Non-genetic risk factors of miscarriage: a comprehensive umbrella review of systematic review and meta-analysis.

We extracted systematic reviews and meta-analyses that included observational studies. We excluded studies that investigated the association between genetic factors, consequences related to recurrent miscarriage and studies related to the treatment of miscarriage. Additionally, we excluded studies that considered miscarriage a risk factor for other diseases, as well as studies related to the incidence and prevalence of miscarriage. Furthermore, we excluded ecological reviews without individual data, narrative reviews, methodological studies, and systematic reviews that did not report any association indicators or report studies. Finally, we excluded studies in which women used assisted reproductive technology to achieve fertility and studies that examined the impact of male factors on fertility. If there were several review studies on a topic, only the most recent review article or meta-analysis with the most complete information was selected. The outcome variable examined in this umbrella review is miscarriage or unintentional termination of a pregnancy. Since the timing of miscarriage varied across different studies, we standardized these timings into four categories: miscarriage before 20 weeks, before 24 weeks, before 28 weeks, and studies in which the definition of miscarriage was not specified. Scopus, PubMed, Embase, and Web of Science were searched until October 26, 2025, using keywords in related articles, and MeSH and systematic review and meta-analysis articles were conducted on observational studies were extracted. The search strategy used the keywords (“non-genetic factors” OR “Risk factor”) AND (“spontaneous abortion” OR “miscarriage” OR “early pregnancy loss”) AND (“systematic review” OR “meta-analysis”). For more search information strategy, refer to the appendix. At this stage, the required information was extracted by two researchers (MA and ML) separately, and in case of disagreement, the opinions of a third expert (MT) in this field were used. The information that was extracted was: the name of the first author of the article, the number of primary studies of each meta-analysis study, the number of participants, the summary effect size using the random effect model with a 95% confidence interval, the 95% prediction interval, P-value of the summary effect size with the random effect model, effect size of the largest study with a 95% confidence interval, heterogeneity index (I 2 ), P-value of Egger's test, and the score of the quality assessment of the studies. Each included systematic review and meta-analysis was assessed using AMSTAR [ 33 ]. This quality assessment tool included 11 questions, and one point was assigned to each question. The scoring method of this tool was done in such a way that an article that got a score of 8 to 11 on the quality assessment score was considered to be in the high-quality range, and if it was between 4 and 7 or below 4, it was considered a moderate score and low. The quality assessment was done independently by two researchers (MA and ML) separately, and in case of disagreement, the opinions of a third expert (MT) in this field were used. For each meta-analysis study, the summary effect size with the 95% confidence interval was calculated using the random effect model, as well as the 95% predictive interval, which was used to evaluate the heterogeneity further and predict the effect in an individual study [ 34 , 35 ], was calculated using the formula proposed by Higgins–Thompson–Spiegelhalter. To re-analyze, in the case of meta-analysis studies with continuous data outcomes (such as the Standard Mean Difference (SMD) effect size), the effect estimate was converted to an odds ratio (OR) using the method proposed by Chinn S [ 36 ]. Additionally, the I 2 index was used to assess the degree of heterogeneity between the studies, with values greater than 50% or 75% considered to indicate large or very large heterogeneity, respectively. One of the biases assessed in this study was the presence of evidence for small study effects [ 37 ]. This bias was calculated using the regression asymmetry test suggested by Egger et al. [ 38 ] and when the P-value of Egger's test was less than 0.10, this bias was present. Another bias that was assessed in this study, was excessive statistical significance [ 39 ]. To report observed studies, the number of studies with statistically significant results within each meta-analysis was counted. Also, the expected studies with significant results in each meta-analysis were calculated by summing the statistical power for each study component within each meta-analysis. The statistical power of each study component was estimated using a non-central t-distribution algorithm assuming that the true effect size was the same as the largest study of each meta-analysis [ 40 , 41 ].Finally, the chi-square test was used to report the P-value. In cases where the observed studies were more than the expected studies and the P-value of the chi-square test was smaller than 0.10, this bias was present. Based on the reanalysis, the studies significant under the random-effects model and included more than 1000 cases; there was no substantial heterogeneity among these studies; the 95% prediction interval did not include the null value and there was no evidence of the two biases mentioned, which were interpreted as indicating a strong association with the risk of miscarriage. Studies that were significant based on the random effect model(p < 0.001 or p < 0.05); they had more than 1000 cases or more than 500 cases; there was no great heterogeneity between their studies; but the prediction interval of 95% include the null value and there was no evidence of two biases that mentioned, that were included as a suggestive evidence with the risk of miscarriage. If even one of the biases was significant or there was high heterogeneity between studies, that factor was classified in the weak evidence group. The analysis of studies was done using STATA version 17 software.

In the present study, a total of 8182 articles (PubMed = 1241, Embase = 1562, Scopus = 3892, Web of science = 1483 and 4 article was found through other sources) were entered into the Endnote version 21 software and after removing duplicate articles, we screened the articles based on the title, abstract and the full text. Finally, 29 systematic review articles and 147 meta-analysis articles qualified with the inclusion criteria remained (Fig.  1 ). In the full-text screening, 179 studies were excluded because a more recent systematic review or meta-analysis with more original articles and complete information had performed the same review. Fig. 1 Flow chart depicting the study selection screening Flow chart depicting the study selection screening The remaining 147 meta-analysis studies with 227 meta-analysis components were included in this umbrella review. These 227 meta-analyses included a wide range of protective and risk factors related to miscarriage. All non-genetic factors of miscarriage that were present in these 227 meta-analyses were included in the groups of Chemical biomarkers, Environmental factors, Microbial, infectious agents and vaccination, Diseases, Medicines and treatments, Behavioral and Emotional factors, Occupational exposures, Semiconductor factors, Nutritional factors, Menstrual factors, Trace elements, and Serum platelets (Tables 1 , 2 , 3 , 4 ). There were continuous data outcomes on 9 risk factors, including zinc, copper, lead, folic acid, vitamin B12, homocysteine, PWD (Platelet Distribution Width), and MPV (Mean Platelet Volume). Table 1 Characteristics of studies on miscarriage factors (< 20 weeks) Author Non-genetic factors Number of primary studies Number of cases/controls Summary effect size based on random model (95%CI) PI 95% P -value random The largest study I 2 P -value Egger test NOS Chemical biomarkers  [ 152 ] Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers 5 - 1/72(1/37–2/17 - 0/0001 - 0% - 11  [ 153 ] High ratio of neutrophils to lymphocytes in blood serum 12 1750 1989 1/84(1/03–3/27) 0/06–5/15 0/037 1/46(1/19–1/75) 94/5% 0/938 8  [ 153 ] High platelet to lymphocyte ratio in blood serum 9 892 1107 1/66(0/65–4/23) 0/007–7/10 0/284 1/46(1/19–1/75) 96/5% 0/418 8 Environmental factors  [ 154 ] Organic Solvent 5 794 2105 1/25(0/99–1/58) - - - - - 7  [ 155 ] Oil and natural gas 4 - 1/82(1/13–2/94) 0/02–6/20 0/013 2/70(1/80–3/90) 72/7% 0/586 9  [ 156 ] higher urinary levels of monobutyl phthalate (MBP) 7 609 4005 1/39(0/99–1/95) 0/32–2/76 0/054 1/00(0/67–1/50) 33/6% 0/571 10  [ 156 ] mono(2-ethylhexyl) phthalate (MEHP) 7 609 4005 1/56(1/29–1/89) 0/59–2/52 0/0001 1/78(1/35–2/34) 0% 0/552 10  [ 156 ] mono(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP) 6 477 3833 1/59(1/22–2/06) 0/38–3/01 0/0001 1/17(0/73–1/88) 0% 0/948 10  [ 156 ] mono(2-ethyl-5-oxohexyl) phthalate (MEOHP) 3 477 3833 1/47(1/14–1/89) 0/37–2/84 0/003 1/28(0/84–1/95) 0% 0/676 10  [ 156 ] mono(2-ethyl-5-carboxypentyl) phthalate (MECPP) 3 309 473 1/10(0/77–1/57) 0/0001–10/15 0/582 0/87(0/58–1/30) 20/3% 0/495 10  [ 156 ] di-2ethylhexyl phthalate (DEHP 3 201 380 1/79(1/26–2/52) 0/0003–10/42 0/001 1/60(0/96–2/70) 0% 0/367 10 Microbial, infectious agents and vaccination  [ 157 ] Smallpox vaccine 2 265 2493 4/82(2/38–9/76) It could not be calculated 0/0001 5/30(2/52–11/17) 0% It could not be calculated 10  [ 158 ] parvovirus B19 infection 4 - 11/50(2/70–49/70) - 0/001 - 0% - 9  [ 159 ] Covid19 Vaccine 5 2525 20,928 1/08(0/90–1/29) 0/32–2/27 0/365 0/94(0/84–1/05) 47/7% 0/25 10  [ 160 ] Chlamydia trachomatis IgG 12 1739 3037 1/60(1/24–2/07) - - - 40/2% - 10  [ 160 ] Chlamydia trachomatis IgA 6 146 1512 1/23(0/56–2/71) - - - 67% - 10 Diseases  [ 161 ] Portal cirrhosis high blood pressure 3 53 87 1/61(0/62–4/14) Including the null value 0/323 0/78(0/25–2/44) 24% 0/338 9  [ 25 ] Hypothyroidism 10 3547 46,306 2/01(1/66–2/44) Including the null value 0/0001 1/67(1/17–2/36) 0% 0/426 10  [ 162 ] Congenital abnormalities of the uterus 8 2188 76,638 1/91(1/35–2/70) 0/28–3/81 0/0001 1/34(1/06–1/70) 73/1% 0/086 10  [ 163 ] Systemic sclerosis 6 694 4929 1/60(1/22–2/22) - - - - - 8  [ 164 ] Intrauterine hematoma 6 894 3589 2/00(1/23–3/23) 0/10–4/93 0/005 1/65(1/10–2/47) 66/4% 0/093 10  [ 19 ] Ulcerative colitis 8 234 538 1/91(1/06–3/43) 0/10–4/83 0/029 1/23(0/47–3/24) 15/1% 0/779 8  [ 19 ] Crohn's 9 201 720 2/42(1/44–4/08) 0/21–4/83 0/001 1/77(0/63–4/95) 0% 0/001 8  [ 19 ] Inflammatory Bowel Disease 16 537 1444 1/7(1/17–3/00) 0/175–4/237 0/009 0/47(0/24–0/91) 45/3% 0/004 8  [ 165 ] Anorexia nervosa 4 373 1960 1/24(0/95–1/63) 0/12–3/51 0/107 1/36(0/98–1/90) 0% 0/973 10  [ 165 ] Bulimia nervosa 3 573 2254 1/60(0/82–3/10) Including the null value 0/162 1/13(0/90–1/41) 46/3% 0/008 10  [ 166 ] Chorionic bump 4 157 589 2/95(2/02–4/31) 0/12–6/14 0/0001 3/01(1/81–4/99) 0% 0/485 10 Medicines and treatments  [ 167 ] Fluconazole 3 3662 27,025 1/08(0/91–1/29) 0/014–5/425 0/363 1/05(0/87–1/26) 0% 0/02 7  [ 168 ] Artemisinin in the treatment of malaria - 670 1072 0/74(0/47–1/17) - 0/19 - - - 11  [ 168 ] Artemether lumefantrine/Quinine - 465 915 0/67(0/37–1/23) - 0/20 - - - 11  [ 169 ] Statin 4 747 3062 1/48(1/11–1/96) 0/10–3/87 0/007 1/60(1/10–2/50) 0% 0/111 8  [ 29 ] Antibiotics 36 - 1/37(1/23–1/53) 0/68–2/06 0/0001 1/03(0/94–1/13) 90/5% 0/001 10  [ 29 ] Sulfonamides 4 - 1/44(0/90–2/92) 0/007–6/71 0/122 0/97(0/92–1/02) 83% 0/105 10  [ 29 ] Penicillin 8 - 0/97(0/83–1/12) 0/42–1/80 0/696 1/00(0/96–1/04) 83/3% 0/967 10  [ 29 ] Macrolides 8 - 1/41(1/04–1/92) 0/24–3/16 0/026 1/03(0/94–1/13) 90/3% 0/328 10  [ 29 ] Quinolones 6 - 2/48(1/46–4/20) 0/08–5/79 0/001 2/45(1/98–3/03) 89/9% 0/388 10  [ 29 ] Cephalosporins 3 - 0/85(0/65–1/12) 0/0009–7/62 0/262 0/88(0/66–1/18) 0% 0/373 10  [ 29 ] Tetracycline 3 - 2/56(1/95–3/37) 0/002–9/37 0/0001 2/81(1/93–4/10) 0% 0/064 10  [ 29 ] Clarithromycin 3 - 1/98(1/45–2/70) 0/0005–10/11 0/0001 2/35(1/90–2/91) 55/7% 0/786 10  [ 29 ] Erythromycin 2 - 0/88(0/61–1/28) It could not be calculated 0/517 1/03(0/94–1/13) 75/7% It could not be calculated 10  [ 29 ] Amoxicillin 4 - 0/83(0/66–1/04) 0/10–2/90 0/111 0/67(0/59–0/77) 77/79% 0/755 10  [ 29 ] Sulfamethizole 3 - 1/21(0/81–1/80) Including the null value 0/333 0/97(0/92–1/02) 57/9% 0/111 10  [ 170 ] Vedolizumab in the treatment of IBD 4 213 628 1/79(1/06–3/01) 0/02–6/19 0/028 1/50(0/76–2/94) 0% 0/093 9  [ 171 ] Metronidazole 2 1499 154,223 1/77(1/52–2/07) It could not be calculated - - 0% It could not be calculated 11  [ 172 ] Tumor necrosis factor alpha inhibitors in arthritic diseases 6 199 253 1/28(0/61–2/70) 0/01–5/54 0/504 1/67(0/67–4/16) 29/3% 0/082 9  [ 173 ] Montelukast for asthma 4 775 27,001 1/14(0/54–2/37) 0/0008 −8/37 0/727 0/44(0/19–0/98) 70/4% 0/589 10  [ 174 ] Lithium 3 321 968 3/77(1/15–12/39) - 0/03 - 86/55% - 10  [ 175 ] H1 antihistamines in the control of nausea and vomiting 13 2522 7276 0/99(0/82–1/20) 0/43–1/81 0/955 0/88(0/63–1/24) 0% 0/819 9  [ 176 ] Treatment for cervical intraepithelial neoplasia 4 519 584 1/16(0/79–1/69) 0/04–4/36 0/434 1/11(0/65–1/91) 0% 0/325 8  [ 177 ] Levothyroxine in the treatment of hypothyroidism 3 893 4580 0/76(0/61–0/96) 0/002–6/40 0/000 0/76(0/60–0/96) 0% 0/400 9  [ 178 ] Non-obstetric surgery with no surgery/conservative treatment 4 - 1/1(1/1–1/2) - - - - - 8 Behavioral factors  [ 10 ] Alcohol drinking 5 2076 9214 1/10(0/90–1/34) 0/27–2/46 0/369 1/18(0/99–1/41) 68/4% 0/896 9 Occupational factors  [ 179 ] Lifting weights ≥ 11 kg - - 1/31(1/08–1/58) 0/64–2/68 0/006 - 79% - 10  [ 179 ] Lifting ≥ 100 kg 6 - 1/19(0/82–1/73) 0/37–3/82 0/36 1/45(1/29–1/63) 63/3% 0/562 10  [ 179 ] Prolonged standing 11 - 1/06(0/92–1/22) 0/37–3/82 0/41 0/89(0/69–1/14) 0% 0/157 10  [ 179 ] Heavy workload 6 - 1/49(0/91–2/45) 0/27–8/32 0/11 3/35(2/72–4/13) 90% 0/125 10  [ 179 ] Bending ≥ 1h 5 - 1/32(0/82–2/12) 0/23–7/74 0/250 2/31(1/88–2/84) 92% 0/914 10 Including the value for PI: PI was too Wide and included the neutral value Table 2 Characteristics of studies on miscarriage factors (< 24 weeks) Author Non-genetic factors Number of primary studies Number of cases/controls Summary effect size based on random model (95%CI) PI 95% P -value random The largest study I 2 P -value Egger test NOS Chemical biomarkers  [ 13 ] Serum vitamin D deficiency 5 3822 1875 1/78(1/11–2/87) 0/07–4/88 0/017 1/21(0/78–1/89) 30/1% 0/361 9 Environmental factors  [ 180 ] Radiofrequency 5 1724 2090 1/05(0/95–1/17) 0/17–2/86 0/269 1/05(0/95–1/17) 0% 0/176 10 Microbial, infectious agents and vaccination  [ 181 ] Influenza vaccine 3 8025 47,818 1/04(0/71–1/52) 0/43–1/78 0/818 1/11(0/71–1/73) 0% 0/186 9  [ 182 ] Covid-19 virus 9 2158 3826 1/09(0/81–1/48) 0/26–2/46 0/549 0/17(0/01–3/61) 0% 0/548 11  [ 183 ] Human cytomegalovirus infection 8 - 2/12(0/98–4/61) 0/035–6/21 0/056 1/32(0/87–1/99) 49/1% 0/437 9  [ 183 ] Herpes simplex virus 1/2 9 - 3/72(1/86–7/46) Including the null value 0/0001 3/80(1/44–9/97) 0% 0/750 9 Diseases  [ 184 ] Psychological stress 8 - 1/41(1/18–1/69) 0/55–2/34 0/0001 1/28(1/05–1/57) 35/6% 0/170 10  [ 185 ] Sub chronic hematoma 5 646 8427 2/18(1/29–3/68) 0/06–5/72 0/004 1/77(1/26–2/47) 58/9% 0/611 10  [ 186 ] Chronic histiocytic intervillositis (CHI) 4 174 84 0/16(0/03–0/85) Including the null value 0/032 0/09(0/03–0/27) 65/2% 0/729 10  [ 24 ] Endometriosis 9 31,775 140,906 1/74(1/28–2/37) 0/31–3/45 0/0001 1/29(1/24–1/34) 93/4% 0/228 10 Medicines and treatments  [ 187 ] Multiple sclerosis medications 8 - 1/14(0/98–1/32) 0/55–1/85 0/078 1/10(0/92–1/32) 0% 0/411 8  [ 187 ] Interferon(MS medication) 4 - 1/13(0/76–1/69) 0/03–4/55 0/533 0/79(0/46–1/37) 13/7% 0/531 8  [ 187 ] Natalizumab (MS medication) 2 - 1/93(0/96–3/85) It could not be calculated 0/064 1/33(0/65–2/73) 51/2% It could not be calculated 8  [ 187 ] Glatiramer acetate (MS medication) 3 - 1/11(0/65–1/91) Including the null value 0/681 1/14(0/57–2/26) 0% 0/371 8  [ 188 ] Antidepressant 29 - 1/24(1/17–1/30) 0/94–1/51 0/0001 1/14(1/10–1/18) 69/4% 0/004 10  [ 188 ] Selective serotonin reuptake inhibitors 12 52,455 1,710,271 1/29(1/15–1/44) 0/73–1/85 0/0001 1/07(1/03–1/12) 81/5% 0/271 10  [ 188 ] Selective norepinephrine reuptake inhibitors 5 2364 986,732 1/41(1/11–1/80) 0/28–3/02 0/005 1/14(0/96–1/35) 46/2% 0/122 10  [ 188 ] Antidepressants vs. untreated depression 6 19,012 23,534 1/16(1/03–1/30) 0/57–1/85 0/011 1/34(1/19–1/50) 58/5% 0/630 10  [ 188 ] Selective serotonin reuptake inhibitors vs. antidepressants 5 10,731 3450 1/03(0/92–1/15) 0/51–1/72 0/538 1/04(0/92–1/16) 0% 0/917 10  [ 188 ] Selective norepinephrine reuptake inhibitors versus antidepressants 3 1632 28,323 1/19(1/02–1/39) 0/025–5/04 0/025 1/23(1/04–1/46) 0% 0/443 10  [ 189 ] Immunosuppressive therapy in Neuromyelitis Optica 8 102 155 0/51(0/39–0/66) 0/14–1/77 0/0001 0/57(0/26–1/25) 0% 0/170 7 Behavioral factors  [ 190 ] Interval between pregnancies less than 6 months 3 18,706 25,203 0/85(0/71–1/03) 0/007–5/61 0/105 0/97(0/80–1/18) 26/1% 0/198 11 Including the value for PI: PI was too Wide and included the neutral value Table 3 Characteristics of studies on miscarriage factors (< 28 weeks) Author Non-genetic factors Number of primary studies Number of cases/controls Summary effect size based on random model (95%CI) PI 95% P -value random The largest study I 2 P -value Egger test NOS Diseases  [ 11 ] Rheumatoid arthritis 6 - 1/32(1/21–1/43) - 0/001 - 25/8% - 10  [ 26 ] Underweight 9 20,512 168,691 1/08(0/98–1/20) - 0/170 - 54% did not exist 7  [ 26 ] Overweight 7 25,618 87,644 1/07(1/02–1/12) - 0/010 - 41% did not exist 7  [ 26 ] Obesity 7 11,934 87,644 1/15(1/08–1/23) - 0/0001 - 8% did not exist 7  [ 26 ] Obesity (Grade 1) 3 2314 23,233 1/08(0/97–1/20) - 0/180 - 47% did not exist 7  [ 26 ] Obesity (Grade 2) 3 1242 2333 1/50(1/32–1/70) - 0/0001 - 23% did not exist 7  [ 191 ] PCOs 3 295 275 2/86(1/64–4/98) Including the null value 0/0001 3/00(1/39–6/46) 0% 0/123 9 Medicines and treatments  [ 192 ] Inhaled corticosteroid-s (anti-inflammatory and immunosuppressive medication) 3 - 1/20(1/15–1/24) 0/46–2/13 0/0001 1/20(1/20–1/30) 0% 0/428 8  [ 192 ] Systemic corticosteroids (anti-inflammatory and immunosuppressive drugs) 2 - 1/41(0/86–2/32) It could not be calculated 0/163 1/70(1/10–2/50) 36/8% It could not be calculated 8  [ 193 ] Endocervical versus decidual polypectomy 2 64 92 0/26(0/04–1/58) It could not be calculated 0/145 0/58(0/07–4/89) 18/4% It could not be calculated 11 Behavioral factors  [ 23 ] Smoking 50 - 1/22(1/16–1/29) 0/90–1/51 0/0001 1/13(1/05 0 1/22) 58/4% 0/010 9  [ 23 ] Second-hand smoke 17 - 1/11(0/94–1/30) 0/51–1/88 0/192 1/01(0/80–1/27) 44/5% 0/262 9  [ 194 ] pregnancy intention 2 - 1/54(1/78–3/03) It could not be calculated 0/213 1/10(0/90–1/50) 90% It could not be calculated 8 Including the value for PI: PI was too Wide and included the neutral value Table 4 Characteristics of studies on miscarriage factors (definition of miscarriage is not mentioned in the study) Author Non-genetic factors Number of primary studies Number ofcases/controls Summary effect size based on random model (95%CI) PI 95% P-value random The largest study I 2 P -value Egger test NOS Chemical biomarkers  [ 195 ] Gamma-aminobutyric acid analog 4 - 1/64(1/13–2/37) 0/07–4/76 0/0001 1/90(1/10–3/40) 0% 0/762 9  [ 196 ] High serum β-human chorionic gonadotropin 4 2558 46,876 2/88(1/76–4/68) 0/042 −7/08 0/0001 2/18(1/38–3/49) 11/4% 0/270 8  [ 197 ] hyperhomocysteinemia 3 117 133 0/53(0/33–0/86) Including the null value 0/010 0/55(0/31–0/99) 0% 0/949 7  [ 197 ] Serum Vitamin B12 deficiency 2 100 102 3/00(1/70–5/28) It could not be calculated 0/0001 2/55(1/43–4/61) 9/3% It could not be calculated 7  [ 197 ] Folic acid deficiency 4 683 1138 0/78(0/48–1/27) 0/009–5/23 0/322 1/06(0/88–1/28) 75% 0/250 7  [ 198 ] Thyroid autoimmune antibodies 10 1426 11,194 4/29(2/22–8/27) 0/10–8/00 0/0001 6/11(3/96–9/38) 87/6% 0/264 7 Semiconductor agents  [ 199 ] fabrication 6 - 1/26(1/03–1/55) 0/40–2/38 0/020 0/94(0/58–1/50) 6/9% 0/210 11  [ 199 ] Photo-lithography work 6 - 1/37(1/10–1/71) 0/41–2/57 0/005 1/43(1/02—2/01) 0% 0/037 11  [ 199 ] Ethylene glycol ether 4 - 1/42(0/94–2/16) 0/03–5/13 0/090 1/43(1/02–2/01) 53/5% 0/145 11  [ 199 ] Fluoride exposure 2 - 1/19(0/73–1/95) It could not be calculated 0/472 1/14(0/66–1/99) 0% It could not be calculated 11 Environmental factors  [ 200 ] Particulate matter PM2.5 5 - 1/21(1/02–1/42) 0/37–2/37 0/025 1/04(1/03–1/06) 98/9% 0/711 10  [ 200 ] Particulate matter PM10 5 - 1/09(1/02–1/17) 0/67–1/56 0/010 1/02(0/99–1/06) 82/1% 0/011 10  [ 201 ] Pesticides 13 192,426 232,214 1/41(1/10–1/80) 0/39–2/68 0/006 1/62(1/58–1/65) 81/8% 0/379 8  [ 202 ] Perfluoro-decanoate 5 - 1/59(0/95–2/66) 0/04–5/09 0/076 1/10(0/81–1/53) 62/8% 0/360 9  [ 202 ] Perfluoro-hexane sulfonate 5 - 1/03(0/96–1/11) 0/65–1/48 0/368 1/04(0/96–1/12) 0% 0/794 9  [ 202 ] Perfluoro-nonanoate 5 - 1/45(0/72–2/90) 0/01–6/34 0/292 0/86(0/70–1/06) 91/5% 0/495 9  [ 202 ] Perfluoro Octanoate sulphonate 9 - 1/00(0/88–1/12) 0/54–1/60 0/999 0/95(0/87–1/04) 34/1% 0/086 9  [ 202 ] Perfluoro-octane sulfonic acid 9 - 1/06(0/94–1/21) 0/55–1/69 0/317 0/99(0/94–1/05) 43/7% 0/283 9  [ 203 ] Antineoplastic agents 9 - 1/56(1/16–2/09) 0/30–3/18 0/003 1/16(0/99–1/35) 78/2% 0/289 10  [ 204 ] Exposure to dioxins in food 2 837 1065 1/04(0/79–1/37) It could not be calculated 0/751 0/98(0/83–1/17) 23/9% It could not be calculated 9 Microbial, infectious agents and vaccination  [ 205 ] Helicobacter pylori 6 1159 4750 1/50(1/05–2/14) 0/18–3/58 0/06 0/92(0/42–2/02) 27/8% 0/764 10  [ 206 ] Serum antibodies against Toxoplasma gondii 11 2207 1717 3/98(2/03–7/77) 0/15–7/24 0/0001 1/59(0/95–2/66) 32/1% 0/005 9  [ 207 ] Toxoplasma gondii IgG antibody 19 4130 3159 2/21(1/54–3/18) 0/28–4/25 0/0001 0/94(0/77–1/15) 83/9% 0/006 10  [ 207 ] Toxoplasma gondii IgM antibody 19 4077 2740 4/33(2/42–7/76) 0/19–7/42 0/0001 1/14(0/73–1/77) 71/3% 0/001 10  [ 208 ] HIV infection 2 155 948 1/35(0/77–2/35) It could not be calculated 0/29 1/58(0/74–3/37) 0% It could not be calculated 8  [ 209 ] Oral cholera vaccine 2 2147 1107 1/54(0/78–3/03) It could not be calculated 0/205 1/62(0/76–3/44) 0% It could not be calculated 9  [ 210 ] HPV Vaccine 7 52,503 3,280,744 0/98(0/85–1/13) 0/43–1/78 0/852 0/94(0/87–1/02) 68/1% 0/724 11  [ 18 ] Neisseria gonorrhoeae 3 - 2/11(0/50–8/92) - 0/307 0/86(0/56–1/33) 72/9% 0/075 10  [ 211 ] HPV 3 124 386 0/99(0/37–2/64) Including the null value 0/990 1/80(0/99–3/30) 66/1% 0/701 10  [ 212 ] hepatitis B virus infection 2 - 1/09(0/51–2/29) It could not be calculated 0/817 0/77(0/53–1/12) 79/1% It could not be calculated 9  [ 213 ] hepatitis E virus infection 3 274 309 1/71(0/79–3/69) Including the null value 0/169 1/49(0/46–4/78) 1/3% 0/080 10  [ 214 ] Genital mycoplasma 19 1584 2020 2/35(1/50–3/67) 0/19–4/83 0/0001 1/17(0/61–2/23) 69/7% 0/471 8  [ 215 ] Vaginal bacteria 7 622 836 1/30(1/04–1/62) 0/37–2/53 0/019 1/04(0/91–1/19) 56/8% 0/104 10  [ 216 ] Dengue infection 5 - 1/90(0/95–3/78) 0/02–6/33 0/067 0/78(0/29–2/11) 17/5% 0/244 9 Diseases  [ 217 ] Thyroid Cancer 3 2373 9838 1/80(1/28–2/52) 0/0002–10/53 0/001 2/07(1/58–2/72) 32/9% 0/056 9  [ 218 ] Breast Cancer 2 - 1/04(0/86–1/26) It could not be calculated 0/696 1/05(0/86–1/27) 0% It could not be calculated 9  [ 219 ] Gestational diabetes 26 2166 24,032 3/01(2/38–3/82) 0/76–4/38 0/0001 4/32(2/95–6/32) 68/2% 0/572 11  [ 220 ] Leiomyomas 5 1394 20,435 0/86(0/72–1/03) 0/28–1/97 0/107 0/83(0/63–1/08) 0% 0/128 10  [ 221 ] Hodgkin lymphoma 2 - 0/78(0/55–1/10) It could not be calculated 0/160 0/80(0/50–1/29) 0% It could not be calculated 9  [ 43 ] Non-Hodgkin lymphoma 2 - 0/94(0/62–1/43) It could not be calculated 0/787 1/10(0/90–1/40) 55/3% It could not be calculated 9  [ 222 ] Psoriasis 4 - 1/62(0/99–2/63) 0/01–6/10 0/050 1/10(1/00–1/20) 80/6% 0/193 9  [ 223 ] Lyme disease 3 - 1/27(0/42–3/85) Including the null value 0/668 2/14(0/50–9/09) 0% 0/147 10  [ 224 ] Uterine artery embolization 3 244 253 0/86(0/72–1/03) Including the null value 0/290 1/40(0/66–2/97) 0% 0/963 10  [ 225 ] Uterine fibroids 15 - 4/51(2/80–7/26) - 0/0001 - 51/3% - 10  [ 17 ] Primary biliary cholangitis 4 596 1058 1/26(1/01–1/57) 0/19–3/12 0/035 1/22(0/86–1/75) 0% 0/880 10  [ 226 ] Familial Mediterranean fever 5 910 1030 0/94(0/71–1/24) 0/16–2/67 0/693 0/93(0/63–1/36) 89% 0/17 10  [ 227 ] Sjögren's syndrome 4 343 902 9/35(4/18–20/91) 0/007–16/47 0/0001 16/32(6/79–39/23) 0% 0/005 11  [ 228 ] Hyperandrogenic 6 572 694 1/47(0/92–2/35) 0/09–4/20 0/100 1/23(0/75–2/01) 42/7% 0/230 10  [ 229 ] Systemic lupus erythematosus 3 4397 8,788,494 4/64(3/02–7/28) Including the null value 0/0001 4/85(4/26–5/51) 66% 0/944 10  [ 230 ] Depression 2 - 1/88(1/50–2/35) It could not be calculated - - 0% It could not be calculated 10  [ 230 ] Anxiety 2 - 1/25(1/11–1/40) It could not be calculated - - 0% It could not be calculated 10  [ 231 ] Non-alcoholic fatty liver 2 1870 4614 1/15(1/02–1/30) It could not be calculated 0/02 - 0% It could not be calculated 9  [ 232 ] Epilepsy 6 - 1/78(1/11–2/85) - - - 78/6% - 11  [ 233 ] Nausea and vomiting during pregnancy 2 - 0/40(0/12–1/27) It could not be calculated 0/122 0/66(0/46–0/99) 77/6% It could not be calculated 6  [ 234 ] Primary antiphospholipid syndrome 3 190 287 2/47(1/11–5/47) Including the null value 0/025 2/58(1/40–4/74) 36/9% 0/679 8  [ 28 ] History of recurrent miscarriage 2 595 989 6/37(3/83–10/57) It could not be calculated 0/0001 5/54(3/17–9/69) 13/1% It could not be calculated 8  [ 235 ] Vaginal bleeding 8 - 4/30(2/00–9/00) - - - 93% - 11  [ 22 ] Diabetes Mellitus 13 110,872 7,266,388 1/23(1/13–1/33) 0/8–1/65 0/0001 1/11(1/07–1/15) 82/6% 0/812 7  [ 22 ] T1DM 7 4806 233,055 1/18(1/01–1/37) 0/51–2/01 0/030 1/11(1/00–1/23) 49/1% 0/846 7  [ 22 ] T2DM 6 - 1/08(1/04–1/12) 0/79–1/38 0/0001 1/04(1/00–1/09) 28/9% 0/062 7  [ 22 ] Insulin-resistant diabetes mellitus 3 27,484 263,731 1/43(1/22–1/68) 0/019–5/71 0/0001 1/26(1/22–1/31) 87/8% 0/890 7  [ 236 ] Hyperthyroidism 4 163 5278 2/11(0/23–18/85) Including the null value 0/501 0/22(0/05–0/89) 78/6% 0/709 10  [ 237 ] Family history of miscarriage 2 338 3450 2/31(1/73–3/09) It could not be calculated 0/0001 2/52(1/74–3/66) 0% It could not be calculated 10  [ 238 ] Celiac disease 7 7978 78,241 1/34(1/10–1/65) 0/43–2/45 0/004 1/05(0/98–1/13) 62/6% 0/006 10 Medicines and treatments  [ 239 ] Itraconazole 3 536 929 1/43(0/37–5/45) Including the null value 0/594 1/44(0/15–13/58) 0% 0/428 9  [ 240 ] Pregravid OCP 4 23,295 76,441 0/62(0/41–0/96) 0/011–4/57 0/032 0/97(0/90–1/05) 92/4% 0/203 9  [ 241 ] Ondansetron 13 188,200 3,936,991 0/52(0/31–0/89) 0/025–3/53 0/017 0/82(0/64–1/05) 92/1% 0/634 9  [ 242 ] Preconcep-tion care for women with diabetes 11 869 1836 0/86(0/70–1/06) - - - - - 11  [ 243 ] Proton pump inhibitors 3 - 0/85(0/70–1/90) - - - 0% - 10  [ 244 ] Nonsteroidal anti-inflammatory drug 11 - 1/19(0/92–1/54) 0/46–3/43 0/163 1/10(0/99–1/22) 91/6% 0/54 10  [ 245 ] Levothyroxine in autoimmunity 5 - 0/12(0/07–0/19) - 0/0001 - 62% - 9  [ 246 ] Treating Chronic Hepatitis B Infection - - 0/47(0/11–1/92) - 0/290 - 31% - 9  [ 247 ] Benzodiazepines in the treatment of anxiety 5 - 1/85(1/42–2/42) 0/29–3/68 0/0001 1/40(1/26–1/57) 76/5% 0/174 10  [ 248 ] Amino Acid Salicylic Acid in IBD 5 - 1/14(0/65–2/01) - 0/738 - - - 8  [ 249 ] Triptans in the treatment of migraine and cluster headaches 2 178 50,865 3/54(2/24–5/59) It could not be calculated 0/0001 3/64(2/22–5/99) 0% It could not be calculated 10  [ 241 ] enoxaparin and aspirin verses aspirin for thromboembolism and thromboprophylaxis during pregnancy 2 - 0/53(0/12–2/23) It could not be calculated 0/392 0/38(0/26–0/56) 34% It could not be calculated 10  [ 250 ] Topical retinoids 3 430 945 1/01(0/63–1/62) Including the null value 0/937 1/41(0/77–2/61) 7/7% 0/156 9  [ 251 ] Antipsychotics 4 1306 2482 1/05(0/61–1/81) 0/007–6/03 0/856 0/81(0/59–1/11) 70% 0/741 9  [ 217 ] Radiotherapy for thyroid cancer 6 6078 7270 1/07(0/99–1/16) 0/69–1/50 0/065 1/07(0/98–1/16) 0% 0/304 9  [ 146 ] Colchine in the treatment of familial Mediterranean fever 4 590 1575 0/64(0/45–0/92) 0/03–3/65 0/003 0/57(0/33–0/99) 0% 0/897 9  [ 252 ] Hemodialysis 2 - 0/38(0/12–1/23) It could not be calculated - - 0% It could not be calculated 9  [ 253 ] Infliximab versus adalimumab in inflammatory bowel disease patients 5 193 97 0/60(0/17–2/09) 0/00024–8/42 0/429 0/32(0/03–4/01) 0% 0/727 11  [ 254 ] surgical therapies for inflammatory bowel disease 5 246 530 2/02(1/13–3/60) 0/05–5/68 0/016 1/20(0/54–2/67) 8/4% 0/317 10  [ 255 ] Hyperthyroidism Treatment (Propylthiouracil vs. Methimazole) 4 1734 1654 0/89(0/71–1/12) 0/13–2/81 0/336 0/90(0/71–1/13) 0% 0/592 9  [ 256 ] Amniocentesis 6 - 1/59(0/98–2/59) - 0/06 - - - 10  [ 256 ] Chorionic villus sampling 2 - 1/61(0/36–7/26) It could not be calculated 0/50 - - It could not be calculated 10  [ 257 ] Cervical cerclage 6 904 622 0/42(0/28–0/63) - 0/0001 - 71% - 10  [ 258 ] Long-acting insulin compared to intermediate-acting basal insulin 2 - 1/26(0/85–1/87) It could not be calculated 0/233 1/25(0/84–1/86) 0% It could not be calculated 9  [ 259 ] Insulin lispro vs. regular 3 - 0/76(0/19–3/02) - - - - - 8  [ 260 ] Laser treatment in monochorionic diamniotic twin pregnancies with twin-to-twin transfusion syndrome 4 - 1/56(1/28–1/91) 0/28–3/24 0/0001 1/44(1/05–2/00) 0% 0/355 10  [ 261 ] Laparoscopy vs. laparotomy in removal of uterine adnexal mass 11 299 504 1/53(0/66–3/54) 0/03–5/25 0/311 1/75(0/33–1/15) 0% 0/153 9  [ 262 ] Poor HbA1C control versus optimal control 4 - 3/23(1/64–6/36) 0/0089–8/99 0/001 5/44(2/23–12/45) 28/2% 0/713 8  [ 263 ] Histamine H2 Blockers in the Treatment of Gastroesophageal Reflux 2 738 1575 0/62(0/36–1/05) It could not be calculated - - - It could not be calculated 9  [ 264 ] Metformin in the treatment of polycystic ovary syndrome 4 326 454 0/19(0/12–0/31) 0/003–4/19 0/001 0/23(0/11–0/48) 0% 0/365 8  [ 265 ] Hydroxychloroquine in the treatment of autoimmune diseases 4 363 707 1/85(1/10–3/14) 0/02–6/33 0/021 2/29(1/13–4/65) 12/4 0/436 9  [ 266 ] hysteroscopic septum resection 7 258 184 0/25(0/07–0/88) Including the null value 0/03 1/67(0/94–2/97) 81% 0/007 9  [ 267 ] Radiation therapy during childhood 2 648 1247 1/20(0/83–1/73) It could not be calculated 0/335 1/44(0/99–2/09) 45/6% It could not be calculated 10 Trace elements  [ 268 ] Cadmium 8 361 398 7/95(2/23–28/33) 0/004–15/40 0/001 2/23(1/40–3/93) 93/8% 0/115 9  [ 268 ] Lead 15 765 1218 14/04(4/95–39/76) 0/032–20/10 0/0001 2/33(1/53–3/52) 96% 0/001 9  [ 268 ] Copper 14 655 2108 0/07(0/02–0/19) 0/0001–6/35 0/0001 0/63(0/39–1/01) 95/4% 0/011 9  [ 268 ] Zinc 14 673 2000 0/14(0/04–0/51) Including the null value 0/003 0/22(0/13–0/36) 96/9% 0/309 9 Behavioral and emotional factors  [ 269 ] IUD removal - - 0/51(0/39–0/66) - 0/0001 - 85% 8  [ 270 ] cocaine use/drug free controls 2 - 5/32(0/22–125/9) It could not be calculated - - - It could not be calculated 5  [ 270 ] cocaine polydrug users/drug free controls 2 - 10/56(1/74–64/1) It could not be calculated - - - It could not be calculated 5  [ 270 ] cocaine polydrug users/polydrug no cocaine controls 2 - 3/50(0/64–19/2) It could not be calculated - - - It could not be calculated 5  [ 271 ] Adverse childhood experiences 3 - 1/82(1/47–2/85) - 0/001 - 0% did not exist 10  [ 272 ] intimate partner violence 9 - 2/03(1/25–3/31) - - - - - 9  [ 272 ] Sexual violence in childhood 6 - 1/35(0/75–2/44) - - - - - 9  [ 273 ] Breastfeeding 3 595 1324 0/88(0/33–2/34) Including the null value 0/805 0/46(0/22–0/95) 63/7% 0/281 10  [ 274 ] Pregnancy interval 6 to 11 months 2 - 1/00(0/80–1/24) It could not be calculated - - 0% It could not be calculated 9 Menstrual disorders  [ 30 ] Short menstrual cycle 3 - 1/86(1/10–3/14) Including the null value 0/020 1/74(0/88–3/44) 0% 0/782 11  [ 30 ] Long menstrual cycles 4 - 1/88(1/02–3/47) 0/008–7/32 0/042 0/99(0/53–1/85) 43% 0/050 11 Occupational factors  [ 20 ] Rotating shifts vs. daily shifts 14 - 1/04(0/84–1/29) 0/34–2/15 0/664 0/82(0/68–0/99) 63/8% 0/534 11  [ 20 ] Working more than 40 h 5 - 1/72(1/27–2/34) 0/25–3/62 0/0001 1/80(1/02–3/19) 0% 0/491 11  [ 20 ] Working the night shift 12 - 1/23(1/03–1/46) 0/52–2/07 0/021 1/32(1/04–1/68) 33/2% 0/507 11  [ 275 ] Anesthetic gases 16 11,623 33,146 1/26(0/98–1/63) 0/29–2/70 0/068 0/85(0/71–1/01) 88/4% 0/735 9  [ 275 ] Chemotherapy agents 6 2529 4866 1/20(0/79–1/83) 0/10–3/62 0/371 1/31(0/75–2/28) 32/4% 0/138 9  [ 276 ] Exposure to noise 5 - 1/07(0/99–1/16) 0/63–1/58 0/087 1/10(1/00–1/20) 28/4% 0/874 8  [ 277 ] Hairdresser 7 5413 1173 1/28(1/11–1/47) 0/39–2/46 0/0001 1/31(1/07–1/60) 0% 0/312 8 Nutritional factors  [ 278 ] Coffee consumption 4 1143 2460 2/76(1/64–4/64) 0/02–7/50 0/0001 3/98(2/55–6/21) 63/8% 0/655 9  [ 278 ] Caffeine consumption 11 3341 10,202 2/38(1/76–3/23) 0/47–3/98 0/0001 2/21(1/53–3/19) 63% 0/110 9  [ 279 ] Fruit consumption 3 - 0/46(0/24–0/87) Including the null value 0/017 0/30(0/20–0/30) 86/7% 0/585 11  [ 279 ] Vegetable consumption 3 - 0/58(0/45–0/75) 0/001–6/94 0/0001 0/60(0/40–0/80) 0% 0/535 11  [ 279 ] Fruit and vegetable consumption 3 - 0/81(0/43–1/53) Including the null value 0/530 0/49(0/36–0/66) 79/7% 0/154 11  [ 279 ] Meat consumption 2 - 0/84(0/44–1/61) It could not be calculated 0/608 1/10(0/90–1/30) 74/4% It could not be calculated 11  [ 279 ] Red meat consumption 3 - 0/99(0/84–1/18) 0/01–5/22 0/994 0/98(0/81–1/18) 0% 0/984 11  [ 279 ] White meat consumption 2 - 0/80(0/64–0/99) It could not be calculated 0/046 0/82(0/65–1/02) 0% It could not be calculated 11  [ 279 ] Seafood consumption 4 - 0/81(0/67–0/98) 0/15–2/48 0/036 0/86(0/71–1/03) 35/7% 0/736 11  [ 279 ] Dairy consumption 5 - 0/65(0/51–0/82) 0/13–2/22 0/0001 0/60(0/50–0/80) 44/8% 0/426 11  [ 279 ] Egg consumption 3 - 0/85(0/62–1/15) 0/0002–9/06 0/307 0/70(0/60–0/80) 79/7% 0/743 11  [ 279 ] Fat and oil consumption 2 - 1/07(0/49–2/31) It could not be calculated 0/858 0/73(0/54–0/99) 90/4% It could not be calculated 11  [ 279 ] Sugar consumption 2 - 2/36(0/37–14/88) It could not be calculated 0/361 1/10(0/87–1/40) 78/8 5 It could not be calculated 11  [ 279 ] Cereal consumption 2 - 0/70(0/44–1/09) It could not be calculated 0/118 0/57(0/41–0/79) 63/5% It could not be calculated 11 Serum platelets  [ 15 ] Average platelet volume 18 2168 1985 1/38(0/61–3/15) 0/009–6/37 0/433 1/43(1/07–1/91) 97/1% 0/682 10  [ 15 ] Platelet distribution width 13 2100 2611 6/42(2/50–16/11) 0/03–11/77 0/0001 1/40(1/10–1/76) 98/2% 0/978 10 Including the value for PI: PI was too Wide and included the neutral value Characteristics of studies on miscarriage factors (< 20 weeks) 1750 1989 892 1107 794 2105 609 4005 609 4005 477 3833 477 3833 309 473 201 380 265 2493 Covid19 Vaccine 2525 20,928 1739 3037 146 1512 53 87 3547 46,306 2188 76,638 694 4929 894 3589 234 538 201 720 537 1444 373 1960 573 2254 157 589 3662 27,025 670 1072 465 915 747 3062 213 628 1499 154,223 199 253 775 27,001 321 968 2522 7276 519 584 893 4580 Non-obstetric surgery with no surgery/conservative treatment 2076 9214 Including the value for PI: PI was too Wide and included the neutral value Characteristics of studies on miscarriage factors (< 24 weeks) 3822 1875 1724 2090 Influenza vaccine 8025 47,818 2158 3826 646 8427 174 84 31,775 140,906 52,455 1,710,271 2364 986,732 19,012 23,534 10,731 3450 1632 28,323 102 155 18,706 25,203 Including the value for PI: PI was too Wide and included the neutral value Characteristics of studies on miscarriage factors (< 28 weeks) 20,512 168,691 25,618 87,644 11,934 87,644 2314 23,233 1242 2333 295 275 64 92 Including the value for PI: PI was too Wide and included the neutral value Characteristics of studies on miscarriage factors (definition of miscarriage is not mentioned in the study) 2558 46,876 117 133 100 102 683 1138 1426 11,194 Particulate matter PM2.5 Particulate matter PM10 192,426 232,214 837 1065 1159 4750 2207 1717 4130 3159 4077 2740 155 948 2147 1107 52,503 3,280,744 124 386 274 309 1584 2020 622 836 2373 9838 2166 24,032 1394 20,435 244 253 596 1058 910 1030 343 902 572 694 4397 8,788,494 1870 4614 190 287 595 989 110,872 7,266,388 4806 233,055 27,484 263,731 163 5278 338 3450 7978 78,241 536 929 23,295 76,441 188,200 3,936,991 869 1836 Amino Acid Salicylic Acid in IBD 178 50,865 430 945 1306 2482 6078 7270 590 1575 193 97 246 530 Hyperthyroidism Treatment (Propylthiouracil vs. Methimazole) 1734 1654 904 622 299 504 738 1575 326 454 363 707 258 184 648 1247 361 398 765 1218 655 2108 673 2000 595 1324 11,623 33,146 2529 4866 5413 1173 1143 2460 3341 10,202 2168 1985 2100 2611 Including the value for PI: PI was too Wide and included the neutral value A total of 31 systematic review articles were included in this study, with two studies overlapping between the meta-analysis and the systematic reviews [ 42 , 43 ]. The non-genetic factors examined were organized into four categories based on their association with miscarriage: factors that statistically significantly increased the risk, factors that statistically significantly decreased the risk, factors with no statistically significant effect, and factors with inconsistent results. Among those that significantly increase risk, examples included anti-Mullerian hormone ≤ 4 [ 44 ], PAPP (pregnancy- associated protein A)[ 45 ], electro-magnetic fields greater than 16 milligauss [ 46 ], unclean fuel and fuel cooking [ 47 ], myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) [ 48 ], Low-calorie sweeteners [ 21 ], barium [ 49 ], mycotoxin [ 50 ], streptococcus agalactiae and lactobacil syphilis [ 51 ], Plague medicines)doxycycline and trimethoprim/sulfamethoxazole TMP-SMX) [ 52 ], refugee women [ 53 ], Exhaust gases from waste incinerators [ 54 ], First sexual intercourse at a young age(< 18) [ 55 ], mirtazapine drug [ 56 ], periodontal disease [ 57 ], Disinfection byproducts of trihalomethanes [ 58 ], radiation ionizing and Ethylene oxide [ 27 ] and Ambient Air Pollution [ 59 ] Those that significantly decreased risk rubidium and copper [ 49 ], ultrasound [ 27 ]. Factors with no statistically significant effects encompassed albendazole or mebendazole [ 60 ], autoimmune hepatitis [ 61 ], soluble fms-like tyrosine kinase (SFLT1) [ 45 ], hepadnavirus (HEDB) [ 51 ], pertussis vaccine [ 62 ], physical activity in Athletes [ 63 ], Terbinafine [ 64 ], vaping versus nicotine replacement therapy [ 65 ], anti TNF in inflammatory Bowel Diseases [ 66 ], Atrazine [ 67 ], Glyphosate [ 68 ], Disinfectants [ 27 ], Tenofovir [ 69 ], Leukemia [ 43 ], Enoxaparin versus no treatment [ 42 ]. Finally, inconsistent results were observed for exercise [ 70 ]. Out of 147 included meta-analysis studies, 141 studies were high-quality and 6 studies were moderate quality and none of them was categorized as low quality. According to Table 1 , 2 , 3 , 4 , in 55-unit meta-analyses the number of cases was more than 1000. 62 meta-analyses had P -values less than 0.001. In 123 of 227 meta-analyses, the heterogeneity was not large, and in 28 meta-analyses there was evidence of a small study effect. Also, there was excessive significance in 8 meta-analyses. In this study, 147 meta-analyses including 227 meta-analysis components, were analyzed. None of the factors met all the necessary criteria and showed a definite and unbiased association miscarriage and were included in the strong evidence group. 45 factors were included in the suggestive evidence group and respectively 60 and 109 factors were included in the weak evidence and lack of evidence group (non-significant associations). In 13 factors, there was not enough information to classify them into the aforementioned categories Table 5 , 6 , 7 , 8 . The 45 factors were included in the Suggestive evidence group, were summarized in Table  9 . Among the 45 factors, which were included in the suggestive-evidence group, three factors—hypothyroidism, high serum β-human chorionic gonadotropin, and hairdressers—contained the most complete information (each with more than 1000 cases and statistically significant under the random-effects model, p  < 0.001, with effect sizes of 2.01 (1.66–2.44), 2.88 (1.76–4.68), and 1.28 (1.11–1.47), respectively). Table 5 Assessment the associations between non-genetic factors with miscarriage (< 20 weeks) Non-genetic factors Sample size (number of cases) Significance threshold reached (under the random-effects model) 95% prediction interval rule Estimate of heterogeneity Small-study effects or excess significance bias Random-effects summary effect size (95% CI) score Associations supported by suggestive evidence Statin  > 500 but < 1000   0·001 Including the null value Not large Neither 1/48(1/11–1/96) high hypothyroidism  > 1000  < 0·001 Including the null value Not large Neither 2/01(1/66–2/44) high Smallpox vaccine  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 4/82(2/38–9/76) high Ulcerative colitis  < 500   0·001 Including the null value Not large Neither 1/91(1/06–3/43) high Chorionic bump  < 500   500 but < 1000  < 0·001 Including the null value Not large Neither 1/56(1/29–1/89) high MEHHP  < 500  < 0·001 Including the null value Not large Neither 1/59(1/22–2/06) high MEOHP  < 500   0·001 Including the null value Not large Neither 1/47(1/14–1/89) high DEHP  < 500   500 but < 1000   1000   0·001 Including the null value Very large Neither 1/84(1/03–3/27) high Oil and natural gas NA   0·001 Including the null value Very large Neither a 1/82(1/13–2/94) High Congenital abnormalities of the uterus  > 1000   500 but < 1000   0·001 Including the null value large Small-study effects 2/00(1/23–3/23) High Antibiotics NA  < 0·001 Including the null value Very large Small-study a effects 1/37(1/23–1/53) High Macrolides NA   0·001 Including the null value Very large Neither a 1/41(1/04–1/92) High Quinolones NA  < 0·001 Including the null value Very large Neither a 2/48(1/46–4/20) High Tetracycline NA  < 0·001 Including the null value Not large Small-study a effects 2/56(1/95–3/37) High Clarithromycin NA  < 0·001 Including the null value large Neither a 1/98(1/45–2/70) High Crohn's  < 500   500 but < 1000   0·001 Including the null value Not large Small-study effects 1/7(1/17–3/00) High Vedolizumab in the treatment of IBD  < 500   0·001 Including the null value Not large Small-study effects 1/79(1/06–3/01) High Lithium  < 500   0·001 NA Very large NA 3/77(1/15–12/39) High Lifting weights ≥ 11 kg NA   0·001 Including the null value Very large NA 1/31(1/08–1/58) High Non-significant associations High platelet to lymphocyte ratio in blood serum  > 500 but  0·05 Including the null value Very large Neither 1/66(0/65–4/23) High Organic Solvent  > 500 but  0·05 NA NA NA 1/25(0/99–1/58) High Covid19 Vaccine  > 1000  > 0·05 Including the null value Not large Neither 1/08(0/90–1/29) High Chlamydia trachomatis IgA   0·05 NA large NA 1/23(0/56–2/71) High Portal cirrhosis high blood pressure   0·05 Including the null value Not large Neither 1/61(0/62–4/14) High Anorexia nervosa   0·05 Including the null value Not large Neither 1/24(0/95–1/63) High Bulimia nervosa  > 500 but  0·05 Including the null value Not large Small-study effects 1/60(0/82–3/10) High MBP  > 500 but  0·05 Including the null value Not large Neither 1/39(0/99–1/95) High MECPP   0·05 Including the null value Not large Neither 1/10(0/77–1/57) High Fluconazole  > 1000  > 0·05 Including the null value Not large Small-study effects 1/08(0/91–1/29) High Artemisinin in the treatment of malaria  > 500 but  0·05 NA NA NA 0/74(0/47–1/17) High Artemether lumefantrine/Quinine   0·05 NA NA NA 0/67(0/37–1/23) High Sulfonamides NA  > 0·05 Including the null value Very large Excess significance bias 1/44(0/90–2/92) High Penicillin NA  > 0·05 Including the null value Very large Neither a 0/97(0/83–1/12) High Cephalosporins NA  > 0·05 Including the null value Not large Neither 0/85(0/65–1/12) High Erythromycin NA  > 0·05 It could not be calculated 0/88(0/61–1/28) Small-study could not be calculated a 0/88(0/61–1/28) High Amoxicillin NA  > 0·05 Including the null value Very large Neither 0/83(0/66–1/04) High Sulfamethizole NA  > 0·05 Including the null value large Neither 1/21(0/81–1/80) High Tumor necrosis factor alpha inhibitors in arthritic diseases   0·05 Including the null value Not large Small-study effects 1/28(0/61–2/70) High Montelukast for asthma  > 500 but  0·05 Including the null value large Small-study effects 1/14(0/54–2/37) High H1 antihistamines in the control of nausea and vomiting  > 1000  > 0·05 Including the null value Not large Neither 0/99(0/82–1/20) High Treatment for cervical intraepithelial neoplasia  > 500 but  0·05 Including the null value Not large Neither 1/16(0/79–1/69) High Alcohol drinking  > 1000  > 0·05 Including the null value large Neither 1/10(0/90–1/34) High Lifting ≥ 100 kg NA  > 0·05 Including the null value large Neither a 1/19(0/82–1/73) High Prolonged standing NA  > 0·05 Including the null value Not large Neither a 1/06(0/92–1/22) High Heavy workload NA  > 0·05 Including the null value Very large Neither a 1/49(0/91–2/45) High Bending ≥ 1 h NA  > 0·05 Including the null value Very large Neither a 1/32(0/82–2/12) High Associations not adequately assessed owing to absence of data Chlamydia trachomatis IgG  > 1000 NA NA Not large NA 1/60(1/24–2/07) High Metronidazole  > 1000 NA NA Not large NA 1/60(1/24–2/07) High Non-obstetric surgery with no surgery/conservative treatment NA NA NA NA NA 1/1(1/1–1/2) High Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers NA  < 0·001 NA Not large NA 1/72(1/37–2/17) High parvovirus B19 infection NA   500 but < 1000 NA NA NA NA 1/60(1/22–2/22) High a Excess statistical significant bias could not be calculated. NA: Not applied Table 6 Assessment the associations between non-genetic factors with miscarriage (< 24 weeks) Non-genetic factors Sample size (number of cases) Significance threshold reached (under the random-effects model) 95% prediction interval rule Estimate of heterogeneity Small-study effects or excess significance bias Random-effects summary effect size (95% CI) Score Associations supported by suggestive evidence Serum vitamin D deficiency  > 1000   0·001 Including the null value Not large Neither 1/78(1/11–2/87) High Herpes simplex virus 1/2 NA  < 0·001 Including the null value Not large Neither a 3/72(1/86–7/46) High Psychological stress NA   1000   0·001 Including the null value Not large Neither 1/41(1/11–1/80) High Selective norepinephrine reuptake inhibitors versus antidepressants  > 1000   0·001 Including the null value Not large Neither 1/19(1/02–1/39) High Associations supported by weak evidence Sub chronic hematoma  > 500 but < 1000   0·001 Including the null value large Neither 2/18(1/29–3/68) High Endometriosis  > 1000  < 0·001 Including the null value Very large Neither 1/74(1/28–2/37) High Chronic histiocytic intervillositis (CHI)  < 500   0·001 Including the null value large Neither 0/16(0/03–0/85) High Antidepressant NA   1000   1000   0·001 Including the null value large Neither 1/16(1/03–1/30) High Immunosuppressive therapy in Neuromyelitis Optica  < 500   1000  > 0·05 Including the null value Not large Neither 1/05(0/95–1/17) High Influenza vaccine  > 1000  > 0·05 Including the null value Not large Neither 1/04(0/71–1/52) High Covid-19 virus  > 1000  > 0·05 Including the null value Not large Neither 1/09(0/81–1/48) High Human cytomegalovirus infection NA  > 0·05 Including the null value Not large Neither a 2/12(0/98–4/61) High Multiple sclerosis medications NA  > 0·05 Including the null value Not large Neither a 1/14(0/98–1/32) High Interferon(MS medication) NA  > 0·05 Including the null value Not large Neither a 1/13(0/76–1/69) High Natalizumab (MS medication) NA  > 0·05 It could not be calculated large Small-study could not be calculated a 1/93(0/96–3/85) High Glatiramer acetate (MS medication) NA  > 0·05 Including the null value Not large Neither a 1/11(0/65–1/91) High Selective serotonin reuptake inhibitors vs. antidepressants  > 1000  > 0·05 Including the null value Not large Neither 1/03(0/92–1/15) High Interval between pregnancies less than 6 months  > 1000  > 0·05 Including the null value Not large Neither 0/85(0/71–1/03) High a Excess statistical significant bias could not be calculated. NA: Not applied Table 7 Assessment the associations between non-genetic factors with miscarriage (< 28 weeks) Non-genetic factors Sample size (number of cases) Significance threshold reached (under the random-effects model) 95% prediction interval rule Estimate of heterogeneity Small-study effects or excess significance bias Random-effects summary effect size (95% CI) Score Associations supported by suggestive evidence Overweight  > 1000   0·001 Including the null value Not large Neither a 1/07(1/02–1/12) Moderate Obesity  > 1000   1000  < 0·001 Including the null value Not large Neither a 1/50(1/32–1/70) Moderate PCOs  < 500  < 0·001 Including the null value Not large Neither 2/86(1/64–4/98) High Inhaled corticosteroid-s (anti-inflammatory and immunosuppressive medication) NA  < 0·001 Including the null value Not large Neither a 1/20(1/15–1/24) High Associations supported by weak evidence Smoking NA  < 0·001 Including the null value large Small-study a effects 1/22(1/16–1/29) High Non-significant associations Second-hand smoke NA   1000  > 0·05 Including the null value Not large Neither a 1/08(0/97–1/20) High Underweight  > 1000  > 0·05 Including the null value large Neither a 1/08(0/98–1/20) High Systemic corticosteroids (anti-inflammatory and immunosuppressive drugs) NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 1/41(0/86–2/32) High Endocervical versus decidual polypectomy   0·05 It could not be calculated Not large Small-study could not be calculated 0/26(0/04–1/58) High pregnancy intention NA  > 0·05 It could not be calculated Very large Small-study could not be calculated a 1/54(1/78–3/03) High Associations not adequately assessed owing to absence of data Rheumatoid arthritis NA   0·001 NA Not large NA 1/32(1/21–1/43) High a Excess statistical significant bias could not be calculated. NA: Not applied Table 8 Assessment the associations between non-genetic factors with miscarriage (definition of miscarriage is not mentioned in the study) Non-genetic factors Sample size (number of cases) Significance threshold reached (under the random-effects model) 95% prediction interval rule Estimate of heterogeneity Small-study effects or excess significance bias Random-effects summary effect size (95% CI) Score Associations supported by suggestive evidence Gamma-aminobutyric acid analog NA   1000  < 0·001 Including the null value Not large Neither 2/88(1/76–4/68) High hyperhomocysteinemia  < 500   0·001 Including the null value Not large Neither 0/53(0/33–0/86) High Serum Vitamin B12 deficiency  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 3/00(1/70–5/28) High fabrication NA   500 but < 1000   0·001 Including the null value Not large Neither 1/26(1/01–1/57) High Primary antiphospholipid syndrome  < 500   0·001 Including the null value Not large Neither 2/47(1/11–5/47) High T1DM  > 1000   0·001 Including the null value Not large Neither 1/18(1/01–1/37) Moderate Family history of miscarriage  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 2/31(1/73–3/09) High Triptans in the treatment of migraine and cluster headaches  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 3/54(2/24–5/59) High Adverse childhood experiences NA  < 0·001 NA Not large Neither a 1/82(1/47–2/85) High Metformin in the treatment of polycystic ovary syndrome  < 500   0·001 Including the null value Not large Neither 0/19(0/12–0/31) High Hydroxychloroquine in the treatment of autoimmune diseases  < 500   0·001 Including the null value Not large Neither 1/85(1/10–3/14) High Colchine in the treatment of familial Mediterranean fever  < 500   0·001 Including the null value Not large Neither 0/64(0/45–0/92) High surgical therapies for inflammatory bowel disease  < 500   0·001 Including the null value Not large Neither 2/02(1/13–3/60) High Laser treatment in monochorionic diamniotic twin pregnancies with twin-to-twin transfusion syndrome NA  < 0·001 Including the null value Not large Neither a 1/56(1/28–1/91) High Poor HbA1C control versus optimal control NA  < 0·001 Including the null value Not large Neither a 3/23(1/64–6/36) High Vegetable consumption NA  < 0·001 Including the null value Not large Neither a 0/58(0/45–0/75) High Seafood consumption NA   0·001 Including the null value Not large Neither a 0/81(0/67–0/98) High Dairy consumption NA  < 0·001 Including the null value Not large Neither a 0/65(0/51–0/82) High White meat consumption NA   0·001 It could not be calculated Not large Small-study could not be calculated a 0/80(0/64–0/99) High Short menstrual cycle NA   0·001 It could not be calculated Not large Neither a 1/86(1/10–3/14) High Hairdresser  > 1000  < 0·001 Including the null value Not large Neither 1/28(1/11–1/47) High Working more than 40 h NA  < 0·001 Including the null value Not large Neither a 1/72(1/27–2/34) High Working the night shift NA   0·001 Including the null value Not large Neither a 1/23(1/03–1/46) High Associations supported by weak evidence T2DM NA   1000  < 0·001 Including the null value Very large Neither 4/29(2/22–8/27) Moderate Photo-lithography work NA   0·001 Including the null value Very large Small-study effects a 1/37(1/10–1/71) High Particulate matter PM2.5 NA   0·001 Including the null value Very large Neither a 1/21(1/02–1/42) High Particulate matter PM10 NA   0·001 Including the null value Very large Neither a 1/09(1/02–1/17) High Pesticides  > 1000   0·001 Including the null value Very large Neither 1/41(1/10–1/80) High Antineoplastic agents NA   0·001 Including the null value Very large Neither a 1/56(1/16–2/09) High Serum antibodies against Toxoplasma gondii  > 1000   1000   1000   1000   1000   0·001 Including the null value Very large Neither 0/62(0/41–0/96) High Ondansetron  > 1000   0·001 Including the null value Very large Neither 0/52(0/31–0/89) High Benzodiazepines in the treatment of anxiety NA  < 0·001 Including the null value Very large Neither a 1/85(1/42–2/42) High Cadmium  < 500   0·001 Including the null value Very large Neither 7/95(2/23–28/33) High Lead  > 500 but < 1000   500 but < 1000   500 but < 1000   0·001 Including the null value Very large Neither 0/14(0/04–0/51) High IUD removal NA  < 0·001 NA Very large NA 0/51(0/39–0/66) High Levothyroxine in autoimmunity NA   1000   0·001 Including the null value Not large Small-study effects 1/80(1/28–2/52) High Diabetes Mellitus  > 1000  < 0·001 Including the null value Very large Neither 1/23(1/13–1/33) Moderate Uterine fibroids NA  < 0·001 NA Very large NA 4/51(2/80–7/26) High Sjögren's syndrome  < 500   1000   500 but < 1000   0·001 Including the null value large Neither 1/30(1/04–1/62) High Vaginal bleeding NA NA NA Very large NA 4/30(2/00–9/00) High Gestational diabetes  > 1000   1000   1000   0·001 Including the null value large both 1/34(1/10–1/65) High hysteroscopic septum resection  < 500   0·001 Including the null value Very large Small-study effects 0/25(0/07–0/88) High Coffee consumption  > 1000   1000  < 0·001 Including the null value large Neither 2/38(1/76–3/23) High Fruit consumption NA   0·001 Including the null value Very large Neither a 0/46(0/24–0/87) High Cervical cerclage  > 500 but < 1000  < 0·001 Including the null value large Neither 0/42(0/28–0/63) High Long menstrual cycles NA   0·001 NA Not large Small-study effects a 1/88(1/02–3/47) High Platelet distribution width  > 1000   1000  > 0·05 Including the null value Not large Neither 0/89(0/71–1/12) High Hyperthyroidism   0·05 Including the null value Very large Neither 2/11(0/23–18/85) High Serum Folic acid deficiency  > 500 but  0·05 Including the null value Very large Neither 0/78(0/48–1/27) High Ethylene glycol ether NA  > 0·05 Including the null value large Neither a 1/42(0/94–2/16) High Fluoride exposure NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 1/19(0/73–1/95) High Perfluoro-decanoate NA  > 0·05 Including the null value large Neither a 1/10(0/81–1/53) High Perfluoro-hexane sulfonate NA  > 0·05 Including the null value Not large Neither a 1/04(0/96–1/12) High Perfluoro-nonanoate NA  > 0·05 Including the null value Very large Neither a 0/86(0/70–1/06) High Perfluoro-octanoate sulphonate NA  > 0·05 Including the null value Not large Small-study effects a 1/00(0/88–1/12) High Perfluoro-octane sulfonic acid NA  > 0·05 Including the null value Not large Neither a 1/06(0/94–1/21) High HPV Vaccine  > 1000  > 0·05 Including the null value large Neither 0/98(0/85–1/13) High Exposure to dioxins in food  > 500 but  0·05 It could not be calculated Not large Small-study could not be calculated 1/04(0/79–1/37) High HIV infection   0·05 It could not be calculated Not large Small-study could not be calculated 1/35(0/77–2/35) High Oral cholera vaccine  > 1000  > 0·05 It could not be calculated Not large Small-study could not be calculated 1/54(0/78–3/03) High Neisseria gonorrhoeae NA  > 0·05 Including the null value large Small-study effects a 2/11(0/50–8/92) High HPV   0·05 Including the null value large Neither 0/99(0/37–2/64) High Helicobacter pylori  > 1000  > 0·05 Including the null value Not large Neither 1/50(1/0–2/14) High hepatitis B virus infection NA  > 0·05 It could not be calculated Very large Small-study could not be calculated a 1/09(0/51–2/29) High hepatitis E virus infection   0·05 Including the null value Not large Small-study effect 1/71(0/79–3/69) High Laparoscopy vs. laparotomy in removal of uterine adnexal mass   0·05 Including the null value Not large Neither 1/53(0/66–3/54) High Histamine H2 Blockers in the Treatment of Gastroesophageal Reflux  > 500 but  0·05 It could not be calculated NA Small-study could not be calculated a 0/62(0/36–1/05) High Breastfeeding  > 500 but  0·05 Including the null value Not large Neither 0/88(0/33–2/34) High Dengue infection NA  > 0·05 Including the null value Not large Neither 1/90(0/95–3/78) High Breast Cancer NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 1/04(0/86–1/26) High Leiomyomas NA  > 0·05 Including the null value Not large Neither 0/86(0/72–1/03) High Hodgkin lymphoma NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 0/78(0/55–1/10) High Non-Hodgkin lymphoma NA  > 0·05 It could not be calculated large Small-study could not be calculated a 0/94(0/62–1/43) High Psoriasis NA  > 0·05 Including the null value Very large Neither a 1/62(0/99–2/63) High Lyme disease NA  > 0·05 Including the null value Not large Neither a 1/27(0/42–3/85) High Uterine artery embolization   0·05 Including the null value Not large Neither 0/86(0/72–1/03) High Familial Mediterranean fever  > 500 but  0·05 Including the null value Very large Neither 0/94(0/71–1/24) High Hyperandrogenic  > 500 but  0·05 Including the null value Not large Neither 1/47(0/92–2/35) High Nausea and vomiting during pregnancy NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 0/40(0/12–1/27) Moderate Itraconazole NA  > 0·05 Including the null value Not large Neither a 1/43(0/37–5/45) High Preconcep-tion care for women with diabetes  > 500 but < 1000 NA NA NA NA 0/86(0/70–1/06) High Proton pump inhibitors NA NA NA Not large NA 0/85(0/70–1/90) High Sexual violence in childhood NA NA NA NA NA 1/35(0/75–2/44) High Pregnancy interval 6 to 11 months NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 1/00(0/80–1/24) High Nonsteroidal anti-inflammatory drug NA  > 0·05 Including the null value Very large Neither a 1/19(0/92–1/54) High Amino Acid Salicylic Acid in IBD NA  > 0·05 NA NA NA 1/14(0/65–2/01) high Treating Chronic Hepatitis B Infection NA  > 0·05 NA Not large NA 0/47(0/11–1/92) High Topical retinoids   0·05 Including the null value Not large Neither 1/01(0/63–1/62) High Antipsychotics  > 1000  > 0·05 Including the null value large Neither 1/05(0/61–1/81) High Radiotherapy for thyroid cancer  > 1000  > 0·05 Including the null value Not large Neither 1/07(0/99–1/16) High Hemodialysis NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 0/38(0/12–1/23) High Infliximab versus adalimumab in inflammatory bowel disease patients   0·05 Including the null value Not large Neither 0/60(0/17–2/09) High Insulin lispro vs. regular NA NA NA NA NA 0/76(0/19–3/02) High Amniocentesis NA  > 0·05 NA NA NA 1/59(0/98–2/59) High Chorionic villus sampling NA  > 0·05 NA NA NA 1/61(0/36–7/26) High Rotating shifts vs. daily shifts NA  > 0·05 Including the null value large Neither a 1/04(0/84–1/29) High Anesthetic gases  > 1000  > 0·05 Including the null value Very large Neither 1/26(0/98–1/63) High Chemotherapy agents  > 1000  > 0·05 Including the null value Not large Neither 1/20(0/79–1/83) High Exposure to noise NA  > 0·05 Including the null value Not large Neither a 1/07(0/99–1/16) High Meat consumption NA  > 0·05 It could not be calculated large Small-study could not be calculated a 0/84(0/44–1/61) High Red meat consumption NA  > 0·05 Including the null value Not large Neither 0/99(0/84–1/18) High Egg consumption NA  > 0·05 Including the null value Very large Neither 0/85(0/62–1/15) High Fat and oil consumption NA  > 0·05 It could not be calculated Very large Small-study could not be calculated a 1/07(0/49–2/31) High Sugar consumption NA  > 0·05 It could not be calculated Very large Small-study could not be calculated a 2/36(0/37–14/88) High Cereal consumption NA  > 0·05 It could not be calculated large Small-study could not be calculated a 0/70(0/44–1/09) High Average platelet volume  > 1000  > 0·05 Including the null value Very large Neither 1/38(0/61–3/15) High Radiation therapy during childhood  > 500 but  0·05 Including the null value Not large Small-study could not be calculated 1/20(0/83–1/73) High enoxaparin and aspirin verses aspirin for thromboembolism and thromboprophylaxis during pregnancy NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 0/53(0/12–2/23) High Long-acting insulin compared to intermediate-acting basal insulin NA  > 0·05 It could not be calculated Not large Small-study could not be calculated a 1/26(0/85–1/87) High cocaine polydrug users/polydrug no cocaine controls NA NA It could not be calculated Not large Small-study could not be calculated a 3/50(0/64–19/2) moderate Fruit and vegetable consumption NA  > 0·05 Including the null value Very large Neither* 0/81(0/43–1/53) High Associations not adequately assessed owing to absence of data History of recurrent miscarriage  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated a 6/37(3/83–10/57) High Depression NA NA It could not be calculated Not large Small-study could not be calculated a 1/88(1/50–2/35) High Anxiety NA NA It could not be calculated Not large Small-study could not be calculated a 1/25(1/11–1/40) High Non-alcoholic fatty liver  > 1000 NA It could not be calculated Not large Small-study could not be calculated a 1/15(1/02–1/30) High cocaine polydrug users/drug free controls NA NA It could not be calculated NA Small-study could not be calculated a 10/56(1/74–64/1) moderate intimate partner violence NA NA NA NA NA 2/03(1/25–3/31) High a Excess statistical significant bias could not be calculated. NA: Not applied Table 9 45 factors were included in the suggestive evidence group Non-genetic factors Sample size (number of cases) Significance threshold reached (under the random-effects model) 95% prediction interval rule Estimate of heterogeneity Small-study effects or excess significance bias Random-effects summary effect size (95% CI) Associations supported by suggestive evidence 1.Statin  > 500 but < 1000   0·001 Including the null value Not large Neither 1/48(1/11–1/96) 2.hypothyroidism  > 1000  < 0·001 Including the null value Not large Neither 2/01(1/66–2/44) 3.Smallpox vaccine  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 4/82(2/38–9/76) 4.Ulcerative colitis  < 500   0·001 Including the null value Not large Neither 1/91(1/06–3/43) 5.Chorionic bump  < 500   500 but < 1000  < 0·001 Including the null value Not large Neither 1/56(1/29–1/89) 7.MEHHP  < 500  < 0·001 Including the null value Not large Neither 1/59(1/22–2/06) 8.MEOHP  < 500   0·001 Including the null value Not large Neither 1/47(1/14–1/89) 9.DEHP  < 500   500 but < 1000   1000   0·001 Including the null value Not large Neither 1/78(1/11–2/87) 12.Herpes simplex virus 1/2 NA  < 0·001 Including the null value Not large Neither a 3/72(1/86–7/46) 13.Psychological stress NA   1000   0·001 Including the null value Not large Neither 1/41(1/11–1/80) 15.Selective norepinephrine reuptake inhibitors versus antidepressants  > 1000   0·001 Including the null value Not large Neither 1/19(1/02–1/39) 16.Overweight  > 1000   0·001 Including the null value Not large Neither a 1/07(1/02–1/12) 17.Obesity  > 1000   1000  < 0·001 Including the null value Not large Neither a 1/50(1/32–1/70) 19.PCOs  < 500  < 0·001 Including the null value Not large Neither 2/86(1/64–4/98) 20.Inhaled corticosteroid-s (anti-inflammatory and immunosuppressive medication) NA  < 0·001 Including the null value Not large Neither a 1/20(1/15–1/24) 21.Gamma-aminobutyric acid analog NA   1000  < 0·001 Including the null value Not large Neither 2/88(1/76–4/68) 23.hyperhomocysteinemia  < 500   0·001 Including the null value Not large Neither 0/53(0/33–0/86) 24.Serum Vitamin B12 deficiency  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 3/00(1/70–5/28) 25.fabrication NA   500 but < 1000   0·001 Including the null value Not large Neither 1/26(1/01–1/57) 27.Primary antiphospholipid syndrome  < 500   0·001 Including the null value Not large Neither 2/47(1/11–5/47) 28.T1DM  > 1000   0·001 Including the null value Not large Neither 1/18(1/01–1/37) 29.Family history of miscarriage  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 2/31(1/73–3/09) 30.Triptans in the treatment of migraine and cluster headaches  < 500  < 0·001 It could not be calculated Not large Small-study could not be calculated 3/54(2/24–5/59) 31.Adverse childhood experiences NA  < 0·001 NA Not large Neither a 1/82(1/47–2/85) 32.Metformin in the treatment of polycystic ovary syndrome  < 500   0·001 Including the null value Not large Neither 0/19(0/12–0/31) 33.Hydroxychloroquine in the treatment of autoimmune diseases  < 500   0·001 Including the null value Not large Neither 1/85(1/10–3/14) 34.Colchine in the treatment of familial Mediterranean fever  < 500   0·001 Including the null value Not large Neither 0/64(0/45–0/92) 35.surgical therapies for inflammatory bowel disease  < 500   0·001 Including the null value Not large Neither 2/02(1/13–3/60) 36.Laser treatment in monochorionic diamniotic twin pregnancies with twin-to-twin transfusion syndrome NA  < 0·001 Including the null value Not large Neither a 1/56(1/28–1/91) 37.Poor HbA1C control versus optimal control NA  < 0·001 Including the null value Not large Neither a 3/23(1/64–6/36) 38.Vegetable consumption NA  < 0·001 Including the null value Not large Neither a 0/58(0/45–0/75) 39.Seafood consumption NA   0·001 Including the null value Not large Neither a 0/81(0/67–0/98) 40.Dairy consumption NA  < 0·001 Including the null value Not large Neither a 0/65(0/51–0/82) 41.White meat consumption NA   0·001 It could not be calculated Not large Small-study could not be calculated a 0/80(0/64–0/99) 42.Short menstrual cycle NA   0·001 It could not be calculated Not large Neither a 1/86(1/10–3/14) 43.Hairdresser  > 1000  < 0·001 Including the null value Not large Neither 1/28(1/11–1/47) 44.Working more than 40 h NA  < 0·001 Including the null value Not large Neither a 1/72(1/27–2/34) 45.Working the night shift NA   0·001 Including the null value Not large Neither a 1/23(1/03–1/46) a Excess statistical significant bias could not be calculated. NA: Not applied Assessment the associations between non-genetic factors with miscarriage (< 20 weeks) Covid19 Vaccine a Excess statistical significant bias could not be calculated. NA: Not applied Assessment the associations between non-genetic factors with miscarriage (< 24 weeks) a Excess statistical significant bias could not be calculated. NA: Not applied Assessment the associations between non-genetic factors with miscarriage (< 28 weeks) a Excess statistical significant bias could not be calculated. NA: Not applied Assessment the associations between non-genetic factors with miscarriage (definition of miscarriage is not mentioned in the study) Particulate matter PM2.5 Particulate matter PM10 Hyperthyroidism Treatment (Propylthiouracil vs. Methimazole) Amino Acid Salicylic Acid in IBD a Excess statistical significant bias could not be calculated. NA: Not applied 45 factors were included in the suggestive evidence group a Excess statistical significant bias could not be calculated. NA: Not applied

This study is the first comprehensive review to investigate non-genetic factors affecting the risk of miscarriage. In this study, non-genetic factors were examined in 147 meta-analyses and 29 systematic reviews. Only 45 non-genetic factors had statistically significant associations without bias and heterogeneity, but were classified in the associations supported by suggestive evidence group because the prediction interval (95%) included the null value. The results of this study can guide health policymakers and professionals in this field to provide necessary suggestions regarding miscarriage.

This is the first and most comprehensive umbrella review to examine non-genetic factors affecting miscarriage. In this study, all 227 non-genetic factors extracted from meta-analyses were grouped into Chemical biomarkers, Environmental factors, Microbial, infectious agents and vaccination, Diseases, Medicines and treatments, Behavioral and Emotional factors, Occupational exposures, Semiconductor factors, Nutritional factors, Menstrual factors, Trace elements and Serum platelets. No study met all the criteria for inclusion in the strong evidence group. Studies related to 45 non-genetic factors were classified as associations supported by suggestive evidence because they contained a null value in the prediction interval (95%) and had a number of cases less than 1000. If the predictive interval includes the null value, it means that some risk factors that were significant in the present study may not be significant in other studies. Studies related to 61 and 102 non-genetic factors were classified as weak and lack of evidence, respectively. In this study, in 27 meta-analyses, there was evidence of a small study effect. Also, there was excessive significance in 9 meta-analyses. Also, in 100 of 227 meta-analyses, the heterogeneity was large and very large. Heterogeneity in this study can be caused by various reasons, including the integration of cohort and case–control studies in a meta-analysis, or geographic and regional differences in the incidence of abortion due to differences in exposure to risk factors. In this study, of the 45 non-genetic factors classified in the group of associations supported by suggestive evidence, statins, hypothyroidism, levothyroxine in the treatment of hypothyroidism, ulcerative colitis, urinary phthalates and chorionic bump were associated with miscarriages before 20 weeks. Serum vitamin D deficiency, selective norepinephrine reuptake inhibitors and selective norepinephrine reuptake inhibitors versus antidepressants were associated with miscarriages before 24 weeks, and polycystic ovary syndrome was associated with miscarriages before 28 weeks. Also, increased human chorionic gonadotropin, decreased serum homocysteine, and serum vitamin B12 deficiency, primary biliary cholangitis, type 1 diabetes, metformin, hydroxychloroquine, colchicine, and surgery for inflammatory bowel disease were associated with miscarriages of uncertain definition. Next, we reviewed the evidence for the 45 non-genetic factors that were categorized as associations supported by suggestive evidence, in relation to miscarriage risk. Statins are drugs that inhibit the enzyme HMG-CoA reductase in the treatment of hypercholesterolemia and prevent complications from cardiovascular diseases [ 71 ]. Not taking lipid-lowering drugs during pregnancy may cause premature birth, preeclampsia, and fetal growth restriction [ 72 – 74 ]. During pregnancy, the fetus requires maternal cholesterol, and maternal HMG-CoA reductase activity is important for placental development [ 75 , 76 ]. From a physiological perspective, statins have adverse effects on the implantation process [ 75 , 76 ]. Studies have shown conflicting results regarding the association between statins and miscarriage. Some studies have found an increased risk, while others have found no association [ 77 – 79 ]. This umbrella review found that statins increase the risk of miscarriage and was classified as suggestive evidence because there were no small-study effects and excess significance bias, and there was not large heterogeneity between studies. Phthalates are chemicals that are widely found in a variety of products, such as clothing and cosmetics, and even in the environment, such as soil, dust, fresh water, and sewage. Therefore, humans are constantly exposed to this group of chemicals [ 80 ]. Studies have shown that high levels of phthalates are present in the urine samples of pregnant women [ 81 , 82 ]. Some studies showed a positive and statistically significant result between increased urinary phthalates and the risk of miscarriage [ 83 , 84 ], while this association was not seen in the study by Liao et al. [ 85 ]. In a recent meta-analysis study included in the present umbrella review, it has been shown that there is a statistically significant relationship between increased urinary phthalates (MEHP, MEHHP, MEOHP, and DEHP) and the risk of miscarriage. Thyroid hormone levels fluctuate during pregnancy due to various changes in endocrine and metabolic functions.[ 86 ]. Hypothyroidism is the most common thyroid disorder during pregnancy, characterized by elevated thyroid-stimulating hormones with normal levels of free thyroxine [ 87 , 88 ]. Studies have shown conflicting results regarding hypothyroidism and the risk of miscarriage [ 86 , 89 – 91 ]. The results of the meta-analysis study on hypothyroidism in this umbrella review showed that hypothyroidism is associated with an increased risk of miscarriage. Levothyroxine is commonly used in the treatment of subclinical hypothyroidism during pregnancy [ 92 ]. Several studies have investigated the effect of SCH. Still, these studies have clinical problems because multiple diagnostic criteria have been defined for subclinical hypothyroidism during pregnancy, which may affect the outcome of levothyroxine treatment. Only one systematic review has used the new diagnostic criteria [ 93 ], but due to methodological problems, definitive conclusions cannot be drawn. The results of the present study, which used the most recent and comprehensive meta-analysis of levothyroxine published to date, suggest that levothyroxine in the treatment of hypothyroidism is associated with an increased risk of miscarriage. Chorionic bumps are defined as a convex, slightly irregular protrusion from the surface of the chorion into the gestational sac [ 94 , 95 ]. Studies have shown that chorionic bumps increase the risk of miscarriage [ 94 , 96 , 97 ]. The results of the present study also confirmed that chorionic bumps are associated with an increased risk of miscarriage. Inflammatory bowel disease (IBD) is a chronic intestinal disease. IBD patients are mostly women of childbearing age [ 98 ]. Studies have shown that the disease increases the risk of adverse pregnancy outcomes, while other studies have not observed this association [ 99 – 102 ]. The results of the present study showed that inflammatory bowel disease increases the risk of miscarriage. Surgery is another option if IBD does not respond to drug therapy to reduce the underlying inflammation [ 103 ]. Studies have reported that women with IBD who undergo surgery experience reduced fertility [ 104 – 106 ]. The results of the present study showed that surgical therapies for inflammatory bowel disease increase the risk of miscarriage. Selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors are the most common medications used to treat depression [ 107 ]. Several studies have shown that the use of antidepressants during pregnancy increases the risk of miscarriage [ 108 – 110 ]. A meta-analysis published in this study also showed a significant association between Selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors and the risk of miscarriage. When fetal placental hypoxia and ischemia occur, placental β-hCG levels increase [ 111 , 112 ]. Studies have shown that high β-hCG levels are associated with adverse pregnancy complications and outcomes [ 113 , 114 ]. When fetal placental hypoxia and ischemia occur, placental β-hCG levels increase [ 259 , 260 ]. Studies have shown that high β-hCG levels are associated with adverse pregnancy outcomes [ 261 , 262 ]. This study also confirmed that elevated levels of human chorionic gonadotropin are significantly associated with an increased risk of miscarriage. Hyperhomocysteinemia causes defective vascularization of chorionic villi, and low serum folate levels and vitamin B12 deficiency are associated with adverse pregnancy outcomes [ 115 – 118 ]. Meta-analyses published in this umbrella review also showed that serum vitamin B12 deficiency, low folic acid levels, and elevated serum homocysteine increase the risk of miscarriage. Primary biliary cholangitis is a chronic autoimmune liver disease that is more common in women [ 119 ]. The disease has been shown to increase miscarriages during pregnancy. The results of a meta-analysis published in this umbrella review also confirmed this association. The high rate of pregnancy complications, such as miscarriages, may be due to genitourinary tract infections associated with their disease [ 51 ]. Antiphospholipid syndrome is a rare autoimmune disorder that can be associated with adverse pregnancy outcomes such as miscarriage, preeclampsia, and premature birth [ 120 ]. Studies have shown that women with APs have a high rate of fetal death [ 121 , 122 ]. Type 1 diabetes is an autoimmune disease that causes the body to destroy cells in the pancreas, resulting in decreased insulin levels [ 123 ]. Conflicting results have been published regarding type 1 diabetes and the risk of miscarriage [ 124 – 126 ]. Polycystic ovary syndrome (PCOS) is the most common endocrine disorder in women that causes disruption of the ovulation process [ 127 , 128 ]. Insulin resistance and hyperinsulinemia are among the symptoms of PCOS [ 129 ]. Conflicting results have been published regarding PCOS and adverse pregnancy outcomes [ 129 – 134 ]. The existence of this contradiction indicates the need for a meta-analysis study. The results of meta-analyses of the factors mentioned in this umbrella review study showed that these factors (Antiphospholipid syndrome, Type 1 diabetes and PCOS) had a significant and unbiased association with risk of miscarriage. Metformin is an antidiabetic drug that reduces insulin concentration and improves tissue insulin sensitivity. Therefore, it is the best treatment. for PCOS [ 135 ]. The result of a review study has shown that the use of metformin during pregnancy is controversial [ 136 ]. The published results of the meta-analysis of this umbrella review showed that the use of metformin during pregnancy in women with polycystic ovary syndrome reduces the risk of miscarriage. Hydroxychloroquine sulfate is an antimalarial drug that is also used to treat autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus [ 137 – 139 ]. A meta-analysis published in 2009 found that use of the drug during pregnancy does not increase the risk of miscarriage [ 140 ], although several studies have suggested an increased risk of miscarriage in women taking the drug [ 141 – 143 ]. Given that half of all pregnancies are unplanned [ 144 ], women may use the drug during pregnancy. Therefore, it is important to evaluate the effect of hydroxychloroquine during pregnancy. The results of the meta-analysis published in this umbrella review showed that there is a significant and positive relationship between the use of Hydroxychloroquine sulfate and miscarriage and that the use of this tablet increases the risk of miscarriage. Colchicine is an anti-inflammatory agent commonly used in the treatment of gout and familial Mediterranean fever [ 145 ]. The U.S. Food and Drug Administration (FDA) has classified colchicine as a drug that may be dangerous to the fetus, but it does not completely prohibit its use [ 146 ]. Because there are no alternative treatments, it is important to investigate the safety of colchicine during pregnancy. Unfortunately, the results of the meta-analysis published in this study indicated that colchicine increases the risk of miscarriage during pregnancy. Hairdressers are exposed to a lot of physical stress every day because they may be working standing for long periods of time. They are also exposed to various chemicals throughout the day that may enter their bodies and threaten their health and cause illness [ 147 – 150 ]. This study showed that hairdressing increases the risk of miscarriage. However, because the studies are cross-sectional and only smoking was a confounder, there is a risk of confounding and temporality of the results, and future cohort studies should be conducted to examine the effect of such occupations on reproductive outcomes. The factors of smallpox vaccine, Herpes simplex virus 1/2, Psychological stress, Overweight, Obesity, Obesity (Grade 2), Inhaled corticosteroid-s (anti-inflammatory and immunosuppressive medication), Gamma-aminobutyric acid analog, Serum Vitamin B12 deficiency, fabrication, Family history of miscarriage, Triptans in the treatment of migraine and cluster headaches, Adverse childhood experiences, Laser treatment in monochorionic diamniotic twin pregnancies with twin-to-twin transfusion syndrome, Poor HbA1C control versus optimal control, Vegetable consumption, Seafood consumption, Dairy consumption, White meat consumption, Short menstrual cycle, Working more than 40 h and Working the night shift also showed a significant association without large heterogeneity with miscarriage, while one of the biases could not be calculated. because the number of cases and controls in the studies was not precisely stated or the number of studies was not sufficient to make the Egger test usable. However, because at least one bias could be calculated in this group of factors, they were classified in the group of associations supported by suggestive evidence. Factors related to Chlamydia trachomatis IgG, Metronidazole, Non-obstetric surgery with no surgery/conservative treatment, Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, parvovirus B19 infection, Systemic sclerosis, Rheumatoid arthritis, History of recurrent miscarriage, Depression, Anxiety, cocaine polydrug users/drug free controls, Non-alcoholic fatty liver and intimate partner violence were not classified into any group due to insufficient evidence and incomplete information from published meta-analyses. It is expected that in the future, the relationship of these factors with miscarriage will be determined by the publication of high quality and unbiased studies. One limitation of this study can be pointed out as follows: firstly, it is true that the 95% confidence interval is useful for further assessing heterogeneity and predicting the effect within an individual study [ 35 ], but its validity strongly depends on the number of initial samples. This is a weakness in meta-analyses that include small studies, so its interpretation should be approached with caution[ 151 ]. Additionally, the prediction interval is sensitive to heterogeneity: high heterogeneity yields a wide interval, and a significant prediction interval requires a large, significant effect. A non-significant prediction interval does not invalidate results; if other strong-evidence criteria are met, it remains valuable. The included studies are observational, so confounding is possible. However, ethical constraints on pregnant-woman interventions mean observational meta-analyses can still suggest associations if heterogeneity is not substantial. Among the 45 factors in the suggestive-evidence group, several (chorionic bump, PCOs, hyperhomocysteinemia, serum vitamin B12 deficiency, primary antiphospholipid syndrome, family history of miscarriage, triptan, surgical therapies for IBD, poor HbA1c control vs. optimal control, Colchicine) had odds ratios or relative risks > 2, but sample sizes and the number of studies were small. Thus, findings should be interpreted cautiously due to limited power and potential random error. The Egger test has limited power to detect publication bias with few studies. Future larger studies are needed to clarify these associations with miscarriage risk. In this umbrella review study, because all factors related to miscarriage have been collected and possible biases have been measured, it is possible to have an overall view of the factors that related to miscarriage and identify the factors that have a significant impact on the risk of miscarriage.

Miscarriage is defined as the death of a fetus before 20 weeks of gestation or weighing ≤ 500 g and is one of the most common complications of pregnancy [ 1 , 2 ]. The incidence of spontaneous abortion is between 10 and 20% [ 3 ]. Approximately 73 million miscarriages occur worldwide each year [ 4 ]. Early studies have shown that approximately 5% of all women who have previously experienced spontaneous abortion later experience recurrent spontaneous abortion, which is a distressing experience for mothers and affects the emotional state of the mothers, and may lead to physical disorders, depression, anxiety, sleep disorders, and general health complications, and reduce the individual's social functioning [ 5 , 6 ]. Genetic disorders and fetal chromosomal abnormalities, which are involved in approximately half of all spontaneous abortions, are the non-modifiable causes of miscarriage, and the remaining causes of miscarriage are largely unknown and may be preventable and modifiable [ 7 – 9 ]. So far, many systematic reviews and meta-analyses have been published on modifiable and non-modifiable risk factors for spontaneous abortion, some of which include smoking and alcohol consumption, night work, obesity, vitamin D deficiency, environmental conditions and maternal lifestyle such as caffeine consumption, artificial sweeteners, exposure to chemicals, microbial infections, use of antibiotics, uterine abnormalities, infectious diseases and untreated maternal diseases, history of previous abortion, irregular menstruation; and underlying endocrine disorders like hypothyroidism and diabetes.[ 10 – 30 ]. With the growing number of systematic reviews, the next step to inform healthcare decision-making is to conduct an overview of existing reviews. Such overviews are also known by various terms in the review literature, and umbrella reviews are one of them [ 31 , 32 ]. So far, no studies have synthesized the evidence from existing systematic reviews and meta-analyses on non-genetic factors associated with miscarriage, nor have they addressed the limitations of these studies, such as potential biases. Therefore, this study aimed to conduct an umbrella review of non-genetic risk factors associated with miscarriage in review studies.