Pertussis outbreak investigation in Janamora district, Amhara Regional State, Ethiopia: a case-control study

Introduction On April 17/2017 Janamora district, Amhara regional state health officials reported an increasing number of people with a cough. The objectives of this study was to investigate the outbreak, describe risk factors and implement control measures. Methods We conducted a community based unmatched 1:1 case-control study April 22-May 10, 2017. We used a probable case definition (≥2 weeks cough with vomiting, apnea, or inspiratory whoop) to identify suspected pertussis cases. Neighbors of cases were considered as controls. We conducted a door-to-door active case search and reviewed medical records, assessed vaccination status by parental interview or vaccination card. We implemented multivariable logistic regression to identify independent factors associated with the outbreak. Results We investigated 60 cases and 60 controls. Most (68.3%) of the cases were under the age of 15. The majority (86.6%) of pertussis suspected cases, and 83.4% controls had not received any pertussis vaccine. The overall attack rate was 0.13% and the case fatality rate was 3.3%. The age-specific attack rate for under-five children was 0.33%. Females were more likely to have pertussis (AOR: 2.91; 95% CI: 1.17-7.22), contact with pertussis suspected person (AOR: 6.29; 95% CI: 2.53-15.62) and living in a relatively poorly ventilated house (AOR: 3.01; 95% CI: 1.17-7.70) were also significant risk factors of pertussis. Conclusion Weak supplementary immunization activities might have contributed to the outbreak. Treating household contacts and integration of diagnostic laboratory test of pertussis into the local health system is of paramount importance to detect outbreaks early on.

Incomplete immunization puts children at greater risk of acquiring the disease [11]. The worldwide increase in vaccination coverage against pertussis has substantially reduced the morbidity and mortality associated with this disease [6,12,13]. Over 10 million cases and as many as 400,000 pertussis-related deaths occur annually, with 90% occurring in infants from developing countries [14,15]. The casefatality rates in developing countries are estimated to be as high as 4% [16]. In 2015, more than half of all estimated pertussis-related deaths of children <5 years of age occurred in Africa [17]. In Ethiopia, as part the expanded immunization program (EPI) children are vaccinated for pertussis through DTP (diphtheria-pertussis-tetanus) given at 6 th , 10 th and 14 th weeks. Even though the vaccination coverage in Ethiopia is increasing, the 2016 Ethiopian Demographic Health Survey (EDHS) showed that in Amhara region, only 46% of children aged 12-23 months received all basic vaccinations including DTP [18]. A study conducted in Ethiopia reported an overall attack rate of 1.3 per 1,000 inhabitants and a case fatality rate of 3.7% [19].
Despite the repeated occurrence of pertussis outbreaks in Ethiopia, data on its risk factors is lacking mainly due to lack of surveillance data, weak surveillance system and absence of diagnostic facilities for confirmation of Bordetella pertussis [10]. This study was initiated after a report from the local health office of an increasing number of individuals with cough.

Operational definitions
Vaccinated: a child who has received at least one dose of pertussiscontaining vaccines, and first single dose vaccination status was documented in their log files and/or the mother or the caretaker reported the child received the vaccine. Verbal informed consent was obtained from all participants.

Results
Demographic characteristics: a total of 60 pertussis cases were found of which 36 (60%) were female. Majority (68.3%) of the cases were under the age of 15. The median age of the cases and the controls was 6.5 years (range -9 month to 28 years) and 11.5 years (range -1 year to 34 years) respectively. The majority (n=56, 93%) of the mothers or caretakers for the cases had no formal education.
Descriptive epidemiology: the overall attack rate (AR%) was 0.13% and the age-specific attack rate for under-five children was 0.33%. The sex-specific attack rate for females was 0.159% and for males was 0.11%. We identified a total of 60 suspected pertussis cases and 2 deaths were reported. The case fatality rate was 3.3%.

Discussion
The delay in notification of the outbreak in this study area indicates that there was a poor active disease surveillance system at the local and district level. In this study, only 13.4% of cases had received pertussis-containing vaccine which is much lower compared with findings from a similar study (66.7%) [20]. A possible reason might be that the health extension worker who is responsible for notification of outbreaks and providing routine immunization had been off work for almost a year. Moreover, the distance of the health facility from the community might have also hindered the community from seeking care at the nearest facility. Overall pertussis containing vaccination coverage reported from the district was 69.3% which is lower as compared with the coverage in a similar study (90%) [20]. In our study, only 8 (13.3%) and 4 (6.6%) vaccinated for the pertussiscontaining vaccine of Penta 1 (given on 6 th week) and Penta 2 (given on 10 th week) respectively which is considerably lower in contrast with Amhara regions' vaccine coverage [21]. In our study, the overall attack rate of the outbreak was 0.13% (130 cases per 100,000 populations) which is higher compared to findings from similar outbreaks with 5.7 and 1 per 100 000 population respectively [22][23][24]. Moreover, the case fatality rate among the cases in our study (3.3%) was relatively higher as compared with a similar outbreak (0.17%) [23]. This discrepancy could be attributed to the lack of laboratory technology in the regional state which rendered the challenge to isolate the agent. Moreover, patients visit health facilities very late after falling severely ill which made the early detection and containment of the outbreak slow.
Having a contact history with pertussis suspected person and living in relatively poorly ventilated houses were associated with pertussis, similar to a study conducted in urban Uganda where individuals that were exposed to a coughing individual in the home or neighborhood were more likely to have pertussis [14]. This might be attributed to the fact that the disease is transmitted by droplets or direct contact with discharges from respiratory mucous membranes. This transmission is higher with cases present in the households and even higher in houses with a lower number of windows [25]. In our study, females were nearly three times at a higher risk of contracting pertussis. However, this finding was in contrast to a study which reported odds of getting pertussis disease was not significantly different among sexes [24]. In our study, the possible reason for higher odds of female cases could be that in this particular community, females spend more time working at home and caring for sick children or adults at home, which puts them at a much higher risk of contracting the disease. The limitations in this investigation included that the outbreak was not confirmed by laboratory for a definitive diagnosis. The controls were not matched with respect to possible risk factors and had a ratio of one to one. Furthermore, even though receiving full dose of vaccination is an important risk factor, in this investigation, vaccination cards were not available in some cases and vaccination status was partly determined based on selfreport of mothers or care takers which could have been subjected to recall bias.

Conclusion
The  Vaccination coverage against pertussis is increasing.

What this study adds
 Pertussis outbreak is affected by absence of local health extension workers who provide routine immunization;  Poorly ventilated housing exacerbated pertussis transmission;  Women were prone to contracting pertussis as they provide care for pertussis cases at home.

Competing interests
The authors declare no competing interests.

Authors' contributions
LA conceived and designed the study, analyzed and interpreted data, prepared manuscript. HB prepared data collection instrument, collected data, interpreted data and reviewed manuscript. All authors read and approved the final manuscript.

Acknowledgments
We would like to thank all the respondents and data collectors involved in this study. Table 1: age and sex distribution of the study subjects, Janamora