Phosphorus 31 magnetic resonance spectroscopy of humanplacenta and quantitation with perchloric acid extracts

Helen H. Kay, MD:'c John D. Gordon, MD, Anthony A. Ribeiro, PhD: andLeonard D. Spicer, PhDb, c

Durham, North Carolina

Phosphorus 31 magnetic resonance spectroscopic studies of fresh placental tissue are reportedthat indicate resonances for adenosine triphosphate, inorganic phosphate, sugar

phosphates-phosphomonoesters, and phosphodiesters, Perchloric acid extract methods were used to

further characterize and quantitate phosphorous metabolites in term human placentas by phosphorus 31magnetic resonance spectroscopy. The perchloric acid extracts give enhanced resolution of phosphorus

signals and allow identification of other phosphorus metabolites including small amounts ofphosphocreatine, Emphasis was placed on quantitating adenosine triphosphate levels in the acid extractswith the use of the external reference standard hexachlorocyclotriphosphazene in a coaxial capillary

system. Adenosine triphosphate levels measured in this way ranged from 0.404 to 0.709 ILmol per gramwet weight. Comparison with an internal standard method with phosphocreatine is also reported.

Contribution to the measured high-energy phosphate pool from blood in the highly vascularized tissue wasfound to be relatively large and could range from 30% to 50% of the total adenosine triphosphatemeasured. (AM J OBSTETGVNECOL 1991 ;164:80-7.)

Key words: Phosphorus 31 nuclear magnetic resonance spectroscopy, human placenta,adenosine triphosphate

Nuclear magnetic resonance spectroscopy is a tech­nique that chemists have used to study chemical struc­ture, More recently it has been used to characterize thebehavior of molecules in biologic tissue, 1 Magnetic res­onance spectroscopy uses the resonant frequencies ofcertain naturally occurring, nonradioactive nuclei (hy­drogen 1, phosphorus 31, carbon 13) that resonate un­der the influence of an external magnetic field, Theexact resonance frequency of the nuclear spin of anatom (e.g" 31P) will depend on the magnetic environ­ment set up by local electrons, This property, termedthe chemical shift, allows the nucleus being studied tobe identified within a molecule in a chemical solutionor in a biologic tissue. The collection of resonance sig­nals at their separate chemical shifts constitutes themagnetic resonance spectrum. The units of chemicalfrequency shift in the spectra are parts per million(ppm) or hertz,

From the Departments of Obstetrics and Gynecology," Biochemistry,'and Radiology,' Duke University Medical Center.Supported by National Institutes of Health Grant No.1K08HD00822-01 to H.K. Nuclear magnetic resonance spectra wereobtained at the Duke University Nuclear Magnetic Resonance Spec­troscopy Center funded by the National Institutes of Health, the Na­tional Science Foundation, the North Carolina Biotechnology Center,and Duke University.Received for publication March 15, 1990; revised August 2, 1990;accepted August 6, 1990.Reprint requests: Helen H. Kay, MD, Box 3373, Duke UniversityMedical Center, Durham, NC 27710,6/1/24392

80

The ability of magnetic resonance spectroscopy tomonitor the atomic signals of molecules makes it animportant tool for metabolic studies in a variety of or­gan systems in both human beings and animals,2 Be­cause it is considered a noninvasive procedure magneticresonance spectroscopy may ultimately also become avery useful clinical tool. Magnetic resonance spectros­copy has recently been used in preliminary studies ofthe human placenta3.4 with promising results, suggest­ing that a better understanding of placental metabolismcan be obtained with this technique,

Placental metabolism and function have been gen­erally understudied even though the placenta plays animportant role in the maintenance of fetal well being,Its primary functions are to transport nutrients (e,g.,oxygen, amino acids, sugars, and lipids) to the fetusand to remove waste from the fetus. Another majorfunction is synthesis of large amounts of steroid hor­mones, primarily estrogen and progesterone, Estro­gens stimulate placental uptake and use of glucose aswell as formation of glycogen and glycerol. 5 The pla­centa also actively synthesizes large amounts of proteinssuch as human chorionic gOhadotropin and human pla­cental lactogen, compounds that maintain maternalproduction of steroid hormones and regulate use offetal glucose much like human growth hormone,6 Fi­nally, the placenta produces large amounts of lactatethat is used by the fetus for gluconeogenesis,7 Alter­ations in any of these and other placental functions may

Volume 164Number I, Part 1

lead to a compromised fetus such as a growth-retardedfetus.

As an initial step toward the study of placental me­tabolism by nuclear magnetic resonance spectroscopicmethods, we used perchloric acid extract methods tocharacterize the phosphorus metabolites in term hu­man placentas by 31p magnetic resonance spectroscopy.In the studies described in this article we focused onthe high-energy phosphate signals from the nucleosidetriphosphates, in particular adenosine triphosphate.Alterations in the adenosine triphosphate levels andother adenosine nucleotides may be associated withpathologic conditions such as preeclampsia.B

•9 In ad­

dition, high levels of placental adenosine triphosphatehave been associated with high glucose levels in vitro JO

and may playa role in diabetic pregnancies, which oftenresult in abnormally large infants. Lastly, ion transport,an important process in the placenta, is thought to beadenosine triphosphate-dependent because the addi­tion of ouabain, an inhibitor of adenine triphosphatase,has been shown to affect such transport." Therefore,abnormalities in placental metabolism may be shownby altered levels of adenosine triphosphate. Here wereport results ofour current efforts in the developmentof quantitation methods for adenosine triphosphateresonances in fresh and frozen human placental sam­ples so that ultimately comparisons can be made be­tween normal and abnormal placentas that may lendfurther insights into pregnancies that result in abnor­mal outcomes.

Material and methods

Fresh tissue preparation. Freshly delivered placentaltissue from both vaginal and cesarean section birthswere studied. Approximately 5 gm oftissue was sharplydissected free from the remainder of the placenta andplaced into a 20 mm nuclear magnetic resonance tubewithin I minute of delivery and maintained on ice forno more than I hour before spectra were acquired.Spectra were obtained with a 65° flip angle (40 Ilsec) ,an acquisition period of 1.12 seconds, a relaxation delayof I second (total pulse interval 2.12 seconds), and 256acquisitions. These conditions were used to qualitativelyassess 31p spectral features and correspond to partiallysaturated inorganic phosphate resonances but to fullyrelaxed adenosine triphosphate signals inasmuch as themeasured longitudinal (Tl) relaxation value that de­termines the degree of relaxation for adenosine tri­phosphate in the tissue at 300 MHz field strength is 0.5to 0.6 sec.

Tissue extracts. Perchloric acid extracts of placentaltissue were prepared with a modification of a previouslydescribed technique. '2 Placentas were harvested by re­moving small pieces of tissue from the maternal sideand freezing them in liquid nitrogen within 1 minute

31p spectroscopy of human placenta 81

of delivery. Tissue trauma was minimized by use ofsharp dissecting instruments, by removal of each piecewith minimal dissection, and by performance of thedissection with the placenta remaining in the dissectingpan. The frozen tissue was pulverized in a custom­designed stainless steel mortar and pestle chilled withliquid nitrogen. A 20 gm sample of the finely groundtissue was mixed with 20 ml of I mollL perchloric acidthat had been precooled to the temperature of dry ice.The perchloric acid-tissue mixture was allowed towarm to _10° C in the ambient air for 15 minutesduring which time the metabolites were extracted. Itwas then centrifuged at 40,000 rpm at - 4° C for 20minutes to separate the remaining solid fraction. Thesupernatant containing the extracted metabolites wasdecanted into a chilled beaker and neutralized to justabove pH 7.0 through the dropwise addition of potas­sium hydroxide. The potassium perchlorate precipi­tated during neutralization was removed by a secondcentrifugation step identical to the first, and the su­pernatant was then passed over a chelating resin (SigmaChemical Co., St. Louis) to remove divalent cations.The extract was again titrated to neutral pH with diluteperchloric acid, frozen in a vacuum flask (Virtis Co.,Gardiner, N.Y.), and lyophilized to dryness. The extractwas then resuspended in 5 ml of 40% deuterated water.For quantitation of spectra, the extracts were placed inthe outer chamber of a coaxial 10 mm nuclear magneticresonance tube with an external reference standard of25 mmollL hexachlorocyclotriphosphazene (AldrichChemical Inc., Milwaukee) 13 in the inner chamber.These spectra were obtained with a 70° pulse angle (21Ilsec) , an acquisition period of 1.12 seconds, a relaxa­tion delay of 15 seconds (total pulse interval 16.12 sec­onds) and 160 acquisitions. These conditions yield fullyrelaxed adenosine triphosphate signals from the ex­tract, inasmuch as the TI relaxation time was measuredto be approximately 1.8 seconds for the adenosine tri­phosphate signals of the extract.

Contribution from blood. To determine the contri­bution of blood to the tissue extract, we performedperchloric acid extracts of maternal blood (peripheral)and fetal blood (cord) from one patient. Perchloric acid(l moll L) was added in a 1: I ratio to 20 ml blood andextracted according to the above protocol.

Nuclear magnetic resonance instrumentation. Allspectra were obtained on a spectrometer (model GN­300WB General Electric, Fremont, Calif.) interfacedwith a wide-bore superconducting magnet (89 mmbore, 7.0 tesla field) in the Duke Magnetic ResonanceSpectroscopy Center. A standard 20 mm broad bandprobe (General Electric) tuned to 121.4 MHz was usedfor 31p spectroscopy. The 90° pulses were: 56 Ilsec fora 20 mm tube with placental tissue; 27 Ilsec for a

82 Kay at al. January 1991Am J Obstet Gynecol

Pi

SP/PME

I10

I-20 PPM

Fig. 1. 121 MHz "P magnetic resonance spectroscopy spectrum of gross placental tissue. Peaksidentified are sugar phosphate-phosphomonoester, SP/PME; inorganic phosphate. Pi; phospho­diester, PDE; a-, ~-. and -y-adenosine triphosphate. Spectrum obtained with 65° flip angle (40 jLsec),acquisition period of 1.12 seconds, relaxation delay of 1 second (total pulse interval, 2.12 seconds),and 256 acquisitions.

10 mm tube with hexachlorocyclotriphosphazene­extract; 25 ,""sec for a 10 mm tube with hexachlorocy­clotriphosphazene-adenosine triphosphate. Spectro­scopic data were collected at the ambient operatingprobe temperature (approximately 22° C) without tem­perature regulation. To avoid radio frequency heatingeffects on the tissue and extracts, no proton decouplingwas used. Sip chemical shifts are referenced to phos­phocreatine at 0.00 ppm.

External reference standard. Quantitation of aden­osine triphosphate levels in the perchloric acid extractswas performed with the external reference standard ofhexachlorocyclotriphosphazene and chromium acetyl­acetonate (Aldrich Chemical Inc.) in a coaxial capillarysystem. IS Hexachlorocyclotriphosphazene has provedto be a useful Sip chemical shift reference because itsSip signal (approximately 24.4 ppm at our conditions)exhibits only a small shift with temperature (0.34 ppmbetween 8° and 50° C) and the narrow hexachlorocy­clotriphosphazene signal is easily observed withoutproton decoupling. The Tl relaxation time of 25mmol/L hexachlorocyclotriphosphazene in deuteratedbenzene at 121 MHz was 15 seconds. Addition of 31mmol/L chromium acetylacetonate reduced the relax­ation time for the hexachlorocyclotriphosphazene res­onance to <200 msec. A calibration curve was con­structed by recording nuclear magnetic resonance spec­tra for the two-compartment coaxial 10 mm nuclearmagnetic resonance tube with 25 mmol/L hexachlo­rocyclotriphosphazene with 31 mmol/L chromium ace-

tylacetonate in the inner chamber of a 10 mm coaxialinsert and varying adenosine triphosphate (SigmaChemical Inc., St. Louis) concentrations in the outerchamber. Spectra were acquired under identical con­ditions to those for the perchloric acid extracts. Peakarea integration of the fully relaxed adenosine tri­phosphate and hexachlorocyclotriphosphazene signalsgenerated the calibration curves. The ordinate is pre­sented as the ratio of the area of the ~-adenosine tri­phosphate resonance to that of hexachlorocyclotri­phosphazene, whereas the abscissa represents theadenosine triphosphate concentration. This calibrationcurve gives a quantitative estimate of the concentrationsof the various metabolites in the perchloric acid ex­tracts. It is also used to estimate concentrations frompeak areas in intact placental tissue although it mustbe recognized the signal attributed to ~-adenosine tri­phosphate in the typical spectrum of whole tissue mayinclude small contributions from other nucleoside tri­phosphates.14 All adenosine triphosphate concentra­tions reported in this work are expressed as micromolesper gram of wet weight placental tissue.

Intemal reference standard. To assess the reliabilityof quantitation with the external reference standard inour coaxial nuclear magnetic resonance tube, we com­pared results with those obtained by an internal stan­dard. Because our spectra show at most very small sig­nals from phosphocreatine in placental extracts, wechose to use it as the internal reference standard byspiking our perchloric acid extracts in a fixed volume

Volume 164Number 1, Part 1

HCCTP

I I

20

4

1213

31p spectroscopy of human placenta 83

15

I I I I I I-20 -30 PPM

Fig. 2. 121 MHz 31p magnetic resonance spectroscopy spectrum of perchloric acid extract of humanplacental tissue. Spectrum acquired with 70° flip angle (21 fJoSec), acquisition period of 1.12 seconds,relaxation delay of 15 seconds (total pulse interval, 16.12 seconds), and 160 acquisitions. Tentativepeak assignments on the basis of literature: ·1, sugar phosphates, 2, 2,3-diphosphoglycerate; 3,phosphoryicholine; 4, inorganic phosphate; 5, glycerophosphorylserine; 6, glycerophosphoryl­ethanolamine; 7, glycerophosphoryicholine; 8, phosphocreatine; 9, a-adenosine triphosphate; 10,13-adenosine diphosphate; 11, a-adenosine diphosphate; 12, a-adenosine triphosphate; 13, the ox­idized form of nicotinamide-adenine dinucleotide; 14, uridinediphosphate glucose; 15, 13-adenosinetriphosphate. External reference standard, hexachlorocyclotriphosphazene, is at approximately 24.4ppm with phosphocreatine at zero.

with a standard aliquot of 5 mmol phosphocreatine.Adenosine triphosphate concentrations were then cal­culated by comparison of the integrated peak areas withthat of the added phosphocreatine.

Statistics. A linear regression analysis and a Pearson'scorrelation coefficient determination was performed tocorrelate the adenosine triphosphate levels determinedwith the use of the external reference standard withthat found by the internal spiking method.

Results

A typical spectrum obtained from our preliminarystudies of fresh placental tissue is shown in Fig. 1. Thea-, 13-, and 'I-nucleoside triphosphates peaks are seenwith the largest contributions to these resonances com­ing from adenosine triphosphate.1 4 There is a largeinorganic phosphorus peak, and significant resonancescorresponding to phosphodiesters and to a sugarphosphate-phosphomonoester envelope. No identifi­able resonance from phosphocreatine is observed inthis spectrum, a result similar to that reported previ­

ously.'Further spectral resolution of components in pla­

cental tissue is revealed in the corresponding spec­trum of a perchloric acid placental extract in coaxialconfiguration with the external hexachlorocyclotri-

phosphazene reference standard as illustrated in Fig.2. Resonances from specific phosphomonoesters andphosphodiesters such as 2,3-diphosphoglycerate, phos­phorylcholine, serine phosphorylethanolamine, glyc­erophosphorylethanolamine, and glycerophosphoryl­choline are clearly distinguishable. Peak assignmentsare made on the basis of previously reported identifi­cation of these compounds. 14 Even with the significantlyhigher signal! noise ratio and better resolution availablein the extracts, only a small peak corresponding tophosphocreatine at 0.00 ppm is observed. Hexachlo­rocyclotriphosphazene appears as a single line at ap­proximately 24.4 ppm downfield from the phospho­creatine position (0.00 ppm).

The calibration curve relating concentrations to rel­ative nuclear magnetic resonance intensities deter­mined for quantitating these studies is shown in Fig. 3.These data were obtained with commercially availableadenosine triphosphate and hexachlorocyclotriphos­phazene in a coaxial configuration identical to that usedto study the placental extracts. Adenosine triphosphateconcentrations from 11 human placentas were deter­mined with both the external hexachlorocyclotriphos­phazene standard and the internal spiking with addedphosphocreatine. The measured values are listed inTable I and range from 0.404 to 0.709 !-Lmol per gram.

84 Kay et al. January 1991Am J Obstet Gynecol

5 -r----------------r-...,

4

Q.I-o 3o:I:--~ 2GI,g

o 10 20 30 40

[ATP] mM

50 60

Calibration curve inset

0.4.------------=--,

0.3

0.2

0.1" "

3

[ATPJmM

0.0 +"~,...~,...~__r~__r~__r~-io

Fig. 3. CalibTation curve for use of external reference standard with constant concentration ofhexachlorocyclotriphosphazene and varying concentrations of adenosine triphosphate. Inset is aden­osine triphosphate at lower concentrations.

The correlation coefficient of comparisons between thetwo methods is 0.755.

Fig. 4 shows a 31p spectrum from maternal blood.The spectrum from fetal cord blood is essentially iden­tical. The significant peaks are those from adenosinetriphosphate, 2,3-diphosphoglycerate, and inorganicphosphate. The adenosine triphosphate calculatedfrom the maternal peripheral blood is 0.256 fLmol permilliliter blood. Because each gram of placenta rep­resents roughly 1 ml in volume, the adenosine tri­phosphate contribution from blood in this placenta canbe estimated to be approximately 0.205 fLmol per gramtissue, assuming that approximately 80% of the pla­centa is composed of blood (maternal and fetal).15 Thisresult suggests that blood with this concentration ofadenosine triphosphate may actually account for asmuch as 30% to 50% of the total adenosine triphos­phate seen in the tissue extracts and presumably asimilar contribution would be observed in in vivo orex vivo spectra of whole placental tissue.

CommentAlthough magnetic resonance spectroscopy has been

applied to other tissues and tissue extracts, its use isonly beginning to be explored for studies of tissue in

the female reproductive tract. Earlier work by Noy­szewski et al.3 has focused on 31p spectroscopy of grossplacental tissue that showed resonances for sugar phos­phate or phosphomonoester, inorganic phosphate, andadenosine triphosphate but no visible phosphocreatine.Before our systematic efforts at more detailed char­acterization and quantitation of metabolites, we alsoexamined fresh tissue samples and obtained similar re­sults as illustrated in Fig. 1.

The overall purpose of this work was to explore theapplicability of the perchloric acid extraction techniquefor quantitative investigations of metabolites in placen­tal tissue and to establish its correlation with gross pla­cental tissue. As can be seen in Fig. 2, the spectrumfrom an extract gives significant peak enhancement andresolution and indeed allows observation of very smallquantities of metabolic substrates such as phospho­creatine that are not easily observed in gross tissue.These extract spectra show that those from placentasare different from those seen in the heart,16 which hasrelatively high concentrations of adenosine triphos­phate and phosphocreatine, but are similar to thoseseen in liver, which has low concentrations of phos­phocreatine. l

? This low phosphocreatine concentrationfound in placenta appears to be an intrinsic property

Volume 164r\umber I, Part 1

HCCTP

2,3-0PG

Pi

3'p spectroscopy of human placenta 85

-30 PPM

Fig. 4. 121 MHz" P spectrum from perchloric acid extract of maternal blood. Fetal blood showedidentical spectrum. External reference standard: hexachlorocyclolriphosphazene. HCCTP; 2.3­diphosphoglycerale, 2,3 DPG; inorganic phosphate. Pi.

Table I. Adenosine triphosphate quantitation in placental extracts

Sample

I23456789

10II

*Mean ± SD,O.536 ± 0.111.t Mean ± SD, 0.659 ± 0.151.

Hexachlorocyclotriphosphaunestandard

(IJ-moll gm wet weight)*

0.4040.4230.4330.4450.4630.5210.5760.6080.6250.6900.709

Phosphocreatinestandard

(IJ-mollgm wet weight)t

0.5540.4960.5150.7470.4860.6770.6830.6750.6390.9950.791

of this tissue. suggesting that other energy metabolismpathways aside from those driven by creatine kinasemay also be significant.

In the performance ofour spectral analysis, we notedthat there were large amounts of inorganic phosphatein placental tissue. Quantitation with the same methoddescribed for quantitation of adenosine triphosphateyielded levels that ranged from 2.045 to 15.575IJ.mol per gram wet weight. Corrections were made forthe inorganic phosphate resonances being only 97.4%relaxed under our spectral conditions. The levels cal­culated are significantly higher than the adenosine tri­phosphate levels (range. 0.404 to 0.709 IJ.mol pergram). This is especially interesting because there arelow levels of phosphocreatine and adenosine triphos-

phate. This large amount of inorganic phosphate mostlikely did not originate from phosphocreatine or aden­osine triphosphate breakdown. It also does not appearthat this is an indirect contribution from blood in thetissue inasmuch as our calculated concentration of in­organic phosphate is 0.647 IJ.mol per millileter of blood.or only 0.518 IJ.mol per gram tissue. This suggests thatinorganic phosphate in placenta may playa major rolein other phosphorus metabolic processes aside fromoxidative phosphorylation. lies et al. 18 suggested thatthe inorganic phosphate seen in rat liver extracts mayderive from labile organic phosphates that liberate in­organic phosphate on extraction or that the inorganicphosphate may exist in an available pool for metabolicprocesses. Inorganic phosphate in placenta probably

86 Kay at al.

behaves in a similar fashion. Because glycolysis is animportant process in the placenta, a large amount ofinorganic phosphate may be needed to stimulate en­zymes such as glyceraldehyde-3-phosphate dehydro­genase and 3-phosphoglycerate kinase, which are im­portant enzymes in glycolysis. Perhaps future studiesof inorganic phosphate levels in normal and abnormalplacentas will prove to give informative biomedical in­sights.

The prominent peaks in the phosphomonoester andphosphodiester regions of the spectra should also bepointed out and may signify enhanced membrane syn­thesis that has previously been noted in spectra of tu­mors. 19 In many ways the placenta has been likened toa large, rapidly growing tumor.

One other aim iiI this study was to perform quanti­tation of phosphorus metabolites in our tissue extractswith a reference standard. Use of either external orinternal intensity standards in our extracts allows quan­titation of adenosine triphosphate levels and thusmeaningful comparative analysis between different pla­centas. We are interested in the development of thesemethods because such analysis may allow metabolicstudies of normal and abnormal placentas, which canbe used to evaluate abnormal conditions such as thoseassociated with growth-retarded infants. The levels ofadenosine triphosphate identified in normal and ab­normal placentas may ultimately be applied to in vivoplacental studies. Clearly the absence of quantitationhas presented a problem in many previous comparativenuclear magnetic resonance studies. The adenosinetriphosphate concentrations measured here forhuman placental tissue fall in the range of 0.40 to 0.70~mol per gram wet weight on the basis of a hexach­lorocyclotriphosphazene standard. Whereas this is lowcompared with some other types of tissues, it agreeswith published values from other placental studies thatwere in the range of 0.30 to 0.70 ~mol per gram wetweigh tOO as determined by spectrophotometric analy­sis and high-performance liquid chromatography onfreeze-clamped tissue.

The external coaxial capillary method with hexa­chlorocyclotriphosphazene has the advantage that it isportable and can be used reproducibly both in extractsand in perfused tissue samples. Use of an external stan­dard rather than an internal standard would minimizevariability in extraction rates, a difficulty that can beencountered in perchloric acid extractions. Reproduc­ibility of the current extraction and quantitation tech­nique was tested with repetitive measurements on sev­eral tissue samples from the same placenta with essen­tially identical results.

The contribution of blood to the total adenosine tri­phosphate measured is an important consideration in

January 1991Am J Obstet Gynecol

studies of highly vascular organs such as kidney, liver,and vascular tumors that may not be commonly ap­preciated. Our calculations indicate that the blood com­ponent may account for as much as 30% to 50% of thetotal adenosine triphosphate in our samples. There­fore, it appears important to make corrections for theadenosine triphosphate component from blood whenmeasuring the total adenosine triphosphate levels invascular tissues. This by necessity will depend on theapproximate blood volume present and an estimate ofblood adenosine triphosphate levels for a representa­tive blood sample. A correction of this type will be im­portant not only in perchloric acid extracts but also inin vivo spectroscopy studies.

REFERENCES

I. Kay HH, Mattison DR. Nuclear magnetic resonance spec­troscopy and imaging in perinatal medicine. In: Nathan­ielsz PW, ed. Animal models in fetal medicine (V). Ithaca,New York: Perinatology Press, 1986.

2. Radda G. The use of NMR spectroscopy for the under­standing of disease. Science 1986;233:640-5.

3. Noyszewski EA, RamanJ, Trupin SR, McFarlin BL, Daw­son MJ. Phosphorus 31 nuclear magnetic resonance ex­amination of female reproductive tissues. AM J OBSTETGYNECOL 1989;161:282-8.

4. Mattison DR, Kay HH, Miller RK, Angtuaco T. Magneticresonance imaging: a noninvasive tool for fetal and pla­cental physiology. BioI Reprod 1988;38:39-49.

5. Sakurai T, Takagi H, Hosoya N. Metabolic pathways ofglucose in human placenta. AM J OBSTET GYNECOL1969; 105: 1044-54.

6. Villee CA. Enzymes, receptors, metabolism, and placentalfunction. Trophoblast Research 1983;1:175-84.

7. Burd L1,Jones MD, Simmons MA, Makowski EL, MeschiaG, Battaglia FC. Placental production and foetal utilisationof lactate and pyruvate. Nature 1975;254:710-1.

8. Hoover CR, Bartholomew RA. Acid-soluble nucleotidesin placentae from normal and toxic patients. Obstet Gy­necoI1959;14:309-2l.

9. Bloxam DL, Bullen BE, Walters BNJ, Lao IT. Placentalglycolysis and energy metabolism in preeclampsia. AM JOBSTET GYNECOL 1987;157:97-10l.

10. Krantz K. Glycolysis and adenosine triphosphate produc­tion by the perfused human placenta. In: Lund CJ.Choate JM, eds. Transcript of the Fifth Rochester Tro­phoblast Conference. Rochester, New York: University ofRochester School of Medicine and Dentistry, 1969: 176.

11. Miller RK, Berndt WOo Mechanisms of transport acrossthe placenta: An in vitro approach. Life Sci 1975;16:7­30.

12. Lowry 0, Passonneau J. A flexible system of enzymaticanalysis. New York: Academic Press, 1972:146.

13. Gard J, Ackerman J. A 31p NMR external reference forintact biological systems.J Mag Res 1983;51:124-7.

14. Kushmerick M, Dillon PF, Meyer RA, Brown TR, Kri­sanda JM, Sweeney JM. 3'p NMR spectroscopy, chemicalanalysis, and free Mg2 + of rabbit bladder and uterinesmooth muscle. J Bioi Chern 1986;261: 14420-9.

15. Bloxam DL. Human placental energy metabolism: its rel­evance to in vitro perfusion. Contrib Gynecol Obstet1985;13:59-69.

16. Stein PD, Goldstein S, Sabbah HN, et al. In vitro evalu­ation of intracellular pH and high-energy phosphate me­tabolites during regional myocardial ischemia in cats using

Volume 164Number I, Part 1

3lp nuclear magnetic resonance. Magn Reson Med 1986;3:262-9.

17. Malloy CR, Cunningham CC, Radda GK. The metabolicstate of the rat liver in vivo measured by 3lp-NMR spec­troscopy. Biochim Biophys Acta 1986;885: 1-11.

18. lies RA, Stevens AN, Griffiths JR, Morris PG. Phosphor­ylation status of liver by 3lP-n.m.r. spectroscopy, and itsimplications for metabolic control. Biochem J 1985;229: 141-51.

JlP spectroscopy of human placenta

19. Sostman HD, Armitage 1M, Fischer J). NMR in cancer.I. High resolutio'n spectroscopy of tumors. Magn ResonImaging 1984;2:265-78.

20. Bloxam DL, 'Bobinski PM. Energy metabolism and gly­colysis in the human placenta during ischaemia and innormal labour. Placenta 1984;5:381-94.

Argon laser versus microscissors for hysteroscopic incision ofuterine septa

Giovanni Battista Candiani, MD, Paolo Vercellini, MD, Luigi Fedele, MD,Salvatore Garsia, MD, Diana Brioschi, MD, and Laura Villa, MD

Milan, Italy

We performed hysteroscopic metroplasty in 21 women with repeated abortion and subseptate uterus. Thepatients were randomly allocated to septal incision with the argon laser (group I, 10 subjects) or

microscissors (group II, 11 subjects) to compare these instruments in terms of surgical feasibility andanatomic results. The mean operating time was 57% longer in group I than in group II (p = 0.001), the

intra- and postoperative morbidity of the whole series was negligible, and the anatomic results atabdominal ultrasonography and hysteroscopy performed 2 months postoperatively were similar in the two

groups. This study confirms that microscissors are the simplest, fastest, most effective, and leastexpensive instrument to correct a septate uterus. The complete agreement of the findings at follow-uphysteroscopy and Ultrasonography suggest the use of the latter as the method of choice to check the

surgical results. (AM J OasTET GVNECOl1991;164:87-90.)

Key words: Uterine malformations, recurrent abortion, hysteroscopy, laser

Hysteroscopic incision seems to be the method ofc:hoice to correct a septate uterus. )·3 The endoscopic useof microscissors and the resectoscope has proved suc­cessful, t·, and hysteroscopic metroplasty with fiberopticlaser has been proposed: We designed a randomizedstudy to compare the surgical and anatomic results ofsection of the uterine septum performed with the argonlaser versus microscissors.

Material and methods

During 1989 we studied 21 women with a mean ageof 28 years (range, 23 to 34) with two or more spon­taneous abortions, a double uterine cavity at hystero­salpingography (Fig. 1) and evidence of a normal uter­ine fundus at ultrasonographic examination with a half-

From the First Department of Obstetrics and Gynecology, Universityof Milano School of Medicine.Received for publication April 23, 1990; revised July 30, 1990; ac­cepted August 3, 1990.Reprint requests: Paolo Vercellini, MD, First Department of Obstet­rics and Gynecology "L. Mangiagalli," University of Milano Schoolof Medicine, Via Commenda 12, Milano, Italy 20122.611124440

full bladder.6• 7 All the uteri were classified as American

Fertility Society class Vb (partial septate uterus).8After giving informed consent, the patients were al­

located according to a randomization list to hystero­scopic metroplasty with argon laser (group I, 10 wo­men) or microscissors (group II, 11 women) betweenday 7 and day 10 of the cycle. Preoperative hystero­salpingography and ultrasonography showed the sep­tum was broad (>3 em wide at the base) in two subjectsin group I and in four subjects in group II. Undergeneral endotracheal anesthesia and laparoscopic con­trol, the cervix was dilated to 7 mm and a rigid hys­teroscope (model 26157 B, Storz Endoscopy, Tuttlin­gen, Federal Republic of Germany) with a 7 mm di­ameter operating sheath (Storz model 26163 C and26163 H) was introduced. The uterine cavity was dis­tended with a 10% solution of dextran of molecularweight 40,000 in normal saline solution (Rheomacro­dex, Baxter-Travenol, Trieste, Italy). In the group Iwomen the septal incision was performed with the ar­gon laser (model 20, HGM Medical Laser Systems Inc.,Salt Lake City, Utah), passing the 0.6 mm flexible quartzfiber through the operating channel of the hystero-

87