Over the last decade, rotavirus (RV) vaccine has been introduced into the immunization schedules of over 80 countries worldwide, leading to 74%–90% reductions in RV hospitalizations and 29%–50% reductions in all-cause gastroenteritis (AGE) admissions for children in both high- and low-income countries.1,2 Direct and indirect vaccine effects have been seen in children and the wider community3 as early as the first year after introduction of vaccine.4 Additional unexpected benefits have included reductions in febrile and nonfebrile seizure pediatric emergency department presentations.5,6 In New Zealand (NZ), RotaTeq (RV-5; Merck & Co., Kenilworth, NJ, live oral attenuated RV vaccine) was introduced to the infant national immunization schedule on July 1, 2014 as 3 doses administered at 6 weeks, 3 and 5 months of age without catch-up and minimal previous private use.7 NZ has a publicly financed healthcare system providing free hospital and outpatient care for residents. Where children require emergency admission, this is provided almost exclusively in public hospitals. Children <13 years old also receive free primary care including scheduled vaccines: in 2016, 94.1% of children at 1 year of age received vaccinations according to the national immunization schedule.8 Before the introduction of RV vaccine in NZ, yearly epidemics occurred with peaks in Winter/Spring (June–November). It was estimated that 43% of AGE admissions in NZ were due to RV infection and that 1 in 43 children were hospitalized for RV by the age of 5.9
Despite many reports worldwide recording the success of RV vaccine on reducing the burden of disease, there have been few that address how vaccine introduction may impact upon the provision and interpretation of RV testing in diagnostic laboratories. RV testing after vaccine introduction presents challenges: there is a risk of generating false positive results,10 and testing may not be targeting other important causes of disease in that patient group, such as norovirus, the predominant cause of gastroenteritis (GE) after vaccine introduction.11
Additionally, changes undertaken in testing algorithms by laboratories may impact on trends of RV detection over time, which are an important component of vaccine surveillance.12 Therefore, it is important to understand how RV testing rates, laboratory methods and test reliability change in the period around vaccine introduction.
We undertook to assess the impact of RV vaccination on GE hospitalizations in the greater Auckland region since the advent of vaccination (January 2015 to December 2015). The greater Auckland region, population 1.4 million, accounts for one third of NZ’s population.13 Our secondary aim was to assess and discuss the impact of changes in laboratory testing over the study period.
MATERIALS AND METHODS
Hospitalization data were obtained from the health intelligence units of the Auckland region district health board (DHB) hospitals. Three DHBs provide healthcare for this region. Data for all ages were included in the study.
Hospitalizations were included if there was an admission date between January 1, 2009, and December 31, 2015, with a primary or secondary diagnosis that had an International Classification of Diseases, 10th Revision, of RV GE or AGE (International Classification of Diseases, 10th Revision, codes; A00–A09). Hospitalization rates for the DHBs were calculated using annual population estimates for the Auckland region from Statistics NZ20 and analyzed by year of age for those 0–4 years old and as a group for those ≥5 years old. For the pre-vaccine period, average annual rates were calculated using the mean of 2009–2013 hospitalizations and relevant population estimates. Data were analyzed by comparing pre-vaccine (January 2009 to December 2013), year of vaccine introduction (January 2014 to December 2014) and post-vaccine (January 2015 to December 2015) periods.
The estimated reduction in hospitalizations were calculated as the difference between mean and the highest and lowest number of hospitalizations in 2009–2013 and those in 2015.
Vaccine coverage data were obtained from the NZ Ministry of Health.21
Laboratory data were obtained from the laboratory information systems of the Auckland region DHBs. Laboratories at these sites provide diagnostic testing for their respective hospitals and outpatient clinics. Samples submitted by primary care providers are tested elsewhere. Data were analyzed by proportion of samples positive for RV over the study period, testing rate over time and stratified by age group.
Annual population estimates for the Auckland region, DHB and age group were used.13
Comparisons were made between pre-vaccine (January 2009 to December 2013), the year of vaccine introduction (January 2014 to December 2014) and post-vaccine period (January 2015 to December 2015) and for annual changes over the study period.
Over the study period, laboratory RV assays and testing algorithms varied by laboratory (Table 1 and results). Positive results presented are those initially reactive by RV immunochromatography (ICT). Polymerase chain reaction (PCR) confirmatory results are also presented separately where these were performed: Counties Manukau DHB (CMDHB) had PCR confirmatory results available on positive RV samples sent to the national reference laboratory (the Institute of Environmental and Scientific Research) for RV genotyping (January 2015 to December 2015), and Auckland DHB (ADHB) performed in-house PCR confirmation of reactive ICTs from September 2015.
Chi square was used to compare categoric variables. Relative risk and 95% confidence intervals (CIs) were calculated to compare rates. Statistical significance was considered where P value <0.05.
Ethical approval was sought and deemed not required for this study in view of anonymized data and audit function (HDEC NZ reference 16/CEN/55).
From July 1, 2014, RV-5 was administered as 3 doses given orally to infants at 6 weeks, 3 and 5 months of age.
By April 1, 2015, 19% of infants <1 year old in the Auckland region had received 3 doses of vaccine. Infant coverage increased rapidly to 78.3% by the end of 2015. From September 2015 to December 2015, the vaccinated cohort also included 20%–35% of 1-year-olds who had received their vaccine in infancy.
Between January 2009 and December 2015, there were 1898 laboratory-confirmed RV hospitalizations and 54,021 AGE admissions (12,705 in those <5 years old) resident in the Auckland region.
Annual winter peaks in RV hospitalization occurred in the pre-vaccine period (2009–2013) with highest monthly incidence in Winter–Spring (June–October) (Fig. 1).
Annual admission rates for children <5 years old were 205–315/100,000 over this period (Table, Supplemental Digital Content 1, http://links.lww.com/INF/C778). The highest rates of admission occurred in children <1 year old (351–683/100,000) followed by those 1–2 years old (338–563/100,000) and then reduced with increasing age. RV hospitalizations for >5-year-olds were uncommon.
Vaccine was introduced onto the immunization schedule in July 2014, and the annual RV season was evident shortly afterwards with rates in <5-year-olds comparable with the pre-vaccine years. In 2015, consistent with an increasing cohort of vaccinated infants, there was a delay in the seasonal onset of RV activity and a significant reduction in RV admissions. Peak hospitalizations were seen in early summer (December 2015). There was a substantial 74% (17–25/100,000 to 7/100,000; P < 0.001) reduction in all-age RV hospitalizations in 2015 compared with the pre-vaccination period. For individual age groups, the greatest reduction (77%; RR: 0.23; P < 0.001) was seen for infants less than a year of age, but statistically significant reductions were also seen for 1–2-year-olds [75%; relative risk (RR): 0.25; P < 0.001] and 2–3-year-olds (41%; RR: 0.59; P = 0.04).
There was no overall change in AGE admissions in 2015 compared with the pre-vaccine period, but a 17% (RR: 0.83; P < 0.001) reduction was seen for children <5 years old. Children 1–2 years old appeared to have the most benefit with a 24% (RR: 0.76; P < 0.001) reduction in this age group though less hospitalizations were also seen for those <1 year old (RR: 0.87; P < 0.001) and those 2–3 years (RR: 0.8; P < 0.001) and 3–4 years old (RR: 0.83; P = 0.002). Notably, there was a 4% (RR: 1.04; P < 0.001) increase in AGE admissions for patients >5 years old in 2015 compared with the pre-vaccine period. The modest reduction in RV hospitalizations for 2–3-year-olds and in AGE hospitalizations for 2–4-year-olds, groups too old to have received vaccine, is consistent with indirect vaccine effects.
We estimate that RV vaccine prevented 316 (range: 289–349) AGE admissions including 180 (142–219) laboratory- confirmed RV admissions for children under 5 years old in the Auckland region in 2015.
All laboratories used ICT throughout the study period, although these differed between sites (Table 1). At the beginning of the study period, testing was automatically performed on all samples submitted from children <3–5 years old, whereas this changed at 2 sites over the study period. From 2012, ADHB tested samples for RV from children only on specific request. From June 2014, in the year of vaccine introduction, CMDHB also ceased automatically performing RV testing of samples submitted from children and performed testing only on request.
At ADHB, restrictions were applied to testing of adults and duplicate samples in 2011–2012.
No changes in testing protocols occurred at Waitemata DHB during the study period.
In the post-vaccine period, confirmatory testing was instituted at CMDHB by the use of a second ICT for reactive samples in August 2015, and at ADHB by the use of a RV PCR from September 2015.
Over the study period, 12,671 RV tests were performed in hospital laboratories in the Auckland region. During the pre-vaccine period, the proportion of laboratory tests positive for RV varied between 11% and 19% for any given year (Table, Supplemental Digital Content 2, http://links.lww.com/INF/C779), and this proportion reduced to 7% in 2015 (P < 0.001).
Positivity rates differed by study site and over time. Rates were significantly lower over the study period at ADHB (11% positive) compared with both CMDHB (20% positive, P < 0.001) and Waitemata DHB (25% positive, P < 0.001).The average annual RV positivity rate at ADHB increased to 14% (range: 11%–19%) in 2012–2014 from 11% (range: 7%–12%) in 2009–2011 (P < 0.001), whereas it was unchanged at the other 2 sites over this period.
The number of RV tests performed in our region reduced over the pre-vaccine period, and rates of testing were 23% (RR: 0.77; 95% CI: 0.72–0.82; P < 0.001) lower in 2015 compared with the 2009–2013. These changes differed between age groups (Fig. 2) and study sites: testing rates were 38% lower in infants (RR: 0.62; P < 0.001) and 31% lower in 1–2-year-olds (RR: 0.69; P < 0.001) in 2013 compared with 2011. By study site, rates were 31% lower at ADHB (RR: 0.69; 95% CI: 0.63–0.75; P < 0.001) in 2013 compared with 2011 and 35% lower at CMDHB (RR: 0.65; 95% CI: 0.6–0.69; P < 0.001) in 2014 compared with 2013.
The declines in testing for 0–3-year-olds, those at ADHB over the pre-vaccine period and CMDHB in the year of vaccine introduction, as well as the increase in positivity rate at ADHB are consistent with changes in laboratory testing protocols during this time (Table 1).
Despite the introduction of vaccine in 2014 and marked reductions in RV hospitalizations in 2015, there was no significant reduction in the overall RV testing between 2015 (104/100,000) and 2013–2014 periods (106/100,000 per year).
Trends were apparent by age group: while testing rates in 0–1- and 2–5-year-olds in 2015 were unchanged compared with 2013–2014, 1–2-year-olds were tested less frequently in 2015 (227/100,000) compared with 2013–2014 (290/100,000 per year; P = 0.002), and there was an increase in testing for those >5 years old (from 20/100,000 per year in 2013–2014 to 24/100,000 in 2015; P = 0.01).
In 2015, 58 (53%) of the 109 positive RV ICT tests in the Auckland region had confirmatory testing by PCR performed. Of these, 19 (33%) were not confirmed as RV positive by PCR.
We compared RV and AGE hospitalizations in the Auckland region before and after the introduction of RV vaccine.
The pre-vaccine mean annual RV hospitalization rate of 258/100,000 per year in children <5 years old is consistent with prior NZ reports of RV hospitalization rates ranging from 146 to 416/100,000.9,14,15 There was a substantial early impact of RV vaccine in the region with a 68% reduction in RV hospitalizations in those <5 years old within 18 months of RV vaccine introduction. This is in line with other high-income countries including the United States,12 United Kingdom4 and Australia16 and a recent NZ report.17
A 17% decline in AGE admissions for children <5 years old, extension of protection to unvaccinated age groups and a later seasonal peak of RV activity were also seen after the introduction of vaccine in common with findings elsewhere.3,18
Using laboratory data, we observed a significant reduction in RV positivity rates between the pre-vaccine (11.9%–18.8%) and post-vaccine periods (7.4%) in the Auckland region. These results are also consistent with those reported in Belgium18 where annual positivity rates fell from 14% to 4% after vaccine introduction. We estimate that RV vaccine prevented 316 (range: 289–349) AGE admissions including 180 (142–219) laboratory-confirmed RV admissions for children under 5 years old in the Auckland region in 2015.
We sought to analyze the implications of vaccine introduction on diagnostic testing for RV and found that diagnostic methods differed between sites over the study period. Laboratory algorithms changed over time; in particular, there was a shift from routine reflex testing in children <5 years old to performing tests only on request. It appears that these changes in algorithms were linked with a reduction in tests performed for children under 2 years old, and may have led to an increase in the positivity rate of RV tests.12 We also saw a difference in test positivity rates between sites which could reflect differing laboratory algorithms, request patterns or patient characteristics.
These changes in laboratory protocols occurred in the pre-vaccine period and year of vaccine introduction. In contrast, there was no change in the number of RV tests requested in the year after vaccine introduction compared with the 2 prior years. This is notable given the changes in RV epidemiology over this period. The modest reduction in AGE hospitalizations in 2015, with an increase in hospitalizations for those >5 years old, offers a partial explanation: because agents of viral GE are clinically indistinguishable,19 trends in RV testing will reflect those performed for AGE. Consistent with this, we observed less testing in 1–2-year-olds, the group where vaccine had the most impact on AGE admissions, and more testing was seen in those >5 years old, who had more hospitalizations in the post-vaccine period. The lack of impact of vaccine introduction on testing in infants and 2–4-year-olds, who also had reduced AGE hospitalization rates in 2015, is intriguing, and may reflect other testing behaviors in these groups. Though we recognize that these changes may represent secular patterns, these findings are not unique: reductions in RV testing of 28%–36% in the United States and 5%–18% in Australia have been reported in the years after introduction of vaccine,20,21 but these declines were smaller and delayed compared with those which occurred in RV activity. It is unclear whether changes in these reports were associated with changes in laboratory practice or clinician testing patterns over time.
If no alterations in RV testing pattern occur with the introduction of vaccine, then continued testing at pre-vaccine rates with a falling prevalence of disease may lead to an increased proportion of falsely positive results, regardless of the specificities of commercially available ICTs and enzyme immunoassays.22,23 In Spain, Lopez-Lacort et al10 estimated that 53% of RV tests, when performed by ICT without PCR confirmation, were likely to be false positive, due to out of season sampling and excessive testing in vaccinated children. In the Auckland region, we found that a third of positive RV tests which underwent confirmatory testing were falsely reactive. While we recognize that confirmatory testing was performed for only a subset of patients, similar proportions of unconfirmed ICT and enzyme immunoassay results have also recently been reported from other NZ laboratories.17 Concerns of generating falsely reactive results led 2 sites in our study to adopt confirmatory testing in the post-vaccine period.
Changes in RV epidemiology are predictable based on the global experience of vaccine efficacy. We believe that in the face of the accompanying challenges to diagnostic testing, guidance should be provided at the vaccine planning stage to address the risks of over testing and generation of false positive results which may occur in the early post-vaccine period.
Laboratories may wish to reconsider their algorithms and consider confirmatory testing for positive tests in vaccinated groups24 in preparation for vaccine introduction. However, it is important that clinicians are also engaged, by reviewing their testing patterns and limiting microbiologic investigation for children presenting with AGE to selected patients such as those with severe disease, immunocompromising conditions and in outbreak management.19,25 As well as directly improving the accuracy of RV testing, these changes may also lead to cost savings by avoiding unnecessary tests, and reduce inappropriate interventions resulting from erroneous results.26,27 As preanalytical decisions to test and laboratory testing algorithms will differ worldwide, a local assessment of practices is warranted.
It is important to note that any changes undertaken around the time of vaccine introduction can influence RV activity trends and must be considered in laboratory-based surveillance programs.12
The introduction of RV vaccination has had a significant early impact on GE hospitalizations for children in the Auckland region in keeping with reports from other high-income countries. Despite this RV testing, rates remained at pre-vaccine levels, with consequent risks of false positive results. Laboratories and clinicians should consider reviewing their RV testing algorithms in preparation for vaccine-related epidemiologic changes.
The authors thank Joanne Hewitt at ESR, New Zealand, for providing results of confirmatory testing.
1. Kollaritsch H, Kundi M, Giaquinto C, et al.Rotavirus vaccines: a story of success. Clin Microbiol Infect. 2015;21:735743.
2. Gruber JF, Hille DA, Liu GF, et al.Heterogeneity of rotavirus vaccine efficacy among infants in developing countries. Pediatr Infect Dis J. 2017;36:7278.
3. Mast TC, Wang FT, Su S, et al.Evidence of herd immunity and sustained impact of rotavirus vaccination on the reduction of rotavirus-related medical encounters among infants from 2006 through 2011 in the United States. Pediatr Infect Dis J. 2015;34:615620.
4. Atchison CJ, Stowe J, Andrews N, et al.Rapid declines in age group-specific rotavirus infection and acute gastroenteritis among vaccinated and unvaccinated individuals within 1 year of rotavirus vaccine introduction in England and Wales. J Infect Dis. 2016;213:243249.
5. Payne DC, Baggs J, Zerr DM, et al.Protective association between rotavirus vaccination and childhood seizures in the year following vaccination in US children. Clin Infect Dis. 2014;58:173177.
6. Sheridan SL, Ware RS, Grimwood K, et al.Febrile seizures in the era of rotavirus vaccine. J Pediatric Infect Dis Soc. 2016;5:206209.
7. Rosie B, Dalziel S, Wilson E, et al.Epidemiology of intussusception in New Zealand
pre-rotavirus vaccination. N Z Med J. 2016;129:3645.
9. Milne RJ, Grimwood KBudget impact and cost-effectiveness of including a pentavalent rotavirus vaccine in the New Zealand
childhood immunization schedule. Value Health. 2009;12:888898.
10. Lopez-Lacort M, Collado S, Díez-Gandía A, et al.Rotavirus, vaccine failure or diagnostic error? Vaccine. 2016;34:59125915.
11. Payne DC, Vinjé J, Szilagyi PG, et al.Norovirus and medically attended gastroenteritis in U.S. children. N Engl J Med. 2013;368:11211130.
12. Aliabadi N, Tate JE, Haynes AK, et alCenters for Disease Control and Prevention (CDC). Sustained decrease in laboratory detection of rotavirus after implementation of routine vaccination—United States, 2000-2014. MMWR Morb Mortal Wkly Rep. 2015;64:337342.
13. Statistics New Zealand
. Census data. 2013. Available at: http://www.stats.govt.nz/Census/2013-census/data-tables/dhb-tables.aspx
. Accessed December 6, 2016.
14. Neuwelt P, Simmons GA Public Health Portrait of Severe Paediatric Gastroenteritis in the Auckland Region: Report of the Auckland Paediatric Gastroenteritis Investigation. 2006. Auckland, New Zealand
: Auckland Regional Public Health Service; Available at: http://www.arphs.govt.nz/Portals/0/About us/Publications and Reports/List of publications and reports/Publication archive/PaedsGastro_Apr06.pdf
. Accessed December 6, 2016.
15. Ardern-Holmes SL, Lennon D, Pinnock R, et al.Trends in hospitalization and mortality from rotavirus disease in New Zealand
infants. Pediatr Infect Dis J. 1999;18:614619.
16. Dey A, Wang H, Menzies R, et al.Changes in hospitalisations for acute gastroenteritis in Australia after the national rotavirus vaccination program. Med J Aust. 2012;197:453457.
17. Public Health Surveillance. ESR, Rotavirus in New Zealand
. 2015. Available at: https://surv.esr.cri.nz/surveillance/Rotavirus.php?we_objectID=4500
. Accessed May 19, 2017.
18. Sabbe M, Berger N, Blommaert A, et al.Sustained low rotavirus activity and hospitalisation rates in the post-vaccination era in Belgium, 2007 to 2014. Euro Surveill. 2016;21:pii=30273.
19. Guarino A, Ashkenazi S, Gendrel D, et alEuropean Society for Pediatric Gastroenterology, Hepatology, and Nutrition; European Society for Pediatric Infectious Diseases. European Society for Pediatric Gastroenterology, Hepatology, and Nutrition/European Society for Pediatric Infectious Diseases evidence-based guidelines for the management of acute gastroenteritis in children in Europe: update 2014. J Pediatr Gastroenterol Nutr. 2014;59:132152.
20. Lambert SB, Faux CE, Hall L, et al.Early evidence for direct and indirect effects of the infant rotavirus vaccine program in Queensland. Med J Aust. 2009;191:157160.
21. Centers for Disease Control and Prevention (CDC). Reduction in rotavirus after vaccine introduction—United States, 2000–2009. MMWR. 2009;58;11461149.
22. Kaplon J, Fremy C, Pillet S, et al.Diagnostic accuracy of seven commercial assays for rapid detection of group A rotavirus antigens. J Clin Microbiol. 2015;53:36703673.
23. Ye S, Lambert SB, Grimwood K, et al.Comparison of test specificities of commercial antigen-based assays and in-house PCR methods for detection of rotavirus in stool specimens. J Clin Microbiol. 2015;53:295297.
24. Australian Government Department of Health. Rotavirus laboratory case definition. 2013. Available at: http://www.health.gov.au/internet/main/publishing.nsf/Content/cda-phlncd-rotavirus.htm
. Accessed December 6, 2016.
25. Polage CR, Solnick JV, Cohen SHNosocomial diarrhea: evaluation and treatment of causes other than Clostridium difficile
. Clin Infect Dis. 2012;55:982989.
26. Fryer AA, Smellie WSManaging demand for laboratory tests: a laboratory toolkit. J Clin Pathol. 2013;66:6272.
27. Hawkins RManaging the pre- and post-analytical phases of the total testing process. Ann Lab Med. 2012;32:516.