Impact of rotavirus vaccination on child mortality, morbidity, and rotavirus-related hospitalizations in Bolivia

Impact of rotavirus vaccination on child mortality, morbidity, and rotavirus-related hospitalizations in Bolivia

Accepted Manuscript Title: Impact of rotavirus vaccination on child mortality, morbidity and rotavirus-related hospitalizations in Bolivia Authors: Lu...

548KB Sizes 0 Downloads 27 Views

Accepted Manuscript Title: Impact of rotavirus vaccination on child mortality, morbidity and rotavirus-related hospitalizations in Bolivia Authors: Lucia Inchauste, Maritza Patzi, Kjetil Halvorsen, Susana Solano, Raul Montesano, Volga I˜niguez PII: DOI: Reference:

S1201-9712(17)30160-1 http://dx.doi.org/doi:10.1016/j.ijid.2017.06.006 IJID 2962

To appear in:

International Journal of Infectious Diseases

Received date: Revised date: Accepted date:

1-3-2017 31-5-2017 1-6-2017

Please cite this article as: Inchauste Lucia, Patzi Maritza, Halvorsen Kjetil, Solano Susana, Montesano Raul, I˜niguez Volga.Impact of rotavirus vaccination on child mortality, morbidity and rotavirus-related hospitalizations in Bolivia.International Journal of Infectious Diseases http://dx.doi.org/10.1016/j.ijid.2017.06.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Impact of rotavirus vaccination on child mortality, morbidity and rotavirusrelated hospitalizations in Bolivia

Lucia Inchaustea, Maritza Patzib, Kjetil Halvorsena, Susana Solanob, Raul Montesanoc, Volga Iñigueza*

a

Instituto de Biología Molecular y Biotecnología, Carrera de Biología. Facultad de Ciencias Puras y Naturales. Universidad Mayor de San Andrés, Campus Universitario de Cota Cota, Calle 27, La Paz, Bolivia. b

Programa Ampliado de Inmunización, Ministerio de Salud, Avenida Capitán Ravelo N°2199, La Paz, Bolivia. c

Organización Panamericana de Salud, Calacoto, Calle 18, La Paz, Bolivia.

*Corresponding author: Volga Iñiguez. Instituto de Biología Molecular y Biotecnología, Universidad Mayor de San Andrés, Campus Universitario de Cota Cota, Calle 27, La Paz, Bolivia, E-mail Address: [email protected], Tel.: 591 2-2612815.

Abstract

Background: The public health impact of rotavirus vaccination in countries with high child mortality rates remains to be established. RV1 rotavirus vaccine was introduced in Bolivia in August 2008. We describe trends in deaths, hospitalizations and healthcare visits due to acute gastroenteritis (AGE) and in rotavirus-related hospitalizations, among children <5 years of age, during pre and post-vaccination periods.

Methods: Data was obtained from the National Health Information System to calculate vaccine coverage and AGE-related health indicators. We examined trend reductions of main health indicators using the pre-vaccine period as baseline. Effect of vaccination on the epidemiology of rotavirus-related AGE was assessed using data from the active surveillance hospitals.

Results: Compared with the 2001-2008 pre-vaccine baseline, the mean number of rotavirus-related hospitalizations was reduced by 40.8% (95% confidence interval (CI%) 21.7-66.4) among children <5 years of age in the post-vaccine period (20092013). Reductions were most pronounced in children <1 year of age, eligible for

vaccination.

The

mean

proportion

of

AGE-related

deaths,

AGE-related

hospitalization and AGE-related healthcare visits during 2009-2014 were reduced by 52.5% (CI% 47.4-56.3), 30.2% (CI% 23.5-36.1) and 12.9% (CI% 12.0-13.2) respectively. The greatest effect in reduction of AGE-related deaths was found during the months of seasonal peaks of rotavirus disease. Along post-vaccine period, changes in rotavirus epidemiology were observed, manifested by variation in seasonality and by a shift in the mean age of rotavirus infection.

Conclusions: The significant decrease in main AGE-related health indicators in children <5 years of age after rotavirus vaccine introduction, provides evidence of substantial public health impact of rotavirus vaccination in Bolivia, as a measure for protecting children against AGE.

Keywords: Acute gastroenteritis, Rotavirus Epidemiology, Rotavirus Vaccine, Gastroenteritis-related Mortality, Gastroenteritis-related Hospitalization.

Introduction

Globally, rotavirus is the leading cause of acute gastroenteritis (AGE) in children <5 years of age, causing annually 37% of hospitalizations and approximately 215.000 deaths per year 1. In 2009, the World Health Organization (WHO) recommended including rotavirus vaccination worldwide 2. Two live attenuated oral rotavirus vaccines are licensed and available globally: a human monovalent strain (RV1) (Rotarix, Glaxo SmithKline Biologics) and a pentavalent bovine-human reassortant (RV5) (Rotateq, Merck) 3. Currently, 81 countries have introduced rotavirus vaccines into their National Immunization Programs 4. Similar to the vaccine efficacy clinical trials, vaccine effectiveness (VE) have been found to be lower in countries with greater levels of child mortality

5-7.

Nevertheless, post-introduction surveillance

studies demonstrated the impact of rotavirus vaccines on reducing the burden of AGE among children <5 years of age, both in developing and developed countries 8-11.

Bolivia, has been characterized for displaying high child mortality rates (39 per 1000) among countries in Latin America and the Caribbean region

12.

AGE is considered

the second leading cause of mortality and morbidity among children <5 years of age. Rotavirus is the main pathogen associated with AGE, accounting annually for a high burden of the disease, representing almost half of hospital admissions

13.

The

country was one of the first Global Alliance for Vaccines and Immunization (GAVI) eligible countries to introduce RV1 vaccine into the National Immunization Program, in August 2008. Later, the performance of rotavirus vaccine has been evaluated by two consecutive case control studies in 2011 and 2013, showing an effectiveness (using negative test controls) of 69% and 59% respectively against rotavirus-related hospitalizations and 62% and 72% against very severe disease

5, 14.

The observed

rotavirus VE was similar to other studies performed in developing countries

15, 16.

In

this study, through a combination of active and passive surveillance, we aimed to assess the impact of RV1 administration in Bolivia, along the period of 2009-2014, on the decline in rotavirus-related hospitalizations and main AGE-related health indicators.

Methods

Study Design and Setting

This was a retrospective study conducted from January 2001 to December 2014 (pre-vaccine period: 2001-2008, post vaccine period: 2009-2014). Sentinel surveillance for rotavirus-related AGE was conducted at 8 pediatric hospitals: Hospital del Niño Ovidio Aliaga and Hospital Materno Infantil at the city of La Paz, Hospital Boliviano Holandés and Hospital Los Andes at the city of El Alto, Hospital Albina Patiño and Hospital German Urquidi at the city of Cochabamba, Hospital San Martin de Porres at the rural community of Cochabamba region and Hospital Mario Ortiz at the city of Santa Cruz. Along the study period, all of these hospitals except the Hospital San Martin de Porres (which was incorporated after the vaccine introduction) participated in the rotavirus surveillance network. Surveillance sites were located across three main geographic regions of Bolivia, where the majority of

the population is located and included population with different socio-economic status. The study protocol was approved by the National Bioethics Committee. Rotavirus surveillance among children <5 years of age hospitalized with AGE, has been in existence since 2001 at the Institute of Molecular Biology and Biotechnology at Universidad Mayor de San Andres. In 2006, this surveillance network became part of the established National Rotavirus Surveillance at the Ministry of Health.

Specimen Collection Storage and Laboratory Testing

A suspected case of rotavirus AGE was defined as a child <5 years of age, who was hospitalized for treatment of acute watery diarrhea (≤3 looser than normal stool in a 24-hour period during the illness, and onset of diarrhea ≤7 days at presentation). Stool samples were collected within 48 hours of hospitalization and tested for group A rotavirus using a commercially available enzyme immunoassay within 24 hours of stool collection (ProSpecT; Oxoid, Cambridge, United Kingdom).

Vaccine Coverage

Annual data (2009-2014) of the national coverage of each dose of RV1 vaccine and pentavalent vaccine was prospectively collected from the records of The National System of Health Information and Epidemiologic Surveillance (Sistema Nacional de Información en Salud y Vigilancia Epidemiológica, SNIS-VE) from the Bolivian Ministry of Health.

Rotavirus-related hospitalizations

For the years 2001-2013, monthly data of AGE- and rotavirus-related hospitalization data were obtained from the Rotavirus Surveillance Program.

Monthly proportion of rotavirus-related hospitalizations was obtained by using the number of rotavirus positive hospitalizations as numerator and number of AGErelated hospitalizations as denominator. To calculate the total number of rotavirusrelated hospitalizations, proportion of rotavirus-related hospitalizations was

extrapolated to the total number of AGE-related hospitalizations, obtained from the SNIS-VE and adjusted for the month of the year and for age.

AGE-related deaths

The number of all-cause and AGE-related deaths among children <5 years of age, were collected monthly from data of SNIS-VE, for the period 2001-2014. Proportion of AGE-related deaths was obtained by using the number of AGE-related deaths as numerator and all-cause of deaths as denominator.

AGE-related hospitalizations

The number of all-cause and AGE-related hospitalizations among children <5 years of age were collected monthly from data of SNIS-VE, for the period 2001-2013. Data from 2014 was not available. Proportion of AGE-related hospitalizations was calculated by using the monthly number of AGE-related hospitalizations as numerator and all-cause hospitalizations as denominator.

AGE-related healthcare visits

The number of AGE-related healthcare visits along the years 2001-2014 were obtained from SNIS-VE. Proportion of AGE-related healthcare visits was obtained by using the monthly number of AGE-related healthcare visits in children <5 years of age as numerator and all-cause healthcare visits as denominator. Pre-vaccine baseline included 2001-2006. SNIS-VE data from 2007 and 2008 were excluded from the analysis due to inconsistency (low quality and lack of data).

Statistical analysis

For all the analysis 2008 was considered a pre-vaccine year, since rotavirus vaccine was introduced after the rotavirus main season, and due to a very low vaccine coverage observed from August to December 2008. The number of rotavirus-related hospitalizations and AGE-related deaths were adjusted to the population growth

factor by using the National Census of Housing and Population. Trends of AGErelated mortality, AGE-related hospitalizations, AGE-related healthcare visits and rotavirus-related hospitalizations among children <5 years of age, before and after rotavirus vaccine introduction were examined by a time series analysis. The time series analysis was performed through exponential smoothing using the HoltWinters procedure, in order to capture the trend and seasonal variation

17, 18.

Time

series analysis was executed using R-studio 3.3.1 with Biwavelet, tserieChaos, fNonlinear and peacots packages. The presence of a structural break at the postvaccine period comparing to a pre-vaccine period was assessed by the Chow Test. The slopes have been evaluated by comparing the sum of squared errors from three regressions -one for each sample period and one for the pooled data, under the null hypothesis, asserting that slopes from the pre and post vaccine periods are not different (Eviews9). In addition, comparison between pre and post-vaccine observed values and between observed and expected means were also analyzed by t-student test using R-studio 3.3.1.

Age distribution of rotavirus positive cases in children <5 years of age was analyzed through a Pearson´s Chi square test for count data using R-studio 3.3.2. Mean rotavirus infection by age group was compared between pre and post-vaccine periods with a t-student test performed using R-studio 3.3.1. Additionally a one way ANOVA, followed by a post-hoc Tukey analysis were performed using IMB SPSS Statistic software.

Results

Vaccine coverage

After vaccine introduction, two-dose coverage of RV1 vaccine increased from 16.4% (2008) to 86.0% (2014), being slightly lower than one-dose coverage. During 2012 and 2013, two-dose RV1 vaccine coverage slightly decreased to 75.7% and 77.6% respectively, compared to the value from previous year. Overall, vaccine coverage of RV1 was below 95% and lower than pentavalent vaccine (Table 1).

Rotavirus-related hospitalizations

During the pre-vaccine years of 2001-2008, the mean number of rotavirus-related hospitalizations in children <5 years of age was 579. After RV1 vaccine introduction, during the period of 2009-2013, this number reduced to 343, corresponding to a reduction of 40.8% (CI% 21.7-66.4; p=0.001) (Figure 1a, Table 2). At the surveillance hospitals, during the pre-vaccine period, annual proportion of rotavirusrelated hospitalizations ranged between 37.2%-51.4% (data not shown), while at the post-vaccine period between 20.1%-28.4% (Table 2). The reduction patterns were more marked among children from 0 to 11 months of age. In this age group, the mean number of rotavirus-related hospitalizations during the pre-vaccine period was 336, reducing to 147 in the post-vaccine period, corresponding to a reduction of 56.3% (CI% 46.5-76.7; p<0.001). Among children from 12 to 23 months, the mean number of rotavirus-related hospitalizations during the pre-vaccine period was 194, reducing to 134 in the post-vaccine period, corresponding to a reduction of 30.9% (CI% -25.5-61.4; p=0.259). Finally, for children from 24 to 59 months, the mean number of rotavirus-related hospitalizations during the pre-vaccine period was 19, increasing to 27 in the post-vaccine period. The increase in this age group, however, was not statistically significant. (p=0.950), Table 2.

At post-vaccine years, there was no difference in the rate reduction of rotavirusrelated hospitalizations between surveillance sites located at different geographic regions of Bolivia (Table 1 Supplementary information).

During the pre-vaccine period, rotavirus-related hospitalizations were present all year long, showing a primary distinctive peak between April and July with a sharp increase during the month of June, followed by a minor peak observed during several years, in the month of September (Figure 1a). During the post-vaccine period different changes in rotavirus epidemiology were observed (Figure 1a): a) reduction in the prevalence and length (one month shorter) of the rotavirus primary peak and blunting of the secondary peak (except for 2011), b) temporal shift of the primary rotavirus peak to the period between July to September, c) severe reduction of rotavirus-related hospitalizations out of the rotavirus season months. In addition, blunting of the primary rotavirus peak, after the first year of RV1 vaccine introduction, among children from 0 to 11 months of age was observed, followed by the blunting

of the peak among children from 12 to 23 months of age in the second post-vaccine year (Figure 1b-c).

Age distribution of rotavirus positive cases

During the post-vaccine period, the distribution of rotavirus positive cases by age among children <5 years, was different than in the pre-vaccine years. By the age of 12 and 23 months 55% and 91% of the child population had already a severe rotavirus infection respectively. In contrast, during the pre-vaccine period, by the same months of age, 67% and 96% of children presented a severe rotavirus infection respectively (Figure 2). Moreover, comparing the mean percentage of rotavirus positive cases among age groups between both periods, significant differences have been observed at 9-11, 18-20 and 24-35 months of age (Table 3).

AGE-related deaths

During the pre-vaccine years 2001-2008, the mean number of AGE-related deaths in children <5 years of age was 16. After RV1 vaccine introduction, during the period of 2009-2014, this number was reduced to 10, corresponding to a reduction of 37.5% (CI% 28.3-44.2; p<0,001) from the mean pre-vaccine baseline (Figure 3a, Table 4).

During the pre-vaccine period, the mean proportion of AGE-related mortality in children <5 years of age was 14.1%, compared to 6.7% in the post-vaccine period, corresponding to a reduction of 52.5% (CI% 47.4-56.3; p<0.001) (Figure 3b, Table 4). For both indicators of mortality, a significant reduction along each year of the post-vaccine period was observed (Table 4).

The reduction in AGE-related deaths was more pronounced during the months of the year with seasonal peaks of rotavirus disease. In the pre-vaccine period, during the rotavirus season months, 825 deaths occurred among children <5 years of age, compared to 177 deaths at the post-vaccine period, corresponding to a reduction of 78.6% (p<0.001) (Figure 4). Conversely, in the pre-vaccine period during the nonrotavirus season months, 734 deaths occurred compared to 545 deaths at the postvaccine period, corresponding to a reduction of 25.7% (p<0.001) (Figure 4). Assessing the post-vaccine rotavirus and non-rotavirus seasons individually, along

each year of the 2009-2014 period, reductions in annual mean number of deaths among children <5 years of age, were significantly larger during rotavirus season months (Table 5).

AGE-related hospitalizations

During the post-vaccine 2009-2013 years, the mean proportion of AGE-related hospitalizations in children <5 years of age was lower than the pre-vaccine baseline (Figure 3c). During the pre-vaccine 2001-2008 years, this proportion was 274.3 per 1000 compared to 191.4 per 1000 at the post-vaccine period, corresponding to a reduction of 30.2% (CI% 23.5-36.1; p<0.001) (Figure 3c, Table 4). Moreover, the magnitude of the reductions along 2009-2012, increased in general with each postvaccine year, been statistically significant (p<0.001) from the baseline 2001-2008 (Table 4).

After the second year of vaccine introduction, the primary peak was reduced, delayed and occurred between June and August. Moreover, the expected proportions of AGE-related hospitalizations showed a primary peak in June (consistent with the rotavirus season) and a secondary peak in January, while the observed proportions showed in general reductions of the primary peak (Figure 3c).

AGE-related healthcare visits

During the post-vaccine 2009-2014 years the mean proportion of AGE-related healthcare visits in children <5 years of age was lower, compared to the pre-vaccine trends (Figure 3d). During the pre-vaccine period, this proportion was 78.2%, compared to 68.1% during the post-vaccine period, corresponding to a reduction of 12.9% (CI% 12.0-13.2; p<0.001) (Table 4). Compared to the baseline period of 2001-2008, the proportion reduction increased with each post-vaccine year, been statistically significant (Table 4).

Discussion

This study evaluates impact of rotavirus vaccination on the disease burden in Bolivia, a country in which rotavirus was responsible for a large proportion of AGE-related cases. By comparing rotavirus-related hospitalizations and main AGE-related health indicators, along 8 years of the pre-vaccine period and 6 years of the post-vaccine period, we provide evidence of substantial public health impact of rotavirus vaccination, as an important measure for protecting children against AGE.

In this study based on active surveillance data over a relatively long term period, we documented changes in rotavirus epidemiology following vaccine introduction, manifested by reductions in rotavirus-related hospitalizations, concomitant with changes in seasonality and with an increase in the mean age of rotavirus infection. The characteristic main rotavirus seasonal peak present at pre-vaccine years, during the transition between autumn and winter, was reduced and delayed in the postvaccine period to late winter and early spring, suggesting a reduction in rotavirus transmission. Moreover, during the first two consecutive post vaccination years, the reduction in rotavirus-related hospitalizations was larger in successive birth cohorts, consistent with the increase in vaccination coverage. Overall, these findings are consistent and comparable to observations from other countries11, 19, 20. Impact of rotavirus vaccination on reducing rotavirus-related hospitalizations was similar across different Bolivian geographic regions (highlands and lowlands) and despite differences in health indicators and socioeconomic levels (data not shown). Of particular relevance, is the rotavirus-related hospitalization decline -comparable to other settings-, observed at the surveillance hospitals located at the city of El Alto a rapidly growing indigenous urban center at the highlands of Bolivia, which display high indicators of poverty, along with high rates of rural-urban migration and chronic undernourishment- 21.

Furthermore, the direct effect of vaccination can be inferred by the largest reduction in rotavirus-related hospitalizations among children from 0 to 11 months of age, consistent with the age group in which the majority of the cases occurred in the prevaccine period. This reduction pattern however, was not seen along 2011-2013 among children from 24 to 59 months of age, when the first cohort of vaccinated children was already included. Although there were no statistical differences among pre and post-vaccine trends along this age group, these data need to be confirmed

by continued surveillance in the future, since vaccine immunity may wane with time 5,

being older children more frequently infected. Vaccine coverage (below 95%)

along 2009-2014, and lower vaccine response due to multiple potential factors such chronic malnutrition, concomitant enteric infections, maternal antibodies and OPV interference, may contribute to the vaccine waning in older children 22-24.

The results from passive surveillance, revealed a significant decline in AGE-related mortality, hospitalizations and healthcare visits observed during the post-vaccine period. Moreover, the Chow test showed that there was a significant structural break between pre-vaccine and post-vaccine period along the time trend of main aforementioned health indicators.

The decline in AGE-related mortality measured both by the number and proportion of deaths is remarkable, considering the dramatic burden of AGE in a country that has shown historically one of the highest child mortality rates, along with high rates of chronic malnutrition and population growth in the region

12, 25,

underscoring the

public health benefits of the vaccination to reduce AGE–related mortality burden in developing countries The specific impact of the rotavirus vaccine on AGE-related deaths can be inferred, from the more pronounced reduction pattern during rotavirus seasonal months, in contrast to the other months of the year. In addition, this study also indicates that despite that VE was lower in Bolivia than in other settings5, 14, its performance on reducing AGE-related mortality is considerable. Moreover, the waning effect of vaccine immunity after the first year of life 5 does not seem to affect the performance of RV1 on reducing mortality rates. The reduction in AGE-related mortality may be associated to the reduction in the severity of the disease and to the shift in the average age of rotavirus infection at the post-vaccine years. During prevaccine period, severe rotavirus cases with the highest Vesikari scores were more prevalent among children less than 1 year, particularly between 3-5 months of age (unpublished data). Across the post-vaccine period, the reductions observed in AGE-related mortality were larger than reductions in AGE-related hospitalization, suggesting that averting severe outcomes of AGE and their consequences at early life, may also indirectly contribute to overall child health, by allowing maturation of the immune system of the gut and preventing deterioration of nutritional status.

In general, these data are comparable to previous studies performed in Latin America and in other low-income countries 11, 26-28. In Brazil after vaccine introduction in 2006, AGE-related deaths reduced overall in 22% during the 2007-2009 period 27, 29.

Likewise in Mexico, AGE-related mortality reduced in 53% among children <5

years of age during the post vaccine period of 2008-2014 9. Moreover, In Botswana, in 2014 AGE-related deaths reduced in 22% among children <2 years of age 10.

Different observations further support the impact of rotavirus vaccination in the decline of AGE-related hospitalizations. First, the magnitude and consistency of the reduction trend that was sustained along the post-vaccine period and was significantly different than the expected secular trend evaluated by time-series analysis. Second, the reduction and delay of the primary peak of AGE-related hospitalizations, which was similar to the pattern observed for rotavirus-related hospitalizations (occurring between autumn and winter season). The specificity of the reduction can be further inferred by the fact that we did not observe a reduction in the secondary peak in the month of January (during the rainy season), where bacterial pathogens are usually more prevalent, as has been shown in a previous study 30. The observed reduction in healthcare visits in the post-vaccine period may in turn reduce the proportion of AGE-related hospitalizations, expanding the benefits of vaccination and further suggesting that rotavirus vaccine may be reducing the transmission rate of the virus in the population.

Similarly, other studies documented the effect of rotavirus vaccination in reducing AGE-related hospitalizations

10, 16, 19, 31and

AGE-related healthcare visits

19, 32.

In El

Salvador, diarrhea related healthcare visits reduced in 35% during 2009 19.

The strength of this study is the combination of active and passive surveillance data, allowing to concomitantly analyze the specific effects of vaccination on rotavirusrelated hospitalizations and main AGE-health indicators. The study however, has some limitations. First, most of the active surveillance sites were located in main urban areas, therefore the rural sites, remain under-represented. Second, the impact of rotavirus vaccine in AGE-related indicators stratified by age was not evaluated due to the lack of data at SNIS. Third, AGE-mortality and morbidity by all causes data might be underreported. However, the effect of underreporting might not be different between the pre and post-vaccine periods and may not significantly bias

the results. Fourth, in some years there was a lack of data for AGE-related hospitalizations (2014) and low quality data for AGE-related healthcare visits (20072008). However, the study period contained monthly data for several years, enough to generate robust inferences. Additionally, other interventions -which were not assessed in this study- such as improvement in sanitation, and reduction of malnutrition in children < 5 years of age, may have also contributed to the observed AGE- related trend declines during post vaccine years. However, these interventions may not explain the magnitude and consistency of the observed declines. Finally, since this is an ecological study, it does not link an exposure to an outcome at an individual level, being therefore prone to ecological fallacy.

Altogether our results encourage other low and middle income countries with high mortality rates to consider introduction of rotavirus vaccination in their National Immunization Programs. Further monitoring of the impact rotavirus-related hospitalizations over next years will provide data on the long-term impact of rotavirus vaccination. Therefore, higher vaccine coverage and continuous prolonged surveillance are needed to establish the full benefits of rotavirus vaccination in Bolivia.

Conflict of interest: The authors report non potential conflict of interest relation to this paper.

Ethical Approval: The study protocol was approved by the National Bioethics Committee.

Acknowledgments 1. This work was part of the UMSA-IBMB Diarrheal Diseases Program, supported by the Swedish International Development Cooperation Agency (SIDA) along the SIDAUMSA research collaboration. We recognize the study staff from the surveillance hospitals and EPI-PAHO network that made completion of this paper possible: Rita Revollo, Yelin Roca, Ericka Cabrera, Fernando Gil, Carlos Terán, Lidia Astroña, Victor Velasco, Luis Tamayo, Leogivildo Alvarez, Ramiro Bustillo, Angel Quispe,

Jaime Rada, Lourdes Carrasco, Rosario Rivera, Vilda Pérez, Reyna Haygua, Aleida Nina, Rodrigo Castedo. 2. We also acknowledge the contribution made by the staff from the SNIS-VE: Max Enriquez Nava and Jhemis Molina and finally Francisco Saavedra and Carlos Veizaga for his help with the time series analysis.

References

1.

Tate JE, Burton AH, Boschi-Pinto C, Parashar UD, World Health Organization-Coordinated Global Rotavirus Surveillance N. Global, Regional, and National Estimates of Rotavirus Mortality in Children <5 Years of Age, 2000-2013. Clin Infect Dis 2016;62 Suppl 2:S96-S105.

2.

WHO. Meeting of the Strategic Advisory Group of Experts on immunization, October 2009 - conclusions and recommendations. Wkly Epidemiol Rec 2009;84:517-32.

3.

World Health Organization. Immunization, vaccines and biologicals. Available at: http://www.who.int/immunization/diseases/rotavirus/en/. Accessed 14 September 2016.

4.

Program for Appropriate Technology in Health. Rotarvirus vaccine access and delivery. Available at: http://sites.path.org/rotavirusvaccine/rotavirusvaccines/#global-intro. Accessed 14 September 2016.

5.

Pringle KD, Patzi M, Tate JE, Iniguez Rojas V, Patel M, Inchauste Jordan L, Montesano R, Zarate A, De Oliveira L, Parashar U. Sustained Effectiveness of Rotavirus Vaccine Against Very Severe Rotavirus Disease Through the Second Year of Life, Bolivia 2013-2014. Clin Infect Dis 2016;62 Suppl 2:S11520.

6.

Gastanaduy PA, Steenhoff AP, Mokomane M, Esona MD, Bowen MD, Jibril H, Pernica JM, Mazhani L, Smieja M, Tate JE, Parashar UD, Goldfarb DM. Effectiveness of Monovalent Rotavirus Vaccine After Programmatic Implementation in Botswana: A Multisite Prospective Case-Control Study. Clin Infect Dis 2016;62 Suppl 2:S161-7.

7.

Gastanaduy PA, Contreras-Roldan I, Bernart C, Lopez B, Benoit SR, Xuya M, Munoz F, Desai R, Quaye O, Tam KI, Evans-Bowen DK, Parashar UD, Patel M, McCracken JP. Effectiveness of Monovalent and Pentavalent Rotavirus Vaccines in Guatemala. Clin Infect Dis 2016;62 Suppl 2:S121-6.

8.

Tate JE, Haynes A, Payne DC, Cortese MM, Lopman BA, Patel MM, Parashar UD. Trends in national rotavirus activity before and after introduction of rotavirus vaccine into the national immunization program in the United States, 2000 to 2012. Pediatr Infect Dis J 2013;32:741-4.

9.

Sanchez-Uribe E, Esparza-Aguilar M, Parashar UD, Richardson V. Sustained Reduction of Childhood Diarrhea-Related Mortality and Hospitalizations in Mexico After Rotavirus Vaccine Universalization. Clin Infect Dis 2016;62 Suppl 2:S133-9.

10.

Enane LA, Gastanaduy PA, Goldfarb DM, Pernica JM, Mokomane M, Moorad B, Masole L, Tate JE, Parashar UD, Steenhoff AP. Impact of Rotavirus Vaccination on Hospitalizations and Deaths From Childhood Gastroenteritis in Botswana. Clin Infect Dis 2016;62 Suppl 2:S168-74.

11.

Ngabo F, Tate JE, Gatera M, Rugambwa C, Donnen P, Lepage P, Mwenda JM, Binagwaho A, Parashar UD. Effect of pentavalent rotavirus vaccine introduction on hospital admissions for diarrhoea and rotavirus in children in Rwanda: a time-series analysis. The Lancet Global Health 2016;4:e129-e36.

12.

UNICEF, WHO, Bank TW, Div UP. Levels and Trends in Child Mortality. Report

2014.

Available

at:

https://www.unicef.org/media/files/Levels_and_Trends_in_Child_Mortality_2 014.pdf. Accessed 21 December 2016.

13.

Smith ER, Rowlinson EE, Iniguez V, Etienne KA, Rivera R, Mamani N, Rheingans R, Patzi M, Halkyer P, Leon JS. Cost-effectiveness of rotavirus vaccination in Bolivia from the state perspective. Vaccine 2011;29:6704-11.

14.

Patel MM, Patzi M, Pastor D, Nina A, Roca Y, Alvarez L, Iniguez V, Rivera R, Tam KI, Quaye O, Bowen M, Parashar U, De Oliveira LH. Effectiveness of monovalent

rotavirus

vaccine

in

Bolivia:

case-control

study.

BMJ

2013;346:f3726.

15.

Patel M, Pedreira C, De Oliveira LH, Tate J, Leshem E, Mercado J, Umaña J, Balmaceda A, Reyes M, Kerin T, McDonald S, Gentsch J, Bowen MD, Parashar U. Effectiveness of Pentavalent Rotavirus Vaccine Against a

Diverse Range of Circulating Strains in Nicaragua. Clinical Infectious Diseases 2016;62:S127-S32.

16.

Armah G, Pringle K, Enweronu-Laryea CC, Ansong D, Mwenda JM, Diamenu SK, Narh C, Lartey B, Binka F, Grytdal S, Patel M, Parashar U, Lopman B. Impact and Effectiveness of Monovalent Rotavirus Vaccine Against Severe Rotavirus Diarrhea in Ghana. Clin Infect Dis 2016;62 Suppl 2:S200-7.

17.

Holt CC, United S, Office of Naval R, Carnegie Institute of T, Graduate School of Industrial A. Forecasting seasonals and trends by exponentially weighted moving averages. Pittsburgh, Pa.: Carnegie Institute of Technology, Graduate school of Industrial Administration; 1957.

18.

Winters PR. Forecasting Sales by Exponentially Weighted Moving Averages. Management Science 1960;6:324-42.

19.

Yen C, Armero Guardado JA, Alberto P, Rodriguez Araujo DS, Mena C, Cuellar E, Nolasco JB, De Oliveira LH, Pastor D, Tate JE, Parashar UD, Patel MM. Decline in rotavirus hospitalizations and health care visits for childhood diarrhea following rotavirus vaccination in El Salvador. Pediatr Infect Dis J 2011;30:S6-S10.

20.

Safadi MA, Berezin EN, Munford V, Almeida FJ, de Moraes JC, Pinheiro CF, Racz ML. Hospital-based surveillance to evaluate the impact of rotavirus vaccination in Sao Paulo, Brazil. Pediatr Infect Dis J 2010;29:1019-22.

21.

Instituto

Nacional

de

Estadística.

Available

at:

http://www.ine.gob.bo/indice/EstadisticaSocial.aspx?codigo=30601. Accessed 21 December 2016.

22.

Emperador DM, Velasquez DE, Estivariz CF, Lopman B, Jiang B, Parashar U, Anand A, Zaman K. Interference of Monovalent, Bivalent, and Trivalent Oral Poliovirus Vaccines on Monovalent Rotavirus Vaccine Immunogenicity in Rural Bangladesh. Clin Infect Dis 2016;62:150-6.

23.

Bautista-Marquez A, Velasquez DE, Esparza-Aguilar M, Luna-Cruz M, RuizMoran T, Sugata K, Jiang B, Parashar U, Patel M, Richardson V.

Breastfeeding linked to the reduction of both rotavirus shedding and IgA levels after Rotarix(R) immunization in Mexican infants. Vaccine 2016;34:5284-9.

24.

Gilmartin AA, Petri WA, Jr. Exploring the role of environmental enteropathy in malnutrition, infant development and oral vaccine response. Philos Trans R Soc Lond B Biol Sci 2015;370.

25.

Ministerio de Salud y Deportes, Instituto Nacional de Estadística, Programa Reforma

de

Salud.

ENDSA

2008.

Available

at:

http://dhsprogram.com/pubs/pdf/FR228/FR228%5B08Feb2010%5D.pdf. Accessed 21 December 2016.

26.

Bayard V, DeAntonio R, Contreras R, Tinajero O, Castrejon MM, OrtegaBarria E, Colindres RE. Impact of rotavirus vaccination on childhood gastroenteritis-related mortality and hospital discharges in Panama. Int J Infect Dis 2012;16:e94-8.

27.

do Carmo GM, Yen C, Cortes J, Siqueira AA, de Oliveira WK, CortezEscalante JJ, Lopman B, Flannery B, de Oliveira LH, Carmo EH, Patel M. Decline in diarrhea mortality and admissions after routine childhood rotavirus immunization in Brazil: a time-series analysis. PLoS Med 2011;8:e1001024.

28.

De Oliveira LH, Giglio N, Ciapponi A, Garcia Marti S, Kuperman M, Sanwogou NJ, Ruiz-Matus C, Marinho de Sousa MF. Temporal trends in diarrhea-related hospitalizations and deaths in children under age 5 before and after the introduction of the rotavirus vaccine in four Latin American countries. Vaccine 2013;31 Suppl 3:C99-108.

29.

Lanzieri TM, Linhares AC, Costa I, Kolhe DA, Cunha MH, Ortega-Barria E, Colindres RE. Impact of rotavirus vaccination on childhood deaths from diarrhea in Brazil. Int J Infect Dis 2011;15:e206-10.

30.

Gonzales L, Joffre E, Rivera R, Sjoling A, Svennerholm AM, Iniguez V. Prevalence, seasonality and severity of disease caused by pathogenic Escherichia coli in children with diarrhoea in Bolivia. J Med Microbiol 2013;62:1697-706.

31.

Mpabalwani EM, Simwaka CJ, Mwenda JM, Mubanga CP, Monze M, Matapo B, Parashar UD, Tate JE. Impact of Rotavirus Vaccination on Diarrheal Hospitalizations in Children Aged <5 Years in Lusaka, Zambia. Clin Infect Dis 2016;62 Suppl 2:S183-7.

32.

Centenari C, Gurgel RQ, Bohland AK, Oliveira DM, Faragher B, Cuevas LE. Rotavirus vaccination in northeast Brazil: a laudable intervention, but can it lead to cost-savings? Vaccine 2010;28:4162-8.

Figure captions:

Figure 1: Observed and predicted trends in the prevalence of rotavirus-related hospitalizations by month of the year among children aged (a) 0 to 59 months, (b) 0 to 11 months, (c) 12 to 23 months, and (d) 24 to 59 months of age, during the prevaccine (2001-2008) and post-vaccine (2009-2013) period. Mean values are represented by a red dotted line. The arrow indicates time of the vaccine introduction. The numbers above the black line in Figure 1a indicate the length of rotavirus seasonal peak in months. Significant differences (p≤0.001) in the slopes of trend lines between the pre-vaccine and post-vaccine period indicating a structural break are denoted by p1**. Significant differences (p≤0.001) during the post-vaccine period between observed and expected values are denoted by p2**.

Figure 2: Number of rotavirus positive cases and cumulative percentage distributed by age group, during the pre-vaccine (2001-2008) and post-vaccine (2009-2014) period among children <5 years of age. Significant differences (p<0,001) between pre and post-vaccine rotavirus positive cases are denoted by p**. p-values are calculated with the use of Pearson Chi-square test for count data.

Figure 3: Observed and predicted trends in (a) AGE-related number of deaths, (b) AGE-related proportion of deaths, (c) AGE-related proportion of hospitalizations, and (d) AGE-related proportion of healthcare visits by month of the year among children <5 years of age, during the pre-vaccine (2001-2008) and post-vaccine (2009-2014) period. Mean values are represented by a red dotted line. The arrow indicates time

of the vaccine introduction. The numbers above the black line in Figure 3c indicate the length of the seasonal peak in months. Significant differences in the slopes of trend lines between the pre-vaccine and post-vaccine period, indicating a structural break are denoted by p1**(p≤0.001) and p1*(p≤0.05). Significant differences during the post-vaccine period between observed and expected values are denoted by p2**(p≤0.001).

Figure 4: Number of AGE-related deaths at the rotavirus and non-rotavirus season among children <5 years of age during the post-vaccine (2009-2014) period, compared to the pre-vaccine (2001-2008) baseline. Percentage reductions at the post-vaccine period are shown in numbers (white colored letters). Significant differences (p≤0.001) during the post-vaccine period between observed and expected values are denoted by p**. p-values are calculated with the use of t- test.

Tables: Table 1 Rotavirus vaccine and pentavalent vaccine coverage among children <1 year of age in Bolivia during the period 2008-2014 Vaccine coverage % Year

Dose 1

Dose 2

Rotavirus

Pentavalent

Rotavirus

Pentavalent

2008

38.2

87.9

16.4

83.2

2009

80.2

87.0

64.7

84.3

2010

83.6

87.3

75.8

82.0

2011

86.1

90.5

80.4

85.8

2012

79.8

85.2

75.7

81.5

2013

82.0

87.5

77.6

83.5

2014

90.8

95.0

86.0

90.0

26

Table 2

Mean number and percentage reduction of rotavirus-related hospitalization among children <5 years of age in Bolivia during the post-va compared with the pre-vaccine (2001-2008) baseline. From 0 to 11 months

From 12 to 23 months

From 24 to 59 months

Mean numbera (%b; 95% CI)

% reductionc (95% CI)

Pd

Mean numbera (%b; 95% CI)

% reductionc (95% CI)

Pd

Mean numbera (%b; 95% CI)

% reductionc (95% CI)

Pd

Pre-vaccine 2001-2008

336 (32; 28 – 34)

-

-

194 (16; 14 – 18)

-

-

19 (2; 2 – 2)

-

-

2009

163 (13; 7 – 19)

51.5 (45.0 - 78.6)

0.049

168 (12; 4 – 20)

13.4 (13.7 - 66.8)

1.000

17 (1; 0 – 2)

10.5 (-516.5 - 129.7)

1.000

2010

150 (15; 7 – 22)

55.4 (45.1 - 78.2)

0.004

124 (11; 4 – 18)

36.1 (6.4 - 72.9)

1.000

14 (1; 0 – 2)

26.3 (-152.7 - 115.3)

1.000

2011

135 (12; 5 – 18)

59.8 ( 26.6 - 82.1)

0.001

129 (11; 5 – 18)

33.5 (-66.0 - 62.3)

0.940

43 (3; 1 – 6)

(+) 126.3 (-1376.2 - 342.8)

0.940

2012

125 (9; 3 – 15)

62.8 (51.4 - 89.7)

0.005

104 (8; 3 – 13)

46.4 (-46.7 - 86.4)

0.262

16 (1; -0 – 2)

15.8 (-300.2 - 132.3)

1.000

2013

163 (11; 5 – 18)

51.5 (37.1 - 82.6)

0.048

144 (10; 4 – 17)

25.8 (-89.0 - 72.7)

1.000

43 (3; 1 - 5)

(+) 126.3 (-2301.0 - 673.0)

0.831

Post-vaccine 2009-2013

147 (11; 9 – 15)

56.3 (46.5 - 76.7)

< 0.001

134 (11; 8 – 13)

30.9 (-25.5 - 61.4)

0.259

27 (2; 1 – 3)

(+) 42.1 (-909.4 - 255.2)

0.950

a Mean

number of rotavirus-related hospitalization per year

b RV+/total

AGE patients tested by enzyme immunoassay

n (%

(43

(26

(27

(28

(20

(25

(25

27

c Percentage dP

reduction of mean numbers

values are calculated with the use of t-test, followed by a Bonferroni correction

Table 3 Distribution of rotavirus positive cases by age group among children <5 years of age pre-vaccine (2005-2008) and post-vaccine (2009-2014). Mean percentage of RV+ cases Age group in months

pre-vaccine period

post-vaccine period

pa

0 to 2

6.9

6.7

0.897

3 to 5

17.5

8.5

0.067

6 to 8

18.2

14.2

0.142

9 to 11

22.9

17.7

0.024

12 to 14

15.5

17.5

0.672

15 to 17

7.4

11.9

0.461

18 to 20

5.3

9.7

0.037

21 to 23

1.6

4.5

0.060

24 to 35

2.9

7.2

0.046

> 35

1.9

2.5

0.754

a

P values are calculated with the use of t-test

28

Tabl e4 Mean number, mean proportion and percentage reduction of AGE-related deaths, AGE-related hospitalizations and AGE-related healthcare visits among children < 5 years of age in Bolivia during the post-vaccine period (2009-2014), as compared with the pre-vaccine (2001-2008) baseline. Number of AGErelated deathsa Me an nu mb ere

% reductio n (95% CI)

Prevacci ne 20012008

16

-

2009

13

2010

10

2011

8

2012

11

18.8 (10.0 45.7) 37.5 (27.1 43.8) 50.0 (41.7 59.1) 31.3 (14.2 43.4)

AGE-related proportion of deathsb

Pd

Mea n prop ortio ne

% reductio n (95% CI)

Pd

-

14.1

-

-

0.9 32

8.4

40.4 (25.5 52.9) 48.2 (39.3 56.1) 61.7 (54.0 68.2) 51.1 (42.1 59.1)

< 0.0 01 < 0.0 01 < 0.0 01 < 0.0 01

< 0.0 01 < 0.0 01 < 0.0 01

7.3

5.4

6.9

AGE-related proportion of hospitalizationsc Mea % n reductio prop n Pd ortio (95% ne CI)

AGE-related proportion of healthcare visitsd Mea % n reductio prop n Pd ortio (95% ne CI)

274. 3

-

-

78.2

230. 8

15.9 (7.3 - 24.9)

0.1 44

69.5

25.9 (18.8 32.9) 35.9 (27.9 42.3) 38.0 (30.3 44.4)

< 0.0 01 < 0.0 01 < 0.0 01

203. 2 175. 8 170. 0

69.0

68.6

68.2

-

-

11.1 (10.0 12.0) 11.8 (10.9 12-2) 12.3 (11.0 13.0) 12.8 (11.6 13.6)

< 0.0 01 < 0.0 01 < 0.0 01 < 0.0 01

29

2013

10

2014

8

Postvacci ne 20092015

10

37.5 (52.7 21.5) 50.0 (39.8 55.9)

< 0.0 01 < 0.0 01

37.5 (28.3 44.2)

< 0.0 01

6.0

6.2

6.7

57.4 (49.3 64.6) 56.0 (45.8 65.2)

< 0.0 01 < 0.0 01

52.5 (47.4 56.3)

< 0.0 01

176. 9

35.5 (23.8 45.3)

< 0.0 01

67.7

NA

NA

NA

65.6

191. 4

30.2 (23.5 36.1)

< 0.0 01

68.1

13.4 (12.5 13.8) 16.1 (14.8 17.0)

< 0.0 01 < 0.0 01

12.9 (12.0 13.2)

< 0.0 01

a

Absolute AGE-related number of deaths AGE-related proportion of deaths per 100 c AGE-related proportion of hospitalizations per 1000 d AGE-related proportion of healthcare visits per 100 e Mean number or mean proportion per year d P values are calculated with the use of t-test, followed by a Bonferroni correction NA: data not available b

Table 5 Number and percentage reduction of AGE-related deaths among children < 5 years of age in Bolivia during the post-vaccine period (2009-2014), as compared with the pre-vaccine (20012008) baseline, as a sum of monthly means. Annual AGE-related deaths % Number CI reduction Prevaccine 20032008

284

-

-

Rotavirus Season AGErelated deaths % Number CI reduction 143

-

-

Non-Rotavirus Season AGE-related deaths % Number CI reduction 141

-

-

30

2009

155

45.4

2010

124

56.3

2011

93

67.3

2012

133

53.2

2013

119

58.1

2014

98

65.5

Postvaccine 20092014

163

42.6

15.7 59.0 30.8 62.9 42.2 74.5 34.1 68.1 32.1 67.7 40.2 72.7 26.3 55.9

28

80.4

50

65.0

25

82.5

13

90.9

45

68.5

16

88.8

52

63.6

36.9 87.5 25.8 78.2 37.2 86.5 43.3 95.4 22.5 83.3 41.1 92.4 26.2 73.9

127

9.9

74

47.5

68

51.8

120

14.9

74

47.5

82

41.8

111

21.3

-17.2 42.4 21.7 61.1 33.3 76.5 15.5 49.9 27.1 66.8 27.0 65.3 15.0 49.2