Clinical Features and CFR
Clinical syndromes were available from 91/92 neonates and 2647/2655 non-neonates from ES sites only. Neonates presented most frequently with meningitis (36/91; 40%) compared with non-neonates (898/2647; 34%, P = 0.3; Table 1). Non-neonates presented most frequently with lower respiratory tract infections (1318/2647; 50%; neonates 28/91; 31%; P < 0.01; Table 1).
The outcomes were available for 90/92 neonates and 2627/2655 non-neonates from the ES sites. Overall, CFR was similar for both neonates and non-neonates (Table 1). CFR was highest for meningitis for both age groups (14/36 (39 %) versus 13/53 (25 %), P = 0.2 for neonates and 327/882 (37%) versus 285/1501 (19%), P < 0.01 for non-neonates.
Viable isolates were available for 76% (5021/6583) of reported cases, 195 neonatal and 4826 non-neonatal isolates. There were 16 isolates that were nontypeable, 1 neonatal and 15 non-neonatal. PCV7 serotypes were responsible for 31% (61/194) neonatal IPD and 59% (2853/4811) non-neonatal IPD (P = 0.05; Table 1). The PCV13 serotypes were responsible for 69% (134/194) of IPD in neonates and 84% in non-neonates (4 042/4 811; Table 1). The proportion of PCV7 and PCV13 serotypes responsible for IPD in neonates was significantly lower than in non-neonates (P < 0.01; Table 1). Forty-six percent (90/194) of neonatal IPD were accounted for by serotypes 5 (n = 18), 1 (n = 17), 19F (n = 15), 3 (n = 14), 8 (n = 13) and 14 (n = 13). These serotypes were responsible for 33% (1572/4811) of non-neonatal IPD serotypes (P < 0.01; Fig. 1). Serotypes 1, 3 and 5 were more frequently isolated among neonates, 25% (49/194), than among non-neonates, 5% (247/4 811; P < 0.01; Fig. 1). The most common non-neonatal serotypes were 14 (n = 805), 6B (n = 618), 6A (n = 580), 23F (n = 542), 19F (n = 520) (Fig. 1). The non-PCV13 serotypes 8, 12F and 13 accounted for 13% (25/194) of neonatal and 4% (183/4811) of non-neonatal isolates (Fig. 1).
Antimicrobial susceptibility testing was performed on all 5021 viable isolates. Among neonates, 76% (148/195) of isolates were susceptible to penicillin, compared with 50% (2424/4826) non-neonatal IPD isolates (P < 0.01; Table 1). Most isolates in this study were susceptible to ceftriaxone, 99% (194/195) and 98% (4757/4826) among neonates and non-neonates, respectively (Table 1). Cotrimoxazole nonsusceptibility was lower among neonates (77/195, 39%) than non-neonates (3542/4826, 73%; P < 0.01; Table 1). Among all tested isolates, 27% (1361/5021) were MDR, of which 15% (30/195) were neonatal and 28% (1331/4826) non-neonatal isolates (P < 0.01; Table 1).
Six serotypes most commonly associated with nonsusceptibility to penicillin were serotypes 14, 19F, 6B, 23F, 6A and 19A. These accounted for 89% (42/47) and 91% (2124/2402) of penicillin nonsusceptible isolates among neonates and non-neonates, respectively (Fig. 2). These 6 serotypes were also the most frequent among the MDR isolates. Serotype 14 was the predominant MDR serotype: 40% (12/30) and 51% (684/1 331) in neonates and non-neonates, respectively (Fig. 2).
Early-onset Versus Late-onset Disease
Fifty-one percentage (149/294) of neonates presented with EOD (Table 2). The median age for EOD was 0 days (IQR, 0–2), with 66% (99/149) presenting within 48 hours of birth. The median age for LOD was 14 days (IQR, 10–22). The EOD patients were more likely to have blood specimen sources than LOD patients (110/149, 74% vs. 76/145, 52%, P < 0.01, Table 2). LOD cases presented with meningitis more frequently than EOD cases (LOD, 25/48; 52% vs. EOD: 11/43; 26%, P = 0.01; Table 2). The high neonatal CFR did not differ by age of onset (EOD, 14/42, 33% and LOD 14/48, 29%, P = 0.7). Cases with meningitis contributed substantially to the CFR in both EOD and LOD (4/11 [36%] in EOD; 10/25 [40%] in LOD), P = 0.7. In addition, pneumonia CFR did differ by age: in EOD (6/14 [43%] versus 1/14 [7%] in LOD), P < 0.03.
In this study, conducted before introduction of PCV7 vaccination in South Africa, neonatal IPD accounted for an estimated 4.5% of IPD cases in children ≤ 2 years of age. Fifty-one percentage of these neonates presented within the first week of life. Meningitis was the most common clinical diagnosis among neonates, particularly among those with LOD. PCV13 serotypes accounted for a substantial portion, 69%, of neonatal cases. The most frequent neonatal serotypes 1, 3 and 5 accounted for 25% of neonatal and only 5% of non-neonatal IPD cases. The high neonatal CFR (31%) varied by site of infection, but not by age of onset.
The national incidence of neonatal IPD, 5 per 100,000 live births in South Africa in 2008, was lower than the estimated global incidence of 36 per 100,000 live births in 2010.4 Billings et al reported an incidence of 16 per 100,000 live births, before the introduction of PCV, in less-developed UN strata countries.4 Our incidence is also much lower than the incidence reported in Chile, of 59 per 100,000 population, and closer to that reported in the United States in 2006 (11 per 100,000 live births), and in England and Wales in 2013 (13 per 100,000 live births) before PCV, among < 90 day old infants.5–7 The incidence in this study is similar to that reported by Cutland,24 from a South African city, Soweto, where the incidence of neonatal sepsis due to the pneumococcus was reported as 8 per 100,000 live births among neonates. In the Sowetan study, S. pneumoniae was noted to occur less frequently than other common causes of neonatal sepsis, such as Streptococcus agalactiae, Staphylococcus aureus, Streptococcus viridans and Escherichia coli.24 Differences in incidence may be attributed to the higher threshold for taking blood culture specimens in neonatal units in South Africa, variation in surveillance methodologies and completeness in reporting.4 , 7 The incidence we report may be an underestimate of true neonatal incidence as infants with clinically evident, but microbiologically negative, sepsis would not have been included in this study. Vaccine probe studies such as that by Palmu et al25 demonstrated that inclusion of these clinically evident, microbiologically negative cases significantly increased the incidence estimates of IPD among children. In addition, the sensitivity of cultures among neonates is low, attributable to inadequate sample volumes being submitted, as well as empiric antimicrobials being commenced before cultures were taken.26 , 27
Among South African neonates, a large proportion of IPD cases, 51%, presented with EOD, similar to high-income countries like the United States and United Kingdom where 70% (19/27) and 77% (101/131), respectively, of neonatal IPD cases had EOD.6 , 7 This contrasts with studies from Utah and Mexico, where only 11% (2/9) and 20% (25/126) of neonatal cases, respectively, were EOD.28 , 29 The higher rates of EOD in South African neonates and those of the United States and England and Wales may be due to similar at risk populations, access to care and specimen-taking practices.4 The variation between and within countries may be attributed to differences in small hospital-based studies, socioeconomic status, access to antenatal care and maternal and infant risk factors.28 , 29
In this study, 66% (99/149) of the EOD neonates presented within the first 48 hours of life, similar to that reported by Ladhani et al7 in the United Kingdom, 67% (84/101), who indicated that these infants were more likely to be premature. Early-onset sepsis has been found to be associated with prematurity, maternal chorioamnionitis, or social factors influencing prenatal care.30 We were unable to analyze for prematurity or other maternal factors as these data were not collected during the study period.
Although the association of IPD and HIV infection in children has been well documented in South African children,31 this was not clear among neonates in this study. The high rates among neonates with IPD, 48%, may be because children who were most ill or had signs of HIV were preferentially tested, or would have presented to a health care setting. In addition, HIV status data were only available for 15% (44/294) of neonates as there was no policy for universal HIV testing at birth at the time of this study.
We observed a female sex preponderance in this study. Two studies, in Mexico and Denmark, reported a male sex preponderance,29 , 32 while others do not report a sex preponderance6 , 7 among neonates with IPD. A male sex predisposition to neonatal sepsis, particularly Gram-negative sepsis, has been attributed to x-linked immunoregulatory genes.33 , 34 This predisposition may be specific to Gram-negative sepsis in neonates and therefore not consistently observed in neonatal IPD.
The predominant clinical presentation among the neonatal group overall (40%) and the LOD group was meningitis (52%). However, the predominant clinical presentation among EOD cases was bacteremia (42%). This was consistent with findings from Ladhani et al7 in England and Wales, and Soto-Noguerón et al29 in Mexico who reported bacteremia as the predominant presentation in the EOD cases and meningitis in the LOD cases. The more frequent diagnosis of bacteremia among EOD cases may relate to an inability of the immature immune system in these very young babies to localize the infection.30 , 35
Although the CFR among neonates (31%) was higher than that among non-neonates (26%), this did not reach statistical significance. The neonatal CFR was also lower than those in other studies in England and Wales and the United States.7 , 11 This may be attributed to an underestimation of the neonatal CFR, as infants who demised at home would not have been included in this database. In addition, only 32% (90/294) of neonates with IPD had outcomes available for analysis. Meningitis, an established risk factor for death in patients with IPD,7 , 36 was associated with the highest CFR among both neonates and non-neonates in this setting. The CFR in neonates with IPD in this study was higher than that of neonates with sepsis due to more frequently encountered pathogens, such as Group B Streptococcus, 16.9%24 or Escherichia coli, 6%,37 in South Africa.
The neonatal isolates were generally more susceptible to antimicrobials tested (penicillin and ceftriaxone) than the non-neonatal isolates, as in the United States and Mexico, before PCV7.11 , 30 This is not unexpected as the neonatal serotypes, unlike the pediatric serotypes, are usually not associated with antimicrobial resistance.13 , 38 , 39
Our findings of 31% PCV7 serotypes in neonatal IPD are consistent with pneumococcal vaccine studies from Mexico (34%), and England and Wales (44%), before PCV7.7 , 29 The PCV13 serotype coverage among neonatal IPD isolates (69%) was also comparable with those in Mexico (64%), and England and Wales (67%).7 , 29 While a substantial proportion of neonatal IPD (69%) were due to PCV13 serotypes, this was significantly less than that observed in the non-neonatal IPD group (84%). The common neonatal IPD serotypes 1, 3 and 5 among South African neonates is consistent with other studies from the United States and Denmark.11 , 32 These serotypes have been reported to occur more frequently among adults than children in the United Kingdom, Denmark and South Africa.7 , 31 , 32 This supports the widely accepted premise of neonatal IPD being acquired via horizontal spread from mother or adult caregiver.9
This study has several limitations. First, the data were collected using a laboratory-based surveillance system, where isolate submission is dependent on diligent local laboratory and surveillance staff. Case ascertainment also suffers from differential access to care and specimen-taking practices throughout the country. Only cases with known ages were included. In addition, audits performed on the surveillance database did not include private sector cases. Therefore, our estimates are an underestimation of actual disease burden in children ≤ 2 years old in South Africa. Second, as the study was performed retrospectively, we were unable to check for maternal factors, such as premature labor, preterm rupture of membranes, maternal HIV infection, vaginal colonization or maternal IPD. Neonatal data, especially relating to HIV infection and outcomes, were also incomplete in our database. Third, susceptibility test results were interpreted using meningitis breakpoints irrespective of the clinical syndrome; therefore, the resistance rates appear higher in this study. This was appropriate as our study looked at trends over time, and not treatment outcomes. Fourth, susceptibility testing for ceftriaxone was revised from an agar dilution method to a Clinical and Laboratory Standards Institute-recommended broth microdilution method, using TREK panels, in 2009,21 as the agar dilution method was found to underestimate beta-lactam resistance.40
Our findings suggest that the pneumococcus, while not as common a cause of neonatal sepsis as other agents like Group B Streptococcus or E.coli, is associated with a higher CFR. Neonatal IPD in this country is found to be similar to neonatal IPD in other countries in terms of clinical presentation, serotype distribution, antimicrobial susceptibility, and CFRs. The findings in this study establish a baseline against which to interpret changes due to herd protection that may occur in neonatal IPD since the implementation of PCV in South Africa.
1. Watt JP, Wolfson LJ, O’Brien KL, et al; Hib and Pneumococcal Global Burden of Disease Study Team. Burden of disease caused by Haemophilus influenzae type b in children younger than 5 years: global estimates. Lancet. 2009;374:903–911.
2. Pneumococcal vaccines WHO position paper 2012. Weekly epidemiological record/Health Section of the Secretariat of the League of Nations. 2012;87:Geneva: World Health Organization. 129–144.
4. Billings ME, Deloria-Knoll M, O’Brien KL. Global burden of neonatal invasive pneumococcal disease
: a systematic review and meta-analysis. Pediatr Infect Dis J. 2016;35:172–179.
5. Lagos R, Muñoz A, San Martin O, et al. Age- and serotype-specific pediatric invasive pneumococcal disease
: insights from systematic surveillance in Santiago, Chile, 1994–2007. J Infect Dis. 2008;198:1809–1817. doi: 10.1086/593334.
6. Poehling KA, Talbot TR, Griffin MR, et al. Invasive pneumococcal disease
among infants before and after introduction of pneumococcal conjugate vaccine. JAMA. 2006;295:1668–1674.
7. Ladhani SN, Andrews NJ, Waight P, et al. Impact of the 7-Valent pneumococcal conjugate vaccine on invasive pneumococcal disease
in infants younger than 90 days in England and Wale. Clinical Infectious Diseases. 2013;56:633–640.
8. Barron P, Pillay Y, Doherty T, et al. Eliminating mother-to-child HIV transmission in South Africa
. Bull World Health Organ. 2013;91:70–74.
9. Malhotra A, Hunt RW, Doherty RR. Streptococcus pneumoniae sepsis in the newborn. J Paediatr Child Health. 2012;48:E79–E83.
10. Rodriguez BF, Mascaraque LR, Fraile LR, et al. Streptococcus pneumoniae: the forgotten microorganism in neonatal sepsis. Fetal Pediatr Pathol. 2015;34:202–205.
11. Gomez M, Alter S, Kumar ML, et al. Neonatal Streptococcus pneumoniae infection: case reports and review of the literature. Pediatr Infect Dis J. 1999;18:1014–1018.
12. Hoffman JA, Mason EO, Schutze GE, et al. Streptococcus pneumoniae infections in the neonate. Pediatrics. 2003;112:1095–1102.
13. Hausdorff WP, Feikin DR, Klugman KP. Epidemiological differences among pneumococcal serotypes. Lancet Infect Dis. 2005;5:83–93.
14. Kim TH, Johnstone J, Loeb M. Vaccine herd effect. Scand J Infect Dis. 2011;43:683–689.
15. Holmlund E, Nohynek H, Quiambao B, et al. Mother-infant vaccination with pneumococcal polysaccharide vaccine: persistence of maternal antibodies and responses of infants to vaccination. Vaccine. 2011;29:4565–4575.
16. Chaithongwongwatthana S, Yamasmit W, Limpongsanurak S, et al. Pneumococcal vaccination during pregnancy for preventing infant infection. Cochrane Database Syst Rev. 2015;1:CD004903.
17. Huebner RE, Klugman KP, Matai U, et al. Laboratory surveillance for Haemophilus influenzae type B meningococcal, and pneumococcal disease. Haemophilus Surveillance Working Group. S Afr Med J. 1999;89:924–925.
18. Pathirana J, Muñoz FM, Abbing-Karahagopian V, et al; Brighton Collaboration Neonatal Death Working Group. Neonatal death: case definition & guidelines for data collection, analysis, and presentation of immunization safety data. Vaccine. 2016;34:6027–6037.
19. Crewe-Brown HH, Karstaedt AS, Saunders GL, et al. Streptococcus pneumoniae blood culture isolates from patients with and without human immunodeficiency virus infection: alterations in penicillin susceptibilities and in serogroups or serotypes. Clin Infect Dis. 1997;25:1165–1172.
21. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. Twenty-third informational supplement. 2013; M100–S23.
22. Magiorakos AP, Srinivasan A, Carey RB, et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect. 2012;18:268–281.
23. Singh E, Cohen C, Govender N, et al. A description of HIV testing strategies at 21 laboratories in South Africa
. Commun Dis Surveill Bull. 2008; 6:16–17.
25. Palmu AA, Kilpi TM, Rinta-Kokko H, et al. Pneumococcal conjugate vaccine and clinically suspected invasive pneumococcal disease
. Pediatrics. 2015;136:e22–e27.
27. Schelonka RL, Chai MK, Yoder BA, et al. Volume of blood required to detect common neonatal pathogens. J Pediatr. 1996;129:275–278.
28. Olarte L, Ampofo K, Stockmann C, et al. Invasive pneumococcal disease
in infants younger than 90 days before and after introduction of PCV7. Pediatrics. 2013;132:e17–e24.
29. Soto-Noguerón A, Carnalla-Barajas MN, Solórzano-Santos F, et al. Streptococcus pneumoniae as cause of infection in infants less than 60 days of age: serotypes and antimicrobial susceptibility. Int J Infect Dis. 2016;42:69–73.
30. Simonsen KA, Anderson-Berry AL, Delair SF, et al. Early-onset neonatal sepsis. Clin Microbiol Rev. 2014;27:21–47.
31. von Gottberg A, Cohen C, de Gouveia L, et al. Epidemiology of invasive pneumococcal disease
in the pre-conjugate vaccine era: South Africa
, 2003–2008. Vaccine. 2013;31:4200–4208.
32. Kaltoft Zeuthen N, Konradsen HB. Epidemiology of invasive pneumococcal infections in children aged 0-6 years in Denmark: a 19-year nationwide surveillance study. Acta Paediatr Suppl. 2000;89:3–10.
33. Nagwa GM, Begum S, El-Batanony MH, et al. Clinical and bacteriological profile of neonatal sepsis in King Khaleed Civilian Hospital, Tabuk, Kingdom of Saudi Arabia. Eur J Prev Med. 2016;4:1–6.
34. Karambin M, Zarkesh M. Entrobacter, the most common pathogen of neonatal septicemia in Rasht, Iran. Iran J Pediatr. 2011;21:83–87.
35. Bulkowstein S, Ben-Shimol S, Givon-Lavi N, et al. Comparison of early onset sepsis and community-acquired late onset sepsis in infants less than 3 months of age. BMC Pediatrics. 2016;16:82.
36. Nyasulu P, Cohen C, De Gouveia L, et al. Increased risk of death in human immunodeficiency virus-infected children with Pneumococcal meningitis in South Africa
, 2003–2005. PIDJ. 2011;30:1075–1080.
37. Motara F, Ballot DE, Perovic O. Epidemiology of neonatal sepsis at Johannesburg Hospital. Southern African J Epidemiol Infect. 2005;20:90–93.
38. Hausdorff WP. The roles of pneumococcal serotypes 1 and 5 in paediatric invasive disease. Vaccine. 2007;25:2406–2412.
39. Von Mollendorf C, Cohen C, Tempia S, et al. Epidemiology of serotype 1 invasive Pneumococcal disease
, South Africa
, 2003–2013. Emerg Infect Dis. 2016;22:261–270.
40. von Mollendorf C, Cohen C, de Gouveia L, et al. Factors associated with ceftriaxone non-susceptibility of Streptococcus pneumoniae: analysis of South African national surveillance data, 2003 to 2010. Antimicrob Agents Chemother. 2014;58:3293–3305.
Keywords:Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
neonates; invasive pneumococcal disease; South Africa