Effect of tarnish on copper release

CONTRACEPTION

EFFECT OF TARNISH ON COPPER RELEASE Y.Y. Tsong and Harold A. Nash Center for Biomedical Research, The Population Council 1230 York Avenue, New York, NY 10021 ABSTRACT The rate of copper loss from bright and tarnished collars from Copper T Model TCu 380A IIJDs has been investieated in amino acid solutions of DH 5.5 and 7.4 and in serum. In all three media, the tar&shed collars quickly became bhght and lost copper at the same rate as the initially bright collars. The single exception was when a high ratio of copper surface to serum was used. Under those conditions the tarnished collars initially became bright but after two days a black precipitate appeared on both the initially bright and tarnished collars and weight loss ceased. When a higher ratio of serum to copper surface was used, the pattern was one of continuing loss although at a lower rate than in the amino acid solutions. It is concluded that tarnish does not compromise the oxidation and dissolution of copper even in serum. Serum is considered a surrogate for uterine fluid. INTRODUCTION Copper-carrying intrauterine devices (IUDs) are widely used as long-term contraceptives in women. Gradual release of copper from their surface is essential to their efficacy as is shown by the inverse correlation between copper surface and pregnancy rates (1,2). A large number of investigators have documented copper loss during use (37). The pattern is one of decreasing rates with time and asymmetric loss, with some areas of the copper surface being much more deeply pitted than others. The copper loss, to the extent it results in copper in solution, must involve oxidation to either the cuprous or cupric ion. In addition to copper loss through oxidation and dissolution, observation of recovered devices suggests occasional detachment of discrete particles of copper metal. It is believed that these fragments, unlike copper ions in solution, do not contribute directly to contraceptive effectiveness. It is assumed that complexation with components of uterine secretions serve to solubilize the copper contained in the oxides or sulfides that may be formed by the oxidation process. Candidate solubilizing agents include amino acids, proteins, and peptides with free amino or sulfhydryl groups, and ascorbic acid. Since the copper in IUDs sometimes becomes darkened in color before insertion, question has arisen as to whether such preinsertion tarnishing affects copper availability. To address this question, we have measured copper release from bright and tarnished copper in amino acid media and in serum Submitted for publication June 25, 1991 Accepted for publication July 25, 1991

OCTOBER 1991 VOL. 44 NO. 4

CONTRACEPTION

MATERIALS AND METHODS Copper Devices: Copper collars recovered from Copper T Model TCu 380A devices were used as experimental materials. Collars used in the study included:a) bright collars without visible tarnish;;;; collars which had become darkened spontaneously;c) collars of dark color recovered fmm subjects after 25 months of used) collars darkened by exposure to hydrogen sulfide;and e) collars darkened by exposure to ammonia. Collars of types a, b, and c were left attached to the polyethylene support of the IUD. Each collar presented 38 mm2 of exposed surface. Collars of types d and e were examined unattached to a support with unattached collars of type a included as controls. Each collar presented 63 mm2 of surface. Table I. Composition of Amino Acid Solution Amino Acid Histidine Isoleucine Leucine Lysine Methionine Phenylalanine Threonine Tiyptophan Valine

Amount (g/liter H,O) 5.5 7.0 11.0 8.0 11.0 11.0 5.0 2.5 8.0

Incubation Conditions: Collars were incubated in an amino acid solution of the composition shown in Table I adjusted to either pH 5.5 or 7.5 with 6 N NaOH. In the instance of units attached to a polyethylene support, five collars of each type were incubated in a stoppered 50 ml serum bottle containing 25 ml of the amino acid solution. There were five such serum bottles for each collar type (25 collars in total). In the instance of collars incubated free of support, three collars were placed in each of five serum bottles. All incubations were performed in a 37°C water bath with shaking at 90 strokes of 3.7 cm amplitude per minute. Amino acids were purchased from Sigma Chemical Co.,&. Louis,Mo. In separate experiments, collars were incubated in serum. Serum was considered a surrogate for uterine fluid. Comparison of aspects of their composition are set forth in Table II (8). Incubation with serum was carried out in two modes. In one, five collars supported on polyethylene were placed in each of three serum bottles containing 20 ml of human serum with 0.01% benzalkonium chloride added as preservative. The bottles were incubated in a 37°C shaking water bath. In the second mode, each of five darkened collars was placed in 20 ml of calf serum containing 0.01% benzalkonium chloride. Measurement of Copper Release: The amount of copper lost was determined gravimetrically. The incubation medium was changed daily and the collars prepared for weighing with five washes of 30 ml of distilled water and three washes of acetone, followed by drying at 80°C for 10 minutes. The five collars placed in serum were pooled together each day for weighing.

OCTOBER 1991 VOL. 44 NO. 4

CONTRACEPTION

Table II. Comparison of the Composition of Midcycle Uterine Fluid to Serum’ Midcycle Uterine Fluid 284 20.1 119 0.77 111 34

Osmolarity @mole/I) K+ (mmole/l) Na+ @mole/l) Ca++ (mmole/l) Cl- (mmole/l) Albumin (g/l) Glucose (mmole/I) Fructose (mmole/l) Urea (mmole/l)

Serum 288 4.3 142 2.55 111 46 5.1

;: 8:3

7.0

(a) Adapted from Casslen B and Nilsson, B. (8). RESULTS Copper Release in Amino Acid Solution at pH 55: The cumulative release from bright and tarnished collars over a 2O-day period is plotted in Fig. 1. The tarnished collars became bright within the first day and the release rates were essentially identical over the 20-day period. The mean release per day per 100 mm2 of copper surface was 0.62 mg. The standard deviation between sets of five collars is seen to be extremely small. The experiment was repeated with essentially identical findings. The release per 100 mm2 of copper surface in the second experiment was 0.61 mglday. Copper release from collars darkened by exposure to hydrogen sulfide or ammonia as compared with release from bright collars is shown in Fig. 2. The tarnished collars quickly became bright and release from all three groups was identical through the 20 days of observation. The mean release rate was 0.86 mg per 100 mm2 of copper surface.

10

15

20

25

INCUBATION PERIOD (day)

Fig. 1. Mean cumulative copper release from sets of 5 bright copper collars (0) and from sets of 5 tarnished copper collars (0). Copper release is expressed as mg + SD per set for 5 sets of each type.

OCTOBER 1991 VOL. 44 NO. 4

387

CONTRACEPTION

0

5

10

15

20

25

INCUBATION PERIOD (day)

Fig. 2. Mean cumulative copper release from sets of 3 bright copper collars (a), sets of 3 copper collars darkened by H,S exposure (0), and sets of 3 copper collars darkened by NH, exposure (8). Copper release in mg zhSD for 5 sets of each type.

Release from devices recovered from in utero exposure was 0.80 mg/lOO mm’ per day. The collars became bright within a few hours and lost 2.9 mg per 100 mm’ in the first day. Copper Release in Amino Acid Solution at pH 75: To determine if the low pH of the amino acid solutions used in the first experiments played a large role in the observed rate of copper loss, release was measured in amino acid solutions adjusted to pH 7.5. Mean loss at this higher pH was 0.58 mg/lOO mm’ of copper surface for both tarnished and non-tarnished devices. Copper Loss in Serum: When five collars were placed in 20 ml of serum w.ith daily changes of the serum, the rate of coppr loss for both bnght and tarnished devrces was approximately 0.2 mg per 100 mm of surface per day durmg the first two days. Thereafter, the collars of both initial conditions became black and gained weight. When only a single tarnished collar was placed in each bottle of 20 ml of serum, the devices became bright within a day and continued to remain bright through at least ten days with daily changes of the serum. Although bright, they did not have the high sheen of devices that had never been darkened. Microscopic examination showed this was due to micropitting--assumedly where tarnish had been removed. The release pattern is represented in Fig. 3. The rate of copper loss was essentially identical for initially tarnished and initially bright collars although the rate for both slowed appreciably over time. The mean release rate was approximately 0.22 mg per 100 mm2 of surface during the first four days and 0.04 mg per 100 mm2 thereafter,

OCTOBER 1991 VOL. 44 NO. 4

CONTRACEPTION

0

1

2

3

4

Incubation

5

6

7

Period

8

91011

(day)

Fig 3. Cumulative copper loss from 5 bright @)or 5 tarnished 0) collars in serum. DISCUSSION The role of oxidation in discoloration of copper has been studied by several investigators and has been summarized by Evans (9). One finding is that surface oxidation can produce a variety of colors ranging through brown, purple, blue, green, yellow, and red depending on the thickness of the films. It is assumed by the present investigators that other agents that promote the development of colored films on copper do so either by accelerating the oxidation process (NH,, SO& or by converting oxides to less soluble compounds (H2S). A basic certainty in considering copper erosion in utero is that it does not go into solution as metallic copper. Oxidation to an ionic form is required, although not necessarily by molecular oxygen. The oxygen concentrations in the uterus have been calculated to be more than sufficient to oxidize copper to the divalent state (10). Lewis et al. (10) applied chemical, photoacoustic, and atomic absorption techniques to darkened corrosion layers on copper IUDs recovered from patients. Photoacoustic spectra were consistent with either CuS or Cu,S being major components of the corrosion layer, but this could not be confirmed by chemical methods. Data from atomic absorption suggested the corrosion layer was largely oxides. The photoacoustic spectra did not match simple oxides and led the investigators to conclude that the “dark color and almost structureless spectra are the result of electron transfer spectra involving copper I, copper II and even a little iron III superimposed on the weaker d-d-transitions of simpler copper Il species.” A small fraction of the layer was judged consistent with copper II amino acid, peptide, or protein complexes. Charmer et al. (11) noted the presence of a predominant sulfur peak in dispersive X-ray analysis of the surfaces of copper devices recovered from patients. Results of the present study confirm the ability of complexing agents such as amino acids or proteins to quickly remove the tarnish layers that are sometimes found on unused devices. This was true at physiological pH and was more rapid at higher acidity. The subsequent pattern of copper dissolution in amino acid solutions was identical for copper surfaces initially tarnished and surfaces initially without visible tarnish. The dissolution

OCTOBER 1991 VOL. 44 NO. 4

389

rate in amino acid solutions was higher than in serum (Table III) and is believed to be a function of a lower concentration of complexing sites in proteins than in the amino acid solutions. When a relatively high concentration of copper surface was present in serum, weight gains eventually occurred coincident with darkening. Similar darkening and decreased release rate over time was observed by Chantler et al. (12) using aqueous fetal calf serum or fetal calf serum plus tissue culture medium [undefined] as the medium. They identified the deposit as a protein layer and showed the copper in solution to be reversibly complexed with albumin. The decreasing release rate in utero may be partially a function of such protein deposits and partially a function of the calcareous deposits that have repeatedly been demonstrated on recovered devices (7,13-16). With the use of a higher ratio of serum to copper surface, the copper loss continued through the period examined (9 days). Table III. Mean Release per Day of Copper from Copper Collars under Various Conditions Amino acid solution pH 5.5:

mg/lOO mmL of copper surface per day

Bright collars Tarnished collars

0.62 0.62

Bright collars H,S darkened collars NHs darkened collars.

0.86 0.86 0.86

Collars tarnished in utero

0.80

0.61 0.61

Amino acid solution pH 7.5: Bright collars Tarnished collars

Tarnished collars

0.58 0.58

0.22 - 0.04

The copper loss under the conditions used in this investigation is summarized in Table III. Bright and tarnished collars released about 0.62 mg/lOOmrr? surface per day when supported on polyethylene but 0.86 mg/lOOmn? per day without a polyethylene support. The reason for the higher release rate from the unsupported collars is believed to lie in the fact that the inner surface of the collars, which is exposed in the unsupported collars, is not polished and has microscopically visible grooves that undoubtedly increase the effective surface. The loss in 20 ml of serum per collar is not markedly different than has been found by analysis of devices after periods of use in women. Moo-Young et al (3) reported rates for copper wire wound devices which converted to loss per 100mm2 of surface ranged from 0.025 to 0.15 mg/day during the first 80 days of use. The highest values reported by Goh andTongfor the Copper T Model 220C work out to about 0.06 mg/day per 100 mm’ of surface (4). The moderately higher rate in our in vitro experiments with serum is perhaps not surprising considering the volumes of serum used as compared with the probable volumes of uterine fluid to which devices are exposed per day in utero. The overall interpretation of the findings is that darkening of copper surface does not compromise copper oxidation and dissolution, even when serum is used as the bathing solution. Serum is considered a surrogate for uterine fluids.

OCTOBER 1991 VOL. 44 NO. 4

CONTRACEPTION

Acknowledgements This research was conducted as part of the contraceptive development program of the Population Council’s International Committee for Contraception Research. The assistance of Mr. Richard Sewell and Ms. Young-Hea Chun in conduct of the study is gratefully acknowledged. REFERENCES 1. Tatum, HJ. Intrauterine 1023.

contraception.

Amer J Obstet Gynecol

1972; 112: lOoo-

2. Sivin, I and Stem, J. Long-acting, more effective copper RIDS: A summary of US experience 1970-75. Stud Fam Plann 1979; 10: 263-281. 3. Moo-Young, AJ, Tatum, HJ, Wan, LS, and Lane, M. Copper levels in tissues of rhesus monkeys and of women bearing intrauterine copper devices. In: F. Hefnawi and S.J. Segal, eds.Analysis of Intrauterine Contraception. New York: American Elsevier Publishing Co., Inc., 1975: 439-457. 4. Goh, TH and Tong, SL.Pattems of copper loss - A comparison of the Multiload Cu250, TCu 220C and Cu7 IUDs.. Contraception 1986; 33: 411-420. 5. Drost, RH, Thiery, M, and Maes, RAALong-term release of copper from two Multiload IUD models: ML Cu250 and ML Cu 375. Adv Contraception 1987; 3: 315-318. 6. Timonen, H.Copper release from Copper-T intrauterine devices. Contraception 14: 25-38.

1976;

7. Thiery, M, Schmidt, F, and Tatum, HJ.Copper loss from the copper T Model TCu 220C. Contraception 1982; 26: 295-302. 8. Casslen, B and Nilsson, B. Human uterine fluid examined in undiluted samples for osmolarity and the concentrations of inorganic ions, albumin, glucose and urea. Amer J Obstet Gynecol 1984; 150: 877-881. 9. Evans, UR. The corrosion and oxidation of metals: scientific principles and practical applications. London: Edward Arnold, Ltd., 1960: 52-53. 10. Lewis, KM, Archer, RD, Ginsberg, AP, and Rosencwaig, A.The corrosion chemistry of copper intrauterine devices. Contraception 1977; 15: 93-104. 11. Chantler, EN, Scott, K, Filho, CI, Elstein, M, Faraher, EB, Lorimer, GW, and Brough, I. Degradation of the copper-releasing intrauterine contraceptive device and its 1984; 91: 172-181. significance. Brit J Obstet Gynecol 12. Charmer, E, Critoph, F, and Elstein, M.Release of copper from copper-bearing terine contraceptive devices. Brit Med J 1977; 2: 288-291.

intrau-

13. Johnson, AB, Maness, RF, and Wheeler, RG. Calcareous deposits formed on IUDs in human exposures. Contraception 1976; 14: 507-517.

OCTOBER 1991 VOL. 44 NO. 4

391

CONTRACEPTION

14. Gosden, C, Ross, A, and Loudon, NB.Intrauterine deposition of calcium on copperbearing intrauterine contraceptive devices. Brit Med J 1977; 1: 202-206. 15. Rosenfeld, J, Williams, WS, and Sharma, NK. SEM analysis of copper containing intrauterine implants. J Biomed Materials Res. 1981; 15: 605-610. 16. Thiery, M and Kosonen, A.The multisleeved Copper T Model T Cu22OC: Effect of long-term use on corrosion and dissolution of copper. Contraception 1987; 35: 163170.

392

OCTOBER 1991 VOL. 44 NO. 4