Università degli Studi di Milano, San Paolo Hospital, Milan, Italy; Homerton University Hospital, London E9 6SR, United KingdomUnited Kingdom
Homerton University Hospital, London E9 6SR, United KingdomUnited Kingdom
Also available as PDF (figures and tables appear in the PDF).
Tuberculosis (TB) affects women, especially in the child-bearing years. TB is associated with a poorer outcome of pregnancy, although this may be due to the general risk factors for TB, namely poverty, malnutrition and overcrowding. New studies have shown that symp-tom screening has a low sensitivity and specificity, but is improved by the addition of a tuber-culin skin test (TST) or interferon-gamma release assay (IGRA) or, in high incidence areas, DNA amplification tests (e.g. Xpert MTB/RIF). TB-HIV co-infection is a common cause of mortality and morbidity in pregnancy. The diagnostic process remains the same in pregnancy, but non-specific symptoms and extra pulmonary disease demand a higher level of suspicion of TB. Standard first-line treatment is safe in pregnancy. Data on second-line drugs in pregnancy is still limited, but injectable drugs may affect the hearing and balance of the fetus. The IGRA responses appear to change during pregnancy, with more positive responses after delivery. The increasing incidence of drug-resistant TB, especially in Eastern Europe and Central Asia, requires an evaluation of the safety of second-line drugs in pregnancy.
Keywords: Tuberclosis; Pregnancy; Mycobacterium tuberculosis; Maternal mortality
Tuberculosis (TB) remains an important disease, despite effective treatment for the last 50 years. The number of cases remains high at 8.6 million new cases and 1.3 million TB deaths (WHO, 2012).1 Although TB affects men more commonly than women, there were still 2.9 million cases and 410,000 deaths among women, predominantly in the 15-44 year age group, which coincides with the age of childbearing.1 In 2011, there were an estimated 216,500 (95% uncertainty range 192,100 to 247,000) active tuberculosis cases in pregnant women.2 Pregnancy itself appears to be a risk factor for developing TB.3 The increased susceptibility to TB may be due to immunological changes in pregnancy. Pregnancy partially suppresses the T-helper 1 (Th1) cell mediated immunity, in favour of the antibody response (Th2 mediated),
with a dominant Th2 immune response.
our knowledge about TB in pregnancy, with particular reference to studies since January 2012.
The biomedical databases MEDLINE, through the search engine PubMed, and EMBASE (Elsevier) were searched. The time limit was May 2014 - Jan 2012. The Medical Subject Heading (MeSH) terms ‘‘tuberculosis”, “pregnancy”, “maternal mortality” and “women’s health” were combined as a major topic in PubMed MEDLINE. A common search strategy was used for all databases, employing a combination of the following terms: “pregnancy”, “maternal”, “tuberculosis”, “congenital”, “latent tuberculosis infection”, “multidrug resistant tuberculosis”. The research was limited to English-language articles and human studies. Case reports and conference abstracts were excluded. Titles and abstracts were reviewed and irrelevant topics were excluded. To complete the search a manual search was made of bibliographic references cited in the original papers included. These articles were then contextualised within current knowledge of TB in pregnancy.
TB disease in pregnancy is associated with adverse pregnancy outcomes.4 Toxaemia (pre-eclampsia), vaginal bleeding, fetal death at 16-28 weeks, acute fetal distress, prematurity (<37 weeks), small for date, low birth weight (< 2.5 kg), and perinatal death have all been described, but are also associated with poverty, malnutrition and overcrowding - factors them- 3,4 selves associated with TB.3,4 Adverse perinatal outcomes have been associated with incomplete treatment, delayed diagnosis and advanced pulmonary involvement.4 A new case-control study conducted in UK showed a lower birth weight in infants born to mothers with TB, especially if pulmonary disease, despite receiving standard treatment.5 Maternal TB is a risk factor for child TB. Congenital TB is rare,6 but the risk of transmission to the infant in the postpartum period is higher due to inhalation of aerial droplets coughed out by the mother. One study in South Africa detected TB in 16% of neonates born to mothers with suspected or proven TB.7 In Kenya, infants with early TB infection (T-SPOT.TB positivity in 6 month old children born from HIV positive mothers) had a 15.5-fold increased odds of having a mother with active TB (P = 0.04).8
An association between acute ectopic gestation and genital tuberculosis has been suggested. In a group of 17 adolescents with acute presentation of ectopic pregnancy, 6 out of 17 (35, 29%) had genital tuberculosis compared with 5% in the control groups (1 out of 20, p=0.03).9
A four symptom screening (cough, fever, night sweat
and weight loss) has been proposed by the WHO as a first step in finding TB in people living with HIV (PLHIV).10A meta-analysis estimated a sensitivity of 79% and specificity of 50% for these symptoms.11 However, when tested in 799 pregnant women from India, the sensitivity was 54.5%, and only reached 100% if combined with a tuberculin skin test (TST).12 A much larger study in South Africa (1415 pregnant women) demonstrated that the WHO four-symptom screen failed to identify most cases of TB in HIV-positive pregnant women13 as most (73%) were asymptomatic. The sensitivity of having any one of the four symptoms for TB disease was 28%, giving a specificity of 84%, positive predictive value 4.4% and a negative predictive value of 98%. Cough, irrespective of duration, was the most sensitive of the symptoms (23%). In Kenya, screening for the four symptoms in HIV-positive pregnant women was not helpful: 3/26 (12%) symptomatic patients had an abnormal chest X-ray (CXR) consistent with TB, while 7/100 asymptomatic pregnant women who had a CXR were then treated for TB.14 The same group screened 2980 HIV-positive and HIV-negative pregnant women, 17 of whom were treated for TB on the basis of clinical judgment and/ or CXR positivity but just 4 had a one of the following features: a cough of >2 weeks, bloody cough in the past year, fever of >3 weeks, past history of TB diagnosis, history of TB contact in the household, weight loss/failure to gain weight if pregnant in the past year.15
The poor performance of symptom screening suggests that additional tests are needed to intensify the effective case finding in HIV pregnant women.
Pregnancy does not appear to alter the performance 5 16 of the TST.16 Tests based on the interferon-γ response to proteins that are not found in Mycobacterium tuberculosis-Bacille Calmette-Guérin (BCG), were designed to improve the specificity of the diagnosis of TB infection.17A comparison of TST and QuantiFERON Gold-in-Tube (QFT) in 199 pregnant women at a US public hospital showed a greater specificity for the IGRA, in that BCG vaccination at birth was an independent predictor of a positive TST but not a positive IGRA.18 Pregnancy itself appeared not to affect the likelihood of a positive QFT, comparing 140 pregnant and 140 non-pregnant adolescents and young women, and exposure to TB correlated better with the IGRA than with TST.19 However, there were twice as many indeterminate results due to a low mitogen response in the pregnant compared with the non-pregnant cohort. By contrast, in India, HIV-negative pregnant women, presenting to the antenatal clinic in their late second or third trimester, showed more positive QFT tests than positive TSTs (37% and 14% respectively), even though 5 TU and a cut-off of 10 mm for TST positivity was used as in the US studies.20
A cross-sectional study suggested that a positive QFT
was more likely in the postpartum period, but attempts to perform a cohort study on 60 women were thwarted by the failure of follow-up, such that the initial finding could not be confirmed.20
The estimated percentage of TB cases living with HIV remains at 13% globally and the African Region accounts for 75% of the estimated number of HIV-positive incident TB cases. Pregnant women with HIV infection have a higher risk of developing active TB than HIV-negative pregnant women. TB increases maternal mortality in HIV co-infection. An analysis of pregnancy-related mortality in western Kenya found that nearly two-thirds of deaths were due to indirect, non-obstetric causes and HIV/AIDS and TB accounted for 45% and 10% respectively of these deaths.23 In a rural sub-district in South Africa from 1992 to 2010, the obstetric mortality ratio averaged 185 per 100,000 live births and deaths during pregnancy were 423 per 100,000 live births, the difference being primarily due to HIV/AIDS and pulmonary TB. Furthermore, mothers with HIV co-infection have a higher rate of infecting their children (30%) compared to those with TB alone (12%).
One group in Kenya has examined the value of the IGRA T-SPOT. TB using cryopreserved peripheral blood mononuclear in a cohort of 327 HIV-positive pregnant women in 26,27,28 Kenya. In the first report, 36% of the HIV-infected women were IGRA-positive during pregnancy, which likely reflects the local rate of LTBI. In multivariate analysis, adjusting for baseline CD4 count, IGRA positivity was associated with a 4.5 (1.1-18.0)-fold increased risk of active TB (p=0.03). Maternal and infant mortality was higher in mothers with a positive IGRA but the difference was not statistically different unless the maternal CD4 count was < 250 cells/mm . Eighteen had a positive test and the number of spot counts was highest at 32 weeks, then fell around delivery, followed by a later increase. Nine (3%) women developed TB within 1 year of childbirth, 6/110 (6%) with a positive test and 3/148 (2%) with a negative test at 32 weeks of pregnancy (no significant difference; indeterminate results were found in 52, of whom 8 had <20 spot counts in the positive control). The prognostic sensitivity and specificity improved if the CD4 count was <250 mm3, although a positive T-SPOT.TB test was more common in those with a higher CD4 count.
A cost-effectiveness analysis of IGRA in HIV-positive pregnant women in low TB incidence countries found that testing with TB-SPOT.TB alone was the most cost-effective strategy where the incidence of TB was ≥1.25%, but if the incidence of TB was less than 1,250 per 100,000, screening with TST and then testing with the T-SPOT.TB test was better.
treatment therapy for all HIV infected individuals, including pregnant women.10 The US Centre for Disease Control and Prevention (CDC) recommend screening for LTBI only in high risk women, i.e. those with known or suspected tuberculosis contact, injection drug use, HIV or other immune suppression, foreign birth and residence in communal settings.30 Antenatal care is an ideal opportunity to identify pregnant women with LTBI, as it is often the first time healthcare has been accessed, especially in these high risk groups that might find accessing healthcare 3,31 difficult. In high-burden countries, the social conditions of the 21,22 woman often limit access to healthcare and antenatal care should 3,4 32 be integrated with TB screening and treatment.
Isoniazid is safe in pregnancy and is not teratogenic.33 Unlike rifampicin, there are no interactions between isoniazid and antiretroviral treatment.34A WHO guideline has recommended 6 24 months preventive treatment on the basis of studies available to them.10 Clinical trials have suggested an increased benefit by in- 25 creasing isoniazid preventive treatment to 36 months duration in 35,36,37 PLHIV, although none reached statistical significance. The safety of long-term isoniazid prophylaxis was examined in HIVinfected women who were pregnant during the course of therapy and no adverse pregnancy outcomes were observed compared to the control group.38
There is no difference in the diagnostic approach between pregnant and non-pregnant women. Sputum examination and a CXR are the most important investigations. Concern about radiation safety in pregnancy has limited the use of chest ra- 3 26 diography, but shielding the abdomen and the lower doses of radiation that are now required to obtain a CXR mean that the 27 exposure to the fetus is considered negligible. However, the symptoms of TB are non-specific and are commonly present during normal pregnancy e.g. general malaise, fatigue, appetite loss; extrapulmonary disease, which is more difficult to detect, 3,4 is also more common in pregnancy.
3 ® Xpert® MTB/RIF, a PCR test on sputum that detects the presence of M. tuberculosis DNA and genetic mutations that indicate resistance to rifampicin, has been proposed as an instrument for active pulmonary case finding40 and in antenatal clinics in a high TB burden setting.41 In Zambia, sputum samples from 94 patients admitted with a primary obstetric or gynaecological problem (67% pregnant or < 6 weeks post natal patients and 74% HIV infected patients) were analysed by sputum smear ® microscopy, culture and Xpert® MTB/RIF assay. Among the 29 participants, 26 had culture-confirmed TB (77% in pregnant or ® postpartum women). In these circumstances, Xpert® had a sensitivity of 81% and a specificity of 97% compared to sputum culture and was more sensitive than sputum smear microscopy
The standard first line treatment for pulmonary and extrapulmonary TB does not differ between pregnant and nonpregnant women. The WHO recommend 2 months of isoniazid, rifampicin, pyrazinamide, and ethambutol, followed by 4 months of isoniazid and rifampicin. This regimen is safe to use during pregnancy. Anti-TB therapy should not be a reason to discontinue breastfeeding.33 Pyridoxine supplementation is recommended for all pregnant or breastfeeding women taking isoniazid. Streptomycin is contraindicated during pregnancy because of damage to the eighth cranial nerve with ototoxicity.
Multidrug-resistant TB (MDRTB) is defined as TB caused by organisms resistant to isoniazid and rifampicin; extensively drug-resistant TB (XDRTB) is defined as MDRTB resistant as well to any one of the fluoroquinolones and to at least one of three injectable second-line drugs.44 Globally, in 2012, an estimated 450,000 people developed MDRTB resulting in 170,000 deaths. The highest levels of MDRTB are found in Eastern Europe and Central Asia. In the treatment of MDRTB, WHO advises giving at least four drugs that are known or likely to be effective against the drug-resistant M. tuberculosis strain isolated, plus pyrazinamide. If possible, WHO group 2 (amikacin, capreomycin or kanamycin) and WHO group 3 (fluoroquinolones) should be core drugs with the rest (etionamide/protionamide and cycloserine) being accompanying drugs.44 New drugs, bedaquiline, an oral diarylquinoline which inhibits the proton pump ATP synthase, and delamanid, which is a nitro-dihydroimidazooxazole derivative with mycobacteria-specific antibacterial activity in vitro that inhibit mycolic acid biosynthesis, have been approved for use in the USA since December 2012 and EU since April 2014 respectively, but their use in pregnancy needs 45,46 evaluation.
Although second-line drugs are used during pregnancy, little is known about the safety of these drugs for the fetus and about the outcome in MDRTB cases during pregnancy. The largest study to our knowledge is of 38 pregnant women treated for MDRTB in Peru and outcomes were similar to those of the general local population. The majority of the second-line drugs are in FDA class C (animal studies suggest a problem, but human studies are inadequate), except for aminoglycosides, which are in class D (definite evidence of fetal risk). A TBNET consensus statement on the management of patients with M/XDRTB in Europe states that “safe treatment of M/XDRTB during pregnancy seems possible but needs individual decision-making” and “Pregnancies should not be terminated because of M/XDR- TB”;“Aminoglycosides/polypeptides are not recommended for M/XDRTB treatment during pregnancy”; “Patients should be advised to maintain double barrier contraception during treat-
There is a need for cohort studies of IGRA in pregnancy in women without HIV to assess the consistency of these tests and any changes that might occur as pregnancy proceeds. The increasing incidence of drug-resistant TB, especially in Eastern 43 Europe and Central Asia, requires an evaluation of the safety of second-line drugs in pregnancy.
33 The authors acknowledge the contribution of Richard Turner (Homerton University Hospital, London) for comments on the manuscript.