lnt
J Gynaecol Obstet
16: 482- 487, 1979
Intrauterine Devices: Medicated and Nonmedicated Daniel R. Mishell, Jr Department of Obstetrics and Gynaecolog)l, School of Medicine, Un iversity of Southern California, Los Angeles, California, USA
ABSTRACT Mishell DR, Jr (Dept of Obstetrics and Gynaecology, School of Medicine, University of Southern California, Los Angeles, CA, USA). Intrauterine devices: medicated and nonmedicated. lnt J Gynaecol Obstet 16: 482- 48 7, / 9 79 Th e main benefits of intrauterine devices (I UDs) are a lack of adverse systemic effects, excellent effectiveness, high continuation rates and the single act of motivation required for use. First year fa ilure rates range from 2% to 3%, but decline steadily thereafter to a cumulative annual failure rate of less than / % after six years. Th e risks of IUDs include increased blood loss, uterine perforation, pelvic infection and pregnancy-related complications. Tlze incidence ofperforation of the uterine fundus ranges from 1:1000 to 1:2500 insertions, while that of cervical perforation with the copper devices ranges from 1:600 to 1:1000. IUD use is associated with about a three-fold increased incidence of developing acute salpingitis in comparison with use oforal contraceptives and diaphragms. If pregnancy occurs with an IUD in place, there is a three-fold increased risk of spontaneous abortion, a ten-fold increased risk of ectopic pregnancy ( 5% of all I UD pregnancies) and a possible increased incidence of sepsis during tlze pregnancy.
BENEFITS The main benefits of IUDs are the lack of associated systemic metabolic effects combined with excellent effectiveness. These two characteristics accoun t for a very high continuation rate. In contrast to other types of contraception, there is no need for the frequent motivation needed to ingest a pill daily or to use consistently a coitus-related method. The method effectiveness and use effectiveness rates for IUDs are similar. Although first year failure rates are generally reported to range from 2% to 3%, the Portions of this presentation will appear in Reproductive Endocrinology, Infertility and Contraception , F. A. Davis Co, Philadelphia, PA, USA, in press.
lnt J Gynaecol Obstet 16
annual incidence of accidental pregnancy decreases steadily after the first year. After six years, the cumulative annual failure rate is less than l% for the loop. The incidence of adverse events with IUDs, such as pregnancy, expulsion or removal for bleeding and/or pain, steadily decreases with increasing age (Table I) . Thus, the IUD is especially suited for older parous women who wish to delay or prevent furth er pregnancies.
RISKS Blood loss Risks associated with IUDs include excessive blood loss, uterine and cervical perforation , pelvic infection and pregnancy-related complications. The amount of blood loss in each menstrual cycle is significantly increased in women wearing an IUD compared with tha t of controls. In a normal cycle, about 35 ml of blood is lost. In women wearing a coil or a loop, about 70 ml to 80 ml of blood is lost ; in those wearing a Cooper-T, 50 ml to 60 ml is lost per cycle. Leidholm and associates (7) reported that the mean increase of menstrual blood loss after insertion of a Copper- T was 26 ml, an increase of 84% of the control cycles. The amount of the increase was not significantly different when measured I month, 6 months or l year after insertion. In the Swedish women studied, there was no significant change in mean values of hemoglobin concentration, serum iron and total iron-binding capaCity determined six and 12 months after IUD insertion , as compared to mean values measured prior to insertion . In a study from England, Morehead and colleagues (I 0) found that IUD use caused a mean hemoglobin decrease of about 0.5 gm% in the first six months of use. This reduction was not large enough to produce anemia in these healthy, wellnourished women. The IUD rarely causes anemia when there are adequate amounts of iron in the diet. If women develop menorrhagia whi le using an IUD, their hemoglobin levels should be measured;
Medicated and nonmedicated /UDs
Table I. Two-year net cumulative event rates with the Lippes Loop 0 per 100 women 8 • Age at Insertion Event
25-29 15-24 30-34 35-49 5 .8 4.7 Pregnancy 2.8 1.5 17.4 9 .8 7.1 Expulsion 5.4 Removal (bleeding/ 18.0 pain) 17.7 16.8 16.2 Continuation 58.0 rate 66 .7 72.0 75.4 First insertion 2753 2082 1397 1187 Woman-months of use 41 758 34 574 23 874 19 912 a Adapted from Tietze and Lewit (16).
if the levels are low, the IUDs should be removed and iron supplementation should be given. IUD users who develop an increased frequency or an excessive amount of vaginal bleeding usually do ~ot benefit from treatment with ancillary therapeutic formul ations. Antifibrinolytic agents, particularly tranexamic acid, have ameliorated the bleeding problems, but these potent agents may be associated with adverse systemic effects. Excessive bleeding in . the first few months after IUD insertion should be treated with reassurance and supplemental oral iron. The bleeding frequently diminishes with time as the uterus adjusts to the presence of the foreign body. Excessive bleeding that continues or develops several months or more after IUD insertion is best treated by removal of the device. After a one-month interval, another device may be inserted if the patient sti ll wishes to use an IUD for contraception. If the original device was plastic, the new device shou ld be a copper or progesteronereleasing IUD as these are associated with less blood loss than the larger plastic devices.
Perforation One of the potentially serious, a lthough uncommon, complications associated with use of an IUD is perforation of the uterine fundus. Perforation initi a ll y occurs at insertion and can best be prevented by straightening the uterine axis with a tenacu lum an d then probing the cavity with a uterine sound before IUD insertion. Sometimes, only the distal portion of the IUD penetrates the uterine muscle at insertion; then, the uterine contractions over the next few months force the IUD into the peritoneal cavity. IUDs correctly inserted entirely within the endometrial cavity do not wander through the uterine muscle into the endometrial cavity. The incidence of perforation is generally related to the shape of the IUD and/or the amount
483
of force used during its insertion. The perforation rate with the Dalkon Shield has been reported to be about 1:350 insertions (8), while perforation rates with the Lippes Loop have been reported to be about 1: 1000 insertions ( 16). The perforation rates for the Copper-7 and the Copper-T in large multiclinic studies are in the same range as those for the Loop (8). One shou ld a lways suspect that perforation has occurred if a patient states that she cannot feel the appendage, but did not notice that the device was expelled. The physician shou ld not assume that an unnoticed expu lsion has occurred. Frequently, the device has rotated 180 degrees and the appendage is withdrawn into the cavity. In this situation, the uterine cavity should be probed after pelvic examination is performed and the possibility of pregnancy excluded. If the device cannot be felt with a uterine sound or biopsy instrument , an x-ray film or sonography should be ordered. It is best to obtain x-ray films with both anteroposterior and lateral views w ith contrast media or a uterine sound inside the uterine cavity. The IUD may be located in the cu lde-sac and the diagnosis may be missed with only an anteroposterior film. If any IUD is found to be outside the uterus, it should be electively removed as complications such as adhesions and bowel obstruction have been reported. Both the copper IUDs and shields have been found to produce severe peritoneal reactions. It is best to remove these devices as soon as possible after the diagnosis of perforation is made. In the absence of severe adhesions, most intrauterine IUDs can be removed by means of laparoscopy, avoiding the necessity of laparotomy. Perforation of the cervix has a lso been reported to occur with devices having a straight vertical arm, such as the Copper-Tor Copper-7. The incidence of downward perforation of these devices into the cervix has been reported to range from about 1:600 to 1: 1000 insertions. When follow-up examinations are performed on patients wit h these devices, the cervix should be carefully inspected and palpated, as frequently the perforation does not extend completely through the ectocervical epithelium. Cervical perforation is not a major problem, but devices that have perforated downward should be removed with a uterine packing forceps because downward displacement is associated with a reduced contraceptive effectiveness.
Congenital anomalies When pregnancy occurs with an IUD in place, implantation takes place away from the device itself,
lntj Gynaecol Obstet / 6
484
Mz shell
so that device is always extra-ammot1c. To date, there is no evidence of an increased incidence of congenital anomalies in infants born with an IUD in utero. Tatum and associates (14) recently reported that of 166 conceptions occurring with an intrauterine Copper-T in place which progressed to a size which permitted adequate examination for anomalies, only one infant had a congenital anomaly (fibroma of the vocal cords). Although these numbers are small , there is no evidence in this series to indicate that the presence of copper in the uterus exerts a deleterious effect on fetal development.
Spontaneous abortion In all st udies of pregnancies with any type of IUD in situ, the incidence of fetal death was not significan tly increased ; however, a significant increase in the incidence of spontaneous abortion has been consistently observed. If a patient conceives with an IUD in place and the IUD is not removed, the incidence of spontaneous abortion is about 55%, approximately three times greater than would occur if the patient conceived without an IUD (6). If, after conception, the IUD is spontaneously expelled or if the appendage is visible and the IUD is removed by traction, the incidence of spontaneous abortion is significantly reduced. Thus, if a woman conceives with an IUD in situ and she wishes to continue the pregnancy, the IUD should be removed if the appendage is visible in order to reduce the chance of spontaneous abortion.
Septic abortion If the IUD cannot be easily removed or if the appendage is not visible, evidence suggests that the risk of septic abortion may be increased if the IUD remains in place. Most of the evidence indicating an increased risk of sepsis is based upon data fromwomen who conceived with a shield type of IUD in the uterus. This device, with its multifilament tail, was widely used throughout the world from 1971 to 1974. A retrospective study from the US Center for Disease Control (1) indicated that the risk of death from spontaneous abortion was 14.8 per 100 000 women per year with an IUD in place compared with a mortality rate of 0.28 per 100 000 pregnant women without an IUD, or a 52.8-fold increased risk of death with the IUD. Although there is theoretical and actual evidence that there is an increased risk of septic abortion if a patient conceives with a shield IUD in place, there is no conclusive evidence that a patient who conceives with a device other than the shield in the uterus has an increased risk of
lntj Gynaeco/ Obslel / 6
septic abortion. Williams and associates ( 18) reported that there was no significant difference in the incidence of septic abortion in women who conceived with an IUD in place compared with the incidence in women who conceived while using other methods. In a study by Tatum and associates of 918 women conceiving with the Copper-Tin situ (14), there were only two cases of septic abortion, both occurring in the first trimester. This evidence does not suggest that there is an increase in sepsis in pregnancy due to the presence of the Copper-T. A patient who conceives with a shield type IUD in utero which cannot be removed without entering the uterine cavity should be fully informed about the increased risk of sepsis if she wishes to continue her pregna ncy. The increased incidence of death due to sepsis in pregnancy reported with this device makes it a dvisable to remove the device and terminate the pregnancy. There is no conclusive evidence that the incidence of sepsis is increased with other types of IUDs, but the patient should be informed of the possibility of an increased incidence of sepsis and of the need to report symptoms of infection promptly if she wishes to continue the pregnancy. If intra uterine infection does occur with an IUD in the pregnant uterus, the endometrial cavity should be evacuated after a short interval of appropriate antibiotic treatment which is similar to the treatment of uterine sepsis without an IUD in place.
Ectopic pregnancy The IUD prevents intrauterine pregnancy more effectively than it prevents ectopic pregnancy. Lehfeldt and associates {5) estimated that, if the 300 000 women enrolled in the Population Council IUD study did not ha ve their rate of fertilization changed by use of the IUD, then during the 45 000 womanyears of use, about 180 000 fertilized ova and 900 ectopic pregnancies would be expected. Of the ectopic pregnancies, 845 would be tubal and approximately five ovarian. Actually, during this time there were 1046 pregnancies with the device in situ, 45 of which were ectopic, 40 tubal and 5 ovarian. From analysis of this data, it was concluded that use of the IUD reduces uterine implantation by 99.5%, tubal implantation by about 95% and ovarian pregnancy not at all. Thus, if pregnancy occurs with an IUD in place, there is a greater chance of it being ectopic than if no IUD were used. If a patient conceives with an IUD in place, her chances of having an ectopic pregnancy range from 3% to 9%. This incidence is about ten times greater than the reported ectopic pregnancy frequency of 0.3% to 0. 7% of total births in similar populations. If a
Medicated and nonmedicated IUDs
patient conceives with an IUD in place, it should be suspected that she has an ectopic pregnancy. Prematurity Tatum and associates (14) studied conceptions with a Copper-T IUD in the uterus and found that the rate of prematurity among live births was four times greater when the Copper-T was left in place than when it was removed. Therefore, if it is not possible to remove the IUD and the patient wishes to continue her gestation, she should be warned of the possible increased risk of prematurity in addition to the increased risk of spontaneous abortion. She should also be informed about an increased risk of ectopic pregnancy and possibly septic abortion, and be reminded to report the first signs of pelvic pain or fever. Infection In the 1960s, despite great concern among gynecologists that use of the IUD would markedly increase the incidence of salpingitis, there was little evidence that such an increase did occur. In 1966, we performed aerobic and anaerobic cultures of endom'e trial homogenates obtained transfundally at varying intervals after insertion of a Lippes Loop (9). In the first 24 hours following insertion, the normally sterile endometrial cavity was consistently infected with bacteria. Nevertheless, the natural defenses destroyed these bacteria within 24 hours in 80% of the cases. In our study, the endometrial cavity, the IUD and the portion of the thread within the cavity were found to be consistently sterile when transfundal cultures were obtained more than 30 days after insertion. These findings indicated that development of salpingitis more than one month after insertion of IUDs other than the shield device is of venereal origin and unrelated to the presence of the IUD. These findings agree with the incidence of clinical diagnosis of pelvic inflammatory disease (PID) found in a group of 23 977 mainly parous women analyzed by Tietze and Lewit (16) in 1970. In this group, there were 239 cases of PID in which the device was removed and 43 7 in which the device remained in situ during and following treatment. When PID rates were computed according to duration of time that the IUD was in place, the rates were higher in the first two weeks after insertion and then steadily diminished. Rates after the first month were on the order of 1.0 to 2.5 per 100 woman-years. The results of both of these studies provide evidence that one should not insert an IUD into a patient who may have been recently infected with a gonococcus, as insertion of the device will transport these
485
pathogens into the upper genital tract. If there is clinical suspicion that a gonococcal endocervicitis exists, cultures should be obtained and the IUD not inserted until negative results are obtained. Following introduction and widespread use of the shield device among nulliparous women in whom IUDs were previously inserted only occasionally, there appeared several studies which suggested that use of the IUD increased the relative risk of developing salpingitis or PID. Tatum and associates (15) carefully examined the sheaths of the appendage of both new shields in their sterile packages and those which were removed from patients. They found that 34% of the former and 9 % of the latter had breaks in the sheath around the knot which attached it to the device. These breaks would allow bacteria continuous access from the vagina into the endometrial cavity and thus increase the risk of upper genital tract infection. For this reason, all of the shield-type IUDs currently in place should be removed. Faulkner and Ory (2) performed a prospective case control study of patients attending an emergency room in Atlanta. They demonstrated a relative risk of 5.1 for IUD-related febrile PID versus non-IUD related febrile PID. In the prospective British Family Planning Study (12) based on inpatient illness, the incidence of salpingitis was about three times greater in IUD users than in diaphragm or oral contraceptive users. In a prospective casecontrol study from Sweden performed by Westrom and associates (17), the diagnosis of salpingitis was made by laparoscopy while the controls were chosen by questionnaire. Analysis of the data by Ory (11) showed that the relative risk for the development of salpingitis "in IUD users versus nonusers was 2.9 for all women, 6.9 for nulliparous women, but only 1. 7 for parous women (Table II) . A recent retrospective case-control study conducted in Seattle by Eschenbach and co-workers found a relative risk for PID among IUD users versus nonusers to be 4.4 for all women, 2.8 for those having gonococcal PID and 6.5 for those having nongonococcal PID. This study also found an increased risk of PID in nulliparous versus parous IUD users. The increased incidence of salpingitis in nulliparous IUD users is possibly related to the greater number of sexual partners in this group of women. There are several problems with these studies. One problem is that uniform guidelines for the diagnosis of salpingitis or PID are not used. Changes in the criteria used to establish the diagnosis of salpingitis may increase the frequency of the diagnosis among IUD wearers. Patients with lower abdominal pain who have only minimal or no temperature evaluation may have the diagnosis of salIn! J Gynaecol Obslet /6
486
Mishell
Table II. Relative risks of PID for IUD users compared to nonusers by age and parity8 • Age (years) Age Standardized
Parity
Nulliparous Multiparous Parity standardizedb 8
b
c
J Gynaecol Obstet
16: 482- 487, 1979
Intrauterine Devices: Medicated and Nonmedicated Daniel R. Mishell, Jr Department of Obstetrics and Gynaecolog)l, School of Medicine, Un iversity of Southern California, Los Angeles, California, USA
ABSTRACT Mishell DR, Jr (Dept of Obstetrics and Gynaecology, School of Medicine, University of Southern California, Los Angeles, CA, USA). Intrauterine devices: medicated and nonmedicated. lnt J Gynaecol Obstet 16: 482- 48 7, / 9 79 Th e main benefits of intrauterine devices (I UDs) are a lack of adverse systemic effects, excellent effectiveness, high continuation rates and the single act of motivation required for use. First year fa ilure rates range from 2% to 3%, but decline steadily thereafter to a cumulative annual failure rate of less than / % after six years. Th e risks of IUDs include increased blood loss, uterine perforation, pelvic infection and pregnancy-related complications. Tlze incidence ofperforation of the uterine fundus ranges from 1:1000 to 1:2500 insertions, while that of cervical perforation with the copper devices ranges from 1:600 to 1:1000. IUD use is associated with about a three-fold increased incidence of developing acute salpingitis in comparison with use oforal contraceptives and diaphragms. If pregnancy occurs with an IUD in place, there is a three-fold increased risk of spontaneous abortion, a ten-fold increased risk of ectopic pregnancy ( 5% of all I UD pregnancies) and a possible increased incidence of sepsis during tlze pregnancy.
BENEFITS The main benefits of IUDs are the lack of associated systemic metabolic effects combined with excellent effectiveness. These two characteristics accoun t for a very high continuation rate. In contrast to other types of contraception, there is no need for the frequent motivation needed to ingest a pill daily or to use consistently a coitus-related method. The method effectiveness and use effectiveness rates for IUDs are similar. Although first year failure rates are generally reported to range from 2% to 3%, the Portions of this presentation will appear in Reproductive Endocrinology, Infertility and Contraception , F. A. Davis Co, Philadelphia, PA, USA, in press.
lnt J Gynaecol Obstet 16
annual incidence of accidental pregnancy decreases steadily after the first year. After six years, the cumulative annual failure rate is less than l% for the loop. The incidence of adverse events with IUDs, such as pregnancy, expulsion or removal for bleeding and/or pain, steadily decreases with increasing age (Table I) . Thus, the IUD is especially suited for older parous women who wish to delay or prevent furth er pregnancies.
RISKS Blood loss Risks associated with IUDs include excessive blood loss, uterine and cervical perforation , pelvic infection and pregnancy-related complications. The amount of blood loss in each menstrual cycle is significantly increased in women wearing an IUD compared with tha t of controls. In a normal cycle, about 35 ml of blood is lost. In women wearing a coil or a loop, about 70 ml to 80 ml of blood is lost ; in those wearing a Cooper-T, 50 ml to 60 ml is lost per cycle. Leidholm and associates (7) reported that the mean increase of menstrual blood loss after insertion of a Copper- T was 26 ml, an increase of 84% of the control cycles. The amount of the increase was not significantly different when measured I month, 6 months or l year after insertion. In the Swedish women studied, there was no significant change in mean values of hemoglobin concentration, serum iron and total iron-binding capaCity determined six and 12 months after IUD insertion , as compared to mean values measured prior to insertion . In a study from England, Morehead and colleagues (I 0) found that IUD use caused a mean hemoglobin decrease of about 0.5 gm% in the first six months of use. This reduction was not large enough to produce anemia in these healthy, wellnourished women. The IUD rarely causes anemia when there are adequate amounts of iron in the diet. If women develop menorrhagia whi le using an IUD, their hemoglobin levels should be measured;
Medicated and nonmedicated /UDs
Table I. Two-year net cumulative event rates with the Lippes Loop 0 per 100 women 8 • Age at Insertion Event
25-29 15-24 30-34 35-49 5 .8 4.7 Pregnancy 2.8 1.5 17.4 9 .8 7.1 Expulsion 5.4 Removal (bleeding/ 18.0 pain) 17.7 16.8 16.2 Continuation 58.0 rate 66 .7 72.0 75.4 First insertion 2753 2082 1397 1187 Woman-months of use 41 758 34 574 23 874 19 912 a Adapted from Tietze and Lewit (16).
if the levels are low, the IUDs should be removed and iron supplementation should be given. IUD users who develop an increased frequency or an excessive amount of vaginal bleeding usually do ~ot benefit from treatment with ancillary therapeutic formul ations. Antifibrinolytic agents, particularly tranexamic acid, have ameliorated the bleeding problems, but these potent agents may be associated with adverse systemic effects. Excessive bleeding in . the first few months after IUD insertion should be treated with reassurance and supplemental oral iron. The bleeding frequently diminishes with time as the uterus adjusts to the presence of the foreign body. Excessive bleeding that continues or develops several months or more after IUD insertion is best treated by removal of the device. After a one-month interval, another device may be inserted if the patient sti ll wishes to use an IUD for contraception. If the original device was plastic, the new device shou ld be a copper or progesteronereleasing IUD as these are associated with less blood loss than the larger plastic devices.
Perforation One of the potentially serious, a lthough uncommon, complications associated with use of an IUD is perforation of the uterine fundus. Perforation initi a ll y occurs at insertion and can best be prevented by straightening the uterine axis with a tenacu lum an d then probing the cavity with a uterine sound before IUD insertion. Sometimes, only the distal portion of the IUD penetrates the uterine muscle at insertion; then, the uterine contractions over the next few months force the IUD into the peritoneal cavity. IUDs correctly inserted entirely within the endometrial cavity do not wander through the uterine muscle into the endometrial cavity. The incidence of perforation is generally related to the shape of the IUD and/or the amount
483
of force used during its insertion. The perforation rate with the Dalkon Shield has been reported to be about 1:350 insertions (8), while perforation rates with the Lippes Loop have been reported to be about 1: 1000 insertions ( 16). The perforation rates for the Copper-7 and the Copper-T in large multiclinic studies are in the same range as those for the Loop (8). One shou ld a lways suspect that perforation has occurred if a patient states that she cannot feel the appendage, but did not notice that the device was expelled. The physician shou ld not assume that an unnoticed expu lsion has occurred. Frequently, the device has rotated 180 degrees and the appendage is withdrawn into the cavity. In this situation, the uterine cavity should be probed after pelvic examination is performed and the possibility of pregnancy excluded. If the device cannot be felt with a uterine sound or biopsy instrument , an x-ray film or sonography should be ordered. It is best to obtain x-ray films with both anteroposterior and lateral views w ith contrast media or a uterine sound inside the uterine cavity. The IUD may be located in the cu lde-sac and the diagnosis may be missed with only an anteroposterior film. If any IUD is found to be outside the uterus, it should be electively removed as complications such as adhesions and bowel obstruction have been reported. Both the copper IUDs and shields have been found to produce severe peritoneal reactions. It is best to remove these devices as soon as possible after the diagnosis of perforation is made. In the absence of severe adhesions, most intrauterine IUDs can be removed by means of laparoscopy, avoiding the necessity of laparotomy. Perforation of the cervix has a lso been reported to occur with devices having a straight vertical arm, such as the Copper-Tor Copper-7. The incidence of downward perforation of these devices into the cervix has been reported to range from about 1:600 to 1: 1000 insertions. When follow-up examinations are performed on patients wit h these devices, the cervix should be carefully inspected and palpated, as frequently the perforation does not extend completely through the ectocervical epithelium. Cervical perforation is not a major problem, but devices that have perforated downward should be removed with a uterine packing forceps because downward displacement is associated with a reduced contraceptive effectiveness.
Congenital anomalies When pregnancy occurs with an IUD in place, implantation takes place away from the device itself,
lntj Gynaecol Obstet / 6
484
Mz shell
so that device is always extra-ammot1c. To date, there is no evidence of an increased incidence of congenital anomalies in infants born with an IUD in utero. Tatum and associates (14) recently reported that of 166 conceptions occurring with an intrauterine Copper-T in place which progressed to a size which permitted adequate examination for anomalies, only one infant had a congenital anomaly (fibroma of the vocal cords). Although these numbers are small , there is no evidence in this series to indicate that the presence of copper in the uterus exerts a deleterious effect on fetal development.
Spontaneous abortion In all st udies of pregnancies with any type of IUD in situ, the incidence of fetal death was not significan tly increased ; however, a significant increase in the incidence of spontaneous abortion has been consistently observed. If a patient conceives with an IUD in place and the IUD is not removed, the incidence of spontaneous abortion is about 55%, approximately three times greater than would occur if the patient conceived without an IUD (6). If, after conception, the IUD is spontaneously expelled or if the appendage is visible and the IUD is removed by traction, the incidence of spontaneous abortion is significantly reduced. Thus, if a woman conceives with an IUD in situ and she wishes to continue the pregnancy, the IUD should be removed if the appendage is visible in order to reduce the chance of spontaneous abortion.
Septic abortion If the IUD cannot be easily removed or if the appendage is not visible, evidence suggests that the risk of septic abortion may be increased if the IUD remains in place. Most of the evidence indicating an increased risk of sepsis is based upon data fromwomen who conceived with a shield type of IUD in the uterus. This device, with its multifilament tail, was widely used throughout the world from 1971 to 1974. A retrospective study from the US Center for Disease Control (1) indicated that the risk of death from spontaneous abortion was 14.8 per 100 000 women per year with an IUD in place compared with a mortality rate of 0.28 per 100 000 pregnant women without an IUD, or a 52.8-fold increased risk of death with the IUD. Although there is theoretical and actual evidence that there is an increased risk of septic abortion if a patient conceives with a shield IUD in place, there is no conclusive evidence that a patient who conceives with a device other than the shield in the uterus has an increased risk of
lntj Gynaeco/ Obslel / 6
septic abortion. Williams and associates ( 18) reported that there was no significant difference in the incidence of septic abortion in women who conceived with an IUD in place compared with the incidence in women who conceived while using other methods. In a study by Tatum and associates of 918 women conceiving with the Copper-Tin situ (14), there were only two cases of septic abortion, both occurring in the first trimester. This evidence does not suggest that there is an increase in sepsis in pregnancy due to the presence of the Copper-T. A patient who conceives with a shield type IUD in utero which cannot be removed without entering the uterine cavity should be fully informed about the increased risk of sepsis if she wishes to continue her pregna ncy. The increased incidence of death due to sepsis in pregnancy reported with this device makes it a dvisable to remove the device and terminate the pregnancy. There is no conclusive evidence that the incidence of sepsis is increased with other types of IUDs, but the patient should be informed of the possibility of an increased incidence of sepsis and of the need to report symptoms of infection promptly if she wishes to continue the pregnancy. If intra uterine infection does occur with an IUD in the pregnant uterus, the endometrial cavity should be evacuated after a short interval of appropriate antibiotic treatment which is similar to the treatment of uterine sepsis without an IUD in place.
Ectopic pregnancy The IUD prevents intrauterine pregnancy more effectively than it prevents ectopic pregnancy. Lehfeldt and associates {5) estimated that, if the 300 000 women enrolled in the Population Council IUD study did not ha ve their rate of fertilization changed by use of the IUD, then during the 45 000 womanyears of use, about 180 000 fertilized ova and 900 ectopic pregnancies would be expected. Of the ectopic pregnancies, 845 would be tubal and approximately five ovarian. Actually, during this time there were 1046 pregnancies with the device in situ, 45 of which were ectopic, 40 tubal and 5 ovarian. From analysis of this data, it was concluded that use of the IUD reduces uterine implantation by 99.5%, tubal implantation by about 95% and ovarian pregnancy not at all. Thus, if pregnancy occurs with an IUD in place, there is a greater chance of it being ectopic than if no IUD were used. If a patient conceives with an IUD in place, her chances of having an ectopic pregnancy range from 3% to 9%. This incidence is about ten times greater than the reported ectopic pregnancy frequency of 0.3% to 0. 7% of total births in similar populations. If a
Medicated and nonmedicated IUDs
patient conceives with an IUD in place, it should be suspected that she has an ectopic pregnancy. Prematurity Tatum and associates (14) studied conceptions with a Copper-T IUD in the uterus and found that the rate of prematurity among live births was four times greater when the Copper-T was left in place than when it was removed. Therefore, if it is not possible to remove the IUD and the patient wishes to continue her gestation, she should be warned of the possible increased risk of prematurity in addition to the increased risk of spontaneous abortion. She should also be informed about an increased risk of ectopic pregnancy and possibly septic abortion, and be reminded to report the first signs of pelvic pain or fever. Infection In the 1960s, despite great concern among gynecologists that use of the IUD would markedly increase the incidence of salpingitis, there was little evidence that such an increase did occur. In 1966, we performed aerobic and anaerobic cultures of endom'e trial homogenates obtained transfundally at varying intervals after insertion of a Lippes Loop (9). In the first 24 hours following insertion, the normally sterile endometrial cavity was consistently infected with bacteria. Nevertheless, the natural defenses destroyed these bacteria within 24 hours in 80% of the cases. In our study, the endometrial cavity, the IUD and the portion of the thread within the cavity were found to be consistently sterile when transfundal cultures were obtained more than 30 days after insertion. These findings indicated that development of salpingitis more than one month after insertion of IUDs other than the shield device is of venereal origin and unrelated to the presence of the IUD. These findings agree with the incidence of clinical diagnosis of pelvic inflammatory disease (PID) found in a group of 23 977 mainly parous women analyzed by Tietze and Lewit (16) in 1970. In this group, there were 239 cases of PID in which the device was removed and 43 7 in which the device remained in situ during and following treatment. When PID rates were computed according to duration of time that the IUD was in place, the rates were higher in the first two weeks after insertion and then steadily diminished. Rates after the first month were on the order of 1.0 to 2.5 per 100 woman-years. The results of both of these studies provide evidence that one should not insert an IUD into a patient who may have been recently infected with a gonococcus, as insertion of the device will transport these
485
pathogens into the upper genital tract. If there is clinical suspicion that a gonococcal endocervicitis exists, cultures should be obtained and the IUD not inserted until negative results are obtained. Following introduction and widespread use of the shield device among nulliparous women in whom IUDs were previously inserted only occasionally, there appeared several studies which suggested that use of the IUD increased the relative risk of developing salpingitis or PID. Tatum and associates (15) carefully examined the sheaths of the appendage of both new shields in their sterile packages and those which were removed from patients. They found that 34% of the former and 9 % of the latter had breaks in the sheath around the knot which attached it to the device. These breaks would allow bacteria continuous access from the vagina into the endometrial cavity and thus increase the risk of upper genital tract infection. For this reason, all of the shield-type IUDs currently in place should be removed. Faulkner and Ory (2) performed a prospective case control study of patients attending an emergency room in Atlanta. They demonstrated a relative risk of 5.1 for IUD-related febrile PID versus non-IUD related febrile PID. In the prospective British Family Planning Study (12) based on inpatient illness, the incidence of salpingitis was about three times greater in IUD users than in diaphragm or oral contraceptive users. In a prospective casecontrol study from Sweden performed by Westrom and associates (17), the diagnosis of salpingitis was made by laparoscopy while the controls were chosen by questionnaire. Analysis of the data by Ory (11) showed that the relative risk for the development of salpingitis "in IUD users versus nonusers was 2.9 for all women, 6.9 for nulliparous women, but only 1. 7 for parous women (Table II) . A recent retrospective case-control study conducted in Seattle by Eschenbach and co-workers found a relative risk for PID among IUD users versus nonusers to be 4.4 for all women, 2.8 for those having gonococcal PID and 6.5 for those having nongonococcal PID. This study also found an increased risk of PID in nulliparous versus parous IUD users. The increased incidence of salpingitis in nulliparous IUD users is possibly related to the greater number of sexual partners in this group of women. There are several problems with these studies. One problem is that uniform guidelines for the diagnosis of salpingitis or PID are not used. Changes in the criteria used to establish the diagnosis of salpingitis may increase the frequency of the diagnosis among IUD wearers. Patients with lower abdominal pain who have only minimal or no temperature evaluation may have the diagnosis of salIn! J Gynaecol Obslet /6
486
Mishell
Table II. Relative risks of PID for IUD users compared to nonusers by age and parity8 • Age (years) Age Standardized
Parity
Nulliparous Multiparous Parity standardizedb 8
b
c
