Use of intravenous immunoglobulin for treatment of recurrent miscarriage: a systematic review


  • Dr Hebert holds a University of Ottawa Chair in Transfusion and Critical Care Medicine.

D Fergusson, Ottawa Health Research Institute, Ottawa Hospital, 501 Smyth Road, Box 201, Ottawa, Ontario, Canada KIH 8L6. Email


Background  Intravenous immunoglobulin (IVIG) is a fractionated blood product whose off-label use for treating a variety of conditions, including spontaneous recurrent miscarriage, has continued to grow in recent years. Its high costs and short supply necessitate improved guidance on its appropriate applications.

Objective  We conducted a systematic review of randomised controlled trials evaluating IVIG for treatment of spontaneous recurrent miscarriage.

Search strategy  A systematic search strategy was applied to Medline (1966 to June 2005) and the Cochrane Register of Controlled Trials (June 2005).

Selection criteria  We included all randomised controlled trials comparing all dosages of IVIG to placebo or an active control.

Data collection and analysis  Two investigators independently extracted data using a standardised data collection form. Measures of effect were derived for each trial independently, and studies were pooled based on clinical and methodologic appropriateness.

Main results  We identified eight trials involving 442 women that evaluated IVIG therapy used to treat recurrent miscarriage. Overall, IVIG did not significantly increase the odds ratio (OR) of live birth when compared with placebo for treatment of recurrent miscarriage (OR 1.28, 95% CI 0.78–2.10). There was, however, a significant increase in live births following IVIG use in women with secondary recurrent miscarriage (OR 2.71, 95% CI 1.09–6.73), while those with primary miscarriage did not experience the same benefit (OR 0.66, 95% CI 0.35–1.26).

Author’s conclusions  IVIG increased the rates of live birth in secondary recurrent miscarriage, but there was insufficient evidence for its use in primary recurrent miscarriage.


Intravenous immunoglobulin (IVIG) is a fractionated blood product made from pooled human plasma that is used to treat a number of medical disorders.1–3 In past years, several authors have proposed using IVIG to treat obstetric conditions such as recurrent miscarriage that are associated with an autoimmune aetiology. Indeed, because its precise mechanism of action has not been well elucidated, IVIG has been described as a non-specific immune suppressant in many complex immunological diseases. Various investigators have documented targeted immunological effects such as the suppression and neutralisation of autoantibodies, a reduction on natural killer cell activity, modification of cytokine production, inhibition of complement binding and activation and inhibition of superantigens among its many plausible effects on immune function.4,5 The US Food and Drug Administration has approved the use of IVIG as a first-line therapy for autoimmune thrombocytopenia, while other conditions such as recurrent spontaneous miscarriage (both primary and secondary), antiphospholipid antibody syndrome and repeated unexplained in vitro fertilisation failure remain off-label indications for this blood product. The high cost of IVIG must be considered in the context of evaluating its role as a first-line therapy. The British Columbia Provincial Blood Coordinating Office indicated that a single course of treatment for an adult commonly exceeds $10 000.1 They also noted that significant and steady increases in Canada’s budget for IVIG paid for by provincial and territorial governments has risen over 30% each year, with a projected 2000–2001 spending of more than $160 million.6 Furthermore, because it is a blood product isolated from large pools of human plasma, there is a theoretical risk of introducing transmissible agents to women receiving treatment with IVIG. Given the prior use of IVIG for treatment of recurrent miscarriage and questions regarding its degree of efficacy, we conducted a systematic review of randomised controlled trials that compared the use of IVIG with other therapies for this condition.


Search strategy

A systematic search strategy applied to Medline (1966 to June 2005), and the Cochrane Register of Controlled Trials (June 2005) was developed to identify randomised controlled trials evaluating IVIG as a mode of therapy, regardless of clinical indication and disease. We, the investigators, will continue our pursuit of systematic reviews of IVIG in multiple clinical fields using results from this search. The strategy combined the text terms ‘ivig’, ‘igiv’, ‘intravenous immune globulin’, ‘intravenous immunoglobulin’, ‘gammagard’, ‘vigam’, ‘gamimmune’, ‘flebogamma’, ‘sandoglobulin’, ‘iveegam’ and ‘gammaglobulin’ with the Dickersin Filter for randomised controlled trials.7 The bibliographies of all identified meta-analyses and trials were also manually reviewed to identify relevant reports. Two investigators (B.H. and R.S.) independently reviewed all citations retrieved from the electronic search to identify all potentially relevant trials for this review. In cases where a unanimous decision could not be reached, a third party was consulted.

Identification of articles and abstraction of data

We included randomised, controlled, parallel-group trials comparing IVIG with either placebo or an active control. We retained only trials that reported clinically relevant outcomes including live births. All dosage regimens of IVIG were considered. Eligibility was not restricted by language of publication, and abstracts were excluded from this review.

A standardised data abstraction form was developed that included the following categories: country of study origin, modes of therapy compared, dosage and duration of treatment, the number of women randomly assigned to each treatment group, length of follow up, patient demographics, all pertinent outcome information and adverse events and withdrawals within each treatment group. Reported outcomes pertaining to recurrent miscarriage included rate of live birth, mean birthweight and mean gestational age. Data pertaining to pregnancy complications were also extracted.

Criteria for diagnosis

Primary recurrent miscarriage referred to women experiencing three or more consecutive early pregnancy losses without ever having successfully given birth to a child. Secondary recurrent miscarriage referred to women suffering from consecutive early pregnancy losses following at least one previous successful birth.

Data synthesis

Measures of effect were calculated independently for each trial. We pooled studies if they were judged to be sufficiently homogeneous. Random effects models were used to synthesise data from included studies according to the DerSimonian–Laird method.8 In circumstances where pooling of trials was inappropriate, a qualitative discussion of the findings is provided. Discrete data were analysed using Review Manager 4.2 (The Cochrane Collaboration, 2004). Dichotomous data measures of effect such as proportion of patients achieving a live birth were expressed as odds ratios (OR) with 95% CI. An OR > 1 suggests that fewer subjects in the control group developed the outcome. Trial arms without an event (i.e. zero cell) had 0.5 added to that cell to allow for computation. Statistical tests for heterogeneity were also carried out for each meta-analysis in this review, and when significant results were observed (a P value <0.10 indicated significance), efforts were made to discover and explain the primary sources of the heterogeneity. When appropriate, subgroup analyses were conducted for important clinical and methodological characteristics. A validated scale was used to evaluate the methodologic quality of trials based on information reported in each manuscript. This scale provides scoring for randomisation (0–2 points), double blinding (0–2 points) and account for withdrawals (1 point), with scores ranging from 0 to 5.9 A score of 3 or more was considered to be indicative of a study of high methodological quality.


Search findings

A total of 3160 citations were identified using the systematic literature search; 2671 from Medline and 489 from the Cochrane Register of Controlled Trials. A total of 25 citations were independently deemed as potentially eligible by two reviewers (B.H. and R.S.). Seventeen studies were excluded for the following reasons: duplicate publication (n = 3), lacking randomisation (n = 10), abstract (n = 4). Thus, eight trials in obstetrics met inclusion criteria for this review (Figure 1). Summaries of sample size and other pertinent characteristics of these studies are given in Table 1.

Figure 1.

Study selection flow chart.

Table 1.  Randomised controlled trials pertaining to recurrent miscarriage
AuthorsIVIG regimenInclusion criteriaSample size enrolledMean age (years)Types of miscarriage
  1. LMWH, low molecular weight heparin.

  2. Details regarding trials evaluating IVIG therapy for treatment of recurrent miscarriage and details pertaining to treatment regimens, trial-specific inclusion criteria, enrolled sample sizes, mean age and distribution of type of miscarriage are given.

German RSA/IVIG Study Group (1994)17Started treatment after pregnancy achieved. Gave 30 g after positive pregnancy test, then 20 g every 3 weeks until gestational week 25 ± 1. (IVIG 5%; Immuno GmbH, Heidelberg, Germany)≥3 spontaneous miscarriagesIVIG: 33, placebo: 31IVIG: 27.2, placebo: 29.7Primary only
Christiansen et al.10Started treatment after pregnancy achieved. Gave 35 g in week 5,6; 25 g in week 7,8; 30 g in week 28,30, 32, 34 (adjusted by 5 g if weight <60 or >80) (Nordimmun; Novo-Nordisk A/S, Bagsvaerd, Denmark).≥3 spontaneous miscarriages at ≤28 weeksIVIG: 17, placebo: 17Not reportedSecondary (n = 24), or women with no live births and ≥1 pregnancy loss after 14th gestational week (n = 10)
Coulam et al.11Started treatment before pregnancy achieved. Gave 500 mg/kg/month every 28 days in the follicular phase until pregnant or for 4 months. After conception, 500 mg/kg every 28 days until delivery or 28–32 weeks of gestation. (Sandoglobulin; Sandoz Pharmaceutical, Princeton, NJ, USA).≥2 spontaneous miscarriages with the same partnerIVIG: 47, placebo: 48IVIG: 35, placebo: 35Primary (n = 49) or secondary (n = 46)
Perino et al.12Started treatment after pregnancy achieved. Gave 25 g/day in two doses for 2 consecutive days, and one-third dose of 25 g administered 3 weeks later when pregnancy was confirmed. (IVIG; Sclavo Pharmaceutical Company, Siena, Italy)≥3 spontaneous miscarriagesIVIG: 22, placebo: 24IVIG: 30, placebo: 29.3Primary only
Stephenson13Started treatment before pregnancy achieved. Gave 500 mg/kg monthly in follicular phase of menstrual cycle for up to six cycles. (Gamimune N; Bayer Canada Inc., Toronto, Canada)≥2 consecutive spontaneous miscarriagesIVIG: 32, placebo: 30IVIG: 31.9 ± 5.4, placebo: 30.8 ± 5Primary (n = 36) or secondary (n = 24)
Jablonowska et al.14Started treatment after (‘at viable pregnancy’) pregnancy achieved. Gave 20 g every 3 weeks on five occasions if a viable pregnancy was confirmed by ultrasound before each treatment. (Gammonativ; Pharmacia and Upjohn Plasma Products, Stockholm, Sweden)≥3 consecutive spontaneous miscarriages before 20th gestational weekIVIG: 22, placebo: 19Not reportedPrimary (n = 20) or secondary (n = 21)
Christiansen et al.15Started treatment after pregnancy achieved. Gave 0.8 g/kg at each infusion until gestational week 20, then 1 g/kg between weeks 20–26. (Note: from inclusion in weeks 5 until week 10, weekly infusions given; biweekly infusions given between weeks 11–26) (Nordimmun; Hemasure A/S, Copenhagen, Denmark).≥4 confirmed miscarriages before the end of 26th gestational weekIVIG: 29, placebo: 29IVIG: 31, placebo: 31Primary (n = 33) or secondary (n = 25)
Triolo et al.16Started treatment after pregnancy achieved. Gave 400 mg/kg/day for two consecutive days, followed by a single dose each month until either 31 weeks of gestation or occurrence of miscarriage (IgVENA; Sclavo Pharmaceutical Company, Siena, Italy).≥3 spontaneous miscarriagesIVIG: 21, LMWH+aspirin: 21IVIG: 32.1 ± 3, LMWH+aspirin: 30.3 ± 5NA


We identified eight studies consisting of 442 women. The characteristics of the studies are given in Table 1. Seven were placebo-controlled trials evaluating IVIG for treatment of recurrent miscarriage,10–15,17 while the eighth and most recent study evaluated treatment with either IVIG or aspirin and low-molecular-weight heparin (LMWH) for the treatment of recurrent miscarriage associated with antiphospholipid antibodies.16 Two additional reports of trial findings of an already included study were excluded, as they did not provide any further reports of outcome data relevant to our analyses18,19. A total of 400 subjects were enrolled in the seven included trials comparing IVIG with placebo. Two trials (n = 110) enrolled only women with primary recurrent miscarriage,12,17 one trial (n = 34) limited enrolment to women with secondary recurrent miscarriage,10 and four trials (n = 256) enrolled women with either primary or secondary recurrent miscarriage. All included studies were double-blind and achieved a quality score of 3 or greater. Dosage and timing of treatment varied between studies, as did patient populations of interest (Table 1).

Live birth rates

All seven placebo-controlled studies reported on successful live births as the primary outcome. Two studies observed a statistically significant improvement in the rate of live births with IVIG compared with placebo,10,11 while the remaining trials did not. A meta-analysis of all seven studies produced a non-significant effect favouring IVIG (OR 1.28, 95% CI 0.78–2.10; P = 0.33) (Figure 2). The forest plot for this pooling suggests some heterogeneity between studies, although the associated test for heterogeneity was not significant (P = 0.32).

Figure 2.

Meta-analyses of live birth rates among categories of recurrent miscarriage.

The seven trials were next stratified by primary or secondary miscarriage status. A meta-analysis from five trials12–15,17 (n = 183) of live birth rates among women suffering from primary recurrent miscarriage achieved a pools OR of 0.66 (95% CI 0.35–1.26, P = 0.21) (Figure 2), a statistically non-significant finding in favour of placebo. Among women suffering secondary recurrent miscarriage, a statistically significant OR of 2.71 (95% CI 1.09–6.73, P = 0.03) (Figure 2) was observed in favour of IVIG in a meta-analysis of four trials10,13–15 (n = 91). The results suggest that IVIG may be effective in women with idiopathic secondary recurrent miscarriage.

Trials were also stratified according to whether therapy in patients was begun prior to pregnancy11,13 or following achievement of pregnancy.10,12,14,15,17 In the five studies (n = 243) beginning treatment after conception, IVIG did not produce a significantly improved birth rate compared with placebo (OR 0.96, 95% CI 0.55–1.65; P = 0.87) (Figure 3). In the two studies (n = 102) starting therapy prior to conception, women receiving IVIG had an increased live birth rate (OR 2.39, 95% CI 1.08–5.33; P = 0.03) (Figure 3).

Figure 3.

Meta-analyses of live birth rates among all types of recurrent miscarriage, stratified by treatment start time.

When simultaneously stratifying by both timing of treatment and diagnostic group, women with primary miscarriage commencing treatment before conception had a greater increase in the live birth rate in the IVIG group (OR 1.50, 95% CI 0.26–8.82; P = 0.65) (Figure 4) compared with women commencing treatment after conception (OR 0.58, 95% CI 0.29–1.16; P = 0.94) (Figure 4). Furthermore, among women with secondary recurrent miscarriage, commencing therapy after conception showed greater improvement in the live birth rate in the IVIG group (OR 3.04, 95% CI 1.06–8.73; P = 0.04) (Figure 5) than commencing prior to conception (OR 1.94, 95% CI 0.32–11.76; P = 0.47) (Figure 5).

Figure 4.

Meta-analyses of live birth rates among all cases or primary recurrent miscarriage, stratified by treatment start time.

Figure 5.

Meta-analyses of live birth rates among all cases of secondary recurrent miscarriage, stratified by treatment start time.

Mean birthweight, gestational age and preterm births

Among the seven placebo-controlled studies, most reported on child’s birthweight10,12–15,17 and/or gestational age.12,14,15,17 Appropriate measures of variation required for meta-analysis were not typically provided, and thus only a qualitative description of these findings was possible. Across all studies, the reported mean/median birthweights varied between studies from 3047 g (range 1100–4070 g) to 3493 g (range 1090–5030 g) in the IVIG group, and from 3012 g (range 700–3600 g) to 3540 g (range 2780–4200 g) in the placebo group. Mean/median measures of gestational age ranged from 37.5 to 40 weeks in the IVIG group, and from 38.3 to 39 weeks in the placebo group. Percentages of term births between IVIG and control groups were comparable among all studies that reported this information,12,14 and were above 80%, and the number of preterm births was consistently reported to be small.12,13,14,17

Pregnancy complications

A total of 60 complications of pregnancy considered minor or moderate in severity were reported in 202 women receiving IVIG among the seven included trials. The description of events included headache, rashes, itching, shivers, fever, nausea, erythema and generalised pruritus (41 of these reports arose from one study that randomised 29 women to IVIG). A total of 14 similar complications of pregnancy were reported in 198 women receiving placebo. A pooled analysis across all seven studies returned a pooled OR of 3.31 (95% CI 0.55–20.01; P = 0.19), indicating a greater frequency of minor or moderately severe pregnancy complications with IVIG as compared with control.

A total of 16 serious neonatal complications were observed among the pool of IVIG women versus 12 serious complications in women treated with placebo. Among these events were a case of mosaic Down syndrome,11 two preterm caesarian sections (week 31) due to the occurrence of pre-eclampsia/severe growth restriction and an acute vaginal haemorrhage (in the latter case, the infant died due to severe anaemia),10 single cases of cleft palate, maldescended testes, alloimmune thrombocytopenia and a slight nose deviation,15 one case of cervical dystocia,12 a second case of intrauterine growth restriction,12 one instance of congenital absence of the radius,12 and one case of premature development.17 In the placebo group, a total of 12 complications were reported, and included one preterm delivery (week 27) due to severe growth restriction and indications of intrauterine asphyxia10 (infant died following delivery), one ruptured tubal pregnancy in week 7 of gestation,10 one case of hygroma in the tongue,15 one requirement of artificial ventilation for 8 weeks that led to normal development,17 one case of cerebral bleeding and mental retardation,17 one infant that was dystrophic at birth proceeded to have normal development,17 one occurrence of death from respiratory distress syndrome on day 7 of life,12 one case of dynamic dystocia incorrectable by oxytocin perfusion,12 one requirement for caesarean section resulting from twins’ discordant presentation,12 one case of pre-eclampsia12 and one occurrence of oligohydramnios and acute fetal distress.12

Therapy of recurrent miscarriage associated with antiphospholipid antibodies

One study16 compared IVIG (n = 21) versus therapy with low-dose aspirin and LMWH (n = 19) for improvement of live birth rate in women suffering recurrent miscarriage associated with antiphospholipid antibodies. After confirmation of pregnancy, participants were randomised to either 400 mg/kg/day IVIG for two consecutive days followed by a single monthly dose each month until 31 weeks of gestation, or to a combination therapy of low-dose aspirin (75 mg/day) and heparin (5700 iu/day via self-administered injection). Aspirin and heparin use was stopped after 34 and 37 weeks of gestation, respectively.

Monitoring of pregnancy outcomes following treatment indicated that women receiving combination therapy with low-dose aspirin and heparin achieved an improved rate of live births (16/19 = 84%) relative to the IVIG group (12/21 = 57%) (OR 0.25, 95% CI 0.05–1.13). Also of note, the IVIG group suffered more fetal losses during the first trimester of pregnancy (6/21 = 28.6%) than the comparison group (2/19 = 10.5%). Mean birthweights (IVIG: 3246 ± 218 g versus aspirin+LMWH: 3298 ± 236 g) and gestational ages (IVIG: 38.3 ± 2.1 weeks versus aspirin+LMWH: 38.7 ± 2.4 weeks) were found to be comparable between groups.


A total of eight trials involving 442 fetal–maternal pairs were identified in our systematic review. The use of IVIG appeared to have a positive effect on the likelihood of a successful live birth in women suffering from secondary recurrent miscarriage relative to treatment with placebo. However, its use for treatment of primary recurrent miscarriage was inconclusive and may require further research.

Based on information from the seven randomised, placebo-controlled trials and 400 women evaluated in this review for treatment of recurrent miscarriage, our findings suggest that IVIG is effective and worthy of further study in couples with idiopathic secondary miscarriage. While the data in four of five studies strongly suggest that IVIG is not effective in primary recurrent miscarriage, further study in autoimmune patients identified through enhanced diagnostic testing may be warranted. With regard to timing of treatment administration, women who were treated during the follicular phase prior to achieving pregnancy experienced less live births than those who began treatment after confirmation of pregnancy. Analysis of other outcomes such as mean birthweight and mean gestational age did not show any significant differences between IVIG and placebo. Two prior meta-analyses pertaining to recurrent miscarriage were carried out by Daya et al.,20,21 one from 1998 using aggregated data from four studies and the other from 1999 using selections of aggregated data and individual patient data (IPD) from six studies. While both reviews showed trends similar to those arising from our systematic review, our study documented a statistically significant effect favouring IVIG in secondary recurrent miscarriage. Also of interest, the aforementioned IPD review also carried out logistic regression analysis of a host of possible covariates (including IVIG use) after pooling selected women from six of our included studies. They noted that increased maternal age and an increased number of previous miscarriages were significant, negative predictors of achieving a successful live birth, and found that neither IVIG use nor pre-/postconception timing had an impact on the probability of success of a live birth. The lack of homogeneity between studies with regard to diagnosis (primary versus secondary miscarriage), severity of condition (represented by the number of prior miscarriages) and treatment timing and regimen serves as a limitation in this study that requires our conclusions be viewed with caution due to small sample sizes within related subgroups. Further randomised trials evaluating the effectiveness of IVIG in women suffering from secondary recurrent miscarriage should be considered, and must include efforts to address the issues of appropriate timing and dosage of treatment.

One study that enrolled a total of 42 women with recurrent miscarriage associated with antiphospholipid antibodies showed that combination therapy with low-dose aspirin and heparin improved live birth rates (84 versus 57%; P = 0.06) in comparison with IVIG treatment. A number of randomised trials support the use of low-dose aspirin with heparin for this indication,22,23 and it remains as the first-line therapy at this time.

Given the widespread off-label use of IVIG in recent years, its elevated cost and short supply, the need for systematic reviews to evaluate effectiveness of its use in all healthcare fields is of great importance and must be addressed.


Funding was provided by the Ontario Ministry of Health and Long Term Care.