Panel: research in context
Evidence before this study
Data describing the burden of endemic healthcare-associated infections (HCAI) and associated antimicrobial resistance (AMR) profile in Africa remain few. In most African countries, HCAI surveillance is not prioritised. The 2009 systematic review of endemic HCAI in Africa included 19 studies, which showed that prevalence of HCAI varied considerably, ranging from 2.5 to 14.8% across different hospital settings, with surgical wards reporting prevalence ranging from 5.7 to 45.8%. Among the reviewed studies, surgical site infection (SSI) was the most reported HCAI, with a prevalence ranging from 2.5 to 30.9%. Limited data on causative pathogens were available; however, a few studies have highlighted the significance of gram-negative rods, particularly in SSI and ventilator-associated pneumonia. Available data on endemic HCAI suggest that its prevalence is substantially higher than that in developed countries.
Added value of this study
This systematic review and meta-analysis provide a comprehensive update on the magnitude and nature of HCAI in Africa, including data from 81,968 patients from 92 studies and 20 countries. We have summarised the recent epidemiology of endemic HCAI in Africa, examined differences in HCAI epidemiology between African Union regions and stratified our analysis by HCAI types. Further, we described bacterial aetiology and prevalence of HCAI associated AMR. By identifying the risk variables associated with HCAI, this review provides information that may be useful in supporting the design of focused and efficient infection prevention and control measures to reduce the burden of these infections in Africa.
Anzeige
Implication of all the available evidence
Surveillance data play a key role in preventing HCAI as they enable healthcare facilities to track, monitor, and respond to trends, detect outbreaks early, identify risk factors, benchmark performance, and support quality improvement efforts. The heterogeneity in both the surveillance methods used and the prevalence of HCAI identified in our review suggests that existing definitions of HCAI types are not fit for purpose in the African context. Collection and analysis of timely and accurate surveillance data are essential for effective infection prevention and control programs. Pragmatic and appropriate HCAI definitions should be developed, validated, and implemented to both for local action and to estimate the burden of HCAI across Africa.
Background
Healthcare-associated infection (HCAI) is a global health challenge that seriously threatens patient safety by significantly increasing the morbidity and mortality associated with healthcare exposure, hospital length of stay, long-term disability, financial burden, and contributing to the spread of multidrug-resistant (MDR) pathogens [1‐6]. HCAI are thought to have a higher burden in African countries than in high-income countries, yet are understudied and underreported [7]. Data summarised in previous reviews report endemic HCAI prevalence of up to 15.5% in general wards and can reach 50% in intensive care units (ICU) in Africa [7‐9]. However, it is important to emphasise that most of these studies focus on a particular institution, often tertiary and/or teaching facilities and may not reflect the situation at a national level.
Maintaining HCAI surveillance is a challenge in well-resourced healthcare settings [10] and even more so in low-resource ones, however defining the magnitude of HCAI is key to placing it in context for policy makers. It is anticipated that HCAI is responsible for a significant burden of disease in contexts where there is a lack of basic infection prevention and control (IPC) capacity. A further complication associated with HCAI is antimicrobial resistance (AMR), which has emerged as a major public health problem worldwide, with bacteria associated with HCAI disproportionately resistant to antibiotics [11, 12]. In low-resource settings, HCAI is likely to be exacerbated by infrastructural problems i.e. lack of water in hospitals (particularly safe water), poor hygiene and sanitation, understaffing, failure to implement or lack of antimicrobial policies, shortage of basic laboratory equipment for diagnosis, suboptimal adherence to safe practices by health care workers, limited compulsion to report HCAI, and limited funding [13, 14].
The World Health Organization (WHO) has developed several IPC guidelines and documents founded on the core components framework, a key aspect of which is HCAI surveillance [15, 16]. Data from HCAI surveillance can be used to quantify the HCAI burden, evaluate HCAI trends over time, pinpoint areas where HCAI prevention efforts need to be targeted and improved, and IPC strategies to reduce HCAI [15, 17].
Anzeige
Despite the significant impact of HCAI in Africa, there is a lack of up-to-date and comprehensive information on the prevalence, risk factors, and AMR of endemic HCAI in the region. This systematic review and meta-analysis of endemic HCAI in Africa is an update of the last 13 years since the last review published in 2011 by Nejad and colleagues [8], containing data published from 1995 to 2009. Here, we aimed to provide an up-to-date and comprehensive overview of the prevalence, risk factors, aetiology, and AMR of endemic HCAI in Africa.
Methods
Search strategy and eligibility criteria
For this systematic review and meta-analysis, we searched the MEDLINE/PubMed, CINAHL, and Global Health (EBSCOhost interface) electronic databases. To ensure literature saturation, the reference lists of the included studies were scanned to identify and capture other relevant studies, and Google Scholar was used to identify and screen studies citing them. Literature published in French and English between January 2010 and December 2022 was considered. The search was limited to human subjects and the most common HCAI encountered in African countries, including surgical site infections (SSI), healthcare-associated urinary tract infections (HA-UTI), healthcare-associated bloodstream infections (BSI), and hospital-acquired pneumonia/ventilator-associated pneumonia. The search strategy was developed based on the outcomes of interest (prevalence, risk factors, and antimicrobial resistance profile of bacteria isolated from HCAI). The complete search strategy with keywords and MeSH terms using the Boolean terms “OR” and “AND” is provided in supplemental materials (Supplement pp3-4).
We included observational studies (case-control, longitudinal, cohort, and cross-sectional) that prospectively or retrospectively explored the outcomes of interest (prevalence, risk factors, aetiologic agent and the antimicrobial resistance profile of bacteria isolated from HCAI) in all age groups in inpatient settings. We excluded the following types of studies due to the lack of relevance to our research question, or due to their limitations in allowing us to identify risk factors for HCAI or assess AMR: those reporting only the prevalence of HCAI without including at least one of the other outcomes of interest, those reporting on specific microorganisms causing HCAI, those reporting risk factors associated with HCAI but not reporting the effect size measures of these factors, or those reporting on HCAI outbreaks. Case series, case reports, editorials, commentaries, conference proceedings, preprints, reviews, previous systematic reviews and meta-analyses, and unpublished articles were excluded. Research letters to the editor containing data that met these criteria were included. We followed the Preferred Reporting Items for Systematic Review and Meta-Analysis 2020 (PRISMA) guidelines [18] (Supplement pp5-6) while conducting this systematic review and meta-analysis. The protocol was registered and published in PROSPERO (CRD42022374559).
Data extraction
Two reviewers (GKB and EM) independently screened the titles and abstracts of studies according to the inclusion criteria. Any disagreements were resolved by discussion and consensus. In cases of further disagreement, a third reviewer (NF) holding a casting vote was consulted. The full-text review then occurred, and each reviewer independently screened the full texts against the inclusion and exclusion criteria, with disagreements resolved by consensus or discussion with a third reviewer (NF), if necessary.
Two reviewers (GKB and EM) extracted the data in a pre-piloted Excel spreadsheet. This included the study’s first author, publication year, region, country, population, study design, study population, surveillance definition used to define HCAI, type of HCAI and its prevalence, risk factors associated with HCAI (if analysed) and their effect size, and isolated bacteria and their antimicrobial resistance profile (if reported). The extracted data were compared, and discrepancies were resolved through discussion.
To assess the risk of bias in the included studies, we used a modified Critical Appraisal Skills Programme (CASP) checklist, designed to fit our research question and the Newcastle-Ottawa scale (NOS) for assessing the quality of non-randomized studies in meta-analyses (Supplement pp7-8) [19]. Both quality scales and domain-based tools were used simultaneously to assess all the included studies. The CASP checklist used in this systematic review assessed four domains: (i) appropriateness of the study population or participant recruitment, (ii) eligibility criteria, (iii) valid methods to identify the HCAI, and (iv) selective non-reporting or under-reporting of outcome measures. The NOS quality instrument score was awarded a star (corresponding to the points) for each area. It assesses the study’s area of selection (maximum of 5 points), comparability (maximum of 2 points), and outcomes (maximum of 3 points). After summing the star points, the studies were classified into three categories: good (7–10 points), moderate (5–6 points), and poor (0–4 points). The risk of bias assessment was performed by GKB and EM, and any disagreements were resolved by consensus.
Data synthesis and statistical analysis
The extracted data were reported as study-level summary estimates, and qualitative and quantitative techniques were used to synthesise them. The primary outcome was the prevalence of HCAI stratified by HCAI type. Secondary outcomes included risk factors associated with HCAI and AMR profile of reported bacteria causing HCAI. All estimates were expressed as proportions with restricted maximum likelihood (REML) 95% confidence intervals (CI) and presented in forest plots using a random-effects meta-analysis. Heterogeneity was assessed using Higgins I2 statistic, Cochran’s Q test, and tau-squared τ2. The 95% CI around τ2 and I2 were calculated to assess confidence in these metrics. We set a stringent I2 threshold of > 75% as indicative of significant heterogeneity, but we also assessed this heterogeneity through the CIs and localisation on the forest plot [20]. Significant heterogeneity in prevalence between HCAI types was expected since some types of HCAI are more commonly reported, and this was addressed by stratification by HCAI type [8]. In the case of high heterogeneity, we determined it was not appropriate to pool estimates of HCAI prevalence [21].
To provide pooled risk estimates for the factors associated with HCAI, an exploratory meta-regression analysis was performed for risk factors found to be significant in at least four of the included studies. This threshold was chosen to ensure that the meta-regression analysis was based on an adequate number of studies to provide a robust and meaningful relationship between the pooled risk estimates and the outcome of interest. This approach helps reduce the risk of random chance or spurious associations and increases the validity of the meta-regression results. We presented the pooled estimated effect size (odds ratio (OR)) and degrees of heterogeneity with their 95% CI and p-values. The OR were computed and reported on a log scale. A statistically significant (p < 0.05) coefficient indicated an association between the effect estimate for HCAI and the associated risk factors.
Anzeige
We used quantile regression to calculate the AMR median rates and interquartile ranges (IQR) of the reported bacteria. We performed post hoc sensitivity analyses stratified by study quality to further estimate the robustness of the relevance estimates. All analyses were performed using the meta (version 6.1-0) and metafor (version 3.8-1) packages in R.
Role of the funding sources
The funders of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript.
Results
A total of 2541 studies were identified after a comprehensive search. Among them, 2499 studies were identified through the database we consulted and the remaining 42 were identified through other sources. Ninety-two studies from 20 African countries were included in the analysis (Fig. 1). The largest number of studies were from Ethiopia (n = 32, 34.8%) [22‐53], Egypt (n = 10, 10.9%) [54‐63], Tunisia (n = 7, 7.6%) [64‐70], Tanzania (n = 6, 6.5%) [71‐76], and Nigeria (n = 5, 5.4%) [77‐81]. Other studies were from South Africa (n = 4, 4.3%) [82‐85], Rwanda [86‐88], Ghana [89‐91], Cameroon [92‐94], Sierra Leone [95‐97], and Morocco [98‐100] (for each, n = 3, 3.3%), Kenya [101, 102], Uganda [103, 104], Democratic Republic of Congo [105, 106], and Benin [107, 108] (for each, n = 2, 2.2%), Malawi [13], Gabon [109], Burkina Faso [110], Mali [111], and Algeria [112] (for each, n = 1, 1.1%). When classifying countries by UN African region, 46 (50%) studies were from Eastern Africa, 21 (22.8%) from Northern, 15 (16.3%) from Western, 6 (6.5%) from Middle, and 4 (4.4%) from Southern Africa (Fig. 2). These studies comprised a total of 81,968 patients. The sample size ranged from 32 to 15,502 patients per study. The geographical distribution of these patients by UN African region is as follow: 37,643 from Northern Africa, 21,102 from Western Africa, 16,384 from Eastern Africa, 3,989 from Middle Africa, and 2,850 from Southern Africa.
Fig. 1
Flowchart of articles screening and selection
×
![]()
Anzeige
Fig. 2
Geographical distribution and number of selected studies in different African countries
×
![]()
Most studies were conducted at university/teaching hospitals (n = 44, 47.8%) and other tertiary hospitals (n = 37, 40.2%). Approximately half of the included studies were conducted in the surgical (n = 31, 33.7%) and obstetrics wards (n = 18, 19.6%). Twenty (21.7%) studies were conducted in ICUs, while 16 (17.4%) reported HCAI from all wards of their hospitals. Most studies used the CDC HCAI surveillance definitions (n = 49, 53.3%), with a few examples of ECDC (n = 3, 3.3%) and WHO (n = 3, 3.3%) definitions being used. Two studies used both the CDC and ECDC definitions, while the IDSA and NINSS HCAI surveillance definitions were each used in one (1.1%) study. Among the studies that used CDC definitions, four adapted them to the local context. A total of 33 (35.9%) studies used their own surveillance definitions or did not report the definitions used in diagnosing HCAI. Study characteristics are summarised in Table 1.
Table 1
Characteristics of included studies
1st Authors, publication year | Region | Country | Type of setting | Study design/type | Targeted ward | Age group | Surveillance definition used | Population | HCAI Prevalence |
|---|---|---|---|---|---|---|---|---|---|
Abdel-Wahab et al. 2013 [54] | Northern | Egypt | University/Teaching hospital | Prospective | Neonatal ICU | Neonates | Adapted CDC | 238 | 21.4 |
Abosse et al. 2021 [22] | Eastern | Ethiopia | Tertiary hospital | Cross-sectional | All | All | NR | 165 | 69.7 |
Abubakar, 2020 [77] | Western | Nigeria | University/Teaching hospital | PPS | All | All | ECDC | 321 | 15.6 |
Ahoyo et al. 2014 [107] | Western | Benin | Tertiary hospital | PPS | All | All | CDC | 3130 | 19.1 |
Akoko et al. 2012 [71] | Eastern | Tanzania | Tertiary hospital | Cross-sectional | Surgical | All | NINSS | 118 | 35.6 |
Alemye et al. 2021 [23] | Eastern | Ethiopia | Tertiary hospital | Cross-sectional | Obstetrics | Women | CDC | 1069 | 12.3 |
Atif et al. 2015 [112] | Northern | Algeria | University/Teaching hospital | Prospective | Surgical | Adults | NNISa | 593 | 5.4 |
Awoke et al. 2019 [34] | Eastern | Ethiopia | University/Teaching hospital | Cross-sectional | Surgical | All | NR | 261 | 13.0 |
Ayala et al. 2021 [45] | Eastern | Ethiopia | University/Teaching hospital | Cross-sectional | Obstetrics | Women | NR | 382 | 8.9 |
Ayed et al. 2019 [64] | Northern | Tunisia | University/Teaching hospital | PPS | All | All | CDC | 752 | 10.9 |
Azeze et al. 2019 [48] | Eastern | Ethiopia | Tertiary hospital | Cross-sectional | Obstetrics | Women | NR | 383 | 7.8 |
Bediako-Bowan et al. 2020 [89] | Western | Ghana | University/Teaching hospital | Cross-sectional | Surgical | Adults | CDC | 3267 | 10.1 |
Bediako-Bowan et al. 2020 (2) [90] | Western | Ghana | University/Teaching hospital | Prospective | Surgical | All | CDC | 358 | 16.2 |
Behari et al. 2015 [82] | Southern | South Africa | District hospital | Prospective | Surgical ICU | Adults | Own | 32 | 25.0 |
Belay et al. 2022 [49] | Eastern | Ethiopia | Tertiary hospital | Retrospective | ICU | Adults | Own | 312 | 27.9 |
Birhanu et al. 2022 [50] | Eastern | Ethiopia | University/Teaching hospital | Cross-sectional | Surgical | All | Own | 373 | 19.3 |
Bizuayehu et al. 2022 [51] | Eastern | Ethiopia | Tertiary hospital | Cross-sectional | ICU | Adults | Own | 220 | 51.4 |
Bizuayew et al. 2021 [52] | Eastern | Ethiopia | District hospital | Cross-sectional | Obstetrics | Women | CDC | 622 | 12.4 |
Bunduki et al., 2021 [13] | Eastern | Malawi | University/Teaching hospital | PPS | Surgical | Adults | ECDC | 105 | 11.4 |
Chernet et al. 2020 [55] | Northern | Egypt | University/Teaching hospital | Longitudinal | All | Adults | WHO | 300 | 14.0 |
Dawit et al. 2021 [53] | Eastern | Ethiopia | Tertiary hospital | Longitudinal | ICU | Adults | NR | 278 | 21.6 |
Dayyab et al. 2018 [81] | Western | Nigeria | Tertiary hospital | Prospective | Medical, Surgical, ICU, Dialysis unit | Adults | CDC | 2334 | 6.2 |
De Nardo et al. 2016 [72] | Eastern | Tanzania | Tertiary hospital | Prospective cohort | Obstetrics | Women | CDC | 467 | 48.2 |
Degbey et al. 2021 [108] | Western | Benin | University/Teaching hospital | Retrospective | Surgical | Adults | NR | 384 | 7.8 |
Di Gennaro et al. 2020 [95] | Western | Sierra Leone | Tertiary hospital | Case–control | Obstetrics | Women | CDC | 1016 | 25.0 |
Dramowski et al. 2015 [84] | Southern | South Africa | University/Teaching hospital | Retrospective | Neonatal ICU | Neonates | CDC | 1145 | 62.6 |
Dramowski et al. 2016 [83] | Southern | South Africa | Tertiary hospital | Prospective | Paediatric ICU | Children | CDC/NHSNa | 1347 | 31.0 |
Endalafer et al. 2011 [24] | Eastern | Ethiopia | University/Teaching hospital | Prospective | Surgical, ICU | Adults | CDC | 215 | 39.1 |
Fisha et al. 2019 [25] | Eastern | Ethiopia | District hospital | Retrospective | Surgical | All | Own | 642 | 10.0 |
Flouchi et al. 2022 [98] | Northern | Morocco | Tertiary hospital | Prospective | Surgical | All | CDC | 2521 | 6.3 |
Gadallah et al. 2014 [56] | Northern | Egypt | University/Teaching hospital | Prospective | Neonatal ICU | Neonates | CDC | 434 | 34.3 |
Galal et al. 2016 [57] | Northern | Egypt | University/Teaching hospital | Prospective | Paediatric ICU | Children | Own | 427 | 30.9 |
Gelaw et al. 2017 [26] | Eastern | Ethiopia | General hospital | Cross-sectional | Obstetrics | Women | Own | 384 | 6.8 |
Ghali et al. 2018 [65] | Northern | Tunisia | University/Teaching hospital | Prospective | Surgical | All | CDC | 349 | 8.6 |
Gomaa et al. 2021 [58] | Northern | Egypt | Tertiary hospital | Case–control | Obstetrics | Women | NR | 15,502 | 5.3 |
Hafez et al. 2012 [59] | Northern | Egypt | University/Teaching hospital | Prospective | Urology, Cardiothoracic | All | CDC | 1062 | 17.6 |
Hajjej et al. 2014 [66] | Northern | Tunisia | Tertiary hospital | Prospective | Surgical ICU | Adults | IDSA | 260 | 12.3 |
Halawi et al. 2018 [27] | Eastern | Ethiopia | University/Teaching hospital | Prospective | Surgical | All | Adapted CDC | 131 | 20.6 |
Hassan et al. 2020 [60] | Northern | Egypt | Tertiary hospital | Prospective | All | All | NR | 6912 | 3.7 |
Iwuafor et al. 2016 [78] | Western | Nigeria | University/Teaching hospital | Prospective cohort | ICU | Adults | Own | 71 | 63.4 |
Jamoussi et al. 2018 [67] | Northern | Tunisia | University/Teaching hospital | PPS | ICU | Adults | Own | 103 | 31.1 |
Kakupa et al. 2016 [105] | Middle | D.R. Congo | University/Teaching hospital | Cross-sectional | All | All | WHO | 171 | 34.5 |
Kallel et al. 2010 [68] | Northern | Tunisia | University/Teaching hospital | Prospective cohort | ICU | Adults | CDC | 261 | 21.1 |
Kefale et al. 2020 [28] | Eastern | Ethiopia | General hospital | Cross-sectional | Surgical | Adults | NR | 281 | 19.6 |
Ketata et al. 2021 [69] | Northern | Tunisia | University/Teaching hospital | PPS | All | All | CDC/NHSNa | 898 | 9.0 |
Ketema et al. 2020 [29] | Eastern | Ethiopia | Tertiary hospital | Prospective cohort | Obstetrics | Women | CDC | 520 | 25.4 |
Kibwana et al. 2022 [73] | Eastern | Tanzania | University/Teaching hospital | Cross-sectional | Urology | All | NR | 182 | 22.0 |
Kisibo et al. 2017 [74] | Eastern | Tanzania | Tertiary hospital | Cross-sectional | Orthopaedics | Adults | Own | 300 | 24.0 |
Labi et al. 2019 [91] | Western | Ghana | Tertiary hospital | PPS | All | All | ECDC | 2107 | 8.2 |
Lakoh et al. 2022 [96] | Western | Sierra Leone | Tertiary hospital | Prospective cohort | Surgical | Adults | CDC, ECDC | 338 | 11.5 |
Lakoh et al. 2022 (2) [97] | Western | Sierra Leone | Tertiary hospital | Prospective | Surgical | Adults | CDC, ECDC | 417 | 8.2 |
Laloto et al. 2017 [30] | Eastern | Ethiopia | University/Teaching hospital | Prospective | Surgical | All | CDC | 105 | 19.0 |
Lijaemiro et al. 2020 [31] | Eastern | Ethiopia | Tertiary hospital | Prospective cohort | Obstetrics | Women | NR | 166 | 15.1 |
Lubega et al. 2017 [104] | Eastern | Uganda | Tertiary hospital | Prospective | Surgical | All | Own | 110 | 16.4 |
Lukuke et al. 2017 [106] | Middle | D.R. Congo | Tertiary hospital | Longitudinal | Obstetrics | Women | WHO | 207 | 15.5 |
Mamo et al. 2017 [32] | Eastern | Ethiopia | University/Teaching hospital | Retrospective | Obstetrics | Women | CDC | 384 | 9.4 |
Maoulainine et al. 2014 [99] | Northern | Morocco | Tertiary hospital | Prospective | Neonatal ICU | Neonates | CDC | 702 | 13.0 |
Mawalla et al. 2011 [75] | Eastern | Tanzania | Tertiary hospital | Cross-sectional | Surgical | All | CDC | 250 | 26.0 |
Melaku et al. 2012 [33] | Eastern | Ethiopia | Tertiary hospital | Cross-sectional | Surgical, gynaecology and obstetrics | Adults | CDC | 1254 | 9.4 |
Mezemir et al. 2020 [35] | Eastern | Ethiopia | University/Teaching hospital | Cross-sectional | Surgical | All | CDC, NNISa | 249 | 24.5 |
Misha et al. 2021 [36] | Eastern | Ethiopia | University/Teaching hospital | Prospective | Surgical | Adults | NR | 251 | 21.1 |
Misha et al. 2021 (2) [37] | Eastern | Ethiopia | University/Teaching hospital | Prospective cohort | Surgical | Adults | CDC | 251 | 21.1 |
Mohamed et al. 2022 [38] | Eastern | Ethiopia | Tertiary hospital | Retrospective | Paediatric ICU | Children | NR | 223 | 20.2 |
Molla et al. 2019 [39] | Eastern | Ethiopia | General hospital | Cross-sectional | Obstetrics | Women | CDC | 334 | 8.1 |
Mpogoro et al. 2014 [76] | Eastern | Tanzania | Tertiary hospital | Prospective cohort | Obstetrics | Adults | CDC | 312 | 10.9 |
Mukagendaneza et al. 2019 [86] | Eastern | Rwanda | University/Teaching hospital | Prospective | Surgical | Adults | CDC | 294 | 11.6 |
Nair et al. 2018 [85] | Southern | South Africa | Tertiary hospital | PPS | All | All | CDC | 326 | 7.7 |
Nanyunja et al. 2022 [103] | Eastern | Uganda | Tertiary hospital | Prospective cohort | Dialysis unit | Adults | Own | 121 | 40.5 |
Nkurunziza et al. 2019 [87] | Eastern | Rwanda | District hospital | Prospective cohort | Obstetrics | Women | NR | 550 | 10.9 |
Nouetchognou et al. 2016 [92] | Middle | Cameroon | University/Teaching hospital | Longitudinal | ICU, Gynaecology, Surgical, Neonatal unit | All | Own | 307 | 19.2 |
Nwanko et al. 2016 [79] | Western | Nigeria | University/Teaching hospital | Prospective | Surgical | All | Own | 5800 | 25.2 |
Olowo-okere et al. 2018 [80] | Western | Nigeria | University/Teaching hospital | Prospective cohort | Surgical | Adults, Children | CDC | 135 | 25.9 |
Ouedraogo et al. 2020 [110] | Western | Burkina Faso | Tertiary hospital | Prospective | Surgical | All | CDC | 964 | 11.8 |
Oumer et al. 2021 [40] | Eastern | Ethiopia | General hospital | Cross-sectional | All | Adults | CDC | 231 | 16.9 |
Raouf et al. 2020 [61] | Northern | Egypt | Tertiary hospital | Prospective | Surgical | Adults | CDC/NHSNa | 2510 | 2.3 |
Sahiledengle et al. 2020 [41] | Eastern | Ethiopia | University/Teaching hospital | Prospective | Paediatrics | Children | CDC/NHSNa | 448 | 12.7 |
Saied et al. 2011 [62] | Northern | Egypt | University/Teaching hospital | Prospective | ICU | All | CDC | 1575 | 38.1 |
Salem et al. 2011 [70] | Northern | Tunisia | University/Teaching hospital | PPS | All | All | CDC | 1373 | 5.4 |
Sattar et al. 2018 [102] | Eastern | Kenya | University/Teaching hospital | Cross-sectional | ICU | All | Own | 92 | 54.3 |
Scherbaum et al. 2014 [109] | Middle | Gabon | District hospital | Prospective | All | All | Adapted CDC | 2925 | 1.6 |
See et al. 2013 [63] | Northern | Egypt | University/Teaching hospital | Prospective | ICU | All | Adapted CDC | 523 | 90.2 |
Shakir et al. 2021 [42] | Eastern | Ethiopia | Tertiary hospital | Cross-sectional | Surgical | All | NR | 306 | 11.8 |
Shimi et al. 2015 [100] | Northern | Morocco | University/Teaching hospital | Retrospective | ICU | All | CDC | 535 | 11.2 |
Titus et al. 2021 [93] | Middle | Cameroon | Tertiary hospital | Cross-sectional | Surgical | Adults | NR | 304 | 12.2 |
Togo et al. 2010 [111] | Western | Mali | University/Teaching hospital | Prospective | Surgical | All | CDC | 460 | 9.6 |
Velin et al. 2021 [88] | Eastern | Rwanda | District hospital | Prospective | Obstetrics | Women | NR | 795 | 5.7 |
Victor et al. 2013 [101] | Eastern | Kenya | University/Teaching hospital | Prospective | Surgical | Adults | CDC | 262 | 6.9 |
Weldu et al. 2018 [43] | Eastern | Ethiopia | District hospital | Cross-sectional | Surgical | Adults | CDC | 280 | 11.1 |
Wendmagegn et al. 2018 [44] | Eastern | Ethiopia | Tertiary hospital | Retrospective | Obstetrics | Women | Own | 206 | 11.7 |
Wodajo et al. 2017 [46] | Eastern | Ethiopia | University/Teaching hospital | Cross-sectional | Obstetrics | Women | CDC | 592 | 11.0 |
Yallew et al. 2016 [47] | Eastern | Ethiopia | University/Teaching hospital | Cross-sectional | All | All | CDC | 908 | 14.9 |
Yaouba et al. 2016 [94] | Middle | Cameroon | Tertiary hospital | Prospective | Surgical, Gynaecology | Adults | CDC | 75 | 30.7 |
The NOS and CASP risk bias assessment findings are detailed in the “Supplemental Materials” section (Supplement pp9-15). Of the 92 included studies, 13 (14.1%), 30 (32.6%), and 49 (53.3%) were assessed to be of poor, moderate, and good quality, respectively. The validity of the methods used to identify the HCAI, the sample size, and selective non-reporting or under-reporting of outcome measures were sources of bias.
Anzeige
HCAI prevalence was extremely heterogeneous across studies ranging from 1.6 to 90.2% with a median of 15% (Table 1). This heterogeneity was also observed when studies were reported by African Union regions: 5.5–69.7% (median 15.7%) for Eastern Africa, 2.3–90.2% (median 12.3%) for Northern Africa, 6.2–63.4% (median 11.8%) for Western Africa, 1.6–34.5% (median 17.3%) for Middle Africa, and 7.7–62.6% (median 28%) for Southern Africa (Fig. 2). When considering high quality studies, prevalence ranges varied from 2.4 to 85.7% (median 20.6%) for pneumonia, 2.4–41.2% (median 20.0%) for BSI, 0.5–28.6% (median 19.8%) for UTI, and 2.3–69.7% (median 12.4%) for SSI (Table 2). We performed post-hoc sensitivity testing by excluding studies of poor and moderate quality when computing the pooled prevalence. Prevalence ranges varied only slightly by HCAI types even amongst high quality studies. Pooled random-effect summary estimates were calculated using good quality studies and stratified by HCAI type to determine if this would explain the heterogeneity observed with the following results: pneumonia (29% [95% CI: 8–51%]), BSI (21% [95% CI: 14–28%]), UTI (17% [95% CI: 10–23%]), SSI (18% [95% CI: 14–21%]). Significant heterogeneity persisted in each stratum, as evidenced by I2 values ranging from 93 to 99% (Supplement pp16-17). There was not enough data to pool estimates for each HCAI syndrome by region.
Table 2
Prevalence ranges of HCAI syndromes in random effects model
Type of HCAI | Prevalence ranges (%) (Median) | I2 (%) | Tau-square | P-value |
|---|---|---|---|---|
Before sensitivity testing | ||||
PN | 2.4–85.7 (20.6) | 98 | 0.0538 | < 0.01 |
BSI | 2.4–62.6 (19.8) | 98 | 0.0192 | < 0.01 |
UTI | 0.5–65.6 (16.9) | 97 | 0.0189 | < 0.01 |
SSI | 1.8–70.5 (13.6) | 98 | 0.0204 | 0 |
Unknowna | 1.6–63.4 (14.9) | 99 | 0.0167 | 0 |
After 1st sensitivity testingb | ||||
PN | 2.4–85.7 (20.6) | 97 | 0.0594 | < 0.01 |
BSI | 2.4–48.9 (19.5) | 89 | 0.0123 | < 0.01 |
UTI | 0.5–51.4 (16.9) | 97 | 0.0151 | < 0.01 |
SSI | 1.8–70.5 (11.7) | 98 | 0.0213 | 0 |
Unknowna | 3.9–63.4 (14.0) | 98 | 0.0178 | < 0.01 |
After 2nd sensitivity testinga | ||||
PN | 2.4–85.7 (20.6) | 98 | 0.0831 | < 0.01 |
BSI | 2.4–41.2 (20.0) | 93 | 0.0124 | < 0.01 |
UTI | 0.5–28.6 (19.8) | 97 | 0.0094 | < 0.01 |
SSI | 2.3–69.7 (12.4) | 97 | 0.0153 | < 0.01 |
Unknowna | 3.9–39.1 (13.5) | 98 | 0.0154 | < 0.01 |
Numerous risk factors associated with HCAI in Africa have been reported (Supplement pp18-29). The exploratory meta-regression analysis identified different significant risk factors (Table 3). These included contaminated and dirty wounds (OR: 1.75, 95% CI: 1.31–2.19), hospital stay more than seven days (1.39, 0.92–1.80), presence of urinary catheter (1.57, 0.35–2.78), endotracheal intubation and mechanical ventilation (1.53, 0.85–2.22), and the presence of vascular catheter (1.49, 0.52–2.45). We did not perform meta-regression of risk factors by HCAI types because we did not have enough data.
Table 3
Exploratory Meta-regression analysis of risk factors associated with HCAI in Africa
Risk factors | Estimated effect size (OR) | 95% CI | Std error | Z-value | P-value | I2 (%) | Q | P-value |
|---|---|---|---|---|---|---|---|---|
Contaminated and dirty wound | 1.75 | 1.31–2.19 | 0.223 | 7.84 | < 0.001 | 0 | 11.33 | 0.88 |
Hospital stays (pre- or post-operative) > 7 days | 1.36 | 0.92–1.80 | 0.225 | 6.02 | < 0.0001 | 72.1 | 94.94 | < 0.0001 |
Presence of urinary catheter | 1.57 | 0.35–2.78 | 0.618 | 2.53 | 0.0114 | 73.8 | 29.4 | < 0.0001 |
Endotracheal intubation and mechanical ventilation | 1.53 | 0.85–2.22 | 0.350 | 4.41 | < 0.0001 | 38.8 | 10.77 | 0.15 |
Presence of PVC or CVC | 1.49 | 0.52–2.45 | 0.492 | 3.02 | < 0.0025 | 62.3 | 16.14 | 0.013 |
Middle incision | 1.49 | 0.66–2.32 | 0.430 | 3.50 | 0.0005 | 0 | 0.46 | 0.98 |
Haemoglobin level < 11 g/dl | 1.42 | 0.78–2.05 | 0.322 | 4.39 | < 0.0001 | 0 | 1.89 | 0.98 |
History of surgery in the past 30 days | 1.26 | 0.76–1.76 | 0.256 | 4.92 | < 0.0001 | 0 | 3.08 | 0.93 |
Blood transfusion | 1.26 | 0.35–2.17 | 0.464 | 2.71 | 0.0067 | 40.9 | 9.41 | 0.09 |
Duration of procedure > 60 min | 1.10 | 0.55–1.65 | 0.281 | 3.91 | < 0.0001 | 2.83 | 12.43 | 0.26 |
Membrane rupture > 12 h | 1.08 | 0.60–1.57 | 0.248 | 4.37 | < 0.0001 | 28.6 | 10.33 | 0.28 |
Multiple vaginal examinations | 1.02 | 0.23–1.81 | 0.402 | 2.53 | 0.0113 | 0 | 2.46 | 0.48 |
Duration of labour > 24 h | 1.01 | 0.39–1.63 | 0.318 | 3.18 | 0.0015 | 0 | 3.16 | 0.79 |
Hypertension and hypertensive disorders | 0.92 | 0.19–1.64 | 0.368 | 2.49 | 0.0128 | 0 | 2.64 | 0.62 |
Diabetes | 0.88 | 0.38–1.39 | 0.256 | 3.42 | < 0.0006 | 0 | 9.72 | 0.37 |
Forty-eight studies reported microbiology results from HCAI. Overall, 6463 isolates were reported: 2145 (33.2%) from SSI, 1995 (30.8%) from BSI, 593 (9.2%) from UTI, and 527 (8.2%) from pneumonia (Table 4). The source of the remaining 1203 (18.6%) isolates was not identified. The most frequently reported bacteria causing HCAI were E. coli (18.3%, n = 1182), S. aureus (17.3%, n = 1118), Klebsiella spp. (17.2%, n = 1115), Pseudomonas spp. (10.3%, n = 671), and Acinetobacter spp. (6.8%, n = 438). The bacteria most isolated from SSI were E. coli (25.3%, 543/2145), S. aureus (20.6%, 443/2145), and Klebsiella spp. (11.8%, 252/2145). The microorganisms most commonly isolated from UTI were E. coli (24.8%, 147/593), Candida spp. (17.4%, 103/593), and Klebsiella spp. (13.8%, 82/593). The bacteria most frequently isolated from BSI were Klebsiella spp. (28.5%, 568/1995), CoNS (11.5%, 229/1995), S. aureus (11.4%, 227/1995), Acinetobacter spp. (10.0%, 200/1995), and E. coli (7.8%, 156/1995). The bacteria most frequently isolated from patients with pneumonia HCAI were Acinetobacter spp. (25.4%, 134/527), Klebsiella spp. (21.6%, 114/527), and Pseudomonas spp. (18.6%, 98/527) (Table 4).
Table 4
Bacteria reported in different HCAI clinical syndromes
Bacteria | SSI, n (%) | UTIa, n (%) | BSIb, n (%) | PNc, n (%) | Unknownd, n (%) | Total, n (%) |
|---|---|---|---|---|---|---|
Acinetobacter spp. | 44 (2.1) | 26 (4.4) | 200 (10.0) | 134 (25.4) | 34 (2.8) | 438 (6.8) |
Citrobacter spp. | 67 (3.1) | 6 (1.0) | 6 (0.3) | . | 45 (3.7) | 124 (1.9) |
CoNS | 119 (5.5) | 21 (3.5) | 229 (11.5) | 3 (0.6) | 54 (4.5) | 426 (6.6) |
E. coli | 543 (25.3) | 147 (24.8) | 156 (7.8) | 40 (7.6) | 296 (24.6) | 1,182 (18.3) |
Enterobacter spp. | 23 (1.1) | 16 (2.7) | 68 (3.4) | 16 (3.0) | 32 (2.7) | 155 (2.4) |
Enterococcus spp. | 92 (4.3) | 52 (8.8) | 117 (5.9) | 6 (1.1) | 107 (8.9) | 374 (5.8) |
Klebsiella spp. | 252 (11.8) | 82 (13.8) | 568 (28.5) | 114 (21.6) | 99 (8.2) | 1,115 (17.2) |
Kluyvera spp. | 5 (0.2) | . | . | . | . | 5 (0.1) |
Morganella morganii | 5 (0.2) | 1 (0.2) | 3 (0.1) | 1 (0.2) | . | 10 (0.2) |
Proteus spp. | 251 (11.7) | 25 (4.2) | 3 (0.1) | 15 (2.8) | 22 (1.8) | 316 (4.9) |
Pseudomonas spp. | 214 (10.0) | 46 (7.8) | 148 (7.4) | 98 (18.6) | 165 (13.7) | 671 (10.3) |
S. aureus | 443 (20.6) | 61 (10.3) | 227 (11.4) | 72 (13.7) | 315 (26.2) | 1,118 (17.3) |
Serratia spp. | 11 (0.5) | 4 (0.7) | 106 (5.3) | 4 (0.8) | 2 (0.2) | 127 (2.0) |
Stenotrophomonas maltophilia | 1 (0.1) | . | 11 (0.6) | . | 15 (1.3) | 27 (0.4) |
Streptococcus spp. | 62 (2.9) | 3 (0.5) | 78 (3.9) | 9 (1.7) | 1 (0.1) | 153 (2.3) |
Haemophillus influenzae | . | . | . | 11 (2.1) | . | 11 (0.2) |
Candida spp. | 13 (0.6) | 103 (17.4) | 75 (3.8) | 4 (0.8) | 16 (1.3) | 211 (3.3) |
Total | 2145 | 593 | 1995 | 527 | 1203 | 6463 |
Twenty studies reported bacterial antimicrobial susceptibility profiles. Bacteria commonly exhibited resistance to multiple antibiotics. A concerning level of resistance to third generation cephalosporins (70.3%, IQR: 50–100) was observed among Enterobacterales; 70.5% (IQR: 58.8–80.3) S. aureus were Methicillin Resistant; and 55% (IQR: 27.3–81.3) Pseudomonas spp. were resistant to all agents tested (Table 5). We did not perform AMR analysis for HCAI types nor by UN African region because most studies reported the overall AMR and not per HCAI types and because there were only 20 studies.
Table 5
Median resistance rates with interquartile ranges of selected bacteria to selected tested antibiotics
Ciprofloxacin | Ceftriaxone | Ampicillin | Coamoxiclav | Gentamicin | Amikacin | Imipenem | Meropenem | Cotrimoxazole | Tetracycline | Chloramphenicol | |
|---|---|---|---|---|---|---|---|---|---|---|---|
Acinetobacter spp. | 44 (16–72) | 100 (36–100) | 100 (99.2–100) | 100 (75–100) | 50 (11.4–75) | 20 (10–72.4) | 11 (0–30) | 0 (0–33.3) | 75 (28.6–80) | 56 (50–80) | 100 (44.4–100) |
Citrobacter spp. | 100 (45–100) | 100 (25–100) | 100 (81.8–100) | 76 (45.5–100) | 50 (26–100) | 0 (0–25) | . | 0 (0–45.5) | 75 (25–100) | 88 (72.7–100) | 36 (27.3–100) |
CoNS | 27.3 (0–80) | . | 20 (0–50) | 60 (23–100) | 13 (0–35) | . | . | . | 75 (63.6–100) | 55 (0–81.2) | 32 (0–60) |
E. coli | 38 (22.3–77.8) | 70.6 (64–100) | 89 (77.3–100) | 72.7 (28.6–85.7) | 45 (31.4–66.7) | 11.8 (0–19.4) | . | 0 (0–16.7) | 75 (57.1–100) | 78 (58.3–87.5) | 50 (28.6–66.7) |
Enterobacter spp. | 4.9 (0–100) | 50 (50–90.2) | 90.2 (85.7–100) | 81 (50–100) | 50 (0–87.8) | 13 (0–25) | 9.8 (9.8–13) | . | 100 (80.5–100) | 57.1 (50–57.1) | 42.9 (0–50) |
Enterococcus spp. | 50 (50–60) | . | 60 (40–72.2) | 67 (23–67) | 62.3 (50–100) | . | . | . | . | 80 (61.1–83.3) | 36 (18–50) |
Klebsiella spp. | 50 (14.3–81.8) | 70 (61–94.6) | 100 (91.6–100) | 75 (60–100) | 62.5 (50–75) | 9.1 (0–50) | 0 (0–9) | 0 (0–11.8) | 76 (62.5–90) | 88 (50–89) | 36 (29.4–67) |
Proteus spp. | 63 (25–100) | 60 (50–64) | . | 55.6 (50–73) | 50 (22.2–50) | 0 (0–63) | 0 (0–38) | 0 (0–33.3) | 88 (0–100) | 100 (77.8–100) | 44.4 (44.4–100) |
Pseudomonas spp. | 30 (19.2–58.3) | 60 (0–75) | 100 (84.6–100) | 100 (53.9–100) | 50 (22–75) | 24.6 (0–25) | 16.7 (0–33.3) | 23.1 (0–33.3) | 88.5 (12.5–100) | 74 (7.5–77.3) | 34.6 (12.5–100) |
S. aureus | 33.3 (16.1–56) | 70.5 (58.8–80.3) | 71 (71–96) | 81 (21.4–90) | 33.3 (0–45) | . | . | . | 71.4 (64–100) | 38.7 (33.3–55) | 36 (15–80) |
Discussion
The prevalence of HCAI in Africa is clearly high, and bacteria associated with infections are frequently antimicrobial-resistant. Available data also suggest that risk factors for HCAI in Africa are entirely predictable and can be mitigated through implementation of IPC programs and many tools to address the challenge of HCAI exist.
Overall, the HCAI prevalence ranged between 1.6 and 90.2% (median 15%). These rates are higher than pooled prevalence reported in Europe (6.5%) [6], Southeast Asia (9%) [113], the United States (4%) [114], Australia (9.9%) [115], and comparable to the pooled prevalence reported in a previous meta-analysis from developing countries (15.5%) [7]. The high prevalence of HCAI in Africa could be due to inadequate infection control and prevention measures which are often hindered by limited capacity for infection prevention and control, poor laboratory support, and limited funding [13]. Previous studies have demonstrated suboptimal adherence to hand hygiene protocols among healthcare workers in Africa that can be attributable to factors such as absence of safe water in healthcare facilities, inadequate healthcare built environment, inadequate knowledge and training, lack of personnel, and heavy workload [13, 116, 117].
Prevalence of HCAI did not vary much when analysed by HCAI types, with pneumonia, BSI, and UTI having medians of 20–21% and SSI 12%. These infections are largely associated with medical devices and can be prevented through appropriate infection prevention and control measures. Hand and environmental hygiene as well as injection safety practices should be promoted in African healthcare facilities. Contrary to previous studies that reported SSI as the most common HCAI in healthcare facilities in Africa, our study showed that SSI has a lower median prevalence than other HCAI, although high heterogeneity was present [8, 118]. After sensitivity testing that excluded studies with poor and moderate quality, this ranking remained the same. There is a need for more routine data to mitigate the impact of bias. Nevertheless, this prevalence is high and lack of adequate infection control before, during and after a surgical procedure coupled with non-compliant surgical antimicrobial prophylaxis are areas that can be addressed to reduce SSI in Africa [116].
Despite the increase in HCAI surveillance studies from the previous systematic review, overall quantity and quality of HCAI data remain poor. CDC and ECDC HCAI surveillance definitions are currently the most widely used, however, applying these definitions in African settings is typically difficult because they require diagnostic facilities such as microbiological laboratories and complex imaging (CT scan, MRI), which are frequently not available [8, 15, 119, 120]. Consequently, HCAI surveillance remains a significant challenge and the burden of HCAI is poorly described in Africa [3]. Comparing HCAI rates within and between countries is critical for raising awareness about HCAI and its prevention and control; however, it necessitates standardised approaches, including uniform definitions. The World Health Organisation (WHO) and US Centers for Disease Control and Prevention are actively addressing this issue by endeavouring to develop and validate a set of definitions and diagnostic criteria for different HCAI syndromes that will be useful in the absence of a full range of diagnostic microbiology or radiology facilities. The WHO has the authority to advocate for the adoption of standard definitions and this will facilitate the integration of African data into broader international datasets.
None of the risk factors for HCAI in Africa were a surprise and were consistent with global data [121]. Patients with multiple comorbidities or complicated chronic illnesses are more susceptible to frequent hospitalisation, which increases their risk of HCAI and colonisation or infection with multidrug-resistant pathogens. To mitigate the burden of HCAI, several strategies can be implemented [121]. There is already evidence globally, that interventions such as hand hygiene, environmental cleaning, surveillance, and multimodal approaches are cost-effective for the prevention and control of HCAI [122]. Overall, implementing a multimodal approach to HCAI prevention and control in Africa is necessary to reduce the burden of these infections and mitigate the risks to patient safety, and surveillance to quantify the problem and guide local action is key. Implementing a multimodal approach to IPC including enhancing healthcare worker training, implementing evidence-based IPC intervention bundles, and establishing effective surveillance systems sits at the heart of the WHO Core Components of IPC strategy and can be used to reduce the burden of these infections [15]. It is anticipated this holistic strategy will help to address the complex challenges associated with HCAI in Africa, promoting patient safety and contributing to the overall improvement of healthcare systems on the continent.
The most frequently reported bacteria were often resistant to multiple antimicrobial agents that are commonly available in most countries in Africa. These bacteria are typical nosocomial pathogens and among the priority AMR bacterial pathogens identified by WHO. Resistance to fluoroquinolones and β-lactam antibiotics (i.e. cephalosporins, penicillins, and carbapenems), which are typically first-line empirical therapy for severe infections poses a significant threat to patient safety, as AMR further complicates HCAI treatment.
Limitations of this study include the limited number of studies, especially of studies that reported bacterial aetiology and associated AMR profile. Most papers reported aggregate data that were not broken down by clinical speciality or hospital department or age. Further there was an absence of studies from the majority of African countries. Many studies have not described the effect size measures (OR, p-value, and confidence interval) of risk factors. In addition, most studies were conducted in teaching hospitals, with minimal data from district hospitals. Quality of the included studies was highly heterogeneous, however, we addressed this by sensitivity tests and ranking of reported HCAI rates was not affected. Lack of consistency in methods for assessing AMR make these data challenging to compare across studies, nor was there enough AMR data to report by HCAI types. This study did not report on the mortality due to HCAI as majority of included papers did not report this measure.
Conclusions
Here, we provide a comprehensive review of the current HCAI situation in Africa. HCAI are a major problem in Africa, with a high prevalence, multiple risk factors, and increasing resistance to antimicrobial agents. It is imperative to build on this work by developing and validating HCAI definitions adapted to African settings and there is a pressing need to move HCAI surveillance beyond the realm of research studies and establish it as part of routine practice including in primary and secondary healthcare facilities. These measures are essential for policymakers to develop, evaluate, and improve appropriate HCAI prevention and control interventions to reduce the HCAI burden in Africa, AMR, and ultimately enhance the quality of healthcare in Africa.
Acknowledgements
We thank the LSTM librarians for their help in refining our search strategies.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.