CLINICAL ARTICLE

Vaginal delivery and serum markers ofischemia/reperfusion injury

E. Conner a,*, R. Margulies b, Mengling Liu c, S.W. Smilen d,R.F. Porges d, C. Kwon d

a Department of Obstetrics and Gynecology, Jersey Shore University Medical Center,Neptune, New Jersey, USAb Division of Gynecology, University of Michigan Health System, Ann Arbor, USAc Division of Biostatistics, New York University Medical Center New York, USAd Department of Obstetrics and Gynecology, Division of Urogynecology,New York University Medical Center New York, NY, USA

Received 13 February 2006; received in revised form 11 April 2006; accepted 19 April 2006

Abstract

Objective: Vaginal deliveries have been associated with pelvic organ prolapse andincontinence. The objective was to show whether markers of ischemia/reperfusioninjury are dependent upon the mode of delivery and length of labor. Method:Complete venipuncture sets were obtained on 62 subjects. All samples collectedwere analyzed for serum creatine phosphokinase (CPK) and lactate dehydrogenase(LDH). Lipid peroxidation was analyzed, using thiobarbituric acid reactive substances(TBARS), on a subset of 37 patients. Results: There was a significant increase in CPKfrom admission to 1 h postpartum and postpartum day 1 in vaginal delivery versuscesarean delivery. Longer second stages were associated with significant increases inCPK. There were no significant changes in either LDH or TBARS from admission to anyother time point regardless of mode of delivery. Conclusion: Vaginal delivery andlonger second stages were associated with a much greater increase in one of theseinjury markers.D 2006 International Federation of Gynecology and Obstetrics. Published by ElsevierIreland Ltd. All rights reserved.

0020-7292/$ - see front matter D 2006 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd.All rights reserved.doi:10.1016/j.ijgo.2006.04.039

* Corresponding author. Tel.: +1 732 776 4389; fax: +1 732 776 4525.E-mail address: connerbennett@msn.com (E. Conner).

KEYWORDSReperfusion injury;Pelvic floor;Creatinephosphokinase;Lipid peroxidation

International Journal of Gynecology and Obstetrics (2006) 94, 96—102

www.elsevier.com/locate/ijgo

1. Introduction

The levator ani muscle group is the largestcomponent of the pelvic floor. This collection ofmuscles is composed of the iliococcygeus, pubo-coccygeus, and puborectalis muscles. The levatorani has two important functions. The first is tomaintain a constant basal tone that keeps theurogenital hiatus closed and prevents descent orprolapse of the pelvic viscera. The second is tocontract reflexively in response to an increase inabdominal pressure (cough, laugh, sneeze) andassist in maintaining continence during suchevents [1]. Childbirth exposes the pelvic floor todirect compression by the fetal head and topressure generated by maternal expulsive efforts.There is a strong association between childbirthand the development of both incontinence andpelvic organ prolapse. Indeed, MRI studies haveclearly shown damage to the levator ani musclesin women who have undergone vaginal deliveries,as opposed to nulliparous women, and the major-ity of women with levator ani abnormalities inthese studies had stress incontinence [2]. Howev-er, the underlying mechanism leading to theseinjuries is unclear. One theory has been thatischemia/reperfusion (I/R) injury is responsiblefor damage to the muscle.

Ischemia activates the complement system andincreases the expression of adhesin molecules onendothelial surfaces, which result in the recruit-ment of neutrophils into the muscle. During thephase of reperfusion, oxygen becomes available. Asa result, the numerous neutrophils found in themuscle produce oxygen radicals, and membranelipid peroxidation ensues, leading to considerablestructural and functional damage [3].

Because I/R reactions are difficult to measure,in vivo serum markers can be measured instead[4]. Creatine phosphokinase (CPK) and lactatedehydrogenase (LDH) are nonspecific markers oftissue and muscle injury, whereas thiobarbituricacid reactive substances (TBARS) and malondial-dehyde (MDA) are more specific markers of lipidperoxidation. Importantly, the elevation of thesemarkers has been correlated with an increase inboth structural (e.g., mitochondrial damage, in-terstitial edema, and dissolution of myofibrils)[3,5] and functional (reduced resting membranepotential and reduced tetanic tension) [6] damageto skeletal muscle.

Because of the direct compression of the pelvicfloor by the fetal head, childbirth may provide anideal model for I/R injury, with the ischemic periodrepresented by the second stage of labor. The aim ofthis study is to determine if the mode of delivery

and length of the second stage affect markers ofskeletal muscle damage and specific markers oflipid peroxidation.

2. Materials and methods

This prospective study was approved by the NYUHospitals School of Medicine Investigational Board ofResearch Associates (IRBA #11282). Nulliparouswomen at term (z37 weeks gestation) who choseto deliver at NYU Tisch Hospital during the period ofstudy (November 2003—December 2004) wereapproached upon their presentation to labor anddelivery for participation in the study. Writteninformed consent was obtained from each subject.Patients who were non-English speaking or hadmultiple gestations, myopathies, or recent historyof aspirin or nonsteroidal anti-inflammatory usewere excluded. Demographic information, includingmother’s age, height, weight, race/ethnicity, pastmedical history, and medication usage, wasobtained. In addition, information on indication foradmission, cervical dilation/station atmultiple timepoints (admission, full dilation, and/or time ofdecision to undergo cesarean delivery), and fetalweight was recorded. The course of the patients’labor and delivery was not altered by their partic-ipation in the study. The intent was to evaluate 5groups among the patients enrolled in the study: (1)spontaneous vaginal delivery, (2) operative vaginaldelivery, (3) elective/scheduled cesarean delivery,(4) cesarean delivery secondary to arrest of dilation,and (5) cesarean delivery secondary to arrest ofdescent. Women who had emergent cesarean sec-tions for fetal indications were also excluded fromstudy as their natural labor curve was truncated andtime did not allow for the venipunctures as perprotocol.

The protocol specified that participants havevenipuncture upon admission to the hospital, 1 hpostpartum, and 1 day postpartum. In addition,those patients attempting vaginal delivery had anadditional blood draw upon reaching full cervicaldilation or at the time of the decision to perform acesarean section (whichever time point wasreached first). The samples were coded withoutany patient-identifying markers and analyzed forcreatine phosphokinase (CPK), lactate dehydroge-nase (LDH), and lipid peroxidation.

Serum CPK and LDH were quantitatively mea-sured in units/liter (u/l) on a Vitros 950 automatedanalyzer (Ortho-Clinical Diagnostics, Inc; Rochester,New York) shortly after the receipt of the specimen.The samples that were to be analyzed for lipidperoxidation were centrifuged (3000 g for 7 min at

Vaginal delivery and serum markers of ischemia/reperfusion injury 97

20 8C) and stored at!80 8C until assayed. Thefollowing thiobarbituric acid reactive substances(TBARS) protocol was used: 100 Al of plasma wasmixed with 500 Al of 0.4% thiobarbituric acid in 10%acetic acid, pH 5.0, and heated to 90 8C for 1 h. Afterbeing cooled under cold tapwater, an equal volume(600 Al) of butanol was added and the mixture wasvigorously shaken. After centrifugation at 7500 rpmfor 10 min, the fluorescence intensities of TBARS inthe butanol phase were measured in arbitrary units(a.u.), using a Gemini fluorescence—chemilumines-cence microplate reader (Molecular Devices, Sunny-vale, California) at an excitation wavelength of 515nm (bandwidth 4.5 nm) and an emission wavelengthof 553 nm (bandwidth 9.0 nm) [7].

Subject demographics were compared amongmodes of delivery, using the Fisher exact test whenthe variables were categorical (e.g., race) and theANOVA when the variables were continuous (e.g.,age, EGA, body mass index, and fetal weight).Within each mode of delivery, differences betweentime points were compared by using the Student-ttests. The ANOVA test was used in comparing theexact current value or change from admission in amarker at specific time points across the variousmodes of delivery. The Student-t test was used tocompare the values of serum markers for those withsecond stage V 1 h versus N 1 h, and linear regressionwas performed by using the length of the secondstage as a continuous variable. All statistical anal-yses were performed with S-PLUS 6.2 for Windows(Insightful Corporation, Seattle, WA).

3. Results

One hundred and eleven subjects were enrolled.One subject was excluded because she was enrolledat a gestation age of 36 2/7 weeks and thereforedid not meet inclusion criteria. Two subjects

withdrew from participation after initially consent-ing. Four subjects had emergent cesarean deliver-ies and were therefore excluded from the study. Ofthe remaining 104 subjects, 42 patients hadincomplete venipuncture sets, with the 24 h post-partum blood draw being the sample most oftenmissed. Complete venipuncture sets were obtainedon 62 subjects. This resulted in analysis of 32spontaneous vaginal deliveries, 9 operative vaginaldeliveries (6 forceps and 3 vacuums), 10 electivecesareans, 11 cesareans for arrest of dilation, and 0cesareans for arrest of descent.

All samples collected were analyzed for serumCPK and LDH. TBARS was initially analyzed on asubset of 37 patients (24 vaginal and 13 cesareandeliveries). Interim analysis revealed that changesin levels of TBARS from admission to any other timepoint were not significant regardless of mode ofdelivery. It appeared that the levels of lipidperoxidation in our samples were below thedetection limit for this assay, and therefore theremaining samples were not analyzed.

The demographic characteristics of the group of62 patients with complete venipuncture sets arepresented in Table 1. Age, body mass index, andrace were not significantly different among the 4modes of deliveries. There was, however, a signif-icant difference in EGA and fetal weight betweenthe delivery modes. Regression analysis revealedthat EGA was significantly higher in both operativedeliveries (p b .02) and cesareans for arrest ofdilation (p b .01) than in elective cesareans andthat the weights of babies born after electivesection (p b .03) or operative delivery (p b .01)were less than those born after cesarean for arrestof dilation.

Changes in levels of LDH from admission to anyother time point were not significant regardless ofmode of delivery. However, there was a significantincrease (pV .01) in CPK from admission to 1 h

Table 1 Demographic information by mode of delivery

Spontaneous vaginaldeliveries

Electivecesareans

Arrest of dilationcesareans

Operative vaginaldeliveries

Number 32 10 11 9Age (years) 31.6 32.6 33.9 29.8EGA (weeks) 39 4/7 38 6/7*y 40 2/7y 40 1/7*BMI (kg/m2) 29.1 27.1 29 29.6Fetal weight (g) 3458 3314z 3773z§ 3181§

Race (%)Caucasian 22 (69%) 8 (80%) 8 (73%) 8 (89%)Asian 4 (12.5%) 1 (10%) 1 (9%) 0 (0%)Hispanic 4 (12.5%) 1 (10%) 1 (9%) 1 (11%)African—American 1 (3%) 0 (0%) 1 (9%) 0 (0%)Other 1 (3%) 0 (0%) 0 (0%) 0 (0%)

[*p b0.02; yp b0.01; zp b0.03; §p b0.01].

E. Conner et al.98

postpartum and from admission to postpartum day1 (pV .001) for all modes of delivery (Fig. 1). Inaddition, from admission to full dilation (or thetime that the cesarean was called), there was nosignificant change in CPK levels among the modesof delivery.

Using the admission time as the baseline, Fig. 2illustrates the change in CPK for each mode ofdelivery. From admission to the 1 h postpartumtime point, there was a significant increase in CPKin the spontaneous vaginal group when comparedto either the elective cesarean or the cesarean forarrest of dilation group (233 u/l vs. 83 u/l and 97 u/l, respectively; pV .02). There was also a statisti-cally significant increase in CPK in the operativevaginal group when compared to either the electivecesarean or cesarean for arrest of dilation groups(348 u/l vs. 83 u/l and 97 u/l, respectively;p V .001). The significant changes in CPK fromadmission to postpartum day 1 between the variousmodes of delivery mirrored those from admission to

1 h postpartum, with a significant increase in CPK inboth the spontaneous vaginal group and theoperative vaginal group when compared to eitherthe elective cesarean or the cesarean for arrest ofdilation groups (372 u/l and 384 u/l vs. 115 u/l and163 u/l, respectively; p b .04). However, there wasno significant difference in CPK between the twovaginal modes or between the two cesarean modesof delivery when either 1 h postpartum or postpar-tum day 1 was compared to baseline.

Because it was theorized that compression of thepelvic floor by the fetal head represented ischemicinjury time, the length of the second stage wasanalyzed, separating the vaginal delivery groupinto those with a second stage N60 min (n =18) andV60 min (n =14). The change in CPK from baselineto 1 h postpartum was significantly greater in thegroup with the longer second stage (p b .02). Asdemonstrated in Fig. 3, regression analysis esti-mated that for every increase of 1 min in thesecond stage of labor, CPK increased by 2.15 u/l in

-50

0

50

100

150

200

250

300

350

400

450

Fully dilated or c/s called Post-partum day 1

Time point

Cha

nge

in C

PK

(uni

ts/li

ter)

1 hour post-partum

†

†

‡§

‡§

¶‹‹

‹‹

*

*

II

¶II

!

!

Elective cesareanArrest dilation cesareanOperative NSVD

Figure 2 Changes in CPK from admission (baseline) to each time point. [*p b .02; yp b .02; zp b .001; §p b .001;||p b .004; bp b .01; lp b .02; Tp b .04].

0

50

100

150

200

250

300

350

400

450

500

Electivecesarean

Arrestdilation

cesarean

Operative Vaginal

Mode of delivery

CP

K (u

nits

/lite

r)

AdmissionFully dilated or cesarean1 hour PP

PPD#1†

†

‡

‡

§

§

*

* ¶II

¶

II

‹‹!

‹‹

!

Figure 1 CPK values at each time point for each mode of delivery. [*p b .01; yp b .001; zp b .01; §p b .001; ||p b .01;bp b .001; lp b .01; Tp b .001].

Vaginal delivery and serum markers of ischemia/reperfusion injury 99

the reperfusion phase (p =.001). This differencewas even more pronounced by postpartum day 1.For every increase of 1 min in the second stage,CPK increased by 2.35 u/l (Fig. 4) by postpartumday 1 (p b .03). These changes were not significantin the operative vaginal delivery group (n =9).There was no significant difference in this changewhen comparing those patients who had a sponta-neous vaginal delivery with a second stage greaterthan 1 h and patients who had an operative vaginaldelivery with a second stage less than 1 h.

4. Discussion

I/R injury in various skeletal muscles has beenpreviously studied both in animal and in humanmodels. Many of these studies have been done inthe context of introducing free radical scavengersor other antioxidants, such as vitamin E, in anattempt to mitigate the damage that occurs. Theaddition of these substances has been shown tosuppress the increase of nonspecific markers ofmuscle damage such as CPK and LDH, as well asspecific markers of lipid peroxidation (e.g., MDA),and ameliorate functional and structural damageto these muscles [3,4,8,9]. However, I/R injury hasnever been studied or documented in the pelvicfloor, as demonstrated by a search of Medline(1966—May 2005; search terms: breperfusioninjuryQ combined with bpelvic floor or pelvis orlevator aniQ).

Even in other skeletal muscles of the body, thetime course of I/R injury has not been completelyelucidated. However, the recent development of acytochrome c-based biosensor has shed some lighton this process by allowing the measurement ofsuperoxide radicals in vivo. This body of work hasshown that during the ischemic episode, superoxide

radicals are not generated. These radicals areproduced only after the ischemic event, taking onaverage 60—90 s for the blood flow to reach themuscle, reoxygenate the cells, and result in theirproduction. Moreover, both maximum superoxideconcentration and the total time of superoxideproduction are strongly dependent on the period ofischemia [10].

In this study, the changes in CPK showed markedincreases during the reperfusion phase (e.g., fromdelivery to 1 h postpartum). In addition, anincrease in the length of the ischemic phase (e.g.,an increase in the length of the second stage oflabor) caused a significant increase in the level ofCPK during the reperfusion phase. Although theexact timing of post-partum day 1 blood drawswere variable (always done the morning afterdelivery, making it between 6 and 30 h afterchildbirth), their continued increase as comparedto 1 h postpartum suggests that the reperfusioninjury changes to the pelvic floor were still ongoingat this time.

Because of the logistical difficulties with draw-ing blood samples at the time of delivery, it was notincluded as a time point. It was felt that knowingexactly when in the course of childbirth theischemia and the elevation in markers of I/R injuryoccurred was not as clinically important as acomparison between the various modes of delivery.In this study both spontaneous and operativevaginal delivery were associated with an increasein a marker of I/R injury compared to both electivecesarean and cesarean for arrest of dilation. Inaddition, the first stage of labor does not appear tocontribute significantly to markers of I/R injury, asthe change in CPK from baseline to full dilation orthe time at which a cesarean was called did notdiffer significantly between the modes of delivery.

Although the cesarean modes did not differ fromone another, perhaps the significant changes in CPK

0

200

400

600

800

1000

1200

1400

1600

1800

500 100 150 200 250Length of second stage (minutes)

Cha

nge

in C

PK

(uni

ts/li

ter)

Figure 4 Change in CPK from admission to 1 daypostpartum in spontaneous vaginal deliveries.

0

200

400

600

800

1000

1200

1400

1600

1800

500 100 150 200 250Length of second stage (minutes)

CP

K (u

nits

/lite

r)

Figure 3 Change in CPK from admission to 1 h postpar-tum in spontaneous vaginal deliveries.

E. Conner et al.100

from baseline to the postpartum time pointswithin each group represent tissue damage thatis inherent in undergoing surgery. Unfortunately,none of the women who had complete venipunc-ture sets had a cesarean for arrest of descent, asthis delivery group would represent an injurygroup combining both a complete ischemic phase(second stage of labor) and the muscular traumaof surgery.

Furthermore, although not a goal of our study, itis interesting to note that the level of CPK in thereperfusion period following operative deliverywith second stages V60 min was not significantlydifferent from that found following spontaneousdelivery with a second stage N60 min. This suggeststhat pelvic floor protection is not necessarilygained by using forceps to shorten the second stageof labor [11]. Perhaps more interesting is thatoperative vaginal delivery was not associated witha significant elevation of CPK compared to sponta-neous vaginal delivery, suggesting that although theformer may lead to clinical damage of the perine-um and vaginal epithelium, it does not necessarilyaugment the inherent pelvic floor damage thatoccurs with vaginal delivery, contrary to populartheory [1,12].

Some statistically significant differences werefound in the demographic variables betweendelivery groups. The elective cesarean group hada significantly lower EGA, probably because elec-tive sections are traditionally performed at 39weeks. In addition, as cephalopelvic disproportionis often the presumed etiology, it follows thatfetal weights might be greater in cesareans forarrest of dilation than in elective sections orvaginal deliveries.

In this study LDH was not a marker that reflectedI/R injury in the pelvic floor. Although in manystudies LDH results mirror those of CPK and TBARS,in a study by Koksal et al., LDH was the only markerthat failed to show a significant difference betweenthe control and treatment group [13]. This mayoccur because LDH is an enzyme present in thecytosol of all human cells, whereas CPK is foundmainly in skeletal and cardiac muscle. The mea-surement of LDH may not have been sensitiveenough to detect damage specific to the skeletalmuscles of the pelvic floor.

The assay used to measure TBARS also did notreflect I/R injury. Further investigation of this assayrevealed a significant flaw in its utility for thisstudy because the amount of lipid peroxidationdetected in the collected samples was consistentlybelow the level of detection for this particularassay. In addition, because this assay has beentypically performed on tissue cell culture (rather

than plasma, as in this study), the appropriateconcentrations of the reagents for plasma couldpotentially be different, and this also contributesto its poor reliability for this study. Future studyusing the more expensive methods needed todetect malondialdehyde (MDA), a stable interme-diate of lipid peroxidation that is widely accept-ed as a marker of reactive oxygen speciesmediated lipid peroxidation [3—6,8,14], couldprovide a better, more specific assay for I/Rinjury.

While the elevations seen in the non-specificmarker (CPK) were not paralleled by elevations inthe marker that is more specific for lipid perox-idation (TBARS) in this study, it should be notedthat previous studies have documented such arelationship. These studies made use of the moreexpensive methods of detecting lipid peroxidation(MDA) that were mentioned above. When suchmethods were used to investigate I/R injury inskeletal muscles other than the pelvic floor, thesignificant changes in plasma CPK did in factmirror those of MDA [5,8]. Moreover, the changesin both types of markers (specific and non-specific) correlated with evidence of histologicaldamage [5].

In summary, vaginal deliveries, both spontaneousand operative, resulted in higher levels of markersof muscle cell damage than did cesarean deliveries.The first stage of labor did not appear to signifi-cantly affect these markers while a longer secondstage of labor was associated with higher levels ofthese injury markers. This study provides prelimi-nary data to suggest that evidence of nonspecificmuscle cell damage correlates with I/R injury ofthe pelvic floor.

References

[1] Handa VL, Harris T, Ostergard DR. Protecting the pelvicfloor: obstetric management to prevent incontinence andpelvic organ prolapse. Obstet Gynecol 1996;88:470–8.

[2] Delancy JO, Kearney R, Chou Q, Speights S, Binno S. Theappearance of levator ani muscle abnormalities in magneticresonance images after vaginal delivery. Obstet Gynecol2003;101:46–53.

[3] Novelli GP, Adembri C, Gandini E, Orlandini SZ, Papucci L,Formigli L, et al. Vitamin E protects human skeletalmuscle from damage during surgical I/R. Am J Surg1997;173:206–9.

[4] Wijnen MHWA, Roumen RMH, Vader HL, Goris RJA. Amultioxidant supplementation reduces damage from is-chaemia reperfusion in patients after lower torso ischae-mia. Eur J Vasc Endovasc Surg 2002;23:486–90.

[5] Nanobashivili J, Neumayer C, Fugl A, Punz A, Blumer R,Prager M, et al. Ischemia/reperfusion injury of skeletalmuscle: plasma taurine as a measure of tissue damage.Surgery 2003;133:91–100.

Vaginal delivery and serum markers of ischemia/reperfusion injury 101

[6] Edrees WK, Young IS, Lau LL, Rowlands BJ, Refsum SE,Soong CV. Accentuated oxidative stress following reperfu-sion injury in diabetic rats. Int Angiol 2002;21:58–62.

[7] Huang X, Zhuang Z, Frenkel K, Klein CB, Costa M. The roleof nickel and nickel-mediated reactive oxygen species inthe mechanism of nickel carcinogenesis. Molec Mech ofMetal Tox Carc 1994;102:281–4.

[8] Bozkurt AK. a-tocopherol and iloprost attenuate reperfu-sion injury in skeletal muscle ishemia/reperfusion injury.J Cardiovasc Surg 2002;43:693–6.

[9] Hirose J, Yamaga M, Takagi K. Reduced reperfusion injury inmuscle: a comparison of the timing of EPC-K1 administra-tion in rats. Acta Orthop Scand 1999;70:207–11.

[10] Buttemeyer R, Andres WP, Mall JW, Ge B, Scheller FW,Lisdat F. In vivo measurement of oxygen-derived free

radicals during reperfusion injury. Microsurgery 2002;22:108–13.

[11] DeLee J. The prophylactic forceps operation. Am J ObstetGynecol 1920;1:34–44.

[12] Heit M, Mudd K, Culligan P. Prevention of childbirth injuriesto the pelvic floor. Curr Womens Health Rep 2001;1:72–80.

[13] Koksal C, Bozkurt AK, Cangel U, Unstundag N, Konukoglu D,Musellim B, et al. Attenuation of iscemia/reperfusion injuryby N-acetylcysteine in a rat hind limb model. J Surg Res2003;111:236–9.

[14] Grisotto PC, dos Santos AC, Coutinho-Netto J, Cherri J,Piccinato CE. Indicators of oxidative injury and alterationsof the cell membrane in the skeletal muscle of ratssubmitted to ischemia and reperfusion. J Surg Res2000;92:1–6.

E. Conner et al.102