Surveillance in this context refers to the systematic observation and data collection to monitor the incidence and outcomes of TB disease and exposure to medications for TB prevention and treatment in pregnant, postpartum and breastfeeding women. This enables the early detection of any signals/adverse events associated with TB or the medicines used for TB prevention and treatment in research and programmatic settings. 

Maternal TB surveillance comprises two elements (Figure 7.1).

1. outine surveillance of the burden and patterns of TB disease, including in pregnant and postpartum women. 
2. Pharmacovigilance – the detection, assessment, understanding and prevention of adverse events related to medicine or vaccine exposure, in this case given for TB prevention or treatment during pregnancy and the postpartum period.

To optimize data collection and comparability, harmonized/standardized definitions and data collection tools around a core set of indicators and variables are used to conserve resources and allow comparison within programmes and research settings as well as between regions and across conditions with overlapping epidemiology (e.g. TB and HIV). 

Routine surveillance

Routine surveillance of TB among pregnant and postpartum women and their neonates may be incorporated into existing national TB programmes and maternal and child health programme surveillance systems and reporting which are well-established (although often separate) in most national health care services, especially in countries who have prioritized the collection of these data and in countries which operate case-based digital surveillance systems. Within programmes, such data can support prioritizing programmatic interventions as well as generating research questions. 

If resources allow, core¹ and expanded² variables related to obstetrics and breastfeeding could be integrated into national tuberculosis programme (NTP) data collection, while core or expanded TB variables could also be included in maternal and child health data collection and reporting (Table 7.1). The use of core and expanded variables to monitor disease, treatment outcomes and adverse effects has been successfully used for maternal HIV in high-burden settings (1); to monitor the burden of HIV in pregnant women; assess the implementation and impact of guidelines for the prevention of vertical HIV transmission and maternal HIV treatment; and monitor adverse effects of antiretrovirals used for HIV prevention or treatment.   

Given that postpartum care is often provided across different facility types (as opposed to antenatal care which is provided in specific facilities), it may be easier to collect data on TB exposures in the postpartum period from NTP programme data. Alternatively, a maternal surveillance system could be integrated with the Expanded Programme of Immunization or well-child services (allowing for collection of relevant data during the postpartum period), although this has been of limited success in the context of HIV. 

Where case-based surveillance systems are used, electronic linkage of NTP and maternal and child health programme data via a unique identifier would obviate the need for additional data collection if the relevant variables are included in respective surveillance systems. In this case TB and obstetric variables can be included in maternal and child health programme and NTP digital surveillance systems, respectively, at the development stage. Collection of such data in all women provides an appropriate denominator to determine the incidence of TB in pregnant and breastfeeding women. In addition to routinely collected data and data linkage, other data-generation approaches may include data collection from a defined sample of medical records and collection of detailed data from sentinel sites.

Pharmacovigilance

Owing to the limited inclusion of pregnant and breastfeeding women in pre-approval clinical trials for medicines and vaccines, the benefits and safety profile of medicines used in pregnancy are usually established through post-licensure surveillance. Post-licensure surveillance is necessary for the evaluation of rare pregnancy outcomes such as congenital disorders or stillbirths, which require larger numbers of observations to determine potential signals compared to data obtained in clinical trials.

Pregnant and breastfeeding women may be exposed to anti-TB medicines in two ways: 

1. A woman taking TB treatment or TPT becomes pregnant (i.e. preconception exposure). TB disease is common in women of childbearing age and TPT is recommended for people living with HIV, household contacts and other at-risk populations. Thus, there is potential for exposures to TB medicines during pregnancy. Because pregnancy is often not recognized until 6–8 weeks after fertilization, TB medicine exposure commonly occurs during early embryogenesis, including organogenesis, when the risk of teratogenic effects is greatest. 
2. A pregnant or breastfeeding woman requires TB treatment or TPT owing to her risk (e.g. household contact, living with HIV, or other risk factor).

The fetus/infant can be affected by anti-TB medicines:

1. Directly during pregnancy: due to potential teratogenic properties of a medication or its disruptive effect on normal fetal growth and development.
2. Indirectly during pregnancy: due to an adverse event experienced by the mother which impacts her health during pregnancy, with potential adverse implications for the fetus. Additionally, suboptimal treatment of TB during pregnancy will expose the fetus to the risks associated with maternal TB (miscarriage, stillbirth, prematurity, low birth weight, asphyxia, perinatal TB (2, 3)).
3. Directly through breastfeeding: Breastfed infants may be exposed to anti-TB medicines excreted in breastmilk with potential toxic effects, especially if the infant is on TB treatment or receiving TPT. Data on the excretion of anti-TB medicines in breastmilk are extremely limited (4). 

Passive surveillance systems monitor the effects of medicine exposures during pregnancy and breastfeeding (e.g. spontaneous reporting of adverse events to national pharmacovigilance programmes or drug-specific registries). This includes clinicians reporting adverse events to a national register or pharmacovigilance system after it occurs. Passive surveillance systems are often limited by reporting bias, duplication, and lack of a denominator for the total exposed population or an unexposed control group. While national pharmacovigilance reporting systems are common, no drug-specific registry exists for passive surveillance of TB medicines. 

Active surveillance approaches include pregnancy-exposure registries, enhanced birth-outcome surveillance, cohort event-monitoring studies, nested case-control studies and data linkage studies. Common to all these approaches is linkage of maternal exposures to certain medicines with pregnancy and infant outcomes. A critical component of such surveillance is the use of standardized data definitions to allow harmonization/linkage between different sources of surveillance data and to facilitate rapid investigation of potential safety signals (1–5). Therefore, active surveillance is preferable to evaluate the potential effects of medicines used during pregnancy on the woman, fetus, neonate and infant.

If an active pharmacovigilance surveillance system is implemented, given resource limitations, birth outcomes with the greatest public health impact should be recorded: namely, preterm birth, low birth weight, small-for-gestational age, stillbirth, neonatal deaths prior to hospital discharge and up to 28 days after birth, and congenital disorders assessed by routine external surface examination at birth. In addition, key maternal variables include death and other serious adverse events (e.g. drug-induced liver injury) (Table 7.2).

High-quality data collection on TB-related exposures (i.e. to TB, medicines for the prevention and treatment of TB, and, in future, TB vaccines) and pregnancy outcomes at all delivery sites is resource intensive and is generally not feasible in high TB burden settings. Therefore, as successfully used for HIV (6–8), enhanced birth outcome surveillance may be implemented at selected high-volume sentinel sites to provide information on the impact of both TB and TB medicines and vaccines on maternal health, pregnancy, and neonatal outcomes. Sentinel-site data may also be used to support assumptions for modelling studies.

Core pregnancy and neonatal/infant variables with standardized definitions (and when possible, expanded and comprehensive variables) should be collected from all women with a pregnancy outcome at sentinel sites regardless of TB exposure status to allow determination of a representative denominator and appropriate comparison groups, enabling the calculation of risk, and stratification by characteristics of interest. A standardized set of variables categorized as birth, neonatal/infant and maternal outcomes, is outlined in Table 7.2. This approach allows data collection on multiple exposures at a single sentinel site.

Successful applications of the enhanced birth outcome surveillance in the context of HIV include the Tsepamo (Botswana (8)) and Eswatini (6) studies in which data on pregnancy exposures and birth outcomes were collected at a single timepoint (at delivery); the Western Cape Pregnancy Exposure Registry (7) and the Ubomi Buhle study in South Africa, in which all women attending antenatal sentinel site facilities were enrolled and followed up until pregnancy outcome. The Western Cape Pregnancy Exposure Registry is incorporated into the Provincial Health Data Centre (9), a linked health information exchange which integrates electronic health data from multiple sources. At the individual level, maternal exposures and birth/infant outcomes can be linked; at the population level, NTP and maternal and child health programme data can be merged.

This toolkit includes TB and obstetric variables that can be incorporated into routine surveillance data if prioritized by a country (and which is more feasible in countries with case-based digital surveillance systems), as well as standardized definitions of maternal, birth and infant outcome variables applicable to pharmacovigilance systems and enhanced birth outcome surveillance.

Key outcome measures and predictors

The following list of pregnancy and infant outcomes is suggested to optimize and harmonize pregnancy safety data across settings. Standardized outcomes would support both public health surveillance and research. These are aligned with the work of the WHO inter-departmental task team to improve maternal and newborn health in low- and middle- income countries. These have been defined in the WHO Antiretrovirals (ARV) in pregnancy research toolkit.

Outcomes were grouped into “tiers” (core; expanded; comprehensive), guided by two main factors: 

• the relative clinical/public health importance of the outcome; and
• the feasibility and resource intensity of collecting data of reasonable quality on the outcome.

The goal of the tier groupings is to help guide the minimum pregnancy data that should ideally be collected, acknowledging that it will not be feasible to collect some of these pregnancy-related outcomes outside of clinical trial settings.

For all outcome and exposure/predictor data:

• basic information on the data source(s) should also be collected (e.g. electronic versus paper, routine clinical sources versus from participants, retrospectively versus prospectively, data based on ultrasounds or other data, etc.). 
• actual numeric dates and outcome values should be collected rather than just categorial data (e.g. date and gestational age at delivery rather than simply term/preterm).

Outcome collection and ascertainment are expected to vary between studies, because of data collection approach, study design, feasibility, resources, and the goals of the projects. 

Table 7.4. Basic outcome definitions 

For additional detail, please refer to WHO harmonized definitions and data elements.  

References

1. Renaud F, Mofenson LM, Bakker C, Dolk H, Leroy V, Namiba A et al. Surveillance of ARV safety in pregnancy and breastfeeding: towards a new framework. J Int AIDS Soc. 2022;25 Suppl 2(Suppl 2):e25922. doi:10.1002/jia2.25922

2. Salazar-Austin N, Hoffmann J, Cohn S, Mashabela F, Waja Z, Lala S et al. Poor obstetric and infant outcomes in human immunodeficiency virus-infected pregnant women with Tuberculosis in South Africa: The Tshepiso study. Clin Infect Dis. 2018;66(6):921–9. doi:10.1093/cid/cix851

3. Sobhy S, Babiker ZOE, Zamora J, Khan KS, Kunst H. Maternal and perinatal mortality and morbidity associated with tuberculosis during pregnancy and the postpartum period: a systematic review and meta-analysis. BJOG. 2017;124(5):727–33. doi:10.1111/1471-0528.14408

4. Algharably EA, Kreutz R, Gundert-Remy U. Infant exposure to antituberculosis drugs via breast milk and assessment of potential adverse effects in breastfed infants: Critical review of data. Pharmaceutics. 2023;15(4)1228. doi:10.3390/pharmaceutics15041228

5. Mofenson LM, Pozniak AL, Wambui J, Raizes E, Ciaranello A, Clayden P et al. Optimizing responses to drug safety signals in pregnancy: the example of dolutegravir and neural tube defects. J Int AIDS Soc. 2019;22(7):e25352. doi:10.1002/jia2.25352

6. Gill MM, Khumalo P, Chouraya C, Kunene M, Dlamini F, Hoffman HJ et al. Strengthening the evidence: Similar rates of neural tube defects among deliveries regardless of maternal HIV status and dolutegravir exposure in hospital birth surveillance in Eswatini. Open Forum Infect Dis. 2023;10(9):ofad441. doi:10.1093/ofid/ofad441

7. Kalk E, Heekes A, Slogrove AL, Phelanyane F, Davies MA, Myer L et al. Cohort profile: the Western Cape Pregnancy Exposure Registry (WCPER). BMJ Open. 2022;12(6):e060205. doi:10.1136/bmjopen-2021-060205

8. Zash R, Makhema J, Shapiro RL. Neural-tube defects with dolutegravir treatment from the time of conception. N Engl J Med. 2018;379(10):979-81. doi:10.1056/NEJMc1807653

9. Boulle A, Heekes A, Tiffin N, Smith M, Mutemaringa T, Zinyakatira N et al. Data centre profile: The provincial health data centre of the Western Cape Province, South Africa. Int J Popul Data Sci. 2019;4(2):1143. doi:10.23889/ijpds.v4i2.1143

¹ Core variables are the minimum datapoints that can be collected within the respective programmes (i.e. NTP and maternal and child health programmes).

² Expanded variables provide more detailed data for in-depth analyses and programme improvement. These may include additional demographic, clinical, or contextual variables that help identify gaps, assess quality of care, and guide targeted interventions.