IJE Advance Access originally published online on January 19, 2005
International Journal of Epidemiology 2005 34(3):540-547; doi:10.1093/ije/dyh392
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Special Theme: Infectious Diseases |
BCG vaccination scar associated with better childhood survival in Guinea-Bissau
1 Bandim Health Project, Apartado 861, Bissau, Guinea-Bissau
2 Danish Epidemiology Science Centre, Statens Serum Institut, Artillerivej 5, 2300, Copenhagen S, Denmark
3 Department of Infectious diseases, Malmö University Hospital, Sweden
* Corresponding author. Adam Roth, Danish Epidemiology Science Centre, Statens Serum Institut, Artillerivej 5, 2300, Copenhagen S, Denmark. E-mail: aro{at}ssi.dk
 |
Abstract |
|---|
 Top Abstract Method and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
Background Recent studies have suggested that Bacille Calmette–Guerin (BCG) vaccination may have a non-specific beneficial effect on infant survival and that a BCG scar may be associated with lower child mortality. No study has previously examined the influence of BCG vaccination on cause of death.
Methods Two cohorts (A and B) were used to describe the mortality pattern for children with and without BCG scar and to determine specific causes of death. In cohort A (n = 1813), BCG scar was assessed at 6 months of age and as previously described children with a BCG scar had lower mortality over the next 12 months than children with no BCG scar. In cohort B, 1617 children aged 3 months to 5 years of age had their BCG scar status assessed in a household-based survey and mortality was assessed during a 12-month period. Causes of death were determined by verbal autopsy (VA) and related to BCG scar status in a cause-specific hazard function.
Results Controlling for background factors associated with mortality, there was lower mortality for children with a BCG scar than without in cohort B, the mortality ratio (MR) being 0.45 (95% CI 0.21–0.96). Exclusion of children exposed to TB did not have any impact on the result. In a combined analysis of cohorts A and B, the MR was 0.43 (95% CI 0.28–0.65) controlling for background factors. There were no large differences in distribution of the five major causes of death (malaria, pneumonia, acute diarrhoea, chronic diarrhoea, and meningitis/encephalitis) according to BCG scar status in the two cohorts. Having a BCG scar significantly reduced the risk of death from malaria [MR 0.32 (95% CI 0.13–0.76)].
Conclusions A BCG scar is a marker of better survival among children in countries with high child mortality. BCG vaccination may affect the response to several major infections including malaria.
Keywords BCG, non-specific effects of vaccines, verbal autopsy, infant mortality, cause of death
Accepted 26 October 2004
The efficacy of Bacille Calmette–Guerin (BCG) against pulmonary tuberculosis (TB) in infants and adults varies from 0 to 80% in different studies.1–4 However, BCG is generally considered to protect against tuberculous meningitis and miliary TB among infants.5,6 When introduced in the 1930s, BCG vaccination was sometimes considered to have a non-TB-related beneficial effect on survival in children.7 Recent findings have also suggested that BCG may have a non-specific beneficial effect on childhood survival:8–11 BCG vaccination was associated with a decrease in infant mortality of 45% in rural Guinea-Bissau9 and of 40% in Benin.11 Whether BCG affects specific diseases remains uncertain, but in a case–control study from Brazil, BCG reduced the risk of death from pneumonia by 50%.10
The implications of a BCG scar in TB protection and vaccine efficacy has often been discussed,12–16 but it is still doubtful whether scar development is linked to the immunological response to BCG14 in spite of studies indicating this.17 The immunological relationship between scar and possible non-specific effects of BCG has not been studied. However, considering the importance of vaccination technique for scar development (injected dose18 and route of administration19), we assumed that if BCG vaccination has a beneficial non-targeted effect, this effect would be strongest among children who had responded to vaccination with a scar. Consistently, we observed in Guinea-Bissau that among BCG-vaccinated children, a BCG scar was related to better child survival,20 and in a hospital study from Malawi a BCG scar was related to less skin infections and sepsis.17 The plausibility of BCG being a non-specific immune stimulator has been strengthened by recent studies showing that BCG produces a Th1-biased immune response at birth.21 BCG has also been shown to reduce atopy 22 and cutaneous anergy 23 and to stimulate both cellular23 and antibody responses24 to unrelated antigens.
To pursue this pattern, we screened children <5 years of age in Bissau for BCG scars to examine whether better survival for children with a BCG scar was a reproducible observation. We were able to exclude children exposed to TB at home, since we have had a continuous TB surveillance system in the study area. We conducted verbal autopsies (VAs) for children who died in the present and a previously presented cohort study20 to identify possible differences in the major causes of death for children with and without a BCG scar.
|
Method and subjects |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
The study population consisted of residents in four districts of Bissau, the capital of Guinea-Bissau. The districts Bandim I, Bandim II, Belem, and Mindará have been followed by the demographic surveillance of the Bandim Health Project (BHP) since 1978, 1984, 1984, and 1994, respectively.25 Deaths were ascertained by BHP through the routine household visits every third month for 0 to 3-year-old children or through the census system with at least yearly visits for older children.
Two partly overlapping cohorts were used to describe the mortality pattern for children with or without a BCG scar and to describe specific causes of mortality:
Cohort A. As part of a two-dose measles vaccine trial, children were recruited at 6 months of age and at the same time screened for the presence of a BCG scar. Between November 1996 and May 1998 we included 1813 children (1676 with scar, 137 without scar) who were BCG-vaccinated at least 1 month before scar reading. As described elsewhere,20 scar-positive children had a lower mortality than scar-negative children the following 12 months, the mortality ratio (MR) being 0.41 (95% CI 0.25–0.67).
Cohort B. Between January 1998 and October 1999, 1679 children between 3 months and 5 years of age were visited in their homes and examined for the presence of a BCG scar (1194 scar, 485 no scar). This was done as part of a scar survey, in which all houses in Bandim I and Bandim II were visited. The surveillance system was used to identify individuals in the households, and all individuals living in the study area who consented to take part in the study were assessed for vaccine scars. In the year following the scar reading, 1673 of the 1679 children were visited at least once to assess mortality while 6 (4 scar, 2 no scar) children were lost to follow-up. Of these 1673 children, 1617 (1159 scar, 458 no scar) had a documented BCG vaccination at least 1 month before scar reading. The other 56 children were either not vaccinated or the BCG status was unknown.
|
BCG vaccination |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
Information on vaccination status was obtained from the vaccination card, either at the time of scar assessment (cohort A) or through the routine surveillance system of the BHP (cohort B). In Bissau, the BCG vaccine is customarily given at birth or as soon as possible after the first health contact. BCG immunization was administered Monday to Friday at the central hospital and on Monday and Friday at the local health centre. Pasteur Mérieux, France, and Statens Serum Institut, Denmark, provided the BCG vaccine via the local Expanded Programme on Immunization (EPI). Following EPI recommendations, BCG was administered by a 0.05 ml (0.1 ml for children >1 year of age) intradermal injection in the left deltoid area. For the present study, children were considered BCG-vaccinated if vaccinated at least 30 days before scar reading.
|
TB surveillance |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
Since May 1996, we have registered all cases of TB in the study area through a TB surveillance system. This was based on passive and active case finding, where all adult patients (
15 years) living in the study area reporting to a health centre with symptoms or signs of active TB were further investigated. In addition two nurses visited houses, where cases had been found, every third month to detect secondary cases.26 Since all scar readings in the present study were done from 1996 onwards, we were able to control for exposure to TB in the household. As recommended by the International Union Against Tuberculosis and Lung Disease,27 patients with clinical signs, symptoms, and X-ray changes compatible with active intrathoracic tuberculosis, but without positive bacteriological tests, were treated with antibiotics (co-trimoxazole or amoxicillin) and then re-evaluated clinically and also with a chest X-ray. If there was no improvement and suspicion remained, the patient was diagnosed as having presumed TB. In the present analysis, TB cases include both smear-positive and smear-negative patients with presumed TB. In order to assess the quality of the TB surveillance in the study area, VAs were performed for all deceased adults in the study area from 1996 to 1998. Only two persons who might have died of TB, but had not been diagnosed and treated, were found.20 |
Mortality |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
Overall mortality for children with a documented BCG vaccination before scar assessment was described for children in cohort B in the same manner as it has earlier been described in cohort A.20 Of the 1617 children in cohort B with documented BCG vaccination at least 1 month before scar reading, 488 had also been included in cohort A. If they were enrolled in cohort B before cohort A (31 scar, 8 no scar), their follow-up was terminated in cohort B on the day of inclusion in cohort A. The 449 (416 scar, 33 no scar) children included in cohort A before being examined as part of the cohort B household survey were excluded from cohort B in order not to replicate results. If the scar status did not correlate between the two assessments made in the two different cohorts, the scar assessment relevant for the follow-up time was used. Children were included in the survival analysis on the day of scar reading or assessment of the vaccination card, whichever came last, and were excluded on the day of moving from the study area or 365 days after scar reading, whichever came first. One child (no scar) dying in an accident was excluded from the analysis. Thus, 1-year mortality from the time of scar reading and onward was described for 1167 (794 scar, 373 no scar) children in cohort B in a Cox regression analysis with age as underlying time (Figure 1). A combined estimate of 1-year mortality for both cohort A and B was based on 2693 person years of observation, not replicating any follow-up time is also presented (Figure 1).
|
The survival analyses controlled for background factors with possible influence on child mortality including sex, year of birth (1993, 1994, 1995, 1996–97), season of vaccination (rainy: July–November or dry:December–June), district of residence (Bandim 1 or Bandim 2), place of birth (home, the central hospital, or the Bandim health centre), electricity in house (yes or no), schooling of mother (0, 1–4 or
5 years), ethnic group (Pepel, or other ethnic group), and age at vaccination (0–30 days, 31 days or older). In univariable and multivariable Cox regression analyses, these background factors were tested for association with mortality in the joint cohort of A and B, and factors with a P-value of <0.20 in the multivariable analysis were adjusted for in all multivariable analyses. TB infection had no impact on the results in cohort A,20 and to assess the possible impact of TB infection on our estimate in cohort B, we excluded children with TB exposure at home in the 6 months prior to scar assessment. |
Specific mortality as assessed by VA |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
Of the children with BCG vaccination at least one month before scar assessment, 488 children were included in both cohort A (1813) and B (1617) resulting in a total of 2942 children (2434 scar, 508 no scar) eligible for assessment of different causes of mortality for scar vs no scar. Within the year following scar reading 149 children died (116 scar, 33 no scar). A VA was developed on the basis of a standard VA recommended by WHO28 and adjusted to the local culture and language through discussions and trials together with physicians and a midwife used to perform VA in the study area.29,30 Once the war was over in Guinea-Bissau, a Guinean physician conducted the VA by visiting the relatives of the deceased children between June 2000 and June 2002, the median time from death to interview being 2.4 years. In order to be considered a reliable informant to the VA, the relative had to have been a member in the same household as the deceased child during the time of the child's disease and death. Of the 149 deaths, we were able to conduct a VA for 77% (90 scar, 25 no scar); for the remaining 34 deaths (26 scar, 8 no scar), all close relatives had moved or died when screening started. The physician conducting the VA was blind to whether the child had had a BCG scar or not. When the 115 VAs were carried out, a group consisting of the interviewing doctor, a paediatrician and another medical doctor, being blind to scar status of the children, decided on one of the following diagnoses as probable primary cause of death; accident/war, anaemia, malnutrition, acute diarrhoea, chronic diarrhoea, pneumonia, malaria, measles, meningitis/encephalitis, other disease, fever unclassified and cause unclassified. To simplify presentation, we categorized the diagnoses that contributed <5% to the total causes of deaths as ‘other disease/cause’, except for one death caused by accident, which was excluded. The expert group deciding on the most probable cause of death took into consideration the whole patient history that could be derived from the VA interview, including duration of a number of symptoms, duration and kind of treatment provided or stopped, and diagnostic tests reported (such as a blood smear test for malaria). The causes of death contributing to
5% of the total number of deaths were: malaria (fever, convulsions, affected responses of the child), pneumonia (cough, difficult breathing, fast breathing, chest in-drawings), acute diarrhoea (liquid and watery stools <14 days, dehydration), chronic diarrhoea (liquid and watery stools 15 days), and meningitis/encephalitis (fever, stiff neck/bulging fontanelle). To demonstrate whether there was a different distribution of scar status among the dead children in whom we did not succeed in performing a VA, these deaths were presented as ‘no VA’. We calculated the risk of dying from each specific disease, according to scar status in the year following scar assessment. This was done through a separate dataset for each diagnosis, where the event was death and where follow-up was terminated at death due to all other causes, thereby creating a cause-specific hazard function.31 Hence, the risk of dying from a particular disease was described in different Cox regression analyses with age as underlying time, one for each major category of disease. |
Statistical software |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
The analyses were done using Stata 7 for Microsoft Windows.
|
Results |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
Scar distribution
There were different distributions of BCG scars in the two cohorts. Among the children with a documented BCG vaccination at least 1 month before scar reading, there were 92% children (1676/1813) with a scar in cohort A and 72% (1159/1617) with a scar in cohort B. Among the children without documented BCG vaccination, 30% (13/43) in cohort A and 55% (31/56) in cohort B had a scar. The median time interval between BCG vaccination and scar reading was 175 days (25–75th percentile: 167–181) in cohort A and 956 days (25–75th percentile: 544–1329) in cohort B. The scar prevalence was more similar when the two cohorts were stratified for time interval between vaccination and scar reading (Table 1).
|
Mortality
In a multivariable Cox regression analysis of both cohort A and B, sex and ethnic group were associated with mortality (P < 0.20, data available on request). During the 12 months following scar reading, 27 of the 1167 children died in cohort B (Table 2). Adjusting for sex and ethnic group, there was lower mortality for children with a scar, the MR being 0.45 (95% CI 0.21–0.96) (Table 2). Taking TB exposure into account, excluding 165 children from the survival analysis and shortening observation time for 30 children, the MR was 0.37 (95% CI 0.15–0.92) (Table 2). The effect of having a BCG scar was not significantly different for boys and girls (Table 2) (test of homogeneity P = 0.281).
|
Combining cohorts A and B, 148 children (116 scar, 32 no scar) died in the 12 months following scar reading, and having a BCG scar was associated with a MR of 0.43 (95% CI 0.28–0.65), adjusting for sex and ethnic group. The effect was largest for children examined in the first year of life (Figure 2). However, in spite of including only children examined after the first year of life, the MR was still 0.47 (95% CI 0.21–1.04) after adjusting for significant background factors.
|
Specific mortality as assessed by VA
There was no differential effect of having BCG scar for different causes of death (Table 3). Adjusting for sex and ethnic group, the cause-specific mortality rate of malaria was significantly lower for children with a BCG scar than without: MR being 0.32 (95% CI 0.13–0.76). Even if deaths due to malaria were combined with deaths due to unclassified fever (as these might well be due to malaria), the mortality rate was still significantly lower for scar-positive children [MR 0.42 (95% CI 0.21–0.82)].
|
|
Discussion |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
In this study we have shown that the mortality in the year following scar assessment was more than halved for children with a BCG scar compared with children without (Table 2). This is comparable with earlier findings.20 The observation would suggest that it is important to monitor BCG scarring among children in high mortality areas. The general distribution of causes of deaths corresponded to what have been described previously from Guinea-Bissau.29,30 There were no major differences in the impact of BCG scarring for different causes of deaths. The present study had limited power to detect differences between causes of death due to the small number of children without a scar, but nothing suggested that such differences might be important. The numbers were too small for a comparison of the potential impact of BCG scarring on the distribution of causes of deaths in the two cohorts, but then again nothing indicated differences of importance between the cohorts (data available on request). The MR in the combined cohort for children with a BCG scar was significantly lower only for malaria, which was the largest group of deaths together with ‘fever unclassified’. While reasoning about different effects of BCG on particular diseases, the limitation of the VA method in determining an exact cause of death should also be stressed. Even though the recall period was delayed due to the outbreak of civil war, the physician carrying out the VA did not find the relatives of the deceased child to have difficulties in remembering the events leading to the death of the child. This observation cannot be validated and despite maternal recall of death of children having been considered quite reliable even years after the child's death,32 the delay is a potential limitation of the present study. Few studies on the validity of VA have been done, and the studies that have been made vary in method, setting and population and it has been reported that the method of VA varies in specificity and sensitivity for different causes of death.33 Hence, the question of what diseases are most affected by having a BCG scar remains unsolved. However, in settings such as that of the present study, VA may often be the best method available and it is used to provide approximate values for relatively common causes of death.28 We therefore believe that our study suggests that the effect of a BCG scar may be related to a general enhancement of the immune response to many infections rather than providing protection against specific infections. This possibility is supported by studies of the impact of vaccination status on the case-fatality rate at the paediatric ward in Bissau; non-specific effects of measles and DTP vaccines were documented, but there were no differences in the causes of deaths for vaccinated and unvaccinated children.34
There were differences in the prevalence of BCG scar between the two cohorts. These could presumably be explained by the different study procedures since scar assessment could vary between readers; in one study a medical doctor knowing the BCG vaccination status assessed the scar as part of a medical examination in a health centre while in the other study, trained research assistants performed the scar reading in the homes of the children often without access to the vaccination card of the child. BCG scar prevalence varies with age of vaccination and time of scar reading after vaccination.12 For BCG vaccinations given within one month of birth, the prevalence of a recognizable BCG scar will decline over time.13 This could only partly explain the lower prevalence of scarring in cohort B (Table 1). Of the 44 children with a BCG scar despite no documented BCG vaccination, 20% were probably due to a rapid scar reaction since vaccination had taken place within 30 days before BCG scar assessment; the remaining were due to failure to mark BCG vaccination on the vaccination card, misclassification of BCG scar or misclassification of BCG vaccination while inspecting the vaccination card. Since BCG vaccination took place in different years for the different cohorts, the differences in scar prevalence may also be related to changes in vaccination technique, vaccine type, vaccinating nurse, or lack of BCG vaccine during certain periods, resulting in delayed vaccination; we have observed that all of these factors may have an impact on BCG scarring (authors' unpublished data).
When assessing the relationship between mortality and BCG scar responses, HIV-1 could be a major confounder considering that HIV-1 infection has been shown to suppress scar reaction to BCG.35,36 During the study period (1993–98), HIV-1 infection was increasing in the study area, from 0% in the late 80s37 among individuals 15 years of age to 2.2% in 1996.38 Expecting a vertical transmission rate of 25%,39,40 we estimate that <1% of the children may have been HIV-1-infected in the present study. Considering the numerous beneficial effects on survival with a BCG scar observed in our study and the relatively low prevalence of HIV-1 infection, we reason that even if HIV-1 infection may have had a modifying effect on our results, it could not explain the major survival benefit observed to be associated with a BCG scar. It was shown consistently in cohort A, that a positive tuberculin skin test (TST) was strongly associated with better survival in the 12 months following TST testing and that excluding deaths of children seropositive to HIV-1 did not change the estimates.20
One could argue that the results of the present study may be skewed as early infant deaths (<3 months) were not included. However, this ought not to be so, unless BCG has a negative effect on survival for the weak and young children which recent findings indicate is not the case.41 Also, since all children included in the survival analyses were BCG vaccinated, confounding due to healthier children being more likely to be vaccinated is not applicable. Instead, it could be speculated that the beneficial effect of having a BCG scar was merely a marker of some children having stronger immune responses and therefore better survival. However, our studies in Guinea-Bissau have clearly found that BCG scarring depends on type of vaccine and vaccination technique. Furthermore, BCG vaccination has been found to be associated with non-specific enhancement of both antibody and cellular immune responses.23,24 Hence, the beneficial effect may in fact be related to the immune response induced by a correctly given BCG vaccination. Future studies should examine the ability of scar-negative children to develop a scar through re-vaccination. This may well be possible since the lack of scarring could be a result of faulty vaccination technique. If the immune stimulation producing a scar, upon re-vaccination, is comparable with the stimulation producing a scar after a primary vaccination, re-vaccination of scar-negative children might reduce infant mortality substantially, depending on the prevalence of scar-negative children. Of interest is also the reason for scar-failure: clearly a child never developing a scar after BCG vaccination may differ from a child with a waned scar reaction. This is also a possible uncontrolled confounder in the present study, although probably not of major importance considering the relatively low rate of waning of the scar reaction (Table 1).
Larger studies relating BCG vaccination, BCG scar, and other markers of BCG vaccination to specific morbidity, immune profiles, and mortality are warranted in order to determine the immunological mechanisms involved and to assess possible modifications in the current vaccination programme. If BCG has non-specific beneficial effects it should be recommended to re-vaccinate children without a BCG scar, especially considering that the beneficial effect of a BCG scar on survival seems to persist through early childhood.
KEY MESSAGES
|
|
Acknowledgments |
|---|
The study received financial support from the Novo Nordisk Foundation, the Danish Council for Development Research, Danish Medical Research, DANIDA and The Graduate Research School of International Health at the University of Copenhagen, supported by the Danish Research Academy. PA holds a research professorship grant from the Novo Nordisk Foundation. The VAs were designed and carried out in co-operation with Gunnar Nylén (MD), doing studies of mortality patterns during an armed conflict in Guinea-Bissau.
A.R., P.A. and M.S. planned the study; P.G. conducted the scar survey in cohort B and supervised the tuberculosis-surveillance programme; A.N., M.S. and A.R. constructed and conducted the VAs; A.P., A.N. and A.R. assessed the VA conducted; M.L.G. carried out the scar survey and survival analyses in cohort A; Q.D., P.A., Am.R., and A.R. did necessary extra follow-up and data cleaning; A.R., P.A., and H.J. performed the analyses; A.R. did the first draft of the paper and the all authors contributed to the final version of the paper. A.R. is guarantor for the paper.
|
Notes |
|---|
* This presentation combines cohort A (presented earlier 17) and cohort B, as described in detail in the paper. Children were followed-up for 12 months after inclusion (scar reading) and age is the underlying time since mortality is highly dependent on age in these age groups.
After exclusion of one child due to death in an accident, as discovered by VA, 2434 children with a scar and 507 without a scan were followed-up for survival 12 months after scar assessment.
|
References |
|---|
|
TopAbstractMethod and subjectsBCG vaccinationTB surveillanceMortalitySpecific mortality as assessed...Statistical softwareResultsDiscussionReferences |
|---|
1 Clarke A, Rudd P. Neonatal BCG immunisation. Arch Dis Child 1992;67:473–74.
2 Houston S, Fanning A, Soskolne CL, Fraser N. The effectiveness of bacillus Calmette-Guerin (BCG) vaccination against tuberculosis. A case–control study in Treaty Indians, Alberta, Canada. Am J Epidemiol 1990;131:340–48.
3 Price JF. BCG vaccination. Arch Dis Child 1982;57:485–86.
4 Tidjani O, Amedome A, ten Dam HG. The protective effect of BCG vaccination of the newborn against childhood tuberculosis in an African community. Tubercle 1986;67:269–81.[CrossRef][Web of Science][Medline]
5 Colditz GA, Berkey CS, Mosteller F et al. The efficacy of bacillus Calmette-Guerin vaccination of newborns and infants in the prevention of tuberculosis: meta-analyses of the published literature. Pediatrics 1995;96:29–35.
6 ten Dam HG, Hitze KL. Does BCG vaccination protect the newborn and young infants? Bull World Health Organ 1980;58:37–41.[Web of Science][Medline]
7 Levine MI, Sackett MF. Results of BCG immunization in New York city. Am Rev Tuberc 1946; 53:517–32.[Web of Science]
8 Kabir Z. Is all-cause mortality a useful epidemiological endpoint in vaccine trials? An example of BCG (Bacille-Calmette-Guerine). Int J Epidemiol 2003;32:161–65.
9 Kristensen I, Aaby P, Jensen H. Routine immunizations and child survival in Guinea-Bissau, West-Africa. BMJ 2000;321:1435–38.
10 Niobey FM, Duchiade MP, Vasconcelos AG, de Carvalho ML, Leal MC, Valente JG. [Risk factors for death caused by pneumonia in children younger than 1 year old in a metropolitan region of southeastern Brazil. A case–control study.] Rev Saude Publica 1992;26:229–38.[Web of Science][Medline]
11 Velema JP, Alihonou EM, Gandaho T, Hounye FH. Childhood mortality among users and non-users of primary health care in a rural west African community. Int J Epidemiol 1991;20:474–79.
12 Fine PEM, Ponnighaus JM, Maine N. The distribution and implications of BCG scars in northern Malawi. Bull World Health Organ 1989;67:35–42.[Web of Science][Medline]
13 Floyd S, Ponnighaus JM, Bliss L et al. BCG scars in northern Malawi: sensitivity and repeatability of scar reading, and factors affecting scar size. Int J Tuberc Lung Dis 2000;4:1133–42.[Web of Science][Medline]
14 Rani SH, Vijayalakshmi V, Sunil K, Lakshmi KA, Suman LG, Murthy KJ. Cell mediated immunity in children with scar-failure following BCG vaccination. Indian Pediatr 1998;35:123–27.[Medline]
15 Santosa G, Syamsuri MM, Gusti I et al. Difference in severity of tuberculosis in children with or without a BCG scar. Paediatr Indones 1985;25:87–92.[Medline]
16 Sterne JA, Fine PE, Ponnighaus JM, Sibanda F, Munthali M, Glynn JR. Does bacille Calmette-Guerin scar size have implications for protection against tuberculosis or leprosy? Tuber Lung Dis 1996;77:117–23.[CrossRef][Web of Science][Medline]
17 Jason J, Archibald LK, Nwanyanwu OC et al. Clinical and Immune Impact of Mycobacterium bovis BCG Vaccination Scarring. Infect Immun 2002;70:6188–95.
18 Edwards LB, Palmer CE, Magnus K. BCG vaccination. WHO Monograph Series 1953;12:.
19 ten Dam HG. Research on BCG vaccination. Adv Tuberc Res 1984;21:79–106.[Medline]
20 Garly ML, Martins CL, Bale C et al. BCG scar and positive tuberculin reaction associated with reduced child mortality in West-Africa. A non-specific beneficial effect of BCG? Vaccine 2003;21:2782–90.[CrossRef][Web of Science][Medline]
21 Marchant A, Goetghebuer T, Ota MO et al. Newborns develop a Th1-type immune response to Mycobacterium bovis bacillus Calmette-Guerin vaccination. J Immunol 1999;163:2249–55.
22 Aaby P, Shaheen S, Heyes C et al. Early BCG vaccination and reduction in atopy in Guinea-Bissau. Clin Exp Allergy 2000;30:644–50.[CrossRef][Web of Science][Medline]
23 Garly ML, Bale C, Martins CL et al. BCG vaccination among West African infants is associated with less anergy to tuberculin and diphtheria-tetanus antigens. Vaccine 2001;20::468–74.[CrossRef][Web of Science][Medline]
24 Ota MO, Vekemans J, Schlegel-Haueter SE et al. Influence of Mycobacterium bovis Bacillus Calmette-Guérin on Antibody and Cytokine Responses to Human Neonatal Vaccination. J Immunol 2002;168:919–25.
25 Aaby P. Bandim: An Unplanned Longitudinal Study. In: Gupta MD, Aaby P, Garenne M, Pison G (eds). Prospective Community Studies in Developing Countries. Oxford: Clarendon Press, 1997, pp. 277–96.
26 Gustafson P, Gomes VF, Vieira CS et al. Tuberculosis in Bissau: incidence and risk factors in an urban community in sub-Saharan Africa. Int J Epidemiol 2004;33:163–72.
27 Enarson DA, Rieder HL, Arnadottir T, Trébucq A. Management of tuberculosis. A guide for low income countries. (5th edn). Paris: International Union Against Tuberculosis and Lung Disease, 2000.
28 Anker M, Black RE, Coldham C et al. A Standard Verbal Autopsy Method for Investigating Causes of Death in Infants and Children. WHO, 2002.
29 Molbak K, Aaby P, Ingholt L et al. Persistent and acute diarrhoea as the leading causes of child mortality in urban Guinea Bissau. Trans R Soc Trop Med Hyg 1992;86:216–20.[CrossRef][Web of Science][Medline]
30 Sodemann M, Jakobsen MS, Molbak K, Alvarenga IC, Jr, Aaby P. High mortality despite good care-seeking behaviour: a community study of childhood deaths in Guinea-Bissau. Bull World Health Organ 1997;75:205–12.[Web of Science][Medline]
31 Prentice RL, Kalbfleisch JD, Peterson AV, Jr, Flournoy N, Farewell VT, Breslow NE. The analysis of failure times in the presence of competing risks. Biometrics 1978;34:541–54.[CrossRef][Web of Science][Medline]
32 Rao MR, Levine RJ, Wasif NK, Clemens JD. Reliability of maternal recall and reporting of child births and deaths in rural Egypt. Paediatr Perinat Epidemiol 2003;17:125–31.[CrossRef][Web of Science][Medline]
33 Chandramohan D, Setel P, Quigley M. Effect of misclassification of causes of death in verbal autopsy: can it be adjusted? Int J Epidemiol 2001;30:509–14.
34 Veirum JE, Sodemann M, Biai S et al. Diphtheria-tetanus-pertussis and measles vaccinations associated with divergent effects on female and male mortality at the paediatric ward in Bissau, Guinea-Bissau. Vaccine 2004; (in press).
35 Anonymous. BCG vaccination and pediatric HIV infection—Rwanda, 1988–1990. MMWR 1991;40:833–36.[Medline]
36 Ota MO, O'Donovan D, Marchant A et al. HIV-negative infants born to HIV-1 but not HIV-2-positive mothers fail to develop a Bacillus Calmette-Guerin scar. AIDS 1999;13:996–98.[CrossRef][Web of Science][Medline]
37 Poulsen AG, Kvinesdal B, Aaby P. Prevalence of and mortality from human immunodeficiency virus type 2 in Bissau, West Africa. Lancet 1989;1:827–31.[CrossRef][Web of Science][Medline]
38 Larsen O, da Silva Z, Sandstrom A et al. Declining HIV-2 prevalence and incidence among men in a community study from Guinea-Bissau. AIDS 1998;12:1707–14.[CrossRef][Web of Science][Medline]
39 Adjorlolo-Johnson G, De Cock KM, Ekpini E et al. Prospective comparison of mother-to-child transimission of HIV-A and HIV-2 in Abidjan, Ivory Coast. JAMA 1994;272:462–66.
40 Prazuck T, Yameogo JM, Heylinck B et al. Mother-to-child transmission of human immunodeficiency virus type 1 and type 2 and dual infection: a cohort study in Banfora, Burkina Faso. Pediatr Infect Dis J 1995;14:940–47.[Web of Science][Medline]
41 Roth A, Jensen H, Garly M, Lisse IM, Sodemann M, Aaby P. Low birth weight infants and Calmette-Guerin bacillus vaccination at birth: community study from Guinea-Bissau. Pediatr Infect Dis J 2004;23:544–50.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. S. Mathurin, G. W. Martens, H. Kornfeld, and R. M. Welsh CD4 T-Cell-Mediated Heterologous Immunity between Mycobacteria and Poxviruses J. Virol., April 15, 2009; 83(8): 3528 - 3539. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. R. Page, A. E. Jedlicka, B. Fakheri, G. S. Noland, A. K. Kesavan, A. L. Scott, N. Kumar, and Y. C. Manabe Mycobacterium-Induced Potentiation of Type 1 Immune Responses and Protection against Malaria Are Host Specific Infect. Immun., December 1, 2005; 73(12): 8369 - 8380. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||