Bacteria and viruses in the upper respiratory tract of Congolese children with radiologically confirmed pneumonia

The study was conducted at Panzi Hospital in Bukavu town in the South Kivu region of DR Congo, which served a population of 470,000 inhabitants including 89,000 children below 5 years in 2017. The hospital is the referral university hospital for three general hospitals, four health centres and eight dispensaries located in the catchment population area, but the hospital also receives patients from outside this area. Panzi Hospital has 69 paediatric beds with an emergency ward which has 12 beds for acute and severe cases requiring nasal oxygen therapy. The department has two trained paediatricians and four resident doctors. The hospital has a radiology department run by a trained senior specialist in radiology.

Out of a total of 2322 children between 2 months and 5 years of age treated for any disease at the Emergency and admissions Unit, Paediatric Department, between June 2015 and June 2017, 184 were diagnosed with acute lower respiratory infections. Of these, 116 cases had pneumonia that was radiologically confirmed and these children were thus included in the study. Among the 116 included children 80 came from the hospital catchment population area whereas 36 children came from other locations, in most cases nearby rural areas. Children admitted for severe malaria with pneumonia as complication were not included due to difficulties differentiating infection from pulmonary oedema. The parents or guardians of five HIV-positive children declined participation in the study and were thus not included while another three HIV-positive children were included. Four patients referred from health centres to Panzi Hospital had X-ray confirmed pneumonia but were not treated at the hospital because of the cost of care that was regarded as too high and were therefore not included. The direct cost of 5 days’ treatment for pneumonia at Panzi Hospital could amount to 150 US dollars.

Data were collected on age, sex, dates of admission, duration of hospital care, as well as underlying conditions, including malnutrition, HIV, sickle cell disease, cerebral palsy, post-neonatal anoxia and congenital diseases. The WHO child growth standards were used to evaluate the nutritional status of children by computing the Z-scores of weight-for-age and height-for-age [18].

Patients were classified according to the revised WHO classification for pneumonia. This encompasses non-severe pneumonia with fast breathing and/or chest indrawings. Fast breathing was present when the respiratory rate was ≥ 50 breaths per minute for infants between 2 and 12 months of age and ≥ 40 breaths per minute for children between 12 months and 5 years of age [18]. Children were considered to have severe pneumonia when there were any additional danger signs (e.g. not able to drink or breastfeed, persistent vomiting, convulsions, lethargy or loss of consciousness, stridor in a calm child or severe malnutrition) [19].

Saturation of peripheral oxygen (SpO₂) was measured on admission by using a pulse oximeter (Handheld Pulse Oximeter 3.7vx1, Li-ion, Model No.AH-M1, Acare Technology Co., Ltd, London, UK). Venous blood samples were collected for measurement of white cell count and C-reactive protein. A nasopharyngeal swab was taken from all children in the study (see below). After a physical examination, radiologist reports and blood test analyses, all admissions were validated by the paediatrician. Lung auscultations were considered as abnormal if crackles and/or rhonchi were present.

Data on pre-hospitalisation use of antibiotics were obtained from the parent or guardian asked if any antibiotics were given to the child during the week in which the symptoms of pneumonia began, or from the referral note from any first-level health care facility. The use of antibiotics and nasal oxygen therapy during hospitalisation was also recorded.

Patient outcomes were grouped into the following three categories: survived without complications, survived with complications (e.g. pneumothorax, empyema, or pleural effusion) and death. The group surviving without complications included patients discharged with improvements and who had switched to oral antibiotic treatment. This group included those discharged against medical advice but who were receiving oral antibiotic treatment.

The study was approved by the Ethics Committee at the Université Catholique de Bukavu, DR Congo (UCB/CIE/NC/22/2014), and the Regional Ethics Committee in Gothenburg, Sweden (Dnr: 504-16). The study was further approved by the South-Kivu province Health Organisation (065/CD/DPS-SK/2015) and by the Director of Panzi Hospital.

The diagnosis of pneumonia was confirmed by X-ray (Siemens-Elema AB, Solna Sweden) with anteroposterior and lateral chest radiographs obtained from all included children. The images were first analysed by the radiologist at Panzi Hospital who was trained in the standard interpretation of chest radiographs for the diagnosis of childhood pneumonia. The second and final validation was performed by a senior specialist radiologist according to the WHO standard interpretation of chest radiographs for the diagnosis of pneumonia in children. Pneumonia was confirmed by the presence of consolidation, infiltrates or effusion and these results were considered as chest X-ray confirmed pneumonia [20]. Pleural effusion was present when there were more than 1 cm fluid in the pleural space between the lung and chest wall [20].

A nasopharyngeal sample was collected from all included children and transported to the Clinical Laboratory at Panzi Hospital for pneumococcal culture as previously described [17]. The nasopharyngeal samples were thereafter stored frozen at − 20 °C before shipment to the Department of Infectious Diseases, University of Gothenburg, Gothenburg, Sweden, where they were stored at − 80 °C prior to the molecular analyses (see below).

Nucleic acids were extracted from 200 µL of the nasopharyngeal sample as described earlier [21]. Eluted nucleic acids were stored at − 20 °C until further analysis.

A multiplex real-time PCR assay capable of detecting 20 different viruses and bacterial species (adenovirus, bocavirus, coronavirus 229E, HKU1, NL63 and OC43, enterovirus, influenza virus A and B, human metapneumovirus, parainfluenza 1–3, rhinovirus, RSV, Bordetella pertussis, Chlamydia pneumoniae, H. influenzae, Mycoplasma pneumoniae and S. pneumoniae) was performed as previously described [22]. A Cycle Threshold (Ct)-value < 40 was classified a positive result with values of Ct < 30 indicating high levels of nucleic acids. Pneumococci were identified by lytA gene detection in nasopharyngeal samples. However, an additional three samples were considered positive for pneumococci consequent to cpsA gene detection (Ct < 40) in the serotype assay described below.

A multiplex real-time PCR assay capable of detecting 40 different pneumococcal serotypes/serogroups in addition to the cpsA gene was used according to a previously published protocol [23].

Descriptive analysis was performed using the SPSS package (version 24.0) for logistic regression analysis of the relationship between nucleic acid identification and symptoms, medical conditions or outcome. Differences in baseline data and clinical factors were analysed with Pearson’s chi-square test, and medians were compared with Mann–Whitney U test. The proportions of detected bacteria and viruses in severe and non-severe pneumonia cases were compared by Pearson’s chi-square or Fisher’s exact test (if n < 5). Potential variables associated with differences between fatal and improving cases were assessed by odds ratios (OR) with 95% CI and were tested by univariable analysis with the Pearson’s chi-square or Fisher’s exact test (if n < 5). Associations to fatal outcome with p < 0.2 were re-analysed by multivariable analysis. A p-value of < 0.05 was considered statistically significant.

Two years after the introduction of PCV13 in the infant immunization program in the Eastern DR Congo, 125/2322 (5.4%) of all children admitted to Panzi Hospital, DR Congo were diagnosed with radiologically confirmed pneumonia. Of these children, 116 were included in the present study (median age 12 months). Ninety-five (82%) of the included children came from Bukavu town while 21 came from rural areas (Table 1).

Almost all the children 101 (87%) were treated with antibiotics before hospital admission, according to the parent or guardian (n = 39) or the referral notes (n = 62). The most common pre-hospitalisation antibiotics were either amoxicillin/ampicillin/phenoxymethylpenicillin or trimethoprim-sulfamethoxazole (Table 1). Two thirds (67%) of the children had received two or three doses of PCV13.

The main signs and symptoms noted during the physical examination at admission were fever or recent history of fever (96%), abnormal lung auscultation (99%), cough or recent history of cough (100%), and rapid or difficult breathing (94%). Malnutrition was present in 11 (9.5%) of the patients. During hospital stay 53% were treated with ceftriaxone and gentamicin whilst 40% received ampicillin and gentamicin (Table 1).

According to the WHO pneumonia classification, 85/116 (73%) of the children met the criteria for having severe pneumonia at admission whereas 31 (27%) had non-severe pneumonia. Between the two groups, there was no statistically significant difference in sex, age, living area or immunisation status, fever, cough or abnormal auscultation (data not shown). However, a white cell count of over 20,000/μL was significantly associated with severe pneumonia (OR 4.75; 95%CI 1.0–21.6, p = 0.043) (Table 1). CRP levels above 75 mg/dL were also significantly associated with severe pneumonia at admission (OR 38.9; 95% CI 5.1–299, p = 0.0004) while CRP levels below 25 mg/dL were negatively associated with severe disease (OR 0.19; 95% CI 0.08–0.48, p = 0.0003) (Table 1). The median CRP level was much higher in children with severe pneumonia at admission as compared with the non-severe pneumonia cases while the difference in median concentration of the white cells were small between the two groups (Table 1). As expected, nasal oxygen therapy was more often used in children with severe pneumonia as compared to children with less severe disease (Table 1). Overall, pre-hospitalisation antibiotic treatment was more common among children with severe pneumonia than in children with non-severe disease. While penicillin and the penicillin derivatives amoxicillin or ampicillin were equally common in the two groups, trimethoprim-sulfamethoxazole was more commonly used in children with severe pneumonia (OR 4.75; 95%CI 1.04–21.6 as above) (Table 1). During hospital stay, ceftriaxone combined with gentamicin was more frequently used to treat the severe pneumonia cases than the children with non-severe disease (Table 1).

From only one nasopharyngeal sample pneumococci could be isolated by culture. However, by real-time PCR, performed directly on the nasopharyngeal sample, pneumococci could be detected in almost all children (96%), whereas H. influenzae was detected in 54% and B. pertussis in 10% (Table 2). When employing a more stringent cut-off level (Ct < 30), bacteria were found in 62% of the samples; S. pneumoniae in 53% of the cases, and H. influenzae in 20% (Table 2).

The most frequently detected virus was rhinovirus (73%), followed by enterovirus (17%), while RSV was found in only 7% of the cases and influenza virus was rare (Table 2). When only high levels of viral nucleic acids were considered (Ct-value < 30), rhinovirus was found in 29% of the cases, and only a few samples were positive for RSV and influenza virus, respectively (Table 2). No statistically significant differences were found in the frequencies nor in the Ct median levels for any of the detected bacterial species or viruses between severe and non-severe pneumonia cases. However, high levels (i.e. Ct-value < 30) of any virus (OR 2.63; p-value = 0.032) or any bacteria (OR 3.14; p-value = 0.008) present in the sample, regardless of type or species was significantly associated with severe pneumonia (Table 2).

Out of the 111 nasopharyngeal samples that were positive for pneumococci (Ct < 40), 83 serotypes/serogroups could be identified in samples from 61 children while no serotype could be detected in secretions from the remaining 50 children. Two different serotypes/groups could be identified in 11 patients, while three children had three, one child had four, and another child had five different serotypes/groups. Among the 83 identified serotypes, 52 (63%) belong to the PCV13 serotypes/serogroups, while the remaining 31 (37%) are non-PCV13 serotypes/serogroups. The serotypes/groups most frequently detected were 19F, 6ABCD and 5 while the predominant non-PCV13 serotypes were 15BC, 10A and 11A (Fig. 1). Two samples were found to be negative for the cpsA pneumococcal capsule gene but were strongly positive for the lytA gene (Ct < 30), indicating non-encapsulated non-typeable pneumococci.

The overall case fatality rate among the hospitalized children with radiological confirmed pneumonia was 9.5% (11 children). Five of these children died within 48 h after hospital admission and another four died after two to five days. Fatal outcome was significantly associated with having a congenital disease as an underling condition (Table 3). Univariable analysis also showed an association between fatal outcome and high nucleic acid levels (Ct < 30) of pneumococci or RSV in the upper respiratory tract of the children, but this could not be confirmed in the multivariable analysis (Table 3). The two RSV-positive cases with fatal outcome were also positive for pneumococci. Two different pneumococcal serotypes could be determined in the nasopharyngeal secretions from the children with fatal outcome; serotype 19F in four cases and 15BC in one case. In the remaining five children with pneumococci detected, no serotype could be identified.