Characteristics of the included studies
| Study identification | Location | Study period | Setting | Definition of neonate | Definition of EONS and LONS | Study design | Sample size associated with risk factors | Risk factors | Limitations | NHLBI score* | |
| Neonatal sepsis: clinically diagnosed | |||||||||||
| 1 | Masanja et al 25 | Tanzania: Kongwa and Mpwapwa District Hospitals and the Dodoma Regional Referral Hospital | May–July 2017 | Secondary-level and tertiary-level hospitals | 0–28 days | EONS: birth to 7 days of age. | Matched, prospective, case–control. | 322 (105 cases and 217 controls). | Neonatal and maternal. | Only examines risk factors related to EONS.21 | Good |
| 2 | Schrag et al 18 | South Africa: Chris Hani Baragwanath Hospital | 1 April 2004–25 October 2007 | Secondary-level and tertiary-level hospitals | 0–28 days | EONS: birth to 2 days of age. LONS: 3–28 days of age. | Prospective cohort. | 8129 (323 with sepsis). | Neonatal and maternal. | Fair | |
| 3 | Shobowale et al 16 | Nigeria: Babcock University Teaching Hospital | August 2014–August 2015 | Tertiary-level hospital | First 28 days of life | EONS: >3 days of age. LONS: older than 3 days. | Retrospective, cross-sectional study. | 100 (34 with sepsis). | Neonatal and maternal. | Poor | |
| 4 |
Cutland et al 17 | South Africa: Chris Hani Baragwanath Hospital | April 2004–October 2007 | Secondary-level and tertiary-level hospitals | 0–28 days | EONS: within 3 days of life. LONS: between 4 and 28 days of life. | Prospective, cohort study. | 8129 (324 with sepsis). | Maternal. | Only examines the association between maternal HIV infection and neonatal sepsis. Same cohort as used in Schrag et al.18 | Good |
| 5 | Chiabi et al 21 | Cameroon: Yaounde Gynaeco-Obstetric and Pediatric Hospital | 18 November 2008–18 May 2009 | Tertiary-level hospital | 0–28 days | EONS: birth to 7 days of age. LONS: 8–28 days of age. | Prospective study. | 628 (218 with sepsis). | Neonatal and maternal. | Poor | |
| 6 | Jabiri et al 19 | Tanzania: Two municipal referral hospitals | 27 August–28 September 2015 | Tertiary-level hospitals | ≤28 days | Prospective, cross-sectional study. | 220 (69 with sepsis). | Neonatal and maternal. | Poor | ||
| 7 |
Subramaniam et al 26 | Cameroon: Three individual Cameroon Baptist Convention Health Services facilities | 10 January–27 April 2017 | Secondary-level hospitals | Not specified | Prospective, cohort study. | 217 mothers giving birth to 219 babies. Of these, 10 were diagnosed with neonatal sepsis. | Maternal. | Only looking at the association between neonatal sepsis and maternal group B streptococcus anogenital colonisation. | Fair | |
| 8 | Gudayu et al 27 | Ethiopia: University of Gondar Comprehensive Specialized Hospital | 1 January–31 December 2017 | Tertiary-level hospital | Not specified | Retrospective study. | 504 (321 with sepsis). | Neonatal, maternal, role of the season. | EONS and LONS are not specified even though they investigate these as risk factors. | Poor | |
| 9 | Woldu et al 28 | Ethiopia: Bishoftu General Hospital | 15 October 2013–15 April 2014 | Secondary-level hospital | 0–28 days | EONS: 0–7 days. LONS: 7–28 days. | Prospective. cross-sectional study. | 306 with sepsis. | Neonatal and maternal. | Poor | |
| 10 | Getabelew et al 29 | Ethiopia: Shashemene Referral Hospital | 5–30 February 2017 | Tertiary-level hospital | 0–28 days | EONS: 0–7 days. LONS: 7–28 days. | Cross-sectional study with retrospective document review. | 244 (190 with sepsis). | Neonatal and maternal. | Poor | |
| Neonatal sepsis (diagnosed by laboratory testing; eg, positive blood culture or haematological criteria) | |||||||||||
| 11 | Silago et al 30 | Tanzania: Bugando Medical Centre | December 2018–July 2019 | Secondary-level hospital | Not specified | Prospective, cross-sectional study. | 200 (69 with sepsis). | Neonatal and maternal. | Only looking at sepsis caused by Gram-negative bacteria. Sample size consists of only 69 neonates. | Poor | |
| 12 | Kabwe31 | Zambia: University Teaching Hospital, Lusaka | October 2013– May 2014 | Tertiary-level hospital | Not specified | EONS: positive culture in neonates aged ≤7 days. LONS: positive culture in neonates aged >7 days. | Prospective. cross-sectional study. | 303 (113 with sepsis). | Neonatal and maternal. | Poor | |
| 13 | Aiken et al 30 | Kenya: Kilifi District Hospital | 16 April 2002–30 September 2009 | Secondary-level hospital | ≤28 days | Prospective, cohort study. | 4668 (53 with nosocomial sepsis). | Neonatal. | Only examines hospital-acquired sepsis. | Fair | |
| 14 | Kayange et al 32 | Tanzania: Bugando Medical Centre | March–November 2009 | Secondary-level hospital | Not specified | EONS: disease occurring in ≤72 hours of age. LONS: disease occurring after more than 72 hours of age. | Prospective, cross-sectional study. | 300 (149 with sepsis). | Neonatal and maternal. | Poor | |
| 15 | Bassingthwaighte and Ballot33 | South Africa: Charlotte Maxeke Johannesburg Academic Hospital | 1 January 2011–31 January 2013 | Tertiary-level hospital | Not specified | EONS: positive blood culture within 72 hours after birth. LONS: positive blood culture >72 hours after birth. | Matched, case–control, retrospective record review. | 356; 178 cases (babies born before arrival) and 178 controls (babies born in hospital). | Born before arrival to hospital (neonatal?). | Only examines the risk of sepsis among babies born before arrival compared with babies born in the hospital. | Fair |
| 16 | Onalo et al 34 | Nigeria: Ahmadu Bello University Teaching Hospital | 25 May 2004–31 May 2005 | Tertiary-level hospital | 0–28 days | EONS: sepsis within the first 48 hours of life. | Prospective study. | 211 (75 with sepsis). | Poor | ||
| 17 | Ogunlesi et al 35 | Nigeria: Olabisi Onabanjo University Teaching Hospital | January 2006– December 2008 | Tertiary-level hospital | 0–28 days | EONS: positive blood culture drawn within the first 72 hours of life. LONS: positive blood culture drawn after 72 hours of life. | Prospective and retrospective observational study. | 1050 (174 with sepsis). | Neonatal and maternal. | Poor | |
| 18 | Pius et al 36 | Nigeria: University of Maiduguri Teaching Hospital | 1 January–31 December 2012 | Tertiary-level hospital | 0–28 days | EONS: first 72 hours of life. LONS: last 72 hours of life. | Prospective study. | 110 (46 with sepsis). | Neonatal and maternal. | Poor | |
| 19 | Shobowale et al 37 | Nigeria: Lagos University Teaching Hospital | Not specified | Tertiary-level hospital | 0–28 days | EONS: first 7 days of life. LONS: after the seventh day of life. | Prospective, cohort study. | 250 (85 with sepsis). | Neonatal. | Poor | |
| 20 | Ekwochi et al 38 | Nigeria: Enugu State University Teaching Hospital | January 2013– December 2016 | Tertiary-level hospital | First 28 days of life | EONS: first 72 hours of life. LONS: after 72 hours of life. | Matched, prospective, case–control study. | 228 (57 cases and 171 controls). | Neonatal and maternal. | Fair | |
| 21 | John et al 39 | Uganda: Kidera Health Center | January–August 2013 | Secondary-level hospital | 1–27 | Prospective, cross-sectional study. | 174 (38 with sepsis). | Neonatal and maternal. | Selection bias, since very sick newborns were transferred to district hospital. | Poor | |
| 22 | Gebremedhin et al 40 | Ethiopia: public hospitals in Mekelle City | December 2014–June 2015 | Secondary-level and tertiary-level hospitals | 0–28 | EONS: <7 days. LONS: 7–28 days. | Unmatched, prospective, case–control study. | 324 (78 cases and 156 controls). | Neonatal and maternal. | Good | |
| 23 | Alemu et al 41 | Ethiopia: Debre Markos Referral Hospital | 1 February–30 March 2018 | Tertiary-level hospital | 0–28 days | Unmatched, prospective, case–control study. | 246 (82 cases and 164 controls). | Neonatal and maternal. | Sepsis diagnosed by haematological criteria. | Good | |
| 24 | Yismaw et al 42 | Ethiopia: University of Gondar Hospital | 1 September–30 November 2017 | Tertiary-level hospital | 0–28 days | Prospective, cross-sectional study. | 423 (47 with sepsis). | Neonatal and maternal. | Poor | ||
| 25 | Adatara et al 43 | Ghana: Trauma and Specialist Hospital, Winneba | January–December 2017 | Tertiary level hospital | First 28 days of life | EONS: <7 days. LONS: 7–28 days. | Unmatched, retrospective, case–control study. | 900 (103 cases and 797 controls). | Neonatal and maternal. | Some patients were only diagnosed based on clinical features. | Fair |
| 26 | Geyesus et al 44 | Ethiopia: University of Gondar Hospital | September 2015–May 2016 | Tertiary-level hospital | 0–28 days | EONS: confirmed infection in the blood or cerebrospinal fluid of patients younger than 3 days of life. LONS: onset of such infection between 3 and 28 days. | Prospective, cross-sectional study. | 251 (117 with sepsis). | Neonatal and maternal. | Poor | |
| 27 | Adatara et al 45 | Ghana: Trauma and Specialist Hospital, Winneba | January–December 2017 | Tertiary-level hospital | First 28 days of life | EONS: <7 days. LONS: 7–28 days. | Retrospective, case–control study. | 383 (67 cases and 316 controls). | Neonatal. | Fair | |
| 28 | Sorsa46 | Ethiopia: Arsi University Teaching and Referral Hospital | April 2016–May 2017 | Tertiary-level hospital | 28 days of life or younger | EONS: first 6 days of life. LONS: 7–28 days of life. | Prospective, cross-sectional study. | 303 (88 with sepsis). | Neonatal and maternal. | Poor | |
| 29 | Kheir and Khair6 | Sudan: Soba University Hospital | October 2011–February 2012 | Tertiary-level hospital | First 28 days of life | Not specified. | Prospective study. | 62 (38 with sepsis). | Maternal. | Small study population. | Poor |
| 30 | Onyedibe et al 47 | Nigeria: Jos University Teaching Hospital | Not specified | Tertiary-level hospital | First 28 days of life | Not specified. | Prospective, cross-sectional study. | 218 (75 with sepsis). | Socioeconomic. | Only examines socioeconomic risk factors for neonatal sepsis. | Poor |
| 31 | Kpikpitse and Siakwa48 | Ghana: St Elizabeth Hospital Asutifi | January 2011 and December 2013 | Secondary-level hospital | Less than a month | Not specified. | Unmatched, prospective, case–control study. | 196 (96 cases and 100 controls). | Neonatal and maternal. | Fair | |
| 32 | Ogundare et al 49 | Nigeria: Wesley Guild Hospital | Conducted over 7 months, ended in March 2009 | Tertiary-level hospital | 0–28 days | EONS: positive blood culture drawn within the first 72 hours of life. LONS: positive blood culture drawn after 72 hours of life. | Prospective study. | 360 (72 with neonatal sepsis). | Neonatal and maternal. | Good | |
| 33 | West and Tabansi50 | Nigeria: University of Port Harcourt Teaching Hospital | 1 July–31 December 2007 | Tertiary-level hospital | 0–28 days | EONS: onset of illness within the first 72 hours of life. LONS: onset of illness after 72 hours of life. | Prospective study. | 406 (169 with sepsis). | Neonatal and maternal. | Poor | |
| 34 | Weldu et al 51 | Ethiopia: Ayder Comprehensive Specialized Hospital | March 2017– September 2018 | Tertiary-level hospital | 1–28 days | Not specified. | Prospective, cross-sectional study. | 317 (116 with sepsis). | Neonatal and maternal. | Poor | |
| 35 | Akalu et al 52 | Ethiopia: Debre Markos and Felege Hiwot Referral Hospitals | March–April 2018 | Tertiary-level hospital | First 28 days after birth | Not specified. | Unmatched, prospective, case–control study. | 231 neonates (77 cases and 154 controls). | Neonatal and maternal. | Good | |
| 36 | Olorukooba et al 53 | Nigeria: Ahmadu Bello University Teaching Hospital | May 2017–May 2018 | Tertiary-level hospital | First 28 days after birth | EONS: onset of symptoms 72 hours or less after birth. LONS: onset of symptoms more than 72 hours after birth. | Retrospective, cross-sectional study. | 465 (175 with sepsis). | Neonatal and maternal. | Poor | |
*Quality assessment score from the NHLBI quality assessment tool (online supplemental Online supplemental appendix 4).
EONS, early-onset neonatal sepsis; LONS, late-onset neonatal sepsis; NHLBI, National Heart, Lung, and Blood Institute.