key: cord-0904261-zob6dvbh authors: Muloiwa, Rudzani; Dube, Felix S.; Nicol, Mark P.; Hussey, Gregory D.; Zar, Heather J. title: Co-detection of Bordetella pertussis and other respiratory organisms in children hospitalised with lower respiratory tract infection date: 2020-10-02 journal: Sci Rep DOI: 10.1038/s41598-020-73462-w sha: f30ec652614b1bee15e6e8c4e519454fb68f9e16 doc_id: 904261 cord_uid: zob6dvbh Multiple potential pathogens are frequently co-detected among children with lower respiratory tract infection (LRTI). Evidence indicates that Bordetella pertussis has an important role in the aetiology of LRTI. We aimed to study the association between B. pertussis and other respiratory pathogens in children hospitalised with severe LRTI, and to assess clinical relevance of co-detection. Nasopharyngeal (NP) swabs and induced sputa (IS) were tested with a B. pertussis specific PCR; additionally, IS was tested for other pathogens using a multiplex PCR. We included 454 children, median age 8 months (IQR 4–18), 31 (7%) of whom tested positive for B. pertussis. Children with B. pertussis had more bacterial pathogens detected (3 versus 2; P < 0.001). While B. pertussis showed no association with most pathogens, it was independently associated with Chlamydia pneumoniae, Mycoplasma pneumoniae and parainfluenza viruses with adjusted risk ratios of 4.01 (1.03–15.64), 4.17 (1.42–12.27) and 2.13 (1.03–4.55), respectively. There was a consistent increased risk of severe disease with B. pertussis. Patterns indicated even higher risks when B. pertussis was co-detected with any of the three organisms although not statistically significant. Improving vaccine coverage against B. pertussis would impact not only the incidence of pertussis but also that of severe LRTI generally. We hypothesise that the conditions produced by B. pertussis toxins not only create a conducive environment for B. pertussis itself but may facilitate the colonisation or infection of the respiratory tract epithelium by other bacteria or viruses. In this study, we aimed to investigate whether the detection of B. pertussis in children hospitalised for LRTI was associated with co-detection of other potential respiratory pathogens. We also explore whether there was an association between co-detection and clinical severity and outcome. Recruitment and specimen collection. Methods for sampling as well as inclusion criteria have been described elsewhere 15 . Briefly, the study recruited inpatient children seen at a referral hospital, Red Cross War Memorial Children's Hospital (RCH) in Cape Town, South Africa over a one-year period (September 2012 to September 2013). Children were recruited if they presented with cough and WHO defined age specific tachypnoea, or apnoea, and were ill enough to warrant admission. Only children whose legal guardians were present to give written consent were enrolled. As the study intended to describe co-infection involving community acquired organisms, we excluded children who had been in touch with the health care services in the preceding two weeks in order to minimise health care-associated infection. To assess the severity of respiratory symptoms, the presence of chest indrawing was noted. In addition, all children had pulse oximetry to assess for oxygen saturation. A cut-off of < 94% was used to define hypoxaemia in children at sea level [16] [17] [18] . A detailed history of the current illness was collected, and participants underwent testing for HIV infection. The diagnosis of HIV infection was made for children less than 18 months of age if they tested positive for two HIV polymerase chain reaction (PCR) tests (COBAS AmpliPrep/COBAS Taqman HIV-1, Roche Molecular Diagnostics, Pleasanton, CA, USA). For children above 18 months of age, HIV infection was diagnosed on the basis of two positive ELISA tests using two different assays (Architect HIV Ag/Ab Combo, Abbott Diagnostics, Wiesbaden; and Enzygnost Anti-HIV 1/2 Plus, Siemens/Dade Behring, Erlange, sequentially). All children had anthropometry (weight and height), performed at enrolment by trained study staff. Nutritional status was classified using WHO weight for age z-scores (WAZ). Children were classified as moderate to severely malnourished if their weight for age fell below − 2 z-scores. The children's vaccination status was sourced from their handheld clinic booklets. The primary schedule according to the South African Expanded Program on Immunisation, in addition to other vaccines, contains an acellular (aP) vaccine against B. pertussis combined with that against Haemophilus influenzae type b at 6, 10 and 14 weeks (with a booster at 18 months), and vaccination with 13-valent pneumococcal conjugate (PCV13) vaccine at 6 and 14 weeks of age (with a booster at 9 months) 19 . A nasopharyngeal (NP) swab was collected first after which an induced sputum (IS) specimen was collected on enrolment as previously described 20 . Molecular diagnostic testing was carried out on batched specimens as described below. No blood culture was done as part of the study. Laboratory methods. Diagnosis of Bordetella pertussis. To diagnose Bordetella pertussis infection, PCR specific for IS481 for Bordetella species was conducted on both NP and IS specimens with a validated commercial kit (Roche LightMix, Basel) using previously published primers 21 . Bordetella holmesii (defined as IS481+ and hIS1001+) which shares the same IS481 target as B pertussis was excluded by further testing all IS481 positive specimens for the presence of insertion site hIS1001 22 . Diagnosis of co-infections. The FTDResp 33 multiplex real-time PCR assay (Fast-Track Diagnostics, Esch-sur-Alzet, Luxembourg) was used to identify presence of a range of viruses and bacteria as well as Pneumocystis jirovecii on IS. As the study was designed specifically to study the epidemiology of pertussis in this population, for the analysis, B. pertussis results from LightMix for B. pertussis (rather than those from FTDResp 33) were used as our assessment indicated that the assay had better sensitivity for B. pertussis than Fast-Track, although both employ the same targets. We used percentages to depict proportions of study participants with organisms detected from respiratory specimens. Continuous data were tested for normality and summarized as medians with interquartile ranges (IQR) or means and standard deviations (SD) as appropriate. The difference in total numbers of organisms detected in participants with and without confirmed B. pertussis was compared using Student's t-test. χ 2 or Fisher's exact tests were used to assess the strength of association between infection with B. pertussis and each co-pathogen. All associations at a two-tailed P < 0.1 were further analysed adjusted for sex, age and HIV status as potential confounders. Generalised linear modelling using Poisson regression with robust error variance was used to estimate adjusted relative risks (aRR) and their 95% confidence intervals in a multivariable analysis. Severity of clinical disease and outcomes were further analysed stratified by a combination of pertussis status and organisms showing strong association with pertussis. Continuous data were tested for normality and comparisons between groups were made with the appropriate test for parametric or non-parametric data as indicated. Statistical significance was set at a two-sided P < 0.05. All analyses were carried out using Stata Statistical Software Release 16 (StataCorp LP, College Station, TX). Ethics Committee of the Faculty of Health Sciences of the University of Cape Town; reference: 371/2011. Written informed consent was sought and received from the parent or legal guardian of each child in order for the child to participate in the study. All methods were carried out in accordance with the relevant guidelines and regulations. Baseline data. Four hundred and sixty children were enrolled. Six children, including four whose IS could not be collected and two whose IS were lost prior to processing during transportation or storage, were excluded, providing 454 participants with sufficient data for analysis, of which 253 (55.7%) were male. The median age of the children was 8 (IQR 4 -18) months. HIV infection was confirmed in 19 (4.2%) of the children. History of asthma or eczema was present in two children, while 26 (5.9%) of the caregivers (all mothers to the children) reported a history of asthma. Nine children (2.0%) did not have their immunization records with them. Of the 445 (98.0%) with known vaccination status, 321(72.1%) were up to date with pertussis and Haemophilus influenzae type b vaccine doses for age, while 418 (93.9%) had received at least one dose of the combination. Similarly, 312 (70.1%) were up to date with PCV13 doses for age with 385 (86.5%) having received at least one dose of the vaccine. Baseline characteristics of the study group are summarized in Table 1 . Confirmed Bordetella pertussis infection. PCR for insertion site IS481 was positive in 16 NP specimens and 25 IS specimens on LightMix. Ten participants had a positive PCR on both NP and IS specimens, therefore 31 (6.8%; 95% CI 4.7-9.6%) participants were confirmed as having B. pertussis infection. The B. holmesii insertion site hIS1001, was not identified in any of the IS481 positive specimens. Only 10 (2.2%) samples, all also found to be positive on LightMix, were positive for B. pertussis on Fast-Track testing of IS samples. The median age of children with confirmed B. pertussis was 8 (IQR 2-22) months while those testing negative had a median age of 8 (IQR 4-18) months; P = 0.559. B. pertussis was isolated in 5 (18.5%) of the 27 who did not get any pertussis vaccine dose compared to 26 (6.2%) of the 418 who received at least one vaccine dose; P = 0.032. For all participants with confirmed B. pertussis infection, a minimum of two organisms were identified. The average number of viruses detected in children with and without confirmed pertussis were 2.5 (SD 1.4) and 2.3 (SD 1.3) respectively; P = 0.665. Children with confirmed pertussis had on average 3.0 (SD 1.6) different bacterial species detected while those without had 2.0 (SD 1.1); P < 0.001. When both bacterial and viral organisms were considered, the average number identified in pertussis positive participants was 5.5 (SD 2.0) and 4.4 (SD 1.9) in the pertussis negative group; P = 0.009 (Fig. 1) . The prevalence of specific organisms identified on IS in participants with and without confirmed pertussis is shown in descending order of frequency in Table 2 . www.nature.com/scientificreports/ The same pattern was noted with the two bacterial organisms whose detection was strongly associated with B. pertussis (Fig. 2) . A higher risk of severe disease was seen when B. pertussis was detected with each organism than when B. pertussis or each of the bacteria was detected on its own. No hypoxaemia was noted in any of the C. pneumoniae positive children who did not have B. pertussis. With parainfluenza viruses, this pattern was noted only with respect to hypoxaemia but not with chest indrawing. The small sample sizes of each stratum were not sufficient to allow for meaningful formal assessment of strength of association. The median WCC was 13.8 (IQR 9.8-16.3) × 10 3 cells/mL in children without confirmed B pertussis compared to 13.8 (10.4-17 .0) × 10 3 cells/mL in those with confirmed infection; P = 0.421, while children without confirmed B. pertussis had median C-reactive protein levels of 11.3 (IQR 2.7-31.5) mg/L compared to 8.4 (IQR 2-37) mg/L in those with confirmed infection; P = 0.5704. The median length of hospital stay was similar in children with confirmed B. pertussis [2 (IQR 1-5) days] and children testing negative [2 (IQR 1-4) days]; P = 0.522. The co-detection of organisms independently associated with pertussis did not have an effect on the length of hospital stay (Fig. 3 ). This study demonstrates a high prevalence of bacteria and viruses involved in LRTI in children requiring hospitalization for severe disease in a low-and middle-income country (LMIC) setting. The study also shows that only a few of these organisms are specifically associated with B. pertussis co-detection. The detection of B. pertussis was however strongly associated with a higher number of co-detected potential respiratory pathogens, specifically bacterial organisms. Although most of the studied potential respiratory pathogens did not show association with the presence of B. pertussis, three, namely C. pneumoniae, M. pneumoniae and parainfluenza viruses, were independently associated with being co-detected with B. pertussis. In addition, the study shows some evidence, albeit weak, that the co-detection of B. pertussis, together with these three organisms may be a risk for severe illness. The finding of other pathogens in a respiratory tract specimen together with B. pertussis is not by itself remarkable. As noted earlier, several studies have shown that multiple potential pathogens are frequently identified from children with respiratory infections, including pertussis 4, 23 . What is of interest, however, is the finding of higher numbers of potential pathogens, in particular bacterial, significantly associated with the presence of confirmed B. pertussis; and the association of specific organisms with confirmed B. pertussis. Participants with B. pertussis had more organisms detected in their sputum samples, compared to those without. This association was significant for bacterial pathogens, a finding which had an impact also on association with overall total number of all detected organisms. In addition, patients with B. pertussis had a fourfold increase in the risk for detection of either C. pneumoniae or M. pneumoniae, as well as twice the risk to detect parainfluenza viruses. The finding of M. pneumoniae in the upper respiratory tract of children in the Netherlands was found not to be associated with clinical disease 24 . The study reported prevalence of 21% and 16% in asymptomatic and symptomatic children, respectively. This shows that carriage of M. pneumoniae may be quite common. Unlike in the current study, the mentioned study used an upper respiratory tract sample, and did not investigate for B. pertussis. Similar to our study, a Swiss study found no association between B. pertussis and detection of viruses via PCR 25 . Again, unlike in our study, a nasopharyngeal specimen was used and not an IS sample. In this study, although the number of viruses associated with B. pertussis was marginally higher, this finding was not of statistical significance. This is in keeping with findings of other studies that show little association between B. pertussis and viral infections 25 . Specifically, our study showed a negative correlation between B. pertussis and RSV although this finding was also not of statistical significance. This pattern was however in keeping with published literature where a negative association was noted between RSV and pertussis or between pertussis and bronchiolitis, a disease mainly attributable to RSV 26, 27 . We did however, find independent association between detection of parainfluenza viruses and the presence of B. pertussis. Children with pertussis generally showed more severe clinical illness (hypoxaemia, chest-indrawing and need for high dependency care). Although these findings were not all statistically significant due to the small numbers, the pattern was consistent. Children who had both strongly associated organisms and B. pertussis detected showed additional risk of severe disease, again without statistical significance. It is of note that the three organisms independently associated with B. pertussis are associated with 'atypical' or interstitial pneumonia which commonly presents with impaired oxygenation. There was no difference in mortality (study registered no deaths) and length of hospital stay. There is evidence associating B. pertussis with pneumonia in children 4 . Although the first vaccine specifically targeting pneumonia was only introduced with the registration of a vaccine against Haemophilus influenzae b in 1985, the decline in pneumonia-associated mortality in the United States was noted three decades earlier 28 . This decline can not be fully explained by the improvement in quality of health care alone, and may in part be explained by the rapid decline in reported pertussis cases following the introduction of DPT in the 1940s. The decline in pneumonia-associated mortality mirrors that of pertussis over the period. Due to the small sample size, our study was limited both in exploring the effect of other factors such as HIV infection and in its ability to establish strong evidence for some associations, even where patterns suggested a correlation. In some instances, in which strong association was demonstrated, the precision of the estimated risk was low due to the same limited sample size. In addition, the study is limited only to the organisms included in the multiplex PCR, noting also that distinguishing between benign colonisation and pathological infection from PCR detection of an organism in the respiratory tract remains a challenge 3 . As such these findings must be interpreted with caution. Further well-powered studies or creative metanalytical systematic reviews are required to study this phenomenon further. 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://creat iveco mmons .org/licen ses/by/4.0/. Disease burden and mortality estimates Viral bacterial co-infection of the respiratory tract during early childhood A preliminary study of pneumonia etiology among hospitalized children in Kenya Aetiology of childhood pneumonia in a well vaccinated South African birth cohort: a nested case-control study of the Drakenstein Child Health Study the PERCH multi-country case-control study. Pneumonia Etiology Research for Child Health Study, G. Causes of severe pneumonia requiring hospital admission in children without HIV infection from Africa and Asia 6World Health Organisation. Immunization, Vaccines and Biologicals Pertussis: review of epidemiology, diagnosis, management and prevention How best to estimate the global burden of pertussis? Compromised defenses: exploitation of epithelial responses during viral-bacterial co-infection of the respiratory tract Bordetella pertussis, Bordetella parapertussis, Mycoplasma pneumoniae, Chlamydia pneumoniae and persistent cough in children. Scand Frequency of serological evidence of Bordetella infections and mixed infections with other respiratory pathogens in university students with cough illnesses A place for Bordetella pertussis in PCR-based diagnosis of community-acquired pneumonia Virulence factors of Bordetella pertussis Bordetella pertussis pathogenesis: current and future challenges Incidence and diagnosis of Pertussis in South African children hospitalized with lower respiratory tract infection Reference values for oxygen saturation by pulse oximetry in healthy children at sea level in Chennai Normal oxygen saturation values in pediatric patients How should oxygen supplementation be guided by pulse oximetry in children: do we know the level The expanded programme on immunisation in South Africa: a story yet to be told Safety and efficacy of induced sputum in young children hospitalised with suspected pulmonary tuberculosis Nested duplex PCR to detect Bordetella pertussis and Bordetella parapertussis and its application in diagnosis of pertussis in nonmetropolitan Southeast Queensland Novel multitarget real-time PCR assay for rapid detection of Bordetella species in clinical specimens Bordetella pertussis and mixed infections Carriage of Mycoplasma pneumoniae in the upper respiratory tract of symptomatic and asymptomatic children: an observational study Bordetella pertussis and concomitant viral respiratory tract infections are rare in children with cough illness Bordetella pertussis is an uncommon pathogen in children hospitalized with bronchiolitis during the winter season Low prevalence of pertussis among children admitted with respiratory symptoms during respiratory syncytial virus season Trends in infectious disease mortality in the United States during the 20th century Mouse and pig models for studies of natural and vaccine-induced immunity to Bordetella pertussis Relative contribution of Th1 and Th17 cells in adaptive immunity to Bordetella pertussis: towards the rational design of an improved acellular pertussis vaccine Recent developments in pertussis Global, regional, and national causes of child mortality in 2000-13, with projections to inform post-2015 priorities: an updated systematic analysis We would like to acknowledge the participants, laboratory personnel and study staff for their immense contribution. The authors acknowledge financial support to their institution by Sanofi Pasteur for the submitted work. HZ is supported by the South African Medical Research Council. R.M. conceptualised and designed the study with the input from G.D.H. and H.J.Z. F.S.D. was responsible for the laboratory aspect of the study under the supervision of M.P.N. R.M. did the data analysis. All authors were involved in the overall interpretation of the data. R.M. drafted the initial manuscript and updated it following inputs from all the authors. All authors reviewed and approved the final manuscript. The authors declare no competing interests. Correspondence and requests for materials should be addressed to R.M.Reprints and permissions information is available at www.nature.com/reprints.Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.