The state of the device and copper levels in the products of conception in women becoming pregnant with a copper-bearing IUD in situ Christine Gosden, A. Ross, A. McGovern and W. Reid MRC Clinical and Population Cytogenetics Unit, Western General Hospital, Crewe Road, Edinburgh EH4 2XU; *MRC Brain Metabolism Unit, 1 George Square, Edinburgh; and ^Department of Obstetrics and Gynaecology, Eastern General Hospital, Seafield Road, Edinburgh, UK.

Summary. Of the 8 pregnancies studied, 2 were of small gestational sacs with blighted ova and were associated with devices in which the copper wire had very high detectable X-ray emissions for copper (>90%). In 2 other pregnancies intrauterine deaths had occurred by the time of termination at 13 and 17 weeks and copper levels in the products of conception were variable. There was no abnormality of the fetus in the term pregnancy but it seems possible that copper can affect the early growth and development of the embryo. On only 1 of the 8 devices was any great amount of calcium deposited and it is therefore considered unlikely that calcium deposition increases the risk of pregnancy by preventing the release of copper.

Introduction

When copper wire is wound round an inert plastic intrauterine device, the copper seems to increase the contraceptive efficiency of the device. The way in which the copper is thought to exert this effect is not known, but it is believed to involve some toxic effect upon spermatozoa, ovum, blastocyst, embryo, uterine milieu or enzyme systems. The role of copper on mammalian reproduction has been comprehensively reviewed by Oster & Salgo (1977). Our recent findings (Ross, Gosden, Loudon & Foxwell, 1978) that the copper wire of a copper-bearing intrauterine device becomes coated with a layer of calcium in some women, but not in others, shows that there are different uterine responses to the device. As the deposition of calcium appears to inhibit the erosion of copper, the question arises whether the risk of pregnancy is greater in women in which the IUD becomes calcified. This paper reports our findings of copper-bearing intrauterine devices removed from women who became pregnant with the device in situ. We also examined the products of conception and measured their copper levels for correlation with the state of the device.

Case histories Patient 1

This woman began vaginal bleeding at 9 weeks gestation and at 10 weeks a gestational sac was spontaneously aborted. This intact blood-filled sac contained a small embryo of 2 mm crown-rump length. The sac was consistent in size with a gestation of 10 weeks. A dilatation and curettage evacuation of the uterus was performed and the intrauterine device was removed. Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access Patient 2

Vaginal bleeding began at 9% weeks gestation and the IUD was removed and sent to the laboratory for analysis. After 3 days of heavy blood loss per vaginam a dilatation and curettage were carried out. The gestation was estimated at 10 weeks from measurements of the fetal limb and 10 weeks by dates.

Patient 3 This patient requested termination of pregnancy at an estimated gestation of 10 weeks by dates. This was carried out by suction evacuation, and the IUD was removed at the same time. There was no evidence of a fetus and only portions of sac estimated at about 8 weeks in size which may have contained a blighted ovum or very small embryo.

Patient 4

This patient requested termination of pregnancy. The device was removed immediately before termination was induced with prostaglandin E-2 administered extra-amniotically. The patient's gestation was 16 weeks by dates but the macerated male fetus which was delivered had a crown-rump length of 8 cm and a heel-toe length of 1-2 cm, giving an estimated gestation based on size of only 13-14 weeks. The fetus had no major external abnormalities and no major organ system abnormalities were detected.

Patient 5

The device was removed at 5^ weeks gestation and sent to the laboratory for SEM and X-ray microprobe analysis. The patient then requested termination of the pregnancy which was carried out at 8 weeks. The products of conception were not available for study.

Patient 6 The tail of the IUD was seen at the first examination after the positive pregnancy test, but the device was not removed. Labour was induced at 41 weeks gestation. The IUD was removed immediately before artificial rupture of the membranes. There was a forceps delivery of a live male child who cried immediately and had Apgar scores of 9 at 1 min and 10 at 5 min. The child had no external abnormalities.

Patient 7

When the pregnancy test was positive at 5 weeks gestation, the device was removed and sent to the laboratory to be processed for SEM and X-ray microprobe analysis. The patient requested early termination of pregnancy. The products of conception were not available for study.

Patient 8

The pregnancy test was positive at 5 weeks. At 20 weeks gestation no fetal heart could be detected. Since the fetus had died in utero, abortion was induced with prostaglandin E-2 administered extra-amniotically after removal of the IUD. After 18 h a macerated female fetus was delivered. The fetus had a crown-rump length of 12-5 cm indicating a gestation of 17 weeks by size. There were no major external or internal abnormalities. The IUD was not available for study.

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access Patient 9

This fetus was used as a control when the pregnancy of a 22-year-old healthy woman who was using no contraception was terminated for social indications. Complete abortion occurred 16 h after prostaglandin E-2 was administered extra-amniotically. The fetus appeared fresh, was not macerated, and careful examination revealed no major external or internal abnormalities of the fetus.

Materials and Methods

When removed from the patient the copper IUDs (Gravigard devices, Searle) were placed into dry sterile containers and immediately transported to the laboratory. The device was cut into small portions and each was mounted onto an aluminium stub, dried in a dessicator and finally coated with a layer of carbon. One portion of the dried IUD was embedded in Araldite (CIBA) and cut to give cross-sections of the device which were then mounted and coated with carbon. All available products of conception, including material from dilatation and curettage in theatre, were collected into dry sterile containers and transported to the laboratory. All specimens were examined, photographed, weighed and measured. Samples were removed for estimation of copper levels and for study of embryonic development, histology, tissue culture and cytogenetic studies (Gosden, Wright, Paterson & Grant, 1976). Small specimens and incomplete portions of tissue were examined under a dissecting microscope before further study. The remainder of the products was preserved in buffered formalin.

Electron microscopy and X-ray analysis All specimens of the IUDs were examined and photographed with a scanning electron microscope (Stereoscan S180, Cambridge Instruments Ltd) fitted with an X-ray microprobe analyser (Link Systems Ltd). The accelerating voltage used was 30 KeV as described previously (Gosden, Ross & Loudon, 1977).

Atomic absorption spectrophotometry to estimate copper Aliquots of tissue were weighed into small plastic tubes (Eppendorf) and digested in 1-5 m- nitric acid at 80°C for 6 h and then at 105°C to dryness. The residues were then re-dissolved in 500 ul distilled water. Flameless atomic absorption analysis was carried out with an atomic absorption spectrophotometer (Model 360: Perkin-Elmer), fitted with a heated graphite atomizer (Regan & Warren 1976). Copper was estimated at a wavelength of 324-7 nm. Aqueous solutions of cupric nitrate (B.D.H. Poole, Dorset) containing 0-1, 0-2 or 0-4 pg copper/ml were prepared as standards. All solutions were made up in de-ionized distilled water. Each sample (50 ul) was injected into the graphite furnace and after the initial pre-drying period to reduce spraying, the material was dried, thermally destroyed and atomized. Argon was the inert gas used (flow rate 300 ml/min), the gas stop mode being selected during atomization. Specimens with a concentration above the standards were further diluted (Sperling, 1976).

Results

The results are summarized in Table 1. As shown in Figs 1(b)—7(b), the surfaces of the copper wire from the 7 devices available for study by scanning electron microscopy were very different from the smooth surface of an unused device. Figures l(c)-7(c) illustrate the cross-sections of the wire and Figs l(a)-7(a) represent the X-ray emissions from the surface of the wire. The of an of a whole mount to the cross-sectional is relationship X-ray analysis Downloaded fromfindings Bioscientifica.comexplained at 10/07/2021by 09:25:11AM Ross et al (1978). via free access r- O < + 3rt <<<<< < ZZZZ Z

(N V) » << < <

a. <<<<<<<<<< o. zzzzzzzzzz o

+, <<<<<<<<<< zzzzzzzzzz . 5£ e ce C cm

m + ¬ <<<<<<<<<< -* - * c/3 zzzzzzzzzz u

X>

o O 3 —' »n m ô vi vi ^ 00 Tt ce -o .-73

o !* + irixm VO \400 Zh-Z«ZZZZZ .S 'S s e o •s < *-» "> .2 3 Z x> 3 -O u ^o vo» " - o « O M ·3< , +, O ,«> 3 > - CU CO u

o e o r*· >n rN ^" >> , + -^ t* Ô O o * S3 — w co zîzzzzzz

3

C co ce 03 3 g £ en e g- '-s E i TÜÍ < 3 &S-3 8.2 c ° D. <*> g-S g s « ce ! -O E s 28 · S3 ri"!«j <« o o 2 tí |3 * c « w « j| 5 E S S . .S s ^ o· Q SfOQS g o cu«-

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access Patient 1

The copper wire had a cracked and peeling surface (Fig. lb). Only small amounts of sulphur and chloride were present with a large amount of copper (Fig. la). The surface appearance of the wire is probably due to erosion of the copper because very little calcium was detected. The degree of erosion and deposition was as shown in Fig. 1 (c). Copper levels in the endometrium and decidua were low but the blood clot contained a lot of copper.

Patient 2

This device had a deposit which not only covered the copper wire but also bridged the gaps between the coils (Fig. 2b). One side of the device had a thick deposit which completely obscured the copper wire. The deposit consisted of sodium, phosphorus, sulphur, chloride and potassium; very little copper was detectable (Fig. 2a). The erosion and deposition layers were extensive (Fig. 2c) but it was not clear whether the erosion had occurred before the deposition of the other elements or the copper erosion had continued in the presence of the deposition layer.

Patient 3 The surface of the wire was only slightly roughened (Fig. 3b) and it consisted of small quantities of sodium, sulphur, chloride, potassium and calcium and much copper (Fig. 3 a). There was little erosion and very little deposition (Fig. 3 c). However, copper levels in the products of conception were low.

Patient 4

The coil was thickly encrusted (Fig. 4b) and 80% of the X-rays emitted were due to calcium (Fig. 4a). Copper erosion was slight, presumably because of the very thick deposit of calcium (Fig. 4c), but the blood clot from the uterus at abortion had high copper concentrations.

Patient 5

The surface of the copper wire was discontinuous (Fig. 5b) and there was little surface deposit (Fig. 5c). There was little calcium present, but 54% of the X-rays emitted were from sulphur, phosphorus, sodium, chlorine and potassium (Fig. 5a). The thin deposition layer would explain the variable erosion of the copper.

Patient 6 The degree of copper erosion was extensive (Figs 6b and 6c) while the deposition layer was minimal (Figs 6c and 6a).

Patient 7

The copper wire surface was fairly uniform in appearance (Fig. 7b). There was little erosion of the copper (Fig. 7c) but, as shown in Fig. 7(a) 28% of the X-rays emitted by the deposition layer were due to calcium and 14% to other elements.

Patients 8 and 9 The tissue levels of copper in the products of conception from these women were similar, although Patient 9 was not using an IUD. Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access Fig. I

10240

Fig. 2

10240 eV

Figs 1-3. Observations on the IUDs removed from Patients 1 (Fig. 1), 2 (Fig. 2), and 3 (Fig. 3) (see Table 1), showing (a) the X-rays emitted from the surface of the wire and their correspon¬ dence with particular elements, (b) the appearance of a whole mount of part of the copper wire round the IUD, and (c) a cross-section of a strand of the wire to indicate the degree of erosion and deposition.

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access (b) (c)

30/ím

300//m

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access (c) (b)

30 tir 300 µ

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access (a)

Fig. 4

Cu Tv I 10240

Fig. 5 Ca

I I -I-I I I I

. . 10240

Cu

Fig. 6

Cu ± * zL· 10240

Cu

Fig. 7

10240 eV Figs 4-7. Observations on the IUDs from Patients 4 (Fig. 4), 5 (Fig. 5), 6 (Fig. 6) and 7 (Fig. 7) (see Table 1), showing (a) the X-rays emitted from the surface of the wire and their correspon¬ dence with particular elements, (b) the appearance of a whole mount of part of the copper wire round the IUD, and (c) a cross-section of a strand of the wire to indicate the degree of erosion and deposition.

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access Discussion

The pregnancy rate with copper 7 IUDs has been reported as being less than 5% per annum (Newton, Illingworth, Elias & McEwan, 1977) but the present findings that devices had allowed pregnancies to occur whilst they were in utero has serious implications both for the suggested methods of contraceptive action of copper-bearing IUDs and for possible teratogenic actions of copper on the fetus. All the postulated mechanisms of action of the copper bearing IUDs involve the release of copper (Chang & Tatum, 1970; Hagenfeldt, 1972; Tatum, 1974). Therefore, in the women who conceived with an IUD in utero, the IUD may not have been releasing sufficient copper for it to have a contraceptive role, perhaps because a layer of elements deposited on the surface of the wire had prevented the release of copper. It might therefore be expected that the thicker the deposit on the wire, the more likely a pregnancy is to occur. However, although some very thick calcium layers (up to 45 pm) have been observed (Gosden et al., 1977), and over 25% of women who have a copper-bearing IUD in situ for longer than 6 months deposit calcium on these devices (Ross et al, 1978), only one pregnancy has so far been observed in a woman with an appreciable amount of calcium on her device (Patient 4 in this present study). This could be because calcium is involved in the mechanism of action of these devices or that a uterus which has this particularly strong 'foreign body' reaction and coats copper-bearing IUDs with calcium also effectively prevents pregnancy from proceeding. Such a reaction becomes an appealing hypothesis when one considers that more than 25% of women have deposited calcium in 2 years whilst the pregnancy rate associated with these devices is less than 5% per annum (Newton et al, 1977). The results for the IUDs from 8 women in whom conceptions had occurred in the present study must be compared with those from 100 normal women who had not become pregnant with the IUD in situ (Ross et al, 1978). These 100 devices showed considerable variation in the elements present and their erosion patterns, but 25% showed some degree of calcification: only 1 of the 8 IUDs in the present study had marked calcification. In 70% of the women in whom pregnancy and calcification had not occurred, copper erosion was quite rapid and the devices had only a small deposit of other elements, namely sodium, phosphorus, potassium and chlorine. The detection of large amounts of sulphur and other elements on the IUDs of 4 of the pregnant women is therefore unusual, occurring in only 8% of non-pregnant women with IUDs. It is difficult to understand why a woman who has a device in situ for 2 years should suddenly become pregnant, but a change in any one of several factors such as size of the uterus, uterine pH, volume of uterine secretion, endometrial thickness, volume of menstrual blood loss, cellular colonization of the device, infection and calcification, may be sufficient to upset the efficiency of the copper-bearing IUD. The individual response of women is difficult to predict. The findings of an elevated cord blood caeroluplasmin level in a neonate (Alderman, 1976) and a spontaneously aborting anencephalic fetus (Gosden et al, 1977) conceived in the presence of a copper-bearing IUD have caused concern because it has been suggested that high copper levels might cause fetal abnormality. Barrie (1976) has discussed the problems of teratogenesis and limb reduction deformity. Many women seek termination of pregnancies or abort spontaneously after conceiving with an IUD in situ, leaving only a small proportion of women to carry their pregnancies to term; some do so with the copper device still in situ, others have the device removed when the pregnancy is recognized. In the present study 2 IUDs had very high levels (>90%) of copper and these pregnancies had not developed normally: there was severe developmental arrest and only a small gestational sac without a fetus could be found. In 2 of the other pregnancies there was intrauterine death and high copper levels may therefore affect growth and development before causing abnormal development. Tatum, Schmidt & Jain (1976) studied 918 women in whom conception occurred with a copper-bearing IUD in situ. Of the total of 918, 643 had the Copper T200; 191 had the Copper Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access T300; 37 had the copper 380A and 47 had the Copper T2200 (the 380A and T2200 devices have copper sleeves on the arms of the device as well as the on the stem). The outcome of 773 of those pregnancies was known: of the 746 uterine pregnancies 578 aborted (465 selective and 113 spontaneous), 147 proceeded to the birth of a full-term liveborn infant and 21 were premature births, 4 of which were stillborn. Thus only 22-5% of the fetuses developed to an age or size at which practical detection of congenital abnormalities could be undertaken. The question of the mental development of these children remains to be established. An examination of infants at term thus represents only a small proportion of the whole. The 2 devices in the present study associated with the severe developmental arrest of the conceptuses (Patients 1 and 3) had the most copper X-rays detected at the surface of their devices, 92% and 95% respectively. Both were relatively new devices. It is possible that high copper levels present in early pregnancy have a severe effect upon the embryos. The final measurable tissue copper levels were relatively low, with the exception of the blood clot in Patient 1 which had exceedingly high copper levels. Thus, in women in which the device was not in association with the gestational sac, the levels of copper at 10 weeks were low. However, the high copper release rates in the uterus early in pregnancy may have a deleterious effect upon the progress of the pregnancy. All the other pregnancies, with the exceptions of Patients 5 and 7 (from whom fetal material was not available) had proceeded to a stage at which there was a recognizable fetus. Perhaps one of the cases demanding most attention is that of Patient 6; here a device which had been in situ for 36 months had been co-existent with a pregnancy proceeding to term. Of all the X-rays detected on the surface of the device 76% were attributable to copper. However, the total erosion of only 25 pm of copper in 3 years, with very little deposition of other elements (only 3% calcium and 20% of other elements including sulphur, potassium, phosphorus, chlorine and sodium) suggests that the main uterine response in this patient was to erode copper from the device very slowly indeed, probably in insufficient quantity for contraceptive effectiveness. Two other cases are rather disturbing. These are those of Patients 4 and 8 in whom the fetus was macerated and intrauterine death had occurred some time previously. The device from Patient 4 was available for study, and only 16% of all the X-rays detected at the surface were still attributable to copper after the device had been in situ for 3 years. However, 80% of the X-rays were emitted from calcium which formed quite a thick deposit as shown in the cross- section. Again, a blood clot from the uterus at abortion had high copper levels. The tissue levels of copper were quite variable. It is clear from the control fetus (Patient 9 who did not have a contraceptive device) that tissue copper levels at the time of abortion can be quite variable, particularly the levels found in the fetal liver. Brain copper levels at gestations between 10 and 17 weeks seem comparatively low.Whether a pregnancy with a copper 7 in situ is likely to end in miscarriage is most probably influenced by high copper levels earlier in pregnancy.

We should like to thank colleagues who helped during the course of this work: all the members of the Department of Obstetrics and Gynaecology, Eastern General Hospital, Edinburgh, particularly Dr J. Loudon, Dr W. Paterson, Dr . Grant, Dr J. Davies and Dr J. Drife; Dr . Loudon and all the staffai the Family Planning Centre, Dean Terrace, Edinburgh; Mr J. Newton, Dr J. McEwan, Mrs J. Deakin and the staff of Kings College Hospital W.H.O. Clinic; Dr J. Grant, Dr A. Shepherd and staff at the Simpson Memorial Maternity Pavilion; Miss A. Fotheringham; Miss C. Gray, Mr A. R. Ross, Mrs P. Eason and Miss C. Bowden for their skilled technical assistance; Mrs E. Smith for secretarial assistance; and Professor H. J. Evans for his help and encouragement during the course of this work.

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access References

Alderman, . (1976) Cord blood copper concentrations Oster, G. & Salgo, M.P. (1977) Copper in mammalian in the presence of a copper releasing intrauterine reproduction. Adv Pharmac. Chemother. 14, 327- contraceptive device. Proc. R. Soc. Med. 69, 823- 409. 825. Regan, J.G. & Warren, J. (1976) A novel approach to Barrie, H. (1976) Congenital malformation associated the elimination of matrix interferencies in nameless with intrauterine contraceptive device. Br. med. J. 1, atomic absorption spectroscopy using a graphite 488-490. furnace. Analyst 101, 220-226. Chang, C.C. & Tatum, HJ. (1970) A study of the Ross, ., Gosden, CM., Loudon, N.B. & Foxwell, M. antifertility effect of intrauterine copper. (1978) The copper-bearing IUD: biological variation Contraception 1, 265-270. in the erosion and deposition of elements and its Gosden, CM., Ross, A. & Loudon, N.B. (1977) contraceptive importance. Br. J. Family Planning 4, Intrauterine deposition of calcium on copper-bearing 1-9. intrauterine contraceptive devices. Br. med. J. 1, Sperling, K.R. (1976) Determination of heavy metals in 202-206. seawater and marine organisms by flameless atomic Gosden, CM., Wright, M.O., Paterson, W.G. & Grant, absorption spectrophotometry III. An improved "gas K.A. (1976) Clinical details, cytogenetics studies and stop". Z. analyt. Chem. 279, 205-208. cellular physiology of a 69XXX fetus with com¬ Tatum, HJ. (1974) Copper-bearing intrauterine devices. ments on the biological effect of triploidy in man. J. Clin. Obstet. Gynaec. 117,93-119. med. Genet. 13,371-380. Tatum, HJ., Schmidt, F.H. & Jain, A.K. (1976) Hagenfeldt, K. (1972) Intrauterine contraception with Management and outcomes of pregnancies the Copper-T device. Contraception 6, 37-53. associated with the copper intrauterine contracep¬ Newton, J.R., IHingworth, R., Ellas, J. & McEwan, J. tive device. Am. J. Obstet. Gynec. 126,869-879. (1977) Continuous intrauterine copper contraception for 3 years; comparison of replacement at 2 years Received 4 August 1978 with continuation of use. Br. med. J. 1, 197-199.

Downloaded from Bioscientifica.com at 10/07/2021 09:25:11AM via free access