Estimation of the Incidence of Guillain-Barré Syndrome During Pregnancy in the United States

Abstract

Guillain-Barré syndrome (GBS) is an adverse event of interest after vaccination, yet few data are available for background rates during pregnancy. We confirmed 2 cases of incident GBS and estimated an incidence of 2.8 confirmed GBS cases per million person-years (95% confidence interval, 0.5–9.3), indicating rare occurrence. Our findings will help inform safety assessments of Zika vaccines in pregnant populations.

Pregnant women are a target population for Zika vaccine candidates due to the consequences of Zika virus infection during pregnancy, including severe fetal brain defects. Because of the increased risk of Guillain-Barré syndrome (GBS) associated with Zika virus infection, there are concerns that Zika vaccine may also enhance the risk of developing GBS [1]. Although rates of GBS in female populations have been estimated in a variety of settings including the United States [2–4], published data for background rates of GBS in a pregnant population are limited to a Swedish cohort study [5].

Little is known about whether immunologic changes during pregnancy can influence rates of neurologic disorders such as GBS; therefore, understanding background rates during pregnancy of outcomes associated with Zika virus infection will be important to quantify any excess risk of such outcomes during safety studies of Zika vaccine in pregnant populations. We describe here the rate of chart-confirmed GBS during pregnancy in the Vaccine Safety Datalink (VSD) population, which includes approximately 2.5 million women of reproductive age in the United States.

METHODS

We identified GBS cases from the inpatient setting using previously described VSD administrative health care data from 7 integrated health care systems [6]. Our study population included pregnant women whose pregnancies ended between January 1, 2004, and July 31, 2015, and who had continuous insurance coverage from start of pregnancy through 42 days after end of pregnancy. We identified pregnancies by using a validated algorithm that utilizes administrative, electronic health record, and claims data [7].

We identified any case of potential GBS during pregnancy and the 42 days after birth using the International Classification of Diseases, 9th Revision (ICD-9), code 357.0 for acute infective polyneuritis in inpatient settings. Trained abstractors reviewed medical records to determine whether each patient with an ICD-9 code was an incident GBS case and recorded relevant clinical information. Two trained personnel independently reviewed the abstracted data and applied Brighton Collaboration case definition criteria to determine diagnostic certainty of GBS [8].

We estimated the rate of Brighton Collaboration level 1–3 GBS in our pregnant population using the Mid-P exact test [9]. Person-time in the study was truncated at symptom onset for GBS and calculated for both total study time (pregnancy plus the 42 days postpartum) and pregnancy only time. The institutional review boards at all participating sites approved the study protocol and abstraction form.

RESULTS

There were 1 203 624 pregnancies among women aged 10 through 55 years (median, 30) that met inclusion criteria, representing 846 427 person-years during the pregnancy and the 42-day postpartum period. Within this population, 35 potential cases of GBS were identified, and 2 cases were confirmed as incident GBS during pregnancy. No cases of incident GBS were diagnosed within the 42 days after delivery. Among the potential cases, 23 were determined to be patients with a history of GBS, 8 were patients with diagnoses of Group B streptococcal infection that were miscoded as GBS, 1 was an apparent miscode not attributable to Group B streptococcal infection, and 1 was determined not to be GBS during clinical follow-up. The majority of patients with a history of GBS (21) had a single ICD-9 code of 357.0 recorded at the hospitalization for delivery. Our overall case confirmation rate for any 357.0 code during pregnancy was 6%.

Both confirmed cases of incident GBS were classified as meeting a Brighton Collaboration diagnostic certainty of level 2 and are described in Table 1. Both recovered with unspecified neurologic sequelae noted at last follow-up, after treatment that included intravenous immune globulin and mechanical ventilation (case number 1) or plasmapheresis and corticosteroids (case number 2). Only case number 2 had maternal immunizations noted in the medical record, having received influenza vaccination early in the first trimester.

Based on the 2 incident cases of GBS in our study population, the rate of incident GBS during pregnancy and the 42 days after delivery was 2.4 per million person-years (95% confidence interval [CI], 0.4–7.8). For pregnancy only time, the rate of incident GBS was 2.8 per million person-years (95% CI, 0.5–9.3).

DISCUSSION

We identified 2 incident cases of GBS in our study, both meeting a relatively high level of diagnostic certainty. Although the highest level of certainty (level 1) for the GBS case definition can be difficult to meet in retrospective studies, the criteria required for the second highest level of certainty were met for both of our incident cases. There were no cases that failed to meet the clinical case definition due to insufficient evidence and no cases that were identified during the postpartum period.

Preceding infections and certain vaccinations have been hypothesized as triggers for the immune-mediated process underlying GBS [8]. One case in our study had a laboratory-confirmed infection of cytomegalovirus concurrent with symptom onset for GBS. The other case reported having no concurrent or antecedent infections; she received seasonal influenza vaccine early in pregnancy and experienced GBS symptom onset >150 days later, well beyond the 42 days generally accepted as the risk window for onset after vaccination. Given the length of time between vaccination and symptom onset, we do not believe that this represents a case of vaccine-associated GBS.

Our chart confirmation rate (or positive predictive value) was <10%, which is lower than previous studies of GBS in the VSD population, where chart confirmation rates have exceeded 50% for hospital settings [3, 10]. There are several reasons for the low confirmation rate in this study. First, to maximize sensitivity for identification of cases during pregnancy, we deliberately chose not to restrict potential case identification to only those with a first ever diagnosis of GBS or first within a set time period. Therefore, we reviewed many more potential cases. Second, miscoding of diagnosis or treatment of Group B streptococcus infection, which is most often associated with pregnancy, represented nearly one-quarter of our potential cases.

Cheng et al. estimated rates of 12 and 40 GBS cases per million person-years in national pregnancy and postpartum (30 days) cohorts in Sweden during 1978–1993, utilizing a national birth registry of live- and still-births and hospital inpatient coded diagnoses further verified to meet diagnostic criteria for GBS [5]. The estimated rate of GBS in our study of a US pregnancy cohort is 4-fold lower, and our study did not find an elevated risk of GBS during the postpartum period. Variations in published incidence rates of GBS have been noted previously, with the caveat that some variability is likely due to differing methods for case ascertainment and clinical case definitions [8].

A limitation of our study is that it may not be fully generalizable to the non-US settings most likely to be targeted for Zika vaccine trials, as their infectious disease burden may be higher and access to health care more limited than in our pregnancy cohort. Also, our estimate of the rate of GBS during pregnancy may be viewed as unstable given that we were only able to confirm 2 cases of GBS after medical record review. However, it is reassuring that in a cohort of >1.2 million pregnancies and using validated methods to identify pregnancies and classify cases of GBS, we found very few cases of GBS occurring during pregnancy.

This study helps fill a critical knowledge gap for a potential adverse event following immunization that will be of particular interest during Zika vaccine trials. Should a signal for GBS be detected during clinical trials of maternal Zika vaccination, it will warrant immediate scrutiny given the low incidence of this syndrome during pregnancy.

Acknowledgments

We gratefully acknowledge all of the Vaccine Safety Datalink investigators and data managers for assistance in collecting data for this study.

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the National Institutes of Health.

Financial support. This work was supported by the Centers for Disease Control and Prevention. Dr. Myers was supported by the National Institute of Allergy and Infectious Diseases (Award Number T32AI074492).

Potential conflicts of interest. All authors: no reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References