with a similar number ofcontractions/unit time, regardless of whether the patient is inlabor or not. The uterine EMG, in contrast, records differentelectrical signal amplitudes and frequencies (Fig. 1 5),depending upon the patient's labor or non-labor state.
Measuring the length of the cervix via endovaginalultrasonography has been used to detect spontaneous pre-term labor [43, 441, with some degree of success, However,even in combination with other factors, with respect topositive predictive capabilities, there is a range of possiblevalues reported, from around 50% to about 71%. And theseresults are obtained only after the onset of symptoms ofspontaneous pre-term labor, so this method, too, is highlylimited in its potential for early diagnosis and resultingtreatment.
Cervicovaginal fetal fibronectin (FFN) has recently beensuggested as a screening method for women at risk ofspontaneous pre-term labor. Several studies [45, 46] haveshown that FFN might be of benefit to predict spontaneouspre-term labor, but other studies indicate that FFN haslimited value [47]. The main value Of the FFN assay seemsto be only in its high NPV (i.e. it has the ability to identifypatients that are not at risk of spontaneous pre-term labor andpre-term birth, but can say nothing of the others). However,a high PPV is potentially important for the improved safetyof the fetus.
Our most recent pilot study characterized differences inuterine EMG and cervical LIF parameters used inconjunction, for patients undergoing successful or failedinduction. 12 patients presenting to the labor and deliveryarea for pitocin induction (for various indications, e.g. post-dates, oligo, etc.), had uterine electrical activity and cervicalcollagen content measured non-invasively using EJVIG andLIF respectively, just prior to treatment with pitocininduction agent (standard start dose, I MU), and then againapproximately 4 hours later. Patients were divided into two
?vMP'I -LSCtx
BS
CFFN
Cx
EMG
LIF
Uterine ElfG and Cervical LIF- Promising Teddrrolagies in Obstetrics
Current Women's Health Reviews, 201)6, Viol 2, No. 3 219
100
Non-labor- low frequencies- low amplitude
-100
Labor- high frequencies- high amplitude
-
1000 2 3 4 5 6
Time (minutes)
Fig. (15). Contraction electrical bursts, as measured by uterine EMG, are generally higher in amplitude and show higher action-potentialfrequencies during labor than during non-labor. This makes the device useful for prediction of labor and delivery.
groups: Gl-successful induction (N=8); G2-failed induction(N=4). Successful induction was deemed to occur for womenwho ultimately delivered vaginally within 24 hours ofinduction (although future studies will explore differentcutoff points). A multivariate expression, namely EMGILIF,was compared for GI vs. G2 between pre-induction vs. post-induction measurements. A composite score, namely theaverage EMGILIF activity of each patient including bothbefore and after dosing, was then made using the mean of thepre- and post-induction values, and was compared betweenG1 and G2. Initially, one-way analysis of variance(ANOVA) was used, and post-hoc pair-wise comparisonswere then made. T-test was used to compare the compositescores. P<.05 was considered significant.
For the multivariate expression EMG/LIF, the pre-induction measurement for G1 was significantly higher thaneither the pre-or post-induction measurements for G2(0.48+0.20 vs. 0.31+0.12 and 0.24+0.08, respectively; Fig.16A). The composite score was significantly higher for GIcompared to G2 (0.45+0.16 vs. 0.28+0.10; Fig. 16B).
In this pilot study, we only included patients whodelivered spontaneously, and therefore c-section patientswere excluded. As a result, we had to use a somewhatartificial criteria for determining successful vs. failedinduction (i.e. delivery or not within 24 hours). In futurestudies, we will also consider c-section patients and cervicaldilation in order to be more rigorous. For example, a womanwho progressed to 9cm dilatation and had a CIS for fetaldistress should not be classified as a failed induction.However, from this pilot study, we concluded that uterineEMG and cervical LIF, in the multivariate expressionEMGILIF, may be indicative of successful induction.Specifically, since the pre-pitocin EMGILIF value was highin those who ultimately had effective treatment, the pre-
treatment EMG/LIF measurement may determine whichpatients are most likely to benefit from pitocin induction andwhich are not. Similarly, in treated patients, the compositescore may indicate imminent success or failure of theinduction treatment when calculated through four hours ofdosing, and could aid clinicians in deciding on whether ornot to continue the treatment.
CONCLUSIONS AND POTENTIAL BENEFITS OFEMG AND LIF
Uterine EMG and cervical LIF monitoring promise to beinvaluable for many clinical situations during pregnancy indeveloped, as well as in developing, nations. In third-worldcountries, for example, one asset of accurate diagnosis ofpreterm labor would be the early transfer of women, whereaccess to adequate facilities is greatly limited relative to thedeveloped world. Medical intervention for pregnant womenand their unborn fetuses under these conditions is mostcritical, since prenatal nutrition and care has probablyalready been sub-par for them by the time pre-term labor isdiagnosed. An easy-to-use, portable system, which poten-tially provides an immediate, on-site diagnosis (such asuterine EMG and cervical LIF), could improve the preg-nancy outcomes for these women. For the modern world,adoption of these technologies into standard pregnancymonitoring would also impact the field greatly therein.
The following is a description of some of the possibleuses for combining EMG and LIF measurements, i.e. if madesimultaneously on the same patients (Table 3):
Considering the tremendous need for (and potentialbenefits from) uterine EMG and cervical LIF technologies,joint implementation of them could likely improve patientmanagement and classification, and could be expected tolead to a reduction in perinatal mortality and morbidity [48].
220 Current Women's Health Reviews, 2006, Vol. 2, Na. 3
Geuy?ehl et al.
GI G2G1 G2
B 1.0
0.8
InH
0.2 -
0.0
b
b
P < .05
Successful
FailedInduction
InductionPostinduction
Fig. (16A,13). Comparison of EtvlUlL[F scores for 12 patients, 8 with successful induction and 4 with failed induction, as obtained both prior
to pitocin treatment and after 4 hours of pitocin treatment: The pre-induction EMGILIF for the patients with successful inductions was
significantly higher than either the pre-induction or post-induction measurements for the patients who failed induction (ANOVA used, P <
.05). A composite EMGILIF score, using the average EMG/LIF value of the pre-induction and post-induction scores, indicated a
significantly higher composite score for the patients having successful inductions compared to those who failed induction (t-test used, P <
.05).
Table 3. Potential Uses of Simultaneous Uterine EMG and Cervical LIP Measurements
Condition Clinical Use Benefits
Antepcx7rrro Predict preterm delivery Prevent preterm delivery and improve perinatal outcome
Early labor Differentiate between false and true labor Prevent unnecessary admissions
Areterm contractions Determine need for tocolysis Prevent unnecessary use of tocolytics and decrease cost
Cervical
Incompetence
Rule out the presence of subclinical uterine electricalactivity. Detect abnormalities in cervical collagen content.
Post-cerclage monitoring
Better define candidates for cerclage, Better understanding ofetiologies. More objective measure of cervical change
Induc•tionlabnr Predict success of induction Prevent unnecessary and prolonged inductions
Tocolysis Predict success of tocolysis. More objective and accurateanalysis of uterine activity. Improved method to monitor and
adjust toeolytic therapy
Prevent complications from tocolytics. Decrease time ontocolytics
[4] Kramer MS. Preventing preterm birth: are we making any progress?ACKNOWLEDGEMENTS Yale J Blot Med 1997; 70: 227-32.
[5] Steer P. The epidemiology of preterm labor--a global perspective. JStudies supported by NIH Grant Number ROl HD37480 Perinat Med 2005; 33(4): 273-6.
and the General Clinical Research Center (GCRC) at the [6j Csapo Al. Force of labor. In: Iffy L, Kamicntzky HA, editors
University of Texas Medical Branch, funded from Grant Principles and Practice of Obstetrics and Perinatology. New YorkWiley, 1981; 76 I -799,
M01
RR00073
by the National
Center for
Research[71 Garfield RE, Yaliampalli C. Control of myometrial contractility and
Resources, NIH, USPHS. labor. In: Chnal iisz K, Garfield RE. Basic Mechanisms Controlling
REFERENCES[8]
Tenn and Pre-term Birth. New York; Springer-Verlag, 1993; 1-28.Chwalisz K, Garfield RE. Regulation of the uterus and cervix
[I] Lewit EM, Baker LS, Corman H, Shiono PH. The direct cost of low during pregnancy and labor: role of progesterone and nitric oxide.
birth weight. Future Child 1995; 5(1): 35-56. Ann N Y Acad Sci 1997; 828: 238-253.
[2] Marlin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, [91 Garfield RE,
Yallampalli
C. Structure and
function of uterine
Munson ML. Births: Final data for 2003. Nat Vital Stat Rep 2005;54(2): 1-24.
muscle. In: Chard T, Grudzinskas JG, Eds. The Uterus. CambridgeReviews in Human Reproduction. Cambridge, UK: Cambridge
[3] Tucker JM, Goldenberg RL, Daivs RO, Copper RL, Winkler CL, University Press, 1994; 54-93, 40-81.
Hauth JC. Etiologies of prerterm birth in an indigent population: is [10] Tezuka N, Ali M, Chwalisa K, Garfield RE. Changes in transcripts
prevention a logical expectation? Obstet Gynecol 199 t; 77: 343-7. encoding calcium channel subunits of rat myometrium duringpregnancy. Am.) Physiol 1995; 269: C1008-C1017,
Preinduction
Uterine E.41G and Cervical LIF- Pramising Technologies in Obstetrics
Current Women 's Health Reviews, 2006, Vol. 2, No. 3 221
[Ill
Danforth DN. The morphology of the human cervix. Clin ObstetGynecol 1983; 26: 7-13.
[12]
Leppert PC. Anatomy and physiology of cervical ripening. ClinicalObstet Gynecol 1995; 38: 267-279,
[131 Leppert PC. The biochemistry and physiology of the uterine cervixduring gestation and parturition. Peewit Neonatal Mcd 1998; 3: 103-ID5.
[141 Unbar( J, Olson G, Goodrum L, Rowe T, Saade G, Hankins G. Pre-term labor at 32 to 34 weeks' gestation: effect of a policy uf'expectant management on length of gestation. Am J Obstet Gynecol1990; [78: S 179.
[ l51 Goldenberg RL, (liver SP, Bronstein J, Cutter GR, Andrews WW,Memncmeyer ST. Bed rest in pregnancy. Obstet Gynecol 1994; 84:131-136.
[16] Henncrsdorf MG, Niebch V, Perings C, Strauer BE. T wavealternans and ventricular arrhythmias in arterial hypertension.flypertension 2001; 37(2): 199-203.
[17] Gasperin CA, Germiniani H, Facin CR, Souza AM, Cunha CL. Ananalysis of electrocardiographic criteria for determining leftventricular hypertrophy. Arq Bras Cardiol 2002; 78(1): 59-82.
118] Perkiomaki JS, Zareba W, Greenberg HM, Moss AJ. ThrombogenieFactors and Recurrent Coronary Events Investigators. Usefulness ofstandard electrocardiographic parameters for predicting cardiacevents after acute myocardial infarction during modern treatmentera. Am J Cardiol 2002; 90(3): 205-9.
[191 Madias JE. Amplitude of the electrocardiographic QRS complexesduring and after severe pulmonary edema. Ann NoninvasiveElectrocardio12004; 9(2): 192-7.
[20] Bacharova L, Michalak K, Kyselovic J, Klimas J. Relation betweenQRS amplitude and left ventricular mass in the initial stage ofexercise-induced left ventricular hypertrophy in rats. Clin ExpFEypertens 2005; 27(6): 533-41.
[21] Nakamura T, Florio H, Chiba Y. Local holder exponent analysis ofheart rate variability in preterm infants. IEEE Trans Biomed Eng2006; 53(1): 83-8.
[22] Maul H, Olson G, Maner W, Wyckoff M, Saade GR, Garfield RE.The intensity of uterine contractions can be estimated usingtransabdominal uterine electromyography. AJOG 2001 (Suppl);185(6): S208.
[231
Undenfriend S. Fluorescence assay in biology and medicine. NewYork: Academic Press, 1962; Vol. 1. -
[24] Rainanujam N, Mitchell MF, Mahadevan A, Thomsen 5, Silva E,Richards-Kortum R. Fluorescence spectroscopy: a diagnostic toolfor cervical intraepithelial neoplasia (CIN). Gynecol Oncol 1994;52: 31-38.
[251 Maner WL, Garfield RE, Maul Olson G, Saade G. Predictingterm and pre-term delivery with Iransabdominal uterine electromyo-graphy. Obstet Gynecol 2003; 101: 1254-1260.
[26] Pajntar M, Verdenik I, Pusenjak S, Rude] D, Leskosek B. Activityof smooth muscles in human cervix and uterus. Eur 1 ObstetGynecol Reprod Biol 1998; 79(2): 199-204.
[27] Snowden S, Simpson NAB, Walker JJ. A digital system forrecording the electrical activity of the uterus. Physiol Meas 2001;22:673-679.
[28] Eswaran H, Wilson JD, Murphy P, Preissl H, Lowery CL.Applieafion of wavelet transform to uterine electromyographicsignals recorded using abdominal surface electrodes, J Matern FetalNeonatal Med 2002; 11(3): 158-66.
[291
Eswaran H, Peeks] H, Wilson JD, Murphy P, Lowery CL.Prediction of labor in term and preterm pregnancies using non-
invasive magnetomyographic recordings of uterine contr actions.Am 1 Obstet Gynecol 2004; 190(6): 1598-602; discussion 1602-3.
[30] Buhimschi C, Boyle MB, Saade GR, Garfield RE. Ulcrine activitydaring pregnancy and labor assessed by simultaneous recordingsfrom the myometrium and abdominal surface in the rat. Am 1Obstet Gynecol 1998; 17814): 811-22.
[31] Figueroa JP, Honnebier MB, Wentworth R, Nathanielsz PW. Powerspectrum analysis of myontetrial electromyogram and intrauterinepressure changes in the pregnant rhesus monkey in late gestation.Am J Obstet Gynecol 1989; 161(2): 467-73.
[32] McGlothlin JA, Lester GD, Hansen P1, e1 at. Alteration in uterinecontractility in mares with experimentally induced placenlitis.Reproduction 2004; 127(1): 57-66.
[33] Chellman GJ, Bee WH, Resendez JC, et at. Telemetric uterinecontraction model in preterm cynomolgus monkeys. ReprodToxicoi 2004; 18(2): 285-93.
[34] Maul H, Olson G, Fittkow CT, Saade GR, Garfield RE. Cervicallight-induced fluorescence in humans decreases throughoutgestation and before delivery: preliminary observations. Am JObstet Gynecol 2003; 188: 537-541.
[35] Fittkow CT, Maul H, Olson G, et at. Light-induced fluorescence ofthe human cervix decreases alter prostaglandin application forinduction of labor at term. Eur J Obstel Gynecol Reprod E3iol 2005;123(1): 62-6.
[36] Shi L, Shi SQ, Saade GR, Chwalisz K, Garfield RE. Changes incervical resistance and collagen fluorescence during gestation inrats. J Perinat Med 1999; 27(3): 188-94.
[37] Fittkow CT, Shi SQ, Bytautiene E, Olson G, Saade GR, GarfieldRE. Changes in light-induced fluorescence of cervical collagen inguinea pigs during gestation and after sodium nitroprussidetreatment. J Perinat Mcd 2001; 29(6): 535-43.
[38] lams JD. Prediction and early detection of pre-term labor. ObstetGynecol 2003; 101: 402-412.
[39] Romero R, Chaiworapongsa T, Espinoza J. Micronutrients andintrauterine infection, pre-term birth and the fetal inflammatoryresponse syndrome. J Nutr 2003; 133(5 Suppl. 2): 1668S-1673S.
[40] Maul H, Maner WL, Olson G, Saade GR, Garfield RE. Non-invasive transabdominal uterine electromyography correlates withthe strength of intrauterine pressure and is predictive of labor anddelivery. J Matern Fetal Neonatal Med 2004; 15(5): 297-301.
[41] Wapner RJ, Cotton DB, Arta! R, Librizzi RJ, Ross MG. Arandomized multicenter trial assessing a home uterine activitymonitoring device used in the absence of daily nursing contact. AmJ Obstet Gynecol 1995; 172: 1026-1034.
[42] Dyson DC, Dange KI-1, Bamber JA, et al. Monitoring women at riskfin pre-term labor. N Engl J Med 1998; 338: 15-19.
(43]
Colombo D, lams JD. Cervical length and pre-term labor. ClinObstet Gynecol 2000; 43: 735-745.
[44] Romero R. Routine obstetric ultrasound. Ultrasound ObstetGynecol 1993; 3: 303-307.
[45] lams JD, Casa! D, McGregor JA, et al. Fetal fibronectin improvesthe accuracy of diagnosis of pre-term labor. Am J Obstet Gynecol1995; 173: 141-145.
[46] Lockwood CJ. Predicting premature delivery-no easy task. N Engl JMed 2002; 346: 282-284.
[47] Hellemans P, Gerris J, Verdonk P. Fetal fibronectin detection forprediction of pre-term birth in low risk women. Br J ObstetGynaecol 1995; 102: 207-2 12.
[48] Garfield RE, Ctswalisz K, Shi L, Olson G, Saade OR.Instrumentation for the diagnosis of term and pre-term labor. JPerinat Med 1998; 26: 413-436.
Received: October 24, 2005
Revised: April 24, 2006
Accepted: May 04, 2006