Abstract
Background
Tuberculosis (TB) disease has been associated with pregnancy complications. However, the potential impact of TB infection (TBI) on pregnancy outcome is unknown. To investigate this, we conducted a register-based study in immigrant women screened with QuantiFERON assays for TBI in antenatal care in Sweden.
Methods
Women with history of immigration from TB-endemic countries were eligible for inclusion if national identification numbers and available QuantiFERON results obtained during pregnancy from 2014 to 2018 were available. QuantiFERON results were linked to data on maternal characteristics and pregnancy outcomes from the national Pregnancy and Patient Registers. TBI was defined as nil-corrected QuantiFERON result ≥0.35 IU/mL, in the absence of TB disease. Pregnancies in women with TB disease or human immunodeficiency virus were excluded, as were multiplex pregnancies, pregnancies resulting in miscarriage, and pregnancies occurring >10 years after immigration. Odds of defined adverse pregnancy outcomes were compared by maternal TBI status using mixed effects logistic regression with adjustment for maternal age and region of origin.
Results
In total, 7408 women with 12 443 pregnancies were included. In multivariable analysis, stillbirth (adjusted odds ratio [AOR], 1.90; 95% confidence interval [CI], 1.13–3.21; P = .016), severe preeclampsia (AOR, 1.62; 95% CI, 1.03–2.56; P = .036), low birthweight (<2500 g; AOR, 1.38; 95% CI, 1.01–1.88; P = .041), and emergency cesarean section (AOR, 1.28; 95% CI, 1.02–1.63; P = .033) were significantly associated with TBI.
Conclusions
Among immigrant women seeking antenatal care in Sweden, TBI was independently associated with adverse pregnancy outcomes. Further studies are needed to corroborate these findings and to explore mechanisms involved.
Keywords: latent tuberculosis infection, tuberculosis, stillbirth, preeclampsia, pregnancy
Among immigrant women seeking Swedish antenatal care, tuberculosis infection was independently associated with stillbirth, severe preeclampsia, low birthweight, and emergency cesarean section, suggesting that tuberculosis infection may be involved in the pathogenesis of pregnancy-related disorders.
The incidence of severe pregnancy complications in Sweden is among the lowest in the world, with stillbirth occurring in <4/1000 births [1]. However, similar to other high-income countries, immigrants from low-income regions have higher rates of stillbirth, as well as preterm birth, with most pronounced excess incidence among women of African origin [2, 3]. The reasons underlying this discrepancy remain unclear; apart from socioeconomic conditions [2, 4], it has been suggested that infectious diseases may be involved [5, 6]. In this context, tuberculosis (TB) is of special interest because it disproportionately affects persons from low-and middle-income countries, with >90% of cases of TB disease in Sweden occurring in individuals originating from TB-endemic areas [7].
Several reciprocal interactions between TB and pregnancy exist. TB is a leading nonobstetric cause of maternal mortality globally [8, 9] and has also been associated with a range of other pregnancy complications, including preterm delivery, low birthweight, and perinatal mortality [10–12]. In addition, the incidence of TB disease is elevated in connection to pregnancy [13, 14].
TB infection (TBI) is commonly defined as immune sensitization to Mycobacterium tuberculosis without evidence of TB disease, and is currently regarded as a continuous spectrum, ranging from spontaneously resolved infection to incipient TB with elevated bacterial activity and risk of progression to TB disease [15, 16]. The increased incidence of TB disease reported during pregnancy and postpartum [13, 14] probably reflects progression of TBI triggered by physiological immune modifications [17]. Furthermore, recent studies (both in pregnant women and in other population groups) imply that TBI can be associated with systemic pro-inflammatory responses [18–20].
In this study, we aimed to explore the association between TBI and pregnancy complications, based on register linkage of women with history of immigration to Sweden from TB-endemic countries who were offered routine screening for TBI during antenatal care.
METHODS
Study Setting
Screening for TBI among women with history of immigration from TB-endemic countries (TB incidence >100/100 000 per year [21]) was implemented in antenatal care in most regions of Sweden in 2014, based on QuantiFERON (QFT) assays (QuantiFERON-TB GOLD IN-TUBE or GOLD PLUS [Qiagen, Hilden, Germany]). Women with positive QFT results and/or clinically suspected TB disease are referred to infectious disease clinics for further evaluation. Pregnancy per se is not an indication for TB preventive therapy (TPT) according to Swedish guidelines. TPT is recommended for women with recent (<2 years) contact with contagious TB or other recognized risk factors for disease progression, regardless of pregnancy status [22].
Study Design
QFT results obtained at pregnancy screening 2014–2018 from 5 Swedish regions (Skåne, Kalmar, Värmland, Östergötland, and Stockholm; total uptake population, 4.8 million) were linked to data from the Patient Register, Pregnancy Register, and Population Register. Data linkage was performed by the National Health and Welfare Board using national identification numbers (available for Swedish citizens and immigrants with permanent residence permit).
The Patient Register was used to extract data on International Classification of Diagnosis-10 (ICD-10) codes from hospital admissions and outpatient visits in specialized healthcare from 2000 to 2019 (including delivery clinics). The Pregnancy Register was introduced in 2014 and covers 97%–98% of deliveries in Sweden. This register contains demographic, medical, and obstetric information, as well as detailed pregnancy outcome data, including ICD-10 codes, from primary and specialized obstetric care for all registered pregnancies. Data were collected from the Pregnancy Register from 2014 until June 2020. Data on immigration dates and region of origin were collected from the Population Register. For individuals with ICD-10 codes indicating TB disease, data on case notification were retrieved from the Public Health Agency of Sweden.
Individuals without valid QFT results or national identification numbers were excluded, as were women who were tested outside of pregnancy, or with residence in Sweden since before 2000, and women with human immunodeficiency virus (HIV). Pregnancies occurring >10 years after immigration to Sweden, multiplex pregnancies, pregnancies ending with miscarriage or induced abortion, and pregnancies in women with TB disease before or in connection to the index pregnancy were excluded from this analysis. All pregnancies occurring for each included participant during the study period were eligible for inclusion. Separately, we characterized pregnancy outcomes in women diagnosed with TB disease in connection to pregnancy.
Study Definitions
QuantiFERON results were categorized as positive or negative, as recommended by the manufacturer, using a cutoff level of M. tuberculosis antigen-stimulated interferon-γ of 0.35 IU/mL (after correction for nil-tube interferon-γ levels). Only individuals with valid QFT result were eligible; indeterminate QFT results were not considered. Because of the low TB transmission rate in Sweden, we assumed that participants with negative QFT results did not acquire TB infection after the time of testing, and that participants with positive QFT results had acquired TB infection before immigration to Sweden. Episodes of TB disease were identified in the Patient Register using ICD codes (Supplementary Table 1). If these ICD codes had been assigned from clinics not routinely involved in management of TB, we also required notification in the Public Health Agency TB Register to define TB disease.
Pregnancies were identified from the Patient Register (2000–2019) as well as from the Pregnancy Register (2014–June 2020) (Supplementary Table 1). The following pregnancy complications were defined by presence of ICD-10 codes obtained from the Patient Register (Supplementary Table 1): preeclampsia, severe preeclampsia (including eclampsia and hemolysis, elevated liver enzymes, low platelet count syndrome syndrome), polyhydramnios, oligohydramnios, emergency cesarean section, stillbirth, and intrauterine growth restriction. From the Pregnancy Register the following additional outcomes were collected: birthweight (analyzed both as continuous and categorical variable [low birthweight, < 2500 g]), gestational length (analyzed both as continuous and categorical variable [prematurity, < 37 weeks of gestation]), stillbirth, neonatal death, and emergency cesarean section.
For descriptive purposes, African origin was further subdivided into Northern Africa and sub-Saharan Africa. Because of the high proportion of immigrants from the Horn of Africa in Sweden, persons with this geographical origin were described separately. Similarly, Asian origin was further subdivided into Central, Eastern, South-eastern, Southern, and Western Asia.
Statistical Analysis
Each outcome was assessed for univariate association with TBI status using simple logistic regression, followed by multivariable analyses using generalized linear mixed modelling with binomial distribution for categorical and linear distribution for continuous outcomes. To adjust for dependence between repeated pregnancies for the same woman (ie, cluster effects), the models contained random effects for maternal identity.
The main multivariable analyses were adjusted for maternal age at delivery (categorized with five 5-year intervals) and region of origin. To allow convergence of the mixed effect regression models, the variable geographical origin was reduced to African versus non-African origin (Table 1).
Table 1.
Characteristics of Included Study Participants at Their Pregnancies
| Total | TB Uninfected | TBI | |
|---|---|---|---|
| Characteristic | N = 12 443 | N = 9907 | N = 2536 |
| N (%) | N (%) | N (%) | |
| Age (y), mean (SD)a | 29.4 (5.4) | 29.2 (5.3) | 30 (5.4) |
| ≤20 | 274 (2.2) | 226 (2.3) | 48 (1.9) |
| >20–25 | 2418 (19.4) | 2001 (20.2) | 417 (16.4) |
| >25–30 | 4020 (32.3) | 3235 (32.7) | 785 (31) |
| >30–35 | 3382 (27.2) | 2631 (26.6) | 751 (29.6) |
| >35–40 | 1530 (12.3) | 1170 (11.8) | 360 (14.2) |
| >40 | 345 (2.8) | 252 (2.5) | 93 (3.7) |
| NA | 474 (3.8) | 392 (4) | 82 (3.2) |
| Time since immigration, y | |||
| ≤2 | 4270 (34.3) | 3536 (35.7) | 734 (28.9) |
| >2–5 | 4400 (35.4) | 3484 (35.2) | 916 (36.1) |
| >5–10 | 3773 (30.3) | 2887 (29.1) | 886 (34.9) |
| NA | 0 (0) | 0 (0) | 0 (0) |
| Highest level of education completedb | |||
| Less than 9 y | 1518 (12.2) | 1107 (11.2) | 411 (16.2) |
| At least 9 y | 2379 (19.2) | 1884 (19) | 495 (19.5) |
| Secondary school | 2965 (23.8) | 2408 (24.3) | 557 (22) |
| University | 2372 (19.1) | 1990 (20.1) | 382 (15.1) |
| NA | 3209 (25.8) | 2518 (25.4) | 691 (27.2) |
| Geographical origin | |||
| Africa | 5700 (45.8) | 4019 (40.6) | 1681 (66.3) |
| Horn of Africab | 3264 (26.2) | 2286 (23.1) | 978 (38.9) |
| Other sub-Saharan Africa | 857 (6.9) | 588 (5.9) | 269 (10.6) |
| North Africa | 441 (3.5) | 387 (3.9) | 63 (2.5) |
| Africa, unspecified country | 1138 (9.1) | 767 (7.7) | 371 (14.6) |
| Asia | 5818 (46.8) | 5076 (51.2) | 742 (29.3) |
| Central | 90 (1.5) | 79 (0.8) | 11 (0.4) |
| Eastern | 267 (4.5) | 204 (2.1) | 63 (2.5) |
| South-Eastern | 512 (8.8) | 404 (4.1) | 108 (4.3) |
| Southern | 1872 (32.2) | 1617 (16.3) | 255 (10.1) |
| Western | 2015 (34.6) | 1834 (18.5) | 181 (7.1) |
| Asia, unspecified country | 1062 (18.2) | 938 (9.5) | 124 (4.9) |
| Europe | 751 (6) | 695 (7) | 89 (3.5) |
| North America | 14 (0.1) | 10 (0.1) | 4 (0.2) |
| Oceania | 4 (0) | 1 (0) | 3 (0.1) |
| South America | 106 (0.9) | 90 (0.9) | 16 (0.6) |
| NA | 17 (0.1) | 16 (0.2) | 1 (0) |
| Parity | |||
| 0 | 4405 (35.4) | 3596 (36.3) | 809 (31.9) |
| 1–2 | 5774 (46.4) | 4618 (46.6) | 1156 (45.6) |
| 3–5 | 1534 (12.3) | 1138 (11.5) | 396 (15.6) |
| >5 | 226 (4.1) | 145 (1.5) | 81 (3.2) |
| NA | 504 (4.1) | 410 (4.1) | 94 (3.7) |
| Sex of offspring | |||
| Female | 4660 (37.5) | 3753 (37.9) | 907 (35.8) |
| Male | 4786 (38.5) | 3848 (38.8) | 938 (37.0) |
| NA | 2997 (24.1) | 2306 (23.3) | 691 (27.3) |
| BMI (kg/m2, mean [SD])b | 25.6 (5) | 25.5 (5) | 26.2 (5.1) |
| ≤25 | 4597 (36.9) | 3793 (38.3) | 804 (31.7) |
| >25–30 | 2857 (23) | 2256 (22.8) | 601 (23.7) |
| >30 | 1609 (12.9) | 1257 (12.7) | 352 (13.9) |
| NA | 3380 (27.2) | 2601 (26.3) | 779 (30.7) |
| Smoking at enrollment | |||
| Yes | 193 (1.6) | 168 (2.3) | 25 (1.4) |
| No | 8857 (71.2) | 7111 (97.7) | 1746 (98.6) |
| NA | 3393 (27.3) | 2628 (26.5) | 765 (30.2) |
Each woman may be represented by more than 1 pregnancy. BMI was measured at antenatal care enrollment.
Abbreviations: BMI, body mass index; NA, not available; SD, standard deviation; TB, tuberculosis; TBI, tuberculosis infection.
aContinuous outcomes are represented by mean, SD, and crude and adjusted difference rather than those indicated at the table header.
bEthiopia, Eritrea, Somalia, and Djibouti were considered to constitute the Horn of Africa.
As a sensitivity analysis, we performed multivariable analyses with adjustment for parity (categorized at 0, 1–2, 3–5, and >5), highest level of completed education, body mass index ([BMI] categorized with thresholds at <25, 25–30, and >30 kg/m2), and smoking status at the inclusion visit, in addition to maternal age and region of origin (for women with available information for all these variables). To allow for convergence of the expanded models, the variable parity was reduced to first versus repeated pregnancy for the analysis of 3 outcomes (severe preeclampsia, polyhydramnios, and placental ablatio). Furthermore, subgroup analyses were performed for women of African and Asian origin, respectively (Supplementary Table 2). These adjusted for age in tertiles to allow model convergence.
The distribution of birthweight and gestational length were explored with respect to maternal TBI. Besides analysis based on categorical definitions of low birthweight and prematurity, fixed effects quantile regression was applied to analyze adjusted differences on the second, third, fifth, and 10th percentiles, with similar adjustments. Quantile regression was performed to explore whether effects on these outcomes were more pronounced in a subset of participants, effects that may be diluted in analysis of average data.
Characteristics associated with stillbirths were presented descriptively by TBI status and gestational length.
Statistical analysis was performed in R, version 1.4.1717 [23]. Mixed effect models and quantile regression were preformed using the lme4 [24] and quantreg [25] R packages, respectively.
Ethical Considerations
Ethical approval was granted by the Swedish Ethical Review Authority (DNR 2019-01448).
RESULTS
Participant Characteristics
Of 10 464 women with valid QFT results and at least 1 registered pregnancy, 3056 (29.2%) were excluded, leaving 7408 women with 12 443 pregnancies for analysis (Figure 1). The proportions of women with TBI were similar among excluded and included women (16.1% vs 19.0%).
Figure 1.
Flow chart of inclusion of women originating from tuberculosis (TB)-endemic countries screened for latent TB infection during pregnancy.
Among included participants, 1408 (19.0%) had TBI. In total, 2536 (20.4%) of 12 443 deliveries occurred in participants with TBI (Table 1). Women with TBI were more frequently of African origin (834/1408, 59.2% vs 2101/6000, 35.0%). Among women of African origin, 71.5% (n = 3264) were from the Horn of Africa. The distribution of TBI was comparable among women from the Horn of Africa (450/1599, 28.1%) versus other sub-Saharan countries (153/503, 30.4%). Participants with TBI were similar to TB-uninfected women with regard to mean age at registration of pregnancy (30.0 vs 29.2 years) and history of previous pregnancy (68.1% vs 64.7%). The proportion of women with >5 years’ residence in Sweden was similar with respect to TBI status (34.9% vs 29.1%), as was mean BMI (26.2 kg/m2 vs 25.5 kg/m2). Smoking at antenatal care enrollment was less frequently reported by women with TBI (25/1771 [1.4%] vs 168/7279 [2.3%]), and women with TBI more frequently reported <9 years of formal schooling (16.2% vs 11.2%).
Pregnancy Outcomes With Respect to TBI
In univariable analysis, pregnancies in women with TBI were more frequently associated with stillbirth (odds ratio [OR], 2.16; 95% confidence interval [CI], 1.34–3.40), preeclampsia (OR, 1.33; 95% CI, 1.03–1.70), severe preeclampsia (OR, 1.68; 95% CI, 1.17–2.38), low birthweight (OR, 1.26; 95% CI, 1.00–1.59), and emergency cesarean section (OR, 1.23; 95% CI, 1.08–1.40) compared with pregnancies in TB-uninfected women (Table 2).
Table 2.
Pregnancy Outcomes by Maternal TBI Status
| TB Uninfected | TBI | Crude Analysis | Adjusted Analysis | Expanded Adjusted Analysis | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Categorical outcomes | N (%) | N (%) | COR | 95% CI | P | AOR | 95% CI | P | AOR | 95% CI | P |
| Stillbirth | 51 (0.51) | 28 (1.1) | 2.16 | 1.34–3.40 | .001 | 1.90 | 1.13–3.21 | .016 | 1.97 | 1.00–3.91 | .051 |
| Neonatal death | 7 (0.09) | 4 (0.22) | 2.35 | .62–7.81 | .17 | 1.90 | .54–6.68 | .32 | 2.98 | .57–15.6 | .20 |
| Preeclampsia | 246 (2.5) | 83 (3.3) | 1.33 | 1.03–1.70 | .028 | 1.27 | .79–2.04 | .32 | 1.22 | .65–2.26 | .54 |
| Severe preeclampsiaa | 103 (1.0) | 44 (1.7) | 1.68 | 1.17–2.38 | .004 | 1.62 | 1.03–2.56 | .036 | 1.85 | .98–3.49 | .060 |
| Emergency cesarean section | 1119 (11.3) | 344 (13.6) | 1.23 | 1.08–1.40 | .002 | 1.28 | 1.02–1.63 | .033 | 1.68 | 1.23–2.28 | .0009 |
| Placental ablation | 37 (0.4) | 13 (0.5) | 1.37 | .70–2.52 | .33 | 1.34 | .61–2.98 | .46 | 1.41 | .62–3.24 | .41 |
| Polyhydramnios | 55 (0.6) | 21 (0.8) | 1.50 | .88–2.44 | .12 | 1.29 | .65–2.58 | .47 | 1.40 | .60–3.23 | .43 |
| Oligohydramnios | 283 (2.9) | 73 (2.9) | 1.01 | .77–1.30 | .95 | .91 | .66–1.25 | .56 | 1.02 | .67–1.53 | .94 |
| Low birthweightb | 319 (4.2) | 97 (5.3) | 1.26 | 1.00–1.59 | .049 | 1.38 | 1.01–1.88 | .041 | 1.38 | .96–1.98 | .080 |
| Prematurityc | 356 (4.7) | 97 (5.3) | 1.13 | .89–1.42 | .29 | 1.23 | .92–1.64 | .16 | 1.32 | .96–1.81 | .087 |
| IUGRd | 486 (4.9) | 126 (5.0) | 1.01 | .83–1.23 | .90 | 1.04 | .79–1.36 | .79 | 1.18 | .82–1.69 | .38 |
| Continuous outcomes | Value | Value | Diff. | 95% CI | P | Adj. diff. | 95% CI | P | Adj. diff. | 95% CI | P |
| Birthweight (g), mean (SD) | 3407 (560) | 3424 (607) | 16 | −13–45 | .27 | −3.5 | −37–30 | .84 | −11 | −48–26 | .56 |
| Birthweight 2nd percentile | 2094 | 1866 | −228 | −514–34 | .086 | −145 | −426–136 | .31 | −269 | −412 – −124 | .0003 |
| Birthweight 3rd percentile | 2345 | 2155 | −190 | −323 – −57 | .0052 | −210 | −333 – −87 | .0008 | −191 | −336 – −48 | .009 |
| Birthweight 5th percentile | 2570 | 2460 | −110 | −244–24 | .10 | −114 | −237–9 | .0069 | −119 | −240–1 | .052 |
| Birthweight 10th percentile | 2790 | 2760 | −30 | −85–25 | .10 | −62 | −123 – −1 | .045 | −45 | −98–8 | .095 |
| Gestation (wk), mean (SD) | 39.7 (2.0) | 39.7 (2.2) | 0.00 | −.10–.10 | .98 | −0.09 | −.20–.02 | .17 | −0.11 | −.23–.01 | .10 |
| Gestation 2nd percentile | 34.6 | 34.0 | −.60 | −2.1–.99 | .47 | −1.29 | −2.7–.16 | .082 | −1.27 | −2.39 – −.15 | .027 |
| Gestation 3rd percentile | 36.0 | 35.1 | −.90 | −1.89 – −.12 | .026 | −1.07 | −1.89 – −.26 | .0099 | −1.14 | −2.34–.05 | .060 |
| Gestation 5th percentile | 37.0 | 36.9 | −.10 | −.73–.44 | .63 | −0.42 | −.99–.13 | .13 | −0.57 | −1.03 – −.11 | .016 |
| Gestation 10th percentile | 38.0 | 37.9 | −.10 | −.40–.11 | .27 | −0.14 | −.37–.09 | .22 | −0.20 | −.44–.04 | .11 |
Adjustments were made for maternal age and African origin, with random effects fitted for maternal study code to account for repeated observations. Because of missing data, the adjusted models excluded 481 (3.9%) pregnancies. The expanded adjusted analyses also include covariates maternal education, parity, body mass index, and smoking; because of missing data, 4964 (39.9%) pregnancies were excluded from these models. Quantile regression models for percentiles of birthweight and gestational age did not include random effects.
Abbreviations: Adj. diff., adjusted difference; AOR, adjusted odds ratio; CI, confidence interval; COR, crude odds ratio; Diff., crude difference; HELLP, hemolysis, elevated liver enzymes, low platelet count syndrome; ICD, International Classification of Disease; IUGR, intrauterine growth restriction; NA, not available; SD, standard deviation; SGA, small for gestational age; TB, tuberculosis; TBI, TB infection.
aSevere preeclampsia, HELLP syndrome, or eclampsia. This model required parity to be categorized at 0 or >0 to allow model convergence.
bIUGR defined by presence of ICD-10 code O365 (maternity care due to known or suspected IUGR).
cLow birthweight was defined as birthweight <2500 g.
dPrematurity was defined as gestational length <37 weeks.
In multivariable analysis adjusted for maternal age and region of origin, TBI remained associated with stillbirth (adjusted odds ratio [AOR], 1.90; 95% CI, 1.13–3.21; P = .016), low birthweight (AOR, 1.38; 95% CI, 1.01–1.88; P = .041), and emergency cesarean section (AOR, 1.28; 95% CI, 1.02–1.63; P = .033; Table 2). The association with preeclampsia did not reach statistical significance in multivariable analysis, but severe preeclampsia (including hemolysis, elevated liver enzymes, low platelet count syndrome and eclampsia) remained significantly associated with TBI (AOR, 1.62; 95% CI, 1.03–2.56; P = .036).
In subgroup analyses of women of Asian and African origin, respectively, TBI was significantly associated with stillbirth, severe preeclampsia, low birthweight, and prematurity in women of Asian origin, whereas these associations did not reach statistical significance in the subset of women originating from Africa (Supplementary Table 2).
Mean birthweight and gestational length were similar with respect to TBI. Despite this, low birthweight remained significantly associated with TBI, suggesting that an effect on birthweight might not be uniform but rather restricted to a subset of women with TBI. For this reason, multivariable quantile regression was performed for the association between maternal TBI and birthweight and for gestational age, respectively. This revealed a difference of 210 g (95% CI, 87–333 g; P = .0008) at the third birthweight percentile, after adjustment for age and origin (Table 2). Similarly, the third percentile of gestational age was 1.07 weeks (95% CI, .26–1.89 weeks; P = .01) shorter in women with TBI after adjustment for age and origin.
Multivariable regressions using the expanded set of covariates was limited by considerable proportions of missing data for some covariates. A total of 7479 (60.1%) pregnancies were included in this analysis. Confidence intervals were wider for these models. Stillbirth (AOR, 1.97; P = .051), severe preeclampsia (AOR, 1.83; P = .060), and low birthweight (AOR, 1.38; P = .080) did not reach statistical significance. However, adjusted OR estimates indicate similar strength of the associations between TBI and the outcomes compared with the main multivariable analysis. Quantile regression analyses with the expanded set of covariates yielded comparable results to the main analyses (Table 2).
Characteristics of stillbirths in women with TBI were compared with those of TB-uninfected women, further disaggregated by prematurity, as shown in Table 3. Proportions of preterm delivery were higher in stillbirths with maternal TBI (14/28, 50% vs 17/51, 33%; not significant). Gestational diabetes mellitus was recorded in 6 (11.8%) stillbirths of TB-uninfected participants and none in participants with TBI (NS). Two cases of chorioamnionitis were recorded in preterm stillbirths of women with TBI.
Table 3.
Characteristics of Pregnancies Complicated by Stillbirths by TBI Status
| Characteristics | TB-uninfected | TBI | ||
|---|---|---|---|---|
| ≥37 wk | <37 wk | ≥37 wk | <37 wk | |
| Age (y) | N = 34 | N = 17 | N = 14 | N = 14 |
| ≤20 | 0 | 1 (5.9) | 0 | 0 |
| 21–25 | 4 (11.8) | 2 (11.8) | 4 (28.6) | 1 (7.1) |
| 26–30 | 6 (17.6) | 5 (29.4) | 5 (35.7) | 4 (28.6) |
| 31–35 | 13 (38.2) | 4 (23.5) | 2 (14.3) | 8 (57.1) |
| 36–40 | 7 (20.6 | 3 (17.6) | 1 (7.1) | 1 (7.1) |
| >40 | 1 (2.9) | 1 (5.9) | 0 | 0 |
| NA | 3 (8.8) | 1 (5.9) | 2 (14.3) | 0 |
| Parity | ||||
| 0 | 13 (38.2) | 5 (29.4) | 8 (57.1) | 4 (28.6) |
| 1–2 | 10 (29.4) | 9 (52.9) | 4 (28.6) | 6 (42.9) |
| 3–5 | 8 (23.5) | 1 (5.9) | 0 | 4 (28.6) |
| >5 | 0 | 1 (5.9) | 0 | 0 |
| NA | 3 (8.8) | 1 (5.9) | 2 (14.3) | 0 |
| Gestational age, wk | ||||
| <27 | 0 | 6 (35.3) | 0 | 3 (21.4) |
| 27–36 + 6 | 0 | 9 (35.3) | 0 | 10 (71.4) |
| 37–41 + 6 | 18 (52.9) | 0 | 5 (35.7) | 0 |
| ≥42 | 2 (5.9) | 0 | 0 | 0 |
| NA | 14 (41.2) | 2 (11.8) | 9 (64.3) | 1 (7.1) |
| IUGRa | 1 (2.9) | 3 (17.6) | 0 | 2 (14.3) |
| Gestational DM | 5 (14.7) | 1 (5.9) | 0 | 0 |
| PPROM | 1 (2.9) | 1 (5.9) | 0 | 1 (7.1) |
| Ablation | 3 (8.8) | 0 | 1 (7.1) | 1 (7.1) |
| Emergency cesarean section | 5 (14.7) | 2 (11.8) | 1 (7.1) | 0 |
| Maternal sepsis | 0 | 0 | 0 | 0 |
| Bacterial chorioamnionitis | 0 | 0 | 0 | 2 (14.3) |
| Puerperal infectionb | 6 (17.6) | 1 (5.9) | 2 (14.3) | 1 (7.1) |
In this table, stratification by preterm birth was based on (1) gestational age and, if absent, (2) presence of ICD-10 codes indicating preterm birth.
Abbreviations: HELLP syndrome, hemolysis, elevated liver enzymes, low platelet count syndrome; ICD, International Classification of Disease; IUGR, intrauterine growth restriction;
NA, not available; PROM, premature preterm rupture of membranes; TB, tuberculosis; TBI, tuberculosis infection.
aIUGR defined by presence of ICD-10 code O365 (maternity care due to known or suspected IUGR).
bPuerperal infection defined by presence of ICD-10 codes O85 and O86.
Thirty-two pregnancies were recorded in women with TB disease in connection to pregnancy. None of these resulted in stillbirth. Three women with TB disease (9.4%) had preeclampsia, 1 had oligohydramnios, 6 (18.8%) were delivered by emergency cesarean section, and 1 had intrauterine growth restriction.
DISCUSSION
In this study, we investigated whether TBI contributes to the excess burden of adverse pregnancy outcomes in women with history of immigration to Sweden from TB-endemic countries. We found significantly increased incidence of stillbirth, severe preeclampsia, emergency cesarean section, and low infant birthweight in pregnancies of women with TBI.
The differences remained in multivariable analysis after adjustment for maternal age and African versus non-African region of origin.
TB disease has been associated with a range of adverse pregnancy outcomes, although the pathogenesis has not been elucidated. Different mechanisms have been implicated for the impact of other microbes on pregnancy; either direct invasion of the fetus or placenta or effects mediated by pathogen-induced immune responses [26, 27]. The placental vascularization may be affected by maternal immune alterations [27], which in turn could have adverse impact on pregnancy outcome [27, 28]. Similar mechanisms have been proposed to be involved in pregnant women with TB disease [29].
Both epidemiological and laboratory data suggest that the balance of immune control in women with TBI changes during pregnancy [13, 14, 17]. The elevated incidence of TB disease in connection to pregnancy observed in low-endemic countries is likely to be due to pregnancy-induced progression of TBI [13, 14]. Furthermore, we and other researchers have found altered M. tuberculosis–specific immune responses during pregnancy [17, 20, 30, 31], suggesting that the host–pathogen interaction changes during pregnancy [17, 20].This register-based study was not designed to investigate mechanisms involved in these phenomena. Yet, we speculate that low-grade inflammation triggered by TBI during pregnancy might be involved in the association between TBI and pregnancy complications observed in our study population [26, 27, 32]. Interestingly, analysis of continuous variables using quantile regression revealed that while mean birthweight and gestational length were almost identical with respect to TBI status, there were substantial differences in the lower-end distribution of these variables, especially around the third percentiles. This suggests that the effect of TBI on gestational length and birthweight is concentrated in a subset of pregnancies.
In subanalysis of women originating from Asia (constituting 47% of the study population and 29% of women with TBI), several pregnancy complications showed strong associations with TBI, supporting the association between TBI and pregnancy complications. However, these associations did not reach statistical significance among women of African origin. The reasons underlying this discrepancy cannot be determined from our study design. Because most these women originated from the Horn of Africa, it is possible that other factors (eg, socioeconomic condition) had a stronger impact on pregnancy outcome in this subpopulation. In line with this, the burden of stillbirth and severe preeclampsia were lower among Asian compared with African TB uninfected women.
To our knowledge, the association between TBI and pregnancy outcome has not been studied previously using a similar design. In a small study performed in Ethiopia, we found higher rates of stillbirth in women with TBI (4.0% vs 2.7%); however, this difference did not reach statistical significance [33]. Further studies are needed, both to corroborate our findings from other settings and to investigate the mechanisms involved in the association between TBI and pregnancy complications observed in our study.
Our study was based on immigrant women seeking antenatal care in Sweden who met indications for routine TBI screening, which was performed using interferon-γ release assays. We excluded women with TB disease before in connection to the pregnancy for a reliable definition of TBI. These data were linked to national registers containing detailed information on pregnancy outcomes as well as factors that may contribute to these outcomes, minimizing the risk of confounding. Importantly, all women in the target population had equal access to antenatal care, with nearly all deliveries in Sweden taking place within the public healthcare system.
Certain limitations should be considered. To reduce the risk of residual confounding, we chose to limit the study population to women originating from TB-endemic countries, all of whom were eligible for TB screening during pregnancy according to Swedish guidelines. Data on potentially confounding variables, such as socioeconomic status, were not available. For some covariates, data were missing from relatively high proportions of women; for example, information on BMI was not available in 27.2% of pregnancies. For this reason, we performed 2 sets of multivariable analyses. Although multivariable analysis based on the expanded set of covariates had reduced power because of lower study sample, the OR estimates were similar, indicating that these covariates did not explain the observed association between TBI and pregnancy outcome. In support of this, a literature review performed by Gissler et al. [4] concluded that socioeconomic background characteristics explain only a minor proportion of the excess risk of perinatal mortality in immigrant groups in industrialized countries.
Similar to other studies on TBI, we were unable to categorize participants with regard to degree of bacterial activity and persistence. It is probable that some study participants with TBI had spontaneously resolved infection [16], whereas others could have had incipient TB that did not progress to clinically overt active disease. In addition, we did not have information on receipt of TPT. Recent trials in women with HIV have shown inconsistent results regarding the effect of TPT on pregnancy outcomes [34, 35]. In light of this, and that TPT is not generally recommended for pregnant women in Swedish guidelines, we consider it unlikely that TPT would explain the observed differences in pregnancy outcomes. Furthermore, our data sources did not allow for analysis of the potential impact of TBI on fetal deaths occurring before 22 gestational weeks.
CONCLUSION
In women immigrating to Sweden from TB-endemic countries and who underwent routine TBI screening during antenatal care, TBI was independently associated with increased risk of stillbirth, severe preeclampsia, emergency cesarean section, and low birthweight. Further studies are needed to corroborate these findings and to explore pathophysiological mechanisms.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Supplementary Material
Contributor Information
John Walles, Clinical Infection Medicine, Department of Translational Medicine, Lund University, Malmö, Sweden; Department of Infectious Diseases, Central Hospital, Kristianstad, Sweden; Department of Clinical Microbiology, Infection Control and Prevention, Skåne University Hospital Lund, Lund, Sweden.
Niclas Winqvist, Skåne Regional Office for Infectious Disease Control and Prevention, Malmö, Sweden.
Stefan R Hansson, Division of Obstetrics and Gynaecology, Department of Clinical Sciences Lund, Lund University, Lund, Sweden; Skåne University Hospital, Lund, Sweden.
Erik Sturegård, Clinical Infection Medicine, Department of Translational Medicine, Lund University, Malmö, Sweden; Department of Clinical Microbiology, Infection Control and Prevention, Skåne University Hospital Lund, Lund, Sweden.
Haitham Baqir, Department of Clinical Microbiology, Linköping University Hospital, Linköping, Sweden.
Anna Westman, Department of Infectious Diseases, Danderyd Hospital, Stockholm, Sweden; Division of Clinical Microbiology, Department of Laboratory Medicine, Karolinska University Hospital Laboratory, Stockholm, Sweden.
Torbjörn Kjerstadius, Laboratory Medicine, Clinical Microbiology, Central Hospital, Karlstad, Sweden.
Thomas Schön, Department of Biomedical and Clinical Sciences, Linköping University, Linköping, Sweden; Department of Infectious Diseases, Kalmar County Hospital, Linköping University, Kalmar, Sweden; Department of Infectious Diseases, Linköping University, Linköping, Sweden.
Per Björkman, Clinical Infection Medicine, Department of Translational Medicine, Lund University, Malmö, Sweden; Department of Infectious Diseases, Skåne University Hospital, Malmö, Sweden.
Notes
Acknowledgments. The authors are grateful for the valuable guidance on the choice and execution of statistical analysis by Gaetano Marrone (Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden), for the assistance in data extraction from the Pregnancy Register by Jonas Söderling (Clinical Epidemiology Division, Department of Medicine, Solna, Karolinska Institutet, Stockholm, Sweden), and of verifying the notification of cases of tuberculosis disease cases from Jerker Jonsson (Swedish Public Health Agency).
Financial support . This work was supported by the Swedish Heart-Lung Foundation (grant number 20170258 to P. B.); the Crafoord Foundation (grant number 20170537 to P. B.); the Alfred Österlund Foundation, Region Skåne research grants (grant number 41509 to P. B.); governmental funding of clinical research within the National Health Services Sweden, The ALF-agreement (grant number 40103 to P. B. and to J. W.), and the Centralsjukhuset Kristianstad (grant number 2019-F016 to J. W.).
References
- 1. Dödfödda barn En inventering och förslag på åtgärder. Stockholm. 2018.
- 2. Ekéus C, Cnattingius S, Essén B, Hjern A. Stillbirth among foreign-born women in Sweden. Eur J Public Health 2011; 21:788–92. [DOI] [PubMed] [Google Scholar]
- 3. Ammoura O, Sehouli J, Kurmeyer C, et al. Perinatal data of refugee women from the Gynaecology Department of Charité University Hospital Berlin compared with German federal analysis. Geburtshilfe Frauenheilkd 2021; 81:1238–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Gissler M, Alexander S, Macfarlane A, et al. Stillbirths and infant deaths among migrants in industrialized countries. Acta Obstet Gynecol Scand 2009; 88:134–48. [DOI] [PubMed] [Google Scholar]
- 5. Barona-Vilar C, López-Maside A, Bosch-Sánchez S, et al. Inequalities in perinatal mortality rates among immigrant and native population in Spain, 2005–2008. J Immigr Minor Heal 2014; 16:1–6. [DOI] [PubMed] [Google Scholar]
- 6. Boga JA, Casado L, Fernández-Suarez J, et al. Screening program for imported diseases in immigrant women: analysis and implications from a gender-oriented perspective. Am J Trop Med Hyg 2020; 103:480–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Folkhälsomyndigheten. Tuberkulos – sjukdomsstatistik—Folkhälsomyndigheten. 2022.
- 8. Menéndez C, Romagosa C, Ismail MR, et al. An autopsy study of maternal mortality in Mozambique: the contribution of infectious diseases. PLoS Med 2008; 5:0220–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Dennis EM, Hao Y, Tamambang M, et al. Tuberculosis during pregnancy in the United States: racial/ethnic disparities in pregnancy complications and in-hospital death. PLoS One 2018; 13:e0194836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Sobhy S, Babiker Z, Zamora J, Khan K, Kunst H. Maternal and perinatal mortality and morbidity associated with tuberculosis during pregnancy and the postpartum period: a systematic review and meta-analysis. BJOG An Int J Obstet Gynaecol 2016:727–33. [DOI] [PubMed] [Google Scholar]
- 11. Figueroa-Damian R, Arredondo-Garcia JL. Neonatal outcome of children born to women with tuberculosis. Arch Med Res 2001; 32:66–9. [DOI] [PubMed] [Google Scholar]
- 12. Jana N, Vasishta K, Jindal SK, Khunnu B, Ghosh K. Perinatal outcome in pregnancies complicated by pulmonary tuberculosis. Int J Gynecol Obstet 1994; 44:119–24. [DOI] [PubMed] [Google Scholar]
- 13. Zenner D, Kruijshaar ME, Andrews N, Abubakar I. Risk of tuberculosis in pregnancy: a national, primary care-based cohort and self-controlled case series study. Am J Respir Crit Care Med 2012; 185:779–84. [DOI] [PubMed] [Google Scholar]
- 14. Jonsson J, Kühlmann-Berenzon S, Berggren I, Bruchfeld J. Increased risk of active tuberculosis during pregnancy and postpartum: a register-based cohort study in Sweden. Eur Respir J 2020; 55:1901886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Migliori GB, Ong CWM, Petrone L, D’ambrosio L, Centis R, Goletti D. The definition of tuberculosis infection based on the spectrum of tuberculosis disease. Breathe 2021; 17:210079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Drain PK, Bajema KL, Dowdy D, et al. Incipient and subclinical tuberculosis: a clinical review of early stages and progression of infection. Clin Microbiol Rev 2018; 31:e00021-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Tesfaye F, Walles J, Winqvist N, et al. Longitudinal Mycobacterium tuberculosis-specific interferon gamma responses in Ethiopian HIV-negative women during pregnancy and postpartum. J Clin Microbiol 2021; 59:e0086821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Huaman MA, Henson D, Rondan PL, et al. Latent tuberculosis infection is associated with increased unstimulated levels of interferon-gamma in Lima, Peru. PLoS One 2018; 13:e0202191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Huaman MA, Deepe GS, Fichtenbaum CJ. Elevated circulating concentrations of interferon-gamma in latent tuberculosis infection. Pathog Immun 2016; 1:291–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Naik S, Alexander M, Kumar P, et al. Systemic inflammation in pregnant women with latent tuberculosis infection. Front Immunol 2021; 11:587617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. World Health Organization . WHO global lists of high burden countries for tuberculosis (TB). TB/HIV and TB (MDR/RR-TB 2021:2021–2025.
- 22. Folkhälsomyndigheten. Rekommendationer för preventiva insatser mot tuberkulos. 2020.
- 23. R Core Team . R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. 2017.
- 24. Bates D, Mächler M, Bolker BM, Walker SC. Fitting linear mixed-effects models using lme4. J Stat Softw 2015; 67. [Google Scholar]
- 25. Koenker R. Quantreg: quantile regression. R package version 5.38. 2018.
- 26. Kalagiri R, Carder T, Choudhury S, et al. Inflammation in complicated pregnancy and its outcome. Am J Perinatol 2016:76508. [DOI] [PubMed] [Google Scholar]
- 27. Weckman AM, Ngai M, Wright J, McDonald CR, Kain KC. The impact of infection in pregnancy on placental vascular development and adverse birth outcomes. Front Microbiol 2019; 10:1924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Redman CWG, Staff AC, Roberts JM. Syncytiotrophoblast stress in preeclampsia: the convergence point for multiple pathways. Am J Obstet Gynecol 2021; 226:S907–S927. [DOI] [PubMed] [Google Scholar]
- 29. Phoswa WN, Eche S, Khaliq OP. The association of tuberculosis mono-infection and tuberculosis-human immunodeficiency virus (TB-HIV) co-infection in the pathogenesis of hypertensive disorders of pregnancy. Curr Hypertens Rep 2020; 22:104. [DOI] [PubMed] [Google Scholar]
- 30. Tesfaye F, Sturegård E, Walles J, et al. Dynamics of Mycobacterium tuberculosis-specific and nonspecific immune responses in women with tuberculosis infection during pregnancy. Microbiol Spectr 2022; 10:e0117822. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Saha A, Escuduero J, Layouni T, et al. Mycobacterium tuberculosis-specific T-cell responses are impaired during late pregnancy with elevated biomarkers of tuberculosis risk postpartum. J Infect Dis 2022; 225:1663–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Shafiq M, Mathad JS, Naik S, et al. Association of maternal inflammation during pregnancy with birth outcomes and infant growth among women with or without HIV in India. JAMA Netw Open 2021; 4:e2140584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Walles J, Otero LG, Tesfaye F, et al. Tuberculosis infection and stillbirth in Ethiopia—a prospective cohort study. PLoS One 2022; 17:e0261972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Gupta A, Montepiedra G, Aaron L, et al. Isoniazid preventive therapy in HIV-infected pregnant and postpartum women. N Engl J Med 2019; 381:1333–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Salazar-Austin N, Team the TS, Cohn S, et al. Isoniazid preventive therapy and pregnancy outcomes in women living with human immunodeficiency virus in the Tshepiso Cohort. Clin Infect Dis 2020; 71:1419–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.