Absence of association between cord specific antibody levels and severe respiratory syncytial virus (RSV) disease in early infants: A case control study from coastal Kenya

Background: The target group for severe respiratory syncytial virus (RSV) disease prevention is infants under 6 months of age. Vaccine boosting of antibody titres in pregnant mothers could protect these young infants from severe respiratory syncytial virus (RSV) associated disease. Quantifying protective levels of RSV-specific maternal antibody at birth would inform vaccine development. Methods: A case control study nested in a birth cohort (2002-07) was conducted in Kilifi, Kenya; where 30 hospitalised cases of RSV-associated severe disease were matched to 60 controls. Participants had a cord blood and 2 subsequent 3-monthly blood samples assayed for RSV-specific neutralising antibody by the plaque reduction neutralisation test (PRNT). Two sample paired t test and conditional logistic regression were used in analyses of log2PRNT titres. Results: The mean RSV log2PRNT titre at birth for cases and controls were not significantly different (P = 0.4) and remained so on age-stratification. Cord blood PRNT titres showed considerable overlap between cases and controls. The odds of RSV disease decreased with increase in log2PRNT cord blood titre. There was a 30% reduction in RSV disease per unit increase in log2PRNT titre (<3months age group) but not significant (P = 0.3). Conclusions: From this study, there is no strong evidence of protection by maternal RSV specific antibodies from severe RSV disease. Cord antibody levels show wide variation with considerable overlap between cases and controls. It is likely that, there are additional factors to specific PRNT antibody levels which determine susceptibility to severe RSV disease. In addition, higher levels of neutralizing antibody beyond the normal range may be required for protection; which it is hoped can be achieved by a maternal RSV vaccine. This study was conducted in Kilifi, the coastal part of Kenya [20]. The study was nested within a previous birth cohort study. Between 1999 and 2007, Kenya Medical Research Institute-Wellcome Trust Research Programme (KEMRI-WTRP), conducted a Kilifi Birth Cohort (KBC) study within the Kilifi Health and Demographic Surveillance System [20–22]. The KBC study was an observational study where participants were followed for a two-year period with a cord blood sample collected at birth and subsequent 3 monthly blood samples during follow ups. Details of the birth cohort study are described elsewhere [16, 21, 22]. The KBC study participants had access to Kilifi County Hospital (KCH) previously referred to as Kilifi District Hospital (KDH). We conducted a case-control study using archived serum or plasma samples from participants of the KBC study. We defined cases as infants who were admitted with severe pneumonia or lower respiratory tract infection (LRTI) within the first 6 months of life and had a positive RSV diagnosis by an Immunofluorescent Antibody Test (IFAT; Millipore, USA). Controls were infants who were not admitted to KCH with RSV associated severe pneumonia during the follow up period. Exposure to RSV infection among the controls was measured by screening cord blood and the 3 monthly subsequent samples for RSV IgA ELISA antibodies. An increase in RSV IgA antibodies detected either in the 3 or 6 month serum sample was used as a marker of exposure. Cases were matched to controls in a ratio of 1:2 by date of birth (within 30 days of date of birth) and geographical location. Every participant had a cord blood sample and two subsequent 3 months follow up blood samples. Continuous surveillance for RSV in paediatric admissions to KCH with syndromic severe or very severe pneumonia (using WHO criteria)[23] was in place throughout the KBC study with screening for RSV antigen by Immunofluorescent Antibody Test (IFAT)[24]. Enhanced detection of RSV cases missed by using the WHO pneumonia criteria was made by inclusion of children with a clinical admission diagnosis of LRTI. During the period for RSV surveillance, all paediatric admissions had blood samples collected for culture to diagnose invasive bacterial pathogens. A computerised system linking residents of the KHDSS and KCH admissions enabled identification of RSV positive hospital admission from the KBC. Furthermore, we investigated whether the level of neutralizing antibodies measured from RSV A2 infection differs according to the circulating strain within the population. To do this, an additional subset of 100 cord blood samples was randomly selected from the Kilifi birth Cohort study regardless of infection status and tested for neutralizing activity to a range of RSV virus strains. All parents and guardians gave written consent to have their children participate in the KBC study or paediatric RSV study at KCH and for storage of blood samples for use in future research. The use of the archived sample set was approved by the KEMRI-Ethical Review Committee. Cord blood and admission samples collected during the KBC study were immediately taken to the microbiology laboratory for processing and storage at -80°C. For this case control study, archived blood samples were retrieved and assayed for RSV neutralising antibody by a plaque reduction neutralisation assay. Serum samples were incubated at 56°C in a water bath for thirty minutes to inactivate complement cascade proteins; thereafter, plaque reduction neutralisation procedures were conducted as described elsewhere[25]. The dilution (and titre reciprocal) of a test serum sample required to induce 50% neutralization of a known titration of RSV A2 virus was determined using the Spearman Karber method. To account for any significant effect of freeze thaw on sample neutralization titres, a validation assay was carried out where a set of KBC samples previously screened (5 years interval); were retrieved, screened and the two PRNT titres results compared. A subset of 100 cord bloods collected over the same time period as for the case-control study were selected at random, stratified by year, from the KBC archive, as previously described [16]. These were screened for RSV specific neutralising antibodies by the method described above using 4 different strains i.e., RSV A2 (Australia, 1961), RSV B860 (Sweden, 1960), RSV A Kilifi (Kenya, 2005) and RSV B Kilifi (Kenya, 2005). Residues of samples from controls were tested for RSV-specific IgA antibodies by ELISA using crude virus extract from laboratory adapted RSV A2 culture[26] Specific antibody concentrations were recorded as log arbitrary units (AU) as determined by a local standards procedure[26]. The crude virus RSV lysate preparation was as previously described by Ochola et al[27]. All data analysis was conducted using STATA version 13.1 (College Station, Texas). Laboratory data for sample PRNT titres were logarithmically transformed (base 2) and merged with the KHDSS and clinical data for analyses. To quantify the level of RSV-specific maternal antibodies at birth that provide infant protection, mean PRNT titres were computed for cord blood samples. The difference in cord blood levels between the cases and controls were analysed using a two sample paired t test. For comparison, a two sample Wilcoxon rank-sum (Mann-Whitney) test was applied to the log2PRNT titres for cases and controls. Further comparison of the distribution of cord blood log-transformed PRNT titres between cases and controls was done using reverse cumulative distribution plots. The absolute reduction in risk of RSV disease per unit increase in cord blood antibody titres was calculated using modified conditional logistic regression methods[28]. The estimated rate of decay of RSV specific log2PRNT titres from birth to 6 months of life was determined by simple linear regression, accounting for clustering of titres for samples from the same individual using the procedure for Huber-White sandwich estimator. Elimination of bias on the rate of decay arising from RSV infection was done as previously described [16]i.e.: the titre of cord, first and second samples for an individual were defined as TC, T1 and T2, respectively. For individuals with T1≥TC, all results for that individual were excluded, and for individuals with T2≥T1 the result for sample T2 was excluded. In addition, results from samples collected from cases after an infection was identified were excluded. Comparison of the estimated rate of decay was conducted using a two sample paired t-test. To further control for the difference in the estimated rate of decay between cases and controls within a match set, a linear regression model with an interaction effect between age and case/control was used. The association between the concentration of maternal antibody titres and gestational age (measured by clinical evaluation or based on date of last menstruation) and birth weight (measured using a weighing scale at birth) was assessed. The odds of severe RSV disease in the first 6 months of life was determined, adjusting for the potential confounders of prematurity (i.e. gestational age < 37 weeks) and low birth weight (i.e. <2.5Kg) as categorical variables, using conditional logistic regression methods[28]. To ascertain exposure among controls, a defined cut off value was computed from the mean cord blood IgA log AU plus 3 standard deviations. Any of the 3 or 6 months serum samples with an IgA level above this cut off value were defined as exposed and those below this cut off as unexposed.