Nasopharyngeal carriage rate, antimicrobial susceptibility pattern, and associated risk factors of Streptococcus pneumoniae among children in Ethiopia: a systematic review and meta-analysis (2024)

Background

Nasopharyngeal carriage of S. pneumoniae is a global health problem that has been associated with the emergence of severe disease and pathogen dissemination in the community. However, summary data on the carriage rate, antimicrobial susceptibility profile, and determinant factors is lacking.

Method

Articles were extensively searched in bibliographic databases and gray literature using entry terms or phrases. Studies meeting eligibility criteria were extracted in MS Excel and exported to STATA version 17 software for statistical analysis. A random-effects model was used to compute the pooled magnitude of the nasal carriage of S. pneumoniae and its multidrug resistance. The heterogeneity was quantified by using the I² value. Publication bias was assessed using a funnel plot and Egger’s test. Sensitivity analysis was done to assess the impact of a single study on the pooled effect size.

Result

Of the 146 studies identified, 8 studies containing a total of 3223 children were selected for meta-analysis of the magnitude of the nasal carriage of S. pneumoniae and its multidrug resistance. The overall pooled prevalence of nasal carriage of S. pneumoniae and its MDR status in Ethiopian children was 32.77% (95%CI: 25.1, 40.44). and 31.22% (95%CI: 15.06, 46.84), respectively. The highest resistant pattern of S. pneumoniae was against tetracycline, which was 46.27% (95%CI: 37.75, 54.79), followed by 45.68% (95%CI: 34.43, 57.28) trimethoprim-sulfamethoxazole, while the least pooled prevalence was against chloramphenicol, which was 16.2% (95%CI: 9.44, 22.95). The pooled effect of age less than 5 years old (pooled OR = 1.97; 95% CI: 1.35, 2.88, P < 0.001), co-sleeping habit with others (pooled OR = 2.36; 95% CI: 1.77, 3.66; P < 0.001), sibling (pooled OR = 1.82; 95% CI: 1.14, 2.91, P = 0.01), history of hospitalization (pooled OR = 4.39; 95% CI: 1.86, 10.34, P = 0.001), and malnutrition (pooled OR = 2.18; 95% CI: 1.49, 3.19; P < 0.001) showed a statistical association with S. pneumoniae nasal carriage rate by using the random effect Sidik-Jonkman model.

Conclusion

The magnitude of the nasopharyngeal carriage rate and multi-drug resistance status of S. pneumoniae alarms the need for immediate interventions such as strengthening antimicrobial stewardship programs, undertaking national antimicrobial surveillance, one-health initiatives, and national immunization programs.

Streptococcus pneumoniae (S. pneumoniae) is a Gram positive, extracellular fastidious bacteria that colonizes the human upper respiratory tract [1]. It is transmitted by having close contact with the airway secretions of carriers, especially younger children. Approximately 27–65% of children are reported to be carriers of S. pneumoniae [2]. Carriers can shed S. pneumoniae in nasal secretions and thereby transmit the bacterium [2]. S. pneumoniae is a major cause of pneumonia, meningitis, sepsis, and other serious infections in children with high morbidity and mortality [3].

In 2000, it was estimated that 11.1 to 18million episodes of serious pneumococcal disease and 826 000 deaths occurred in children, with the majority of causes reported in Africa and Asian countries. It is responsible for the deaths of 294,000 human immunodeficiency virus (HIV)-uninfected children aged 0–59 months old from 2000 to 2015 [4]. Approximately 1.6million people, including 1million under five children, died due to pneumococcal pneumonia, meningitis, and sepsis, of which, majority occurred in developing countries [5].

Nasopharyngeal carriage of S. pneumoniae plays a vital role in the development and transmission of pneumococcal diseases [6]. Nasopharyngeal colonization of S. pneumoniae is the prerequisite for invasive pneumococcal disease and plays a vital role in the development of antimicrobial resistance [7]. The invasive capacity of S. pneumonia depends on the serotype and the age of the infected children [8]. Antibiotic resistance and evasion of vaccine-induced immunity by S. pneumoniae are favored by its ability to remodel its genome by taking and incorporating exogenous DNA from other pneumococci and closely related streptococci [2]. Approximately 100 pneumococcal serotypes are identified, and most of them are virulent and are able to cause invasive pneumococcal disease in children [9, 10]. Inclusion of pneumococcal conjugate vaccine (PCV) in the national immunization strategy has reduced the incidence of pneumococcal infections and the circulation of drug-resistant strains in both developed and developing countries [5].

Different studies were done in Ethiopia. However, summary data on the carriage rate among healthy children, the antimicrobial susceptibility profile, serotype distribution, and determinant factors is lacking. Additionally, the current estimates of S. pneumoniae carriage in Ethiopia are not known. Thus, the main objective of this systematic review and meta-analysis is to determine the national pooled nasopharyngeal carriage rate of S. pneumoniae, its antimicrobial susceptibility profiles, and to identify risk factors associated with the nasopharyngeal carriage rate of S. pneumoniae in Ethiopia. The data obtained from this study indicate the current burden of S. pneumoniae and its antimicrobial susceptibility pattern, which is vital for assessing the effectiveness of PCV and antimicrobial agents against S. pneumoniae, and its associated morbidity and mortality. It will also help to monitor the changing trends of S. pneumoniae and attempt interventions to cease these trends.

Search strategy

Electronic databases such as Google Scholar, PubMed. Web of Science, Scopus, ResearchGate, and Google were used to systematically search to identify relevant articles published from 1/1/2000 to 31/2/2024 (Supplementary File 1). The searching was done according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [11]. The comprehensive searching strategy employed used condition, context, population, and outcome of interest (CoCoPop) formulating questions and searching terms, which were nasopharyngeal carriage rate, magnitude, prevalence, epidemiology, incidence, burden, antimicrobial resistance, antimicrobial susceptibility profiles, serotype distribution, Streptococcus pneumoniae, S. pneumoniae, pneumococcus, pneumococcal diseases, invasive pneumococcal disease, invasive pneumococcal infection, pneumococcal pneumonia, PCV, determinants, risk factors, factors associated, predictors, children, pediatrics, all regions and administrative cities of Ethiopia, and Ethiopia combined by the balloon operators. Boolean operators such as “OR” and “AND” were appropriately used to combine search terms to fit the advanced searching of articles based on the formulation of the research questions. In addition, the conference proceedings and bibliography of the identified studies were checked to include studies that had been missed during the database searches.

Research question

What is the pooled prevalence of S. pneumoniae and its antimicrobial susceptibility pattern among Ethiopian children?

• Condition (Co): S. pneumoniae.

• Context (Co): Ethiopia.

• Population (Pop): Children.

Study selection and quality assessment

The quality of the included studies was evaluated independently by two authors (DGF and HE) using the Joanna Briggs Institute (JBI) critical appraisal checklist customized for cross-sectional studies reporting prevalence [12] (Supplementary File 2). The JBI checklist for quantitative studies assesses study quality based on eight criteria, including bias, confounding, the validity of measurement of exposures and outcomes, and the validity of methods of analysis. Using the critical appraisal checklists, studies were reviewed, and articles with an average score of 50–75% were considered good quality, while those with a score greater than 75% were defined as high quality. Disagreements between the two reviewers were resolved by discussion. Finally, articles of good and high quality were included in this systematic review and meta-analysis.

Eligibility criteria

Two reviewers (HB and HD) independently identified eligible studies based on the indicated inclusion and exclusion criteria. The two reviewers select the article first based on the information in the title, abstract, and full text. Disagreements were resolved through discussion and reassessing the studies together. In this study, original articles; cross-sectional, prospective, or retrospective studies that reported the nasopharyngeal carriage rate of S. pneumoniae, its antimicrobial susceptibility profiles, and serotype distribution in Ethiopia among children aged from 2 months to 12 years old were included. However, studies involving adults as their study subjects, data on specific diseases, reviews, case reports, and meta-analyses were excluded. Studies missing the relevant information to be extracted were also excluded.

Outcome variables

The outcomes of this study are the magnitude of the nasopharyngeal carriage rate of S. pneumoniae; the magnitude of multidrug resistant S. pneumoniae, and the antimicrobial susceptibility pattern of S. pneumoniae, as well as factors associated with the nasopharyngeal carriage rate of S. pneumoniae among children in Ethiopia.

Data extraction

The data was extracted by two independent authors (HB and AG) using a predefined extraction sheet in Microsoft Excel. The extracted data were the name of the first author, year of publication, region where the study was conducted, study group, sample size, study design, diagnostic methods, nasopharyngeal carriage rate, and factors associated with the carriage rate such as age, living room, co-sleeping habit, sibling, history of hospitalization, malnutrition and passive smoking history, exclusive breast feeding, passive exposure to cigarette smoking, vaccination status, residence, and antibiotic use.

Data management and analysis

Microsoft Excel was used for data cleaning and imputation. Statistical analysis and modeling were performed using STATA version 17 software. Due to the observed heterogeneity, a random effect meta-analysis model was used to calculate the overall pooled effect size with the corresponding 95% confidence interval [13]. The Cochrane chi-square (x²) test and quantification with the I² statistic were used to calculate the source of heterogeneity [14, 15]. Despite the DerSimonian and Laird (DL) approach being the simplest and most commonly applied between-study variance estimator, it can’t provide a reliable statistical estimation due to the small number of studies included in the meta-analysis, unequal size of studies, and the high level of heterogeneity between studies. Thus, the Sidik and Jonkman (SJ) method was applied to overcome the limitations of the DL method. SJ is a more reliable two-step estimator that yields positive estimates of heterogeneity variance [16,17,18]. Publication bias was assessed using the symmetry of the funnel plot and the Egger’s test statistics, with the symmetry of the funnel plot and a p-value > 0.05 in the Egger’s test indicating the absence of publication bias [19, 20]. Sensitivity analysis was done to evaluate the impact of a single study on the pooled effect size. The findings were presented in tables and figures.

Literature search and eligible studies

During the initial electronic search and manual search of reference lists, a total of 146 articles were retrieved. Thirty-nine articles were excluded because of duplication. After reviewing their title and abstract, 96 articles were excluded due to unrelated titles, not being studied in Ethiopia, and being review articles. Then, the full text of 11 studies was evaluated for eligibility criteria. Then 3 articles were excluded because the study participants were all in all age groups, the study groups were patients with acute respiratory infection, the samples were not nasopharyngeal swabs; and they assessed community acquired sepsis. Finally, after methodological quality assessment, eight studies were included in this systematic review and meta-analysis (Fig. 1).

Characteristics of the included studies

The studies were conducted from 2013 to 2024 in three national regional states of Ethiopia. Of these, four studies were conducted in the Southern Nations, Nationalities, and People Region (SNNPR), two in the Oromia regional state, and two in the Amhara national regional state. All studies were cross-sectional and published in English. All studies enrolled in this study were conducted on healthy individuals and assessed the nasopharyngeal carriage rate of S. pneumonia. In all the studies, a total of 3,223 children with ages ranging from 2 months to 13 years old were involved. From the 3,223 children, 1,625 were males and 1,598 were females. Five studies use STGG, and two studies use Amens transport media for sample transport. All the studies used culture for the diagnosis of S. pneumoniae and disc diffusion for antimicrobial susceptibility testing. One thousand and eighty-one S. pneumoniae isolates were evaluated for antimicrobial susceptibility patterns (Table1). Eight studies reported the antimicrobial susceptibility patterns of S. pneumoniae to erythromycin and chloramphenicol and its multidrug-resistance pattern (MDR). One study reported pan-drug resistance (non-susceptibility to all agents in all antimicrobial categories). Seven studies assessed the antimicrobial susceptibility pattern of S. pneumoniae to tetracycline, and five studies assessed oxacillin and trimethoprim sulfamethoxazole. The antimicrobial susceptibility pattern of S. pneumoniae to penicillin and cotrimoxazole was reported (Table2).

Nasopharyngeal carriage rate

The prevalence of S. pneumoniae nasal carriage ranged from 18.4% in Oromia regional state to 43.8% in SNNPR. The overall national pooled prevalence of nasopharyngeal carriage of S. pneumoniae among children using a random effect model was 32.77% (95% CI: 25.1, 40.44). A significant heterogeneity was observed with I² value of 95.7% and a heterogeneity chi-squared of 164.05 (Fig. 2). Subgroup analysis was done for the region where the studies were conducted and the gender of the study participants. The pooled nasopharyngeal carriage rate of S. pneumoniae was highest in the Amhara region (35.21%), followed by SNNPR (32.45%), and Oromia region (31.03%) (Fig. 3). A comparable nasopharyngeal carriage rate of S. pneumoniae was seen among male (16.56%) and female (16.01%) participants (Fig. 4).

Antimicrobial resistance profile

In this systematic review, an antibiotic resistant pattern was evaluated for 1,081 S. pneumoniae isolates. All the studies use the disc diffusion method. Antimicrobial resistant patterns of S. pneumoniae were assessed for tetracycline, oxacillin, erythromycin, penicillin, trimethoprim-sulfamethoxazole, and chloramphenicol (Table2). The results of the included studies revealed a variable resistance pattern to different antimicrobial agents. the highest resistant pattern of S. pneumoniae was against tetracycline, which was 46.27% (95% CI: 37.75, 54.79), followed by 45.68% (95% CI: 34.43, 57.28) trimethoprim sulfamethoxazole. The pooled prevalence of S. pneumoniae resistance to oxacillin and erythromycin was 37.41% (95% CI: 30.87, 43.95) and 31.22% (15.6, 48.64), respectively. The least pooled prevalence of S. pneumoniae resistance was against chloramphenicol, which was 16.2% (95% CI: 9.44, 22.95). A high level of heterogeneity was observed in all analyses, with an I² value ranging from 72.4 to 97.8% (Table3).

Multidrug resistance pattern of S. pneumoniae

All the included studies reported the prevalence of multidrug resistant (MDR) S. pneumoniae. Six studies reported MDR if the bacterial isolates were resistant to three or more antimicrobials. While two studies classify MDR as if the bacteria were resistant to two or more antimicrobials. The highest prevalence of MDR was reported in Amhara regional state (68.8%). On the other hand, the lowest prevalence of MDR was reported in SNNPR at 2.9%. The overall national pooled prevalence of MDR was 31.22% (95% CI: 15.06, 46.84), but the pooled estimate had significantly high heterogeneity with I² of 97.8% (Fig. 5).

Subgroup analysis was done for the way of classification of MDR and the region where the studies were conducted. The pooled prevalence of MDR was higher among studies that defined MDR as resistant to three or more antimicrobials, which was 32.19% (Fig. 6). The lowest pooled prevalence of MDR was in SNNPR (23.11%), while the highest was in the Amhara region (41.65%) (Fig. 7).

Risk factors associated with nasopharyngeal carriage rate

The effect size of risk factors was computed from the number of studies reporting that the factors were two and above. Six studies reported a significant association between infancy age (age less than 5 years old), four studies reported co-sleeping habits with parents, and a history of passive smoking with the S. pneumoniae nasal carriage rate among Ethiopian children. The pooled effect of infancy age (age less than 5 years old), co-sleeping habit with others, sibling, history of hospitalization, and malnutrition showed a statistical association with S. pneumoniae nasal carriage rate by using the random effect Sidik-Jonkman model. The overall analysis showed a significantly increased risk of nasal carriage among children less than 5 years of age and history of hospitalization, despite moderate heterogeneity (pooled OR = 1.97; 95% CI: 1.35, 2.88; I² = 69.65%, p < 0.001) and (pooled OR = 4.39; 95% CI: 1.86, 10.34; I² = 71.88%, p = 0.003), respectively. Likewise, there was no heterogeneity in analyzing the pooled effect of malnutrition (pooled OR = 2.18; 95% CI: 1.49, 3.19; I² = 0.0%, p = 0.98) and co-sleeping habit with parents (pooled OR = 2.36; 95% CI: 1.77, 3.66; I² = 10.78%, p = 0.75) (Table4).

Publication bias and sensitivity analysis

Publication bias was assessed using visual observation of the funnel plot (Fig. 8) and Egger’s test. The symmetry of the funnel plot and the Egger’s test statistics with a p-value of 0.452 confirm the absence of publication bias. In addition, sensitivity analysis was done to see the impact of the individual effect size on the overall pooled effected size. When individual studies were omitted, the resulting pooled effect size was within the 95%CI of the original pooled prevalence (Table5). This confirms the absence of a single study impact on the pooled effect size.

The asymptomatic nasopharyngeal carriage of S. pneumoniae is common in young children. The antibiotic resistance rate of S. pneumoniae, particularly in children, is also an important global issue. Antimicrobial-resistant S. pneumoniae is becoming more common, and its nasal colonization in children increases the risk of pneumococcal illness [29]. Even though the impact of PCV vaccination on the prevention of nasal carriage of S. pneumoniae is a complex issue, effective vaccination coverage is expected to reduce the magnitude of nasal carriage. However, 23million children did not have adequate access to vaccines in 2020, which might be responsible for aggravated nasal carriage rates. Thus, this study evaluated the magnitude of the nasopharyngeal carriage rate, antimicrobial susceptibility pattern, and associated risk factors for S. pneumoniae among children in Ethiopia. In this systematic review and meta-analysis, eight studies that recruited a total of 3,223 children were included. All studies in this review used pneumococcal culture, which was not standardized according to the WHO working group recommendation [30]. Moreover, in this study, Skim Milk Tryptone Glucose Glycerol (SMTGG) was the most commonly used transport medium for the recovery rate of S. pneumoniae.

The pneumococcal nasopharyngeal carriage rate varies according to the different geographical regions. The highest carriage rate was reported in SNNPR, and the lowest was in the Oromia region. The overall nasopharyngeal carriage rate of S. pneumoniae among children was 32.77% (95% CI: 25.1, 40.44). The findings of this study were consistent with previous evidence reported from the South East Asia Region (35%) [31] but higher than 2% [32] and 18% prevalence in Iran [33]. The variation could be due to variations in the study setting, carriage prevalence between studies, type of swabs and transport media used to collect the specimen, and type of specimen itself, as well as the methods used during collection that might influence the recovery rate of S. pneumoniae.

The findings of this meta-analysis showed that the carriage rate of S. pneumoniae was significant, which alarms the need for immediate interventions such as strengthening antimicrobial stewardship programs, undertaking national antimicrobial surveillance, one-health initiatives, and national immunization programs. The implication of this study is that three in every 10 children had nasopharyngeal carriage of S. pneumoniae. Moreover, this nasopharyngeal colonization by the bacterium might be the source of spreading between hosts and cause an invasive disease, depending on the density of the bacterium and the sufficient period for colonization and persistence [29].

The result of this systematic review showed that the Kirby-Bauer disk diffusion method was entirely used for testing the antimicrobial sensitivity of S. pneumoniae isolates. Moreover, the finding revealed that there was a variable resistance pattern of S. pneumoniae isolates to different antimicrobial agents. The highest pooled resistance prevalence of tetracycline and trimethoprim sulfamethoxazole (TMP-SMX) was 46.27% (95%CI: 37.75, 54.79) and 45.68% (95%CI: 34.43, 57.28), respectively.

Now a days, antibiotic resistance among S. pneumoniae isolates is accelerating in Ethiopia. In this meta-analysis, the highest rate of resistance among S. pneumoniae isolates was against tetracycline and trimethoprim sulfamethoxazole (TMP-SMX), while the lowest proportion of antibiotic resistance was against chloramphenicol. The level of resistance among S. pneumoniae isolates against tetracycline and TMP-SMX was consistent with previous evidence from African [34], Middle Eastern countries [35], and SSA [36]. Tetracycline and TMP-SMX are the most widely used antimicrobials in most public and private health care settings to treat respiratory tract infections empirically. These conditions might be linked to the highest rate of resistance to these antimicrobials. Thus, according to this finding, tetracycline and TMP-SMX might not be effective treatment regimens for pneumococcal infections. This finding implies the utilization of antimicrobials based on their level of efficacy, the use of other alternative antimicrobials, and strict regulation of pharmacies and drug stores to limit over-the-counter access and ensure prescription-based treatment. Moreover, effectively implementing antimicrobial stewardship could be crucial to preserving the efficacy of our current drugs. The overall national pooled prevalence of multi-drug resistance among S. pneumoniae isolates was 31.22%. This finding implies that nearly one-third of S. pneumoniae isolates have been developing resistance to antimicrobials, which could be linked to mobile genetic elements such as plasmids and transposons [37], the sharing of medication, a lack of effective therapeutics, a lack of medical practitioners with proper training, the evolution of bacteria, environmental conditions, as well as poor-quality and unhygienic sanitary conditions [38]. Moreover, the emergence of antimicrobial-resistant strains might be due to the indiscriminate use of antibiotics in managing bacterial infections without having determined the local antimicrobial sensitivity result. The findings of this study alarm the need for immediate actions to reduce the burden of MDR through different strategies such as the design and implementation of antimicrobial stewardship programs; multisectoral collaboration like ONE HEALTH initiatives, and the implementation of national antimicrobial surveillance programs.

In this study, certain socio-demographic and clinical factors have been statistically associated with the S. pneumoniae carriage rate. The odds of children under 5 years of age are nearly twice (pooled OR = 1.97; 95% CI: 1.35, 2.88) higher than those of younger children. An increased S. pneumoniae carriage rate associated with decreasing age might be due to the immaturity of the immune status of children, which will be alleviated through the gradual development of mucosal immunity and a progressive increase in the clearance rates of microbes. In addition, children who have more siblings than their counterparts had higher odds of developing S. pneumoniae. The history of hospitalization was found to be a significant predictor of a higher rate of S. pneumoniae carriage. This could be supported by evidence that antibiotic resistant S. pneumoniae infections require higher doses of antibiotics, a longer duration of treatment and hospitalization, the use of more expensive medications, or the use of medications with greater side-effect potential [39].

Furthermore, malnutrition (such as stunting, wasting, and being underweight) had a higher probability of pneumococcal colonization, which might be due to the impairment of the immune system and lead to persistent and recurrent colonization with pneumococci. Additionally, children who had a habit of sleeping with parent(s) or guardians were 2.36 times more likely to carry S. pneumoniae in their nasopharynx compared to those who did not (pooled OR = 2.36; 95% CI: 1.77, 3.66). This could be due to the fact that the likelihood of horizontal transmission of respiratory secretions; and inhaling the breath and nasal secretions containing the causative agent, as well as prolonged exposure to the causative agent, will increase the likelihood of nasal colonization.

Publication bias was assessed by using the visual funnel plots and Egger’s test statistics. Both methods confirmed the absence of publication bias. Future studies shall focus on circulating serotype distribution and molecular mechanisms of resistance to inform effective vaccination strategies, monitor the impact of PCV, and understand the genetic basis of antimicrobial resistance in S. pneumoniae. This study has certain limitations, such as the lack of studies in all regions of the country, which might affect the generalizability of the findings. Although I² is not an absolute measure of heterogeneity, high heterogeneity was observed. Therefore, to limit the influence of the study heterogeneity, the random effect Sidik-Jonkman model [16,17,18] was used, and subgroup analyses and sensitivity analyses were performed to rule out study variability.

AMR:

Antimicrobial Resistance

MDR:

Multidrug-resistance

PCV:

Pneumococcal Conjugate Vaccine

PRISMA:

Preferred Reporting Items for Systematic Review and Meta-analysis

SMTGG:

Skim Milk Tryptone Glucose Glycerol

SNNPR:

Southern Nations, Nationalities, and People Region

TMP-SMX:

trimethoprim sulfamethoxazole

Authors and Affiliations

1. Department of Medical Laboratory Sciences, College of Medicine and Health Sciences, Wollo University, Dessie, Ethiopia

   Habtye Bisetegn,Habtu Debash,Ousman Mohammed,Ermiyas Alemayehu,Hussen Ebrahim,Mihret Tilahun&Alemu Gedefie

2. Institute of Biosciences, Department of Chemical and Biological Sciences, Sao Paulo State University, Sao Paulo, Brazil

   Habtye Bisetegn

3. Department of Microbiology, Immunology and Parasitology, College of Medicine and Health Sciences, Addis Ababa University, Addis Ababa, Ethiopia

   Daniel Getacher Feleke

Contributions

Habtye B, Alemu G, Habtu D, Ousman M, Ermiyas A, Mihret T, Danie Getacher F and Hussen E. involved in the conception and design, data collection, analysis, and interpretation of the data. Habtye B, Alemu G, Habtu D, Ousman M, Ermiyas A, Mihret T, Danie Getacher F and Hussen E took part in drafting the article or revising it critically for important intellectual content. Finally, all authors read and approve the final manuscript before submission.

Corresponding author

Correspondence to Alemu Gedefie.

Ethical approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.

Reprints and permissions