Limited data exists regarding the relationship between obesity and miscarriage in spontaneous conceptions. In 2004, Lashen et al. reported that obesity is significantly associated with an increased risk of first-trimester miscarriage, defined as 6–12 weeks of gestation (OR 1.2; 95 % CI 1.01–1.46; P = 0.04) [10].

In 2011, Boots and Stephenson published a systematic review and meta-analysis on obesity and miscarriage following spontaneous conception. Overall, six studies were included with a total of 3,800 obese women, 3,792 overweight women, and 17,146 normal BMI women. The percentages of women with one or more miscarriage were 16.6 % in the obese, 11.8 % overweight, and 10.7 % in the controls. The odds of having one or more miscarriage were increased for obese women (OR 1.31; 95 % CI 1.18–1.46) when compared to normal BMI women [12]. One of the limitations of this systematic review was the broad definition of “miscarriage” used in the studies. For example, two studies did not describe the gestational age at the time of the pregnancy loss, while another study included all pregnancy losses of less than 24 weeks, and another study included only pregnancy losses of less than 13 weeks. Additionally, five of the six studies were retrospective in design, and most of the studies used self-reported height and weight rather than an objective measurement, thereby limiting the reliability.

Recurrent pregnancy loss (RPL) is a devastating disease that affects up to 5 % of couples [14]. Recently, the American Society of Reproductive Medicine (ASRM) redefined RPL as two or more failed pregnancies [15]. Previously, RPL was more strictly defined as three or more consecutive miscarriages, which the Royal College of Obstetricians and Gynecologists term “recurrent miscarriage” [16].

ASRM recommends initiating clinical evaluation after two ultrasounds or histopathologic documented pregnancy losses [15, 17]. Recently, Bernardi et al. and Foyouzi et al. concluded that chromosome testing of the second miscarriage should be performed, based on cost-effectiveness studies, to select which couples would benefit from a thorough RPL evaluation [18, 19].

Standard evaluation of RPL presently includes cytogenetic analysis of both partners, maternal thyroid stimulating hormone (TSH), prolactin, glucose, anticardiolipin IgG/IgM, beta-2 glycoprotein-1 IgG/IgM, lupus anticoagulant, and assessment of the uterine cavity [15, 17]. With a history of a fetal demise of at least 10 weeks size, evaluation also includes inherited thrombophilia testing, including Factor V Leiden, prothrombin gene mutation, protein C, protein S, antithrombin, and homocysteine [20]. Patients should be advised to avoid conception during the evaluation. Despite this extensive evaluation, almost half of the couples are left without an answer [14].

Recently, the association between obesity and pregnancy loss in the RPL population has been investigated in three studies. In 2004, as a subanalysis of a large case–control study, Lashen et al. evaluated the risk of recurrent early miscarriage, defined as three or more miscarriages of less than 12 weeks of gestation [10]. Despite a small sample size (n = 10), an association between obesity and recurrent early miscarriage was found to be significant (OR 3.5; 95 % CI 1.03–12.01; P = 0.04).

In 2010, Metwally et al. reported a retrospective analysis of 844 pregnancies in 491 women with a history of recurrent miscarriage, defined as three or more consecutive miscarriages. Sixty-nine obese women (BMI ≥ 30 kg/m²) were found to have a significantly increased risk of miscarriage in the subsequent pregnancy as compared to normal-weight controls (OR 1.71; 95 % CI 1.05–2.8). Logistic regression analysis showed that an increased BMI was the second most important factor in predicting subsequent pregnancy outcome, after advanced maternal age [13].

Most recently, Lo et al. published a retrospective analysis of 696 women with unexplained recurrent miscarriage, of which 90 were obese (BMI ≥ 30 kg/m²). The investigators found that obese women with a history of unexplained recurrent miscarriage had an increased risk of miscarriage in their subsequent pregnancy when compared to normal-weight women (OR 1.73; 95 % CI 1.06–2.83) [11].

Although only three studies have evaluated the association of obesity and miscarriage in women with a history of RPL all three studies have shown comparable results. BMI is a modifiable risk factor for women with RPL, and, therefore, we recommend measurement of height and weight at initial consultation, assistance with preconception weight loss, and measurement of weight at all subsequent pregnant and nonpregnant visits.

The association between obesity and miscarriage has been extensively studied in women undergoing advanced reproductive technology (ART). Although many observational studies support this association [21–23], there are several studies that did not find a relationship [24–26]. Between 2007 and 2011, three systematic reviews and meta-analyses were performed. Each included 12–33 observational studies and used either a BMI of 25 or 30 kg/m² as a cutoff [7–9]. Rittenberg et al. and Maheshwari et al. both concluded that obese women undergoing ART had a higher rate of miscarriage [8, 9]. Maheshwari et al. found moderate heterogeneity when comparing BMI < 25 to ≥25 kg/m² (I ²  = 46 %), but no heterogeneity when comparing BMI < 30 to ≥30 kg/m² (I ²  = 0 %) [9]. Conversely, Rittenburg et al. showed moderate heterogeneity when comparing both BMI < 25 to ≥25 kg/m² (I ²  = 48 %) and BMI < 30 to ≥30 kg/m² (I ²  = 56 %) [8].

However, Metwally et al. found conflicting results. In their primary analysis, which included all methods of conception (spontaneous conception, ovulation induction, and IVF/ICSI), there was a significant increase in miscarriage rates in women with a BMI ≥ 25 kg/m² compared to controls (OR 1.67; 95 % CI 1.25–2.25). In the subanalysis of only IVF/ICSI pregnancies, the association was not significant. The limitations of this meta-analysis are the inclusion of only one study with spontaneous conceptions and the absence of an assessment for heterogeneity [7].

Increasing evidence suggests that obesity affects oocyte quality, which subsequently impacts embryo quality [37]. The studies of Wang et al. and Luzo et al. showed that oocytes of high-fat-fed mice had increased mitochondrial dysfunction and abnormal morphology, as well as defects in spindle formation and chromosomal misalignment, when compared to lean controls. These abnormalities led to an increased frequency of oocyte aneuploidy [31, 38].

However, human studies have not associated obesity with an increased frequency of aneuploid miscarriages. In fact, two studies have shown that obese women are more likely to have euploid (46,XX or 46,XY) miscarriages compared to normal BMI women [32, 33]. In a retrospective case–control study from an infertility practice, consisting of 204 miscarriages, Landres et al. compared the frequency of euploid miscarriages in overweight or obese women (BMI ≥ 25 kg/m²) to normal-weight women (BMI < 25 kg/m²). There were 51 overweight or obese women with a significantly increased frequency of euploid miscarriages (56.9 % vs. 36.6 %, P = 0.04) [32].

Recently, Boots et al. reported miscarriage chromosome results in women with a history of recurrent early pregnancy loss, defined as two or more losses of less than 10 weeks gestation [33]. The mean maternal age at miscarriage was similar between the obese and nonobese groups, 35.0 vs. 34.1 years. The frequency of euploid miscarriage among obese women was significantly increased compared to nonobese controls, 58 % vs. 37 % (RR = 1.63; 95 % CI 1.08– 2.47; P = 0.02) [33]. Since euploid miscarriage increases the risk of subsequent miscarriage, whereas an aneuploid miscarriage is considered a random event and does not increase the risk for subsequent miscarriage, these concordant studies suggest that obesity may be an independent risk factor for further miscarriage in women with RPL [39, 40]. To date, there are no studies that have reported on the impact of weight loss on the frequency of euploid miscarriage in women with RPL.

Although human studies suggest that obesity does not increase the frequency of aneuploid miscarriage, obesity maybe impacting the embryo’s epigenetics [41, 42]. Studies show that metabolic insults from maternal dietary imbalance, increased fatty acids, and/or glucose intolerance during embryonic development can cause potentially stable epigenetic modifications, which could explain the lifelong effects seen in offspring [41–43].

Dabelea et al. compared siblings that were born before and after their mother was diagnosed with diabetes. The risk of developing diabetes was significantly higher in the siblings born after the mother developed diabetes, 73 % vs. 33 % (OR 3.7, 95 % CI 1.3–11.3, P = 0.02). BMI was also higher in siblings exposed to intrauterine diabetes with a mean increase of 2.6 kg/m² (0.9–4.3; P = 0.003) [44]. This work proposes that a hyperglycemic environment increases the risk of diabetes beyond the genetic risk alone.

The findings of Debalea et al. are supported by the study of Jungheim et al., who reported that pups from obese murine mothers have increased body fat and are at a higher risk of glucose intolerance and hypercholesterolemia [45]. Stothard et al. compared outcomes of embryos transferred from diabetic female mice to nondiabetic female mice to embryos transferred from nondiabetic to nondiabetic female mice [46]. The fetuses from diabetic female mice had significantly decreased implantation rates and higher rates of malformations. The effects of early, transient exposure to maternal obesity and/or the diabetic milieu appear to influence metabolic and morphologic consequences in offspring [46, 47].

Jungheim et al. reported that embryonic insulin resistance in a type II diabetic mouse model has been shown to increase the risk of miscarriage [48]. In addition, Eng et al. suggested that metformin, which increases insulin sensitivity, reverses this association in the mouse model [49]. In human studies, results are convergent. In 2009, Palomba et al. published a meta-analysis of 17 randomized controlled trials (RCT), which evaluated the pregnancy outcomes in 566 women with PCOS preconceptionally treated or not treated with metformin [50]. There was no difference in the miscarriage rate, 20.9 % vs. 21.2 % (OR 1.02, 95 % CI 0.65–1.21).

In 2012, Morin-Papunen et al. published a multicenter, double-blind, RCT of 320 women with PCOS, defined by Rotterdam criteria, who were randomized to receive preconception metformin or placebo [51]. There was no difference in miscarriage rates between cohorts (15.2 % vs. 17.9 %, P = 0.8). However, live birth rates were significantly increased in the metformin cohort (41.9 % vs. 28.8 %, P = 0.006), and a nonsignificant trend was present in a subanalysis of obese (BMI ≥ 27 kg/m²) women treated with metformin preconceptionally (35.7 % vs. 21.9 %, P = 0.07). Although insulin resistance appears to contribute to the development and maintenance of early pregnancy, it remains unclear whether improvement in insulin sensitivity through the use of metformin alleviates this impact.

In addition to the oocytes and embryos, obesity and its hormonal and metabolic effects may negatively influence the endometrium. Obesity is associated with increased systemic estrogen, which has been shown to decrease endometrial receptivity [34]. Increased acute phase proteins and inflammatory cytokines, such as IL-6, PAI 1, and TNF, are elevated in obese women and have been shown to decrease implantation and impair maintenance of pregnancy [52–54].

Plasminogen activator inhibitor (PAI), produced by adipose tissue, is elevated in obesity and has been correlated with metabolic syndrome [52]. Glueck et al. has published two studies correlating an elevated plasminogen activator inhibitor activity (PAI-Fx) with PCOS and RPL [54, 55]. In 1999, Glueck et al. demonstrated that women with PCOS who have had prior miscarriages but no live births had a significantly higher PAI-Fx than women with PCOS who have had live births but no prior miscarriages, 67 % vs. 29 % (P = 0.05) [54]. In 2003, Glueck et al. studied 33 women with PCOS and RPL and compared them to 16 women with RPL but without PCOS and to 116 normal controls (no PCOS or RPL). The patients with PCOS and RPL were statistically more likely to have higher PAI-Fx, 38 % vs. 8 % (P = 0.004) [55]. These studies suggest that elevated PAI is associated with an increased risk of miscarriage.