Keywords
preterm delivery - preeclampsia - vaginosis - social stressors - access to care
Preterm birth is defined by the World Health Organization (WHO) as delivery before
37 weeks of pregnancy. According to the 2012 “Born Too Soon” report, 15 million children were born prematurely in the world in the 2010.[1]
[2]
[3] Thus, in most industrialized countries prematurity has continuously increased in
the past three decades[4]
[5] with between 5 and 11% of children being born preterm.[6] There are however, important differences between countries with 12 and 13% of preterm
births in the U.S.A. and 5 to 7% in Europe.[7]
In France and French territories, it was shown that 7.4% of children were born before
37 weeks of pregnancy with marked differences between mainland France (6.6%) and its
overseas territories (13.9%).[8]
French Guiana is the largest French department (one-sixth of France) which is 7,000 km
away from France. It is located on the South American continent between Suriname and
Brazil.[9]
French Guiana is mostly covered with primary forest (90%) and most of the population
(279,933) lives on the coastal area. In the past 50 years, the population has grown
7-fold due to one of the highest birth rates (26.2‰) in Latin America and massive
immigration of persons in search of better economic prospects in this ultraperipheral
region of Europe.[9]
[10]
Given the high incidence of preterm births in French Guiana since the 1990's, a registry
has been set up to record birth outcomes. It allows to register perinatal health indicators
and to objectively monitor preterm births on the territory and to study trends and
local risk factors.[11]
The registry has shown that in the past 20 years, the incidence of preterm births
has remained stable at 13.5%. Preterm birth is considered as the first cause of neonatal
deaths and belowfive of mortality worldwide. Life expectancy at birth is lower in
the overseas French territories than in mainland France largely because of the higher
child mortality. In France, both mainland and overseas, the frequency of type IV,
III, II, and I intraventricular hemorrhages of preterm infants, a major cause of lifelong
disability, was 3.8% (95% CI [confidence interval], 3.2–4.5%), 3.3% (95% CI, 2.7–3.9),
12.1% (95% CI, 11.0–13.3) and 17.0% (95% CI, 15.7–18.4), respectively.[12]
[13] Thus, preterm births remain a major health challenge and health indicator in French
Guiana.[14]
[15]
A first retrospective study in French Guiana study was conducted including all preterm
births and the factors associated with preterm delivery.[16] However, the registry does not collect some key variables that may be associated
with preterm delivery. Early preterm births represent 3.7% of births in French Guiana
relative to 1.3% in mainland France and thus still represent a major public health
problem.[12]
[17]
[18] The objective of the present study was thus to study factors associated with early
preterm births in French Guiana.
Material and Methods
Study Design
The study was a monocentric case control study. Inclusions were prospective.
Study Site and Conduct
The study was conducted between February 2016 and February 2017 in the only type 3
maternity in French Guiana. In utero fetal deaths and multiple pregnancies were not
included.
Comparison Groups
Cases were defined as giving birth prematurely between 22 and 32 weeks of pregnancy.
Controls were defined as women delivering on term. For each case, three controls were
matched on age.
Induced preterm delivery was defined as the induction of delivery or a cesarean section
between 22 and 32 weeks of pregnancy. This usually occurred in case of hypertension,
preeclampsia, intrauterine growth retardation, or placenta praevia.
Spontaneous preterm delivery was a spontaneous delivery between 22 and 32 weeks of
pregnancy or a cesarean section in a context of premature membrane rupture with or
without chorioamniotitis.
Exclusion Criteria
In utero deaths, medical pregnancy interruptions, and multiple pregnancies (major
known causes of preterm delivery) were excluded from the study.
Sample Size
The sample size was defined by the number of medical records for early preterm deliveries
available over a 1-year period and the matched records (three controls per case).
Study Variables
Sociodemographic variables, medical and obstetrical history, the notion of complications
during the current pregnancy, and the results of the last available vaginal swab before
delivery were recorded in the second or in the third trimester.
There is variability in pregnancy follow-up with 40% of women delivering in French
Guiana being foreign and only having health insurance after at least 3 months of presence
on the territory.
Statistical Analysis
A descriptive analysis was first performed. Socioeconomic data, medical history, and
bacteriological results were compared by delivery term (preterm vs. normal term).
Statistical comparisons used the Chi-squared test or when cell numbers were <5, Fisher's
exact test.
Quantitative variables were compared using Student's t-test or when the distribution was non-Gaussian, nonparametric tests.
Multivariate analysis using conditional logistic regression was performed to identify
the main factors associated with early preterm birth and to adjust for potential confounding.
First, thematic models were computed: one for obstetrical history, one for current
pregnancy events, and one for the bacteriological findings. The significant variables
identified were then included in global models.
Statistical significance was set at p< 0.05. Data was analyzed with Stata 12 (College Station, TX).
Ethical and Regulatory Aspects
The retrospective analysis of anonymized monocentric data from medical records is
authorized by French authorities. The project was approved by the local Ethical committee
(n°3–2016-V1). The database was declared to the regulatory authorities, the Commission
NationaleInformatiqueetLibertés CNIL (1914209v 0 le 18/12/2015).
Results
Overall there were 94 cases and 282 matched controls.
Sociodemographic Data
The average age of the study population was 29 years old (range: 13–45 years), 7.5%
(28/376)were aged less than 18 years old, and 6.7%(25/376) were 40 years old or more.
Overall, 60.7%(228/376) of the study population was not born in French Guiana and
77.6%(292/376) were without a job. Finally, 15.5% (58/375) of the women had no health
insurance.
Medical Data
During the study period, extreme or very preterm delivery was spontaneous in 47.9%
(45/94) of the cases and induced in 52.1% (49/94) of the women. Over a quarter of
the women were multiparous (27.4% [103/376]). Among pregnant women with a history
of preterm delivery, over half (59.6%[28/47]) delivered before 33 weeks of pregnancy.
In the study sample, 61.7% (69/112) of cesarean sections were performed between 22
and 32 weeks of pregnancy and only 26.6% (25/94) of grand preterm newborns were delivered
vaginally and none by instrumental vaginal delivery.
Among women with hypertension or preeclampsia (14.6% [55/376]), most pregnancies resulted
in grand preterm delivery 83.6% (46/55). Smoking was relatively rare 24/376 and was
not associated with preterm delivery (data not shown). In 15.6% of women, there was
no vaginal swab data (51/376). In 23.1% of cases (75/325), the vaginal swab culture
result showed polymicrobial culture which would have required a new sample.
[Table 1] shows that in the bivariate analysis a history of preterm delivery, a scarred uterus,
preeclampsia, and vaginal infections were significantly associated with early preterm
birth.
Table 1
Medical Obstetrical factors, prevalence of various obstetrical pathologies in the
population and Prevalence of vaginal infection of women having delivered in French
Guiana 2016–2017: crude odds ratios
|
Obstetrical history
|
Number (%) of extreme/very preterm delivery/term delivery relative to term delivery
|
Crude OR (95% CI)
|
Current obstetrical problems
|
Number (%) of extreme/very preterm delivery/term delivery relative to term delivery
|
Crude OR (95% CI)
|
Vaginal infections before delivery
|
Number (%) of extreme/very preterm delivery/term delivery relative to term delivery
|
Crude OR (95% CI)
|
|
History of spontaneous abortion
|
|
|
Gravidic hypertension
|
|
|
Polymicrobial culture
|
|
|
|
No
|
68/281 (24.2%)
|
1
|
No
|
80/355 (22.5%)
|
1
|
No
|
61/281 (21.7%)
|
1
|
|
Yes
|
26/95 (27.4%)
|
1.2 (0.7–2.2)
|
Yes
|
14/21 (66.7%)
|
8.2 (2.8–24.2)
|
Yes
|
14/44 (31.8%)
|
1.4 (0.7–2.9)
|
|
History of abortion
|
|
|
Preeclampsia
|
|
|
Gardnerella vaginalis
|
|
|
|
No
|
71/298 (23.8%)
|
1
|
No
|
62/342 (18.1%)
|
1
|
No
|
53/253 (20.9%)
|
1
|
|
Yes
|
23/78 (29.5%)
|
1.4 (0.8–2.5)
|
Yes
|
32/34 (94.1%)
|
75.7 (16.8–340.3)
|
Yes
|
22/72 (30.6%)
|
1.8 (1–3.3)
|
|
History of medical pregnancy interruption
|
|
|
Gestational diabetes
|
|
|
Ureaplasma urealyticum
|
|
|
|
No
|
92/372 (24.7%)
|
1
|
No
|
86/353 (24.4%)
|
1
|
No
|
75/319 (23.5%)
|
1
|
|
Yes
|
2/4 (50%)
|
2.6 (0.4–18.7)
|
Yes
|
8/23 (34.8%)
|
1.6 (0.6–4.2)
|
Yes
|
0/6 (0%)
|
1
|
|
History of ectopic pregnancy
|
|
|
Premature membrane rupture
|
|
|
Streptococcus B
|
|
|
|
No
|
93/375 (24.8%)
|
1
|
No
|
69/351 (19.7%)
|
1
|
No
|
57/276 (20.7%)
|
1
|
|
Yes
|
1/1 (100%)
|
1
|
Yes
|
25/25 (100%)
|
1
|
Yes
|
18/49 (36.7%)
|
2.6 (1.3–5.3)
|
|
History of preterm delivery
|
|
|
Placentapraevia
|
|
|
Candida albicans
|
|
|
|
No
|
66/329 (20.1%)
|
1
|
No
|
91/372 (24.5%)
|
1
|
No
|
66/263 (25.1%)
|
1
|
|
Yes
|
28/47 (59.6%)
|
5.3 (2.7–10.3)
|
Yes
|
3/4 (75%)
|
6 (0.6–60.1)
|
Yes
|
9/62 (14.5%)
|
0.5 (0.2–1)
|
|
Scarred uterus
|
|
|
Intra uterine growth retardation
|
|
|
Trichomonas vaginalis
|
|
|
|
No
|
71/324 (21.9%)
|
1
|
No
|
77/354 (21.8%)
|
1
|
No
|
74/322 (23%)
|
1
|
|
Yes
|
23/52 (44.2%)
|
3.3 (1.7–6.3)
|
Yes
|
17 /22 (77.3%)
|
14.2 (4.9–41.9)
|
Yes
|
1/3 (33.3%)
|
3.9 (0.3–44.9)
|
|
|
|
Hydramnios
|
|
|
Candida (nonalbicans)
|
|
|
|
|
|
No
|
94/282 (33.3%)
|
1
|
No
|
74/324 (22.8%)
|
1
|
|
|
|
Yes
|
0
|
1
|
Yes
|
1/1 (100%)
|
1
|
|
|
|
|
|
|
Chlamydiae
|
|
|
|
|
|
|
|
|
No
|
75/323 (23.2%)
|
1
|
|
|
|
|
|
|
Yes
|
0/2 (0%)
|
1
|
|
|
|
|
|
|
Candida dubliniensis
|
|
|
|
|
|
|
|
|
No
|
75/324 (23.1%)
|
1
|
|
|
|
|
|
|
Yes
|
0/1 (0%)
|
1
|
|
|
|
|
|
|
Gonococcus
|
|
|
|
|
|
|
|
|
No
|
75/324 (23.1%)
|
1
|
|
|
|
|
|
|
Yes
|
0/1 (0%)
|
1
|
|
|
|
|
|
|
Mycoplasma
|
|
|
|
|
|
|
|
|
No
|
75/324 (23.1%)
|
1
|
|
|
|
|
|
|
Yes
|
0/1 (0%)
|
1
|
Abbreviations: CI, confidence intervals; OR, odd ratio.
[Table 2] presents different thematic multivariate models showing a history of preterm delivery,
gravidic hypertension, preeclampsia, placenta praevia, growth retardation, and vaginal
bacteriosis, and group B Streptococcus infections were significantly associated with early preterm birth.
Table 2
Predictors of extreme/very preterm delivery French Guiana 2016–2017: adjusted odds
ratios
|
Obstetrical history
|
Adjusted OR (95% CI)
|
p-Value
|
Current obstetrical problems
|
Adjusted OR (95% CI)
|
p-Value
|
Vaginal infections before delivery
|
Adjusted OR (95% CI)
|
p-Value
|
|
History of spontaneous abortion
|
|
|
Gravidic hypertension
|
|
|
Polymicrobial culture
|
|
|
|
No
|
1
|
|
No
|
1
|
|
No
|
1
|
|
|
Yes
|
1.1 (0.6–2.1)
|
0.7
|
Yes
|
8.4 (1.5–45.5)
|
0.01[a]
|
Yes
|
2 (0.9–4.7)
|
0.09
|
|
History of abortion
|
|
|
Preeclampsia
|
|
|
Gardnerella vaginalis
|
|
|
|
No
|
1
|
|
No
|
1
|
|
No
|
1
|
|
|
Yes
|
1.7 (0.9–3.2)
|
0.1
|
Yes
|
121.7 (21.6–684.7)
|
0.0001[a]
|
Yes
|
2.2 (1.1–4.4)
|
0.02[a]
|
|
History of medical pregnancy interruption
|
|
|
Gestational diabetes
|
|
|
Ureaplasma urealyticum
|
|
|
|
No
|
1
|
|
No
|
1
|
|
No
|
1
|
|
|
Yes
|
2.8 (0.4–21.7)
|
0.3
|
Yes
|
0.9 (0.1–6.8)
|
0.9
|
Yes
|
1
|
|
|
History of ectopic pregnancy
|
|
|
Premature membrane rupture
|
|
|
Streptococcus B
|
|
|
|
No
|
1
|
|
No
|
1
|
|
No
|
1
|
|
|
Yes
|
1
|
|
Yes
|
1
|
|
Yes
|
3.5 (1.6–7.5)
|
0.001[a]
|
|
History of preterm delivery
|
|
|
Placenta praevia
|
|
|
Candida albicans
|
|
|
|
No
|
1
|
|
No
|
1
|
|
No
|
1
|
|
|
Yes
|
4.9 (2.4–10)
|
0.0001[a]
|
Yes
|
18.3 (1.6–209.6)
|
0.02[a]
|
Yes
|
0.6 (0.2–1.3)
|
0.2
|
|
Scarred uterus
|
|
|
Intra uterine growth retardation
|
|
|
Trichomonas vaginalis
|
|
|
|
No
|
1
|
|
No
|
1
|
|
No
|
1
|
|
|
Yes
|
1.8 (0.8–3.7)
|
0.1
|
Yes
|
17.3 (3.1–98.5)
|
0.001[a]
|
Yes
|
7.3 (0.6–86.9)
|
0.1
|
|
|
|
Hydramnios
|
|
|
Candida (nonalbicans)
|
|
|
|
|
|
No
|
1
|
|
No
|
1
|
|
|
|
|
Yes
|
1
|
|
Yes
|
1
|
|
|
|
|
|
|
|
Chlamydiae
|
|
|
|
|
|
|
|
|
No
|
1
|
|
|
|
|
|
|
|
Yes
|
1
|
|
|
|
|
|
|
|
Candida dubliniensis
|
|
|
|
|
|
|
|
|
No
|
1
|
|
|
|
|
|
|
|
Yes
|
1
|
|
|
|
|
|
|
|
Gonococcus
|
|
|
|
|
|
|
|
|
No
|
1
|
|
|
|
|
|
|
|
Yes
|
1
|
|
|
|
|
|
|
|
Mycoplasma
|
|
|
|
|
|
|
|
|
No
|
1
|
|
|
|
|
|
|
|
Yes
|
1
|
|
Abbreviations: CI, confidence intervals; OR, odd ratio.
a
p < 0.05.
[Table 3] shows that women born on the Latin American continent, including French Guiana,
where more likely to deliver early preterm babies when compared with women from mainland
France. [Table 3] also shows that women without any health coverage were at higher risk of spontaneous
early preterm delivery but not induced early preterm delivery.
Table 3
Mothers' origin and health coverage in relation to extreme/very preterm delivery in
French Guiana 2016–2017
|
Number
(%) of extreme/very preterm delivery per category relative to all deliveries
|
Crude OR (95% CI)
|
Adjusted OR (95% CI)
For extreme/very preterm birth
|
p-Value
|
AOR (95% CI) for spontaneous extreme/very preterm birth
|
p-Value
|
AOR (95% CI) for induced extreme/very preterm birth
|
p-Value
|
|
Place of birth
|
|
|
|
|
|
|
|
|
|
Africa
|
0/3 (0%)
|
1
|
|
|
1
|
|
1
|
|
|
Antilles
|
0/4 (0%)
|
1
|
|
|
1
|
|
1
|
|
|
Brazil
|
5/30 (14.3%)
|
1.2 (0.2–5.7)
|
0.9 (0.1–6.9)
|
0.9
|
0.5 (0.03–9.3)
|
0.7
|
2.3 (0.05–97.6)
|
0.7
|
|
China
|
0/3 (0%)
|
1
|
|
|
1
|
|
1
|
|
|
Guyana
|
2/6 (33.3%)
|
4.8 (0.5–41.5)
|
4.8 (0.4–59.2)
|
0.2
|
1
|
|
21.4 (0.5–841.2)
|
0.1
|
|
French Guiana
|
48/100 (32.4%)
|
4.8 (1.3–18.4)
|
6 (1.3–27.5)
|
0.02[a]
|
6.9 (1–46.9)
|
0.05
|
14.2 (0.6–354.8)
|
0.09
|
|
Haïti
|
24/87 (21.6%)
|
2 (0.5–7.6)
|
2.5 (0.5–13.5)
|
0.3
|
1.1 (0.1–10.5)
|
0.08
|
7.3 (0.3–176.5)
|
0.2
|
|
Mainland France
|
3/21 (12.5%)
|
1
|
|
|
1
|
|
1
|
|
|
Dominican Republic
|
3/10 (23.1%)
|
2.4 (0.4–15.3)
|
4.3 (0.5–35.5)
|
0.2
|
10.5 (0.6–174.1)
|
0.1
|
4 (0.1–169.6)
|
0.5
|
|
Reunion Island
|
0/1 (0%)
|
1
|
|
|
1
|
|
1
|
|
|
Suriname
|
9/17 (34.6%)
|
5.3 (1.2–24.1)
|
1 (0.1–9.8)
|
0.9
|
0.8 (0.02–21.4)
|
0.9
|
2.6 (0.1–97.6)
|
0.6
|
|
Other
|
0/2 (0%)
|
1
|
|
|
1
|
|
1
|
|
|
Town of residence
|
|
|
|
|
|
|
|
|
|
Cayenne
|
29/133 (17.9%)
|
1
|
|
|
1
|
|
1
|
|
|
Kourou
|
6/1 (85.7%)
|
30.8 (3.2–296.4)
|
73.2 (6.4–84.2)
|
0.001[a]
|
1
|
|
54.2 (3.1–957.3)
|
0.006
|
|
Matoury
|
16/57 (21.9%)
|
1.2 (0.6–2.6)
|
1.2 (0.6–2.7)
|
0.6
|
2.3 (0.7–7.3)
|
0.2
|
0.9 (0.4–2.6)
|
0.9
|
|
Rémire-Montjoly
|
9/27 (25%)
|
1.4 (0.6–3.4)
|
1.3 (0.5–3.7)
|
0.6
|
5.8 (1.4–23.1)
|
0.01
|
0.2 (0.03–2)
|
0.2
|
|
Saint-Laurent du Maroni
|
15/1 (93.8%)
|
69 (8.6–552.3)
|
107.6 (9.6–1205.1)
|
0.0001[a]
|
82.3 (6.5–1034.3)
|
0.001
|
1
|
|
|
Profession
|
|
|
|
|
|
|
|
|
|
No
|
72/220 (24.7%)
|
1
|
|
|
1
|
|
1
|
|
|
Yes
|
22/62 (26.2%)
|
0,7 (0.4–1.1)
|
1.1 (0.6–2)
|
0.7
|
2.6 (0.8–8.6)
|
0.1
|
0.9 (0.3–2.9)
|
0.8
|
|
Health coverage
|
|
|
|
|
|
|
|
|
|
Yes
|
20/38 (34.5%)
|
1
|
1
|
|
1
|
|
1
|
|
|
No
|
74/243 (23.3%)
|
1.9 (1–3.5)
|
3.3 (1.4–7.9)
|
0.006[a]
|
9.1 (2.2–38.3)
|
0.002
|
2.1 (0.6–6.7)
|
0.2
|
Abbreviations: CI, confidence intervals; OR, odd ratio.
a
p < 0.05.
[Table 4] distinguishes between spontaneous and induced early preterm delivery and shows that
a history of preterm birth was associated with both spontaneous and induced preterm
delivery.
Table 4
Predictors of spontaneous and induced preterm birth delivery in French Guiana 2016–2017:
adjusted odds ratios (AOR)
|
Obstetrical history
|
Spontaneous extreme/very preterm birth
|
Induced extreme/very preterm birth
|
Current pregnancy problems
|
Spontaneous extreme/very preterm birth
|
Induced extreme/very preterm birth
|
Vaginal infections
|
Spontaneous extreme/very preterm birth
|
Induced extreme/very preterm birth
|
|
Adjusted OR (95% CI)
|
p-Value
|
Adjusted OR (95% CI)
|
p-Value
|
|
Adjusted OR (95% CI)
|
p-Value
|
Adjusted OR (95% CI)
|
p-Value
|
|
Adjusted OR (95% CI)
|
p-Value
|
Adjusted OR (95% CI)
|
p-Value
|
|
History of fetal loss
|
1.5 (0.7–3.5)
|
0.3
|
0.9 (0.4–2)
|
0.8
|
Gravidic hypertension
|
4.3 (0.4–50.3)
|
0.2
|
22.9 (1.8–299)
|
0.02
|
Polymicrobial culture
|
1.6 (0.6–4.5)
|
0.3
|
3.1 (0.9–10.7)
|
0.05
|
|
History of abortion
|
1.5 (0.6–3.9)
|
0.9
|
2 (0.9–4.4)
|
0.08
|
Preeclampsia
|
1
|
|
746.3 (127.3–4373.6)
|
0.0001
|
Gardnerella vaginalis
|
1.3 (0.5–3.1)
|
0.6
|
4 (1.5–11.1)
|
0.008
|
|
History of medical pregnancy interruption
|
4 (0.3–50.6)
|
0.3
|
2.4 (0.2–29.3)
|
0.5
|
Gestational diabetes
|
1.1 (0.1–10.5)
|
0.9
|
1.2 (0.2–8.5)
|
0.9
|
Ureaplasm aurealyticum
|
1
|
|
1
|
|
|
History of ectopic pregnancy
|
1
|
|
1
|
|
Premature membrane rupture
|
1
|
|
1
|
1
|
Streptococcus B
|
2.3 (0.9–5.9)
|
0.07
|
6.3 (2–19.7)
|
0.001
|
|
History of preterm birth
|
5.7 (2.1–15.8)
|
0.001
|
5.2 (2.3–12.3)
|
0.0001
|
Placenta praevia
|
1
|
|
236.3 (19.7–2840.4)
|
0.0001
|
Candida albicans
|
0.5 (0.2–1.5)
|
0.2
|
0.6 (0.2–1.9)
|
0.4
|
|
Scarred Uterus
|
2.6 (0.9–7)
|
0.06
|
1.4 (0.5–3.5)
|
0.6
|
Intra uterine growth retardation
|
1
|
|
43.9 (8.7–222.3)
|
0.0001
|
Trichomonas vaginalis
|
1
|
|
24.6 (1.4–340.3)
|
0.03
|
|
|
|
|
|
|
|
|
|
|
Candida (nonalbicans)
|
1
|
|
1
|
|
|
|
|
|
|
|
|
|
|
|
Chlamydiae
|
1
|
|
1
|
|
|
|
|
|
|
|
|
|
|
|
Candida dubliniensis
|
1
|
|
1
|
|
|
|
|
|
|
|
|
|
|
|
Gonococcus
|
1
|
|
1
|
|
|
|
|
|
|
|
|
|
|
|
Mycoplasma
|
1
|
|
1
|
|
Abbreviations: CI, confidence intervals; OR, odd ratio.
[Table 4] also shows that obstetrical complications such as gravidic hypertension, placenta
praevia, intrauterine growth retardation, and mostly preeclampsia were linked to induced
preterm delivery but not spontaneous delivery.
[Table 4] shows that G.vaginalis and group B Streptococcus infections were significantly associated with induced early preterm delivery but
not spontaneous early preterm delivery. Finally, [Table 4] shows that spontaneous preterm birth a history of preterm birth was associated with
spontaneous early preterm birth.
Discussion
Preterm birth is heterogeneous in terms of delivery and in mechanism (spontaneous,
following a preterm premature rupture of membranes, and it may be iatrogenic, induced
for medical reasons).[19] In the present study, we mostly identified factors associated with induced early
preterm labor. As others before, the present study showed that variables, such as
a history of preterm delivery, social vulnerability, hypertension, and preeclampsia
were associated with early preterm delivery.[6]
[14]
[20] Induced preterm births represented over half of early preterm births. As elsewhere,
this has been an increasing trend with between 40 and 55% of preterm births being
induced.[21]
[22]
[23]
[24]
[25] The proportion of induced early preterm births seemed higher than in France (52.5
vs. 42.5%, p< 0.001) but what was remarkable in the present study was the importance of preeclampsia
as a cause of early preterm delivery. In mainland France,15.3% of induced preterm
births were the result of preeclampsia; in the U.S.A., preeclampsia led to preterm
medically delivery in 30 to 43% of women; in Holland, the proportion was 41.1%; whereas
in French Guiana the proportion was the highest at 65.3%(32/49) of induced early preterm
deliveries.[26]
[27]
[28] Thus overall, 33.6% of all early preterm deliveries and 65.1% of induced early preterm
deliveries could be attributed to preeclampsia. In French Guiana, 4.1% of pregnancies
are complicated by preeclampsia, a figure that is double than observed in mainland
France (2%), or the overseas French territory of Reunion Island (2.3%).[29]
[30] In the Reunion island study, among women with preeclampsia, 59.8% delivered before
37 weeks of pregnancy, 28.6% between 34 and 36 weeks, and 31.2% before 34 weeks of
pregnancy.[31] Given the frequency of preeclampsia in French Guiana and its association with induced
preterm delivery, preventive interventions in at risk women (previous history of preeclampsia,
high blood pressure, multiple gestation, first pregnancy, obesity, age <18, or >40,
sickle cell disease…) could have some impact; however, this would require a close/serious
follow-up of pregnancy which is not always the case in socially disadvantaged women.
Studies have shown a relation between psychosocial stress[32]
[33] and preterm delivery as delivery before 37 weeks of pregnancy. We have observed
in French Guiana that social vulnerability, the absence of health coverage was associated
with delivery before 37 weeks of pregnancy. Here we show that the relation is also
true for early preterm birth notably for spontaneous preterm delivery but not induced
preterm delivery.
The last available vaginal swab results before delivery were analyzed because 21.9%
of women initiate pregnancy follow-up late, and 7.6% have no follow-up at all.[29]
[34] The most common genital infections observed here were G.vaginalis, Streptococcus B and Candida albicans. Several studies have suggested a link between preterm birth and genital infections[35]
[36]
[37] some studies showing that up to 30% of preterm births are associated with genital
infections.[37]
[38]
In our study, the prevalence of G.vaginalis was 23.1% (75/325), and the infection was associated with a 2-fold increase of the
odds ratio for extreme/severe preeclampsia which is similar to what has been observed
in a previous meta-analysis showing odds ratios ranging between 1.5 and 2.[39]
[40] The hypothesized pathophysiology of the association is that the infections leads
to the fragilization of the membranes which are at increased risk of rupture. It was
also suggested that cytokines in infected women may lead to uterine contractions.[36]
[41]
[42]
[43] Nadeau et al estimated that genital infections may be associated with 25 to 40%
of preterm deliveries. The prevalence of vaginal bacteriosis varies widely between
populations.[34]
[39]
[44]
[45]
[46] In our population, it was significantly higher than in mainland France 23.1 versus
15.2%;p = 0.006.
However, the view that the link between vaginal bacteriosis and preterm birth is causal
is not unanimous. Indeed vaginal bacteriosis is associated with several factors that
may confound the link between vaginosis and preterm delivery. A study in Canada suggested
that vaginal douching was associated with vaginosis and significant risk of preterm
birth. It is noteworthy that in French Guiana, vaginal douching/steam baths are quasi
systematic in some populations.[47] As shown elsewhere, after adjustments for social variables and smoking, group B
Streptococcuswas significantly associated with induced extreme/very preterm delivery.[35] It is noteworthy that vaginal infections leading to premature membrane rupture and
delivery were classified as spontaneous preterm delivery. However, when differentiating
between spontaneous and induced preterm delivery, the numbers of infections seemed
balanced between types of prematurity but the statistical association was observed
for induced preterm delivery only. Given the low study power, this should be verified
in larger cohorts.
The present study had several limitations. First the sample size was insufficient
to construct large multivariate models. The results of the vaginal swab were the last
ones, the presumption that they apply to the whole pregnancy may not be true. Studies
evaluating the role of treatment of bacterial vaginosis have suggested that this intervention
may most effective when applied early in the pregnancy.[48]
[49]
[50] We did not have any information on cultural practices, such as steam baths using
astringent herbs,[51] which may be associated with changes in the vaginal flora that may promote bacterial
vaginosis or other colonizations. In addition, some authors have suggested that there
may be an association with preterm birth when vaginal douching is frequent.[52] This information would be potentially important to adapt primary prevention messages
to the cultural practices on the territory.
Conclusion
In the Amazonian region, preterm delivery is often studied in relation to the specific
effects of malaria or mercury exposure.[53]
[54] Preterm delivery remains a major problem in French Guiana. We have previously observed
the importance of social factors when considering deliveries before 37 weeks of pregnancy.
The present study provides an overview of factors associated with early preterm birth
in French Guiana. The first observation was that social factors were significantly
associated with spontaneous early preterm delivery. The present results also show
that preeclampsia was a major contributor to induced early preterm delivery. Vaginal
infections by G.vaginalis or group B Streptococcus were also associated with induced early preterm delivery. These factors may be amenable
to prevention.[32]
[33] Overall, the reduction of the incidence of early preterm birth will require specific
efforts to improve timely follow-up of most at risk women.
However, this requires a close and rigorous pregnancy follow-up which is still a challenge
for the most economically vulnerable women in French Guiana who often combine poor
follow-up and psychosocial stressors.[32]
[33] Finally, given the frequency of preeclampsia in French Guiana, it is unfortunately
difficult to identify women who are at risk to give preventive treatment.
Summary
Preterm delivery remains a major public health problem in French Guiana. While trying
to reduce its occurrence, a better and contextualized understanding of its features
and its risk factors is important. A prospective study compared women with early preterm
deliveries with women of the same age group with term delivery. Preterm deliveries
were classified as spontaneous or as induced.
This distinction showed that social vulnerability was associated with spontaneous
preterm delivery and that induced preterm delivery was linked to maternal pathologies
of pregnancy, notably preeclampsia which was the major contributor to induced preterm
delivery, significantly more so than in other published series.
Vaginal infections were also associated with induced preterm labor suggesting that
early diagnosis and treatment could reduce induced early preterm delivery. The present
results suggest that reducing psychosocial stressors could lead to potential improvements
in terms of spontaneous preterm births. Vigilance on possible vaginal infections and
subsequent treatment may also yield some benefits. Finally, preeclampsia remains a
challenge, it is hard to predict and to prevent, and should be the focus of more research.
