Unsafe abortion in Brazil: a systematic review of the scientific production, 2008-2018

Domingues, Rosa Maria Soares Madeira; Fonseca, Sandra Costa; Leal, Maria do Carmo; Aquino, Estela M. L.; Menezes, Greice M. S.

Acessibilidade / Reportar erro

Brasil

Cadernos de Saúde Pública

Español English

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

REVIEW • Cad. Saúde Pública 36 (Suppl 1) • 2020 • https://doi.org/10.1590/0102-311X00190418 copy

Unsafe abortion in Brazil: a systematic review of the scientific production, 2008-2018

Authorship SCIMAGO INSTITUTIONS RANKINGS

This study sought to update knowledge on unsafe abortion in Brazil. We carried out a systematic review with study search and selection on MEDLINE and LILACS, with no language restriction, from 2008 to 2018. We evaluated article quality using the Joanna Briggs Institute instruments. We evaluated 50 articles. The prevalence of induced abortion in Brazil was estimated by a direct method to be 15% in 2010 and 13% in 2016. Higher prevalences were observed in more socially vulnerable populations. There was a decrease in the ratio of induced abortions by 1,000 women of reproductive age in the period 1995-2013, reaching 16 per 1,000 in 2013. Half of all women reported using medications for terminating pregnancies and the number of hospital admissions due to complications from abortion, especially severe complications, decreased from 1992 to 2009. Maternal morbimortality from abortion had a reduced frequency but reached high values in specific contexts. It is likely that maternal deaths from abortion are under-reported. Common mental disorders during pregnancy and postpartum depression were more frequent among women who unsuccessfully attempted to induce an abortion. Findings indicate that abortion is frequently used in Brazil, especially in less-developed regions and by more socially-vulnerable women. Access to safer methods probably contributed to the reduction in hospitalizations due to complications and to the reduction in morbimortality from abortion. However, half of all women still resort to other methods and the number of admissions due to complications from abortion is still high.

Keywords:
Induced Abortion; Systematic Review; Health Surveys

Reference	Location (year)/ period	Study design/Base and estimate	Population	Methodological limitations	Results
Population base, national scope, direct estimate
Cecatti et al. ¹⁰10. Cecatti JG, Guerra GV, Sousa MH, Menezes GM. Aborto no Brasil: um enfoque demográfico. Rev Bras Ginecol Obstet 2010; 32:105-11.	Brazil (1996).	Household population survey. Direct estimate. 1996 PNDS.	12,612 women aged 15-49 years.	Information on abortion obtain through direct interview, subject to under-reporting; Does not present CI of estimates; 85.5% response rate, with no information provided to evaluate whether refusals were selective.	Unsafe abortions prevalence: 2.4%. Regional inequalities: Northest (3.1%), North (2.3%), South (1.7%) and Central (1.3%). Rio de Janeiro (6.5%). Associated factors: - Age: growing prevalence from 15-19 years (0.5%) to > 40 years (4.5%); - Residence: urban > rural; - Non-religious > religious; - Higher educational level > lower educational level (3.7%).
Camargo et al. ¹¹11. Camargo RS, Santana DS, Cecatti JG, Pacagnella RC, Tedesco RP, Melo Jr. EF, et al. Severe maternal morbidity and factors associated with the occurrence of abortion in Brazil. Int J Gynaecol Obstet 2011; 112:88-92.	Brazil (2006).	Household population survey. Direct estimate. 2006 PNDS.	15,775 women aged 15-49 years, 4,340 of whom with children born alive in the past 5 years.	Information on abortion obtain through direct interview, subject to under-reporting; Does not present CI of estimates; 11% loss of eligible women in urban regions and 9.4% in rural regions, with no information provided to evaluate whether refusals were selective.	Unsafe abortions prevalence over reproductive life: 2.3% and among women with previous pregnancies: 3.3%; Regional inequalities: North (4.3%), Northeast (3.5%), Southeast (1.8%); Central (1.3%) and South (0.8%); unsafe abortions prevalence: 1.8% among women who had been pregnant in the previous 5 years.
Diniz & Medeiros ¹²12. Diniz D, Medeiros M. Aborto no Brasil: uma pesquisa domiciliar com técnica de urna. Ciênc Saúde Colet 2010; 15 Suppl 1:959-66.	Brazil (2010).	Household population survey using ballot box technique. Direct estimate. 2010 PNA.	2,002 literate women residing in urban areas. Age: 18-39 years.	Does not inform all parameters used to calculate sample size (only error margin and estimate precision); does not include the entire reproductive period (only 18-39 years), or illiterate women or women from rural regions; does not inform if the design effect was incorporated into the analysis (cluster sampling); does not inform proportion of losses and refusals.	Induced abortion prevalence: 15%; By age group: smaller among women aged 18-19 years (6%) and increases to 22% in the 35-39 years; Among those who had abortions: peak 20-24 years; Among those with low educational levels: 23% and high educational levels: 14%; No differences according to religion; Half used medication and 55% needed hospitalization.
Diniz et al. ¹³13. Diniz D, Medeiros M, Madeiro A. Pesquisa Nacional de Aborto 2016. Ciênc Saúde Colet 2017; 22:653-60.	Brazil (2016).	Household population survey using ballot box technique. Direct estimate. 2016 PNA.	2,002 literate women residing in urban areas. Age: 18-39 years.	Does not inform all parameters used to calculate sample size (only error margin and estimate precision); does not include the entire reproductive period (only 18-39 years), or illiterate women or women from rural regions; does not inform if the design effect was incorporated into the analysis (cluster sampling); does not inform proportion of losses and refusals.	Induced abortion prevalence: 13%; By age group: smaller among women aged 18-19 years (9%) and increases to 18% in the 35-39 years; Among those who had abortions: peak 20-24 years; Among those with low educational levels: 22% and high educational levels: 11%; No differences according to religion; According to race/color: black and brown (14 to 15%) and white (9%); Half used medication; 48% needed hospitalization (< than in 2010); Estimate of 503 thousand unsafe abortions in 2015.
Massaro et al. ¹⁴14. Massaro LTS, Abdalla RR, Laranjeira R, Caetano R, Pinsky I, Madruga CS. Alcohol misuse among women in Brazil: recent trends and associations with unprotected sex, early pregnancy, and abortion. Braz J Psychiatry 2019; 41:131-7.	Brazil (2006 e 2012).	Household population survey. Direct estimate. LENAD 2012.	2,537 women aged ≥ 14 years.	Does not inform parameters used to calculate sample size; unsafe abortions measured by direct interview, subject to under-reporting; Does not describe characteristics of the studied sample; it is not possible to evaluate temporality between alcohol use and unsafe abortions.	Lifetime abortion prevalence - 26.3% (14.5-18.3) in the entire sample; 15% (13.2-17.0) among women with no binge drinking or AUD; 20.4% (15.3-26.7) among women with binge drinking, but without AUD; 24.9% (16.3-36.2) among women with binge drinking and AUD.
Population base, national scope, indirect estimate
Martins-Melo et al. ¹⁵15. Martins-Melo FR, Lima MS, Alencar CH, Ramos Jr. AN, Carvalho FH, Machado MM, et al. Tendência temporal e distribuição espacial do aborto inseguro no Brasil, 1996-2012. Rev Saúde Pública 2014; 48:508-20.	Brazil (1996-2012).	Ecological. Temporal and spacial dimension. Hospital study (SUS users). Indirect estimate.	Around 4 million hospitalizations from abortion in SIH/SUS (ICD-10 codes:O-00-O-08).	No limitations identified. One item (response rate) not applicable.	Estimate of 994 thousand unsafe abortions/year; Mean UA/WRA coefficient: 17/1,000. UA/LB ratio: 33/100 live birth; Higher elevated indicators in the Northeast (21/1,000 women of reproductive age and 40/100 live birth); Decline over time at the national level, with regional differences; For UA/WRA: constant decreasing trend in Northeast, Central and Southeast; stability in the South; non-constant growing trend in the North; For the UA/LB ratio: constant growing in the North, Central and S, decreasing in the Southeast and Northeast; For both indicators, differences were also found between states; Spatial distribution, clusters in North, Northeast and Southeast regions.
Monteiro et al. ¹⁶16. Monteiro MFG, Adesse L, Drezett J. Atualização das estimativas da magnitude do aborto induzido, taxas por mil mulheres e razões por 100 nascimentos vivos do aborto induzido por faixa etária e grandes regiões. Brasil, 1995 a 2013. Reprod Clim 2015; 30:11-8.	Brazil (1995-2013).	Time series. Hospital-based study (SUS users). Indirect estimate.	N ≥ 4 million hospitalizations. Age: 15-49 years.	Does not inform ICD-10 codes used for the indirect estimate of abortion cases; does not use statistical tests to assess time variations; One item (response rate) not applicable.	Maximum abortion estimate: 1,086,708 in 1995 and 865,160 in 2013. Minimum estimate: 864,628 and 697,347 in 2013; Global reduction of unsafe abortions/1,000 women of reproductive age: from 27-16. In regions: North (29-21); Northeast (38-18) and Southeast (25-14); stable South; Central unstable reduction. 27% reduction in the number of hospitalizations; greater for those aged 20-29 years (-38%) and 15-19 years (-35%). Reduction in the number of abortions/100 live birth: from 35/100 in 2004 to 30/100 live birth in the Northeast (38-28) and Southeast (25-14). Reduction of unsafe abortions/1,000 women of reproductive age: from 30 to 17 in the age group 15-19 years; from 43 to 22 (20-29 years); from 23 to 17 (30-39 years).
Hospital base, national scope, direct estimate
Machado et al. ¹⁷17. Machado CJ, Lobato AC, Melo VH, Guimarães MD. Perdas fetais espontâneas e voluntárias no Brasil em 1999-2000: um estudo de fatores associados. Rev Bras Epidemiol 2013; 16:18-29.	Brazil (2006).	Multicenter cross-sectional study. Hospital-based (SUS reference maternities). Direct estimates.	1,838 puerperae with previous pregnancies (of 3,047).	Sample obtained from reference maternity hospitals in the national STI/AIDS program. It is not clear if this sample can be considered representative of women of reproductive age; Parameters used to calculate sample size related to prevalence of syphilis during pregnancy; sub-sample of women with previous pregnancies used to estimate frequency of unsafe abortions; Proportion of losses not reported, only proportion of refusals; Induced abortions estimated based on the difference between total losses and miscarriages, with the latter information obtained from direct interviews, subjected to under-reporting; Did not incorporate design effect into the analysis; did not present CI of obtained estimates.	Unsafe abortions prevalence: 9,7%.
Population base, local scope, indirect estimate
Mello et al. ¹⁹19. Mello FM, Sousa JL, Figueroa JN. Magnitude do aborto inseguro em Pernambuco, Brasil, 1996 a 2006. Cad Saúde Pública 2011; 27:87-93.	Pernambuco State, Brazil (1996-2006).	Ecological. Temporal and spatial dimensions. Restricted to SUS users. Indirect estimate.	147,205 hospitalizations from abortion in SIH-SUS. (ICD-10 codes: O00 and O03 -O06)	Does not use all ICD-10 codes related to abortion and does not justify the criterion they used; One item (response rate) not applicable.	Estimates of 65,457 abortions/year; Greater concentration in GERES I and IV. Higher indicator in GERES I and VIII, around 40/100 live birth; Unequal temporal decline, more intense in GERESI; GERES X and XI had an increase in the indicator; Decline in the total number of abortions in the state, last year of the series: 51,853 abortions.
Madeiro et al. ²⁰20. Madeiro AP, Rufino AC, Santos IS, Carvalho MS. Estimativas e tendências de aborto provocado no Piauí: um estudo ecológico no período de 2000-2010. Rev Bras Promoc Saúde 2015; 28:168-75.	Piauí, Brazil (2000-2010).	Ecological. Temporal and spatial dimensions. Restricted to SUS users. Indirect estimate.	55,678 hospitalizations from abortion in SIH-SUS. (ICD-10 codes: O03 -O07).	Does not justify ICD-10 codes used to identify hospitalizations from abortion; One item (response rate) not applicable.	Estimated number of unsafe abortions dropped from 10,362 (2000) to 6,738 (2010); UA/LB ratio for the state dropped from 17.6 to 13.5 with significant annual reduction of 2.2%. Region TD4 was the most responsible, with 40% population and reduction from 16.7 to 6.9; Other regions with increase or stability.
Population base, local scope, women of reproductive age, direct estimate
Silva & Vieira ²¹21. Silva RS, Vieira EM. Frequency and characteristics of induced abortion among married and single women in São Paulo, Brazil. Cad Saúde Pública 2009; 25:179-87.	São Paulo, Brazil (1993).	Household survey. Population base. Direct and RRT estimates.	1,749, age 15-49 years; Two sub-samples: one with measurement of unsafe abortions in the previous year by RRT (n = 876) and another with measurement by direct interview (n = 873); In both sub-samples, measurement of lifetime unsafe abortions through direct interview; Exclusion of 317 women who were separated, in consensual unions or widowed.	Non-probabilistic samples; Unclear parameters for calculating sample size and sampling procedure; Uninformed losses and refusals. Does not describe sample characteristics; Lifetime unsafe abortions measured through direct interviews, subjected to under-reporting; Does not present CI of unsafe abortions estimates.	Unsafe abortions prevalence in the previous year: Measurement through direct interview: 1 per 1,000; Measurement by RRT: 42 per 1,000; Prevalence of lifetime abortion (direct interview): 45 unsafe abortions per 1,000 women, 4.4% of pregnancies ended in unsafe abortions; Among married women (n = 764): 45 unsafe abortions per 1,000, 90% of women with previous pregnancy, 2% ended in unsafe abortions; Among single women (n = 658): 49 unsafe abortions per 1,000, 16% of women with previous pregnancy, 18% ended in unsafe abortions.
Souza et al. ²²22. Souza MG, Fusco CLB, Andreoni SA, Souza e Silva R. Prevalência e características sociodemográficas de mulheres com aborto provocado em uma amostra da população da Cidade de São Paulo, Brasil. Rev Bras Epidemiol 2014; 17:297-312.	São Paulo, Brazil (2008).	Household survey. Population base. Direct estimate.	683 women with previous pregnancies; age: 15-59 years.	Does not inform parameters used to calculate sample size; Apparent underestimation of young women, indicating probable failure in the sample selection; Information on abortion obtained through direct interview, subject to under-reporting; Does not inform proportions of loss and refusal; Does not present CI of estimates.	Unsafe abortions prevalence: 4.5%. Higher among single women (10.1%) in those with 5 or more live birth (14.3%) and the ones who use ineffective contraception (7.7%).
Fusco et al. ²³23. Fusco CLB, Silva RS, Andreoni S. Unsafe abortion: social determinants and health inequities in a vulnerable population in São Paulo, Brazil. Cad Saúde Pública 2012; 28:709-19.	Favela Inajar, São Paulo, Brazil (2005/2006).	Household survey. Population base. Direct estimate.	375 women, 278 with previous pregnancies; Age = 15-54 years.	Unsafe abortions measured through direct interviews, subjected to under-reporting. Does not present confidence intervals for the estimates.	Global unsafe abortions prevalence: 13.6%; Among black women and with low educational levels - 35.7%, black with low income - 40%; black and single - 36.7%; Prevalence among women with some pregnancy - 18.35.
Santos et al. ²⁴24. Santos TF, Andreoni S, Souza e Silva R. Prevalência e características de aborto induzido - Favela México 70, São Vicente - São Paulo. Rev Bras Epidemiol 2012; 15:123-33.	Favela México 70, São Vicente, São Paulo State, Brasil (2008).	Household survey. Population base. Direct estimate.	735 women with previous pregnancies; Age: 15-49 years.	It is not clear if the method used for the selection of women may have generated some selection bias; Measurement of abortion through direct interviews, subjected to under-reporting; Does not inform the proportion of losses and refusals; Does not present a confidence interval of estimates.	Prevalence 6.9%. In the group 40 to 45 years: 12.9%; with low educational level: 11.4%; no children: 15.2% and with 6 or more children: 12.2%.
Population base, local or regional scope, young population, direct estimate
Silva & Andreoni ²⁵25. Silva RS, Andreoni S. Fatores associados ao aborto induzido entre jovens pobres na cidade de São Paulo, 2007. Rev Bras Estud Popul 2012; 29:409-19.	Freguesia do Ó community, São Paulo State, Brazil (2007).	Household survey. Population base. Direct estimate.	Sexually active youths: 102 men and 99 women; Age: 15-25 years.	Measurement of induced abortion through direct interviews, subjected to under-reported; Does not present confidence intervals for the estimates.	Unsafe abortions prevalence men (partners): 10.8%; Unsafe abortions prevalence among women: 6.1%
Silva & Fusco ²⁶26. Silva RS, Fusco CLB. Comportamento do aborto induzido entre jovens em situação de pobreza de ambos os sexos - Favela México 70, São Paulo, Brasil, 2013. Reprod Clim 2016; 31:13-21.	Favela México 70, São Vicente, São Paulo, Brazil (2013).	Household survey. Population base. Direct estimate.	Sexually active youths: 327 women and 253 men; Age: 15-24 years.	Measurement of induced abortion through direct interviews, subjected to under-reported; Does not present confidence intervals for the estimates.	Unsafe abortions prevalence men (partners): 2.8%; Unsafe abortions prevalence among women: 1.2%.
Pilecco et al. ²⁷27. Pilecco FB, Knauth DR, Vigo A. Aborto e coerção sexual: o contexto de vulnerabilidade entre mulheres jovens. Cad Saúde Pública 2011; 27:427-39.	Rio de Janeiro, Porto Alegre (Rio Grande do Sul State) e Salvador (Bahia State), Brazil (2001-2002).	Household survey. GRAVAD study. Population base. Direct estimate.	870 young women (18-24 years) with previous pregnancies.	Does not inform parameters used when calculating the sample size; Measurement of induced abortion through direct interviews, subjected to under-reported; Does not present data that enabled an evaluation of whether losses (14.8%) were selective; Does not present confidence intervals for the estimates.	Prevalence: 21.5% among those who had become pregnant; Rio de Janeiro (52.8%), Salvador (42.1%), Porto Alegre (5.1%).
School base, local scope, young population, direct estimate
Correia et al. ²⁸28. Correia DS, Cavalcante JC, Egito ES, Maia EMC. Prática do abortamento entre adolescentes: um estudo em dez escolas de Maceió - AL, Brasil. Ciênc Saúde Colet 2011; 16:2469-76.	Maceió, Alagoas State, Brazil (2005).	Cross-sectional. Restricted to schools (public and private). Direct estimate, self-administered questionnaires, in the school environment.	Age: 12-19 years, 2,592 students. In this analysis, N = 559 (sexually active girls).	Only adolescent students; Unclear sampling procedure; Unsafe abortions measured through a self-administered questionnaire, in a school environment, with a possibility of under-reporting; No confidence interval for estimates; Does not incorporate design effect in the analyses; Did not inform losses and refusals.	Unsafe abortion prevalence: 26.7% for the total of sexually active girls and 81.9% among those who became pregnant.
Population or service base, local scope, specific populations, direct estimate
Neiva-Silva et al. ²⁹29. Neiva-Silva L, Demenech LM, Moreira LR, Oliveira AT, Carvalho FT, Paludo SDS. Pregnancy and abortion experience among children, adolescents and youths living on the streets. Ciênc Saúde Colet 2018; 23:1055-66.	Porto Alegre e Rio Grande (Rio Grande do Sul State), Brazil (2008).	Cross-sectional, RDS sampling.	307 homeless children, adolescents and youths of both sexes, aged 10-21 years. Two hundred and four in Porto Alegre, 103 in Rio Grande.	Abortion measured through direct interview. Abortion type (induced or miscarriage) asked only for the first abortion, with 32% reporting more than one abortion. Possibility of unsafe abortions under-reporting; CI of estimates not presented.	81,1% male sex, 93.8% between 12-21 years, 29.3% with experience with pregnancy, 26.7% of whom aged ≤ 14 years; 15.6% unsafe abortions prevalence among individuals with previous experience of pregnancy; Greater abortion experience (miscarriage or unsafe abortion) among girls than among boys (17.4% vs. 8.8%).
Barbosa et al. ³⁰30. Barbosa RM, Pinho AA, Santos NS, Filipe E, Villela W, Aidar T. Induced abortion in women of reproductive age living with and without HIV/Aids in Brazil. Ciênc Saúde Colet 2009; 14:1085-99.	Thirteen Brazilian municipalities from the 5 macro-regions (2003-2004).	Cross-sectional. Self-administered questionnaire in HIV/AIDS reference services and in primary health units and women’s health care services. Direct estimates.	Women aged 18-49 years, literate, sexually active; 1,777 WLHA; 2,045 WNLHA.	Convenience sample; Under-representation of WLHA in the North Region and over-representation in the Southeast Region; CI of estimates not presented.	Lifetime unsafe abortion prevalence: WLHA: 17.5%; WNLHA: 10.4%, p < 0.001; Non-significant differences after adjusting for age, marital state, number of children and lifetime sexual partners: WLHA 13,3% vs. WNLHA 11,0%; p > 0,05).
Pilecco et al. ³¹31. Pilecco FB, Teixeira LB, Vigo A, Dewey ME, Knauth DR. Lifetime induced abortion: a comparison between women living and not living with HIV. PLoS One 2014; 9:e95570.	Porto Alegre, Rio Grande do Sul State, Brazil (2011).	Cross-sectional. Interviews in 7 and 27 primary health services. Direct estimates.	Women with previous pregnancies; 18-49 years; 625 WLHA; 498 WNLHA.	Unclear procedure for selecting primary health units; unsafe abortions measured through direct interviews, subject to under-reporting; CI of estimates not presented.	Lifetime unsafe abortion prevalence: WLHA: 13.0%; WNLHA: 4.9%; Pregnancies resulting in unsafe abortion: WLHA: 6.5% (7.7% before diagnosis of the infection, 3.8% after diagnosis of the infection, p = 0.024); WNLHA: 2.9%.
Pinho et al. ³²32. Pinho AA, Cabral CDS, Barbosa RM. Differences and similarities in women living and not living with HIV: contributions by the GENIH study to sexual and reproductive healthcare. Cad Saúde Pública 2017; 33:e00057916.	São Paulo, Brazil (2013-2014).	Cross-sectional. Interviews in 18 HIV/AIDS reference services and 38 primary health services. Direct estimates.	Sexually active women; 18-49 years; 918 WLHA;1,003 (WNLHA).	High percentage of refusals (27% among WLHA, 26.5% among WNLHA); Characteristics not describe, uncertain selection bias; unsafe abortions measured through direct interviews, subject to under-reporting; CI of estimates not presented.	Lifetime unsafe abortions prevalence among women with previous pregnancies: WLHA: 14.1%; WNLHA: 3.02.
Friedman et al. ³³33. Friedman RK, Bastos FI, Leite IC, Veloso VG, Moreira RI, Cardoso SW, et al. Pregnancy rates and predictors in women with HIV/AIDS in Rio de Janeiro, Southeastern Brazil. Rev Saúde Pública 2011; 45:373-81.	Rio de Janeiro, Brazil (1996-2003).	Prospective cohort STI/AIDS reference center Direct estimate through interview	225 women living with HIV/AIDS > 18 years.	Unsafe abortions measured through direct interview, subject to under-reporting.	Unsafe abortions incidence: 2.1% (95%CI: 1.2%-3.0%) women year; Of the 60 pregnancies that occurred during follow up, 31% resulted in unsafe abortions.
Madeiro & Rufino ³⁴34. Madeiro AP, Rufino AC. Aborto induzido entre prostitutas: um levantamento pela técnica de urna em Teresina - Piauí. Ciênc Saúde Colet 2012; 17:1735-43.	Teresina, Piauí, Brazil (2011).	Cross-sectional. Workplace interviews. Population. Direct estimate.	310 female sex workers; 18-39 years.	Non-probabilistic sample, not including the entire reproductive range, with the exclusion of illiterate women and majority inclusion of women working in brothels (under-representation of women working on the streets). Did not present CI of estimates.	52.6% prevalence. Age groups: 18-19 years: 27.1%; 35-39 years: 71.2%; 3 or more previous pregnancies: 94.7%; More than 10 years working in prostitution: 61.3%; Misoprostol used in over 70% of cases.
Diehl et al. ³⁵35. Diehl A, Pillon SC, Santos MA, Laranjeira R. Abortion and sex-related conditions in substance-dependent Brazilian patients. Cad Saúde Pública 2017; 33:e00143416.	São Paulo, Brazil (2009-2011).	Cross-sectional; Interviews 15 days after admission to a clinic specialized in addiction treatment.	Patients aged ≥ 18 years, with confirmed clinical diagnosis of addiction; Using DSM-IV-TR N = 616; 82.5% male sex.	Does not inform parameters used to calculate sample size; unsafe abortions measured through direct interviews, subject to under-reporting (in men, experience with induced abortion measured); Did not present CI of estimates.	26.8% prevalence; In men: 23.8%; In women: 40.7%

Reference	Location (years)	Study design	Population	Methodological limitations	Results
Population studies, national scope
Cecatti et al. ¹⁰10. Cecatti JG, Guerra GV, Sousa MH, Menezes GM. Aborto no Brasil: um enfoque demográfico. Rev Bras Ginecol Obstet 2010; 32:105-11.	Brazil (1996).	Household survey (PNDS, 1996).	12,612 women (15-49 years) from urban and rural areas.	Induced abortion measured through direct interview, subject to under-reporting; Analysis of factors associated with unsafe abortions in comparison with women who never had abortions or who had miscarriage; Inadequate comparison group.	Unsafe abortions over the course of reproductive life: Age (OR = 1.06/year), residing in urban areas (OR = 1.56), non-white race/color (OR = 1.41), working (OR = 1.41), number of children > 1 (OR = 2.22); being religious (OR = 0.59); ideal number of children > 1 (OR = 0.68); Educational level and marital status: no statistically significant association.
Camargo et al. ¹¹11. Camargo RS, Santana DS, Cecatti JG, Pacagnella RC, Tedesco RP, Melo Jr. EF, et al. Severe maternal morbidity and factors associated with the occurrence of abortion in Brazil. Int J Gynaecol Obstet 2011; 112:88-92.	Brazil (2006).	Household survey (PNDS, 2006).	4,340 of the 15,755 women (15-49) with children 5 years before the survey.	Induced abortion measured through direct interview, subject to under-reporting; Analysis of factors associated with unsafe abortions in comparison with women who never had abortions or who had miscarriage; Inadequate comparison group.	None of the assessed factors (age at time of interview, urban/rural residence, race/color, work, number of children, religion, planned pregnancy) was associated with unsafe abortions in the previous 5 years of life.
Massaro et al. ¹⁴14. Massaro LTS, Abdalla RR, Laranjeira R, Caetano R, Pinsky I, Madruga CS. Alcohol misuse among women in Brazil: recent trends and associations with unprotected sex, early pregnancy, and abortion. Braz J Psychiatry 2019; 41:131-7.	Brazil (2006 and 2012).	Household population survey. Direct estimate. LENAD 2012.	2,537 women aged ≥ 14 years.	Induced abortion measured through direct interview, subject to under-reporting; Does not describe sample characteristics, only the prevalence of the analyzed outcomes according to women’s characteristics.	Outcome = lifetime unsafe abortions. OR adjusted by age and educational level, reference category: women with no binge drinking or AUD: Women with binge drinking, but not AUD: 1.9 (95%CI: 1.3-2.8); Women with binge drinking and AUD: 2.5 (95%CI: 1.5-4.4); Lack of evidence of direct effect of alcohol use on unsafe abortions (direct effect 0.142, 95%CI: -0.103-0.386) when early pregnancy is excluded from analysis.
Population studies, local scope
Souza e Silva & Vieira ²¹21. Silva RS, Vieira EM. Frequency and characteristics of induced abortion among married and single women in São Paulo, Brazil. Cad Saúde Pública 2009; 25:179-87.	São Paulo Brazil (1993).	Household population survey. Direct estimate.	1,749; Age 15-49 years; For this analysis, exclusion of 317 women who were separated, in a consensual union or widows.	Non-probabilistic sample; Unclear parameters used to calculate sample size and sampling procedure; Sample characteristics not informed; lifetime unsafe abortions measured through direct interviews, subjected to under-reporting; Exposure variables measured at the time of the interview, not at the time of the Unsafe abortions; Only non-adjusted analysis.	Associated factors: Among single women: greater proportion of unsafe abortions/pregnancies among those who did not reach the desired number of children (29.5%) with regard to those who did (4.3%) and in women aged 15-19 years (60%); Non-significant differences according to contraception used; Among married women, greater proportion also among younger women, but in lower proportions (4.3% among women aged 15-19 years) and among women who used effective contraception methods (2.7%).
Souza et al. ²²22. Souza MG, Fusco CLB, Andreoni SA, Souza e Silva R. Prevalência e características sociodemográficas de mulheres com aborto provocado em uma amostra da população da Cidade de São Paulo, Brasil. Rev Bras Epidemiol 2014; 17:297-312.	São Paulo, Brazil (2008).	Household survey	683 women with previous pregnancies. Age: 15-49 years.	Induced abortion measured through direct interview, subject to under-reporting; Variables used in the multivariate model were measured at the time of the interview, not at the time when unsafe abortions occurred, generating incoherent results.	Factors associated with unsafe abortions at the time of the interview: age 40-44 years (PR = 2.76); being single (PR = 2.79); ≥ 5 live birth children (PR = 3.97); use of pill or IUD (PR = 2.70), low-efficacy method (PR = 4.18) compared with permanent sterilization.
Fusco et al. ²³23. Fusco CLB, Silva RS, Andreoni S. Unsafe abortion: social determinants and health inequities in a vulnerable population in São Paulo, Brazil. Cad Saúde Pública 2012; 28:709-19.	Favela Inajar, São Paulo, Brazil (2005).	Household survey.	375 women; 278 with previous pregnancies; Age: 15-49 years.	Study subjects and context not adequately described; Induced abortion measured through direct interview, subjected to under-reporting; Some variables used in the multivariate model measured at the moment of the interview and not when the unsafe abortions occurred.	Factors associated with UA among women with reported pregnancy: age at sexual initiation < 16 years (OR = 3.91); number of partners in the last year ≥ 2 (OR = 3.30); black race/color (OR = 2.27); low educational level (OR = 2.85); and accepting abortions for economic reasons (OR = 3.35); Not associated: being single at the time of abortion, religion, work, income, contraception use, being a migrant; For the entire sample, the same factors were associated, in addition to number of children > ideal (OR = 3.08).
Santos et al. ²⁴24. Santos TF, Andreoni S, Souza e Silva R. Prevalência e características de aborto induzido - Favela México 70, São Vicente - São Paulo. Rev Bras Epidemiol 2012; 15:123-33.	Favela México 70, São Vicente, São Paulo, Brazil (2008).	Household survey.	735 women with previous pregnancies; Age: 15-49 years.	Induced abortion measured through direct interviews, subject to under-reporting; Variables used in the multivariate model measured at the moment of the interview and not at the time when the abortion occurred, generating incoherent results.	Associated factors: low income (OR = 1.76); live birth = 0 (OR = 12.2)/2 to 5 (OR = 4.5)/6 or + (OR = 5.2); acceptance of abortion (OR = 5.7) and age (OR = 1.04/year); Not associated: Paid activity, educational level, marital status, contraception.
Population studies restricted to youths/adolescents
Correia et al. ²⁸28. Correia DS, Cavalcante JC, Egito ES, Maia EMC. Prática do abortamento entre adolescentes: um estudo em dez escolas de Maceió - AL, Brasil. Ciênc Saúde Colet 2011; 16:2469-76.	Maceió, Alagoas, Brazil (2005).	Cross-sectional, 10 public and private schools. Self-administered questionnaire.	2,592 students of both sexes aged 12-19 years. Analysis restricted to 559 sexually active students (12-19 years old) of both public and private schools.	Schools and students included in the study not adequately described; Although the questionnaire is self-administered, the authors themselves admit the possibility of unsafe abortions under-reporting; Only non-adjusted analysis.	Non-adjusted analysis of associated factors among sexually active female students: Public school (OR = 1.41); Age 12-14 (OR = 0.22); Marital union (OR = 3.31); Other factors not tested.
Silva & Andreoni ²⁵25. Silva RS, Andreoni S. Fatores associados ao aborto induzido entre jovens pobres na cidade de São Paulo, 2007. Rev Bras Estud Popul 2012; 29:409-19.	Favela Inajar, São Paulo, Brazil (2007).	Household survey.	102 men and 99 women (aged 14-25) who were sexually active.	Some exposure variables measured in a non-validated manner; Induced abortion measured through direct interviews, subject to under-reporting; Variables used in the multivariate model measured at the time of the interview and not at the time of the unsafe abortions.	Associated factors: Model without the variable “current partner”: being male (OR = 9.12) and number of pregnancies (at each new pregnancy, OR = 7.29, 95%CI: 3.33-15.98); Model with the variable “current partner”: being male (OR = 13.9), higher number of pregnancies (OR = 7.30), living alone (OR = 4.32) and increased age (OR = 0.73 per year).
Pilecco et al. ²⁷27. Pilecco FB, Knauth DR, Vigo A. Aborto e coerção sexual: o contexto de vulnerabilidade entre mulheres jovens. Cad Saúde Pública 2011; 27:427-39.	Rio de Janeiro, Porto Alegre (Rio Grande do Sul State) and Salvador (Bahia State), Brazil (2001-2002).	Multi-center household survey.	870 women (18-24 years) with accounts of pregnancy, participants of the GRAVAD study.	Unsafe abortions assessed through direct interviews, subject to under-reporting.	Associated factors: sexual coercion (PR = 1.60); living in Rio de Janeiro (PR = 2.16) or Salvador (PR = 2.75); having obtained information on sexual relations from people other than parents (PR = 1.9); number of partners: 2-4 (PR = 2.21) and > 5 (PR = 2.66); education (the higher the education, the higher the PR - higher education - PR = 6.47); number of pregnancies (PR = 1.65). The following were not associated: young woman’s age, mother’s educational level, remunerated occupation and characteristics of coercion, age and partner at sexual initiation and contraceptive use.
Hospital/health service based studies
Nader et al. ³⁶36. Nader PRA, Macedo CR, Miranda AE, Maciel ELN. Aspectos sociodemográficos e reprodutivos do abortamento induzido de mulheres internadas em uma maternidade de Serra - ES. Esc Anna Nery Rev Enferm 2008; 12:699-705.	Serra, Espírito Santos State, Brasil (2005-2006).	Case control. Classification of abortion according to WHO criteria.	21 women hospitalized for certainly induced abortions and 83 women with pregnancies carried to term.	Does not inform adequately case identification and inclusion process; Certainly induced abortion measured through direct interviews, with possibility of under-reporting and case misclassification; Variables included in the final model probably highly correlated.	Factors associated with certainly induced abortion: Being married: adjOR 0.= 241 (95%CI: 0.061-0.951); Wanted pregnancy: adjOR = 0.168 (95%CI: 0.042-0.669).
Silva et al. ³⁷37. Silva DFO, Bedone AJ, Faundes A, Fernandes AMS, Moura VGAL. Aborto provocado: redução da frequência e gravidade das complicações. consequência do uso de misoprostol? Rev Bras Saúde Mater Infant 2010; 10:441-7.	Campinas, São Paulo State, Brazil (2008-2009).	Cross-sectional. Interviews and extraction of data from charts. Classification of abortion according to WHO criteria.	259 women hospitalized for abortion.	Possibility of unsafe abortions under-reporting with misclassification error; There was no control of confounding factors.	87.7% of certainly induced unsafe abortions, 4.3% probably induced and 10% possibly induced; No differences for the types of abortion regarding age, educational level, number of pregnancies and live-born children. Probably/certainly induced abortion more frequent among women without a stable union than those with a fixed partner (21.8% vs. 9.5%, p = 0.010).
Ramos et al. ³⁸38. Ramos KS, Ferreira ALC, Souza AI. Mulheres hospitalizadas por abortamento em uma Maternidade Escola na Cidade do Recife, Brasil. Rev Esc Enferm USP 2010; 44:605-10.	Recife, Pernambuco State, Brazil (2005-2006).	Cross-sectional. Classification of abortion according to WHO criteria.	160 women hospitalized for abortion. IMIP Hospital.	It is not clear if there were losses and refusals; There was no control of confounding factors.	Hospitalizations for abortion: 3.1% of all obstetric hospitalizations; 14.3% induced abortions and 56.3% possibly induced; Association between higher number of children (p<0.001) and lack of partner (p=0.008) and type of abortion (induced vs. miscarriage).
Chaves et al. ³⁹39. Chaves JHB, Pessini L, Bezerra AFS, Rego G, Nunes R. Abortamento provocado na adolescência sob a perspectiva bioética. Rev Bras Saúde Mater Infant 2010; 10 Suppl 2:S311-9.	Maceió, Alagoas, Brazil (March 2008/April 2009).	Cross-sectional. Classification of abortion according to WHO criteria.	201 adolescents with incomplete abortions admitted to a SUS hospital for curettage.	It is not clear whether all adolescents admitted in the period were considered eligible for the study; There was no control of confounding factors.	81.5% certainly induced abortions, 9.9% probably induced and 6.4% possibly induced; predominance of young people aged 15-19 years and the brown race/color. Most frequent age at first induced abortion = 16 years; No differences between types of abortion regarding ethnicity, marital status, sexual initiation, number of partners, contraception.
Borsari et al. ⁴⁰40. Borsari CMG, Nomura RMY, Benute GRG, Lucia MCS, Francisco RPV, Zugaib M. Aborto provocado em mulheres da periferia da Cidade de São Paulo: vivência e aspectos socioeconômicos. Rev Bras Ginecol Obstet 2013; 35:27-32.	São Paulo, Brazil (2008-2010).	Case control.	33 patients (11 cases and 22 controls) from 2 public hospitals in the periphery.	Unclear sampling procedure; Identification of cases (unsafe abortions) and controls (miscarriage) through direct interviews, subject to misclassification; inadequate comparison group; There was no control of confounding factors.	Cases: induced abortion; Controls: miscarriage; No difference between groups regarding age, marital status and reproductive variables; Differences regarding educational level and income (lower among those with induced abortion) and religion (no evangelicals in the induced abortion group).
Machado et al. ¹⁷17. Machado CJ, Lobato AC, Melo VH, Guimarães MD. Perdas fetais espontâneas e voluntárias no Brasil em 1999-2000: um estudo de fatores associados. Rev Bras Epidemiol 2013; 16:18-29.	Brazil (1999-2000).	Cross-sectional Interviews in the puerperium.	Random sample of 1,838 puerperal women with previous pregnancies from 24 reference maternity hospitals.	Unsafe abortions measured by subtraction of reported miscarriages from total abortions, with the possibility of measurement bias; Does not report the inclusion of the design effect in the data analysis.	Associated factors: age 18-23 years (OR = 0.54); non-white race/color (OR = 1.68); age at 1st relationship < 16 (OR = 1.56) partners >1 (OR = 3.07); absence of prenatal care (OR = 1.97) and STI (OR = 2.25). Number of children: = 1 (0.07) and > 1 (0.08); Not associated: schooling, income, age 1st pregnancy, contraception, marital status.
Fusco et al. ⁴²42. Fusco C, Akerman M, Drezett J, Silva RS. Social determinants of health: from the concept to the practice in outcomes of unintended pregnancies which result in induced abortion. Reprod Clim 2016; 31:22-30.	São Paulo State, Brazil (2005 and 2009-2012).	Cross-sectional. Interviews and extraction of data from medical charts.	153 women (51 from Favela Inajar, São Paulo, 2005, 51 from a public hospital that provides legal abortion and 51 from a private clinic).	Unclear sampling procedure; Different sources of information for the exposure variables used in the analysis; Proportion of incomplete variables not informed; Measurement of unsafe abortions in the community through direct interview, subject to under-reporting.	Compared to abortions performed in the private hospital, illegal and unsafe abortions (Favela) were associated with: lower income (OR = 49.26), lower educational level (15.64), black skin color (OR = 8.61), migrant status (OR = 19.16) and age (OR = 0.88); Legal and safe abortion (public hospital) also associated with lower income (OR = 31.16) and lower educational level (OR = 9.65) and being a migrant (OR = 5.18).
Studies with specific populations
Barbosa et al. ³⁰30. Barbosa RM, Pinho AA, Santos NS, Filipe E, Villela W, Aidar T. Induced abortion in women of reproductive age living with and without HIV/Aids in Brazil. Ciênc Saúde Colet 2009; 14:1085-99.	Thirteen Brazilian municipalities from the 5 macro-regions (2003-2004).	Cross-sectional. Self-administered questionnaire in HIV/AIDS reference services and primary health units and women's health care services.	3,822 sexually active, literate women aged 18-49 years. 1,777 women living with HIV/AIDS; 2,045 users of primary care services.	Serological status of users of primary care units and women’s health services measured through women’s accounts; More than 50% of the women did not know their serological status with the possibility of classification error; Exposure variables measured at the time of the interview and not at the time the abortion occurred.	Factors associated with unsafe abortions among WLHA: Age: linear trend with increase in unsafe abortions as age increases. In women aged 40-49 years (OR = 3.20, 95%IC: 1.60-6.39); Region of residence: Northeast (OR = 2.07, 95%CI: 1.15-3.71), Southeast (OR = 2.38; 95%IC: 1.36-4.15), North (OR = 3,19; 95%IC: 1,39-7,35); Age in the first relationship: increase in OR with decrease in age; ≤ 15 years (OR = 1.80; 95%CI: 1.13-2.86); Number of sexual partners in life: increase in OR with the increase in the number of partners: 3 or + partners: OR = 6.38, 95%CI: 2.49-16.29; Drug use: OR = 1.84, 95%CI: 1.01-3.35; Previous diagnosis of STI: OR = 1.84, 95%CI: 1.31-2.62; Among WNLHA, the same factors as observed among WLHA, except absence of association in the Northeast and Southeast regions; Association with non-white race/color (OR = 1.79, 95%CI: 1.17-2.73) and occasional sexual partner (OR = 1.66, 95%CI: 1.05-2.63).
Pilecco et al. ³¹31. Pilecco FB, Teixeira LB, Vigo A, Dewey ME, Knauth DR. Lifetime induced abortion: a comparison between women living and not living with HIV. PLoS One 2014; 9:e95570.	Porto Alegre, Rio Grande do Sul State, Brazil (2011).	Cross-sectional. Interviews in 7 HIV/AIDS reference services and 27 primary care services.	625 women living with HIV/AIDS (WLHA) and 498 women not living with HIV/AIDS (WNLHA). 18-49 years old, women with previous pregnancies.	Unclear procedure for selecting primary care units; unsafe abortions assessed through direct interview, subject to under reporting; Small number of abortions among WNLHA may have reduced the study’s power to detect significant associations.	Factors associated with unsafe abortions: WLHA: Age at interview (OR = 3.44; 95%CI: 1.18-10.04); 12 or more years of schooling (OR = 3.29; 9%CI: 1.84-5.88); number of sexual partners in life ≥ 5 (OR = 2.34; 95%CI: 1.33-4.14), live children (OR = 3.42, 95%IC: 2.10-5.57), not living with sexual partner (OR = 5.00, 95%CI: 3.35-7.47); WNLHA: 12 or more years of schooling (OR = 8.69, 95%CI: 1.91-39.33), 5 or more sexual partners in life (OR = 3.85, 95%CI: 1.44-10.30).
Madeiro & Rufino ³⁴34. Madeiro AP, Rufino AC. Aborto induzido entre prostitutas: um levantamento pela técnica de urna em Teresina - Piauí. Ciênc Saúde Colet 2012; 17:1735-43.	Teresina, Piauí State, Brazil (2011).	Cross-sectional. Interviews at participants’ workplaces.	310 female sex workers (age: 18-39 years).	No limitations identified.	Only number of pregnancies had significant association: preg III (OR = 3.99) and preg > III. (OR = 27.0); Factors such as age, education, religion, marital status and income were investigated, but did not show statistically significant differences.
Dias et al. ⁴¹41. Dias TZ, Passini Jr. R, Duarte GA, Sousa MH, Faúndes A. Association between educational level and access to safe abortion in a Brazilian population. Int J Gynaecol Obstet 2015; 128:224-7.	São Paulo, Brazil (2010).	Survey.	296 state public servants (both sexes), with reports of unwanted pregnancies.	Response rate of only 10%. It is not clear whether the sample is representative of the population of state public servants; It is not clear whether the variables used in the analysis model refer to the time of the interview or of the unsafe abortions.	Factors associated with unsafe abortions: high educational level only (PR = 1.56); The following factors were not associated: sex, age at the time of pregnancy, marital status, number of children.
Diehl et al. ³⁵35. Diehl A, Pillon SC, Santos MA, Laranjeira R. Abortion and sex-related conditions in substance-dependent Brazilian patients. Cad Saúde Pública 2017; 33:e00143416.	São Paulo, Brazil (2009-2011).	Cross-sectional, interviews 15 days after admission to a specialized clinic for addiction treatment.	616 patients, 82.5% male. Patients aged ≥ 18 years, with confirmed clinical diagnosis of addiction according to DSM-IV-TR.	Unsafe abortions assessed through direct interviews, subject to under-reporting (in men, measurement of experience of induced abortion); Exposure variables measured at the time of the interview and not at the time of the abortion; evaluation of temporality not possible.	In the adjusted analysis, induced abortion associated with: Female: OR = 2.9 (95%CI: 1.75-4.76); single OR = 1.8 (95%CI: 1.13-3.12); unemployment OR = 2.4 (95%CI: 1.46-3.82); tobacco use OR = 1.6 (95%CI: 1.03-2.49); sexual activity in the last 12 months OR = 2.0 (95%CI: 1.17-3.69), irregular condom use (never or occasionally) OR = 1.7 (95%CI: 1.09-2.75), history of STI OR = 2.0 (95%CI: 1.35-3.23), HIV testing OR = 2.2 (95%CI: 1.32-3.53), use of emergency contraception OR = 3.2 (95%CI: 1.29-5.73)

Reference	Location (years)	Study design	Population	Methodological limitations	Results
Hospitalizations
Singh et al. ⁴⁴44. Singh S, Monteiro MF, Levin J. Trends in hospitalization for abortion-related complications in Brazil, 1992-2009: why the decline in numbers and severity? Int J Gynaecol Obstet 2012; 118 Suppl 2:S99-106.	Brazil (1992-2009).	Trend study using data on abortion from SIH/SUS (ICD 10 O00-O08); Adjustments made (changes in ICD in the period: ICD-9 1992-1997, ICD-10 from 1998 to 2009) and separate analysis of abnormal cases (ectopic pregnancies, hydatidiform mole, other abnormal products from conception)	Hospitalizations from abortions in public hospitals - women aged 15-44 years.	No limitations identified.	Reduction of complications from abortion (41%) and in hospitalization rate (57%). Rates in 2009: 3.1/1,000; Sharper decline among more severe complications (69%) than among less severe complications (52%), especially from infection and hemorrhage; Sustained reduction from 1992 to 1997, fluctuating rate from 1998 to 2005 and new, less sharp reduction from 2005 to 2008; Regions N and NE have higher rates of hospitalization from complications, while S region has the lowest.
Mariutti et al. ⁴⁵45. Mariutti MG, Silva HLR, Costa Jr ML, Furegato ARF. Abortamento: um estudo da morbidade hospitalar no país. Rev Bras Med 2010; 67:97-103.	Brazil (2006).	Cross-sectional study using data from SIH-SUS; Does not inform ICD-10 codes used.	Hospitalizations from abortion in public hospitals - women aged 10-49 years.	Unclear method used to identify abortion cases (does not inform which ICD codes were selected for identifying abortion cases).	Rate of 3.81 abortions/1,000 women in the country, higher in North and Northeast regions, lower in South region; States with the highest hospitalization rates: Roraima State and Amapá State; Of the 50 municipalities with the highest rates, 17 located in Bahia State; Several municipalities with no recorded hospitalization due to abortion.
Veras et al. ⁴⁶46. Veras TCS, Mathias TAF. Principais causas de internações hospitalares por transtornos maternos. Rev Esc Enferm USP 2014; 48:401-8.	Paraná State, Brazil (2010).	Cross-sectional study using SIH-SUS data; Cases identified using the main diagnosis of hospitalization, type of discharge, procedures, secondary diagnosis and Intensive Care Unit stay.	Obstetric hospitalizations of women aged 10-49 years in public hospitals of the State of Paraná.	No limitations identified.	Hospitalizations from abortion: 24.1% of hospitalization diagnosis; Ratio of 9.1/100 deliveries; Increase in rate of hospitalization for obstetric complications along with increase in woman’s age; Among women aged 35 to 49 years: 36.9% of hospitalizations for abortion; rate of 20.2/100 deliveries; 3.6% of deaths that occurred in the hospitalizations had abortion as the main hospitalization diagnosis.
Bonassa et al. ⁴⁷47. Bonassa RT, Rosa MI, Madeira K, Simões PW. Caracterização de casos de internação por abortos complicados na Macrorregião Sul Catarinense. Arq Catarin Med 2015; 44:88-100.	Santa Catarina State, Brazil (1999-2010).	Ecological study, using data from SIH-SUS (ICD-10 codes O03, O04 and O05).	Hospitalizations for abortion of women aged 15-49 years in public hospitals in Santa Catarina.	Did not include all ICD codes related to abortion; Time series with no statistical tests to evaluate trends.	Reduction in the abortion rate in the period; mean rate 2.1/1,000; Lower rate in 2007 (1.7/1,000); Higher among women aged 20-29 years and those who had “other pregnancies ending in abortion”; Regional differences found.
Complications from abortion
Araújo & Mochel ⁴⁸48. Araújo MCR, Mochel EG. Aborto provocado: fatores associados em mulheres admitidas em maternidades públicas em São Luís, Maranhão, Brasil. Rev Paul Enferm 2008; 27:79-86.	São Luiz, Maranhão State, Brazil (2006).	Case series in 2 public maternity hospitals; Interviews + chart data.	80 unsafe abortions cases	Unsafe abortions measured through direct interview, subject to under-reporting; It is unclear if complications were obtained from interviews or charts; Data presented jointly, with no characterization of the services; Only descriptive case analysis.	17.5% of unsafe abortions in 456 hospitalizations for complications from abortion; 56.25% used misoprostol alone and 8.75% used misoprostol in combination with other method; 41.25% with complaints of cramps and bleeding, 27.5% only bleeding, 25% only cramps, 2.5% fever; 46.25% with hospitalization of only 1 day. Mean hospitalization duration of 2.5 days.
Correia et al. ⁴⁹49. Correia DS, Monteiro VG, Egito ES, Maia EM. Aborto provocado na adolescência: quem o praticou na Cidade de Maceió, Alagoas, Brasil. Rev Gaúcha Enferm 2009; 30:167-74.	Maceió, Alagoas, Brazil (2005).	Cross-sectional; 10 public and private schools; Self-administered questionnaire.	2,592 students of both sexes, 12-19 years old; This analysis only included reported cases of unsafe abortions (N = 149).	Inadequate description of schools and students; Possibility of unsafe abortions under-reporting; Complications and hospitalizations measured through interviews with students, subject to information bias.	149 abortions reported, 16.1% with reports of complications and 10.1% with hospitalization.
Silva et al. ³⁷37. Silva DFO, Bedone AJ, Faundes A, Fernandes AMS, Moura VGAL. Aborto provocado: redução da frequência e gravidade das complicações. consequência do uso de misoprostol? Rev Bras Saúde Mater Infant 2010; 10:441-7.	Campinas, São Paulo State, Brazil (2008-2009).	Cross-sectional, hospital-based study; Abortion classified according to WHO criteria.	538 women hospitalized due to abortion; Emphasis on 259 women with induced abortions or abortions suspected of being induced.	Likely classification error of certainly induced cases of abortion (under-estimation due to women’s low accounts); Analysis of association between misoprostol use and complications limited by the small number of women who reported inducing the abortion; Analysis limited by lack of identification and adjustment by confounding variables; Limitation in the measurement of the outcome variable (complications), no specification if occurrence was before or after the abortion.	48% of women with induced abortion or abortion suspected of being induced (85.7% possibly induced, 4.3% probably induced, 10% certainly induced); Complications from abortion: infectious 10% and hemorrhagic 13%; Significantly higher prevalence of hemorrhagic (24.0% vs. 8.5%, p = 0.027) and infectious (32.0% vs 11.5%, p = 0.010) complications among women with certainly induced abortions; Use of misoprostol: 36% of certainly induced abortions, 3.5% of total suspected of being induced; Misoprostol use not associated with lower occurrence of infectious (22% vs. 38%, NS) and hemorrhagic complications (11% vs. 31%, NS).
Chaves et al. ³⁹39. Chaves JHB, Pessini L, Bezerra AFS, Rego G, Nunes R. Abortamento provocado na adolescência sob a perspectiva bioética. Rev Bras Saúde Mater Infant 2010; 10 Suppl 2:S311-9.	Maceió, Alagoas, Brasil (2008-2009).	Cross-sectional, hospital-based; Abortion classified according to WHO criteria.	221 pregnant adolescents hospitalized with an abortion diagnosis who underwent uterine curettage; 9.9% of loss.	It is unclear if all women hospitalized in the study period were considered eligible, nor if there were losses and refusals; Authors compared type of abortion and complications without identification and adjustment of confounding factors; Complications measured using woman’s account, without consulting chart data, subject to information bias; Small number of outcomes limited statistical analyses.	81.5% of certainly induced abortions and 9.9% of probably induced abortions; Misoprostol use: 77.4% in certainly induced abortions; Complications: 1.5% uterine perforation, 4.4% blood transfusions; All cases of uterine perforation and blood transfusion happened to women with certainly or probably induced abortions.
Dias et al. ⁴¹41. Dias TZ, Passini Jr. R, Duarte GA, Sousa MH, Faúndes A. Association between educational level and access to safe abortion in a Brazilian population. Int J Gynaecol Obstet 2015; 128:224-7.	Campinas, São Paulo State, Brazil (2010).	Cross-sectional study, restricted base; Non-probabilistic sample of public servants; Self-administered questionnaire.	1660 public servants (1,240 women, 444 men); Response rate of 10%.	Sample of 10% of public servants, unclear if it is representative of the study universe. General characteristics of the sample not described, only that of women with unplanned pregnancies and induced abortion; characteristics of women with terminations performed by doctors or another person not described, preventing the identification of potential confounding variables; outcome variable (complications) measured only through woman’s account; theoretical model for choice of analysis variables not explained.	17.8% reported unplanned pregnancies, of which 55.7% resulted in abortion, 70% of which before 1990; 62% of abortions performed by doctors; 22.9% of women with induced abortions required medical care and 16.6% were hospitalized; Abortion performed by a doctor associated with fewer complications (11.8% vs. 40.0%, p < 0.001) and hospitalizations (7.5% vs. 29.7%, p < 0.002).
Gomperts et al. ⁵⁰50. Gomperts R, van der Vleuten K, Jelinska K, Costa CV, Gemzell-Danielsson K, Kleiverda G. Provision of medical abortion using telemedicine in Brazil. Contraception 2014; 89:129-33.	Brazil (2011).	Case series; Restricted base; Information provided by women.	370 women who had a medical abortion provided by Women on Web; Information on outcome available for 307 women; 17% loss.	27.6% loss to follow-up of women who registered and received guidance and medication for terminating pregnancy; Information on gestational age at termination and occurrence of complications provided only by the woman, subjected to information bias;	67.1% without children, 66.1% sought a health service due to method failure; Outcome: 2.3% remained pregnant, 76.9% had complete abortions, 20.9% underwent surgical procedures (only 12.5% with symptoms compatible with procedure recommendations); Greater proportion of procedures among women who had medical abortions at 13 weeks of gestation or more.
Severe maternal morbidity, maternal near miss and maternal mortality
Camargo et al. ¹¹11. Camargo RS, Santana DS, Cecatti JG, Pacagnella RC, Tedesco RP, Melo Jr. EF, et al. Severe maternal morbidity and factors associated with the occurrence of abortion in Brazil. Int J Gynaecol Obstet 2011; 112:88-92.	Brazil (2006).	Cross sectional study (PNDS 2006); population base, national scope, urban and rural area; Maternal morbidity measured through interviews with women.	15,575 women aged 15-49 years.	Complications from miscarriages and induced abortions analyzed jointly, which may have underestimated the frequency of complications among induced abortions. Analysis without adjustment for confounding variables.	Greater occurrence of complications among women who had an abortion as the outcome of the pregnancy when compared with those who had a delivery as the outcome: hemorrhage (PR = 2.54, 95%CI: 1.85-3.49), infection (PR = 2.89, 95%CI: 1.34-6.24), any complication defined as severe maternal morbidity (PR = 2.29, 95%CI: 1.73-3.04).
Santana et al. ⁵¹51. Santana DS, Cecatti JG, Parpinelli MA, Haddad SM, Costa ML, Sousa MH, et al. Severe maternal morbidity due to abortion prospectively identified in a surveillance network in Brazil. Int J Gynaecol Obstet 2012; 119:44-8.	Brazil (2009-2010).	Cross-sectional, hospital-based study; 27 reference hospitals located in all regions of the country; Data from hospital charts.	9555 women hospitalized with obstetric complications.	Authors did not inform the criteria used for classification of miscarriages and induced abortions; Theoretical model used for the selection of confounding variables not explained; Adjustment made for a large number of variables, which may have limited the statistical analysis.	2.5% of cases of severe maternal morbidity, maternal near miss and maternal death related to abortion; Abortion cases with a higher proportion of maternal near miss than other complications (PR = 1.93, 95%CI: 1.12-3.31); Mortality rate in cases of abortion higher (16.3%) than in other complications (11.6% ectopic pregnancy, 15.3% other causes); Higher prevalence of infection among unsafe abortions than among safe abortions (42.9% x 3.3%, p < 0.002); Factors associated with greater severity of complications in women who had abortions: preexisting diseases (sickle cell anemia and thalassemia, PR = 4.78, 95%CI: 2.85-8.01), low maternal weight (PR = 18.78 95% CI 5.44-64.81), pre-existing conditions (neoplasms, PR = 2.33; 95%CI: 1.37-3.94), access to services (transfer or referral, PR = 2.53, 95%CI: 1.39-4.62), previous uterine scarring (PR = 2.07, 95%CI: 1.08-3.97) and any delay in receiving adequate attention (PR = 2.80, 95%CI: 1.09-7.16).
Souza et al. ⁵²52. Souza KV, Almeida MRCB, Soares VMN. Perfil da mortalidade materna por aborto no Paraná: 2003-2005. Esc Anna Nery Rev Enferm 2008; 12:741-9.	Paraná State, Brazil (2003-2005).	Cross-sectional study; Maternal deaths assessed by the state committee for the prevention of maternal mortality; MMR (total and abortion-specific) in the trienniums 1997-1999, 2000-2002, 2003- 2005 and analysis of the deaths from abortion in the triennium 2003-2005.	306 maternal deaths from 1997 to 2005 and 17 maternal deaths from abortion from 2003 to 2005.	8% of deaths of women of reproductive age from 2003 to 2005 not investigated, possible under-reporting of maternal deaths; Descriptive analysis of the cases (proportional distribution without calculating MMR from abortion according to women’s characteristics); Some variables with a high proportion of uninformed data.	MMR for abortion per 100,000 live birth: 1997-1999 = 3.7; 2000-2002 = 4,3; 2003-2005 = 3,6; Of the 17 deaths from abortion in 2003-2005, 43.7% aged 20-29 years, 75% white, 64.3% less than 8 years of schooling, 40% 2 to 4 children, 33% first pregnancy. 59% of deaths caused by infection.
Souza et al. ⁵³53. Souza ML, Ferreira LAP, Burgardt D, Monticelli M, Bub MBC. Mortalidade por aborto no Estado de Santa Catarina: 1996 a 2005. Esc Anna Nery Rev Enferm 2008; 12:735-40.	Santa Catarina State, Brasil (1996-2005).	Cross-sectional study, SIM data; Maternal deaths from abortion excluding ectopic pregnancies, hydatidiform mole and other abnormal products of conception.	31 maternal deaths associated with abortion.	Does not provide data on the investigation of deaths of womens of reproductive age that would enable assessment of possibility of under-reporting of maternal deaths in general and from abortion specifically; Descriptive analysis with calculation of MMR from abortion only according to region of residence.	MMR for abortion: 1.5 per 100,000 live birth, ranging from 1.3 to 5.1 in the six regions of the state; 45.16% among women aged 20-29 years, 51.6% married women, 38.7% with 1 to 8 years of schooling.
Figueiredo et al. ⁵⁴54. Figueiredo YMD, Malta DC, Rezende EM. Análise da mortalidade materna no município de Governador Valadares, 2002-2004. REME Rev Min Enferm 2010; 14:376-85.	Governador Valadares, Minas Gerais State, Brasil (2002-2004).	Cross-sectional study, data from SIM, SINASC and reports from maternal mortality committees.	5 maternal deaths.	One item not applicable (response rate).	5 identified maternal deaths, 4 direct obstetric deaths, 3 from complications from abortions (2 infected abortions and 1 from hemorrhage); Abortion as the cause of 60% of maternal deaths and 75% of direct obstetric deaths in the three-year period.
Martins et al. ⁵⁵55. Martins EF, Almeida PF, Paixão CO, Bicalho PG, Errico LS. Causas múltiplas de mortalidade materna relacionada ao aborto no Estado de Minas Gerais, Brasil, 2000-2011. Cad Saúde Pública 2017; 33:e00133115.	Minas Gerais, Brasil (2000-2011).	Ecological study, time series; data from the SIM, underlying and associated causes.	82,790 deaths of women of reproductive age (10-49 years), of which 1,219 were maternal deaths.	No limitations were identified; One item not applicable (response rate).	15% of maternal deaths caused by abortion (n = 183); RMM from abortion stable in the period; 70.5% of the deaths from abortion among black women, 72.7% among women 20 to 34 years, single (68%) and with low educational level (40% missing); Multiple causes/underlying cause ratio of 1.38: 38% increase of abortion-related deaths after inclusion of associated causes; Main basic causes: unspecified abortion (33.8%), tubal pregnancy (21.1%), attempt failure (16.5%); Among the deaths identified by associated causes, 44% were not declared as maternal deaths in the underlying cause; Most frequent diagnoses: genitourinary disorders (22.7%) and unspecified sepsis (18.2%).
Galvão et al. ⁵⁶56. Galvão LP, Alvim-Pereira F, Mendonça CM, Menezes FE, Góis KA, Ribeiro Jr. RF, et al. The prevalence of severe maternal morbidity and near miss associated factors in Sergipe, Northeast Brazil. BMC Pregnancy Childbirth 2014; 14:25.	Sergipe, Brazil (2011-2012).	Cross-sectional nested case-control study. Hospital database, hospital chart data.	16,243 deliveries, 1,102 cases of severe maternal morbidity, 77 cases of maternal near miss and 17 maternal deaths. In the case-control study, 77 cases of maternal near miss and 151 controls.	Limitations in statistical analysis: theoretical model for selection of confounding variables not explained; some variables poorly defined (for example, complications in pregnancy; type of delivery, with no differentiation between antepartum and intrapartum cesarean sections); variable “degree of consciousness”, used in adjustment, is one of the components of the analyzed outcome (maternal near miss).	Maternal near miss ratio 5.8/1,000, severe maternal morbidity/maternal near miss 72.6/1,000, maternal near miss/maternal morbidity ratio 4.5/1, mortality rate 18%; Abortion as a cause of 11.8% of maternal deaths in the period; Previous abortion associated with maternal near miss in the current pregnancy.
Madeiro et al. ⁵⁷57. Madeiro AP, Rufino AC, Lacerda ÉZ, Brasil LG. Incidence and determinants of severe maternal morbidity: a transversal study in a referral hospital in Teresina, Piauí, Brazil. BMC Pregnancy Childbirth 2015; 15:210.	Teresina, Piauí Brazil, (2012-2013).	Cross-sectional with a nested case-control study. Hospital database, hospital chart data.	5,841 births, 343 cases of severe maternal morbidity, 56 cases of maternal near miss and 10 cases of maternal death.	Limitations related to the analysis of factors associated with outcomes: absence of a theoretical model, problems in variable selection and definition.	Maternal near miss ratio 9.6/1,000; MMR = 171.2/100,000; Ratio of severe outcomes 11.2/1,000; Abortion as a cause of 1.8% of maternal near miss cases and 0.3% of cases of severe maternal morbidity; Infected abortion as the most frequent single cause of maternal death (30%); Cesarean section in current pregnancy was the only factor associated with maternal near miss in the final model.
Kale et al. ⁵⁸58. Kale PL, Jorge MHPM, Fonseca SC, Cascão AM, Silva KS, Reis AC, et al. Deaths of women hospitalized for childbirth and abortion, and of their concept, in maternity wards of Brazilian public hospitals. Ciênc Saúde Colet 2018; 23:1577-90.	Rio de Janeiro, Niterói and São Paulo, Brazil (2011).	Cross-sectional hospital-based study (public hospitals: 4 in São Paulo, 1 in Niterói, 1 in Rio de Janeiro); Interviews with women, in addition to extraction of data from medical charts, prenatal card and SIM.	7,845 women.	Non-random sample of services; Type of participant selection in each service not informed. Abortion classification criteria not informed; Study design not suitable for rare events, resulting in only one identified maternal death.	One maternal death was identified (MMR 13.6 per 100,000 live birth), not associated with abortion; 498 abortions; Proportion of unsafe abortions: Rio de Janeiro:11.9%; São Paulo: 1%; Misoprostol as the most commonly reported method of termination.
Mental health outcomes
Benute et al. ⁵⁹59. Benute GR, Nomura RM, Pereira PP, Lucia MC, Zugaib M. Abortamento espontâneo e provocado: ansiedade, depressão e culpa. Rev Assoc Med Bras 2009; 55:322-7.	Location not informed, (2001-2002).	Cross-sectional, hospital-based, interviews.	50 women with spontaneous abortion and 50 women with unsafe abortions admitted to the Emergency Room of a University Hospital. Interview 30 days after abortion.	No information on parameters for sample calculation, procedures for case identification and inclusion, number of losses and refusals, and technique used to measure unsafe abortions; Inadequate analysis model, with no strategies for controlling confounding.	Depression and anxiety scores measured through the Hospital Anxiety and Depression scale, higher in women with unsafe abortions than in women with miscarriage; Anxiety: 11.0 (presence of anxiety) x 8.7 (likely presence), p < 0.05; Depression: 8.3 (likely presence of depression) x 6.1 (absence of depression), p < 0.05.
Ludermir et al. ⁶⁰60. Ludermir AB, de Araújo TV, Valongueiro SA, Lewis G. Common mental disorders in late pregnancy in women who wanted or attempted an abortion. Psychol Med 2010; 40:1467-73.	Recife, Pernambuco, Brazil (2005-2006).	Cross-sectional (prospective cohort baseline); Population base (low income, 15% of the Recife population); Interviews using SRQ-20 to measure CMD.	1,133 pregnant women aged 18 to 49 years in the third trimester of pregnancy; 1,104 with information for all variables.	It is unclear if measuring unsuccessful abortion attempts during pregnancy using direct interviews is a valid and reliable method.	13,7% of pregnant women attempted abortion; Prevalence of CMD: 43.1% (total sample), 63.6% (women who attempted abortion); Attempted abortion associated with common mental symptoms: score >7 (adjusted OR = 2.05, 95%CI: 1.3-3.1), score > 11 (adjusted OR = 1.73, 95%CI: 1.1-2.8).
Ludermir et al. ⁶¹61. Ludermir AB, Araya R, Araújo TV, Valongueiro SA, Lewis G. Postnatal depression in women after unsuccessful attempted abortion. Br J Psychiatry 2011; 198:237-8.	Recife, Pernambuco State, Brazil (2005-2006).	Prospective cohort. Population base (low-income women, 15% of the Recife population); Interviews with women; Use of the SRQ-20 to measure CMD and the Edinburgh scale to measure postpartum depression.	1,057 pregnant women aged 18 to 49 years who completed the postnatal interview (94.3% of the baseline).	It is unclear if measuring unsuccessful abortion attempts during pregnancy using direct interviews is a valid and reliable method.	Prevalence of postpartum depression: women who did not consider abortion (22%); women who wanted an abortion but did not attempt one (32%); women who attempted abortion (41%); Attempted abortion associated with postpartum depression: adjOR 1.59 (95%CI: 1.0-2.5); Average score on the Edinburgh scale: women who did not consider abortion (7.25); women who wanted an abortion but did not attempt one (9.49); women who attempted abortion (10.66); Non-significant differences after adjusting for other variables (adjusted OR = 1.43, 95%CI: 0.4-2.4).

Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz Rua Leopoldo Bulhões, 1480 , 21041-210 Rio de Janeiro RJ Brazil, Tel.:+55 21 2598-2511, Fax: +55 21 2598-2737 / +55 21 2598-2514 - Rio de Janeiro - RJ - Brazil
E-mail: cadernos@ensp.fiocruz.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] Correspondence
R. M. S. M. Domingues Instituto Nacional de Infectologia Evandro Chagas, Fundação Oswaldo Cruz. Av. Brasil 4365, Rio de Janeiro, RJ 21040-360, Brasil. rosa.domingues@ini.fiocruz.br