Knowledge gaps on paediatric respiratory infections in Morocco, Northern Africa

A systematic review was conducted using the PRISMA proposed methodology 13]. Pubmed®, Google Scholar®, Hinari®, Sciences Direct®, EM Premium® online databases
were screened, in addition to physical libraries of three Moroccan medical schools
(Rabat, Casablanca and Fez) in search of unpublished PHD theses. Finally, official
reports of the Moroccan Ministry of Health (MOH) were also searched and publications
from 1997 to 2014 reviewed in both French and English languages. The search query
terms used included: (“Severe Acute Respiratory Infections/epidemiology” OR “Severe
Acute Respiratory Infections/etiology” OR “Severe Acute Respiratory Infections/microbiology”
OR “Severe Acute Respiratory Infections/mortality” OR “pneumonia”) AND “Morocco”).
The data found were identified and selected by two readers: a public health professional
and a paediatrician using the PRISMA flow chart. (Figure 1 summarizes the flow diagram). After the screening, the two readers merged the documents
to study their eligibility using items from the PRISMA checklist (title, objective,
study population, study setting, information sources, process of data collection,
and outcome measures) 13]. Then, they made consensus decision about the inclusion of the eligible documents
in the final analysis. The studies inclusion’s criteria were: reporting original data
on the incidence, distribution, or a clinical description of the diseases or their
etiology or describing medical treatment or national preventive strategies. Studies
were excluded if the documents implied duplication or when, full texts of abstracts
were not found.

Figure 1. Flow diagram for the extraction of the records related to ARI in Morocco among children
published from 1996 to 2014.

Results and discussion

This first screening yielded seventeen original articles published in peer review
journals, 12 in English language and five in French, one original paper was published
in French language in a Moroccan peer reviewed journal. Four abstracts were published
in French language and two in English. Finally three additional MOH reports were found
related to the topic. From the school of medicine libraries’, eight unpublished doctoral
theses were extracted. After analysis of each of those publications, two reporting
duplicate results were excluded and one was found to be not eligible, as it did not
report any original result.

These documents describe paediatric ARI-related surveys in Morocco conducted in the
last 17 years. The data could be classified according to the following three groups:
1) those documents describing severe pneumonia cases admitted to the university hospitals
(7 published, 8 unpublished). 2) Those describing samples obtained from pneumonia
patients and analyzed by laboratories of university hospitals (10, all published),
and finally, 3) National data provided from the annual statistics report from the
MOH including a survey evaluating the IMCI program (one study, published) and a study
conducted within the community (three studies, all published).

Description of national health statistics

The World Health Organization (WHO) reports that ARI remains in Morocco the leading
individual cause of under five mortality, accounting for a total of 13% of all deaths
in this age group in 2012 2]. Figure 2 illustrates the proportion of annual deaths among under five according to age group
(infant and older children) which can be attributable to ARIs, from the period 2005
to 2012. Data show an important decrease of ARI-associated deaths from 2008 to 2011
among older children. No changes can be observed for infants during the same period.
No differences in the distribution of death gender were evidenced (data not shown).
WHO has also estimated that in the year 2012, only half of the children with suspected
pneumonia were taken to an appropriate health provider, and only 49% suffering from
pneumonia got antibiotics 2].

Figure 2. Proportion of under five mortality due to ARI by age group from 2005 to 2012. (Data
extracted from the National Health Statistics annual reports and estimated from death
certificates).

The summarized data from 2007–2012 obtained from Morocco’s National annual health
statistics (Figure 3), show that the most commonly notified types of ARI both in urban and rural areas
are pneumonia cases, followed by severe pneumonia cases. The very severe cases are
less prevalent. The most commonly affected age group included toddlers age 24–59 months.
But the most severe cases occurred in a higher proportion among infants (0–11 months).
The figure also shows the proportion of ARI cases which subsequently were prescribed
antibiotics, with an increasing trend in all age groups throughout the study period.

Figure 3. Evolution of the number of outpatient visits (per age group) to primary health care
centers for acute respiratory infections (ARI) and the proportion of those visits
being prescribed antibiotics among children under five years old at the national level
from 2007 to 2012. Data extracted from the National Health Statistics annual reports.

Figure 4, shows the distribution of the incidence of ARI for the different country regions
for which data are available. Almost 54% of all ARIs are reported from rural areas,
whereas the remaining (46.4%) occur in urban setting.

Figure 4. Map of Moroccan regions for which incidence of pediatric ARI have been summarized
for the year 2012. Data extracted from the National Health Statistics annual report.

In the places where the IMCI program has been implemented, there are some data collected
from the MOH since 1997. The indicators collected estimate the performance of trained
nurses in the diagnosis of pneumonia (whether these episodes diagnosed were severe
or not), and the proportion of those receiving antibiotics. A study of the quality
of IMCI six months after its implementation in two Moroccan regions as a pilot project
showed that following the IMCI program allowed a better quality of care, measured
by better indicators of correct antibiotic prescription, and a better overall classification
of the severity of illness 11].

Clinical trends and characteristics of ARI in Morocco from 1997 to 2014

Different studies were found describing the clinical characteristics of patients (range
7–700) admitted to public paediatric hospitals, including clinical patterns of ARI,
duration of hospitalization, treatments received and outcomes. The reported observational
period in each of these studies ranged from one to ten years. What emerges from these
studies is that: the majorities were descriptive studies; only one being a comparative
study and using statistical analysis. Study populations were also quite distinct,
including in certain studies hospitalized pneumonia cases only, whereas in others,
outpatients with ARI. The inclusion criteria were based on the clinical case definition,
or on radiological or microbiological criteria. Age at recruitment in these different
paediatric studies ranged from a minimum of two months to up to 15 years of age. Most
of these studies, have never been published and are available as doctoral theses,
and only one reports the bacteriological evidence of the causative organism of ARI
in half of its studied cases (5/10= 50%). Table 1, summireses the information collected through these unpublished studies.

Table 1. Unpublished sources of information on ARI in children in Morocco: from 1997 to 2014

The clinical patterns of ARI throughout the past 17 years have varied in the different
studies available. For instance, studies looking at acute lobar pneumonia and pleural
effusion are common from 1998 onwards. However, a thesis produced at the university
hospital of Fez, reported 53 cases of purulent pleurisy between 2006 and 2009 14]. Studies conducted between 1994 and 1998 show that the frequency of viral bronchoalveolitis
accounted for 10% of cases of ARI admitted to university hospital. Patients with acute
lobar pneumonia and bronchopneumonia represented at that time 38% and 40% of the ARI
respectively. However, a recent study 15] reported that among 700 patients admitted to a paediatric ward in Rabat, fulfilling
WHO clinical severe pneumonia case definition, only 28% of them had a clinical diagnosis
of suspected bacterial pneumonia (as given by the clinician in charge) at the time
of discharge. The same study reported that the nasopharyngeal carriage of respiratory
virus was almost 92% and that invasive bacterial disease was identified only in 2%
of patient 15], in spite of universal blood culture and PCR investigation of invasive bacterial
diseases screening among all study participants (n=700). Beside this, and from year
2000 onwards, the information available from University Hospital of Rabat, has only
focused on cases of viral bronchoalveolitis. Conversely, the reports coming from Hospital
of Casablanca, presented nine blood culture confirmed cases of pneumococcal community-acquired
pneumonia and 145 further pneumonia cases without bacteriological confirmation during
an observation period of two years. A further case of pneumonia associated to pleural
effusion secondary to Streptococcus Pneumoniae was reported in a one year-long survey. In 2000, ARIs accounted for 30% of all emergency
consultations and 20% of the total number of admissions in the paediatric infectious
diseases ward. 16] The most affected groups of age were infants and children less than 2 years of age,
representing up to 75% of all paediatrics patients admitted for ARI. This distribution
did not seem to change across the 17 years, since a published study in 2014 conducted
in the same ward showed similar findings, and reported that almost 30% of children
attending the paediatric emergency department had respiratory symptoms, and that 42%
of those would be eventually admitted 15].

In terms of seasonality, ARI’s admissions for bronchiolitis episodes usually peaked
during the winter season, whereas pneumonia and bronchitis are distributed throughout
the year 15]. The most consistent clinical feature of pneumonia described in these studies during
the 17 years long period of assessment included respiratory distress, fever and cough.
The average length of stay ranged usually from two to five days and the associated
in-hospital case fatality rates from one to five percent. Only one study 15], documented the nutritional status of children admitted for ARI, and showed that
5% of these cases occurred in severe malnourished children (according to a weight
for-age Z score WAZ 3- SD).

Description of laboratory data

Morocco lacks a national established system for the monitoring of antimicrobial resistance
of pathogens causing ARI. However, there is some information provided by laboratory
university hospitals. During the period 1994 to 1998, most isolated germs included
in order of frequency Staphylococcus aureus, followed by Streptococcus pneumoniae and Haemophilus influenzae type b. These pathogens were isolated from blood or pleural fluid. The potential
role played by viruses in the etiology of pneumonia was not at all explored before
the epidemic episodes of influenza. Few published studies reported the clinical features
of the epidemic episodes of influenza H1N1 in 2009 17],18], and showed that children had the highest proportion of laboratory-confirmed influenza
through the influenza sentinel surveillance system. Data on other virus (Respiratory syncitial virus (RSV), parainfluenza, 1,2,3 and Adenovirus) explored through this sentinel surveillance are still not available 19].

More recently in 2014, data from a paediatric university hospital reported that the
invasive pneumococcal disease was infrequent, but the nasal and nasopharyngeal carriage
of S. pneumonia was common. 15] In this same study, a full panel of respiratory viruses was screened among all study
participants (n=700), yielding extremely high (90%) nasopharyngeal mono or multiples
infection rates. Most commonly identified viruses included Rhinovirus followed by
respiratory syncytial virus (RSV) and Adenovirus. Importantly, Human metapneumovirus
was also frequently detected and the only of those viral pathogens independently and
significantly associated with an adverse outcome 15].

Regarding the monitoring of antimicrobial susceptibility, some data are available
for Streptoccoccus pneumoniae. Active laboratory surveillance was held from 1994 onwards in the microbiology laboratory
of University Hospital of Casablanca, including all the strains collected from the
wards (adult and paediatric) of three university hospitals located in Casablanca 20]. Among 1152 strains collected over these 14 years of ongoing surveillance, antibiotic
susceptibility for this pathogen decreased significantly, and importantly, antibiotic
resistance was more prevalent among isolates from children. From 1994–1997, 12.5%
of the strains were found to have reduced susceptibility to penicillin, this figure
increasing to 15. 3% from 1998 to 2001 21], to 18. 9% from 2002 to 2005, and to 23. 5% from 2006–2008 22],23]. This study included all types of strains. The same center had contributed in the
collection of data within a regional network (ARMed) 24] during the period 2003 to 2005. Notably, Morocco’s Streptococcus pneumoniae isolates had one of the highest erythromycin resistances within the network. In the
city of Rabat two independent surveys were conducted in two distinct university hospitals
during the periods 1997 to 2001 and 2006 to 2007, also trying to assess the antibiotic
susceptibility patterns of Streptocococcus pneumoniae25] and other respiratory pathogens 26]. The first study showed a low susceptibility to beta-lactams (7,8%), but one of the
resistant isolates had been isolated from the cerebrospinal fluid of an infant with
meningitis secondary to meningeal breach 25]. The second study showed that 2/6 (33.3%) of the pneumococci were fully susceptible to amoxicillin, and 80% to erythromycin. 91 strains of H. influenzae type b were also obtained and investigated 27] with the vast majority (97, 9%) being susceptible to amoxicillin+/clavulanic acid,
and the totality (100%) to cephalosporins. This study showed the lowest resistance
rates to penicillin from Streptococcus pneumoniae observed in Northern Africa 24].

Samples collected in 2010 within a hospital setting, from patient’s under 5 years
of age admitted for severe clinical pneumonia 28], showed that S.pneumoniae isolates had moderately low resistance rates to commonly available antibiotics (Erythromycin
20%, cotrimoxazole 24%, penicillin G 10% and amoxicillin 14.6% ) 28].

The surveillance of pneumococcal serotype distribution among admitted patients in
Casablanca, before the introduction of the pneumococcal vaccine, showed that the most
prevalent circulating serotypes were 19F, 14,6,18 and 9 22]. The same distribution was also confirmed among healthy children, in whom the prevalence
of nasopharyngeal carriage was 45, 8% 29]. The analysis of the carriers, suggested that carriage was more likely in children
who had exclusively breastfed for less than two months, had a low socio.economic background,
or lived in crowded houses 29]. The commonly identified serotypes included 19 F,6,14,23,18 and 9. In a study conducted
in Rabat city 15], coinciding with the year of pneumococcal vaccine implementation, the most prevalent
serotypes isolated among these admitted patients were 6A, 19F and 6B. The current
13-valent vaccine, introduced into the national immunization program in 2011, would
cover up to 85% of isolates.

Clinical management of ARI

Outside the public sector, management of ARI in Morocco is difficult to document,
due to the widespread availability of private practitioners’ in urban areas and some
of rural areas in the country. In these ambulatory private settings there are no official
documents which reflect the private physician’s practices. IMCI guidelines, which
propose a standardized and documented approach to diagnosis and management, have not
been properly assessed at national scale in Morocco, and thus, little information
exists on how they are truly applied in the rural areas. Conversely, in university
hospital settings, specific protocols for the management of ARI based on consensus
exist and are followed.

A recent study conducted in the university hospital of Rabat 28], reports that before their admission, 30% of patients with ARI received antibiotics
within the two weeks preceding their hospitalization. These antibiotics were prescribed
in 86.5% of the cases by a physician. The most common antibiotics used prior to arrival
to hospital, were Amoxicillin/Clavulanic acid, followed by amoxicillin and Macrolides.
The same study documented that once admitted, these same patients often received in
decreasing order of frequency, cephalosporines, macrolides and gentamicin. Most of
these antibiotics were given parenterally.

Determinants, risk factors and prognostic of ARI

Specific data on independent risk factors for ARI are poorly described, and often
never explored in the unpublished studies. However, infants and males appear to be
consistently more at risk of being admitted as result of their ARI.

A recent survey 30], reported the risk factors of poor outcomes among children admitted for clinically
severe pneumonia. The main independent risk factors for death or the requirement of
the admission to the intensive care unit included prematurity, exposure to passive
smoking at home, history of fever, the presence of cyanosis, pallor, ronchi at chest
auscultation, unconsciousness on admission and human metapneumovirus infection. Other
common risk/protective factors described in the literature (breastfeeding, age, malnutrition,
parental and socio economical and educational level) were also explored in this study,
but were not found to be associated in multivariate analysis with an adverse outcome.

One study was conducted within the community to assess the exposure of children to
passive smoking and showed that around 34% of children participating were exposed
at home 31]. Another study conducted in a hospital setting described slightly higher exposure
rates (up to 40%) among children admitted with ARI 15].

The role of indoor pollution in relation to ARI was never explored. However, a single
study exploring the air pollution as a determinant of asthma among schoolchildren
in an industrialized city, was found and suggested that the children living in the
most polluted industrial zone were most affected by respiratory diseases and, in particular,
asthma and that the majority of the affected children were from underprivileged families
and were malnourished 32].

Little anthropological research has been conducted in Morocco to assess whether family
behaviors could affect or influence the management of ARI among children. A single
study was done during 1998 in the region of Marrakech 33]. The study showed that the parental socio economical and educational level, and that
access to the health system, or to transportation had an important influence in the
parent’s behaviors. Illiterate parents predominantly preferred to start treating their
sick children at home and herbal remedies were typically used. Additionally, visits
to the traditional healers, which may also use herbs or more aggressive practices,
including burning of the chest using a piece of burning charcoal through a blanket
were also common. Such parents would only visit the physician as a last resource,
when the previous methods had shown to fail, possibly coinciding with a worsening
condition of their child. In Morocco, the abundance of a private medical sector has
been related to an abuse or misuse of antibiotics, and to challenges related with
the radiography examination of the chest among patients with suspected pneumonia,
possibly in relation to high costs 33].

Available data reviewed regarding acute respiratory infections in Morocco in the past
17 years, offer a blurred picture (Table 2). Indeed, the scarcity of official published or unpublished data regarding the number
one killer of children in Morocco is disquieting. Among the few available reports,
consisting mostly of ad hoc descriptive studies from admitted patients to university hospitals, data are difficult
to compare and to generalize, because the National Health Statistics do only represent
the data collected from the public sectors. The private sector does not participate
to the national health statistics, which cuts off the national statistics of important
data, relating to the characteristics of the population attending private practices
and to their care practices. Moreover, access to the health care system is mostly
depending on the geographical characteristics of the regions and the contribution
of the private health care sector in these areas. So, the distribution of the ARI
in Morocco shown in figure 4, and the distribution of ARI between urban and rural areas, may not reflect the real
burden and distribution of ARI within the country.

Table 2. Keys knowledge gaps regarding paediatric ARIs in Morocco

Additionally, authors of the different papers, have used different case definition,
or involved different study populations. Data produced by hospitals, and in particular
that of the University Hospital of Casablanca, arise from local laboratories offering
microbiological point monitoring data of patients admitted to the clinical services
of their structure. While these local data are not representative of the rest of the
Moroccan population under 5 years of age, the evident scarcity of available data is
the reason behind its usefulness to guide the countries’ first line treatment against
ARI, or similarly, the desired serotype composition of the newly introduced pneumococcal
conjugate vaccine. Better, more geographically representative, detailed etiological
and epidemiological data of the principal pathogens responsible for paediatric ARI
are urgently needed, so as to base in evidence current recommendations for its prevention
and management.

The mortality rate of ARI reported by the national health statistics in 2010, among
children less than one year old (2,5%) is similar to the mortality rate reported globally
(2,3%) and in developing countries (2,4%) 6]. The same report estimates that 4% of death occurs within a hospital setting. However,
these statistics, like in many other countries, should be interpreted with caution,
as the sources from which they are gathered may vary, and because death certificates
are not filled uniformly, leading to potential over or underestimating the IRA as
causes of death.

In Morocco, ARI-associated mortality statistics do not seem to show significant differences
between boys and girls, such as those observed in many Asian countries 6].

The description of the clinical patterns of ARI throughout the last of 17 years shows
a changing profile of the reported types of cases. More than representing a real epidemiological
transition, sporadic reports seem to translate the particular focus or scientific
interests of the team of the ward where the data were extracted. Needless to say,
they do represent an important reporting bias.

The laboratory studies are few in numbers, are recent, and are unrelated, using different
methods, different populations of study and different specimens’ types. As a result,
they are difficult to compare, also, and trends in antimicrobial resistance are challenging
to infer. The laboratory of Casablanca seems to be the single reference laboratory
where long-term surveillance of ARI has been implemented, and has been providing aggregate
data since the year 1999. This essentially microbiological surveillance, relatively
basic at the moment, should be optimized by including data on patient’s characteristics,
their clinical manifestations, the prescribed treatments and outcomes, so as to better
describe the burden of microorganisms isolated and the population at risk. Furthermore,
a careful monitoring should also be conducted in the pilot settings where IMCI is
implemented to follow trends of the antimicrobial sensitivity patterns, particularly
in the light of the suspected overuse of antibiotics, as antibiotic prescription seems
to be also very common practice. Such data could provide evidence-based orientation
among the identity of antibiotics recommended through IMCI, their dose adjustments,
or a possible need for replacement.

Only one recently published study documents the complete picture of clinical severe
pneumonia among admitted children in a university setting, by collecting individual
data, related to the demographic status, clinical and biochemical characteristics,
microbiological patterns and prognosis factors and the evolution of the disease 15]. The limitation of this study is that its results cannot be extrapolated to the entire
country, because this area is better than average covered in terms of health infrastructures,
and is a predominantly urban.

Regarding the pneumococcal conjugate vaccine, the lack of consistent and geographically
representative data on the circulating serotypes known to cause paediatric invasive
bacterial disease in Morocco hampers our understanding of the potential effectiveness
of the introduction of such vaccine. It would appear essential to monitor prospectively
what pneumococcal serotypes are still causing disease among Moroccan children, not
only to understand whether this pathogen remains the major public health concern,
but also demonstrate whether serotype replacement is really occurring 34].

The risk factors identified among admitted children with ARI are common to those documented
in other settings 35]. Risk factors identified among children in a hospital setting were related to the
management of the disease, and also relate to the need of the improvement of the quality
of care of patients admitted for ARI. Among these risk factors an easily preventable
one relates to control tobacco exposure. One study reported that passive smoking is
a real concern as a risk factor for ARIs in Morocco, having this country still not
ratified the convention on the tobacco control 36]. Therefore, such data can be used as a strong argument to help the scientific community
lobby in the activation of this process for the protection of the health of children
and citizens in general.

The few available data sources that we have been able to identify, may adequately
describe local patterns of ARI, but clearly have important limitations as to their
use to describe the burden of ARI at national level or their extrapolation to the
general population to estimate the burden of disease or epidemiological trends. Moreover,
knowledge gaps found of ARI in Morocco identified by this review seem to be common
to other developing countries 4],7], where the real burden and impact of childhood ARI and/or pneumonia is difficult
to assess and to document.

To our knowledge, this is the first review that has collected evidence from a variety
of sources of data for ARI in Morocco. This is a critical step which has allowed identifying
and summarizing already existing relevant data, (Figure 5), but also shows the weaknesses of the current national surveys designed to estimate
the real burden of ARI caused by Streptococcus pneumoniae and to monitor the antibiotic resistance among the prescribed antibiotics for ARI.

Figure 5. Main information identified in the documents included in the systematic review analysis
of ARI among children in Morocco. From 1997 to 2014.

Morocco should promptly develop a set of common indicators for tracking progress in
its fight against pneumonia and use these data to identify groups at greater risk
or missed by Health care Services and develop integrative approaches to reach them.