Vaccine 33 (2015) 1815–1823
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Pandemic influenza A vaccination and incidence of Guillain–Barré syndrome in Korea Changsoo Kim a , Seonkyeong Rhie b , Mina Suh c , Dae Ryong Kang d , Yoon Jung Choi e , Geun-Ryang Bae f , Young-Chul Choi h , Byung Yool Jun i , Joon Soo Lee j,∗ a
Department of Preventive Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea Department of Pediatrics, CHA Bundang Medical Center, CHA University, Seongnam, Republic of Korea c National Cancer Center, Goyang, Republic of Korea d Biostatistics Collaboration Unit, Yonsei University College of Medicine, Seoul, Republic of Korea e Health Insurance Review and Assessment Service, Seoul, Republic of Korea f Korea Centers for Disease Control and Prevention, Osong, Republic of Korea h Department of Neurology, Yonsei University College of Medicine, Gangnam Severance Hospital, Seoul, Republic of Korea i Department of International Health Science, Graduate School of Public Health, Yonsei University, Seoul, Korea j Department of Pediatrics, Severance Children’s Hospital, Epilepsy Research Center, Yonsei University College of Medicine, Seoul, Republic of Korea b
a r t i c l e
i n f o
Article history: Received 17 September 2014 Received in revised form 16 February 2015 Accepted 16 February 2015 Available online 26 February 2015 Keywords: Guillain–Barré syndrome Influenza Incidence rate Vaccination
a b s t r a c t Background: Many studies have investigated the association between Guillain–Barré syndrome (GBS) and vaccinations during the influenza A H1N1 pandemic of 2009 (pH1N1). Based on a nationally representative sample, we estimated the incidence of GBS during the pandemic period in Korea. Materials and methods: All medical records of GBS cases were reviewed in 28 randomly selected hospitals during 2008–2010, and the number of GBS cases at the national level was extrapolated using emergency care utilization data. The GBS rate per 100,000 person-years was estimated in the reference and pandemic periods. Results: The incidence of GBS was 0.63 (95% CI: 0.37–0.89) per 100,000 person-years in the reference period and 0.87 (0.49–1.26) in the pandemic period. During the vaccination season, the pandemic period GBS incidence rate was not significantly higher than the reference period incidence rate (rate ratio: 1.52; 0.99–2.32), but difference was observed among persons aged 20–34 years. Rate of GBS increased after pH1N1 vaccination compared to the reference period (1.46, 1.26–1.68). Discussion: The incidence of GBS increased slightly but not significantly during the pandemic period, although pH1N1 vaccination increased the GBS rate. Therefore, mass influenza vaccination programs should not be precluded on the basis of GBS. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Guillain–Barré syndrome (GBS) is an inflammatory, demyelinating disorder of the spinal nerve roots and peripheral nerves that has an acute to sub-acute onset [1]. The overall incidence of GBS is approximately 1–2 cases per 100,000 person-years. Of all GBS
Abbreviations: GBS, Guillain–Barré syndrome; pH1N1, influenza A H1N1 pandemic of 2009; CI, confidence interval. ∗ Corresponding author at: Department of Pediatrics, Severance Children’s Hospital, Epilepsy Research Center, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea. Tel.: +82 2228 2050. E-mail address: [email protected] (J.S. Lee). http://dx.doi.org/10.1016/j.vaccine.2015.02.035 0264-410X/© 2015 Elsevier Ltd. All rights reserved.
cases, 65–74% are preceded by a recognized infection and 8% are preceded by vaccination [2,3]. Several decades ago, it was reported that the risk of GBS increased substantially in the 6 weeks following vaccination with inactivated swine influenza in the United States, 1976–1977 [4]. Since then, many epidemiologic and experimental studies have investigated the causal relationship between seasonal influenza vaccination and GBS. However, the evidence is still insufficient to definitively confirm or reject a causal relationship. After the emergence of pandemic influenza A H1N1 2009 (pH1N1) in April 2009, many countries contemplated mass vaccination as an important pharmaceutical intervention strategy to prevent the spread of influenza. After the first Korean pH1N1 case was identified on May 2, 2009, the Korea Centers for Disease Control and Prevention (KCDC) began a program of mass influenza
1816
C. Kim et al. / Vaccine 33 (2015) 1815–1823
vaccination on October 27, 2009. Although several studies have reported that pH1N1 vaccination is not associated with an increase in the risk of GBS, other epidemiologic studies have reported exactly the opposite [5–12]. However, these studies were limited by small GBS and pH1N1 vaccination sample sizes. Consequently, it is possible that these studies could miss a small or moderate association between pH1N1 vaccination and GBS risk. Additionally, the population background rates of GBS were rarely considered by these studies, which may introduce bias. In Korea, the adverse events following immunization (AEFI) management system, which is conjoined with the vaccine injury compensation program, is operated by the Korea Centers for Disease Control and Prevention (KCDC). The KCDC reviews AEFI, which are reported through a web-based passive surveillance system handled by physicians and victims or guardians [13,14]. However, due to the limitations of the passive surveillance system, AEFI data should be interpreted with caution when estimating AEFI incidence. Thus, in this study, to estimate GBS incidence and explore its relationship to pH1N1 vaccination, we reviewed all potential GBS cases in a nationally representative sample of hospitals during both the pH1N1 pandemic period (May 2009–April 2010) and an earlier reference period (May 2008–April 2009). 2. Materials and methods 2.1. Study sample Representative hospitals were sampled based on medical utilization data, which were mandatorily and unfailingly reported to the Health Insurance Review & Services (HIRA) between Jan 2006 and June 2009. HIRA includes records of medical utilization, including disease code (International Classification of Diseases, 10th revision [ICD-10]), date of visit, diagnostic tests, drug prescriptions, and a variety of other details. During January 2006–June 2009, we identified 850 hospitals that reported GBS claims to HIRA. For each hospital, the ability to diagnose and manage GBS patients was reviewed by a research committee in our study. Because South Korea is a relatively small country, we assumed that the preferences of patients with GBS might largely depend on the sizes or characteristics of hospitals, rather than their geographic locations. Thus, during the sampling process, we did not consider the geographic location of each hospital. After excluding primary care or geriatric hospitals without the ability to manage GBS patients by a research committee, 71 hospitals were finally identified. They covered 75.6% of all GBS claims data during January 2006–June 2009. Of the 71 hospitals, we randomly selected 28 hospitals, with the exception of one hospital in Jeju Island; in Jeju Island, one representative university hospital was arbitrarily selected. This study was conducted with the approval of the Institutional Review Board of the Yonsei University Health System, Seoul, Korea. 2.2. Medical record review and case identification We retrospectively reviewed all medical records of suspected GBS cases in the 28 hospitals, 2008–2010. Retrospective review was conducted based on the main or sub-diagnoses (G610; ICD10) associated with each medical claim. To facilitate the review of medical records, a case report form was developed by a research committee of neurologists, pediatric neurologists, and epidemiologists. During February 2011–May 2011, trained physicians visited the majority of the selected hospitals, reviewing their medical records. For 15 hospitals, medical records were reviewed by medical record professionals or neurologists who were hospital staff members, on the basis of the committee case report guidelines. The research committee reevaluated each case report form.
Incomplete case reports were reinvestigated by trained physicians via telephone, fax, or the direct review of medical records. The research committee was responsible for the final assignment of GBS levels (1, 2, or 3), to ensure uniform and reliable interpretations of all medical records, according to the Brighton criteria [15]. However, because of the retrospective nature of our study’s design, confirmation could be limited by incomplete medical record information (9 cases of 970). In cases of incomplete information, the research committee reevaluated patient medical records to make a final determination. During January 2008–December 2010, 823 GBS cases were reported in the 28 hospitals. However, the review procedures determined that 128 cases (15.6%) were not eligible for GBS; although cases were initially reported as GBS, GBS was not the final physician-confirmed diagnosis, and instead comprised other diseases (32), prevalent cases (72), and transferred cases (24). Of the 695 incident cases, 370 (53.2%) met the case definition of GBS (levels 1, 2, or 3); during review of medical records by the research committee, 325 cases were finally confirmed as nonGBS or Brighton criteria level 4. To estimate the incidence rate of GBS associated with pH1N1 and mass vaccination, we restricted our attention to the 245 cases that were diagnosed during May 2008–April 2010. 2.3. Influenza vaccination and antiviral drug prescription The first pH1N1 case in Korea was identified on May 2, 2009. The mass vaccination program began on October 27, 2009, consisting of unadjuvanted 15 g pH1N1 monovalent vaccine (GREEN FLU-S® ; Green Cross Corporation, Yong-in, Korea) for persons aged
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Pandemic influenza A vaccination and incidence of Guillain–Barré syndrome in Korea Changsoo Kim a , Seonkyeong Rhie b , Mina Suh c , Dae Ryong Kang d , Yoon Jung Choi e , Geun-Ryang Bae f , Young-Chul Choi h , Byung Yool Jun i , Joon Soo Lee j,∗ a
Department of Preventive Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea Department of Pediatrics, CHA Bundang Medical Center, CHA University, Seongnam, Republic of Korea c National Cancer Center, Goyang, Republic of Korea d Biostatistics Collaboration Unit, Yonsei University College of Medicine, Seoul, Republic of Korea e Health Insurance Review and Assessment Service, Seoul, Republic of Korea f Korea Centers for Disease Control and Prevention, Osong, Republic of Korea h Department of Neurology, Yonsei University College of Medicine, Gangnam Severance Hospital, Seoul, Republic of Korea i Department of International Health Science, Graduate School of Public Health, Yonsei University, Seoul, Korea j Department of Pediatrics, Severance Children’s Hospital, Epilepsy Research Center, Yonsei University College of Medicine, Seoul, Republic of Korea b
a r t i c l e
i n f o
Article history: Received 17 September 2014 Received in revised form 16 February 2015 Accepted 16 February 2015 Available online 26 February 2015 Keywords: Guillain–Barré syndrome Influenza Incidence rate Vaccination
a b s t r a c t Background: Many studies have investigated the association between Guillain–Barré syndrome (GBS) and vaccinations during the influenza A H1N1 pandemic of 2009 (pH1N1). Based on a nationally representative sample, we estimated the incidence of GBS during the pandemic period in Korea. Materials and methods: All medical records of GBS cases were reviewed in 28 randomly selected hospitals during 2008–2010, and the number of GBS cases at the national level was extrapolated using emergency care utilization data. The GBS rate per 100,000 person-years was estimated in the reference and pandemic periods. Results: The incidence of GBS was 0.63 (95% CI: 0.37–0.89) per 100,000 person-years in the reference period and 0.87 (0.49–1.26) in the pandemic period. During the vaccination season, the pandemic period GBS incidence rate was not significantly higher than the reference period incidence rate (rate ratio: 1.52; 0.99–2.32), but difference was observed among persons aged 20–34 years. Rate of GBS increased after pH1N1 vaccination compared to the reference period (1.46, 1.26–1.68). Discussion: The incidence of GBS increased slightly but not significantly during the pandemic period, although pH1N1 vaccination increased the GBS rate. Therefore, mass influenza vaccination programs should not be precluded on the basis of GBS. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Guillain–Barré syndrome (GBS) is an inflammatory, demyelinating disorder of the spinal nerve roots and peripheral nerves that has an acute to sub-acute onset [1]. The overall incidence of GBS is approximately 1–2 cases per 100,000 person-years. Of all GBS
Abbreviations: GBS, Guillain–Barré syndrome; pH1N1, influenza A H1N1 pandemic of 2009; CI, confidence interval. ∗ Corresponding author at: Department of Pediatrics, Severance Children’s Hospital, Epilepsy Research Center, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea. Tel.: +82 2228 2050. E-mail address: [email protected] (J.S. Lee). http://dx.doi.org/10.1016/j.vaccine.2015.02.035 0264-410X/© 2015 Elsevier Ltd. All rights reserved.
cases, 65–74% are preceded by a recognized infection and 8% are preceded by vaccination [2,3]. Several decades ago, it was reported that the risk of GBS increased substantially in the 6 weeks following vaccination with inactivated swine influenza in the United States, 1976–1977 [4]. Since then, many epidemiologic and experimental studies have investigated the causal relationship between seasonal influenza vaccination and GBS. However, the evidence is still insufficient to definitively confirm or reject a causal relationship. After the emergence of pandemic influenza A H1N1 2009 (pH1N1) in April 2009, many countries contemplated mass vaccination as an important pharmaceutical intervention strategy to prevent the spread of influenza. After the first Korean pH1N1 case was identified on May 2, 2009, the Korea Centers for Disease Control and Prevention (KCDC) began a program of mass influenza
1816
C. Kim et al. / Vaccine 33 (2015) 1815–1823
vaccination on October 27, 2009. Although several studies have reported that pH1N1 vaccination is not associated with an increase in the risk of GBS, other epidemiologic studies have reported exactly the opposite [5–12]. However, these studies were limited by small GBS and pH1N1 vaccination sample sizes. Consequently, it is possible that these studies could miss a small or moderate association between pH1N1 vaccination and GBS risk. Additionally, the population background rates of GBS were rarely considered by these studies, which may introduce bias. In Korea, the adverse events following immunization (AEFI) management system, which is conjoined with the vaccine injury compensation program, is operated by the Korea Centers for Disease Control and Prevention (KCDC). The KCDC reviews AEFI, which are reported through a web-based passive surveillance system handled by physicians and victims or guardians [13,14]. However, due to the limitations of the passive surveillance system, AEFI data should be interpreted with caution when estimating AEFI incidence. Thus, in this study, to estimate GBS incidence and explore its relationship to pH1N1 vaccination, we reviewed all potential GBS cases in a nationally representative sample of hospitals during both the pH1N1 pandemic period (May 2009–April 2010) and an earlier reference period (May 2008–April 2009). 2. Materials and methods 2.1. Study sample Representative hospitals were sampled based on medical utilization data, which were mandatorily and unfailingly reported to the Health Insurance Review & Services (HIRA) between Jan 2006 and June 2009. HIRA includes records of medical utilization, including disease code (International Classification of Diseases, 10th revision [ICD-10]), date of visit, diagnostic tests, drug prescriptions, and a variety of other details. During January 2006–June 2009, we identified 850 hospitals that reported GBS claims to HIRA. For each hospital, the ability to diagnose and manage GBS patients was reviewed by a research committee in our study. Because South Korea is a relatively small country, we assumed that the preferences of patients with GBS might largely depend on the sizes or characteristics of hospitals, rather than their geographic locations. Thus, during the sampling process, we did not consider the geographic location of each hospital. After excluding primary care or geriatric hospitals without the ability to manage GBS patients by a research committee, 71 hospitals were finally identified. They covered 75.6% of all GBS claims data during January 2006–June 2009. Of the 71 hospitals, we randomly selected 28 hospitals, with the exception of one hospital in Jeju Island; in Jeju Island, one representative university hospital was arbitrarily selected. This study was conducted with the approval of the Institutional Review Board of the Yonsei University Health System, Seoul, Korea. 2.2. Medical record review and case identification We retrospectively reviewed all medical records of suspected GBS cases in the 28 hospitals, 2008–2010. Retrospective review was conducted based on the main or sub-diagnoses (G610; ICD10) associated with each medical claim. To facilitate the review of medical records, a case report form was developed by a research committee of neurologists, pediatric neurologists, and epidemiologists. During February 2011–May 2011, trained physicians visited the majority of the selected hospitals, reviewing their medical records. For 15 hospitals, medical records were reviewed by medical record professionals or neurologists who were hospital staff members, on the basis of the committee case report guidelines. The research committee reevaluated each case report form.
Incomplete case reports were reinvestigated by trained physicians via telephone, fax, or the direct review of medical records. The research committee was responsible for the final assignment of GBS levels (1, 2, or 3), to ensure uniform and reliable interpretations of all medical records, according to the Brighton criteria [15]. However, because of the retrospective nature of our study’s design, confirmation could be limited by incomplete medical record information (9 cases of 970). In cases of incomplete information, the research committee reevaluated patient medical records to make a final determination. During January 2008–December 2010, 823 GBS cases were reported in the 28 hospitals. However, the review procedures determined that 128 cases (15.6%) were not eligible for GBS; although cases were initially reported as GBS, GBS was not the final physician-confirmed diagnosis, and instead comprised other diseases (32), prevalent cases (72), and transferred cases (24). Of the 695 incident cases, 370 (53.2%) met the case definition of GBS (levels 1, 2, or 3); during review of medical records by the research committee, 325 cases were finally confirmed as nonGBS or Brighton criteria level 4. To estimate the incidence rate of GBS associated with pH1N1 and mass vaccination, we restricted our attention to the 245 cases that were diagnosed during May 2008–April 2010. 2.3. Influenza vaccination and antiviral drug prescription The first pH1N1 case in Korea was identified on May 2, 2009. The mass vaccination program began on October 27, 2009, consisting of unadjuvanted 15 g pH1N1 monovalent vaccine (GREEN FLU-S® ; Green Cross Corporation, Yong-in, Korea) for persons aged
