The existing high child mortality rate (85·5/1000 in children aged <5 years) has become a great challenge bearing in mind Pakistan's high population growth rate (>1·8% per year) [1]. This problem becomes even more acute in the absence of appropriate healthcare considering the fact that children are highly vulnerable in parts of Khyber Pukhtoonkhawa (KPK) province of Pakistan to vaccine-preventable diseases, due particularly to the political unrest after the 9/11 sequence of events.

Although the incidence of whooping cough (pertussis) remains highest in neonates who have not received any of their scheduled pertussis-containing vaccinations, which are generally associated with Bordetella pertussis, it is known that another Bordetella species, B. parapertussis also contributes to the disease burden worldwide [Reference Edwards2]. Previous researches have already shown up to 35% of pertussis-like cases in highly vaccinated populations of some European countries as well as in Pakistan to be caused by B. parapertussis [Reference He3, Reference Bokhari4]. The exact prevalence of B. parapertussis infections due to the mild or sub-clinical course of the disease probably go unnoticed and may be underestimated [Reference He5]. For that reason community-based data on the occurrence of B. parapertussis infection is scarce [Reference He5, Reference He6].

Pakistan is a predominantly rural developing country with its ⩾190 million inhabitants facing political turmoil since the 9/11 incident. This has resulted in inaccessibility of vaccinators to susceptible populations in many parts of the country [7]. Therefore, it is a major concern for health sector workers to have some understanding of disease outbreak and its aetiological agent for better risk assessment. In this regard we report a case study from the geographically well isolated small town of Bisham in Shangla district near Swat in KPK province where a local epidemic due to B. parapertussis occurred during 2009–2010.

In order to determine the prevalence of pertussis in Bisham we collected a total of 50 samples from 12 houses in adjacent streets during the months of December 2009 and January 2010, from patients with a short history of cough, i.e. up to 4 or 5 days. Household contacts where two or more siblings were available were recruited for the study. Prior approval by the Ethics Committee of the Department of Biosciences, COMSATS Institute of Information Technology, Islamabad as well as parental consent was obtained before sample acquisition. Information on age, gender and vaccination history pertaining to each patient was recorded. The age of patients recruited in this case study ranged from 6 months to 5 years. Nasopharyngeal swabs were obtained carefully as described in standard CDC protocols in order to minimize cross-contamination and transferred in transport medium. Samples were cultured on Bordet–Gengou (BG) agar containing 10% horse blood and 40 μg/ml cefalexin to minimize growth of other bacteria and routine biochemical tests were performed before DNA extraction and polymerase chain reaction (PCR). Bacterial culture, identification, DNA extraction and PCR were performed as described previously [Reference Schmidt-Schlapfer8, Reference Boursaux-Eude and Guiso9]. Routine identification tests included nitrate reduction, urease, citrate, indole, oxidase, and catalase. Growth on MacConkey agar was used for initial identification. Further confirmation based on the formation of β-haemolytic colonies as seen on BG agar supplemented with 5% defibrinated sheep blood and a positive reaction of these colonies with antiserum to B. parapertussis ZM11 (Remel, The Netherlands) was achieved. Total genomic DNA templates isolated were screened for the presence of insertional sequences IS481, IS1001 and IS1002 [Reference Schmidt-Schlapfer8, Reference Boursaux-Eude and Guiso9]. The presence of IS1001 was first detected in these to confirm B. parapertussis positivity, next amplification of IS1002 was used to confirm the origin of the isolate, i.e. human B. parapertussis (Fig. 1). B. parapertussis (B0024) was used as a positive control.

Fig. 1. Agarose gel electrophoresis showing PCR amplification products (a) IS1001, (b) IS1002, respectively. The first lane represents the positive control, B. parapertussis B0024. Lanes 2–7 are some of the representative B. parapertussis isolates collected during December 2009–January 2010.

During the course of this study we observed a minor outbreak of B. parapertussis in Bisham (2009–2010), in patients with cough-like illness, in particular children aged ⩽5 years (December–January, 2009–2010) (Table 1). B. parapertussis was isolated from 21 patients (mean age 3·46 years). During the first peak of the outbreak (December 2009), 12 children proved to be culture positive for B. parapertussis whereas during the second peak (January 2010), nine children tested positive for B. parapertussis. One hundred percent B. parapertussis cases were detected during the surveillance period.

Table 1. Summary of B. parapertussis isolates and affected clinical cases

PCR, Polymerase chain reaction.

The investigation was performed with the objective of evaluating the species of Bordetella circulating in the population that had not been vaccinated during the past 3-4 years. Vaccination history (dates of vaccinations) for all subjects was obtained in order to evaluate vaccine coverage and efficacy. Coincidentally, all isolates were recovered from non-vaccinated patients.

Additional data from adjacent towns would have provided a more comprehensive overview of the situation in the region. However, further follow-up could not be carried out due to concerns regarding prevailing political unrest in the region at that particular time. Since we detected B. parapertussis in a large number of clinical samples from patients that had not been vaccinated, we speculate that this was due to the deteriorating political events and lack of security, compromising outreach of vaccinators in the region [7].

In order to evaluate vaccination status in the region, vaccine coverage in KPK with reference to the rest of the country was analysed. Vaccine coverage data obtained from the Extended Immunization Programme (EPI) show about 44–55% pertussis vaccine coverage in the relevant major district, Swat, during the years 2008–2010 compared to overall coverage in Pakistan (Supplementary Figs S1 and S2). This could be a reflection of the vaccinators being unable to reach the susceptible population in that area due to the prevailing political unrest in the KPK region during the study period. Furthermore, with the existing health system in Pakistan, like many other countries in the region, effective control of disease is difficult due to lack of trained health experts, poorly equipped laboratories and facilities lacking the capacity to distinguish B. pertussis and B. parapertussis. This situation is further exacerbated due to the poor outreach of vaccinators in times of political turmoil.

Furthermore, data pertaining to the overall prevalence of Bordetella infections in the Pakistani population, whether immunized or not, is lacking. Despite overall high vaccine coverage in children, clinical cases have been reported consistently throughout the country (Supplementary Figs S3 and S4).

Previous reports depict a relatively smaller percentage of B. parapertussis existence in countries like France and Iran where pertussis vaccine coverage has been high [Reference Ghanaie10, Reference Mooi11]. While in countries like Finland where diphtheria-pertussis-tetanus (DTP_w) vaccine is used, a large percentage of pertussis cases are caused by B. parapertussis. Therefore, it can be suggested that factors other than vaccination such as geographical elements may also contribute to the higher incidence of B. parapertussis disease [Reference Kurova12]. This fact is further supported by surveillance data suggesting decreased effectiveness of typical pertussis vaccines against B. parapertussis [Reference Mughal13]. A recent study by Mughal et al., revealed the presence of pertussis in vaccinated children in the Khairpur district of Sindh [Reference Bokhari14]. Moreover, pertussis outbreak data from the National Institute of Health (NIH), Pakistan show a total of 158 suspected pertussis cases during the years 2011-2013 in KPK alone, which may be possibly due to B. parapertussis (Supplementary Fig. S4). This study, as well as pertussis outbreak data from NIH, further affirms the inefficiency of current vaccine protocols in elimination of the disease. Another study comparing the prevalence of B. pertussis vs. B. parapertussis surprisingly showed all isolates to be B. parapertussis [Reference Zawadka15]. It is known that efficacy and protection against B. parapertussis is significantly higher with the B. pertussis whole-cell vaccine DTwp rather than the acellular vaccine DTap. Furthermore, data obtained from studies on mice suggest that the level of protection and course of the infection are strain specific [Reference Long16]. The available acellular vaccines target B. pertussis antigens and show little or no effectiveness against B. parapertussis. Furthermore, these in vivo infection studies reveal that DTap vaccination against B. pertussis can interfere with the normal host response leading to clearance of B. parapertussis, increasing susceptibility to infection by this bacterium [Reference Long16]. Consequently, there is a possibility of improving pertussis vaccines by including B. parapertussis antigens as adopted by some East European countries [Reference He5, Reference Mughal13]. The likelihood of B. parapertussis emerging in populations where B. pertussis is endemic and DTP_w is the routinely used vaccine is high. This is also supported by the fact that B. parapertussis-induced immunity protects against both species in contrast to B. pertussis-induced immunity which protects only against B. pertussis infection [Reference Wolfe17].

By contrast, considering that routine whole-cell vaccines may offer protection against B. parapertussis; we believe that the local outbreak of B. parapertussis in Bisham, near Swat, may actually have been a result of non-compliance with the vaccine schedule in this area due to ‘war-on-terror’ before 2009–2010. This is supported by the observation that increased vaccine coverage in certain districts in KPK limited disease incidence as seen in a decrease of reported cases in the following years (Supplementary Fig. S2). Vaccination in the KPK region according to a recent survey conducted by the National Institute of Population Studies during the period 2012–2013 shows increased coverage of the DTP3 vaccine compared to other vaccines. However, overall coverage still remains about 10% lower than that reported in Punjab with less than 40% of children having vaccination cards. The authors also report threats received during the survey in KPK and Balochistan, where security issues run deeper and hence vaccine coverage is even lower [18]. Several other studies from war-affected and conflict-ridden regions show that security threat within the country is an important factor affecting health services and overall immunization coverage [Reference Cetorelli19, Reference Mashal20]. Providing increased security to vaccinators in the region as well as maintaining the vaccines in provincial rather federal headquarters will enable a more efficient response.

This high percentage of B. parapertussis cases reported from Bisham creates a disturbing situation for healthcare professionals as well as policy makers. Moreover, B. parapertussis infections may also be associated with their atypical manifestation in the adult population of Pakistan; therefore, we propose adult vaccination with whole-cell vaccines containing both B. pertussis and B. parapertussis to reduce the risk of infection in infants and children through adult reservoirs.

A convincing explanation of this puzzle is difficult because both B. pertussis and B. parapertussis infections are not systematically recorded in Pakistan. However, a preschool booster to increase herd immunity and to reduce transmission to susceptible subjects is a useful strategy, a practice that has been adopted in many countries [Reference Zepp21]. Additionally, adult vaccination is also an important approach to control pertussis infection [Reference Forsyth22]. We strongly recommend ensuring that infants are vaccinated at the correct age and that all scheduled doses comprising of both B. pertussis and B. parapertussis (DTP_wP_aw) are administered on time. Here the role of primary healthcare units should be emphasized, especially in rural areas of the country. Quick diagnosis and treatment of pertussis cases in children and adults having contact with infants would limit the infection and should be provided in primary healthcare units.