Measles cases among adolescents in southern Pakistan 2012–2015: The case for revisiting vaccination strategies

ABSTRACT

Aims: Surveillance of adult measles in Pakistan is a challenge as it does not enjoy the status of a reportable disease unlike childhood cases and therefore cases remain undetected and unreported or misdiagnosed. Consequently no data or estimates of young adult cases, seroprevalence, or estimates of susceptible preadolescent or young adult population exist. We have presented both laboratory conformed and clinically suspected cases of measles occurring in adolescents and adults in the southern province of Sindh in Pakistan. Methods: Through an examination of 2 independent databases, i.e. a laboratory database of measles IgM positive cases and clinically detected cases on surveillance performed by the Disease Early Warning System, we have analyzed and reported age-specific positivity rates from 2012 to 2015 in Sindh, Pakistan. Results: High rates of laboratory confirmed measles were observed in those aged 9 y and younger. Among adolescents and adults, significantly higher positivity rates were observed among those aged 10–19 y. Clinically detected cases from Sindh showed similar distribution of cases. Conclusions: High burden of cases among children <9 y of age confirm that supplementary immunization activities (SIAs) among this age group are inadequate and need to be strengthened. Cases among those 10–19 y further demonstrate the need for consolidating SIAs with an additional strategy to vaccinate those who remain non-immune at college entry and in institutions where outbreaks can be prevented. Such measures are essential to achieving the goal of measles elimination in the country and region.

KEYWORDS:

• adults
• adolescents
• measles
• Pakistan

Introduction

Measles remains a high-priority public health problem for Pakistan.¹ Though the reported incidence of measles has decreased from 4.4 per 100 000 population in 2013 to 0.74 per 100 000 population in 2014, over the last 3 years, several outbreaks have been reported.² The national strategic plan for measles control is childhood vaccination under the Expanded Program for Immunizations (EPI) at the ages of 9 months and 15 months with 2 doses of measles containing vaccine (MCV-1 and MCV-2) respectively.³ Vaccination coverage remains at a stagnant 61% (based on vaccine registers maintained at district and provincial level) despite government led efforts to increase rates.⁴ The national response plan for prevention of outbreaks is to increase MCV coverage among all children <23 months of age (included in the government run EPI) and follow-up supplementary immunization activities (SIAs) among children 9 months – 9 y.⁵ This underachieves the level of protection required for prevention of measles cases and outbreaks in Pakistan as SIAs are suboptimal given the civil unrest in the country. The resulting high burden among children results in sporadic cases among adolescents and adults.⁶

Through this report we would like to share our observation of instances of measles seropositive cases and occurrence of clinically suspected cases among preadolescents and adolescents in the southern province of Sindh, Pakistan, through 2 independent databases. Laboratory results are from a laboratory database and clinical cases have been collected from the disease surveillance system in the province of Sindh.

Results

Of 975 measles IgM tests performed at the Aga Khan University clinical microbiology laboratory from 2012 to 2015, 18.1% were positive. Tests were requested by physicians due to variable reasons, including fever alone (especially in immunocompromised children to rule out acute atypical measles), fever with rash, or as supportive evidence of vaccination (if negative) along with a positive measles IgG performed for immigration purposes.

Age groups for the tested population based on positivity-specific quintiles showed high positivity rates in infants and children 9 y and younger. Among the adolescent and young adult age groups, the 10–19 y age quintile showed significantly higher positivity rates than the 20–25, and the 26–45 age quintiles (p < 0.001 z-test for sample proportions). Fig. 1 demonstrates percentages of positive samples by age quintiles. Total samples and positive proportions for this calculation are available in Supplementary file 1. Fig. 2 shows cumulative rates of clinical measles reported among children <9 y and in those 10 y of age and older reported in the eDEWS database. Most cases occurring in ages outside of the SIA age range are seen in children 10–15 y (n = 64; 0.6%).

Figure 1. Proportion of measles IgM positive tests at the Clinical Microbiology laboratory, Aga Khan University Hospital Karachi, Pakistan from 2012–2015 by age quintiles. Outside of the SIA age range, highest rates are observed among those 10–19 y of age.

Figure 2. Electronic DEWS data from Sindh (southern Pakistan) show reported measles cases by age groups; 64 of 77 cases in the 10–19 y age group were seen in those aged 10–15 y. Only 4 cases were reported in those aged 20 and above (not included in chart).

Discussion

Age-specific positivity rates show very high laboratory confirmed cases among those 9 y and younger, evidently due to low MCV1 and MCV2 coverage. The incidence of measles among those aged 10 y and older indicates pockets of susceptibility in these age groups which are outside the target SIA age range.

Unstable political and economic milieu, poor governance and weak health immunization system endangers Pakistan to future possible outbreaks as routine immunization and SIAs may fail to deliver intended outcomes.⁷ While MCV coverage among children has not increased, it is hoped that SIAs conducted in 2015 and 2016 will supplement EPI activities and enhance coverage among those aged <10 y. We have observed a very high positivity rate and incidence among infants and young children <5 y. One reason may be the low immunogenicity of the measles vaccine in the Pakistani population.¹ This may be related to the very early age for MCV1 when maternal antibodies may interfere with vaccination. While there is no evidence to support this, there are also several limitations to the very early age for MCV1 which was originally based on assumptions that have been summarized by Aaby et al.⁸ They recommend that age for MCV1 be increased to 12 months.

Meanwhile, preadolescents and young adults remain at risk and outbreaks among susceptible populations especially in educational institutions and healthcare are possible if cases are not recognized and quarantined timely. It is therefore important to increase awareness among the general population regarding risk perception among adolescents and adults as it is often perceived as a childhood illness. Such occurrences are not separately notified to the government and retrospective examination of data miss the window period for action. Although genotyping data are not available for the cases presented, virology reference laboratories in Pakistan have the capacity for genotyping which can be applied to clusters of cases to establish the presence of and investigate outbreaks. Moreover, preadolescents, adolescents, and young adults should be regarded as a susceptible subpopulation for sporadic measles infection by physicians in Pakistan. To detect cases and prevent outbreaks in healthcare, physician training to recognize adult and atypical measles must be imparted by health ministries in Sindh. An additional strategy for immunizing at-risk preadolescents and young adults with no prior history of vaccination, or absence of evidence of immunity can be introduction of public awareness programs to increase risk perception among the general masses.⁹

Our data has several limitations. First, we have used a secondary data source in the form of laboratory data where denominators for population are not available, and cases are passively identified based on physician requests. Furthermore, eDEWS rates could also not be translated into incidence data for Sindh as age-disaggregated population figures were not available. Confirmation rate of clinically suspected cases also remains unknown to us as this data was not provided to us by eDEWS or the provincial government. We did not consolidate data from these 2 sources as eDEWS is not integrated with all measles cases notified to the provincial government by sentinel laboratories and therefore this variation across the numerals as well as denominators produces problems with validity of any combined data. However, data we present are still a useful indicator of the susceptibility of adolescents to measles. For future analyses, data from these sources can be merged by individual-level normalization.

According to the 2009 WHO guidelines on Response to Measles Outbreaks in Measles Mortality, all age groups contributing to measles cases should be vaccinated, especially the older age groups as even low age-specific attack rates indicate a large susceptible population.¹⁰ Further analysis of age-disaggregated surveillance data from Pakistan with strategic planning models is important to determine the suitable age range for future SIAs to reliably cover all susceptible populations at risk of outbreaks. Alonso et al recently describe an excellent example of temporally structured matrices using historical MCV coverage, demographics, and seroprevalence data to construct dynamic mathematical models for measles vaccination strategies.¹¹

Measles control and ultimate regional elimination and global eradication are important global health concerns and lack of control in countries such as Pakistan with weak health systems is a major hindrance in achieving these goals. We therefore recommend that seroprevalence and age-specific incidence data be generated for all provinces in Pakistan for use in strategic mathematical models that can effectively be used to determine the need for a new, wider target age range for SIAs, especially covering school-going children. Increasing the target age range to 9 months −15 y is likely to adequately cover several birth cohorts and support non-selective SIAs in young adults through school and college vaccination programs. Additionally, reinstating eDEWS, physician training, and introducing vaccination of non-immune individuals at college entry will supplement ongoing activities in monitoring and controlling the measles epidemic in Pakistan.

Methods

Laboratory confirmed cases were extracted from the laboratory management systems of the Clinical Microbiology Laboratory of the Aga Khan University Hospital Karachi (the largest private diagnostic laboratory in southern Pakistan, with 122 collection units in Sindh) from 2012 to 2015. Data retrieval and analysis was granted ethical review exemption by the Ethical Review Committee at the Aga Khan University. All IgM positive cases were clinically compatible with measles as detailed by the requesting physician or confirmed post-hoc by laboratory personnel through telephonic histories.

Since laboratory sentinel surveillance in Pakistan may be limited by the restricted access to diagnostic services¹² posing generalizability concerns for the laboratory data, we also examined health facility surveillance data in Sindh over 3 y (2012–2014) from the electronic Disease Early Warning System (eDEWS) established by the World Health Organization (WHO) and operant between 2005 to 2014.¹³ Clinical and confirmed case definitions used by the eDEWS system were based on the 2009 definitions put forward by the National Institutes of Health (NIH), Islamabad, Pakistan, and are detailed in supplementary file 1.¹⁴ Briefly, all index cases were reported serologically or virologically in the reference laboratory (NIH) but no laboratory confirmation was sought for epidemiologically linked cases. As both databases are independent and although elements used (age and evidence of measles) are comparable, data could not be linked deterministically for 2 reasons: a) no individual level identifiers were used in either database (such as national identity card number) and b) denominators for both databases were unknown (laboratory and eDEWS catchments); therefore data was processed and analyzed separately and not linked. Duplicates were removed from each of the databases exported in MS Excel. Data was analyzed to create age-specific quintiles for measles IgM positive cases, and frequencies and graphs derived in MS Excel. Sample proportion differences were analyzed by applying the z test in MS Excel. eDEWS data was also extracted in MS Excel and age-specific frequencies calculated.

Disclosure of potential conflicts of interest

The authors received no funding for this work and have no conflicts of interest.