- Email: [email protected]
Simplified antibiotic regimens for neonatal sepsis—AFRINEST
Correspondence
workers might address the issue of specificity in the diagnosis of severe bacterial infection and should be assessed in a clinical trial. If effective, this C-reactive protein-guided pathway would reduce the risk of antibiotic resistance and complement the use of simplified antibiotics regimens at the community and primary health-care level. We declare no competing interests.
*Juan Emmanuel Dewez, Harish Kumar Chellani, Abdul Halim, Nynke van den Broek [email protected] Centre for Maternal and Newborn Health, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK (JED, NvdB); Department of Paediatrics, Safdarjung Hospital and Associated Vardhman Mahavir Medical, New Delhi, India (HKC); and Reproductive and Child Health Division, Centre for Injury Prevention & Research Bangladesh, Mohakhali, Dhaka, Bangladesh (AH) 1
2
3
4
5
6
See Online for appendix
1338
7
African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Simplified antibiotic regimens compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with clinical signs of possible serious bacterial infection when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1767–76. Seale AC, Blencowe H, Manu AA, et al. Estimates of possible severe bacterial infection in neonates in sub-Saharan Africa, south Asia, and Latin America for 2012: a systematic review and meta-analysis. Lancet Infect Dis 2014; 14: 731–41. Lee AC, Chandran A, Herbert HK, et al. Treatment of infections in young infants in low- and middle-income countries: a systematic review and meta-analysis of frontline health worker diagnosis and antibiotic access. PLoS Med 2014; 11: e1001741. WHO. Antimicrobial resistance; global report on surveillance. Geneva: World Health Organization, 2014. http://www.int/ drugresistance/documents/surveillancereport/ en/ (accessed April 10, 2015). NICE. Antibiotics for early-onset neonatal infection: antibiotics for the prevention and treatment of early-onset neonatal infection. 2012. http://www.nice.org.uk/guidance/ cg149/chapter/key-priorities-forimplementation#/investigations-beforestarting-antibiotics-in-the-baby (accessed March 5, 2015). Papa F, Rongioletti M, Majolini MB, et al. Fast bedside measurement of blood count and C-reactive protein in newborns compared with conventional methods. Clin Lab 2012; 58: 951–57. Aydin B, Dilli D, Zenciroglu A, et al. Comparison of a rapid bed-side test with a central laboratory analysis for C-reactive protein in newborn infants with suspicion of sepsis. Clin Cancer Res 2013; 59: 1045–51.
The two AFRINEST field trials1,2 of simplified antibiotic regimens for suspected bacterial infections in neonates and young infants successfully showed that a team of community health workers and nurses can deliver antibiotic treatment, including injectable drugs, very safely and with high coverage in the community. Thus, the investigators confirm that the home-based management of neonatal infections with antibiotics, which we first reported,3 is feasible in programme conditions. We congratulate the authors. However, we disagree with the conclusion that the various antibiotic regimens tested were equally effective. They were probably all equally unnecessary. The babies they treated were a selectively low-risk group, mostly false positive cases of neonatal sepsis or pneumonia, because of three factors. First, 67% of neonates were enrolled after the period of major risk of neonatal mortality (0–6 days) had passed. Neonates weighing more than 1500 g were excluded. Second, 3038 (27%) of the 11 153 ill babies were either referred or admitted to hospital or excluded as critically ill.1 Third, of the 3564 babies that were treated for bacterial infection, 3111 (87%) babies had only one clinical sign of infection.1 The use of only one clinical sign to diagnose sepsis results in the inclusion of a large proportion of false positives,4 and therefore we have previously suggested that the presence of minimum two clinical signs should be used to diagnose bacterial infection, thereby reducing the chance of false positivity.4 In the second trial,2 2333 treated babies had only fast breathing.2 The incidence of only fast breathing without other signs of bacterial infection in the community-based data on neonates in the Gadchiroli district in Maharashtra, India, during 1995–2014 was 138 (1%) of 14 767 babies (appendix). When these neonates were treated with oral co-trimoxazole, the case fatality was zero, similar to the very low (<1%) case fatality reported
when neonates were treated with amoxicillin or procaine penicillin plus gentamicin by the AFRINEST group.2 We suspect that the fast breathing alone in neonates is an innocuous sign which might not need any antibiotic. How many of the 5897 babies treated in the two AFRINEST trials really needed antibiotics? Of the 763 neonates observed in the community in the pre-intervention year (1995–96) in Gadchiroli,4 we find that by applying the AFRINEST methods, 329 (43%) neonates would be identified as having bacterial infection, but only 19 of these cases would be true positive. The methods of the two AFRINEST studies1,2 would have identified 44% (43%+1%) of neonates in the community for treatment with antibiotics whereas only 3% (22 of 763 neonates) of the neonates in the community needed it.4 No wonder all antibiotic regimens compared showed equal and low treatment failure and a very low case fatality. In the absence of an untreated control group—not possible for ethical reasons—these trials proved the feasibility and safety but not the efficacy of the antibiotic regimens tested. In view of the growing global concern about antibiotic overuse and resistance,5 it will be desirable to use more rigorous criteria to selectively treat those at high risk of death. We declare no competing interests.
*Abhay T Bang, Mahesh Deshmukh [email protected] Society for Education, Action and Research in Community Health (SEARCH), Gadchiroli, India 1
2
African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Oral amoxicillin compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with fast breathing when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1758–66. African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Simplified antibiotic regimens compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with clinical signs of possible serious bacterial infection when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1767–76.
www.thelancet.com Vol 386 October 3, 2015
Correspondence
4
5
Bang AT, Bang RA, Baitule SB, et al. Effect of home-based neonatal care and management of sepsis on neonatal mortality: field trial in rural India. Lancet 1999; 354: 1955–61. Bang AT, Bang RA, Reddy MH, et al. Simple clinical criteria to identify sepsis or pneumonia in neonates in the community needing treatment or referral. Pediatr Infect Dis J 2005; 24: 335–41. Laxminarayan R, Duse A, Wattal C, et al. Antibiotic resistance—the need for global solutions. Lancet Infect Dis 2013; 13: 1057–98.
Authors’ reply We thank our colleagues for their remarks on our studies. In the AFRINEST studies,1,2 done in a programme setting, we assessed regimens for outpatient treatment of young infants with clinical signs of severe infection whose families could not comply with advice to admit their child to hospital. Hospital treatment of such infants remains the treatment of choice, and about half of infants with signs of severe infection who were referred to hospital were not the target of this intervention. In our study context, we could not assess aetiology. Although not all enrolled infants would have bacterial infections, differentiating between bacterial and viral infections is challenging in the best hospitals, and antibiotic treatment is initiated on the basis of clinical signs, even in high-income countries.3 Blood cultures for infectious organisms have low yields and do not provide immediate results for use by clinicians. Screening tests, such as serial quantitative C-reactive protein measurements, must be assessed further before use becomes widespread. Neither test can be easily applied in the population of newborn babies whose families cannot accept that their infant must be admitted in primary care settings in developing countries. Gary Darmstadt and colleagues4 found that the Integrated Management of Childhood Illness (IMCI) algorithm, which requires the presence of one of seven signs to identify infants with possible serious bacterial infection, had a better balance between sensitivity and specificity (56·5% and 93·1%, respectively) than other algorithms, including those cited by Juan Emmanuel Dewez and colleagues. www.thelancet.com Vol 386 October 3, 2015
We therefore chose a diagnostic approach consistent with IMCI. We enrolled all 0–59-day-old infants with signs of possible serious bacterial infection that met our eligibility criteria, and one-third of the enrolled infants were in their first week of life, which reflects the prevailing pattern of morbidity in the populations. Mortality is indeed very high in the first week after birth, but the predominant causes of death are preterm birth complications and asphyxia.5 We classified infants with possible serious bacterial infection into three groups on the basis of clinical signs of increasing severity: fast breathing alone, signs of severe infection (stopped feeding well, severe chest indrawing, fever, hypothermia, and no spontaneous movements), and signs of critical illness such as unconsciousness and convulsions. The latter group was excluded from the study for ethical reasons, and we believe that this should be done in any future programmes. We show that the first group of infants can be safely and effectively treated with oral antibiotics, and the second group of infants can be treated with a combination of oral and once-daily intramuscular antibiotics. We agree that the mortality was low in enrolled young infants who all received early and assured treatment. Although it is possible that the low mortality could be due to exclusion of infants who were critically ill or who had very low birthweight, or because a proportion of enrolled infants did not have bacterial sepsis, it is equally probable that this low mortality was the result of early and effective treatment and close follow-up. Abhay Bang and Mahesh Deshmukh suggest that antibiotic treatment of sick young infants, such as those included in our study, was probably unnecessary. We believe that not treating such infants would be a very risky decision. A randomised, placebo-controlled trial (NCT01533818) comparing oral amoxicillin with placebo in young infants with fast breathing alone (the least severe illness group in our study)
was terminated early by the data safety monitoring board because of high treatment failure with placebo relative to oral amoxicillin. We declare no competing interests.
*Fabian Esamai, Adejumoke I Ayede, Ebunoluwa A Adejuyigbe, Antoinette Tshefu, Robinson D Wammanda, and the African Neonatal Sepsis Trial (AFRINEST) group
Ton Koene/dpa/Corbis
3
[email protected] Department of Child Health and Paediatrics, School of Medicine, Moi University, Eldoret, Kenya (FE); Department of Paediatrics, College of Medicine, University of Ibadan, and University College Hospital, Ibadan, Nigeria (AIA); Department of Paediatrics and Child Health, Obafemi Awolowo University, Ile-Ife, Nigeria (EAA); Department of Community Health, Kinshasa School of Public Health, Kinshasa, DR Congo (AT); and Department of Paediatrics, Ahmadu Bello University Teaching Hospital, Ahmadu Bello University, Zaria, Nigeria (RDW) 1
2
3
4
5
African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Oral amoxicillin compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with fast breathing when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1758–66. African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Simplified antibiotic regimens compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with clinical signs of possible serious bacterial infection when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1767–76. Cohen J, Vincent JL, Adhikari NK, et al. Sepsis: a roadmap for future research. Lancet Infect Dis 2015; 15: 581–614. Darmstadt GL, Baqui AH, Choi Y, et al. Validation of clinical algorithm to identify neonates with severe illness during routine household visits in rural Bangladesh. Arch Dis Child 2011; 96: 1140–46. Saleem S, McClure EM, Goudar SS, et al, and the Global Network Maternal Newborn Health Registry Study Investigators. A prospective study of maternal, fetal and neonatal deaths in low- and middle-income countries. Bull World Health Organ 2014; 92: 605–12.
“Off the pulse” about incidental atrial fibrillation In their Comment, Brian Potter and Jaques LeLorier ( July 11, p 113)1 criticised the authors of the Framingham Heart Study2 on trends in atrial fibrillation for concluding 1339
workers might address the issue of specificity in the diagnosis of severe bacterial infection and should be assessed in a clinical trial. If effective, this C-reactive protein-guided pathway would reduce the risk of antibiotic resistance and complement the use of simplified antibiotics regimens at the community and primary health-care level. We declare no competing interests.
*Juan Emmanuel Dewez, Harish Kumar Chellani, Abdul Halim, Nynke van den Broek [email protected] Centre for Maternal and Newborn Health, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK (JED, NvdB); Department of Paediatrics, Safdarjung Hospital and Associated Vardhman Mahavir Medical, New Delhi, India (HKC); and Reproductive and Child Health Division, Centre for Injury Prevention & Research Bangladesh, Mohakhali, Dhaka, Bangladesh (AH) 1
2
3
4
5
6
See Online for appendix
1338
7
African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Simplified antibiotic regimens compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with clinical signs of possible serious bacterial infection when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1767–76. Seale AC, Blencowe H, Manu AA, et al. Estimates of possible severe bacterial infection in neonates in sub-Saharan Africa, south Asia, and Latin America for 2012: a systematic review and meta-analysis. Lancet Infect Dis 2014; 14: 731–41. Lee AC, Chandran A, Herbert HK, et al. Treatment of infections in young infants in low- and middle-income countries: a systematic review and meta-analysis of frontline health worker diagnosis and antibiotic access. PLoS Med 2014; 11: e1001741. WHO. Antimicrobial resistance; global report on surveillance. Geneva: World Health Organization, 2014. http://www.int/ drugresistance/documents/surveillancereport/ en/ (accessed April 10, 2015). NICE. Antibiotics for early-onset neonatal infection: antibiotics for the prevention and treatment of early-onset neonatal infection. 2012. http://www.nice.org.uk/guidance/ cg149/chapter/key-priorities-forimplementation#/investigations-beforestarting-antibiotics-in-the-baby (accessed March 5, 2015). Papa F, Rongioletti M, Majolini MB, et al. Fast bedside measurement of blood count and C-reactive protein in newborns compared with conventional methods. Clin Lab 2012; 58: 951–57. Aydin B, Dilli D, Zenciroglu A, et al. Comparison of a rapid bed-side test with a central laboratory analysis for C-reactive protein in newborn infants with suspicion of sepsis. Clin Cancer Res 2013; 59: 1045–51.
The two AFRINEST field trials1,2 of simplified antibiotic regimens for suspected bacterial infections in neonates and young infants successfully showed that a team of community health workers and nurses can deliver antibiotic treatment, including injectable drugs, very safely and with high coverage in the community. Thus, the investigators confirm that the home-based management of neonatal infections with antibiotics, which we first reported,3 is feasible in programme conditions. We congratulate the authors. However, we disagree with the conclusion that the various antibiotic regimens tested were equally effective. They were probably all equally unnecessary. The babies they treated were a selectively low-risk group, mostly false positive cases of neonatal sepsis or pneumonia, because of three factors. First, 67% of neonates were enrolled after the period of major risk of neonatal mortality (0–6 days) had passed. Neonates weighing more than 1500 g were excluded. Second, 3038 (27%) of the 11 153 ill babies were either referred or admitted to hospital or excluded as critically ill.1 Third, of the 3564 babies that were treated for bacterial infection, 3111 (87%) babies had only one clinical sign of infection.1 The use of only one clinical sign to diagnose sepsis results in the inclusion of a large proportion of false positives,4 and therefore we have previously suggested that the presence of minimum two clinical signs should be used to diagnose bacterial infection, thereby reducing the chance of false positivity.4 In the second trial,2 2333 treated babies had only fast breathing.2 The incidence of only fast breathing without other signs of bacterial infection in the community-based data on neonates in the Gadchiroli district in Maharashtra, India, during 1995–2014 was 138 (1%) of 14 767 babies (appendix). When these neonates were treated with oral co-trimoxazole, the case fatality was zero, similar to the very low (<1%) case fatality reported
when neonates were treated with amoxicillin or procaine penicillin plus gentamicin by the AFRINEST group.2 We suspect that the fast breathing alone in neonates is an innocuous sign which might not need any antibiotic. How many of the 5897 babies treated in the two AFRINEST trials really needed antibiotics? Of the 763 neonates observed in the community in the pre-intervention year (1995–96) in Gadchiroli,4 we find that by applying the AFRINEST methods, 329 (43%) neonates would be identified as having bacterial infection, but only 19 of these cases would be true positive. The methods of the two AFRINEST studies1,2 would have identified 44% (43%+1%) of neonates in the community for treatment with antibiotics whereas only 3% (22 of 763 neonates) of the neonates in the community needed it.4 No wonder all antibiotic regimens compared showed equal and low treatment failure and a very low case fatality. In the absence of an untreated control group—not possible for ethical reasons—these trials proved the feasibility and safety but not the efficacy of the antibiotic regimens tested. In view of the growing global concern about antibiotic overuse and resistance,5 it will be desirable to use more rigorous criteria to selectively treat those at high risk of death. We declare no competing interests.
*Abhay T Bang, Mahesh Deshmukh [email protected] Society for Education, Action and Research in Community Health (SEARCH), Gadchiroli, India 1
2
African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Oral amoxicillin compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with fast breathing when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1758–66. African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Simplified antibiotic regimens compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with clinical signs of possible serious bacterial infection when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1767–76.
www.thelancet.com Vol 386 October 3, 2015
Correspondence
4
5
Bang AT, Bang RA, Baitule SB, et al. Effect of home-based neonatal care and management of sepsis on neonatal mortality: field trial in rural India. Lancet 1999; 354: 1955–61. Bang AT, Bang RA, Reddy MH, et al. Simple clinical criteria to identify sepsis or pneumonia in neonates in the community needing treatment or referral. Pediatr Infect Dis J 2005; 24: 335–41. Laxminarayan R, Duse A, Wattal C, et al. Antibiotic resistance—the need for global solutions. Lancet Infect Dis 2013; 13: 1057–98.
Authors’ reply We thank our colleagues for their remarks on our studies. In the AFRINEST studies,1,2 done in a programme setting, we assessed regimens for outpatient treatment of young infants with clinical signs of severe infection whose families could not comply with advice to admit their child to hospital. Hospital treatment of such infants remains the treatment of choice, and about half of infants with signs of severe infection who were referred to hospital were not the target of this intervention. In our study context, we could not assess aetiology. Although not all enrolled infants would have bacterial infections, differentiating between bacterial and viral infections is challenging in the best hospitals, and antibiotic treatment is initiated on the basis of clinical signs, even in high-income countries.3 Blood cultures for infectious organisms have low yields and do not provide immediate results for use by clinicians. Screening tests, such as serial quantitative C-reactive protein measurements, must be assessed further before use becomes widespread. Neither test can be easily applied in the population of newborn babies whose families cannot accept that their infant must be admitted in primary care settings in developing countries. Gary Darmstadt and colleagues4 found that the Integrated Management of Childhood Illness (IMCI) algorithm, which requires the presence of one of seven signs to identify infants with possible serious bacterial infection, had a better balance between sensitivity and specificity (56·5% and 93·1%, respectively) than other algorithms, including those cited by Juan Emmanuel Dewez and colleagues. www.thelancet.com Vol 386 October 3, 2015
We therefore chose a diagnostic approach consistent with IMCI. We enrolled all 0–59-day-old infants with signs of possible serious bacterial infection that met our eligibility criteria, and one-third of the enrolled infants were in their first week of life, which reflects the prevailing pattern of morbidity in the populations. Mortality is indeed very high in the first week after birth, but the predominant causes of death are preterm birth complications and asphyxia.5 We classified infants with possible serious bacterial infection into three groups on the basis of clinical signs of increasing severity: fast breathing alone, signs of severe infection (stopped feeding well, severe chest indrawing, fever, hypothermia, and no spontaneous movements), and signs of critical illness such as unconsciousness and convulsions. The latter group was excluded from the study for ethical reasons, and we believe that this should be done in any future programmes. We show that the first group of infants can be safely and effectively treated with oral antibiotics, and the second group of infants can be treated with a combination of oral and once-daily intramuscular antibiotics. We agree that the mortality was low in enrolled young infants who all received early and assured treatment. Although it is possible that the low mortality could be due to exclusion of infants who were critically ill or who had very low birthweight, or because a proportion of enrolled infants did not have bacterial sepsis, it is equally probable that this low mortality was the result of early and effective treatment and close follow-up. Abhay Bang and Mahesh Deshmukh suggest that antibiotic treatment of sick young infants, such as those included in our study, was probably unnecessary. We believe that not treating such infants would be a very risky decision. A randomised, placebo-controlled trial (NCT01533818) comparing oral amoxicillin with placebo in young infants with fast breathing alone (the least severe illness group in our study)
was terminated early by the data safety monitoring board because of high treatment failure with placebo relative to oral amoxicillin. We declare no competing interests.
*Fabian Esamai, Adejumoke I Ayede, Ebunoluwa A Adejuyigbe, Antoinette Tshefu, Robinson D Wammanda, and the African Neonatal Sepsis Trial (AFRINEST) group
Ton Koene/dpa/Corbis
3
[email protected] Department of Child Health and Paediatrics, School of Medicine, Moi University, Eldoret, Kenya (FE); Department of Paediatrics, College of Medicine, University of Ibadan, and University College Hospital, Ibadan, Nigeria (AIA); Department of Paediatrics and Child Health, Obafemi Awolowo University, Ile-Ife, Nigeria (EAA); Department of Community Health, Kinshasa School of Public Health, Kinshasa, DR Congo (AT); and Department of Paediatrics, Ahmadu Bello University Teaching Hospital, Ahmadu Bello University, Zaria, Nigeria (RDW) 1
2
3
4
5
African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Oral amoxicillin compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with fast breathing when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1758–66. African Neonatal Sepsis Trial (AFRINEST) group, Tshefu A, Lokangaka A, et al. Simplified antibiotic regimens compared with injectable procaine benzylpenicillin plus gentamicin for treatment of neonates and young infants with clinical signs of possible serious bacterial infection when referral is not possible: a randomised, open-label, equivalence trial. Lancet 2015; 385: 1767–76. Cohen J, Vincent JL, Adhikari NK, et al. Sepsis: a roadmap for future research. Lancet Infect Dis 2015; 15: 581–614. Darmstadt GL, Baqui AH, Choi Y, et al. Validation of clinical algorithm to identify neonates with severe illness during routine household visits in rural Bangladesh. Arch Dis Child 2011; 96: 1140–46. Saleem S, McClure EM, Goudar SS, et al, and the Global Network Maternal Newborn Health Registry Study Investigators. A prospective study of maternal, fetal and neonatal deaths in low- and middle-income countries. Bull World Health Organ 2014; 92: 605–12.
“Off the pulse” about incidental atrial fibrillation In their Comment, Brian Potter and Jaques LeLorier ( July 11, p 113)1 criticised the authors of the Framingham Heart Study2 on trends in atrial fibrillation for concluding 1339