In 2009, after decades of using maternal mortality as the most important health indicator for women experiencing the reproductive process, the definition of maternal near miss (MNM) with its correspondent criteria was published. Women suffering a severe complication during pregnancy, childbirth, or within 42 days of the postpartum period, who almost died, but survived due to luck or effective interventions are now considered
The need for better exploring the concept of maternal morbidity instead of maternal mortality arose because, fortunately, maternal deaths became rare in several settings when using absolute numbers and, therefore, it became more difficult to understand and identify factors or conditions that could possibly be associated with its occurrence. There are also a number of reasons already publicized on the advantages of looking into cases of morbidity and not only to maternal deaths [
However, until recently, the concept of maternal morbidity and the levels of severity it could present were definitely not standardized, and therefore they were used indistinctly in the scientific literature with different meanings. This, of course, represented a restriction for using the concept as a real health indicator that could be used for policy changes or even as a starting point for interventions and for comparisons between different settings or in the same location across different periods. With the WHO definition and criteria [
Taking these limitations into account and the availability of several possible sources of information on maternal morbidity, a considered approach was to build a pragmatic definition of maternal near miss that could be even retrospectively applied to obtain relevant information on the condition, in order to direct next steps for policy changes.
The opportunity came with the information already collected from the WHO Global Survey on Maternal and Perinatal Health. Several different combinations of conditions that were hypothesized as possible predictors of maternal death due to its severity were tested in their predictive capacity of identifying cases of women who died from a maternal cause. The group of most common criteria with the highest accuracy for this prediction included hysterectomy due to haemorrhage or infection, admission to intensive care unit (ICU), blood transfusion, and eclampsia. Altogether, they are now the pragmatic criteria established by the WHO [
Although already officially recommended by WHO for gathering information on maternal morbidity and also for appraising the quality of maternal healthcare [
Keeping these points into account, the objective of the current study is to perform an exploratory analysis of the database on maternal and neonatal information for childbirths occurring in several maternity hospitals located in the Latin America and Caribbean region and coordinated by CLAP, the Latin American Centre for Perinatology, Women and Reproductive Health from the Department of Family, Gender and Life Course of PAHO. The product of this analysis may be useful for identifying a group of severity markers taken as a proxy or pragmatic criteria for levels of severity in maternal morbidity and as well for building a general profile of maternal morbidity for Latin America and the Caribbean region.
Information on childbirth in the region of Latin America and the Caribbean, for both maternal and neonatal health conditions, has been routinely stored in the SIP (Perinatal Information System) database, when shared by countries or health institutions, during more than 25 years [
Therefore, although the full SIP database includes more than 4 million records, for the specific objectives currently addressed we assessed and analysed a more recent SIP database containing standardized information on over 700 thousand childbirths which occurred between 2009 and 2012. In this database, there is information from childbirths which occurred in some health facilities from 12 countries from the South Cone, Andean, Central America, and Caribbean subregions from America (Argentina, Bolivia, Colombia, El Salvador, Ecuador, Guatemala, Guiana, Honduras, Haiti, Nicaragua, Paraguay, and Uruguay). The sample for each country does not necessarily represent all births in the country for the period nor is proportional to the population size and therefore is not supposed to be representative of the respective country. This is the reason why the data does not allow for analysis of specific country’s reports. With data routinely collected from all births in the period occurring in the participating health facilities, the methodological approach of the current exploratory analysis is that of a cross-sectional study. No sample size was previously estimated because of the huge number of women with data available, although this was not a population-based study. Missing information for variables used was assumed to be randomly distributed among centres, countries, and time and associated neither with predictors nor with outcomes. Although the system used for data collection was built to immediately check for internal consistency and also those in charge of feeding the online platform with data from health facilities received training following standardized operating procedures and instructions, for additional quality control, several cross-checking instances between variables of interest were performed for assessing the consistency of the database. This was performed to improve the quality of data collected. However, of course, the data refers only to cases from centres that voluntarily applied for using the system. No information at all is available for childbirths occurring at home, at the health centre, or at any other health facility not using the system.
This period was chosen because data were gathered with a data collection form standardized for all participating facilities and countries and because this is the period immediately before the introduction of some specific information for identifying maternal near miss cases with the WHO definition and criteria [
In order to achieve the objectives of identifying different levels of severity for maternal morbidity, in the current analysis as below outlined, we built the groups using some possible identifiers from the database: Group of maternal death (MD): with the purpose of identifying the cases of maternal deaths, this group included the women with information that death had occurred at the health facility or who died during transport or in the place they were transferred to. This was used descriptively to estimate maternal mortality ratio in the sample, however with no classification and no attribution of cause of maternal death because this information was not reliably available in the database. Group of maternal near miss (MNM): to identify the cases with more severe conditions of maternal morbidity that could serve as a proxy for maternal near miss (supposedly with an organ dysfunction or failure), this group (MNM) was identified using some pragmatic criteria already described and even developing a new set of criteria within the database. This was the group with greater difficulty to be identified. According to the WHO pragmatic criteria as a proxy for MNM [ Group of potentially life-threatening conditions (PLTC): to identify the cases that could be classified as PLTC in a proxy to these exact conditions as defined by WHO [ Group of less severe maternal morbidity (LSMM): to identify the cases of less severe maternal morbidity, including all other remaining conditions identified as maternal morbidities with specific information collected in the SIP, this group constituted cases presenting any morbidity recorded in the SIP database, other than those above described. For this group, the following conditions were taken into account: anaemia, HIV+, diabetes mellitus, hospital admission during pregnancy, ovular infection, urinary tract infection, another maternal pathological condition, retained placenta, or need to be referred. This was mainly based on the recently issued maternal morbidity matrix, in its dimension one, including symptoms, signs, investigations, and managements related to both direct and indirect causes of maternal morbidity [ Group of any maternal morbidity (AMM): to identify all cases of maternal morbidity (joining groups 1 to 4), from the most severe (death) to the less severe ones, this is also aligned with the recent WHO proposal for identifying, classifying, and building a full profile of maternal morbidity occurring worldwide [ Group of women with no maternal morbidity: this group represents the remaining women who did not experience any maternal morbidity during childbirth and postpartum period as previously described and survived the event. The remaining women who experienced no identifiable maternal morbidity during the childbirth process were used to build a comparison group for the above-mentioned maternal morbidity and mortality groups, to generate information on factors possibly associated with worse outcomes.
Abruptio placentae Accreta/increta/percreta placenta Ectopic pregnancy Postpartum Haemorrhage Ruptured uterus
Endometritis Pulmonary oedema Respiratory failure Seizures Sepsis Shock Thrombocytopenia <100.000 Thyroid crisis
Acute cyanosis Gasping Respiratory rate > 40 or <6/min Shock Oliguria non-responsive to fluids or diuretics Clotting failure
Oxygen saturation < 90% for ≥60 minutes PaO2/FiO2 < 200 mmHg Creatinine ≥ 300 mmol/l or ≥ 3.5 mg/dl Bilirubin > 100 mmol/l or >6.0 mg/dl
Use of continuous vasoactive drugs Hysterectomy following infection/haemorrhage Transfusion of ≥5 units red cell transfusion
The main purpose of the current analysis was to explore the capacity of a big database of a birth registry from Latin America and the Caribbean region to retrospectively identify different levels of maternal morbidity according to what is recently recommended by WHO. It was a methodological exploration approach. The database was not originally built with the objective of an in-depth analysis of maternal mortality and morbidity. Therefore, the analyses allow for neither country variations nor a general and representative scenario of the whole region regarding maternal morbidity. No woman, health facility, or countries were identified and this information is not available in the database. The dataset is regularly fed with the previous agreement of countries that analyses of data are performed at regular intervals. The protocol for the current analyses was previously ethically evaluated and approved by CLAP/WR. All the recommendations of the Declaration of Helsinki for studies involving human beings were strictly followed.
Initially, descriptive analyses were performed, with the prevalence of these adverse maternal outcomes determined together with their correspondent health indicators. Some of these health indicators had already been described and used [
For categorical variables, chi-square tests were used, with
The database from SIP used for the current analysis has information on over 712 thousand women from 12 countries of Latin America and Caribbean region, who were admitted to any health facilities using SIP routinely, for delivery or management of any complication associated with pregnancy. How these women were classified is presented in Figure
Prevalence of adverse maternal outcomes and correspondent health indicators. CLAP 2009–2012.
Maternal outcomes |
|
% | Health indicators |
---|---|---|---|
MD | 1,028 | 0.14 | MMR = 147.3/100,000 LB |
MNM | 21,985 | 3.1 | MNMR = 31.5/1000 LB |
|
|
|
|
PLTC | 110,038 | 15.5 | PLTCR = 157.7/1000 LB |
LSMM | 137,589 | 19.3 | LSMMR = 197.2/1000 LB |
AMM | 270,640 | 38.0 | AMMR = 387.8/1000 LB |
No morbidity | 441,441 | 62.0 | |
|
|||
Total | 712,081 | 100.0 | Total LB: 697,820 |
AMM: any maternal morbidity; AMMR: any maternal morbidity ratio; LB: live births; LSMM: less severe maternal morbidity; LSMMR: less severe maternal morbidity ratio; MD: maternal death; MMR: maternal mortality ratio; MNM: maternal near miss; MNMR: maternal near miss ratio; PLTC: potentially life-threatening condition; PLTCR: potentially life-threatening condition ratio; SMO: severe maternal outcome; SMOR: severe maternal outcome ratio.
Flow chart of women in the study (AMM: any maternal morbidity; LSMM: less severe maternal morbidity; MD: maternal death; MNM: maternal near miss; PLTC: potentially life-threatening condition; SMO: severe maternal outcomes).
In Table
Prevalence of some previous pathological maternal conditions according to the type of adverse maternal outcomes. CLAP 2009–2012.
Previous conditions | MD | MNM | PLTC | LSMM | Any (AMM) | No morbidity |
|
---|---|---|---|---|---|---|---|
Diabetes |
8 (0.8) | 230 (1.1) | 861 (0.8) | 1681 (1.3) | 2780 (1.1) | 1043 (0.2) | <.001 |
Hypertension |
32 (3.2) | 952 (4.5) | 1971 (1.8) | 2310 (1.7) | 5265 (2.0) | 4704 (1.1) | <.001 |
Preeclampsia |
27 (2.7) | 730 (3.4) | 1403 (1.3) | 2136 (1.6) | 4296 (1.6) | 3750 (0.9) | <.001 |
Eclampsia |
12 (1.2) | 125 (0.6) | 176 (0.2) | 213 (0.3) | 526 (0.2) | 594 (0.1) | <.001 |
Other severe conditions |
26 (2.6) | 873 (4.1) | 5014 (4.7) | 3603 (2.7) | 9516 (3.6) | 5730 (1.3) | <.001 |
Cardiac disease |
7 (0.7) | 227 (1.1) | 194 (0.2) | 180 (0.1) | 608 (0.2) | 374 (0.1) | <.001 |
Renal disease |
1 (0.1) | 105 (0.5) | 173 (0.2) | 141 (0.1) | 420 (0.2) | 274 (0.1) | <.001 |
Any previous condition |
90 (9.1) | 2542 (13.0) | 8723 (8.3) | 8834 (10.4) | 20189 (9.6) | 14493 (3.5) | <.001 |
The estimated risks for occurrence of maternal death (MD) were higher for women from ethnic group other than white (indigenous or black people, 19-fold), with any previous pathological condition (2-fold), if labour was induced (1.6-fold) or birth was through an elective C-section (2-fold), as shown in Table
Crude (PR) and adjusted (APR) estimated risks of maternal death (MD) and of maternal near miss (MNM) according to some maternal and obstetric characteristics. CLAP 2009–2012.
Characteristics | No morbidity | MD | PR (95% CI) | APR (95% CI) | MNM | PR (95% CI) | APR (95% CI) |
---|---|---|---|---|---|---|---|
Maternal age (years) |
|||||||
10–19 | 108,907 | 276 | 1.15 [0.89–1.48] | 1.43 [0.87–2.35] | 5,689 |
|
|
20–24 | 128,968 | 285 | Ref. | Ref. | 5,530 | Ref. | Ref. |
25–29 | 96,493 | 187 | 0.88 [0.74–1.04] |
|
4,433 | 1.07 [0.98–1.16] | 1.06 [0.97–1.17] |
30–34 | 64,458 | 169 | 1.19 [0.89–1.58] | 0.79 [0.61–1.02] | 3,432 |
|
1.12 [0.95–1.32] |
35–55 | 40,585 | 109 | 1.21 [0.67–2.19] | 0.72 [0.37–1.40] | 2,811 |
|
|
|
|||||||
Ethnicity/skin colour |
|||||||
White | 91,116 | 175 | Ref. | Ref. | 6,556 | Ref. | Ref. |
Mixed | 298,612 | 231 | 0.40 [0.11–1.42] |
|
12,082 | 0.58 [0.29–1.14] | 0.70 [0.47–1.04] |
Others | 19,328 | 594 |
|
|
1,749 | 1.24 [0.76–2.00] | 1.09 [0.84–1.42] |
|
|||||||
Literacy |
|||||||
No or primary | 203,185 | 379 | 0.68 [0.41–1.12] | 0.90 [0.58–1.37] | 8,418 |
|
0.96 [0.79–1.16] |
Secondary or university | 217,307 | 600 | Ref. | Ref. | 12,616 | Ref. | Ref. |
|
|||||||
Marital status |
|||||||
Married + stable part | 364,380 | 825 | Ref. | Ref. | 17,369 | Ref. | Ref. |
Single + other | 56,037 | 142 | 1.12 [0.66–1.89] | 0.87 [0.64–1.18] | 3,695 |
|
|
|
|||||||
Parity |
|||||||
Nullipara | 153,659 | 350 | 0.93 [0.71–1.23] | 0.64 [0.36–1.12] | 9,963 |
|
|
Multipara | 250,299 | 612 | Ref. | Ref. | 10,577 | Ref. | Ref. |
|
|||||||
Any previous condition |
|||||||
Yes | 14,493 | 90 |
|
|
2,542 |
|
|
No | 398,396 | 897 | Ref. | Ref. | 17,059 | Ref. | Ref. |
|
|||||||
Number of prenatal care visits |
|||||||
0 | 30,445 | 60 | 0.84 [0.47–1.51] | 1.05 [0.52–2.11] | 1,910 | 1.22 [0.85–1.74] |
|
1–4 | 115,059 | 328 | 1.21 [0.91–1.62] | 1.00 [0.81–1.22] | 5,623 | 0.96 [0.79–1.18] |
|
>4 | 252,273 | 592 | Ref. | Ref. | 12,822 | Ref. | Ref. |
|
|||||||
Smoking, drugs, alcohol, or violence |
|||||||
Yes | 30,559 | 76 | 0.95 [0.16–5.46] | 0.87 [0.48–1.58] | 2,255 |
|
|
No | 288,779 | 760 | Ref. | Ref. | 12,176 | Ref. | Ref. |
|
|||||||
Previous C-section |
|||||||
Yes | 50,662 | 149 | 1.27 [0.98–1.63] | 0.73 [0.46–1.18] | 2,996 | 1.12 [0.94–1.34] |
|
No | 303,471 | 705 | Ref. | Ref. | 15,929 | Ref. | Ref. |
|
|||||||
Onset of labour |
|||||||
Spontaneous | 340,966 | 627 | Ref. | Ref. | 12,910 | Ref. | Ref. |
Induced | 20,578 | 66 | 1.74 [0.70–4.34] |
|
1,436 |
|
1.20 [0.89–1.62] |
Elective C-section | 50,673 | 289 |
|
|
6,408 |
|
|
|
|||||||
Mode of delivery |
|||||||
C-section | 123,423 | 447 |
|
1.19 [0.94–1.50] | 12,877 |
|
|
Vaginal (any) | 310,183 | 563 | Ref. | Ref. | 8,864 | Ref. | Ref. |
|
|||||||
Total | 441,441 | 1028 | 21,985 |
Missing information for a: 0.4%; b: 6.9%; c: 4.1%; d: 4.2%; e: 7.6%; f: 12.4%; g: 7.6%; h: 31.1%; i: 16.8%; j: 5.4%; k: 1.2% of cases; APR: adjusted prevalence ratio (adjusted for cluster effect and all other predictors); MD: maternal death; MNM: maternal near miss.
Table
Crude (PR) and adjusted (APR) estimated risks of potentially life-threatening condition (PLTC) and of less severe maternal morbidity (LSMM) according to some maternal and obstetric characteristics. CLAP 2009–2012.
Characteristics | No morbidity | PLTC | PR (95% CI) | APR (95% CI) | LSMM | PR (95% CI) | APR (95% CI) |
---|---|---|---|---|---|---|---|
Maternal age (years) |
|||||||
10–19 | 108,907 | 25468 | 1.03 [0.92–1.15] | 1.05 [0.93–1.19] | 35482 | 1.04 [0.97–1.11] |
|
20–24 | 128,968 | 29098 | Ref. | Ref. | 40029 | Ref. | Ref. |
25–29 | 96,493 | 24018 | 1.08 [1.00–1.18] |
|
28105 | 0.95 [0.85–1.07] |
|
30–34 | 64,458 | 18924 |
|
|
19737 | 0.99 [0.89–1.10] |
|
35–55 | 40,585 | 12150 |
|
|
13714 | 1.07 [0.94–1.21] | 1.00 [0.95–1.06] |
|
|||||||
Ethnicity/skin colour |
|||||||
White | 91,116 | 48,646 | Ref. | Ref. | 38,882 | Ref. | Ref. |
Mixed | 298,612 | 48,499 |
|
0.44 [0.19–1.01] | 75,497 | 0.67 [0.40–1.15] |
|
Others | 19,328 | 5,790 | 0.66 [0.39–1.12] | 0.83 [0.54–1.27] | 15,303 | 1.48 [0.97–2.25] | 0.99 [0.67–1.47] |
|
|||||||
Literacy |
|||||||
No or primary | 203,185 | 40,699 | 0.71 [0.50–1.02] |
|
48,548 | 0.69 [0.45–1.05] | 0.90 [0.74–1.09] |
Secondary or university | 217,307 | 66,638 | Ref. | Ref. | 84,547 | Ref. | Ref. |
|
|||||||
Marital status |
|||||||
Married + stable part | 364,380 | 89,465 | Ref. | Ref. | 110,699 | Ref. | Ref. |
Single + other | 56,037 | 17,616 | 1.21 [0.99–1.48] | 1.04 [0.93–1.15] | 22,171 | 1.22 [0.97–1.53] | 1.10 [1.00–1.22] |
|
|||||||
Parity |
|||||||
Nullipara | 153,659 | 43,923 |
|
|
56,559 |
|
0.95 [0.80–1.13] |
Multipara | 250,299 | 56,737 | Ref. | Ref. | 75,554 | Ref. | Ref. |
|
|||||||
Any previous condition |
|||||||
Yes | 14,493 | 8,723 |
|
|
8,834 |
|
|
No | 398,396 | 96,144 | Ref. | Ref. | 76,313 | Ref. | Ref. |
|
|||||||
Number of prenatal care visits |
|||||||
0 | 30,445 | 6,820 | 0.78 [0.31–1.99] | 0.99 [0.66–1.50] | 9,831 | 0.95 [0.56–1.60] |
|
1–4 | 115,059 | 22,285 | 0.69 [0.39–1.22] | 0.87 [0.73–1.02] | 35,221 | 0.91 [0.71–1.17] | 0.92 [0.79–1.06] |
>4 | 252,273 | 77,050 | Ref. | Ref. | 87,566 | Ref. | Ref. |
|
|||||||
Smoking, drugs, alcohol, or violence |
|||||||
Yes | 30,559 | 17,576 | 1.76 [0.84–3.72] | 1.01 [0.89–1.14] | 14,097 |
|
|
No | 288,779 | 75,450 | Ref. | Ref. | 48,689 | Ref. | Ref. |
|
|||||||
Previous C-section |
|||||||
Yes | 50,662 | 8,748 |
|
|
19,847 | 1.09 [0.97–1.21] | 0.89 [0.80–1.00] |
No | 303,471 | 83,961 | Ref. | Ref. | 106,305 | Ref. | Ref. |
|
|||||||
Onset of labour |
|||||||
Spontaneous | 340,966 | 87,326 | Ref. | Ref. | 97,812 | Ref. | Ref. |
Induced | 20,578 | 10,826 |
|
1.13 [0.87–1.47] | 9,894 |
|
|
Elective C-section | 50,673 | 9,082 | 0.75 [0.47–1.18] |
|
24,656 |
|
|
|
|||||||
Mode of delivery |
|||||||
C-section | 123,423 | 27,657 | 0.88 [0.66–1.18] | 0.83 [0.68–1.02] | 50,768 |
|
|
Vaginal (any) | 310,183 | 81,745 | Ref. | Ref. | 85,770 | Ref. | Ref. |
|
|||||||
Total | 441,441 | 110,038 | 137,589 |
Missing information for a: 0.4%; b: 6.9%; c: 4.1%; d: 4.2%; e: 7.6%; f: 12.4%; g: 7.6%; h: 31.1%; i: 16.8%; j: 5.4%; k: 1.4% of cases; APR: adjusted prevalence ratio (adjusted for cluster effect and all other predictors); LSMM: less severe maternal morbidity; PLTC: potentially life-threatening condition.
Neonatal outcomes stratified by groups of maternal morbidity are reported in Table
Neonatal outcomes according to the adverse maternal outcomes. CLAP 2009–2012.
Neonatal outcome | MD | MNM | PLTC | LSMM | Any (AMM) | No morbidity |
|
---|---|---|---|---|---|---|---|
Gestational age at birth |
|
||||||
<32 weeks | 68 (6.9) | 1438 (7.0) | 3,125 (3.0) | 5,116 (3.9) | 9,747 (3.8) | 10,620 (2.6) | |
32–36 | 141 (14.2) | 4,114 (20.0) | 8,978 (8.5) | 13,049 (9.9) | 26,282 (10.1) | 28,853 (7.0) | |
≥37 weeks | 783 (78.9) | 14,985 (73.0) | 93,534 (88.5) | 113,707 (86.2) | 223,009 (86.1) | 369,957 (90.4) | |
|
|||||||
Birth weight |
|
||||||
<2500 g | 181 (18.0) | 5,364 (24.8) | 9,135 (8.4) | 14,786 (10.9) | 29,466 (11.1) | 35,103 (8.1) | |
2500–3999 g | 755 (75.0) | 15,451 (71.4) | 93,650 (86.4) | 113,928 (84.3) | 223,784 (84.1) | 378,050 (87.6) | |
≥4000 g | 71 (7.1) | 822 (3.8) | 5,571 (5.1) | 6,495 (4.8) | 12,959 (4.8) | 18,218 (4.2) | |
|
|||||||
Apgar 5th min <7 |
49 (5.0) | 789 (3.7) | 2,105 (2.0) | 2,319 (1.7) | 5,262 (2.0) | 7,363 (1.7) | 0.088 |
Vital status at birth | 0.726 | ||||||
Alive | 986 (95.1) | 21,431 (97.5) | 107,788 (98.0) | 134,556 (97.8) | 264,761 (97.8) | 433,059 (98.1) | |
Fetal death | 42 (4.9) | 554 (2.5) | 2,250 (2.0) | 3,033 (2.2) | 5,879 (2.2) | 8,382 (1.9) | |
|
|||||||
Child condition at maternal discharge |
|||||||
Alive | 736 (95.9) | 16,906 (96.7) | 97,206 (98.4) | 112,886 (98.2) | 227,734 (98.1) | 365,715 (99.0) |
|
Neonatal death | 28 (3.6) | 358 (2.0) | 804 (0.8) | 915 (0.8) | 2,105 (0.9) | 2,066 (0.6) | |
Referred | 4 (0.5) | 227 (1.3) | 824 (0.8) | 1,144 (1.0) | 2,199 (1.0) | 1,792 (0.5) | |
|
|||||||
Neonatal near miss |
128 (13.1) | 2,964 (14.5) | 5,516 (5.2) | 8,440 (6.4) | 17,048 (6.6) | 19,757 (4.9) |
|
The main results of this analysis showed that the overall MMR for the sample was high and over one-third of the total cases presented with any morbidity. Following some midiatic figures, this means that, for each woman who died in this sample, 262 others experienced any degree of morbidity and survived, although they may have had impairments and functioning disabilities lasting for different periods. While a minority had severe morbidity, almost one-fifth of them experienced less severe morbidity, which emphasizes the need for surveillance and timely and adequate diagnosis of complications. Apart from these main results, the study also showed that it is possible to explore a database of big birth registries in the search for variables or reported diagnosis with the specific purpose of identifying a gradient of morbidity. This would enable building a full profile of all pregnancies that could be a proxy for the theoretical continuum of morbidity, from normal pregnancy to maternal death. This gradient seems to work considering that the related adverse neonatal outcomes matched accordingly. The study also demonstrated that the history of previous pathological conditions played an important role in increasing the risk of severe maternal outcomes, and the same occurred for other factors already known to be associated, including extremes of maternal reproductive age, low literacy, absence of a steady partner, nulliparity, low number of prenatal visits, smoking, drug or alcohol use or violence, induction of labour, and elective caesarean section.
The present study has some clear limitations. It developed an operational definition for some degrees of severity on maternal morbidity with information on variables already available in a big database of an international birth registry, which was not built for that specific purpose. It is not a population-based study and does not allow generating estimates for maternal morbidity and mortality stratified for regions and countries. Unfortunately, there is no information available on the distribution of births in the settings providing information for the database. This is the main reason why data were not stratified by countries. Although the adoption of a rigorous process of checking and assuring the quality of data collection and management exists, there is still some degree of incompleteness of some information and this is also a limitation. However, this is a retrospective analysis of a database and therefore we could no longer ask for additional corrections or completeness. The amount of missing information is anyway provided for each variable assessed.
On the other hand, our study has some important strengths. To the best of our knowledge, this is the first study trying to capture a full profile of morbidity occurring during pregnancy using a routine birth registry. According to the new initiatives from WHO, this is a recommended procedure for building a full scenario of the burden of maternal morbidity for the women’s lives. Using classifications for a gradient of maternal morbidity as the currently employed, additional secondary analysis will also be possible in databases from birth registries, focusing for instance on maternal morbidity linked with specific causes like hypertension or postpartum haemorrhage, twin pregnancy, obesity and overweight, ethnic groups, and other hot topics that could be easily assessed in the database, thus generating strong evidence supported by huge numbers.
The overall MMR found in this analysis was high, almost 150 per 100,000 live births. Despite the fact that recently an increased trend in the estimated MMR has been shown for the US in the last decade [
The analysis of previous morbid conditions (diabetes, hypertension, preeclampsia, eclampsia, cardiac disease, renal disease, and any other severe condition) shows a clear statement of the risks involving pregnancies with baseline complications. They presented a clear increase in the occurrence of any degree of morbidity. This highlights the importance of understanding and identifying this gradient or spectrum of complications as a way to provide appropriate management for any morbidity that could potentially develop furthermore severe complications. The Global Burden of Disease Study showed that around 70% of Latin American maternal deaths occurred in intrapartum and postpartum periods. The surveillance of maternal complications during hospitalization for childbirth might play a key role on prevention of maternal mortality [
The sociodemographic and pregnancy characteristics confirmed known risk factors for poor maternal outcomes, such as extreme age groups, nonwhite ethnicity, no stable partner, no prenatal care, smoking, drug and alcohol use, previous morbid conditions, elective C-section, or induction of labour. In addition, approximately the same results were found among cases with severe or less severe morbidity. Perhaps the best example is that of maternal age, whose extreme groups, adolescents and older women, are showed to be at higher risk of developing any degree of severity for maternal morbidity. Two big multicentre studies using the new WHO concepts and criteria for severe maternal outcomes showed similar results [
Over one-third of women displayed any maternal morbidity. This figure is quite higher than the mean 9.2% of adverse outcome index (AOI) described as a way of assessing the quality of obstetrical care [
Neonatal outcomes matched maternal condition when considering its degree of severity. MD and MNM were clearly associated with worse perinatal outcomes and increased preterm birth rates (especially late preterm). They are due to either the severity of maternal condition itself that is able to impair the intrauterine fetal health condition and as a result of an early interruption of pregnancy because of the maternal condition, resulting in preterm birth and its negative consequences. The association of maternal complications with preterm delivery is an emerging topic in maternal and perinatal health. Provider-initiated preterm birth, for instance, is closely associated with maternal complications during pregnancy [
Considering the high proportion of women experiencing maternal morbidity with varying degrees of severity and the need to ensure clinical responses to these differing morbidity profiles, it is important to develop and sustain active surveillance systems that would theoretically follow women in real time while they are experiencing the process of pregnancy and childbirth and possibly some related complications. Following the WHO recommendation, the SIP birth registry recently adopted the same concepts and criteria for maternal near miss [
Any maternal morbidity
Any maternal morbidity ratio
Adverse outcome index
Latin American Center for Perinatology, Women and Reproductive Health
Intensive care unit
Live births
Less severe maternal morbidity
Less severe maternal morbidity ratio
Maternal death
Maternal mortality ratio
Maternal near miss
Maternal near miss ratio
Pan American Health Organization
Potentially life-threatening condition
Potentially life-threatening condition ratio
Perinatal Information System
Severe maternal outcome
Severe maternal outcome ratio
World Health Organization.
The protocol for the current analyses was previously technically and ethically evaluated and approved by CLAP/WR. The data were retrospectively collected from a routine birth registry without any identification of women and therefore the need for consent was waived.
The sponsor PAHO/CLAP was responsible for the current manuscript publication. The institution itself did not play any role in planning, performing, analysing, or interpreting results; however, some of their staff members actively participated in all phases of the study as researchers.
The authors declare they have no conflicts of interest regarding the publication of this article.
Suzanne J. Serruya, Bremen de Mucio, and Jose G. Cecatti had the idea for the study and first outlined its main objectives. Suzanne J. Serruya, Bremen de Mucio, Gerardo Martinez, Luis Mainero, and Jose G. Cecatti developed a detailed plan of analysis. Maria H. Sousa coded the data and carried out the analysis. Jose G. Cecatti, Renato T. Souza, Jussara Mayrink, and Maria L. Costa saw the first results, discussed their implications, and drafted the first version of the manuscript. Andres de Francisco and Lale Say detailed and reviewed the results and tables. All authors carefully reviewed the manuscript, made suggestions, and approved the final version of it.
Currently, the analysis was possible under the PAHO/CLAP support (Service Agreement CL/CNT/1400059.001).