INTRODUCTION — Bacille Calmette-Guérin (BCG) is a live strain of Mycobacterium bovis developed by Calmette and Guérin for use as an attenuated vaccine to prevent tuberculosis (TB) and other mycobacterial infections. The vaccine was first administered to humans in 1921 and remains the only vaccine against TB in general use.
BCG vaccine is the most widely administered vaccine in the world; it has been given to over three billion individuals, principally in the setting of routine newborn immunization (as dictated by guidelines of the World Health Organization) [1]. There are multiple BCG vaccines in use around the world, produced by different manufacturers and administered via different schedules.
BCG vaccine is also effective for protection against other mycobacterial infections including leprosy and Buruli ulcer. In addition, it is used as an immunostimulant in the treatment of superficial carcinoma of the bladder. These issues are discussed separately. (See "Leprosy: Treatment and prevention" and "Buruli ulcer (Mycobacterium ulcerans infection)" and "Infectious complications of intravesical BCG immunotherapy".)
TERMINOLOGY — TB terminology is inconsistent in the literature [2]. Relevant terms are defined in the table (table 1)
MYCOBACTERIA AND HOST IMMUNITY — Virtually any prior mycobacterial infection (whether naturally acquired or vaccine induced) appears to produce some level of protection against subsequent disease due to TB and, in some cases, to other mycobacteria [3]. Natural infections that confer protection against TB include prior contained infection with Mycobacterium tuberculosis itself or prior infection with nontuberculous mycobacteria [4,5]. These observations suggest that protection is conferred by the immune response to common mycobacterial antigens. (See "Immunology of tuberculosis".)
Prior infection with M. tuberculosis that has been contained provides as much as 80 percent protection against disease after subsequent exposure [6]. However, prior active disease is associated with an increased risk of a second episode of active TB due to a different strain in both HIV-infected and HIV-uninfected persons [7-13]. (See "Tuberculosis: Natural history, microbiology, and pathogenesis", section on 'Host factors'.)
Bacille Calmette-Guérin (BCG) has also been associated with an overall reduction in childhood mortality not attributable to TB; this effect is not fully understood but may be related to epigenetic and metabolic reprogramming of innate immune cells [14].
BCG vaccine is administered either intradermally or percutaneously. Since natural infection and sensitization to M. tuberculosis in humans usually occurs by the respiratory route, research is being conducted on aerosol administration of BCG [15,16].
BACILLE CALMETTE-GUÉRIN VACCINE
Immune response to BCG — Numerous reports have evaluated the immune response to primary Bacille Calmette-Guérin (BCG) immunization. Studies among infants demonstrate BCG-associated induction of CD4+ and CD8+ T cells, interferon (IFN)-gamma+, interleukin-2+, tumor necrosis factor (TNF)-alpha+, and polyfunctional CD4+ T cells [17]. However, none of these responses were found to correlate with protection against TB in BCG-immunized infants [18] nor have any studies identified a surrogate marker of BCG-induced protection against TB. Among BCG-immunized infants in Uganda, lower rates of purified protein derivative (PPD)-specific T cell responses have been found among infants of mothers with TB infection than infants of mothers without TB infection [19]. In contrast, a study from South Africa found no difference in BCG-specific T cell proliferative responses or cytokine/chemokine induction between infants of mothers with and without TB infection [20].
Infant exposure to maternal HIV does not appear to have a significant effect on infant immune response to BCG administered at birth. In one study, infant immune responses to BCG did not differ between HIV-exposed and HIV-unexposed infants [20]. In another study, responses to M. tuberculosis PPD were marginally lower at 14 weeks in the HIV-exposed group but responses to BCG were similar; all measured responses were similar at 24 and 52 weeks [21]. However, actual infant infection with HIV is a contraindication to BCG immunization (see below).
Studies in adults indicate that BCG induces CD4+, IFN-gamma responses, and IFN-gamma and TNF-alpha secreting CD8+ cells with cytotoxic activity; data on polyfunctional T cells have been conflicting [22]. In a study among BCG-naïve, interferon-gamma release assay-negative adults from the Netherlands, immune responses could be separated into two distinct patterns: a predominant proinflammatory response and a predominant T regulatory cell response with no cytokine induction [22]. An increasing body of literature is exploring the potential role that BCG-induced humoral immunity may play a role in protection against TB disease [23].
Timing of infant immunization — Several studies have assessed immune responses among infants whose BCG immunization was delayed for several weeks after birth. In one study in South Africa, HIV-unexposed infants immunized at eight weeks of age had higher frequencies of BCG-specific polyfunctional T cells at one year of age compared with infants immunized at birth [24]. In another study among HIV-unexposed infants in the Gambia, cytokine responses were lower among infants immunized at 18 weeks of age than among infants immunized at birth; however, immune responses in the two groups were comparable at 36 weeks [25]. In a study among HIV-exposed infants in South Africa, immune responses were compared in infants immunized at birth versus eight weeks; at six weeks following immunization, BCG-specific T cell responses were similar, but at 14 weeks the frequency of IFN-gamma expressing CD4+ T cells and multifunctional BCG-specific responses were higher in the delayed immunization arm [26].
These immunologic studies of delayed infant BCG immunization cannot yet be applied to immunization policy since a surrogate immune marker of BCG-induced protection has not been identified [27]. Routine immunization at birth remains the preferred practice since delayed immunization may reduce BCG compliance rates and subject infants to an increased risk of interim TB exposure and all-cause mortality [28].
BCG vaccine efficacy — The efficacy of BCG varies depending on whether it is administered as a primary or booster vaccine, and whether the endpoint of interest is TB infection or TB disease. These categories are considered separately below.
Efficacy against TB infection
Primary immunization — Most studies of BCG have evaluated the efficacy for protection against TB disease. In older BCG trials, TB infection was not used as an efficacy endpoint since the presence of TB infection could only be assessed using the tuberculin skin test [TST], which itself is affected by BCG. (See "Tuberculosis infection (latent tuberculosis) in adults: Approach to diagnosis (screening)".)
Interferon-gamma release assays (IGRAs) have been shown to be useful tools for evaluating the efficacy of BCG in preventing TB infection [29-35] (see "Use of interferon-gamma release assays for diagnosis of latent tuberculosis infection (tuberculosis screening) in adults"):
●In one systematic review and meta-analysis including 14 studies and more than 3800 individuals who received BCG vaccination, BCG had an overall protective efficacy of 19 percent (risk ratio [RR] 0.81, 95% CI 0.71-0.92) [32]. In a restricted analysis of studies in which BCG was administered at birth, protective efficacy of 28 percent (RR 0.72, 95% CI 0.56-0.93) was observed.
●In a retrospective study from Greenland including nearly 1700 participants, BCG was associated with reduced risk of TB infection (hazard ratio [HR] 0.52) and risk of TB disease (HR 0.50); these effects persisted beyond 15 years of age [33].
●In a cross-sectional analysis of a cohort including more than 3400 TB contacts in the United Kingdom, a strong association was found between BCG immunization and protection against TB infection. The overall vaccine efficacy was 30 percent and diminished with time since immunization; however, the prevalence of TB infection at >20 years was still lower among those with prior BCG immunization [34].
●In a household contact study from Indonesia, BCG immunization was associated with protection against both baseline TB infection (positive IGRA) and incident TB infection (conversion from negative baseline IGRA to positive IGRA at 14 weeks). Protection was strongest against incident infection (relative risk 0.56, 95% CI 0.40-0.77) and decreased with increasing age [35].
●An observational study among 1404 children and adolescents immigrating to Sweden showed that a BCG scar was associated with a 59 percent reduced risk of TB infection, as documented by a positive TST or IGRA [36].
Booster immunization — In a randomized trial more than 900 IGRA-negative adolescents in South Africa who received neonatal BCG vaccination were randomized to receive BCG booster, the H4:IC31 vaccine (an investigational vaccine containing mycobacterial antigens), or placebo [37]. Neither vaccine was effective in preventing initial IGRA conversion to positive. However, the rate of sustained IGRA conversion (a potential surrogate for TB infection defined as conversion to positive IGRA without reversion to negative IGRA at six months following vaccination) was reduced among recipients of the BCG vaccine by 45.4 percent and among recipients of the H4:IC131 vaccine by 30.5 percent.
Efficacy against TB disease
Primary immunization — BCG is administered routinely at birth in TB-endemic countries as a component of the global Essential Program on Immunization (EPI). Because efficacy in mycobacteria-naïve newborns is higher than in older children and adults, these groups are considered separately here. BCG is 70 to 80 percent effective against all forms of TB when administered at birth to mycobacteria-naïve infants [38]. Primary vaccination of older children and adults is considerably less effective [39-41]. These differences are likely due to the immune status of the recipient and degree of prior exposure to mycobacteria (both M. tuberculosis and nontuberculous mycobacteria [NTM]).
Mycobacteria-naïve infants and newborns — The efficacy of BCG immunization in providing protection against subsequent TB in mycobacteria-naïve newborns and infants has been evaluated in four prospective trials conducted before effective treatment of TB was available [42-47]. Collectively, these trials demonstrated an efficacy of 73 percent for protection against active disease and 87 percent against death [42]. Infant BCG immunization is also very effective in preventing tuberculous meningitis and disseminated disease in children (75 to 86 percent efficacy) [38-42,48,49]. Infant BCG immunization provides long-term protection against pulmonary TB.
●In a 40-year population-based survey from Norway, childhood BCG immunization provided 67 percent protection against pulmonary TB at 10 years and 63 percent protection at 20 years [50].
●In a cohort study from Greenland, infant BCG immunization provided 50 percent protection against TB disease in older children and adults, the age groups where most disease is pulmonary [33].
●Other studies have documented significant efficacy for least 15 years, although efficacy typically begins to wane after this interval [50,51]. Waning BCG-induced protection has led to studies of BCG boosters and to studies of investigational TB vaccine boosters. (See 'Booster immunization' below and 'TB vaccines under development' below.)
Mycobacteria-exposed older children and adults — The protective efficacy of primary BCG immunization appears to decrease as prior exposure to mycobacteria increases, which in turn is a function of age and rates of endemic TB [38,52]. Thus, BCG is less effective when primary immunization is provided to older individuals than when it is provided to mycobacteria-naïve newborns [53]. Two factors may be relevant: first, natural mycobacterial infection may provide protection equal to BCG (“masking”) and second, BCG may have reduced replication (and therefore reduced efficacy) in the face of prior mycobacterial immunity (“blocking”) [3,54]. The second hypothesis is supported by an animal model in which prior infection with NTM blocks the protective effects of BCG [55].
Analyses that have combined trial results in newborns with trial results in older children and adults have led to the misleading conclusion that the efficacy of BCG is “variable”. As noted above, BCG has been uniformly effective in all prospective trials conducted among mycobacteria-naïve infants. Variable efficacy has also been incorrectly attributed to higher rates of NTM infection in trials conducted near the equator, most notably based on the results of a trial in South India with methodologic flaws.
●The South India 1980 trial was a large randomized trial designed to investigate the efficacy of BCG against TB disease [56,57]. The study objectives of this trial were: (1) to compare the efficacy of different BCG strains and doses and (2) to assess the efficacy of BCG in individuals with and without prior latent TB (determined by TST).
In reality, however, the trial was largely a study of the effect of primary BCG immunization in older children and adults, many of whom were already tuberculin positive and all of whom lived in an area where prior infection with Mycobacterium leprae was prevalent. Over 270,000 patients were enrolled, but only 0.6 percent were mycobacteria-naïve newborns. Surveillance for TB was based on positive chest radiographs, which were only performed at age ≥5 years; those with positive radiographs had sputum microbiology. Criteria for TB were limited to positive sputum culture or positive acid-fast bacilli stain, and there were no methods for detecting extrapulmonary TB. These methods would be insensitive for the diagnosis of TB in children. Further, the observed rate of TB in this large study cohort was half the predicted rate.
A critical review of BCG trial methodology concluded that this trial had major methodologic flaws [40]. At most, this trial demonstrated that BCG vaccination of mycobacteria-experienced older children and adults in India does not lead to a reduction in pulmonary TB defined by positive sputum cultures and chest radiographs.
●BCG efficacy does not correlate with latitude or higher baseline rates of exposure to NTMs near the equator, as suggested previously [58]. Studies indicate that population exposure to NTMs is high in northern as well as southern latitudes [38,59]. Further, no well-designed prospective trials among mycobacteria-naïve infants (the population in which BCG is most effective) have been conducted in southern latitudes. When identical methods were applied to surveillance for disseminated NTM infection in patients with AIDS, rates were high in the northern hemisphere and low in the southern hemisphere: 11 to 22 percent in the United States and Finland versus 2 to 3 percent in Kenya and Trinidad [60]. This disparity may be related to protection associated with the higher rate of TB in southern regions. A United States study showed that prior TB was associated with a reduced risk of disseminated Mycobacterium avium in patients with AIDS [61].
Mycobacteria-naïve older children and adults — Some older children and adults remain sufficiently mycobacteria naïve to respond to BCG. Trials in which TSTs were used to exclude children or adults with prior mycobacterial immunity have been able to demonstrate efficacy against TB disease in some cases.
●A randomized trial initiated in the 1950s that involved over 25,000 TST-negative British teenage students demonstrated BCG vaccine efficacy of 76 percent against TB disease at 15 years [49].
●A randomized trial of BCG vaccination was conducted between 1935 and 1938 among 3287 TST-negative Native Americans age 0 to 20 (28 percent ≤5 years). Evaluation of vaccine recipients 11 years after vaccine administration demonstrated 75 percent reduction in radiographically diagnosed TB; evaluation at 20 years demonstrated an 82 percent reduction in overall mortality due to TB. The protective efficacy of BCG against disease due to TB was 70 percent.
●Evaluation of vaccine recipients 60 years after vaccine administration (the longest follow-up period of any BCG trial) included data for 1998 of the original participants and demonstrated protection against pulmonary TB was 52 percent at 50 years [48].
This trial also demonstrated that BCG may confer protection against multiple episodes of TB. The rate of two or more episodes of TB was 34 per 100,000 person-years in the placebo group and 4 per 100,000 person-years in the BCG group, consistent with a vaccine efficacy of 89 percent (95% CI 53-99 percent) [48].
●In a subsequent analysis of the South India trial, BCG was found to be 32 percent effective in subjects who had negative NTM skin test results (PPD from Mycobacterium intracellulare [PPD-B]) at baseline [62].
●In a study from Brazil, where approximately 70 percent of children are tuberculin negative, BCG efficacy was 25 percent in subjects who received their first BCG vaccine at age 7 to 14 [63].
●Efficacy does not appear to be related to the particular BCG strain used in the vaccine or to latitude, as explained (see 'Mycobacteria-exposed older children and adults' above) [58]:
Booster immunization — Waning immunity 15 to 20 years after primary BCG vaccination in infancy has been the rationale for efforts to develop an effective booster for adolescents and adults [38]. BCG revaccination is generally well tolerated and carries minimal risk [64]. However, large studies in Malawi and Brazil have shown that a booster dose of live BCG does not increase the efficacy of protection against TB disease [52,65,66]. A follow-up study indicated that revaccination also had no demonstrable effect on long-term mortality [67]. The host immune response induced by the initial BCG dose may prevent replication of live organisms administered in a subsequent vaccine dose ("blocking") [55] or the acquisition of infections due to nontuberculous mycobacteria after infancy may provide protection equivalent to a BCG booster ("masking").
Efficacy against other mycobacterial infections — Protection against mycobacterial disease does not appear to be species-specific; BCG may confer protection against infections due to nontuberculous mycobacteria [68]. Further, NTM infection is associated with protection against TB [69].
●Mycobacterium leprae – Studies have demonstrated that BCG reduces the risk of disease due to M. leprae by 50 to 80 percent, and this effect may be increased with booster doses of BCG [65,70]. (See "Leprosy: Treatment and prevention", section on 'Prevention'.)
●Mycobacterium ulcerans – BCG immunization of infants induces M. ulcerans-specific immune responses [71]. Some data suggest BCG confers some protection preventing Buruli ulcer disease due to M. ulcerans [72,73] and BCG may reduce the risk of osteomyelitis, a major complication of M. ulcerans infection [74]. However, other retrospective data have found no evidence for BCG protection against Buruli ulcer disease [75]. (See "Buruli ulcer (Mycobacterium ulcerans infection)".)
●M. avium complex – BCG provides cross protection against childhood lymphadenitis due to M. avium complex [76]. Cessation of routine childhood BCG immunization in Sweden and Finland was associated with a marked increase in the rate of childhood adenitis due to NTM [77,78]. (See "Nontuberculous mycobacterial pulmonary infections in children", section on 'M. avium complex'.)
Effect on child morbidity and mortality — Neonatal BCG administration provides protection against TB and reduces all-cause mortality among neonates and infants. In Guinea-Bissau, BCG reduced all-cause neonatal mortality by >40 percent among low birthweight infants [79]. Protection against septicemia and respiratory infection has also been observed [80]. In a randomized controlled trial in Guinea-Bissau, infants <2500 g who received BCG at birth had a 43 percent reduction in 28-day infectious disease mortality compared with infants with delayed administration of BCG [81]. A randomized trial comparing early (pre-1931) and later strains of BCG-based vaccines showed that effects on childhood morbidity did not differ between “early” (BCG Russia) and “late” (BCG Japan/BCG Denmark) BCG strains [82].
Another study in Guinea-Bissau (where as many as 50 percent of BCG-immunized children do not develop an associated scar) noted that development of a BCG scar was associated with lower mortality and morbidity among children <5 years of age. Protection was strongest against respiratory infection and in the first year of life [83].
Additional analysis of data from randomized trials in Guinea-Bissau has shown that the beneficial effect of BCG at birth on neonatal mortality in low-birth weight infants is maximal during days 1 to 7 of life among boys and during days 8 to 28 among girls [84]. In one meta-analysis evaluating BCG administration to low birth weight or premature infants when immunization was given early (≤7 days) or later (>7 days), the safety and immunogenic profiles were similar [85]. In addition, a modeling study of infant BCG immunization found that universal immunization at birth (rather than six weeks or later) could reduce global pediatric TB mortality by 16.5 percent [28].
A study from Uganda has demonstrated that BCG immunization on the first day of life reduces neonatal infectious disease morbidity compared with immunization at six weeks and suggests that the benefits documented in Guinea-Bissau may be seen in other regions where neonatal infectious diseases are common. The study documented a 25 percent reduction in episodes of clinician-diagnosed infectious disease in the first six weeks of life. Effects were more pronounced in low-birthweight babies (≤2500 g versus >2500 g) and in boys (versus girls) [86].
These findings emphasize the importance of administering BCG at birth rather than at later follow-up health visits, and that a multiple-dose BCG vial should be opened even if it will only be used to administer a single dose of BCG to a newborn.
Effect on risk of respiratory infection and COVID-19 in adults — Evidence of an effect of BCG on non-TB respiratory infections in infants has led to trials to determine if similar effects and be detected in adults.
A randomized trial in Greece involved administration of BCG or placebo to 200 older hospitalized adults on the day of discharge. The one-year incidence of new infections (predominantly respiratory) was 25 percent in the BCG group and 42 percent in the placebo group [87]. Studies are underway to determine if BCG immunization will protect adults against coronavirus disease 2019 (COVID-19) [88].
Effect on risk of malignancy — Studies have been conflicting on whether BCG vaccination in childhood might have an association with later development of leukemia or lymphoma. Some studies have suggested that BCG vaccination might be associated with increased incidence of and mortality due to leukemia and lymphoma, especially non-Hodgkin lymphoma [89-91], while others have suggested no difference or decreased leukemia incidence and mortality [92-94].
A published retrospective, 60-year follow-up of a multicenter clinical trial of BCG vaccination in American Indian and Native Alaskans in the United States, reported that BCG vaccination in childhood was associated with a significant reduction in lung cancer. In this study, 2963 subjects, median age 8 years, were randomized to receive a single dose of BCG (strain 317 or 575) or placebo. While there was no difference in overall mortality or cancer diagnosis between groups, including lymphoma or leukemia, the rate of lung cancer was significantly lower in BCG vaccinated recipients (18.2 versus 45.4 cases per 100,000 person-years; hazard ratio 0.38, 95% CI 0.20-0.74), controlling for sex, region, alcohol overuse, smoking, and TB. The mechanism of this observed protection is unknown [95].
Areas of uncertainty — The efficacy of BCG vaccination has not been studied in the setting of drug-resistant TB; however, drug susceptibility would not be expected to influence the efficacy of BCG. There are no prospective data on the efficacy of BCG for health care workers or travelers. (See 'Groups to consider for vaccination' below.)
Safety and adverse effects
General principles — Injection site reactions following BCG vaccination are common. More serious adverse effects include osteitis, osteomyelitis, and disseminated infection. Potential factors affecting the rate of adverse reactions include the BCG dose, vaccine strain, and method of vaccine administration.
As many as 95 percent of BCG recipients have a local reaction at the site of injection characterized by a bluish-red pustule accompanied by pain, swelling, and erythema within two to three weeks after vaccination. Ulceration with drainage occurs at the vaccine site in about 70 percent of cases, and about 75 percent of recipients experience myalgia. After about six weeks, the pustule ulcerates, forming a lesion approximately 5 mm in diameter. Lesions typically heal by three months with permanent identifiable scar in approximately 80 percent of recipients (picture 1).
Less common manifestations include abscess and regional lymphadenitis (1 to 2 percent). These may be accompanied by draining sinus tracts or fistulae; the risk for neonates is higher than for older children [96,97].
Management of local reactions consists of attentive wound care. Since viable organisms can be recovered from ulcer drainage, vaccination sites should be covered to reduce transmission of the vaccine strain [98]. Abscess or fistula formation may rarely require surgical drainage. In the setting of suppurative lymphadenitis, some favor antimycobacterial therapy, although antimicrobial therapy has not been shown to be beneficial in meta-analyses [99-102]. Nonsuppurative lesions are best managed with observation alone [99,100].
●Osteitis and osteomyelitis – BCG osteitis and osteomyelitis are rare; these entities may occur as a result of direct spread from the administration site; less commonly, these may also occur as a result of dissemination [103]. Osteitis affecting the epiphyses of the long bones (particularly the leg) can occur 4 to 24 months after vaccination. It has been reported in 0.01 per million vaccinees in Japan (multipuncture technique) and 30 per million in Finland (intradermal technique) [97]. There is an increased incidence of genetic variants in the gene encoding Toll-like receptor 2 among infants who develop BCG osteitis following vaccination, compared with those that do not [104]. (See "Toll-like receptors: Roles in disease and therapy", section on 'Specific TLRs'.)
Treatment of osteomyelitis usually consists of surgical debridement plus administration of isoniazid (INH) and rifampin (RIF) for 6 to 12 months [97,105] (BCG is resistant to pyrazinamide and, unlike treatment of wild strain M. bovis infection, ethambutol is not generally needed). (See "Mycobacterium bovis" and "Infectious complications of intravesical BCG immunotherapy".)
●Disseminated disease – Disseminated disease following BCG vaccination occurs most commonly in the setting of immunosuppression. Populations at risk for dissemination include infants with severe combined immunodeficiency, individuals with HIV infection or other immunodeficiency, and individuals who have received intravesical BCG for bladder cancer [106]. The risk of disseminated BCG in HIV-infected infants ranges from 403 to 1300 per 100,000 doses administered. Mortality among HIV-infected children with disseminated BCG is about 75 percent, although this figure is from studies in which most children were not on current antiretroviral therapy (ART) [107-109]. In the pre-HIV era, the rate of disseminated BCG was reported to be between 0.19 and 1.56 per million vaccinees [107].
Although most cases of disseminated BCG have been reported in the weeks and months following childhood immunization, case reports suggest that reactivation can occur years later in the setting of immunosuppression [107]. Clinical manifestations and treatment of disseminated disease due to BCG are outlined separately. (See "Infectious complications of intravesical BCG immunotherapy", section on 'Systemic complications' and "Mycobacterium bovis", section on 'Treatment'.)
HIV infection — Safety of BCG administration in HIV-infected individuals is an important concern given that it is a live vaccine. Adverse effects due to BCG vaccination may be more frequent among persons who have advanced or symptomatic HIV infection than among persons who have asymptomatic HIV infection, are on effective ART, or are not HIV infected [110-114].
Case reports of BCG-related lymphadenitis, local ulceration, and disseminated BCG disease (which can occur several years after BCG vaccination) have been described in HIV-infected patients. In a retrospective study of 352 HIV-infected children started on ART, clinically significant complications of BCG immunization were observed in 6 percent of cases [115]. In some cases, these reactions appeared to be due to immune reconstitution inflammatory syndrome [109,115,116]. (See 'Safety and adverse effects' above and "Immune reconstitution inflammatory syndrome".)
Management of disseminated BCG — Patients with disseminated BCG infection should be treated with INH and RIF for at least several months. In the setting of immunosuppression, addition of ethionamide may be beneficial. This was illustrated in a small series of HIV-infected children with BCG-induced complications for whom treatment with INH, RIF, and ethionamide was administered [115]. HIV-infected patients with disseminated BCG who are not already on combination ART should be started on ART. Vaccine-strain BCG is sensitive to INH and RIF; however, some suggest that resistance to INH or RIF may develop in HIV-infected children treated for complications of BCG [117,118].
Administration and dosing — There is no evidence that one strain of BCG is preferable or superior to another. Numerous strains of BCG have been used to produce a BCG vaccine, and none of these strains is demonstrably superior in terms of their efficacy or immunogenicity. In the United States, BCG for prevention of TB is manufactured by Merck as the Tice strain [119] but is often difficult to obtain. Intravesical forms of BCG are available for treatment of bladder cancer but not for prevention of TB
BCG can be administered intradermally or by multiple percutaneous puncture using a device with multiple tines. The clinical efficacy of intradermal and percutaneous administration appears comparable; this was illustrated in a study of 11,680 newborn infants in South Africa randomized to receive intradermal or percutaneous BCG; the rate of TB was equivalent between the groups in the first two years of life [120].
The World Health Organization (WHO) favors intradermal administration of BCG. Percutaneous puncture is the method recommended by the manufacturer for BCG distributed in the United States; we favor intradermal administration where available for ease of administration and standardization of dose [120]. Intradermal administration appears to result in greater in vitro immunogenicity, skin test conversion, and scarring than percutaneous administration [121].
BCG is supplied as a lyophilized product and must be reconstituted according to product specific instructions (different dilutions for adults and infants). The standard volume for intradermal administration of BCG vaccine for adults after reconstitution is 0.1 mL. The standard volume for multiple puncture administration is 0.2 to 0.3 mL. The dose for infants and neonates up to 12 months of age is 0.05 mL for intradermal administration or 0.2 to 0.3 mL for multiple puncture administration. Other childhood vaccines can be administered simultaneously with BCG.
Groups to consider for vaccination — The approach to BCG vaccination policy depends on the regional prevalence of TB and is variable around the world. In countries where the prevalence of TB is moderate to high, neonatal vaccination is recommended by the WHO and is administered routinely. In some circumstances, BCG is also administered for health care workers and close contacts of patients with TB (particularly multidrug-resistant [MDR] TB) with negative tuberculin tests. In most TB-endemic countries, BCG is given to children soon after birth. Boosters are administered later in some countries despite the fact that they have been proven ineffective in adding to protection against disease. Further, they are not recommended for BCG vaccine recipients who remain negative by subsequent tuberculin testing.
In countries with a low burden of TB, universal BCG vaccination is not recommended or required. For example, routine BCG vaccination has never been implemented in the United States; instead, TB control measures have focused on detection and treatment of latent TB. Universal BCG vaccination was utilized for all school children at age 13 and all neonates in high-risk groups in the United Kingdom between 1953 and 2005. However, routine BCG vaccination was discontinued in 2005 because of diminishing incidence of TB.
The WHO does not recommend use of BCG vaccine in the countries meeting the following criteria [122]:
●Average annual rate of smear-positive pulmonary TB below 5 per 100,000
●Average annual rate of tuberculous meningitis in children under five years below 1 per 10 million population
●Average annual risk of TB infection below 0.1 percent
Worldwide — In settings where TB is highly endemic or where there is high risk of exposure to TB, the World Health Organization recommends that a single dose of BCG vaccine should be given to all infants [122]. BCG vaccination should be administered to healthy neonates as soon as possible after birth [123]. In addition, immunization of BCG-naïve school-age children (aged 7 to 14) not previously vaccinated has been shown to confer partial protection against TB [63].
BCG immunization practices vary by region depending on the prevalence of TB [124-126]. In countries with high prevalence of TB, childhood BCG immunization should be administered routinely. For countries with intermediate to low rates of TB (<5/100,000 smear-positive cases per year), selective childhood BCG immunization for children at particular risk of TB exposure is appropriate. [125,126]. For example, BCG immunization may be reasonable for children with exposure to drug-resistant disease.
Caution in administering BCG must be exercised in regions with high prevalence of both TB and HIV. The risk of disseminated BCG in HIV-infected infants ranges from 403 to 1300 per 100,000 doses administered; in one series, a mortality rate of 75 percent was reported among HIV-infected children with disseminated BCG [107-109]. Therefore, BCG vaccination is not appropriate for infants or adults with known HIV infection (or other immunodeficiency) nor for infants with symptoms consistent with HIV infection in the absence of laboratory confirmation of actual HIV infection [127-130].
In contrast, BCG vaccination should be administered to asymptomatic infants born to mothers with unknown HIV status in countries with high TB prevalence [108,130-132]. In infants whose HIV status is unknown and who are born to HIV-positive mothers and who lack symptoms suggestive of HIV, BCG vaccine should be given after considering local factors: coverage and success of the prevention of mother to child transmission of HIV programs, possibility of deferring BCG vaccination in HIV-exposed infants until HIV infection status has been established, availability of early diagnosis of HIV infection in infants, and provision of early ART to HIV-positive infants [122].
An international advisory group has recommended that routine childhood BCG immunization be continued until all elements of an HIV testing program can be implemented [108]. Although some favor delaying BCG vaccination for HIV-exposed infants until HIV testing results are available, implementation of this approach requires careful coordination [108]. In addition, delaying BCG vaccine may reduce BCG immunization rates for children who are not HIV infected.
BCG-vaccinated infants born to known HIV-infected mothers should be followed clinically to evaluate for signs of disseminated BCG. For infants with exposure to smear-positive pulmonary TB in the neonatal period, BCG vaccination should be deferred until six months of preventive INH therapy have been administered to the infant (so that INH does not inactivate the live organisms in the BCG vaccine).
United States — In the United States, BCG vaccination may be considered in the following circumstances [133]:
Children — BCG vaccination should only be considered for children who have a negative TB test, who are continually exposed, and cannot be separated from adults who:
●Are untreated or ineffectively treated for TB disease, and the child cannot be given long-term primary preventive treatment for TB infection or
●Have TB disease caused by strains resistant to isoniazid and rifampin
Health care workers — BCG vaccination of health care workers should be considered on an individual basis in settings in which:
●A high percentage of TB patients are infected with TB strains resistant to both isoniazid and rifampin
●There is ongoing transmission of drug-resistant TB strains to health care workers and subsequent infection is likely or
●Comprehensive TB infection-control precautions have been implemented, but have not been successful
Exposure to MDR-TB — The efficacy of BCG vaccination for health care workers, travelers, and individuals in the community with expected exposure to multidrug-resistant (MDR-) TB is uncertain. However, given the potentially significant risk of MDR-TB treatment failure, together with the relatively low rate of complications related to BCG vaccination in immunocompetent individuals, some favor administering BCG vaccination to unvaccinated, tuberculin-negative individuals exposed to MDR-TB [130]. Further study is needed to reconcile the protective efficacy of BCG vaccination in the setting of MDR-TB exposure among older children and adults.
Health care workers — The protective efficacy of BCG vaccination in health care workers is not certain [134]. In settings with low risk of M. tuberculosis transmission, BCG vaccination for health care workers is not warranted [134]. In regions with high risk for TB transmission, careful adherence to TB infection control practices should be emphasized. Despite data demonstrating the limited efficacy of BCG in adults, consideration of BCG vaccination may be appropriate for health care workers from low-risk countries caring for patients or refugees in TB-endemic countries [135,136]. (See "Tuberculosis transmission and control in health care settings".)
In the setting of substantial risk for exposure to MDR-TB strains, BCG vaccination for health care workers should be considered on an individual basis [130]. In such circumstances, counseling should be offered regarding the variable data for BCG vaccination efficacy, including discussion of risks and benefits associated with BCG vaccination.
Groups not to vaccinate
Immunocompromised patients — Safety of BCG administration in immunocompromised individuals is an important concern given that it is a live vaccine. BCG vaccination should not be administered to individuals with immune compromise due to HIV infection, congenital immunodeficiency, malignancy, or immunosuppressive drugs such as corticosteroids and tumor necrosis factor-alpha blockers. (See 'Safety and adverse effects' above.)
In addition, caution should be observed among immunocompromised patients with household or other close contact with individuals recently immunized with BCG; drainage from the injection site contains live organisms and should be covered to avoid transmission.
Adults with HIV infection and individuals with HIV infection in areas of low TB prevalence should NOT receive BCG vaccination [137]. The role of BCG vaccination for HIV-infected children in areas with endemic TB is discussed in the preceding section. (See 'Children' above.)
Pregnant women — Although BCG vaccination has not been associated with harmful fetal effects, it should not be administered in pregnancy since it is a live vaccine [131].
Interpreting TST and IGRA after BCG — Most individuals who have received BCG vaccine have a positive tuberculin skin test (TST) two to three months following vaccination [138]. The reaction wanes with time; at more than 10 years after infant vaccination, it is generally <10 mm. However, with repeated tuberculin testing, such reactions may be boosted to >10 mm.
●In a study from Montreal among 5952 persons who received BCG 10 to 25 years previously, 8 percent vaccinated in infancy had significant tuberculin reactions, compared with 18 percent among those vaccinated between age 1 and 5, and 25 percent among those vaccinated after the age of 5 (p<0.001) [139].
●Long-term follow-up from a vaccine study among Native Americans showed that administration of BCG was associated with an increased risk of TST reactions ≥10 mm as long as 55 years later. BCG vaccination after infancy was associated with an increased risk of TST reactivity in the first 15 years after vaccination (adjusted hazard ratio [aHR] 2.33). This association remained during the interval 16 to 55 years after vaccination, although the effect was attenuated (aHR 1.26) [140].
Therefore, previous BCG vaccination may influence decisions regarding interpretation of TST results in individuals vaccinated at birth. Since interferon-gamma release assays (IGRAs) are not affected by BCG administration, this test can be used to interpret positive TSTs in persons with a history of BCG immunization or can be used as the initial test for TB infection. (See "Use of interferon-gamma release assays for diagnosis of latent tuberculosis infection (tuberculosis screening) in adults".)
In the setting of serial tuberculin testing, prior BCG vaccination may be associated with boosting. The booster phenomenon is defined as a TST that is initially defined as negative (in a patient whose previous TST reactivity has diminished over time) and is then boosted to a positive test by the skin testing procedure itself. If repeated tuberculin testing is planned (such as annual screening for health care workers), initial two-step testing should be performed to distinguish tuberculin conversion from boosting. (See "Use of interferon-gamma release assays for diagnosis of latent tuberculosis infection (tuberculosis screening) in adults".)
Intravesical administration of BCG for bladder cancer has been reported to result in conversion of the TST in ≥50 percent of patients but should not result in a positive IGRA [141,142].
TB VACCINES UNDER DEVELOPMENT — A major international effort is underway to develop an improved vaccine strategy for prevention of TB, an objective that is considered critical for elimination of TB by 2035.
Several candidate vaccines are in clinical trials; these include booster vaccines (designed to boost the effect of Bacille Calmette-Guérin [BCG]) and priming vaccines (intended to replace BCG) [143,144]. Types of candidate vaccines include subunit vaccines (in which antigens are delivered in a viral or other non-mycobacterial vector), live attenuated vaccines, and inactivated whole-cell vaccines [145,146]. Three designs are being used in clinical trials: Prevention of Infection (POI), Prevention of Disease (POD), and Prevention of Recurrence (POR) [143].
Two candidate vaccines have shown efficacy or proof-of-concept in phase 2 or phase 3 POD trials:
●M72/AS01E − M72/AS01E is a subunit booster vaccine consisting of two M. tuberculosis proteins (Mtb32A and Mtb39A) and an adjuvant (AS01E). In a randomized trial including more than 3500 HIV-uninfected adults in South Africa, Kenya, and Zambia with latent TB infection (as defined by a positive interferon-gamma release assay) and likely ongoing TB exposure, the majority of whom had received BCG vaccination as children, the incidence of TB disease was lower among those who received M72/AS01E than among those who received placebo after three years of follow-up (0.3 versus 0.6 cases per 100 person-years); the overall vaccine efficacy was 49.7 percent (95% CI 2.1-74.2) [147,148]. Protection was greater among participants ≤25 years of age than among those >25 years of age, and greater among men than in women. Serious adverse events and deaths occurred with similar frequencies in the two groups.
●SRL-172 and DAR-901 − SRL-172 is an inactivated whole-cell booster vaccine derived from a nontuberculous mycobacterium. In a randomized phase 3 trial including more than 2000 HIV-infected adults in Tanzania with a history of childhood BCG randomized to receive the vaccine (five intradermal doses) or placebo, a reduction in the rate of TB disease was observed among those who received the vaccine (hazard ratio 0.61, 95% CI 0.39-0.96) [149,150]. Vaccine strain DAR-901 is a scalable manufacturing method for SRL-172 and has completed phase 1 and 2b trials [151,152].
TB vaccines are also being studied as potential therapeutic adjuncts to chemotherapy of active TB disease.
SUMMARY AND RECOMMENDATIONS
●Bacille Calmette-Guérin (BCG) is a live strain of Mycobacterium bovis developed by Calmette and Guérin for use as an attenuated vaccine (intradermal or multiple percutaneous puncture) to prevent tuberculosis (TB) and other mycobacterial infections. The magnitude of protection appears to be in the range of 70 to 80 percent in the first 15 years of life but falls to the range of 50 percent or lower subsequently. (See 'Mycobacteria and host immunity' above.)
●BCG protects against pulmonary TB and is also highly effective in preventing tuberculous meningitis and disseminated disease in children. Vaccination of mycobacteria-naïve newborns and infants appears to confer greater benefit than vaccination of older children and adults who have often had other mycobacteria exposure. (See 'Efficacy against TB disease' above.)
●Injection site reactions are common following BCG vaccination. Management of local reactions consists of attentive wound care, and indolent lesions are typically resolve with observation alone. Since viable organisms can be recovered from ulcer drainage, vaccination sites should be covered to reduce transmission of the vaccine strain. (See 'Safety and adverse effects' above.)
●Serious adverse effects of BCG include osteitis, osteomyelitis, and disseminated infection. Disseminated BCG infection following BCG vaccination occurs most commonly in the setting of HIV infection and other forms of immunosuppression. (See 'Safety and adverse effects' above.)
●Most individuals who have received BCG vaccine have a tuberculin reaction of 3 to 19 mm in size at two to three months following vaccination. The reaction wanes with time; at more than 10 years after vaccination, it is generally <10 mm. Testing with an interferon-gamma release assay may permit distinction between positive tuberculin reactions due to BCG versus TB. (See 'Interpreting TST and IGRA after BCG' above.)
●In areas where the prevalence of TB is high, we recommend that a single dose of BCG be administered to healthy neonates as soon as possible after birth (Grade 1A). This includes administration of BCG to neonates born to mothers with unknown HIV status. (See 'Worldwide' above.)
●BCG should NOT be administered to individuals with immune compromise due to HIV infection, congenital immunodeficiency, malignancy, or immunosuppressive drugs. For asymptomatic infants with unknown HIV status born to mothers known to be HIV infected, the optimal approach to BCG vaccination is uncertain and should be considered in conjunction with local factors. (See 'Immunocompromised patients' above.)
●We suggest NOT administering BCG in the setting of pregnancy (Grade 2C). (See 'Groups not to vaccinate' above.)
●TB vaccines under development are discussed above. (See 'TB vaccines under development' above.)
