Abstract: A number of factors can contribute to a delay in the diagnosis of tuberculosis in pregnant women, including the presence of nonspecific symptoms, such as fatigue and cough; extrapulmonary manifestations; and asymptomatic disease. The diagnostic evaluation is the same as for nonpregnant patients and includes tuberculin skin testing and, when indicated, chest radiography (with appropriate shielding) and acid-fast bacillus stain and culture. Antituberculous therapy during pregnancy is generally safe and effective, although streptomycin should not be used because of the risk of vestibular or auditory damage to the fetus. For patients with active tuberculosis, treatment should be initiated as soon as the diagnosis is established. The treatment of latent infection is somewhat more controversial. The timing of the initiation of therapy is based on the risk of progression to active disease. (J Respir Dis. 2006;27(8):338-347)
After years of consistent decline, the incidence of tuberculosis in the United States increased from the mid-1980s to a peak in 1992.From 1992 to 2002, the number of cases of tuberculosis and related mortality decreased,1 but the number of cases among foreign-born residents increased.2 Worldwide, tuberculosis remains a leading cause of death from infectious disease, and tuberculosis control remains an important and unresolved global health issue.3,4
In addition, the prevalence of tuberculosis is increasing in younger persons. From 1985 to 1992, the largest increase in incidence occurred in the 25- to 44-year-old age group, and the average age of the patient with tuberculosis continues to decline.2,5
For women, this equates to an increasing incidence of disease during childbearing years. Although the overall rate of tuberculosis in pregnant women is relatively low (0.1% to 1.9%),2 it has increased. In New York City, the incidence of tuberculosis in pregnant women rose from 12.4 per 100,000 deliveries between 1985 and 1990 to 94.8 per 100,000 deliveries between 1991 and 1992.6 This increase is a reflection of the general prevalence of disease; it does not represent an increased risk of tuberculosis resulting from pregnancy. Pregnancy, birth, puerperium, and lactation do not predispose women to tuberculosis or render treatment of tuberculosis ineffective.
In this article, I review the diagnostic challenges of patients who are pregnant and who may have latent Mycobacterium tuberculosis infection (LTBI) or active disease. I then discuss the treatment options, examining the efficacy and safety of first- and second-line therapies.
The impact of pregnancy on the occurrence and clinical course of tuberculosis has been debated since the time of Hippocrates. At that time, pregnancy was thought to have a beneficial effect on the course of tuberculosis. During the mid-19th century, however, pregnancy in women with tuberculosis was believed to have deleterious effects such that it was recommended that these women were not to marry; if married, they were not to conceive; if they conceived, the pregnancy was terminated; and if they delivered, they were not permitted to breast-feed. Currently, it appears that pregnancy has minimal, if any, influence on the pathogenesis of tuberculosis.
Tuberculosis refers to active disease that may be pulmonary or extrapulmonary. LTBI is an infection that is contained by a person's innate host defense mechanisms but carries a risk of progression to tuberculosis. Congenital tuberculosis is caused by the transfer of mycobacteria from the mother to the fetus via the placenta, either through the umbilical vein or by aspiration of contaminated amniotic fluid. Although congenital tuberculosis is rare, it is associated with a high mortality rate. Perinatal tuberculosis is caused by the transmission of airborne infection from the mother or other care provider to the infant in the neonatal environment.
A number of factors can contribute to a delay in the diagnosis and treatment of tuberculosis. Consideration of the possibility of tuberculosis may be the rate-limiting step in the diagnosis in any situation.
Because tuberculosis and pregnancy can have some similar symptoms, there may be a delay in seeking medical attention or in initiating diagnostic testing. The non- specific symptoms associated with tuberculosis, such as fatigue and cough, can mimic the physiologic changes expected in pregnancy. In fact, for a patient who is pregnant, a failure to gain weight may be the only clue that a diagnosis of tuberculosis should be considered. A significant percentage of pregnant women with active tuberculosis may be asymptomatic.2
In addition, studies have suggested that there is a higher proportion of extrapulmonary tuberculosis in women in some populations, unrelated to HIV status.6-8 This further complicates the diagnosis,since patients presenting with extrapulmonary tuberculosis are less likely to have chronic cough than those with pulmonary involvement.
The evaluation of tuberculosis in women who are pregnant is no different than that in any other patient. Women with known or suspected exposure to tuberculosis, those with symptoms suggestive of tuberculosis, and those in a high-risk group should be evaluated. These women should have a tuberculin skin test (TST) with a 0.1-mL intradermal injection of 5-tuberculin-unit-strength purified protein derivative.
A chest radiograph, with appropriate shielding, is indicated if the TST results are positive or if the patient has had recent contact with an active tuberculosis case regardless of the TST result. An acid-fast bacillus stain and culture of material from the area of suspected infection should be obtained.9,10 For suspected pulmonary disease, sputum is generally the first specimen to be analyzed. Depending on the results and clinical circumstances, other pulmonary or extrapulmonary specimens may be collected. TST reactivity is not affected by pregnancy5,11; interpretation of the result is based on risk factors for tuberculosis and is not altered based on pregnancy (Table 1).
A positive TST result in a targeted TST program indicates infection. Clinical assessment must then determine whether there is active disease, which requires a multidrug treatment regimen, or LTBI, which requires treatment with a single agent. A negative TST result does not rule out either active disease or LTBI because false-negative results are possible.
Adherence to tuberculosis therapy is a well-recognized challenge that is compounded by concerns--both among health care providers and patients--about possible adverse effects of any medication on the outcome of the pregnancy and on the developing fetus. Some symptoms, such as nausea, may be related to pregnancy or to the effect of a drug, and determining the specific cause may be difficult.
Major therapeutic concerns can be categorized according to maternal toxicity, teratogenicity, and maternal-fetal outcomes. In each patient, the relative risk-to-benefit ratio of the treatment of active tuberculosis or LTBI must be carefully considered.
Much has been written about the toxicity of antituberculous medications. These drugs are typically divided into first-line therapies, which are generally effective and relatively safe, and second-line agents, which are less effective and have increased potential for toxicity.
First-line medications include isoniazid, rifampin, ethambutol, pyrazinamide, and streptomycin. Of these, isoniazid, rifampin, ethambutol, and pyrazinamide are generally considered safe for use during pregnancy. Streptomycin should not be used because of a risk of vestibular or auditory damage.7 The second-line drugs have been used less extensively in patients who are pregnant (Table 2).
The incidence of preeclampsia, vaginal bleeding, and early fetal death is higher in pregnant women with tuberculosis than in pregnant women without tuberculosis. Pregnant women with tuberculosis are more likely to have infants who are premature, are small for date, have low birth weight, and have low Apgar scores. Their infants are also at increased risk for perinatal death. Statistically significant increases in neonatal morbidity and incidence of adverse obstetric outcomes have been demonstrated in pregnant women with tuberculosis compared with a matched control group of nontuberculous pregnant women.6 The risk of adverse events was increased in cases of pulmonary tuberculosis compared with extrapulmonary tuberculosis6 and in cases in which treatment was delayed.6,12
Treatment-related complications are less common than reported tuberculosis-associated complications (Table 3).7 Isoniazid-induced elevations in liver enzyme levels are common; estimates of elevation above normal range vary between 10% and 25%. These elevations are transient and generally considered harmless; they normalize even in the presence of ongoing therapy. Clinically significant isoniazid- induced hepatitis is relatively rare.
Isoniazid readily crosses the placental barrier and is not considered teratogenic. In a study by Wilson and colleagues,13 there was no excess of fetal malformations in children born to mothers treated with isoniazid during pregnancy. A review of the literature published before 1980 demonstrated a lower rate both of miscarriage and of perinatal death in pregnant women treated with isoniazid compared with the general population.14 The risk of isoniazid-induced liver disease is 2 orders of magnitude smaller than the effect of tuberculosis on pregnancy.7
Similarly, no excess of birth defects has been noted in infants of mothers taking rifampin. However, rifampin induces cytochrome P-450 microsomal hepatic enzymes, which affect the metabolism of many medications, including contraceptive hormones. This effect can result in unplanned pregnancies. Rifampin-induced hepatic abnormalities can also occur; the risk is additive (not synergistic) to that of isoniazid.
Pyrazinamide has been endorsed by many for use in pregnancy, despite the absence of animal studies to support a lack of teratogenicity. Its use in the early stages of therapy in cases of potentially drug-resistant tuberculosis makes it less likely that initial treatment will be ineffective. It also allows for a 6-month treatment regimen in the event that adherence to a 9-month course proves problematic. These potential benefits must be weighed against the lack of animal studies; however, there have been no reports linking pyrazinamide to fetal malformations.
There has also been a lack of animal studies to assess the teratogenicity of ethambutol. However, clinical reports suggest that there is no increase in maternal complications or fetal malformations with the use of this drug. Although retrobulbar neuritis has been reported in patients treated with high doses or prolonged dosing,7 no visual problems have been reported in children born to mothers treated with ethambutol during pregnancy.
Streptomycin should not be given during pregnancy because of an association with damage to the eighth cranial nerve. Up to 17% of all children born to mothers treated with streptomycin may experience problems with hearing or balance.7
Pyridoxine (vitamin B6) should be added to any regimen that contains isoniazid. Isoniazid is a competitive inhibitor of coenzymes pyridoxal phosphate and pyridoxamine. Pyridoxine is a cofactor in the production of amines, which act as neurotransmitters. Competitive inhibition leads to neurotoxic conditions, such as peripheral neuropathy and optic neuritis, and can be prevented with daily vitamin B6 supplementation (10 to 25 mg). Higher doses can be used to treat neuropathy should it develop.
The treatment of multidrug- resistant tuberculosis (MDRTB) is more complicated, even in the absence of pregnancy, and overall direction of care should always be referred to a physician with expertise in the management of MDRTB. Similarly, coinfection with HIV and tuberculosis renders treatment regimens more problematic because of the potential for both drug interactions and an overall increase in the frequency of treatment-related adverse effects.
Not all studies demonstrate a negative outcome in pregnancy with concomitant tuberculosis. Differences in reports may be related to such factors as the impact of socioeconomic status, country of origin, advanced maternal disease, late diagnosis, HIV coinfection, and irregular treatment. In general, outcome appears to be affected primarily by the timing of the diagnosis and treatment and by the site of infection, with pulmonary tuberculosis evincing a worse outcome than extrapulmonary tuberculosis does.
Certainly, treatment for active tuberculosis is essential and should be initiated as soon as the diagnosis is ascertained. Treatment should not be suspended if pregnancy occurs or is discovered after treatment has been initiated. Given the overall safety of first-line medications and the potential for good maternal-fetal outcomes with prompt and effective therapy, there is no basis for recommending therapeutic abortion. Treatment outcomes with respect to sputum conversion, clinical stabilization, and relapse do not appear to be different in women with tuberculosis who are pregnant than in those who are not.12
For active tuberculosis, treatment with first-line drugs is generally safe and improves maternal-neonatal outcomes. Treatment strategies currently recommend directly observed therapy (DOT) whenever possible.
Treatment duration in pregnancy should be 9 months if the regimen contains both isoniazid and rifampin and if there is no first-line drug resistance. A 6-month course can be administered if pyrazinamide is included for the first 2 months of therapy.
Ethambutol is routinely added to the initial drug regimen pending drug susceptibility test results, unless the documented rate of isoniazid resistance within the community in which the patient has been infected is lower than 4%.
Intermittent drug therapy should not be used unless the patient is in a DOT program. Successful outcomes with treatment of MDRTB have also been reported, but treatment regimens and duration of therapy must be individualized.
For all patients, the ideal treatment team should be multidisciplinary, with supervision guided by experienced chest or infectious disease physicians. If HIV coinfection is present, a physician with expertise in HIV should also be involved. Participation by the patient's primary care physician and obstetrician, as well as a pediatrician, is essential to optimize outcomes.
The treatment of LTBI is somewhat more controversial than is the treatment of active tuberculosis. Isoniazid therapy has been shown to be both effective--with a 68% (range, 22% to 90%) reduction in the development of active disease-- and safe, with no demonstrated teratogenic potential.15 Although an elevation in liver enzyme levels is common, clinically significant isoniazid-induced hepatitis has been reported in only 0.15% to 2% of patients, with a 0.001% risk of fatal hepatitis.
A decision analysis model of treatment of LTBI in pregnancy projected that in the absence of treatment, 3300 cases of active tuberculosis per 100,000 cases of LTBI would occur. This incidence could be decreased with postpartum or antepartum treatment to 1800 and 1400 cases, respectively.15
Another argument for treatment of LTBI during pregnancy is that women are already involved in ongoing prenatal care, and adherence to treatment may therefore be greater than during the postpartum period. Prenatal care allows for the increased monitoring of adherence to therapy as well as monitoring for toxicity.
However, a pervasive concern that appears in the literature suggests that pregnancy increases the risk of isoniazid-related hepatotoxicity, with a 2.5 times greater risk of isoniazid-induced hepatitis and a 4-fold risk of death. These frequently cited statistics are based on a single retrospective study, in which the difference between the study group and previously collected data on 3948 nonpregnant women was not statistically significant.16 There were 5 women in whom isoniazid-induced hepatitis developed and there were 2 deaths in the study cohort of 3681 patients. This mortality rate of 0.0005% actually represents a lower rate of isoniazid-associated fatal hepatitis than was reported previously.
Early studies of isoniazid-associated hepatotoxicity demonstrated a correlation with increasing age. Subsequently, treatment guidelines were adjusted to limit treatment in older patients unless there was an increased risk of reactivation. The change in guidelines was followed by a decrease in the risk of fatal hepatotoxicity from 92 cases of probable hepatitis with 8 fatalities in 13,838 persons treated17 to 11 of 11,141 and 15 of 1427 persons receiving isoniazid alone for LTBI or combination therapy for active tuberculosis, respectively (with 1 fatality in the latter group).18
Although it was not statistically significant, there was a trend in the later study toward an increased frequency of toxicity in women that was consistent with previous reports. Of those patients in the 2 groups who experienced toxicity, 8 of 11 and 9 of 15, respectively, were women.
Official recommendations for the treatment of LTBI during pregnancy generally favor delaying treatment until the postpartum period, despite the apparent overall safety and the potential benefit.7 However, if treatment was started before pregnancy, therapy can be continued. Recommendations vary as to whether to initiate treatment immediately9 or wait until after the first trimester7 in cases of recent contact, concurrent HIV disease, or other risk factors for progression of LTBI to active disease.
Given the lack of any demonstrated teratogenicity and the added burden of treating active tuberculosis, initiating treatment at the time of clinical assessment, rather than delaying, appears logical. Treatment of LTBI in women who are at low risk for progression can be delayed until after pregnancy; however, this probably serves merely as a concession to the general fear of taking any medication during pregnancy.
Monitoring for hepatotoxicity
Given the high frequency of clinically insignificant and transient elevations in liver enzyme levels, routine laboratory testing is not recommended. Instead, evaluation should include questions regarding side effects and a physical assessment to look for signs of hepatitis. Laboratory testing is generally reserved for patients exhibiting signs or symptoms associated with hepatotoxicity. However, baseline testing is recommended for women who are either pregnant or in the immediate postpartum period.
Routine laboratory monitoring during treatment should be done in patients who have abnormal baseline liver function test results or who are at risk for liver disease. Routine monitoring during pregnancy may be considered, given the potential similarity between symptoms associated with pregnancy and with hepatotoxicity. Discontinuation of therapy should be considered if elevation of hepatic enzyme levels exceeds 3 times the upper limit of normal in symptomatic patients or 5 times the upper limit of normal in asymptomatic patients.
Groups including the World Health Organization, United Nations Children's Fund, and the American Academy of Pediatrics have advocated universal breast-feeding. Given the well-established maternal and infant benefits of breast-feeding, the recommendation to limit breast-feeding in patients with tuberculosis is related to the potential transmission of infection via respiratory contact from an infected mother with pulmonary disease to her infant--not to the method of feeding.
The safety of breast-feeding is not an issue in women who have a positive TST result and no active disease. Women with active tuberculosis who are still potentially infectious should not be in direct contact with their infants.
Since breast milk does not contain tubercle bacilli, women may pump their milk to be fed to their infants pending initiation of therapy. Once the patient is noninfectious, she can breast-feed her infant directly.19
Breast-feeding during the course of antituberculous therapy is not contraindicated. Toxic effects of antituberculous medications delivered in breast milk have not been reported. Breast-fed infants have serum levels that measure less than 20% of the usual therapeutic levels for isoniazid and less than 11% for other antituberculous medications.20 Pyridoxine should be given to the mother as well as to the infant during the course of an isoniazid-containing regimen.
The exception to all of these scenarios is the case of a mother with an active tuberculous breast lesion. In this situation, breast milk can be pumped and discarded until the lesion is fully healed, after which breast-feeding can resume.
The care of a neonate born to a mother with active infectious tuberculosis must be carefully evaluated. In general, in the absence of drug resistance, neonates are given isoniazid prophylaxis (5 mg/kg). They should be reassessed between the ages of 6 weeks and 3 months.
For neonates with a positive TST result, a thorough investigation must be undertaken to rule out active disease. If it is present, full multidrug therapy should be initiated. If there is no evidence of active disease, however, treatment for LTBI should be continued for a full course. If the neonate has a negative TST result and a BCG vaccine has not been administered, it should be considered and chemoprophylaxis should be stopped. The amount of drug present in breast milk is inadequate for treatment purposes.
Women should be evaluated for risk of tuberculosis early in their prenatal care and assessed accordingly. TST and chest radiography with appropriate shielding should be performed. Treatment of active tuberculosis during pregnancy is considered safe and effective, and it should be initiated as soon as the diagnosis is ascertained.
Treatment of LTBI during pregnancy is also considered to be generally safe and effective, but it is more controversial. The timing of the initiation of therapy for LTBI is based on the risk of progression of disease; therapy should be initiated immediately in any patient at increased risk. The only contra- indication to breast-feeding is active tuberculous infection of the breast and the need for respiratory precautions.
In the more problematic situation of MDRTB, with the need to use potentially less effective and more toxic medications, there are fewer guidelines available to clinicians. Reported options include termination of the pregnancy, withholding all treatment pending delivery, "holding" therapy using some medications pending delivery, or treatment with informed discussion.
Limited data demonstrating the safety of many of the second-line drugs, as opposed to evidence of toxicity, may lead to a tendency toward undertreatment. However, the need for aggressive treatment is more urgent for patients with MDRTB than it is for those with nonresistant tuberculosis.
Management must be individualized, considering the relative risk versus benefit for both mother and fetus, and the patient must be involved in decision making. Although a limited number of cases have been reported in the literature, effective treatment with good maternal-neonatal outcomes is possible.21
1. Arias E, Smith BL. Deaths: preliminary data for 2001.
Natl Vital Stat Rep.
2. Vo QT, Stettler W, Crowley K. Pulmonary tuberculosis in pregnancy.
Prim Care Update Ob Gyns
3. Thillagavathie P. Current issues in maternal and perinatal tuberculosis: impact of the HIV-1 epidemic.
4. Small PM, Fujiwara PI. Management of tuberculosis in the United States.
N Engl J Med
5. Vallejo JG, Starke JR. Tuberculosis and pregnancy.
Clin Chest Med
6. Figueroa-Damian R, Arredondo-Garcia JL. Neonatal outcome of children born to women with tuberculosis.
Arch Med Res
7. Bothamley G. Drug treatment for tuberculosis during pregnancy: safety considerations.
8. Llewelyn M, Cropley I, Wilkinson RJ, Davidson RN. Tuberculosis diagnosed during pregnancy: a prospective study from London.
9. ATS Official Statement. Targeted tuberculin testing and treatment of latent tuberculosis infection.
Am J Respir Crit Care Med
. 2000;161: S221-S247.
10. ATS Official Statement. Diagnostic standards and classification of tuberculosis in adults and children.
Am J Respir Crit Care Med
11. Eriksen NL, Helfgott AW. Cutaneous anergy in pregnant and nonpregnant women with human immunodeficiency virus.
Infect Dis Obstet Gynecol
12. Tripathy SN, Tripathy SN. Tuberculosis and pregnancy.
Int J Gynaecol Obstet
. 2003;80: 247-253.
13. Wilson EA, Thelin TJ, Dilts PV Jr. Tuberculosis complicated by pregnancy.
Am J Obstet Gynecol
14. Snider DE Jr, Layde PM, Johnson MW, Lyle MA. Treatment of tuberculosis during pregnancy.
Am Rev Respir Dis
. 1980;122: 65-79.
15. Boggess KA, Myers ER, Hamilton CD. Antepartum or postpartum isoniazid treatment of latent tuberculosis infection.
16. Franks AL, Binkin NJ, Snider DE Jr, et al. Isoniazid hepatitis among pregnant and postpartum Hispanic patients.
Public Health Rep
17. Kopanoff DE, Snider DE Jr, Caras GJ. Isoniazid-related hepatitis: a U.S. Public Health Service cooperative surveillance study.
Am Rev Respir Dis
18. Nolan CM, Goldberg SV, Buskin SE. Hepatotoxicity associated with isoniazid preventive therapy: a 7-year survey from a public health tuberculosis clinic.
. 1999;281: 1014-1018.
19. Lawrence RM, Lawrence RA. Given the benefits of breastfeeding, what contraindications exist?
Pediatr Clin North Am
. 2001;48: 235-251.
20. Ormerod P. Tuberculosis in pregnancy and the puerperium.
21. Shin S, Guerra D, Rich M, et al. Treatment of multidrug-resistant tuberculosis during pregnancy: a report of 7 cases.
Clin Infect Dis