91J Med Genet 1992; 29: 91-97

Fetal hydrops

P A Boyd, J W Keeling

AbstractSeventy-two fetuses or neonates withnon-immune hydrops were examinedbetween 1983 and 1988. The commonestassociation was chromosome abnormal-ity; 11 fetuses had a 45,X karyotype and 11

autosomal trisomy. Chromosome abnor-mality was suspected in a further 20 on

necropsy findings but chromosome cul-ture was not possible or unsuccessful. In11 cases there was histological evidence ofinfection; seven babies had major struc-tural anomalies and six affected fetuseswere twins. In six (8%) the cause of hyd-rops was not determined compared witheight (16%) of cases examined between1976 and 1982. Hydrops was diagnosedmore frequently while the fetus was alive,before 20 weeks' gestation, and associatedwith chromosome anomaly than foundpreviously.

Departments ofMedical Genetics,Churchill Hospital,Headington, OxfordOX3 7LJ.P A Boyd

Department ofPaediatric Pathology,John RadcliffeMaternity Hospital,Oxford.J W Keeling**Present address:Department of PaediatricPathology andCytogenetics, RoyalHospital for SickChildren, EdinburghEH9 lLF.

Correspondence toDr Boyd.Received 17 April 1991.Accepted 2 July 1991.

Effective prevention of rhesus disease hasshifted the emphasis of the investigation of thepregnancy complicated by fetal hydrops to-wards non-immunological causes.' The increas-ing availability of obstetric ultrasound and itsuse in the investigation of pregnancies whichare 'large for dates' or have a clinical diagnosisof polyhydramnios has resulted in a largerproportion of cases of fetal hydrops beingdiagnosed while the fetus is still alive. Themore frequent use of ultrasound for datingpurposes in the second trimester, togetherwith opportunities for a detailed fetal anomalyscan around 20 weeks' gestation, means thatmany hydropic fetuses are now recognised atan asymptomatic stage. This enables a widerrange of investigations to be undertaken. Elu-cidation of the underlying cause or associationof hydrops should be possible in a largerproportion of cases.

We report our experience with fetal hydropsduring the last five years and compare gesta-tional age at diagnosis, method of diagnosis,and pathological findings in the hydropic fetuswith those examined in the same unit between1976 and 1982.2

Material and methodsCases of fetal hydrops from pregnancies whichterminated in both the second and third tri-mesters were identified from the files of theDepartment of Paediatric Pathology, JohnRadcliffe Maternity Hospital (JRH), Oxford.These comprised fetuses from women whowere booked for delivery at that hospital andthose referred for pathological examination.Referral was often prompted by prenatal in-vestigations undertaken in Oxford, either

ultrasound examination at the JRH or throughthe Department of Medical Genetics.

Fetal hydrops was defined as generalisedsubcutaneous oedema with an effusion in atleast one body cavity. After pathological ex-amination each case was allotted to one of sixgroups: proven chromosome anomaly, dys-morphic features (two or more) suggestive ofchromosome anomaly but without cytogeneticdiagnosis, infection, visceral anomaly, compli-cation of twin pregnancy, and idiopathic.Clinical and pathological details were con-sidered first within each group and then thegroups were compared with each other. Theywere finally compared with cases of fetal hyd-rops seen between 1976 and 1982.

ResultsPertinent clinical details of each pregnancyand the results of prenatal ultrasoundexamination and chromosome culture are pre-sented in table 1, grouped according to patho-logical findings in the fetus. A comparisonbetween the groups is made in table 2 andgestational age at diagnosis, status at delivery,and pathological associations are comparedwith cases reported previously in table 3.

CHROMOSOME ANOMALYThe commonest association of fetal hydropsseen during the present study was chromo-some anomaly. There were 22 such fetuses(table 1). The hydropic state in all fetuses inthis group was identified by ultrasound scan.In seven this was during a scan before amnio-centesis. Amniocentesis was indicated becauseof advanced maternal age in six and for knownmaternal balanced translocation in one case. In13 of the remaining 15 women, a hydropicfetus was identified on routine scan, and in twowomen a detailed scan was performed becauseof raised maternal serum alphafetoproteinlevel and a history of intrauterine growth re-tardation, respectively.Half of the fetuses with chromosome anom-

aly exhibited a 45,X karyotype. In 10 of the 11,postnuchal fluid accumulation was describedin the scan report and two were already deadwhen the scan took place. Examination of thesefetuses after delivery showed that 10 out of 11exhibited tubular hypoplasia of the distal aor-tic arch (Haa). In case 11, material was notavailable for review. The only other cardiacanomaly was a persistent left superior venacava (PLSVC). Three fetuses had horseshoekidney and a further fetus had unilateral in-complete renal rotation, the renal pelvis facingarteriorly.Three of five fetuses with trisomy 21 had

cardiovascular malformations. These were

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from

Table I Clinical details of pregnancies complicated by fetal hydrops.

Case Maternal Parity Diagnosed Scan details Pregnancy details Fetal karyotypeage scan (w gest)

Chromosome anomaly1 18 0+0

2 32 2+0

3 23 0+0

4 36 1+15 18 1+0

6 21 0+07 19 0+08 25 0+09 23 0+110 29 NK11 29 3+2

12 37 0+0

13 40 3+0

14 42 4+215 39 NK

16 34 0+117 36 1+0

18 36 0+3

19 35 1+1

20 32 0+0

21 43 1+0

22 34 1+0

Suspected chromosome anomaly1 24 0+1

2 26 NK3 39 0+0

4 34 1+05 34 2+26 29 2+07 27 1+2

8 19 0+0

9 28 0+010 26 4+011 22 NK12 27 0+013 21 1+1

14 23 1+015 31 2+016 26 0+017 2218 19 0+ 1

19 27 1+020 19 0+0

Infection1 25 0+02 24 1+03 36 1+0

4 29 1+0

5 35 0+2

6 26 1+0

7 30 0+0

8 26 3+0(1 NND)

9 26 1+0

10 29 2+1

11 21 0+1

Visceral anomaly1 33 0+1

18

18

18

1818

181819161923

15

16

1617

BPD 18. Ventricles, spine normal, gross oedema, large fluid filled se AFPswelling round neck, ascites. ?Turner'sBPD 16. Gross occipital oedema, no ascites, no cardiac anomaly.?Turner'sBPD 15. Cystic hygromata, severe hydrops, hydrothorax, lowcardiac outputBPD 18. Gross oedema, ascites, large occipital sac AmniocentesisCystic hygroma, gross subcutaneous oedema, ascites, Rhydroureter, ?double outlet R ventricle. ?Turner'sBPD 18. Cystic hygroma, hydrops, small baby. ?Turner'sBPD 18. Massive oedema, large cystic hygroma. ?Turner'sIUDCystic hygromaSevere non-immune hydrops with massive cystic hygroma. IUDBPD 18. Gross fetal ascites, oedema, pleural effusion, cystic areas Placental biopsyat neck. Heart normalHydropic fetus. Massive subcutaneous oedema, normal cardiacrateGross hydrops, ascites, subcutaneous oedema particularly over Amniocentesiscervical spine. ?Turner'sHydrops and ascites, normal heart rate, ventricles normalNK Diagnosed at

22 BPD 22. Small omphalocele16 BPD 16. Massive subcutaneous oedema head and trunk, little

ascites, hydrops17 Hydropic fetus, gross subcutaneous oedema, pleural effusions,

ascites, heart normal. ?Turner's13 Gross subcutaneous oedema from head to buttocks. No ascites or

pleural effusions, heart rate normal. ?Down's18 Cystic mass over back of neck. Occiput, spine normal, oedema

cord, probable chromosome problem16 BPD 16. Hydropic fetus, normal cardiac rhythm, no obvious

cause of hydrops13 IUD

15 CR 13, IUD, small encephalocoele. ?Hydrops

13 IUD16 FH identified, no movements, no bladder seen. Membrane

attached to fetal head and abdomen. Liquor. ?Cystic hygroma16 Missed abortion. ?Hydatidiform mole13 BPD 16. Oedema, ascites, cystic hygroma17 Encephalocele20 BPD 16. Abnormal fetus, cystic areas round head, little fluid,

abnormal tissues surrounding body, IUD18 BPD 18. IUD, massive oedema, large cystic hygroma, bilateral

pleural effusions. Probable Turner's18 ?Missed abortion. ?Turner's18 IUD19 BPD 19. Cystic lesion from cervical spine to occiput20 Gross fetal abnormalities. ?Turner's20 BPD 13. No FH, IUD

21 ?Meningomyelocele, IUD21 IUDat21w25 IUD at 25 w25 IUD. ?Turner's25 BPD 18. 1 liquor, no FH, IUD

- No scan27 IUD, polyhydramnios

15 BPD 15. Hydropic, no FH15 BPD 15. Hydropic, no FH16 IUD on scan

18 Hydropic fetus, massive subcutaneous oedema, normal cardiacrhythm. ?Turner's

22 No diagnostic scan

22 BPD 22. Oedema, cardiomegaly

25 Gross hydrops, large heart, no structural abnormality, bulkyplacenta. ?a thalassaemia

26 BPD 20. Massive ascites, subcutaneous oedema over face, nocardiac anomaly, normal rate. ?Turner's

32 No FH, some oedema, stomach high

33 Polyhydramnios, fetal ascites, bilateral pleural effusions

34 Gross subcutaneous oedema with ascites and bilateral pleuraleffusions, kidneys normal, small heart, no positive evidence ofcardiac anomaly

18 Cystic lesion arising posterior head/neck region, ?encephaloceleor NTD, only 2 heart chambers, abdominal protrusion

amniocentesisPlacental biopsyKnown maternal 14;21translocationAbnormal scan atamniocentesis

Diagnosed atamniocentesis

t se AFP, 18 w size,molar placentaTOPt se AFP. Abnormalcholinesterase amn AFPClomiphene inducedTOPTOP

TOPTOP

20 w spontaneousabortion. T se AFP

Cannabis smoker till12wPrem labour at 27 wNo FMs

45,X

45,X

45,X

45,X45,X

45,X45,X45,X45,X45,X45,X

47,XY, + 21

47,XX, + 21

47,XX, + 2147,XY, + 21

47,XY, + 1846,XX, - 14, + t(14q;21q) mat47,XX, + 18

47,XX, + 18

47,XX, + 18

47,XX, + 18

46,XY/47,XY,+ 13

Failed

Not doneFailed

Not doneFailedNot doneNot done

Failed

Not doneFailedFailedFailedNot done

FailedFailedNot doneNot doneNot done

FailedFailed

Not done46,XY

Maternal influenza B Not doneinfection at 13 wt se AFP 46,XY

Clomid induced 46,XYpregnancy, pv bleeding8-13 w. Anti Mantibodies. Prem labour23wPolyhydramnios 28 w 46,XYsizeT se AFP. Parents FailedChineseN se AFP. IUD at 26 w 46,XY

No problems until 32 w. Not doneParents ChinesePolyhydramnios at 33 w. 46,XXDelivered 38 w. D 1 dyInduced 34 w for 46,XYmaternal distress.Polyhydramnios. SB

T se AFP Not done

(continued)

92 Boyd, Keeling

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from

Case Maternal Parity Diagnosed Scan details Pregnancy details Fetal karyotypeage scan (w gest)

2 19 0+ 1 25 Gross hydrops, grossly abnormal fetus Threatened abortion 46,XY25 w. Delivered 27 w

3 29 0+ 0 25 BPD 25. Hydropic fetus with enlarged heart, normal rhythm, 30 w size at 24 w. 46,XXright ventricle small, ascites, subcutaneous oedema Placental biopsy

4 29 0 + 1 27 BPD 27. Polyhydramnios, heart normal, abdominal distension, Polyhydramnios 46,XY?cystic structure, ascites, no oedema, normal placenta APH 27 w

5 29 1+0 No scan APH 27w. Preterm Not donelabour CS. Hydropicmale

6 24 0 + 1 28 Breech. Irregular heart rhythm, gross hydramnios Pregnancy telangectasia. 46,XXPolyhydramnios.SB29w

7 22 0+0 31 Abnormal brain with dilated ventricle, soft tissue mass. 37w size at 31 w. 46,XX?haematoma, hydrops, pericardial effusion Proteinuria + +

Twins1 28 1 + 0 19 Twin gestation. I. BPD 19. FH seen, spine and abdomen normal. t se AFP 46,XY

II. No FH, grossly hydropic. BDP 12. No membrane between Placental biopsy. Ifetuses. normal male, delivered

36w2 27 2+0 23 Polyhydramnios I. d 1 h, hydropic I 46,XX

II. IUD 10 dy beforedelivery at 31 w

3 36 1 + 1 26 Hydrops in 1 twin only, ascites Polyhydramnios I 46,XXPrem labour, deliv 26 w.I hydropic d 40 h. II d5 dy. IVH

4 25 0 + 0 27 I. Massive hydrops, ascites, distended bladder, hydrocalyces, no Thr abortion following Not doneother structural anomaly. II. Squashed amnio. I. SB hydropic.

II. d 10 dy5 27 1+ 0 19 Twins. One twin omphalocele or gastroschisis, gut only Prem labour, APH at 46,XY

32 w 46,XY32 I hydropic, II omphalocele. IUD

6 24 1 + 1 No diagnostic scan Spont lab 39 w, ?IUD 1 ?I 46,XXweek previously. Isurvived

Idiopathic1 26 3+ 1 16 Large cystic spaces in neck, bilateral pleural effusions Amniocentesis 46,XY2 31 2 + 0 17 Moderate bilateral pleural effusions, subcutaneous oedema over Placental biopsy 46,XY

scalp and trunk, cardiac rate normal3 24 1+0 20 Hydrops. ?Pelvic teratoma IUD at 20 w Failed4 36 0+2 24 BPD 20. No FH Pergonal induced 46,XX

pregnancy5 38 2+ 1 32 No FH. Massive pleural effusions, ascites, hydrops Maternal age amnio, SB 46,XY

at 32 w6 26 1 + 0 33 Head too low to measure, large amount of fluid in chest, thick Hydramnios at 33 w, 46,XX

layer over abdomen term size. Live born.Survived, mild mental &motor delay

BPD(18) = biparietal diameter equivalent to (18) weeks gestation.IUD = intrauterine death.NK =not known.w = weeks' gestation.d= died.dy = days.se AFP = serum alphafetoprotein.FMs = fetal movements.FH = fetal heart.CR13 = crown-rump measurement equivalent to 13 weeks' gestation.

ventricular septal defect (VSD), VSD plusHaa, and Haa in addition to external dysmor-phic features. Three of five fetuses with tri-somy 18 karyotype also had cardiac defects(atrial septal defect plus VSD, VSD plus Haa,and Haa plus aberrant right subclavian artery).The mosaic trisomy 13 fetus exhibited many

dysmorphic features including cleft lip andpalate; the heart had a globular appearance butnormal anatomy, a finding in trisomy 13 at

term. Eight out of 22 of the women whosehydropic fetus had a chromosome anomalyhave achieved one or more subsequent suc-

cessful pregnancies.

SUSPECTED CHROMOSOME ANOMALY

There were 20 fetuses where necropsy showedtwo or more anomalies suggestive of an abnor-mal karyotype but where tissue culture forchromosome analysis was either unsuccessfulor not possible because the fetus was receivedafter fixation. Hydrops was found on ultra-sound examination in 19 of 20 cases during thesecond trimester. Eight scans were undertakenfor a variety of clinical reasons, three becauseof raised maternal serum alphafetoprotein(AFP) levels, and hydrops was detected duringa routine scan in eight women. Thirteen ofthese fetuses were dead at the time of diagnosis

Table 2 Clinical details of hydropic pregnancies. Comparison between different diagnostic groups.

Pathological diagnosis Chromosome Suspected Infection Visceral Twin Idiopathicanomaly chromosome anomaly

anomaly

No of cases 22 20 11 7 6 6Mean maternal age (y) (SD) 31 (8) 26 (5) 28 (5) 26 (5) 25 (4) 30 (5)Mean (SD) gestation at 17 (3) 19 (4) 24 (7) 26 (5) 23 (4) 24 (7)diagnosticscan (w) (n=20) (n= 19) (n= 10) (n= 5) (n= 5) (n= 6)Chromosome analysis 0/22/0 0/0/20 5/0/6 5/0/2 5/0/1 3/0/3normal/abnormal/failed

93Fetal hydrops

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from

94Boyd, Keeling

Table 3 Comparison between previously reported casesfrom this centre and present series.

Period seen 1976-822 1983-87

No 50 72Before21 w 3 41

21-27 w 16 2228 w to term 31 9

Abortus 19 56Stillbirth 18 9Live birth (survival) 13 (3) 7 (1)Diagnosis after pathexamination 42 (84%) 66 (92%)Chromosome anomalydiagnosed 5 22Chromosome anomalysuspected 3 20Infection 8 11Visceral anomaly 23 7Twin 1 6Other 2 -

Idiopathic 8 (16%) 6 (8 0%)

which in part explainsanalysis.

failure of cytogenetic

There was very prominent postnuchal fluidaccumulation in addition to generalised hyd-rops and effusion in body cavities in 17 of the20 fetuses in this group; this had been noted onscan in six cases. In 12 of the 17, the distalaortic arch was hypoplastic and six hadanother cardiovascular anomaly (bicuspid aor-tic valve, PLSVC plus aberrant right sub-clavian artery, PLSVC plus single umbilicalartery, bicuspid aortic valve plus PLSVC,single coronary origin, and atrioventricularcanal defect). Two further fetuses with post-nuchal fluid had cardiovascular anomalies,VSD with right sided aortic arch and VSD,respectively. The three other fetuses hadgeneralised oedema and effusions withoutgross postnuchal fluid accumulation, althougha fetus with overfolded fingers had a postnu-chal bleb. One fetus had cleft lip and palate,meningomyelocele, VSD, and transposition ofgreat arteries, and the last had syndactyly,septum primum atrial septal defect, and amolar placenta.

Eight of these women have had a subsequentsuccessful pregnancy.

INFECTIONEleven of our cases had clinical and/or histolo-gical findings suggestive of an infective aeti-ology for the hydropic state. Hydrops wasidentified prenatally by ultrasound examina-tion in 10 of 11 cases. In seven, this was before28 weeks of gestation and in four after thattime. Seven women were scanned for clinicalreasons, one because of raised maternal serumAFP, and two abnormal fetuses were identi-fied during routine scan. Table 1 shows clin-ical details and scan findings. Both parents ofcase 7 were of Chinese origin; a thalassaemiawas suspected and apparently excluded by theresults of haematological investigations. Rein-vestigation in a subsequent pregnancy showedthat both parents were carriers for ox thalassae-mia. Three women had normal scans per-formed before the diagnostic scan. In fourcases, ultrasound scan examination showed nofetal heart contractions and two more (cases 8and 1 1) died before delivery.

Evidence of infection took the form of roundcell infiltration, most readily apparent in thestroma of placental villi, lung parenchyma, andmyocardium. Large numbers of circulatingnucleated cells were apparent in many organsin four cases and prominent haemopoiesis wasobserved in placental villous capillaries infour. In three of the former (cases 1, 2, 7, and9), intranuclear inclusions were apparent inerythrocyte precursors and macrophageswhich reacted positively for parvovirus RNAusing an in situ hybridisation technique. Incase 3 intrauterine death occurred during thecourse of antibody proven influenzal infection;round cell infiltration was present in placenta,lung, myocardium, and skeletal muscle. Themost mature baby in this group (case 11)exhibited a striking frog leg posture after de-livery. Skeletal muscle from several sitesshowed a pattern of groups of fibres of normalsize separated by groups of small size fibresand scattered very large fibres. A silverimpregnation technique showed regenerationof nerve endings within muscle suggestive ofanterior horn cell disease, similar to that in theacute stage of polio virus infection. There wasdense cellularity of the lamina propria atseveral sites in the small intestine and calcifica-tion of intestinal contents suggesting previousenteroviral infection.

VISCERAL ANOMALY GROUPSeven cases were identified and in six prenataldiagnosis of hydrops was made by ultrasoundscan (table 1). Four were diagnosed before 28weeks of gestation. Three cases (1, 6, and 7)had normal level one scans during the secondtrimester. Four of these fetuses deliveredspontaneously between 27 and 31 weeks ofgestation after polyhydramnios and antepar-tum haemorrhage in three and pre-eclampsiaand polyhydramnios in one case. In case 1,pregnancy was terminated after recognition ofanomaly.The anomalies in this group were diverse.

Case 1 had laryngeal atresia with overdis-tended lungs and a complicated cardiac defectcomprising situs inversus, atrial septal defect,double outlet right ventricle, and mitral atresiawith a midline abdominal wall defect.

Case 2 had facial dysmorphism and shortlimbs with curved tibiae which had suggested aprenatal diagnosis of osteogenesis imperfecta.Internal examination showed a large liver andgrossly enlarged, cystic kidneys. Histologicalexamination of both showed cystic dysplasiadiagnostic of Meckel-Gruber syndrome.

In case 3, a large heart with a small rightventricle was observed on ultrasound scan inaddition to hydrops and ascites. Necropsy con-firmed cardiomegaly and showed biventricularhypertrophy in an anatomically normal heart.Histological examination of horizontal slicesthrough the ventricular mass showed myocar-dial hypertrophy at several levels, particularlyin the anterior wall of the right ventricle, butalso present in the interventricular septum. Apale subendothelial layer was apparent in bothventricles on H & E staining, which comprised

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from

Fetal hydrops

collagen and elastic fibres, characteristic ofsubendocardial fibroelastosis.

Case 4 had disproportionate ascites and onlymild subcutaneous oedema. The lungs weregrossly distended and completely filled thechest. The upper airways were patent buthistological examination of the lungs showed apaucity of intrapulmonary bronchi with noairways identifiable between lobar bronchi andterminal bronchioles. There were no anomal-ies in other systems.

In case 5, the only anomaly was a greatlyexpanded right lobe of the liver which wascompletely replaced by a vascular hamartoma.

Case 6 also had a vascular anomaly. Inaddition to fetal hydrops, prominent neck ves-sels were apparent on ultrasound scan andwere sufficiently abnormal to suggest a cervicalvascular malformation. Necropsy showed thatthese vessels were much enlarged but other-wise normal. There was diffuse angiodysplasiaof the whole of the cerebral and cerebellarleptomeninges, apparent on naked eyeexamination as an increase in both size andnumber of leptomeningeal vessels. This vascu-lar abnormality was confirmed on histologicalexamination.The last baby in this group was found to be

hydropic on ultrasound scan during investiga-tion of extremely large uterine size accompan-ied by signs of pre-eclampsia. In addition tofetal hydrops and polyhydramnios, hydro-cephalus and cerebral haematoma wereobserved. Necropsy confirmed hydrocephalusand both intraventricular and intracerebralhaemorrhage. One portion of the 'haematoma'appeared different; histological examinationshowed that this was a very vascular primitiveneuroectodermal tumour, presumably thesource of haemorrhage which had obstructedthe aqueduct of Sylvius and caused hydro-cephalus. Immunohistochemical staining ofthe tumour did not show any type of cellulardifferentiation.

TWINSAll six cases in this group were products ofmonochorionic, diamniotic twin gestation(table 1). In five, hydrops was diagnosed pre-natally by ultrasound scan. Three had normalscans earlier in pregnancy. Only one of eachtwin pair was hydropic but four of the non-hydropic twins also died. In five pairs, hydropswas the result of twin-twin transfusion. In thelast case there was an acardiac, acephalic co-twin. Two of the hydropic twins and two non-hydropic co-twins who died showed evidenceof longstanding ischaemic brain injury.3 Inone of the latter this was complicated by intra-cerebral and intraventricular haemorrhage.One hydropic and one non-hydropic twinfrom different gestations had cardiac ventricu-lar hypertrophy and endocardial fibroelastosis,probably reflecting myocardial ischaemia com-plicating high output failure.

Five women had successful singleton preg-nancies before the index pregnancy. No subse-quent pregnancy is, as yet, recorded.

IDIOPATHIC GROUPExamination of six fetuses showed no patholo-gical abnormality other than hydrops. Fourwere identified by ultrasound examinationduring the second trimester. Two were alreadydead. In two pregnancies second trimesterultrasound scans had been normal.These six fetuses showed no macroscopic or

microscopic evidence of infection and nomajor visceral anomaly was detected. Threefetuses had a normal karyotype including one(46,XY) with dysmorphic features of the headand limbs and a septum secundum atrial septaldefect. Another fetus had massive postnuchalfluid accumulation but no cardiac defect. Pla-cental or disproportionate hepatic erythro-poiesis was not apparent in any case, althoughnucleated cells were prominent in both fetaland placental vessels in case 3.Comparisons between the different dia-

gnostic groups are shown in table 2. Whileaverage maternal age in the chromosome an-omaly group was not high, the mean age ofwomen with trisomic 21 and 18 pregnancieswas 37 years. Scan diagnosis in this group wasmade earlier than in other groups, averagegestation at diagnosis was 17 weeks in thisgroup, and 19 weeks among fetuses with sus-pected chromosome anomaly.Compared with cases of non-immunological

hydrops reported previously,2 hydrops wasdiagnosed more frequently while the fetus wasstill alive, before 20 weeks' gestation, and morefrequently associated with chromosome anom-aly. A cause or association of fetal hydrops wasfound more often in the cases reported here(92%) compared with 84% of those investig-ated between 1976 and 1982.

DiscussionThe hydropic state results from the com-bination of a total increase in body water and arelative increase in fluid in the interstitialspace. When the latter becomes severe there isusually free fluid in body cavities as well.Maintenance of normal fluid distribution re-quires a balance of osmotic and hydrostaticforces on either side of cell membranes and anintact sodium pump. The common mechan-isms, severe chronic anaemia, fetal heart fail-ure, and obstruction to venous return from theplacenta, act through a common pathway atthe cellular level of hypoxia but severe hypo-proteinaemia may be the starting point ofhomeostatic failure.4Non-immunological fetal hydrops has come

to the fore as a clinical problem as effectiveprevention and treatment of rhesus hydropshas been widely implemented. A wide range ofdisorders have been reported in associationwith non-immune hydrops including a diverserange of malformations, chromosome anom-aly, genetic metabolic disease, tumours, andinfection,4 but when the associations observedin larger series are combined, cardiac andpulmonary malformations, cardiac arrhyth-mias, chromosome anomalies, and monochor-ionic twin pregnancy are the commonest asso-ciations encountered.4 In a review of published

95

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from

Boyd, Keeling

reports of fetal hydrops in the 1980s, Machin'noted identification of an associated conditionin 80 to 85% of cases. Only 10% had abnormalchromosomes. The higher proportionreported here reflects the lower gestation atdiagnosis among our cases. Many of thefetuses described here had cardiovascular an-omalies. A causal relationship between mal-formation and hydropic state is put forward bymost authors, supported by the in utero de-monstration of cardiac arrhythmias in the hyd-ropic fetus with cardiac anomaly67 and anexcess of complex cardiac lesions withdisproportionate representation of ambiguouscardiac situs.26 McFadden and Taylor8 cautionthe uncritical acceptance of such relationshipsin respect of cardiac malformations, a sugges-tion which has implications for management offetal hydrops in the presence of a demon-strated cardiac anomaly.

Interest in the causes, associations, andmanagement of non-immunological hydropsin this unit was stimulated by the resourceimplications, both to the obstetric and neo-natal intensive care units, of delivery by emer-gency Caesarean section of known hydropicinfants when they exhibited signs of fetal dis-tress both before and during labour. Thesebabies required a great deal of medical andsurgical attention but their mortality washigh.'29 In 31 of 50 pregnancies describedpreviously, hydrops became apparent in thelast trimester of pregnancy. In 13 pregnanciesdiagnosis was made by ultrasound examina-tion, although in three the fetus was alreadydead. Since 1982, many changes have takenplace in obstetric practice including the fre-quent recourse to ultrasound scan in thesecond trimester of pregnancy for a variety ofreasons including estimation of gestation, rou-tine fetal anomaly screen, and investigations ofintrauterine growth retardation. In future firsttrimester vaginal scanning might permit veryearly detection of some hydropic fetuses.'0This study was undertaken to examine thecauses and associations of fetal hydrops and tocompare these with our previous experience tosee whether changes in the management ofpregnancy resulted in more frequent positivediagnoses or affected the survival of the hydro-pic fetus.We drew attention to the difference in asso-

ciations of fetal hydrops between those fetuseswhere the hydropic state was diagnosed before28 weeks' gestation and those where it wasobserved in the last trimester of pregnancy.2The larger proportion of cases where chromo-some anomaly was found or suspected becauseof a combination of dysmorphic features andstructural anomalies reported here reflects thegreater number of second trimester diagnoses.It is likely that most of these fetuses would diein utero if termination of pregnancy had notbeen undertaken after demonstration of anabnormal karyotype.

Eleven of the hydropic fetuses had evidenceof infection. The preceding maternal infectionusually produces only minor symptoms or mayeven go unnoticed. The existence of pairedsera for antibody studies is unlikely. The

advent of techniques, such as in situ hybridisa-tion for identification of viral protein in fixedtissue samples, has already permitted unequi-vocal diagnosis of parvovirus infection in thehydropic fetus." The extension of this tech-nique will allow us to characterise other fetalinfections fully. Although the parents of case 7were subsequently shown to be carriers for athalassaemia, the fetus exhibited round cellinfiltration of the myocardium, and both focalvillitis and intervillositis were seen in the pla-centa; the excessive haemopoiesis in the liverand in villous capillaries is compatible withchronic anaemia. We feel therefore that it islikely that the fetus was homozygous for athalassaemia.Among the more mature hydropic fetuses,

three showed evidence of infection, two hadvisceral anomaly, two were twins, and twowere unexplained. The hydropic state in theonly survivor of this group was unexplained. Itis in the group of more mature fetuses wherecareful ultrasound examination of the fetus isparticularly important'2 so that those withmajor, untreatable malformations can be iden-tified, cardiac arrhythmias sought and treated,7or treatment with albumin or blood productsgiven to try to restore normal fluid balance."When contemplating treatment, it is importantthat the assessment is thorough and treatmentprompt because of the risk of cerebral hypoxicinjury.3 14

Most of the associations of fetal hydropsdescribed here have a low risk of recurrence.Occasionally, diagnoses of disorders of known,higher recurrence risk, such as Meckel-Grubersyndrome, infantile arterial calcification,2 andsome skeletal dysplasias,4 are made in associ-ation with fetal hydrops. It is most importantthat the hydropic fetus, whether from termina-tion of pregnancy or spontaneous abortion, isexamined with great care so that any associ-ation is demonstrated. In most cases, the par-ents can be confidently, optimistically coun-selled about prospects for future pregnanciesand the unfortunate few prepared for early,appropriate investigations.

We are grateful to Shirley Martin for typingthe manuscript.

1 Gough JD, Keeling JW, Castle B, Iliff PJ. The obstetricmanagement of non-immunological hydrops. Br 3 ObstetGynaecol 1986;93:226-34.

2 Keeling JW, Gough DJ, Iliff P. The pathology of non-rhesus hydrops. Diagn Histopathol 1983;6:89-1 1 1.

3 Squier MV, Keeling JW. The incidence of prenatal braininjury. Neuropathol Appl Neurobiol 1991;17:29-38.

4 Keeling JW. Hydrops fetalis and other forms of excess fluidcollection in the fetus. In: Wigglesworth JS, Singer DB,eds. Textbook of fetal and perinatal pathology. Oxford:Blackwell Scientific Publications, 1990:429-54.

5 Machin GA. Hydrops revisited: literature review of 1,414cases published in the 1980s. Ant _7 Med Genet1 989;34:366-90.

6 Kleinman CS, Donnerstein RL, DeVore GR, et al. Fetalechocardiography for evaluation of in utero congestiveheart failure: a technique for study of nonimmune fetalhydrops. N Engl 3 Med 1982;306:568-75.

7 Maxwell DJ, Crawford DC, Curry PVM, Tynan MJ, AllanLD. Obstetric importance, diagnosis, and management offetal tachycardias. BMJ 1988;297:107-10.

8 McFadden DE, Taylor GP. Cardiac abnormalities andnonimmune hydrops fetalis: a coincidental, not causal,relationship. Pediatr Pathol 1989;9:11-17.

9 IliffPJ, Nicholls A, Keeling JW, Gough JD. Non-immuno-logic hydrops fetalis: a review of 27 cases. Arch Dis Child1 983;58:979-82.

10 Szab6 J, Gellen J, Szemere G. Non-immune hydrops intrisomy 18. Diagnosis by vaginosonography and chorion

96

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from

Fetal hydrops 97

villus sampling in the first trimester. BrJ Obstet Gynaecol 13 Shimokawa H, Maeda H, Miyamoto S, Koyanagi T,1990;97:955-6. Nakano H. Intrauterine treatment of idiopathic hydrops

11 Porter HJ, Khong TY, Evans MF, Chan VTW, Fleming fetalis. J Perinat Med 1988;16:133-8.KA. Parvovirus as a cause of hydrops fetalis: detection by 14 Kobori JA, Urich H. Intrauterine anoxic brain damage inin situ DNA hybridisation. J Clin Pathol 1988;41:381-3. nonimmune hydrops fetalis. Biol Neonate 1986;49:311-7.

12 Watson I, Campbell S. Antenatal evaluation and manage-ment in nonimmune hydrops fetalis. Obstet Gynecol1986;67:589-93.

on June 14, 2020 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.29.2.91 on 1 F

ebruary 1992. Dow

nloaded from