Nasopharyngeal carriage of Streptococcus pneumoniae in children under 5 years of age before introduction of pneumococcal vaccine (PCV10) in urban and rural districts in Pakistan

This was a cross-sectional survey done in two union councils of rural Matiari and one union council of urban Karachi, both in the southern province of Sindh, Pakistan. These sites were selected because of pre-existing infrastructure and ongoing research in the community. A union council is the smallest administrative unit in the government structure. Matiari is around 180 km away from Aga Khan University (AKU) in Karachi, which houses the research unit and the Infectious Disease Research Laboratory (IDRL). The field site in urban Karachi was located around 20 km from AKU. This survey was done during the months of January and February 2013, immediately before the introduction of PCV10 vaccine in Sindh.

All children aged 3 to 11 months resident in the study area in the Matiari site and all children under the age of 5 years in the Karachi site were eligible to be enrolled in the study. A random sample of children 3 to 11 months of age was selected from a household listing of two union councils, Khyber and Seekhat, in Matiari. In Karachi an additional random sample of children 12 to 59 months was selected from a line listing of all children living in the area. Children with a history of severe acute respiratory illness in the last 2 weeks, presence of chest wall indrawing, blue skin discoloration (cyanosis) and fast breathing were excluded from the study. Children with moderate to severe cerebral palsy, neurological disorders affecting swallowing and nose and throat disorders were also excluded. If a child was not found, or found to be ineligible in the random list, the next eligible child in the list was approached.

Baseline demographic data and clinical characteristics of children were collected through a questionnaire (available as a supplementary file). The questionnaire was developed, translated and back-translated between English, Urdu and Sindhi languages and pretested in the field before implementation. A two-day training was held for staff involved in the study to standardize data and specimen collection methods. Crowding index was developed by dividing number of people by number of rooms in the household, axillary temperature was measured using a digital thermometer and vaccination history was collected either verbally or verified through card.

Nasopharyngeal swabs were collected by trained staff as per World Health Organization methods [18]. Swabs were immediately placed into skim milk tryptone glucose glycerol (STGG) media in cryovials and transported to the Infectious Disease Research Laboratory (IDRL) in Karachi at 2-8°C within 6-8 hours of collection. Upon reaching the central laboratory, inoculated STGG was vortexed for 10-20 seconds to disperse organisms from the swab and frozen in an upright position at -70°C until further processing. Samples were then cultured in batches of 20-40 samples and sub-cultured as described in the CDC protocol (https://​www.​cdc.​gov/​streplab/​downloads/​pcr-pneumo-carriage-march2010.​pdf). Briefly, an aliquot of the thawed STGG mixture was enriched in Todd Hewitt broth supplemented with rabbit serum (20%) and yeast extract (0.5%) and subcultured onto TSA medium with 5% sheep blood. After 18-24 hours of incubation, plates were examined for the presence of alpha-haemolytic colonies with morphological characteristic features of pneumococci. There was only one case where two colonies were identified and the dominant one was elected. Identification was confirmed by susceptibility to Optochin and bile solubility testing.

Serotypes for the pneumococcal isolates were deduced by sequential conventional multiplex PCR method as described in CDC Streptococcus laboratory protocols [19]. Briefly, DNA was extracted by boiling a loop full of bacterial culture in 1 ml of TE Buffer at 100 °C for 10 minutes followed by centrifugation at 10,000 rpm for 10 minutes; the resulting supernatant was collected in a new sterile 1.5 ml microfuge tube and 2 μl of the extracted DNA was directly used in the PCR reaction. The cpsA gene was included in each multiplex reaction to serve as an internal positive control. PCR master mix (of 23 μl) was prepared using 25 μM working stock of primers, 2X Qiagen multiplex PCR buffer, Qiagen Q solution and nuclease free water, with 2 μl of DNA template added to the reaction mixture. Amplification was carried out in an Eppendorf Master Cycler Gradient with the following temperature profile: initial denaturation at 95 °C for 15 minutes, then 35 cycles consisting of 94 °C for 30 seconds, 54 °C for 90 seconds, and 72 °C for 60 seconds, followed by a final extension step at 72 °C for 10 minutes. The amplified PCR products were separated on 2% (w/v) agarose gel electrophoresis, stained with 1X SYBR-green staining solution, and documented under BioRad Gel Doc imager. Serotypes detected through sequential multiplex conventional PCR were further confirmed by monoplex PCR, using the same PCR conditions, to avoid misidentification due to non-specific bands [20].

Serogroup 6 was further discriminated into individual serotypes, 6A, 6B, 6C, and 6D by conventional PCR as described by Jin et al. [19, 20] with slight modifications, using 25 μM working stocks of primers, 2X Qiagen multiplex PCR buffer, Qiagen Q solution and nuclease free water, with 2 μl of DNA template added to the reaction mixture. Amplification reactions were conducted using a Eppendorf Master Cycler Gradient with the following temperature profile: initial denaturation at 95°C for 15 minutes, then 35 cycles consisting of 94°C for 30 seconds, 66.2°C for 60 seconds, and 72°C for 60 seconds, followed by a final extension step at 72°C for 10 minutes. Appropriate pneumococcal serotype controls obtained from the CDC streptococcal lab were included in every reaction. The amplified PCR products were separated on 2.5% (w/v) agarose gel electrophoresis, stained with 1X SYBR -green staining solution, and documented under a Biorad Gel Doc imager.

Identification of non-typeable pneumococcal isolates (those with no positive results for any tested serogroups/types) was confirmed with a lytA real time PCR as described by Carvalho et al. [21].

For Optochin susceptibility and bile solubility reactions, the following reference American Type Culture Collection (ATCC) strains were used for quality control: Streptococcus pneumoniae ATCC 49619 and Enterococcus faecalis ATCC 29212. In addition, a total of 176 randomly selected pneumococcal isolates were sent to CDC to cross check PCR serotying results with the Quellung reaction; 95% concordance was found between PCR and CDC Quellung reaction results.

Sample size was calculated based on a predictive decline in NP carriage in subsequent years after the introduction of vaccine. There are no recent data regarding prevalence of pneumococcal carriage in Pakistan. One study from 1993 reported prevalence of PCV7 serotypes in children 2-24 months as 50% and 65 % [22]. To obtain maximum sample size, a baseline carriage rate of 50% was assumed. This gave a sample size of 220 in order to show a decrease in prevalence from 0.5 to 0.3 after introduction of PCV10 with 80% power and 5 % level of significance.

Data consisting of demographic and socio-economic indicators as well as general health status and common symptoms were collected on paper forms by trained personnel at the designated fieldwork sites. All participants were assigned a unique identifier and a linking identification number from the demographic surveillance system was also noted. All completed forms were reviewed by the study supervisor at the end of the day for completeness and consistency. Data were dual entered in an MS Access database and validation and consistency checks were run. Analysis was done using STATA version 12 and MS EXCEL. Confidence bound for predictors such as age, gender, educational status etc. were created using a logistic regression model. All variables that were found to be significant at a p-value of ≤ 0.10 at univariate level were added in a multivariate model. A stepwise backward elimination approach was used to derive a parsimonious model retaining only variables with a P-value of <0.05. Odds ratios with 95% confidence intervals were calculated. Analysis was done looking at overall nasopharyngeal carriage as well as carriage of PCV10 specific serotypes as the outcome and 95% confidence intervals were presented.