Complications of IVF and ovulation induction

Reija Klemetti1,3, Tiina Sevón1, Mika Gissler2 and Elina Hemminki1

1 Research on Practices and 2 Information Unit, STAKES, National Research and Development Centre for Welfare and Health, Helsinki, Finland

3 To whom correspondence should be addressed: STAKES, P.O.Box 220, 00531 Helsinki, Finland. E-mail: reija.klemetti{at}stakes.fi


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: The frequency and importance of complications of IVF and other ovulation induction (OI) are poorly known. We examined the occurrence of serious complications and miscarriages leading to hospitalization or operation after IVF (including microinjections and frozen embryo transfers) and OI treatment (with or without insemination). METHODS: Women who received IVF (n = 9175) or OI treatment (n = 10 254) 1996–1998 in Finland were followed by a register linkage study until 2000. RESULTS: After the first IVF treatment cycle, 14 per 1000 women had a serious case of OHSS (ovarian hyperstimulation syndrome), with 23 per 1000 throughout the study period (mean of 3.3 treatments). The corresponding values after OI were very low. The rates of registered ectopic pregnancies and miscarriages after IVF were nine and 42 respectively per 1000 women, with corresponding rates after OI of eight and 42. Infections and bleeding were not common after IVF and even rarer after OI. Overall, 15% of IVF and 8% of OI women had at least one hospital episode during the study period. CONCLUSIONS: Though there was a low risk of complications after each IVF treatment cycle, repeated attempts resulted in serious complications for many women, and these occurred much more often than after ovulation induction alone.

Key words: complications/IVF/OHSS/ovulation induction/register-based study


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Impaired fertility has increasingly become a health service issue because of the availability of new treatments, especially IVF and its related procedures, such as ICSI and frozen embryo transfer (FET) (called here IVF). Older treatments, including ovulation induction with or without insemination (OI), are still in wide use. In Finland, currently ~5% of infants are born as a result of IVF or ovulation induction (Gissler, 2003Go). According to our unpublished data, the estimated yearly number of treatment cycles between 1996 and 1998 was 8200 for IVF and 6550 for OI, compared to 1320 and 1360 resultant births per year.

The novelty of IVF has attracted a large number of studies on the health of the newborn (Helmerhorst et al., 2004Go), but less is known about the long-term health effects of IVF on children (Hampton, 2004Go) or about the health effects on the women. Complications can occur during the ovulation induction, the oocyte collection procedure, and post-operatively. The pregnancy achieved can be ectopic (embryo outside the uterus) or heterotopic (embryo both in and outside the uterus), and it can end in a miscarriage.

Ovarian hyperstimulation syndrome (OHSS) is considered the most serious complication of ovulation induction. It can vary from being a mild illness to a severe, life-threatening disease requiring hospitalization. OHSS can occur as soon as a few days after receiving HCG (‘early OHSS’) or later (‘late OHSS’). Multiple pregnancy has been shown to be associated with a higher risk of late OHSS (Mathur et al., 2000Go). The incidence of severe OHSS has been reported to vary from 0.7 to 1.7% per initiated cycle (Bergh and Lundkvist, 1992Go; Serour et al., 1998Go; Westergaard et al., 2000Go; Nyboe Andersen et al., 2005Go). Some case reports (Cluroe and Synek, 1995Go; Koo et al., 2002Go), studies (Bergh and Lundkvist, 1992Go; Serour et al., 1998Go; Abramov et al., 1999aGo), and reviews (Whelan and Vlahos, 2000Go; Delvigne and Rozenberg, 2003Go; De Sutter, 2004Go) describe some serious aspects of OHSS, such as thromboembolic events, pulmonary manifestations, and death, but the magnitude of the risk is unclear.

The frequency of IVF complications other than OHSS has been only sporadically reported. For oocyte collection, bleeding has been reported in 0.03–0.5% and infections in 0.02–0.3% of embryo transfers (Bergh and Lundkvist, 1992Go; Nyboe Andersen et al., 2005Go). Two to 5% of IVF pregnancies have been reported to be ectopic and 0.1–0.3% heterotopic, and estimates of IVF pregnancies ending in miscarriage have varied from 15 to 23% (Roest et al., 1996Go; Serour et al., 1998Go; Westergaard et al., 2000Go; Kupka et al., 2003Go; Schieve et al., 2003Go; Bryant et al., 2004Go).

Most studies reporting complications after ovulation induction with or without insemination are based on old data (St Clair Stephenson, 1991Go; Venn et al., 1994Go), and the frequency of OHSS after OI is unknown (Unkila-Kallio, 2001Go). In one small clinical study, the OHSS hospitalization rate per initiated cycle was 0.07% (Quasim et al., 1997Go).

Even though various adverse effects of IVF and OI on treated women have been identified, many of the published studies and reports are deficient. They are based on old data, voluntary reporting, or a small number of cases or treatment cycles and concentrate on only one complication, or they lack information on the severity of the complications. The purpose of this study is to estimate the frequency of serious complications and miscarriages following IVF and OI. The criterion of seriousness is the need for hospital care, either in the form of hospitalization or an operation at a hospital outpatient clinic. This study is the first extensive study examining the different serious complications of both IVF and OI.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The study is a historic cohort study based on prospectively collected register data on the following two exposed cohorts: 20–59 year old women having had IVF treatments (IVF, ICSI and FET, n = 9175) and women having had other infertility treatments including drug-based treatments (ovulation inductions with or without artificial insemination, OI, n = 10 254) between 1996 and 1998 in Finland. The women were identified with a pre-designed algorithm using the reimbursement files of the Finnish Social Insurance Institution (SII) (Hemminki et al., 2003). Population controls, matched by age and municipality, were randomly selected for IVF women (n = 9175) from the SII population record (which covers the entire Finnish population).

The data included detailed information on the use of infertility drugs, including dates of prescription and purchase. The beginning of a cycle was defined by the date of the first purchase of the drug. All drugs bought within 35 days of the first purchase were considered part of the same treatment cycle with the exception of clomiphene citrate; a new prescription of clomiphene was considered the beginning of a new cycle regardless of the time interval. Consecutive cycles without a new prescription could not be separated.

The women’s background characteristics were obtained from the Central Population Register, information about care episodes in hospitals from the Hospital Discharge Register (HDR), and dates and causes of death from the Cause-of-Death Register. The HDR collects information on inpatient care as well as on those visits to outpatient clinics that included an operation. It gathers information on diagnoses (10th revision of the International Classification of Diseases, ICD-10), operation codes, and dates of admission and discharge. The diagnoses include the main diagnosis (which can consist of separate diagnoses of the reasons and the symptoms leading to hospitalization) and two secondary diagnoses (which can also also consist of two separate diagnoses of the reasons and the symptoms) for each hospital episode. Miscarriages and ectopic pregnancies can be identified if they lead to inpatient care or an operation (such as laparoscopic surgery or curettage). From 1983 to 1995, operations were registered according to a national coding system; since 1996, according to the Finnish version of the NOMESCO classification system for surgical operations (Toimenpideluokitus, STAKES, 1996Go).

All hospital episodes due to OHSS, pelvic infections and abscesses, bleeding, other complications, miscarriages, and ectopic pregnancies were identified using the HDR (codes in Appendix); the dates for the start and end of the hospital stays were recorded. In cases of miscarriage and ectopic pregnancy, the hospitalization can be registered with an ICD-10 code and/or an operation code. A hospital outpatient visit is registered only when an operation has been carried out. By using ICD-10 codes, 71% of miscarriages after IVF and 74% after OI could be identified, and the rest were identified by using only operation codes. Almost all ectopic pregnancies were identified by using ICD-10 codes (98% of IVF and 97% of OI cases).

In addition to the diagnoses defined in the HDR to be OHSS-related, we searched for women having had symptoms or diseases potentially related to OHSS (‘potential OHSS’). Our definition was based on one unpublished Finnish clinical study of women with diagnosed OHSS, on consultations with experienced clinicians, on symptoms and diseases described in the literature, and on diagnoses given to women in our data just before OHSS diagnosis (ICD-10 code N98.1). Five of the 48 women listed in the Finnish clinical OHSS data could be found in the HDR with the ICD-10 code N98.1. Two other women were hospitalized after IVF treatment with a diagnosis of pain localized in other parts of the lower abdomen (R10.3), and one due to other and unspecified ovarian cysts (N832). The diagnoses women received just before they were hospitalized due to OHSS were E15 (non-diabetic hypoglycaemic coma), R10.3, and J90 (pleural effusion, not elsewhere classified).Combining these diagnoses with those found in the literature, we made a list of diagnoses probably related to OHSS by consulting experienced clinicians (see Appendix).

Using all diagnoses and operation codes for all hospital episodes, we searched for those due to an IVF/OI complication (hereafter called a complication episode). To be eligible, the episode had to have occurred within 120 days from the beginning of the IVF/OI treatment in the case of OHSS, infections, abscesses, bleeding, and other complications. A time lag of 240 days was chosen in the case of miscarriages and ectopic pregnancies. These time lags were defined after studying the shapes of distribution curves of each complication to find a point when incidences clearly decreased. In addition, we consulted experienced clinicians in calculating the probable time lag from the start of treatment (first purchase date of the drug) to the occurrence of potential complications. If a woman had several treatments within this time frame, the latest was defined to be the treatment that led to the complication. If the same episode included different complications (e.g. OHSS and bleeding), they were all counted.

The complication risk was calculated in two ways. First, the proportion of women whose first complication occurred after the first treatment cycle (in our study window) was calculated separately for each type of complication (risk after the first treatment). Secondly, all treatment cycles were considered, but still only the first occurrence of each complication was taken into account, and the proportion of women having at least one complication (of each type) was calculated (risk of a complication after an average of 3.3 IVF and 2.7 OI treatment cycles). Furthermore, the risk of OHSS in each treatment cycle was calculated. In this calculation, individual women can appear more than once. The proportion of women having had any complication episode during the study period was calculated as well as the proportion of women whose hospital visit had lasted >5 days.

To calculate whether OHSS is more common in multiple than singleton pregnancies (among pregnancies ending in birth) and the rate of miscarriages and ectopic pregnancies per 100 births, we linked the data to the nationwide Medical Birth Register by using the women’s personal identification numbers. For IVF and OI births, time limits of 44 and 48 weeks respectively were used to define whether births were the result of IVF or OI or spontaneous pregnancies; times were calculated from the beginning of treatment (the first purchase date of the drug) to the date of birth. In addition, to be able to estimate the risk of each complication per initiated cycle and to find comparable rates for earlier studies, the number of women having had each type of complication was divided by the total number of treatment cycles. This calculation produced only raw estimates because only the first occurrence of each complication was counted.

The numbers of deaths during and after the exposure to IVF and OI until the end of 2000 (after an average of 3.7 and 3.8 years for IVF and OI women respectively) were obtained from the Cause-of-Death Register. The follow-up time of the control group was as long as that of the IVF women (from the first date of IVF exposure to the end of 2000). The causes of death were classified according to the following eight categories: reproductive mortality (Fortney et al., 1986Go) including methods related to achieving pregnancy, diseases of the circulatory system, cancers, suicides, homicides, accidents, other, and unknown causes.

The differences were tested by using a {chi}2-test or a test of relative proportions. The statistical analyses were performed in SAS, version 9.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The background characteristics of the women are presented in Table I.


View this table:
[in this window]
[in a new window]
 
Table I. Background characteristics of women in the IVF, ovulation induction (OI) and controla groups at the beginning of the 1996–98 follow-up in Finland

 

Depending on the definition of OHSS, the number of women having OHSS after the first treatment cycle varied from 14 to 19 per 1000 women and after all treatment cycles from 23 to 35 (Table II). As a specified (N98.1 in ICD-10) and potential diagnosis, OHSS was most common 1–2 and 2–3 months respectively after the beginning of IVF treatment. Depending on the definition of OHSS, the total rate of hospitalization due to OHSS per initiated IVF treatment cycle varied from 0.9% (210 cases/24 318 cycles) to 1.4% (330 cases/24 318 cycles). The mean length of hospitalization due to OHSS and potential OHSS was 4.1 and 3.3 bed days respectively.


View this table:
[in this window]
[in a new window]
 
Table II. Serious complications in the firsta IVF or ovulation induction (OI) treatment cycle, and including allb treatment cycles in the study period, according to the time lag from the start of the cycle to hospitalization, per 1000 women

 

After OI (94% of women treated with clomiphene citrate) hospitalization due to OHSS was rare (Table II). Depending on the definition of OHSS, the total rate of hospitalization per OI treatment cycle varied from 0.04% (eight patients/18 000 cycles) to 0.5% (100 patients/18 000 cycles). Most cases of OHSS occurred in the first month after the treatment cycle. Potential OHSS typically appeared 2–3 months after the treatment cycle.The mean length of hospitalization due to ICD-10 OHSS and potential OHSS was 2.4 and 1.9 bed days respectively.

The risk of OHSS was highest after the first and after the fifth or more IVF treatment cycles (Table III).


View this table:
[in this window]
[in a new window]
 
Table III. Proportion (%) of IVF womena having ovarian hyperstimulation syndrome (OHSS) leading to hospitalization and the proportion with long hospitalization in each treatment cycle

 

In our data, 21.2% of IVF pregnancies ending in birth (total of 3737 births) were twin births and 0.4% were triplets. The risk of OHSS (ICD-10 code) was more common among twin than among singleton pregnancies ending in birth (3.2 versus 1.4%, P < 0.001 for a test of relative proportions). In triplet pregnancies, no OHSS was registered.

After IVF treatment, OHSS (ICD-10) was much more common among younger than older women; 3.2% of women <35 years of age were hospitalized due to OHSS, but only 1% of older women. This was also the case when potential OHSS was taken into account.

Bleeding that necessitated hospital care was a rare complication (Table II). Most instances of bleeding occurred within 2 months of starting the treatment cycle. The total rate per initiated treatment cycle was 0.09%. Ten per 1000 IVF women were hospitalized due to infections; infections were diagnosed somewhat later than bleeding. The total rate per treatment cycle was 0.4%. The mean length of hospitalization after all treatment cycles was 2.1 bed days for bleeding and 3.7 for infections. After OI, hospitalizations due to bleeding and infections were even rarer than after IVF (Table II).

Only 17 IVF women were hospitalized due to complications other than OHSS, bleeding, and infections, and these mainly occurred 1–2 months after starting the treatment (Table II). Six women (four registered as pregnant) had a thromboembolic event. Three of these six cases were serious: one cerebral infarction and two pulmonary embolisms. None of the six women was registered as having OHSS in their thromboembolic episodes, but the cerebral infarction occurred just after hospitalization due to OHSS. One pulmonary embolism was registered as a complication of assisted reproduction (ICD-10 N98.8) that had included excision of the ovary and Fallopian tube.

After all OI treatment cycles, a total of five women had another complication. Four women (two registered as pregnant) had a thromboembolic event, but there was no information in the register about any of them possibly having OHSS. One was a serious case with a pulmonary embolism.

After the first IVF treatment cycle, 42 per 1000 women had received hospital care due to miscarriage; after all treatment cycles, the value was 93 per 1000 women (Table IV). The need for hospital care increased steadily until 4 months after IVF treatment, but the hospitalizations were short; the mean length of hospitalization was 1.2 bed days. The number of miscarriages per 100 births after all IVF treatment cycles was 23 (854/3737).


View this table:
[in this window]
[in a new window]
 
Table IV. Miscarriage or ectopic pregnancy leading to hospitalization or an operation after the firsta IVF or ovulation induction (OI) treatment cycle, and including allb treatment cycles in the study period, according to the time lag from the start of the cycle to hospitalization, per 1000 women

 

After the first IVF treatment cycle, nine per 1000 women, and after all IVF treatment cycles, 21 per 1000 women had needed hospital care due to an ectopic pregnancy (Table IV). Ectopic pregnancies had led to hospitalization most commonly during the first 3 months after the beginning of the IVF cycle, and the mean length of hospitalization was 2.1 bed days. The rate of ectopic pregnancies per 100 births after all IVF treatment cycles was 5.0 (187/3737).

The percentage of women having a miscarriage was similar after the first OI and IVF treatment cycle, but after all treatment cycles, miscarriage was 1.6 times more common among IVF women than among OI women (Table IV). The number of miscarriages per 100 births (15.0) after all OI treatment cycles (634/4188) was lower than the corresponding rate after IVF (23.0). Ectopic pregnancies were almost equally common after the first treatment cycle in both groups, but they were more common among IVF women than OI women after all cycles. The number of ectopic pregnancies per 100 births after all OI treatment cycles was 2.6 (107/4188).

Overall, after all treatment cycles, 1354 IVF (15%) and 824 OI (8%) women were hospitalized for complications. Of these hospital episodes, 10.5% of IVF and 1.6% of OI women’s episodes lasted >5 days.

A total of 12, 15 and 37 women died in the IVF, OI and control groups respectively during the follow-up as a whole (after an average of 3.7 years (IVF) and 3.8 years (OI) from the time of exposure). The causes of deaths are presented in Table V. One death in both the IVF and the OI group was related to reproduction.


View this table:
[in this window]
[in a new window]
 
Table V. Deaths by cause from 1996 to 2000 [after an average 3.7 years for women in IVF and control groups, and 3.8 for ovulation induction (OI) women after the first exposure]

 


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The risk of complications after each IVF treatment cycle was low, but cumulatively repeated attempts led to hospital care in the case of many women. After ovulation induction (OI) treatment, there were far fewer complications. OHSS and miscarriages were the most common reasons for hospital care. OHSS occurred after IVF much more often than after OI alone, but miscarriages and ectopic pregnancies were equally common after OI and IVF after the first treatment cycle. Other serious complications, infections, and instances of bleeding were quite rare, especially among OI women.

Are our estimates of complication risks reliable? The identification of IVF and OI cohorts from the reimbursement records went well (Hemminki et al., 2003Go). We also believe that the data covered most Finnish women treated from 1996 to 1998. The cohorts are large enough to study the frequency of even rare complications caused by IVF and OI. However, our method of identification of serious complications and miscarriages was based on a register covering only hospital care, inpatient care, and operations in outpatient clinics; therefore our identification of outcomes depended on care practices and the diagnosis recording style of physicians.

Even though the validity of the HDR has been high (Keskimäki and Aro, 1991Go), it is a routine care register that does not have the accuracy of ad hoc epidemiological studies. For example, OHSS is not always easy to diagnose. Symptoms may be atypical, and there may be no questions about a history of IVF or OI treatments; or a physician may not enter the code for a specific form of OHSS and instead list its symptoms, such as abdominal pain, enlarged ovaries, or dyspnoea. Likewise, the identification of miscarriages using the register is difficult (Hemminki, 1998Go). The time lags (between the start of a cycle and hospitalization due to a complication) were defined on the basis of drug purchases, and the chosen thresholds were relative. We only had data on the purchase day, not the actual start of the cycle. There may have been intervals between drug purchases and use. On the other hand, due to the character of IVF treatment and the costs of the drugs, it is unlikely that women would have bought the drugs a long time before their use.

Overall, we think that our results underestimate the real complication risks. Mild (less serious) cases are missing in our data. To calculate how reliable our definition of OHSS is, we correlated our data with unpublished Finnish clinical data from the same time period. This showed that only 16% of OHSS cases are in the HDR (10% with the OHSS code and 6% with codes related to OHSS symptoms). Most likely the rest of the cases are missing because the women had had a less serious OHSS that did not lead to hospitalization or to an operation. Some cases may be missing because OHSS was not correctly diagnosed or registered. The overall frequency of OHSS, including mild forms, is no doubt much higher than our results show. On the other hand, some symptoms of OHSS such as abdominal pain are common and can relate to many other diseases. Thus some of our OHSS cases (‘potential OHSS’) may have actually been other diseases.

Our results cannot be extrapolated to countries with a less advanced health care system or less developed practices, in which treated women are less strictly screened or more risks are taken. For example, multiple pregnancies are suspected to entail a higher risk of late OHSS (Mathur et al., 2000Go). In our population, plural births were lower than in many other countries (Bryant et al., 2004Go; Nyboe Andersen et al., 2005Go).

In our study, the hospitalization rates of OHSS per initiated IVF treatment cycle are somewhat higher and the length of hospital stays shorter than reported earlier (Quasim et al., 1997Go; Serour et al., 1999Go). Earlier studies are, however, based on less extensive or older data than ours (from 1980s to mid 1990s). It is possible that serious OHSS became more common from the 1980s to the late 1990s, when more potent treatments were used. At least in Israel, the incidence of severe OHSS after IVF increased from 1987 to 1996 (Abramov et al., 1999bGo). On the other hand, it might be that OHSS has become better known and more readily diagnosed. Shorter stays at hospital for OHSS may relate to the general trend towards shorter hospital stays.

As earlier reported (Schenker, 1999Go), women aged <35 years are at greater risk of developing OHSS. According to that study, the risk is also greater among women who receive only one treatment cycle or many cycles. It is generally assumed that a woman with a previous OHSS is at greater risk of OHSS in a following cycle (Whelan and Vlahos, 2000Go). For that reason, it may be assumed that women with previous OHSS are more carefully monitored during the next cycle to prevent OHSS. However, the women’s first cycle was the first in our study window. We do not know how many cycles the women had received before 1996. After many repeated cycles, a woman may be at greater risk due to the repeated doses of drugs (especially gonadotrophin). This should be considered when many attempts are needed. Furthermore, the risk was also higher among twin pregnancies than among singleton pregnancies ending in birth. This was also found in an earlier study on late OHSS (Mathur et al., 2000Go). We were not able to classify the OHSS cases as late or early.

Only a few studies (Quasim et al., 1997Go; Abramov et al., 1999bGo) have reported the frequency of OHSS after ovulation induction (OI) alone or compared it to the frequency after IVF. Our study, as well as the older ones from the USA and Israel, shows serious OHSS to be more common after IVF than after OI. In addition, according to our study, the length of hospital stay was longer after IVF than after OI. This suggests that OHSS after IVF is more serious than after OI.

Our study confirmed earlier results that thromboembolic events exist after IVF but are rare complications, and even rarer after OI. In our study, the proportion of women reported as having a thromboembolic event leading to hospitalization (0.07%) was much lower than in an earlier Egyptian study (0.2%) (Serour et al., 1998Go). The difference between the studies might be explained by different data collection methods (register/clinical study) and different time periods. In the Egyptian study, all cases were related to severe OHSS. In contrast, none of the women in our study were registered as having OHSS at the same time as thromboembolic complications. We do not know whether this is because thromboembolic events related to OHSS were not specifically recorded (but considered a part of OHSS), or whether OHSS was not diagnosed or not present. In one case report (Ulug et al., 2003Go), a woman had venous thrombosis without OHSS after ovulation induction and ICSI. According to an extensive review, many studies have reported thromboembolic phenomena related to IVF or ovulation induction without any signs of OHSS (Delvigne and Rozenberg, 2003Go). However, severe OHSS was diagnosed in >76% of thromboembolic cases. Some case reports have described a cerebral infarction complicating OHSS (Koo et al., 2002Go).

The frequency of bleeding per initiated IVF cycle was the same in our study as in an Egyptian study with 3500 cycles (Serour et al., 1998Go). But it was much lower than in a recent report from the ESHRE (Nyboe Andersen et al., 2005Go), covering all bleeding complications, even those not leading to hospitalization. The frequency of infections in our study was similar to that reported earlier in the Nordic countries (Bergh and Lunkvist, 1992Go) and Egypt (Serour et al., 1998Go), but higher than reported by ESHRE (Nyboe Andersen et al., 2005Go). The very low bleeding and infection rates after OI suggest that bleeding and infections were complications of IVF technique.

We found that >9% of IVF and >6% of OI women received hospital care due to miscarriages. We could not identify how many of the treated women had become pregnant, but the number of miscarriages per 100 IVF births (23.0) suggests similar miscarriage rates found earlier (15–23%; Serour et al., 1998Go; Kupka et al., 2003Go; Schieve et al., 2003Go; Wang et al., 2004Go). However, the real miscarriage rate must have been higher, for not all miscarriages lead to hospital care. The miscarriage rate per 100 OI births was lower than the rate per 100 IVF births (15 versus 23). Whether this was due to characteristics of the individual women or the procedure itself could not be judged on the basis of this study. Previously, greatly varying miscarriage rates have been reported after clomiphene-induced pregnancies (9–27%; Venn et al., 1994Go).

Were the miscarriage rates after IVF higher than in natural pregnancies? We did not have a pregnant control group, but 8% of women with spontaneous pregnancies needed hospital care due to a miscarriage according to one Finnish study (Hemminki and Meriläinen, 1996Go). In one study from the USA, the miscarriage rate was similar (15%) among assisted reproduction treatment women and the rest of the female population (Schieve et al., 2003Go). However, in another study from the USA, the risk of miscarriage slightly increased after assisted reproductive treatment (Wang et al., 2004Go).

The rate of ectopic pregnancies per 100 IVF births (5.0) in our study is twice that of earlier studies in Finland (Hemminki and Heinonen, 1987Go; Mäkinen, 1996Go). The frequency per initiated IVF cycle (0.8%) is also somewhat higher than in earlier studies of IVF treatments (Bergh and Lundkvist, 1992Go; Serour et al., 1998Go; Bryant et al., 2004Go). We do not have information on the reasons for infertility among the women in our study. We also could not examine how many of the women studied had tubal occlusion, which is considered to be one reason for the high frequency of ectopic pregnancies among IVF women. In the case of OI, which requires open tubes in order to be effective, clomiphene citrate could be a possible reason for the observed high rate of ectopic pregnancies (Venn et al., 1994Go).

Hardly any data have been published about maternal mortality or other deaths occurring as IVF complications. In one study from Australia, the maternal mortality of IVF women was higher than the national rate (Venn et al., 2001Go). We did not have a control group consisting of spontaneously pregnant mothers or women trying to become pregnant. The best control group for IVF women that we had was OI women, although IVF women were older than OI women. Total mortality within an average of 3.7 years (IVF) and 3.8 years (OI) of follow-up was nearly equal among IVF and OI, and one death in both groups was connected with reproduction. The overall mortality in our study was lower than in the general female population (matched by age and municipality). In particular, the cardiovascular deaths were rarer. This indicates a ‘healthy patient effect’ among IVF and OI women, i.e. sick women are less likely to be married or to receive infertility treatment than healthy women. The socioeconomic position of IVF women in our study was somewhat higher than that of the women in the control group, which can explain some of the lower mortality. Lower mortality has been reported in Australia among women who received IVF compared to women who had registered for IVF but never received the treatment (Venn et al., 2001Go).

Register-based studies with sufficiently large populations enable the examination of rare events. However, such studies have their limitations. Registers provide only limited information, and the coding of diagnoses are very likely to vary. Estimates of the frequencies of complications are needed to help clinicians in choosing safer methods, in applying new methods, and in informing women who contemplate IVF or OI treatment. It would be important to establish a routine follow-up system for IVF and OI treatments and their complications. This should also provide information on the duration and causes of infertility, the exact nature and duration of maternal drug exposure, and maternity background data. In countries with computerized health care and IVF registers, it would be easy to implement such a system. However, even before the possible establishment of such a new follow-up system, current estimates of the complication risk should be available both to women and physicians.


View this table:
[in this window]
[in a new window]
 
Appendix. ICD-10 and operation codes and their explanations used in identifying the complications of IVF and ovulation induction (OI)

 

    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The study was supported financially by the Academy of Finland, the Finnish Ministry of Education (the School of Doctoral Programs in Public Health), the Social Insurance Institution, and the National Research and Development Centre for Welfare and Health.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Abramov Y, Elchal U and Schenker JG (1999a) Pulmonary manifestations of severe ovarian hyperstimulation syndrome: a multicenter study. Fertil Steril 71,645–651.[CrossRef][ISI][Medline]

Abramov Y, Elchal U and Schenker JG (1999b) Severe OHSS. An ‘epidemic’ of severe OHSS: a price we have to pay? Hum Reprod 14,2181–2185.[Free Full Text]

Bergh T and Lundkvist Ö (1992) Clinical complications during in-vitro fertilization treatment. Hum Reprod 7,625–626.[Abstract]

Bryant J, Sullivan EA and Dean JH (2004) Supplement to assisted reproductive technology in Australia and New Zealand 2002. Assisted reproductive technology series. Number 8. Publication of AIHW National Perinatal Statistics Unit, no. PER 26, Sydney, Australia.

Cluroe AD and Synek BJ (1995) A fatal case of ovarian hyperstimulation syndrome with cerebral infarction. Pathology 27,344–346.[CrossRef][ISI][Medline]

De Sutter P (2004) Miscellaneous risks and complications. In Gerris J, Olivennes F and De Sutter P (eds) Assisted Reproductive Technologies: Quality and Safety. Taylor & Francis, London, UK, pp 247–255.

Delvigne A and Rozenberg S (2003) Review of clinical course and treatment of ovarian hyperstimulation syndrome (OHSS). Hum Reprod Update 9,77–96.[Abstract/Free Full Text]

Fortney JA, Susanti I, Gadalla S, Saleh S, Rogers SM and Potts M (1986) Reproductive mortality in two developing countries. Am J Publ Hlth 76,134–138.[Abstract]

Gissler M (2003) Hedelmällisyys Suomessa ennen, nyt ja tulevaisuudessa [Fertility in Finland in the past, present and future]. Yhteiskuntapolitiikka 6,553–565.

Hampton T (2004) Panel reviews health effects data for assisted reproductive technologies. J Am Med Assoc 292,2961–2962.[Free Full Text]

Helmerhorst FM, Perquin DAM, Donker D and Keirse MJNC (2004) Perinatal outcome of singletons and twins after assisted conception: a systematic review of controlled studies. Br Med J doi:10.1136/bmj.37957.560278.EE (published 23 January 2004)

Hemminki E (1998) Treatment of miscarriage—current practice and rationale. Obstet Gynecol 91,247–253.[Abstract/Free Full Text]

Hemminki E and Heinonen PK (1987) Time trends of ectopic pregnancies. Br J Obstet Gynecol 94,322–327.[ISI][Medline]

Hemminki E and Meriläinen J (1996) Long-term effects of cesarean sections: ectopic pregnancies and placental problems. Am J Obstet Gynecol 174,1569–1574.[ISI][Medline]

Hemminki E, Klemetti R, Rintapaavola M and Martikainen J (2003) Identifying exposures from drug reimbursement files: a study on unintended effects of in-vitro-fertilization in Finland. Med Informatics Internet Med 28,279–289.[CrossRef][ISI]

Keskimäki I and Aro S (1991) Accuracy and data on diagnoses, procedures and accidents in the Finnish Hospital Discharge Register. Int J Health Sci 2,15–21.

Koo EJ, Rha JH, Lee BI, Kim MO and Ha CK (2002) A case of cerebral infarct in combined antiphospholipid antibody and ovarian hyperstimulation syndrome. J Korean Med Sci 17,574–576.[ISI][Medline]

Kupka MS, Dorn C, Richter O, Felderbaum R and van der Ven H (2003) Impact of reproductive history on in vitro fertilization and intracytoplasmic sperm injection outcome: evidence from the German IVF Registry. Fertil Steril 80,508–516.[CrossRef][ISI][Medline]

Mäkinen J (1996) Ectopic pregnancy falls in Finland. Lancet 348,129–130.[CrossRef]

Mathur RS, Akande AV, Keay SD, Hunt LP and Jenkins JM (2000) Distinction between early and late ovarian hyperstimulation syndrome. Fertil Steril 73,901–907.[CrossRef][ISI][Medline]

Nyboe Andersen A, Gianaroli L, Felderbaum R, de Mouzon J and Nygren KG (2005) Assisted reproductive technology in Europe, 2001. Results generated from European registers by ESHRE. Hum Reprod 20,1158–1176.[Abstract/Free Full Text]

Quasim SM, Karacan M and Kemmann E (1997) An eight year review of hospitalization for ovarian hyperstimulation syndrome. Clin Exp Obstet Gynecol 24,49–52.[Medline]

Roest J, Mous HVH, Zeilmaker GH and Verhoeff A (1996) The incidence of major clinical complications in a Dutch transport IVF programme. Hum Reprod Update 2,345–353.[Abstract/Free Full Text]

Schenker JG (1999) Clinical aspects of ovarian hyperstimulation syndrome. Eur J Obst Gynecol Reprod Biol 85,13–20.[CrossRef][ISI][Medline]

Schieve LA, Tatham L, Peterson HB, Tones J and Jeng G (2003) Spontaneous abortion among pregnancies conceived using assisted reproductive technology in the United States. Obstet Gynecol 101,959–967.[Abstract/Free Full Text]

Serour GI, Aboulghar M, Mansour R, Sattar MA, Amin Y and Aboulghar H (1998) Complications of medically assisted conception in 3,500 cycles. Fertil Steril 70,638–642.[CrossRef][ISI][Medline]

St Clair Stephenson PA (1991) The risk associated with ovulation induction. Iatrogenics 1,7–16.

Toimenpideluokitus (The Finnish version of NOMESCO’s Classification of Surgical Procedures) (1996) STAKES. Kirjapaino Oy West Point, Rauma.

Ulug U, Aksoy E, Erden H, Bayazit N and Bahceci M (2003) Bilateral internal jugular venous thrombosis following successful assisted conception in the absence of ovarian hyperstimulation syndrome. Eur J Obstet Gynecol 109,231–233.[CrossRef][ISI][Medline]

Unkila-Kallio L (2001) Infertility and its treatment: association with ovarian granulosa cell tumor and impact on vascular endothelial growth factor, leptin and selected tumor markers in serum. Academic dissertation, University of Helsinki, Finland. Yliopistopaino, Helsinki.

Venn A, Grad DE and Lumley J (1994) Clomiphene citrate and pregnancy outcome. Aust NZ Obstet Gynaecol 34,56–65.

Venn A, Hemminki E, Watson L, Bruinsma F and Healy D (2001) Mortality in a cohort of IVF patients. Hum Reprod 16,2691–2696.[Abstract/Free Full Text]

Wang JX, Norman RJ and Wilcox AJ (2004) Incidence of spontaneous abortion among pregnancies produced by assisted reproductive technology. Hum Reprod 19,272–277.[Abstract/Free Full Text]

Westergaard HB, Tranberg Johansen AE, Erb K and Nyboe Andersen A (2000) Danish National IVF Registry 1994 and 1995. Treatment, pregnancy outcome and complications during pregnancy. Acta Obstet Gynecol Scand 79,384–389.[CrossRef][ISI][Medline]

Whelan III JG and Vlahos NF (2000) The ovarian hyperstimulation syndrome. Fert Steril 73,883–896.[CrossRef][ISI][Medline]

Submitted on April 20, 2005; resubmitted on July 10, 2005; accepted on July 14, 2005.





This Article
Abstract
Full Text (PDF )
All Versions of this Article:
20/12/3293    most recent
dei253v1
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Request Permissions
Google Scholar
Articles by Klemetti, R.
Articles by Hemminki, E.
PubMed
PubMed Citation
Articles by Klemetti, R.
Articles by Hemminki, E.