Interventions to prevent neonatal healthcare-associated infections in low-resource settings: a systematic review ================================================================================================================ * Felicity C. Fitzgerald * Walter Zingg * Gwendoline Chimhini * Simbarashe Chimhuya * Stefanie Wittmann * Helen Brotherton * Ioana D. Olaru * Samuel R. Neal * Neal Russell * André Ricardo Araujo da Silva * Mike Sharland * Anna C Seale * Mark F Cotton * Susan Coffin * Angela Dramowski ## Abstract **Background** Clinically suspected and laboratory-confirmed bloodstream infections are frequent causes of morbidity and mortality during neonatal care. The most effective infection prevention and control (IPC) interventions for neonates in low-and-middle-income countries (LMIC) are unknown. **Aim** To identify effective interventions in the prevention of hospital-acquired bloodstream infections in LMIC neonatal units. **Methods** Medline, PUBMED, The Cochrane Database of Systematic Reviews, EMBASE, and PsychInfo (January 2003 – October 2020) were searched to identify studies reporting single or bundled interventions for prevention of bloodstream infections in LMIC neonatal units. **Results** Our initial search identified 5206 articles; following application of filters, 27 publications met the inclusion and ICROMS assessment criteria and were summarised in the final analysis. No studies were carried out in low-income countries, only one in sub-Saharan Africa and just two in multiple countries. Of the 18 single intervention studies, most targeted skin (n=4) and gastrointestinal mucosal integrity (n=5). Whereas emollient therapy and lactoferrin achieved significant reductions in proven neonatal infection, glutamine and mixed probiotics showed no benefit. Chlorhexidine gluconate for cord care and kangaroo mother care reduced infection in individual single-centre studies. Of the nine studies evaluating bundles, most focused on prevention of device-associated infections and achieved significant reductions in catheter- and ventilator-associated infections. **Conclusion** There is a limited evidence-base for the effectiveness of IPC interventions in LMIC neonatal units; bundled interventions targeting device-associated infections were most effective. More multi-site studies with robust study designs are needed to inform IPC intervention strategies in low-resource neonatal units. **Sources of Support** FF is supported by the Academy of Medical Sciences, the funders of the Starter Grant for Clinical Lecturers scheme and UCL Great Ormond Street NIHR Biomedical Research Centre. AD is supported by the Fogarty International Center of the National Institutes of Health, Emerging Global Leader Award Number K43-TW010682. Keywords * Infection prevention and control * low-and-middle income countries * systematic review * neonatal infection * hospital-acquired infection ## Introduction The World Health Organization (WHO) estimates that bacterial infections cause ∼25% of the 2.8 million annual neonatal deaths and long-term neurodevelopmental disabilities in survivors 1. Hospital-acquired infection (HAI) is a major cause of neonatal morbidity and mortality with prevalence ratios in low and middle-income countries (LMICs) 3–20 times higher than high-income countries 2. Traditional definitions, applied in high-income countries, use a 72-hour cut-off to differentiate early-from late-onset infection: the former associated with vertical transmission of pathogens such as Group B Streptococcus, the latter with horizontal transmission of hospital-acquired pathogens, often associated with prematurity and invasive procedures such as intravenous catheterisation. However, particularly in LMICs, there is recognition that facility-based delivery is itself a risk for HAIs, with pathogens such as *Klebsiella pneumoniae* (previously associated with late-onset infection) commonly isolated in the first 24 hours of life2, 3. This observation informs the Strengthening the Reporting of Observational Studies in Epidemiology for Newborn Infection (STROBE-NI) guidelines, which recommend recording the timing of symptom onset rather than the binary early/late-onset dicohotomy1. It also raises questions about fundamental differences in the mechanisms of neonatal infections in LMICs, as compared to high-income countries. The leading neonatal pathogens are increasingly resistant to first and second-line antimicrobials, with substantial resistance to commonly used agents including ampicillin (89% of *Escherichia coli)*, ceftriaxone (49% of *Klebsiella* spp. Isolates) and cloxacillin (40% of *Staphylococcus aureus*)3. In this context, effective, feasible and affordable interventions to enhance infection prevention and control (IPC) in LMIC neonatal units are critical to prevent both neonatal mortality, and emerging antimicrobial resistance. However, even in high-income settings, implementing effective prevention measures are challenging, and a robust evidence-base on what tools to use is limited. Randomised controlled trials are considered the gold-standard for generating evidence in general. However, best practice procedures and quality improvement interventions must be contextual for maximum impact. As interventions are seldom identical across trial sites, patient-level randomisation is often not possible. Trials within hospitals (randomizing wards for example) are at risk of bias due to movement between wards of staff and patients. Furthermore, matching hospitals for randomisation can be complex 4. To address these methodological challenges, new study designs, such as interrupted time series for cohorts and hospital-level stepped-wedge cluster-randomisation have been adopted. In addition, qualitative research aiming at understanding behaviour change is increasingly used to complement quantitative data 4. For neonates in LMICs, various HAI prevention strategies have been suggested but only studied in small and single centre studies. To date, the evidence-base in these settings has not yet been systematically assessed. We set out to review a broad range of potential interventions (both single and bundled), aiming to reduce healthcare-associated infections, with a focus on bloodstream infections (BSIs) in LMIC neonatal units. ## Methods This systematic review was conducted in adherence with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statements of evaluations of healthcare interventions5. We registered the search strategy on the international prospective register of systematic reviews (CRD42018112346 on PROSPERO, see supplementary files). ### Search strategy We searched Medline, The Cochrane Database of Systematic Reviews, EMBASE, and PsychInfo (1 January 2003 – 31 October 2020) to identify studies reporting on the effectiveness of interventions to prevent infections in LMIC neonatal wards and neonatal intensive care units. We selected the year 2003 to reflect the rapid evolution and spread of resistant bacteria causing HAIs in the last 17 years. IPC interventions were defined as any intervention aiming to prevent the development of a healthcare-associated bacterial or fungal infection such as BSI, meningitis, laboratory-confirmed urinary tract infection, or clinically suspected but culture-negative infections. We limited results by age (neonates 0-27 days or 0-89 days if admitted on a neonatal ward or NICU), by location (LMIC as defined by the 2021 World Bank classification6), by language (articles written in English, German, French, Italian, Portuguese and Spanish were included), and by relevant filters as per exclusion criteria (For a full list of terms and filters see supplementary files). Our primary outcome was the effect of the interventions on (a) incidence of infection, or (b) attributable mortality, depending on study definitions. Fungal or bacterial hospital-acquired invasive infections in hospitalised neonates were the primary events for study. Secondary outcomes included impact on incidence of laboratory-confirmed urinary tract infection, thrombophlebitis, necrotizing enterocolitis (NEC), device-associated infections (clinically suspected or culture proven) and clinically suspected infection where laboratory cultures were negative or not available. ### Inclusion criteria Studies were eligible for full-text review if conducted in hospitalised neonates, including neonatal ward and/or NICU settings, with a detailed description of the intervention. We included both single interventions (e.g. probiotics, kangaroo mother care (KMC), breastfeeding, fluconazole prophylaxis) and bundled interventions (e.g. vascular device care, hand hygiene, and healthcare worker education combined). Studies conducted in several countries including both high- and low- or middle-income countries (as per World Bank 2021 regions) could be included if possible to extract data from the LMIC settings. Study designs included randomised controlled trials, controlled and non-controlled before-after, controlled and non-controlled interrupted time series and cohort studies. ### Exclusion criteria We excluded letters, opinion articles and reviews that did not report primary data. IPC interventions conducted during maternal care, in community-based settings and during outbreaks were excluded. We also excluded studies conducted exclusively in high-income countries as per World Bank 2021 regions6. Interventions targeting viral infections (including HIV), infants older than 3 months, or involving vaccination, diagnostic tools, infection prediction scores were excluded. We also excluded studies addressing IPC interventions on mixed neonatal/paediatric populations where extraction of neonatal data was not possible; and where only abstracts were available despite contacting the corresponding author. Finally, we excluded studies where bacterial colonisation as opposed to invasive infection was the outcome, if bloodstream infection was not also included. ### Study selection process The initial eligibility assessment of titles and abstracts identified by our search was conducted independently by FF and AD using the pre-determined inclusion and exclusion criteria. Disagreements on eligibility were resolved by consensus, if needed by consulting a third party. The reference lists of all eligible publications were screened for cross-referencing. After finalising articles for full-text review, two authors evaluated the quality of each eligible publication using the Integrated quality Criteria for the Review Of Multiple Study designs (ICROMS) tool7, with disagreements resolved as explained above. The ICROMS tool was designed to allow the systematic integration and assessment of differing study types including both quantitative and qualitative designs for reviews of public health interventions such as those targeting IPC 7. The ICROMS tool provides a list of quality criteria with a set of requirements specific for the study design. Studies are evaluated by a ‘decision matrix’ where mandatory criteria must be met. The robustness of the study is measured by a score (See Supplementary Tables for criteria and scoring). To pass to the final analysis, studies must meet the minimum score and the mandatory ICROMS criteria, after duplicate review. ### Data abstraction We extracted data using a standardised data-collection form already independently piloted by FF and AD on a representative sample of studies. Study details collected on the form included: author/s, year of publication, country or countries where the study was performed, study design, study timeframe, setting (neonatal ward, NICU or both), intervention type, intervention details and effect. We grouped studies by intervention type: IPC bundles; catheter care; skin integrity and bacterial colonisation (umbilical cord care, skin cleansing, emollients and/or massage); fluconazole prophylaxis; hand hygiene; KMC; rooming-in/parental involvement in neonatal care; and gastrointestinal integrity (probiotics and feeding practices). Data synthesis involved the collation and tabulation of results by intervention type, summarizing the key intervention/s and their effectiveness in IPC for hospitalised neonates (using either relative risk, odds ratios or hazard ratios as reported by each study). We did not undertake a meta-analysis due to diversity of study type, interventions and outcomes i.e. although all studies targeted reduction of neonatal infections, each study had different modes of action for the intervention and/or major differences in study design that precluded combining data. ## Results We identified 5206 articles on initial searching, after removal of duplicates (Figure 1). Filter application (see appendix) reduced this to 1799 titles and abstracts then reviewed independently by two study authors (FF and AD) for relevance. Of these, 124 were selected for full text review in duplicate and ICROMS scoring, leading to another 97 exclusions and 27 selected for inclusion in the final review (Tables 1 and 2). Forty studies were excluded for either missing mandatory ICROMS criteria or ICROMS scores below the cut-off for the particular study design. Of the included studies, 8 were conducted in lower middle-income countries and 19 in upper middle-income countries (only two studies were multi-country). None were conducted in low-income countries. Including multi-site studies and using the 2021 World Bank regions, 14 study sites were in Latin America/Caribbean, fourteen in South-East Asia/Pacific, five in the Middle East/North Africa, three in Europe/Central Asia and one in sub-Saharan Africa6. Eighteen studies evaluated single interventions and nine evaluated bundled interventions (two of which were conducted in multiple countries). View this table: [Table 1:](http://medrxiv.org/content/early/2020/12/19/2020.12.17.20248047/T1) Table 1: Studies reaching ICROMS Criteria for inclusion describing single interventions for the prevention of hospital-acquired neonatal bloodstream infections and clinically-suspected infection in low-resource settings (January 2003 – January 2020) View this table: [Table 2:](http://medrxiv.org/content/early/2020/12/19/2020.12.17.20248047/T2) Table 2: Bundled interventions for the prevention of hospital-acquired neonatal bloodstream infections and clinically-suspected infection in low-resource settings (January 2003 – September 2018) ![Figure 1](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2020/12/19/2020.12.17.20248047/F1.medium.gif) [Figure 1](http://medrxiv.org/content/early/2020/12/19/2020.12.17.20248047/F1) Figure 1 Search strategy for the identification and selection of publications reporting the effectiveness of interventions to prevent infections in neonatal wards and intensive care (January 2003-October 2020) ### Single Intervention studies Of the single interventions (Table 1), probiotics/feeding interventions were the most commonly evaluated (five), followed by emollients (four), chlorhexidine cord cleansing (two) and KMC (two). Three of the 5 probiotic/feeding interventions evaluated oral bovine lactoferrin versus placebo in a total of 370 neonates with birthweights <2500g 8-10. Varying bovine lactoferrin dosage (from 80-200mg/kg/day) and weight/gestational age thresholds made data incomparable and meta-analysis inappropriate. Two studies showed reduction in HAI in the intervention groups, one documenting 4.4 infections per 1000 patient days in the intervention arm versus 17.3 (p=0.007), the other finding a risk ratio of 0.211; (95% confidence intervals [CI] 0.044-1.019; p = 0.036) in those receiving the intervention versus placebo8,9. Two studies evaluated enteral supplements but neither reduced infection incidence (parenteral glutamine supplementation[p=0.518]11 nor mixed probiotic administration[p=0.4] 12). For emollients, one group conducted two studies using sunflower seed oil in 103 Egyptian and 497 Bangladeshi neonates <72 hours of age, born at <34 or <33 weeks’ gestational age respectively13, 14. Both studies found that sunflower seed oil massage was associated with a significant decrease in the adjusted incidence rate ratio (aIRR, adjusted for weight on admission, gestational age and sex) of culture proven BSI than control (aIRR 0.46; 95% CI, 0.26–0.81; and 0.59, 95% CI 0.37-0.96). Notably, the Bangladeshi study showed no difference in the rate of clinically suspected infection triggering taking of blood cultures or antibiotic treatment rates between groups, although culture-proven BSI decreased in the intervention arm. Topical coconut oil was used in a Pakistani study in 270 neonates (26-34 weeks gestational age), first in the neonatal unit (NNU) and then at home15. Neonates randomised to the control arm had an increased risk of hospital-acquired BSI (adjusted hazard ratio 6.0, 95% CI 2.3-16). A Turkish study of 197 preterm neonates (<34 weeks’ gestation and <24 hours old) found no difference in mortality, incidence of culture-proven or clinically-suspected infection in patients randomised to receive Aquaphor emollient versus standard skin care16. Two studies from India examined the impact of topical application of chlorhexidine gluconate; one in 140 neonates ≥32 weeks’ gestational age and ≥1500g using chlorhexidine 2.5% to clean the umbilical stump; the other in 140 neonates comparing whole body cleansing with chlorhexidine 0.25% versus tepid water17, 18. The first demonstrated a significant decrease in culture-proven BSI with chlorhexidine cord care (2 versus 15, p= 0.02, absolute risk 21%, versus 3%, absolute risk reduction 19%, confidence intervals not shown), although clinically-suspected infections increased in intervention versus control subjects (Table 1)17. The second study found a non-significant decrease in blood culture positivity with whole body cleansing (6/168 blood cultures positive in intervention group versus 12/175, p=0.195), possibly owing to a small sample size and that blood cultures were taken at set intervals regardless of clinical indication18. Studies on KMC were carried out in Colombia and Malaysia, in 746 neonates <2000g and 126 neonates <1500g respectively19, 20. These studies evaluated substantially different KMC interventions (∼24 hours per day of KMC versus ≥1 hour per day of KMC (Table 1). The Colombian study found similar numbers of infectious episodes 49/382 (intervention) versus 44/364 (controls), although they describe a milder phenotype in the intervention arm, and a reduction in nosocomial infections (8% versus 4% in interventions/controls, p=0.026, absolute figures not given), without a clear distinction of the definition of ‘nosocomial’ versus other infections. In the Malaysian study, there were 2/64 infections in the intervention group versus 1/64 (controls, p=1.0). A large cohort study in Colombia (6655 neonates) evaluating a hand hygiene intervention (alcohol-based handrub dispensers, daily surveillance and quarterly feedback), found a decreased incidence density of neonatal methicillin-resistant *Staphylococcus aureus* bloodstream infections (from 2.2 to 0.6 per 1000 patient days, p=0.01), although no decrease in *Acinetobacter baumannii*21 (0.6-0.2 per 1000 patient days, p not given). A small Brazilian study of massage therapy versus no intervention (n=104) reported lower incidence of late onset infections in the intervention versus control groups22. No study evaluating ‘rooming-in’ (defined as continuous presence of parent caregivers in the neonatal unit23), peripherally inserted central catheters versus standard intravenous catheters24 and fluconazole prophylaxis25 found differences in infection rates between the study arms (Table 1). ### Bundled interventions Five of the 9 studies reporting the impact of IPC bundles (Table 2) focused on preventing device-associated infection26-30. One small, single-centre study in an Egyptian NICU, achieved significant reduction in ventilator-associated pneumonia (VAP) rates and mechanical ventilation days, with a trend towards reduction in NICU length of stay and overall mortality26. A multi-country study in 10 NICUs demonstrated significant reduction in VAP rates (RR 0.67, 95%CI 0.50-0.91), after implementation of a multi-modal strategy including hand hygiene, oral antiseptics, ventilator circuit management and enhanced infection surveillance27. A tertiary hospital, 50-bed NICU in China significantly reduced VAP rates, as well as overall mortality following implementation of a bundle including hand hygiene, ventilator disinfection, education and rational antibiotic use28. Two studies targeted prevention of central line-associated bloodstream infection (CLABSI). A multi-country study in 4 NICUs demonstrated significant reduction in CLABSI rates following a multi-modal intervention strategy including education, enhanced process and outcome surveillance and staff feedback (rate ratio 0.45; 95%CI 0.33–0.63)29. A single-center Brazilian NICU significantly reduced CLABSI rates (24 versus 15 per 1000 catheter days; p=0.04) following implementation of a bundle including education, hand hygiene, CHG skin preparation and removal of unnecessary catheters30. The first of two studies utilizing education/training interventions was a non-controlled ‘before after’ study conducted in 2 NICUs in the Philippines. The bundle focused on quality improvement in blood culture collection, hand hygiene compliance, use of infection control checklists and staff education. Although there was no change in the primary outcome (proportion of neonates newly colonised with resistant pathogens) or in the secondary outcome of bacteraemia, the study achieved improved hand hygiene compliance rates and reduction in overall mortality31. A Brazilian study in 5 neonatal units conducted an interrupted time-series analysis following introduction of a nurse training package including IPC measures. Despite improvement in nurses’ knowledge and practices, there was no change in mortality or rates of hospital-acquired BSI (11.3 vs. 12.3 cases/1,000 infant days)32. A single-centre cohort study at a large, academic centre NICU in China enrolled outborn neonates <1500g to assess the impact of a hypothermia prevention bundle on admission temperature, rates of NEC and neonatal infection. Mean axillary temperature on arrival increased and overall mortality rates decreased significantly, however there was no difference in either NEC or infection incidence following the intervention33. A recent, large cohort study in a Zambian neonatal unit evaluated the impact of IPC training, text message reminders for staff, hand hygiene promotion with alcohol-based handrub, enhanced environmental cleaning and weekly whole-body bathing of neonates >1.5 kg with 2% chlorhexidine gluconate. The bundle achieved significant reduction in overall mortality, clinically-suspected infection and culture-proven BSI for all birth weight groups except those <1kg.34 In a subsequent sub-analysis of the intervention group data, CHG bathing reduced the hazard rate of bloodstream infection among inborn babies >1.5 kg by a factor of 0.58 (p = 0.10, 95% CI: 0.31, 1.11)35. ## Discussion Although infection is the most frequent complication of hospitalisation in LMIC neonates, the most effective IPC interventions remain unknown. We therefore conducted a systematic review of published studies describing the impact of various IPC interventions on healthcare-associated infection rates in LMIC NNUs. We identified 27 eligible publications that assessed single (n=18) and bundled IPC interventions (n=9). None were carried out in low-income countries, only one in sub-Saharan Africa and just two had sites in multiple countries. We found considerable heterogeneity of study design, analysis and outcomes selected, as well as diversity in the modes of infection prevention targeted (skin and gastrointestinal mucosal integrity, promotion of normal flora acquisition, reduction of bacterial pathogen colonisation). The evidence-base we have identified for the effectiveness of IPC interventions in LMIC neonatal units is limited, but appears most promising for bundled interventions targeting device-associated infections. Limitations of this review include the paucity of published research on neonatal IPC from LMIC, the lack of multi-centre studies or large sample sizes and the failure to use optimal study interventional study designs. Although we endeavoured to be as inclusive as possible in our search terms, we only searched four databases and in six languages, so it is possible that we missed some relevant studies. It was not appropriate to do meta-analyses due to heterogeneity of both interventions and outcomes. Most studies were carried out in tertiary or academic neonatal units, which further limits the generalisability of the findings. Of note, although our initial search captured a large number of potentially eligible studies, full text review led to 40/120 (33%) papers being excluded due to not including mandatory criteria required by ICROMS, or having a low score for study design/analysis quality. Thus some geographical areas were not well represented-in particular sub-Saharan Africa with only one study included34. This highlights the challenges for clinicians in LMIC settings to identify and implement contextually appropriate evidence-based guidelines. It also demonstrates the difficulties of designing and analysing high quality IPC studies where facility, laboratory and statistical support may be lacking. IPC studies are notoriously complex to design and implement, with issues of contamination between arms, the need for large scale randomisation (e.g. cluster randomisation of hospitals) and use of study designs unfamiliar to many academic clinicians e.g. interrupted time series analysis. IPC interventions also frequently involve behaviour change, which does not lend itself to RCT evaluation. In recognition of the importance of evaluating effective behaviour change in interventions in fields such as IPC, the UK Medical Research Council has developed guidance on how these studies should be designed and implemented36. Similarly the ICROMS score was developed to allow the inclusion of studies such as controlled before -after, non-controlled before-after studies and qualitative studies in assessing evidence, the exclusion of which from standard systematic reviews undermines their potential contribution to the evidence base7. A major challenge in selecting the primary endpoint for neonatal IPC studies is the very low yield of blood cultures (the current gold standard for confirmation of BSI) in both high- and low-income settings. This necessitates recruitment of large numbers of neonates to conclusively demonstrate an intervention’s impact, which is often particularly challenging in LMIC owing to budgetary and logistic constraints. Sensitive and specific neonatal infection diagnostic tools that are accessible and affordable in LMIC settings are needed. In addition, standardised and validated definitions for clinically-suspected, culture-negative neonatal infections are required, to allow for comparison of findings across study sites. Use of multiple study outcomes (proven infection, clinically-suspected infection and mortality) may complicate interpretation of findings, particularly where the results are discrepant14. Until there is consensus on definitions of clinically-suspected neonatal infection, particularly in settings where cultures have limited availability, the issue of quantifying reduction in infection rates will persist. Despite these inherent limitations in the available data, endpoint definitions and study methodologies used, we have conducted the first systematic review of IPC interventions for LMIC NNUs. We used a robust search strategy, long inclusion timeframe and ICROMS quality assessment to ensure we have identified all relevant and rigorously conducted research on this topic. Among the single intervention studies, emollient therapy (sunflower oil) in low birthweight babies had the strongest evidence supporting its use, demonstrating reduced healthcare-associated infection rates in both studies13, 14. There was also evidence to support the use of oral bovine lactoferrin, although the studies were small and there was inconsistency in dosage used. This finding is echoed in a recent Cochrane review of studies in high and low-resource settings which concluded there was low-certainty evidence that lactoferrin supplementation could reduce late-onset sepsis, though not necrotising enterocolitis or all cause mortality37. Contrary to another previous Cochrane review, we did not find strong evidence for KMC as an intervention to reduce BSI in LMICs- only two studies fulfilled ICROMS criteria and only one had some evidence of impact on BSI20, 38. For studies that analyzed the impact of bundled interventions, the strongest evidence was generated from studies aiming to prevent device-associated infection. Bundles incorporating other interventions (education, infection surveillance with feedback, hand hygiene promotion and chlorhexidine gluconate bathing) were also effective, but the evidence was generated from single-centre or small studies. Particular areas that appear promising for future research on neonatal IPC in LMIC are the use of chlorhexidine gluconate body washing and/or emollient therapy. Bundles that target neonatal BSI (the most common neonatal HAI) should be developed, utilising lessons learned from the success of bundles targeting device-associated infections. The ideal bundled intervention should target all portals of entry for pathogenic bacteria causing neonatal BSI. It could include avoidance of hospitalisation and/or invasive procedures, promotion of mucosal integrity (gut and skin), promotion of colonisation with normal flora and reduced colonisation with pathogenic bacteria. Future studies in LMIC should utilise multi-national collaborations, standardise definitions (or at least clearly elucidate what criteria have been used) and use robust study designs e.g. individual randomised or cluster-randomised controlled trials and interrupted time-series analysis to generate evidence for IPC interventions that can be adopted in neonatal practice. Wherever possible guidelines such as Strengthening the Reporting of Observational Studies in Epidemiology for Newborn Infection (STROBE-NI) should be followed to allow for future comparisons between studies1. ## Conclusion There is a limited evidence-base for IPC interventions in LMIC neonatal units. Overall, bundled interventions targeting prevention of device-associated infection are supported by the strongest evidence to date. More multi-site studies using standardised neonatal infection definitions and robust study designs are needed to inform IPC interventions for use in low-resource neonatal units. ## Data Availability Systematic Review of Previously published data View this table: [Supplementary Table 1:](http://medrxiv.org/content/early/2020/12/19/2020.12.17.20248047/T4) Supplementary Table 1: ICROMS Quality Criteria applied for each study, by study design (from Zingg *et al*., 2016)1 View this table: [Supplementary Table 2:](http://medrxiv.org/content/early/2020/12/19/2020.12.17.20248047/T5) Supplementary Table 2: ICROMS Decision matrix: mandatory criteria and minimum score for study type to be included in review (from Zingg *et al*., 2016)1 ## Appendix Supplementary Appendix Filters applied to initial search results with number of records identified View this table: [Table3](http://medrxiv.org/content/early/2020/12/19/2020.12.17.20248047/T3) * Received December 17, 2020. * Revision received December 17, 2020. * Accepted December 19, 2020. * © 2020, Posted by Cold Spring Harbor Laboratory This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/) ## References 1. 1.Fitchett EJA, Seale AC, Vergnano S et al. Strengthening the Reporting of Observational Studies in Epidemiology for Newborn Infection (STROBE-NI): an extension of the STROBE statement for neonatal infection research. Lancet Infect Dis 2016; 16: e202– e13. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S1473-3099(16)30082-2&link_type=DOI) 2. 2.Zaidi AKM, Huskins WC, Thaver D, et al. Hospital-acquired neonatal infections in developing countries. Lancet 2005; 365: 1175–88. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0140-6736(05)71881-X&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15794973&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000227909500031&link_type=ISI) 3. 3.Okomo U, Akpalu ENK, Le Doare K et al. Aetiology of invasive bacterial infection and antimicrobial resistance in neonates in sub-Saharan Africa: a systematic review and meta-analysis in line with the STROBE-NI reporting guidelines. Lancet Infect Dis 2019; 19: 1219–34. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31522858&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) 4. 4.Gould DJ, Creedon S, Jeanes A, Drey NS, Chudleigh J, Moralejo D Impact of observing hand hygiene in practice and research: a methodological reconsideration. J Hosp Infect 2017; 95: 169–74. 5. 5.Liberati A, Altman DG, Tetzlaff J et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ (Clinical research ed) 2009; 339: b2700. [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjE3OiIzMzkvanVsMjFfMS9iMjcwMCI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDIwLzEyLzE5LzIwMjAuMTIuMTcuMjAyNDgwNDcuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 6. 6.Bank W. World Bank Country and Lending Groups. 7. 7.Zingg W, Castro-Sanchez E, Secci FV et al. Innovative tools for quality assessment: integrated quality criteria for review of multiple study designs (ICROMS). Public Health 2016; 133: 19–37. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.puhe.2015.10.012&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) 8. 8.Akin IM, Atasay B, Dogu F et al. Oral lactoferrin to prevent nosocomial sepsis and necrotizing enterocolitis of premature neonates and effect on T-regulatory cells. Am J Perinatol 2014; 31: 1111–20. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1055/s-0034-1371704&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24839144&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) 9. 9.Kaur G, Gathwala G Efficacy of Bovine Lactoferrin Supplementation in Preventing Late- onset Sepsis in low Birth Weight Neonates: A Randomized Placebo-Controlled Clinical Trial. J Trop Pediatr 2015; 61: 370–6. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/tropej/fmv044&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26224129&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) 10. 10.Ochoa TJ, Zegarra J, Cam L et al. Randomized Controlled Trial of Lactoferrin for Prevention of Sepsis in Peruvian Neonates Less than 2500 g. Pediatric Infectious Disease Journal 2015; 34: 571–6. 11. 11.Li ZH, Wang DH, Dong M Effect of parenteral glutamine supplementation in premature infants. Chinese Medical Journal 2007; 120: 140–4. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17335657&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) 12. 12.Wang Y, Gao L, Zhang YH, Shi CS, Ren CM Efficacy of probiotic therapy in full-term infants with critical illness. Asia Pacific Journal of Clinical Nutrition 2014; 23: 575–80. 13. 13.Darmstadt GL, Badrawi N, Law PA et al. Topically applied sunflower seed oil prevents invasive bacterial infections in preterm infants in Egypt: a randomized, controlled clinical trial. Pediatric Infectious Disease Journal 2004; 23: 719–25. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/01.inf.0000133047.50836.6f&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15295221&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000223391400005&link_type=ISI) 14. 14.Darmstadt GL, Saha SK, Ahmed AS et al. Effect of topical treatment with skin barrier- enhancing emollients on nosocomial infections in preterm infants in Bangladesh: a randomised controlled trial. Lancet 2005; 365: 1039–45. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0140-6736(05)71140-5&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15781099&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000227731800028&link_type=ISI) 15. 15.Salam RA, Darmstadt GL, Bhutta ZA Effect of emollient therapy on clinical outcomes in preterm neonates in Pakistan: a randomised controlled trial. Archives of Disease in Childhood Fetal & Neonatal Edition 2015; 100: F210–5. 16. 16.Erdemir A, Kahramaner Z, Yuksel Y et al. The effect of topical ointment on neonatal sepsis in preterm infants. J Matern Fetal Neonatal Med 2015; 28: 33–6. 17. 17.Gathwala G, Sharma D, Bhakhri B Effect of topical application of chlorhexidine for umbilical cord care in comparison with conventional dry cord care on the risk of neonatal sepsis: a randomized controlled trial. J Trop Pediatr 2013; 59: 209–13. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/tropej/fmt003&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23407285&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000319814500008&link_type=ISI) 18. 18.Gupta B, Vaswani ND, Sharma D, Chaudhary U, Lekhwani S Evaluation of efficacy of skin cleansing with chlorhexidine in prevention of neonatal nosocomial sepsis - a randomized controlled trial. Journal of Maternal-Fetal & Neonatal Medicine 2016; 29: 242–7. 19. 19.Boo NY, Jamli FM Short duration of skin-to-skin contact: effects on growth and breastfeeding. Journal of paediatrics and child health 2007; 43: 831–6. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.1440-1754.2007.01198.x&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17803670&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2020%2F12%2F19%2F2020.12.17.20248047.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000251202200009&link_type=ISI) 20. 20.Charpak N, Ruiz-Pelaez JG, Figueroa de CZ, Charpak Y Kangaroo mother versus traditional care for newborn infants