Early predictors of Guillain-Barré syndrome in the life course of women

Early predictors of Guillain-Barré syndrome in the life course of women Abstract Background We sought to determine if immune disorders early in life were associated with the later risk of Guillain-Barré syndrome, a neurological disorder thought to be infection-related. Methods We conducted a longitudinal cohort study with 16 108 819 person-years of follow-up for a population of 1 108 541 parous women in Quebec, Canada (1989–2014). The outcome was Guillain-Barré syndrome. We identified women with potential risk factors for future Guillain-Barré syndrome, including immune-mediated and rheumatological diseases, cancer, transfusion, surgical procedures and pregnancy-specific disorders. We estimated hazard ratios (HR) and 95% confidence intervals (CI) for the association of risk factors with later onset of Guillain-Barré syndrome, adjusted for personal characteristics of women. Results The overall incidence of Guillain-Barré syndrome was 1.42 per 100 000 person-years. Incidence was higher for women with immune-mediated (8.79 per 100 000 person-years) and rheumatological disorders (9.84 per 100 000 person-years), transfusion (4.41 per 100 000 person-years), and preeclampsia (2.62 per 100 000 person-years). Immune-mediated disorders were associated with six times the risk of Guillain-Barré syndrome (HR 6.57, 95% CI 3.58 to 12.04), rheumatological disorders with seven times the risk (HR 7.23, 95% CI 3.21 to 16.28), transfusion three times the risk (HR 3.58, 95% CI 1.83 to 6.98) and preeclampsia two times the risk (HR 2.01, 95% CI 1.29 to 3.12). Women with other potential risk factors did not have an increased risk of Guillain-Barré syndrome. Conclusions Immune-related conditions that occur early in life are associated with an increased risk of Guillain-Barré syndrome. The pathophysiology of Guillain-Barré syndrome may extend beyond infectious triggers. Autoimmune disease, autoimmune neuropathy, blood transfusion, Guillain-Barré syndrome, preeclampsia, rheumatic diseases Key Messages Pre-existing risk factors for Guillain-Barré syndrome are poorly understood. Using a cohort of 1 million parous women, we examined the association of immune problems early in life, with later risk of Guillain-Barré. Immune-mediated and rheumatological disorders were associated with seven times the risk of Guillain-Barré syndrome after 25 years of follow-up. Blood transfusion and preeclampsia were associated with two to three times the risk of Guillain-Barré. Immune conditions early in life can influence the future risk of Guillain-Barré syndrome beyond infectious triggers. Introduction A growing number of studies indicate that infectious agents are important triggers for Guillain-Barré syndrome,1–3 but there is a paucity of information on other predisposing risk factors. Guillain-Barré syndrome is the most common cause of acute flaccid paralysis worldwide, with an overall incidence of 1–2 cases per 100 000 individuals annually,1,2,4 and the Zika outbreak has recently drawn more attention to this disorder.3 Most patients with Guillain-Barré syndrome make a full recovery, but a proportion suffer permanent disability.2,4 Guillain-Barré syndrome is an autoimmune disease thought to have both environmental and genetic triggers.1 Previous research has focused primarily on environmental exposures such as infectious agents, vaccination or surgical procedures,1–3,5 but the role of preexisting immune disorders as risk factors for Guillain-Barré syndrome has received less attention.2,4 Immune disorders occur in 5% of the population and are more common in women than men.6,7 Although the pathophysiology of immunological diseases varies widely, some of the aetiological pathways recently identified in Guillain-Barré syndrome are present in a range of diseases with immunological features.6,8 Molecular mimicry is critical to the pathogenesis of Guillain-Barré, but is also implicated in multiple sclerosis and type I diabetes mellitus.7 Guillain-Barré syndrome shares genetic properties and characteristics of B cell dysfunction with diseases such as systemic lupus erythematosus and multiple sclerosis.8–10 Biomarker studies show that immune cytokines such as interleukin-22, which is elevated in patients with Guillain-Barré, also play a role in the pathogenesis of myasthenia gravis and rheumatoid arthritis.11 Yet no study has investigated whether immune-related diseases are related to the risk of Guillain-Barré syndrome. Our objective therefore was to test whether medical disorders with immunological features were associated with a greater risk of later onset of Guillain-Barré syndrome. Methods Study population We used a population-based cohort comprising 1 108 541 women who delivered in the province of Quebec, Canada, between 1989 and 2013. The decision to use a cohort of pregnant women was 2-fold. First, the cohort limits the potential for Berkson’s selection bias at cohort inception because hospital delivery in Quebec is independent of disease status. Nearly all women in Quebec deliver in hospital,12 enabling us to construct a representative sample of all parous women in the province. Second, risk of immunological disease is generally higher in women than men,6 allowing us to focus on the population most pertinent to the research question. Women were eligible to enter the cohort if they delivered an infant in any hospital in Quebec between 1989 and 2013, and did not have Guillain-Barré syndrome recorded in their medical history. We followed the women over time for subsequent hospitalizations, with follow-up extending to 31 March 2014. Follow-up ended with hospitalization for Guillain-Barré syndrome (event), death (competing event) or the end of the study (censored). We obtained data on the clinical characteristics of women from the Maintenance and Use of Data for the Study of Hospital Clientele registry, a dataset containing discharge summaries for all hospitalizations in Quebec.13 Discharge summaries are compiled by trained hospital personnel and are validated with rigorous algorithms.13 Each summary includes information on up to 26 diagnoses and 20 procedures, and indicates if admission to an intensive care unit or death occurred during hospitalization. Diagnoses are coded with the 9th and 10th revisions of the International Classification of Diseases (ICD). Procedures are coded with the Canadian Classification of Diagnostic, Therapeutic, and Surgical procedures, and beginning in 2006, the Canadian Classification of Health Interventions. Guillain-Barré syndrome We identified women hospitalized for Guillain-Barré syndrome using the ICD codes 357.0 and G61.0 marked on the hospital discharge summary.14 To corroborate the diagnosis of Guillain-Barré, we determined the proportion of women who were intubated or admitted to an intensive care unit. Approximately 20–30% of individuals with Guillain-Barré syndrome require artificial ventilation and monitoring in an intensive care setting.3 Risk factors We focused on exposures that were common, as Guillain-Barré syndrome has a relatively low incidence and the number of cases was insufficient in women with rare risk factors. We classified potential predictors of Guillain-Barré into two general categories, medical and pregnancy-specific (see Supplementary Table S1, available as Supplementary data at IJE online). We included disorders that had immune and non-immune features in both categories. Medical disorders included immune-mediated diseases (autoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease),6,7,16 rheumatological diseases (systemic lupus erythematosus, rheumatoid arthritis, scleroderma, ankylosing spondylitis),6,7,16 malignant cancer and endometriosis. Rheumatological diseases have an immune component,7 but cancer and endometriosis are less directly linked with immunological pathophysiology. We identified if women had any of these disorders at baseline or during follow-up, and considered them time-fixed as they reflect a phenotype present throughout life. We identified women who received a transfusion any time during follow-up. Evidence is mounting that transfusion has effects on the immune system through immunization and immunomodulation.17 Because previous surgery has been linked to Guillain-Barré,5 we evaluated surgical procedures that commonly occur at delivery or during follow-up, including caesarian section, hysterectomy and cholecystectomy. We analysed transfusion and surgery as time-varying covariates. Pregnancy-specific risk factors included preeclampsia, placental disorders (placenta abruption, accreta, praevia, infarction or retention), gestational diabetes mellitus, preterm birth before 37 weeks of gestation, intrauterine growth restriction, postpartum haemorrhage and multiple pregnancy. Preeclampsia is a hypertensive disorder of pregnancy with a suspected immune component.18,19 Immunological processes are also suspected in preterm birth.20 The remaining pregnancy-specific exposures are less directly linked with immune pathophysiology, and were used as comparison groups. These exposures were analysed as time-fixed covariates. Covariates We considered baseline age (continuous, 12 to 56 years), year of cohort entry (continuous, 1989–2013), socioeconomic deprivation (poorest fifth of the population, no, unknown)12 and total lifetime parity (1, 2, ≥3 deliveries) as potential confounders. To improve adjustment, we transformed age and year into quadratic splines with knots at the 5th, 50th and 95th percentiles.21 Accounting for age as accurately as possible is important, as the incidence of Guillain-Barré syndrome increases with age.2 Data analysis In primary analyses, we calculated the annual incidence of Guillain-Barré syndrome per 100 000 person-years, as well as the cumulative incidence after 25 years of follow-up with death as a competing event.22 We estimated hazard ratios (HR) and 95% confidence intervals (CI) for the risk of Guillain-Barré syndrome using Cox proportional hazards regression, comparing women with and without each risk factor. We used the number of days since cohort entry as the time scale, and censored women who died or never developed Guillain-Barré syndrome by the last day of the study. We analysed exposures as fixed covariates, except for transfusion and surgery which were time-varying. We adjusted all models for age and year at cohort entry, socioeconomic status and total parity. We ensured that the proportional hazards assumption was not violated using log(‐log survival) plots. In secondary analyses, we used Weibull accelerated failure time models to estimate the adjusted median number of years needed for the exposed and unexposed groups to attain a cumulative incidence of 20 cases of Guillain-Barré per 100 000 women.23 In sensitivity analyses, we tested models in which we excluded women who entered the cohort after 2005, to ensure that we had a minimum of 10 years of follow-up for all women. We did so because women who entered the cohort towards the end of the study had significantly shorter follow-up. In addition, use of anti-tumour necrosis factor (TNF) therapies for rheumatological diseases was more likely after 2005. Finally, we examined regression models in which we did not adjust for parity. Women with high fertility may be healthier and less likely to have immunological diseases, making parity a potential mediator rather than confounder. We performed the analysis in SAS v9.3 (SAS Institute Inc., Cary, NC). As the data were de-identified and the study conformed to ethical requirements for research involving humans in Canada, the University of Montreal Hospital Centre’s institutional review board determined that ethical review was not necessary. Results There was a total of 228 cases of Guillain-Barré syndrome during 16 108 819 person-years of follow-up (Table 1). The overall incidence rate of Guillain-Barré in the cohort was 1.42 per 100 000 person-years (95% CI 1.24 to 1.61). Slightly more than a quarter of women with Guillain-Barré syndrome required intensive care unit admission or intubation (28.1%, N = 64) and three died (1.3%). Among medical risk factors, women with rheumatological disease had the highest incidence of Guillain-Barré at 9.84 cases per 100 000 person-years, followed by women with immune-mediated disorders at 8.79 cases per 100 000 person-years. Women who received transfusions had an incidence of Guillain-Barré syndrome of 4.41 per 100 000 person-years. Among pregnancy-specific risk factors, incidence rates were not elevated for any disorder except preeclampsia (2.62 per 100 000 person-years). Table 1 Incidence of Guillain-Barré syndrome according to characteristics of women   No. women  No. person- years  No. Guillain- Barré syndrome  Incidence per 100 000 person-years (95% confidence interval)  Medical disorders   Immune-mediateda  7 632  125 175  11  8.79 (4.87 to 15.87)   Rheumatologicalb  3 578  60 988  6  9.84 (4.42 to 21.90)   Cancer  29 236  544 555  10  1.84 (0.99 to 3.41)   Endometriosis  37 123  737 420  6  0.81 (0.37 to 1.81)  Medical procedures   Transfusion  23 054  204 092  9  4.41 (2.29 to 8.48)   Caesarian section  272 754  3 597 379  51  1.42 (1.08 to 1.87)   Hysterectomy  50 694  515 507  8  1.55 (0.78 to 3.10)   Cholecystectomy  69 299  713 573  11  1.54 (0.85 to 2.78)  Pregnancy-specific   Preeclampsia  64 350  840 607  22  2.62 (1.72 to 3.97)   Placental disordersc  93 442  1 399 364  13  0.93 (0.54 to 1.60)   Gestational diabetes mellitus  73 710  999 006  15  1.50 (0.90 to 2.49)   Preterm birth, <37 weeks  120 281  1 769 983  24  1.36 (0.91 to 2.02)   Intrauterine growth restriction  58 421  816 232  10  1.23 (0.66 to 2.28)   Postpartum haemorrhage  89 909  1 252 527  13  1.04 (0.60 to 1.79)   Multiple pregnancy  25 424  362 029  5  1.38 (0.57 to 3.32)  Covariates   Age at baseline, years            <25  309 235  4 570 103  68  1.49 (1.17 to 1.89)    25–34  686 935  10 031 560  144  1.44 (1.22 to 1.69)    ≥35  112 371  1 507 156  16  1.06 (0.65 to 1.73)   Total parity            1  501 746  6 820 960  120  1.76 (1.47 to 2.10)    2  444 399  6 590 047  86  1.31 (1.06 to 1.61)    ≥3  162 396  2 697 813  22  0.82 (0.54 to 1.24)   Socioeconomic deprivation            Disadvantaged  208 992  2 919 806  41  1.40 (1.03 to 1.91)    No  809 728  11 381 497  167  1.47 (1.26 to 1.71)   Year at baseline            1989–96  497 311  10 752 794  147  1.37 (1.16 to 1.61)    1997–2004  287 942  3 751 490  47  1.25 (0.94 to 1.67)    2005–12  323 288  1 604 535  34  2.12 (1.51 to 2.97)  Total  1 108 541  16 108 819  228  1.42 (1.24 to 1.61)    No. women  No. person- years  No. Guillain- Barré syndrome  Incidence per 100 000 person-years (95% confidence interval)  Medical disorders   Immune-mediateda  7 632  125 175  11  8.79 (4.87 to 15.87)   Rheumatologicalb  3 578  60 988  6  9.84 (4.42 to 21.90)   Cancer  29 236  544 555  10  1.84 (0.99 to 3.41)   Endometriosis  37 123  737 420  6  0.81 (0.37 to 1.81)  Medical procedures   Transfusion  23 054  204 092  9  4.41 (2.29 to 8.48)   Caesarian section  272 754  3 597 379  51  1.42 (1.08 to 1.87)   Hysterectomy  50 694  515 507  8  1.55 (0.78 to 3.10)   Cholecystectomy  69 299  713 573  11  1.54 (0.85 to 2.78)  Pregnancy-specific   Preeclampsia  64 350  840 607  22  2.62 (1.72 to 3.97)   Placental disordersc  93 442  1 399 364  13  0.93 (0.54 to 1.60)   Gestational diabetes mellitus  73 710  999 006  15  1.50 (0.90 to 2.49)   Preterm birth, <37 weeks  120 281  1 769 983  24  1.36 (0.91 to 2.02)   Intrauterine growth restriction  58 421  816 232  10  1.23 (0.66 to 2.28)   Postpartum haemorrhage  89 909  1 252 527  13  1.04 (0.60 to 1.79)   Multiple pregnancy  25 424  362 029  5  1.38 (0.57 to 3.32)  Covariates   Age at baseline, years            <25  309 235  4 570 103  68  1.49 (1.17 to 1.89)    25–34  686 935  10 031 560  144  1.44 (1.22 to 1.69)    ≥35  112 371  1 507 156  16  1.06 (0.65 to 1.73)   Total parity            1  501 746  6 820 960  120  1.76 (1.47 to 2.10)    2  444 399  6 590 047  86  1.31 (1.06 to 1.61)    ≥3  162 396  2 697 813  22  0.82 (0.54 to 1.24)   Socioeconomic deprivation            Disadvantaged  208 992  2 919 806  41  1.40 (1.03 to 1.91)    No  809 728  11 381 497  167  1.47 (1.26 to 1.71)   Year at baseline            1989–96  497 311  10 752 794  147  1.37 (1.16 to 1.61)    1997–2004  287 942  3 751 490  47  1.25 (0.94 to 1.67)    2005–12  323 288  1 604 535  34  2.12 (1.51 to 2.97)  Total  1 108 541  16 108 819  228  1.42 (1.24 to 1.61)  No., number. aAutoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease. bLupus, rheumatoid arthritis, scleroderma, ankylosing spondylitis. cPlacenta: abruption, accreta, praevia, infarction or retention. Women with immune-mediated disease had the highest cumulative incidence of Guillain-Barré syndrome after 25 years (282.4 per 100 000), followed by women with rheumatological disease (208.1 per 100 000) (Table 2). Compared with no disease, immune-mediated disease was associated with six times the risk of Guillain-Barré syndrome (adjusted HR 6.57, 95% CI 3.58 to 12.04), and rheumatological disease with seven times the risk (adjusted HR 7.23, 95% CI 3.21 to 16.28). Transfusion was associated with three times the risk of Guillain-Barré syndrome (adjusted HR 3.58, 95% CI 1.83 to 6.98), but surgical procedures did not increase the risk. No other medical risk factor was associated with a higher risk of Guillain-Barré syndrome. Among transfused women with Guillain-Barré syndrome, there were no cases of rheumatological disease and only one case of immune-mediated disease. Table 2 Association of medical disorders and procedures with risk of Guillain-Barré syndrome   No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Immune-mediatedb   Yes  11  7 632  282.4 (119.4 to 598.2)  6.56 (3.58 to 12.02)  6.57 (3.58 to 12.04)   No  217  1 100 909  29.4 (25.4 to 33.9)  Ref  Ref  Rheumatologicalc   Yes  6  3 578  208.1 (85.0 to 454.9)  7.25 (3.23 to 16.32)  7.23 (3.21 to 16.28)   No  222  1 104 963  30.9 (26.5 to 35.8)  Ref  Ref  Cancer   Yes  10  29 236  45.8 (23.8 to 83.8)  1.35 (0.72 to 2.55)  1.43 (0.75 to 2.71)   No  218  1 079 305  30.7 (26.3 to 35.7)  Ref  Ref  Endometriosis   Yes  6  37 123  19.2 (8.2 to 41.5)  0.59 (0.26 to 1.32)  0.56 (0.25 to 1.27)   No  222  1 071 418  32.1 (27.5 to 37.3)  Ref  Ref  Transfusion   Yes  9  23 054  96.1 (34.7 to 233.4)  3.35 (1.72 to 6.52)  3.58 (1.83 to 6.98)   No  219  1 085 487  30.4 (26.1 to 35.3)  Ref  Ref  Caesarian section   Yes  51  2 72 754  29.9 (21.4 to 41.2)  1.00 (0.73 to 1.37)  1.04 (0.76 to 1.42)   No  177  835 787  31.9 (26.9 to 37.6)  Ref  Ref  Hysterectomy   Yes  8  50 694  34.5 (14.3 to 76.3)  1.33 (0.65 to 2.74)  1.30 (0.63 to 2.67)   No  220  1 057 847  32.2 (27.5 to 37.4)  Ref  Ref  Cholecystectomy   Yes  11  69 299  41.7 (20.1 to 81.2)  1.21 (0.66 to 2.23)  1.25 (0.68 to 2.30)   No  217  1 039 242  31.7 (27.2 to 36.9)  Ref  Ref    No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Immune-mediatedb   Yes  11  7 632  282.4 (119.4 to 598.2)  6.56 (3.58 to 12.02)  6.57 (3.58 to 12.04)   No  217  1 100 909  29.4 (25.4 to 33.9)  Ref  Ref  Rheumatologicalc   Yes  6  3 578  208.1 (85.0 to 454.9)  7.25 (3.23 to 16.32)  7.23 (3.21 to 16.28)   No  222  1 104 963  30.9 (26.5 to 35.8)  Ref  Ref  Cancer   Yes  10  29 236  45.8 (23.8 to 83.8)  1.35 (0.72 to 2.55)  1.43 (0.75 to 2.71)   No  218  1 079 305  30.7 (26.3 to 35.7)  Ref  Ref  Endometriosis   Yes  6  37 123  19.2 (8.2 to 41.5)  0.59 (0.26 to 1.32)  0.56 (0.25 to 1.27)   No  222  1 071 418  32.1 (27.5 to 37.3)  Ref  Ref  Transfusion   Yes  9  23 054  96.1 (34.7 to 233.4)  3.35 (1.72 to 6.52)  3.58 (1.83 to 6.98)   No  219  1 085 487  30.4 (26.1 to 35.3)  Ref  Ref  Caesarian section   Yes  51  2 72 754  29.9 (21.4 to 41.2)  1.00 (0.73 to 1.37)  1.04 (0.76 to 1.42)   No  177  835 787  31.9 (26.9 to 37.6)  Ref  Ref  Hysterectomy   Yes  8  50 694  34.5 (14.3 to 76.3)  1.33 (0.65 to 2.74)  1.30 (0.63 to 2.67)   No  220  1 057 847  32.2 (27.5 to 37.4)  Ref  Ref  Cholecystectomy   Yes  11  69 299  41.7 (20.1 to 81.2)  1.21 (0.66 to 2.23)  1.25 (0.68 to 2.30)   No  217  1 039 242  31.7 (27.2 to 36.9)  Ref  Ref  Ref, reference. aAdjusted for age and year in splines, socioeconomic status and total parity. bAutoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease. cLupus, rheumatoid arthritis, scleroderma, ankylosing spondylitis. Among pregnancy-specific disorders, preeclampsia was associated with the highest cumulative incidence of Guillain-Barré syndrome (55.4 per 100 000 women), with a 2-fold increase in risk (adjusted HR 2.01, 95% CI 1.29 to 3.12) compared with no preeclampsia (Table 3). No other pregnancy risk factor was associated with a greater risk of Guillain-Barré syndrome. Table 3 Association between pregnancy-specific disorders and risk of Guillain-Barré syndrome   No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Preeclampsia   Yes  22  64 350  55.4 (34.7 to 86.1)  1.91 (1.23 to 2.96)  2.01 (1.29 to 3.12)   No  206  1 044 191  30.2 (25.8 to 35.3)  Ref  Ref  Placental disordersb   Yes  13  93 442  18.0 (10.1 to 31.0)  0.63 (0.36 to 1.11)  0.73 (0.42 to 1.28)   No  215  1 015 099  32.8 (28.1 to 38.2)  Ref  Ref  Gestational diabetes mellitus   Yes  15  73 710  41.2 (23.1 to 70.4)  1.05 (0.62 to 1.78)  1.23 (0.73 to 2.09)   No  213  1 034 831  31.0 (26.6 to 36.1)  Ref  Ref  Preterm birth to <37 weeks   Yes  24  120 281  30.1 (19.4 to 45.9)  0.95 (0.62 to 1.45)  1.04 (0.68 to 1.59)   No  204  988 260  31.7 (27.0 to 37.0)  Ref  Ref  Intrauterine growth restriction   Yes  10  58 421  28.8 (14.5 to 54.1)  0.86 (0.45 to 1.61)  0.91 (0.48 to 1.71)   No  218  1 050 120  31.7 (27.2 to 36.9)  Ref  Ref  Postpartum haemorrhage   Yes  13  89 909  22.7 (12.6 to 39.5)  0.71 (0.41 to 1.25)  0.77 (0.44 to 1.35)   No  215  1 018 632  32.3 (27.6 to 37.5)  Ref  Ref  Multiple pregnancy   Yes  5  25 424  46.6 (15.0 to 125.2)  0.97 (0.40 to 2.35)  1.01 (0.42 to 2.46)   No  223  1 083 117  31.3 (26.9 to 36.3)  Ref  Ref    No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Preeclampsia   Yes  22  64 350  55.4 (34.7 to 86.1)  1.91 (1.23 to 2.96)  2.01 (1.29 to 3.12)   No  206  1 044 191  30.2 (25.8 to 35.3)  Ref  Ref  Placental disordersb   Yes  13  93 442  18.0 (10.1 to 31.0)  0.63 (0.36 to 1.11)  0.73 (0.42 to 1.28)   No  215  1 015 099  32.8 (28.1 to 38.2)  Ref  Ref  Gestational diabetes mellitus   Yes  15  73 710  41.2 (23.1 to 70.4)  1.05 (0.62 to 1.78)  1.23 (0.73 to 2.09)   No  213  1 034 831  31.0 (26.6 to 36.1)  Ref  Ref  Preterm birth to <37 weeks   Yes  24  120 281  30.1 (19.4 to 45.9)  0.95 (0.62 to 1.45)  1.04 (0.68 to 1.59)   No  204  988 260  31.7 (27.0 to 37.0)  Ref  Ref  Intrauterine growth restriction   Yes  10  58 421  28.8 (14.5 to 54.1)  0.86 (0.45 to 1.61)  0.91 (0.48 to 1.71)   No  218  1 050 120  31.7 (27.2 to 36.9)  Ref  Ref  Postpartum haemorrhage   Yes  13  89 909  22.7 (12.6 to 39.5)  0.71 (0.41 to 1.25)  0.77 (0.44 to 1.35)   No  215  1 018 632  32.3 (27.6 to 37.5)  Ref  Ref  Multiple pregnancy   Yes  5  25 424  46.6 (15.0 to 125.2)  0.97 (0.40 to 2.35)  1.01 (0.42 to 2.46)   No  223  1 083 117  31.3 (26.9 to 36.3)  Ref  Ref  Ref, reference. aAdjusted for age and year in splines, socioeconomic status, and total parity. bPlacenta: abruption, accreta, praevia, infarction or retention. Analyses of the time taken to reach a cumulative incidence of 20 cases of Guillain-Barré syndrome per 100 000 women indicated that immune-mediated or rheumatological disorders led to faster onset of Guillain-Barré syndrome (Table 4). The median estimated time to reach an incidence of 20 Guillain-Barré cases per 100 000 was less than 2 years for immune-mediated or rheumatological disorders, compared with more than 14 years for women without these disorders. Women who were transfused took around 4 years and women with preeclampsia took around 7 years to reach an incidence of 20 cases of Guillain-Barré per 100 000. Table 4 Number of years to reach a cumulative incidence of Guillain-Barré syndrome of 20 per 100 000 women   Median no. years (interquartile range)a   Risk factor present  Risk factor absent  Medical disorders   Immune-mediatedb  1.7 (1.3 to 2.3)  14.5 (10.5 to 20.0)   Rheumatologicalc  1.5 (1.1 to 2.0)  14.2 (10.3 to 19.6)   Cancer  10.2 (8.1 to 13.9)  14.0 (10.1 to 19.3)   Endometriosis  25.0 (20.1 to 34.8)  13.6 (9.9 to 18.9)  Medical procedures   Transfusion  3.5 (2.6 to 5.0)  14.2 (10.3 to 19.7)   Caesarian section  14.1 (10.4 to 19.5)  13.6 (9.8 to 18.6)   Hysterectomy  13.2 (10.6 to 17.9)  13.4 (9.9 to 18.7)   Cholecystectomy  13.3 (10.2 to 18.7)  13.5 (9.7 to 18.5)  Pregnancy-specific   Preeclampsia  6.8 (4.9 to 9.8)  14.8 (10.7 to 20.3)   Placental disordersd  23.0 (17.2 to 31.9)  13.3 (9.6 to 18.1)   Gestational diabetes mellitus  13.2 (9.6 to 18.6)  14.0 (10.1 to 19.2)   Preterm birth, <37 weeks  14.7 (10.8 to 21.0)  13.8 (9.9 to 18.8)   Intrauterine growth restriction  16.2 (11.7 to 23.3)  13.8 (10.0 to 18.9)   Postpartum haemorrhage  19.6 (14.5 to 28.1)  13.5 (9.8 to 18.5)   Multiple pregnancy  14.2 (10.6 to 20.4)  13.9 (10.0 to 19.1)    Median no. years (interquartile range)a   Risk factor present  Risk factor absent  Medical disorders   Immune-mediatedb  1.7 (1.3 to 2.3)  14.5 (10.5 to 20.0)   Rheumatologicalc  1.5 (1.1 to 2.0)  14.2 (10.3 to 19.6)   Cancer  10.2 (8.1 to 13.9)  14.0 (10.1 to 19.3)   Endometriosis  25.0 (20.1 to 34.8)  13.6 (9.9 to 18.9)  Medical procedures   Transfusion  3.5 (2.6 to 5.0)  14.2 (10.3 to 19.7)   Caesarian section  14.1 (10.4 to 19.5)  13.6 (9.8 to 18.6)   Hysterectomy  13.2 (10.6 to 17.9)  13.4 (9.9 to 18.7)   Cholecystectomy  13.3 (10.2 to 18.7)  13.5 (9.7 to 18.5)  Pregnancy-specific   Preeclampsia  6.8 (4.9 to 9.8)  14.8 (10.7 to 20.3)   Placental disordersd  23.0 (17.2 to 31.9)  13.3 (9.6 to 18.1)   Gestational diabetes mellitus  13.2 (9.6 to 18.6)  14.0 (10.1 to 19.2)   Preterm birth, <37 weeks  14.7 (10.8 to 21.0)  13.8 (9.9 to 18.8)   Intrauterine growth restriction  16.2 (11.7 to 23.3)  13.8 (10.0 to 18.9)   Postpartum haemorrhage  19.6 (14.5 to 28.1)  13.5 (9.8 to 18.5)   Multiple pregnancy  14.2 (10.6 to 20.4)  13.9 (10.0 to 19.1)  aMedian number of years to reach a cumulative incidence of Guillain-Barré syndrome of 20 per 100 000 women, adjusted for age and year in splines, socioeconomic status and total parity. Transfusion and surgery are modelled as time-fixed. bAutoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease. cLupus, rheumatoid arthritis, scleroderma, ankylosing spondylitis. dPlacenta: abruption, accreta, praevia, infarction or retention. For immune-mediated and rheumatological disorders, the incidence of Guillain-Barré syndrome increased sharply and almost immediately compared with women who did not have these diseases (Figure 1). This was also the case for transfusion and preeclampsia, though the increase was not as abrupt or pronounced. Figure 1 View largeDownload slide Cumulative incidence of Guillain-Barré syndrome in women with immune-mediated disease, rheumatological disease, transfusion and preeclampsia. Solid line, exposure present; dotted line, exposure absent. Figure 1 View largeDownload slide Cumulative incidence of Guillain-Barré syndrome in women with immune-mediated disease, rheumatological disease, transfusion and preeclampsia. Solid line, exposure present; dotted line, exposure absent. In sensitivity analyses, exclusion of women who entered the cohort after 2005 did not change the pattern of results. Similarly, exclusion of parity from the regression models did not affect the findings. Discussion In this study comprising up to 25 years of follow-up for a cohort of 1.1 million parous women, we found an association between immunological medical conditions and the subsequent risk of Guillain-Barré syndrome. Women with disorders such as multiple sclerosis, autoimmune haemolytic anaemia, rheumatoid arthritis and systemic lupus erythematosus had seven times the risk of Guillain-Barré syndrome compared with women who had no immune-mediated or rheumatological disease. In addition, conditions with a suspected immune component, including transfusion and preeclampsia, were associated with two to three times the risk. Thus, past history of existing or possible immune dysfunction may be a novel risk factor for Guillain-Barré syndrome. Risk factors for Guillain-Barré syndrome may not be limited to infectious triggers, as previously thought. Further research is necessary to obtain insight into the underlying pathophysiological processes that connect immunological diseases and infectious triggers with Guillain-Barré syndrome. Guillain-Barré syndrome is an acute paralytic neuropathy typically preceded by infection, most commonly Campylobacter jejuni.2 Guillain-Barré neuropathy is caused by a dysfunctional autoimmune response targeting peripheral nerves and spinal roots.2 The two variants of Guillain-Barré syndrome, including acute inflammatory demyelinating polyneuropathy and acute motor axonal neuropathy,4 both involve an aberrant immune response.2 In North America, the demyelinating subtype predominates.1,4 The immunological mechanisms are not completely understood, but molecular mimicry is thought to be an important component of Guillain-Barré, based on studies of C. jejuni demonstrating that surface polysaccharides of the microbe resemble glycoconjugates of human nerve tissues.2 Cross-reactive antibodies generated by B cells in patients with C. jejuni subsequently lead to an autoimmune response against peripheral nerves.2 Guillain-Barré is currently thought to be primarily humorally mediated, although a T cell-mediated response cannot be ruled out.2,4 Interleukin-22, part of the interleukin-10 cytokine family which is produced by T helper cells, is elevated in cerebrospinal fluid of Guillain-Barré patients.11 Although autoimmune diseases tend to cluster in individuals and families,6,7 few studies have investigated if immune diseases could be risk factors for Guillain-Barré syndrome. In a descriptive analysis of comorbidity in 496 patients with Guillain-Barré syndrome, immune-mediated or rheumatological diseases were not specifically identified, although diabetes mellitus, anaemia and hypothyroidism were frequent comorbidities.14 A case-control study of 5296 patients in California found that individuals with multiple sclerosis were more likely to develop Guillain-Barré syndrome compared with controls.24 A large body of research on shared immunological pathways implicates molecular mimicry in the pathogenesis of Guillain-Barré syndrome,1,2,4 type 1 diabetes,7 and less concretely, multiple sclerosis.7,24 Altered expression of the cytokine interleukin-22 is also common to multiple sclerosis, systemic lupus erythematosus, rheumatoid arthritis and Guillain-Barré.11 Anti-TNF therapy for rheumatological disease has been associated with numerous case reports of Guillain-Barré syndrome.25 Moreover, individuals with autoimmune or rheumatological disease may be more prone to adverse immune reactions to infectious triggers such as C. jejuni. Infectious triggers may thus be mediators of upstream risk factors that predispose to Guillain-Barré syndrome. The association of transfusion with Guillain-Barré syndrome is impressive in light of growing evidence that transfusion has lasting immune effects with persisting immunomodulation.17 Transfusion-related immunomodulation has been associated with adverse clinical outcomes such as infection, acceleration of cancer growth, transfusion-associated graft versus host disease, and mortality.26,27 Mechanisms responsible for such effects are unclear, but leukocyte dysfunction has been incriminated for many transfusion-related adverse events.17,27 The mechanisms by which transfusion-related immunomodulation could increase the risk of Guillain-Barré are unknown; however, both involve altered leukocyte responses. We found no association between surgery and Guillain-Barré, but the evidence for surgery is based on case studies, not cohort data,5 and involvement of autoimmune pathways after surgery has yet to be shown. Our results therefore provide novel evidence that transfusion has implications for subsequent risk of Guillain-Barré syndrome. Among pregnancy-specific disorders, only preeclampsia was associated with Guillain-Barré syndrome. The pathophysiology of preeclampsia involves reduced placental perfusion due to abnormally shallow trophoblast invasion with inadequate vascular remodelling of the uterine spiral arteries.19 Imbalance in T-cell subtypes and autoantibody production is implicated in this process.18,28 Growing evidence suggests that preeclampsia is associated with an increase in pro-inflammatory T-cells,28 as well as production of autoantibodies against the angiotensin II receptor.19,28 Immune dysregulation and inflammation in preeclampsia are believed to persist past pregnancy.18 Some studies indicate that preeclampsia increases the risk of rheumatological diseases.29 Many of the immunological pathways in preeclampsia have yet to be firmly established, but the association we found with later risk of Guillain-Barré, a disorder with established autoimmune pathways, is compelling. Limitations This study covered a long period of follow-up in a large sample of women, providing sufficient power to detect associations with Guillain-Barré syndrome, a rare neurological disorder. Although the findings are novel and the biological rationale is robust, a number of limitations deserve mention. First, we could not audit charts or validate cases of Guillain-Barré using standard diagnostic criteria,4 though the proportion requiring intensive care (28%) was consistent with existing research.2 We did not have information on preceding infections, or whether patients had the demyelinating or axonal subtype of Guillain-Barré. We identified risk factors at baseline for all women, but exposures may nonetheless be misclassified, particularly if women who developed disease were never readmitted. Due to small case numbers, we could not study specific immune-mediated or rheumatological disorders separately. There is also the possibility that cases of Guillain-Barré syndrome were over-detected in women with immune-mediated or rheumatological disorders due to more intense medical surveillance, or because these women were more likely to be hospitalized for milder forms of Guillain-Barré. We could not identify women who left the province during the study or died out of hospital, although the proportion is unlikely to differ between women with and without risk factors. The results are likely generalizable to parous women in other settings, but we do not know if similar associations would be observed in the general population, including men who may have greater risk of Guillain-Barré syndrome.1 More research is needed to determine if nonparous women with infertility due to severe immune-mediated or rheumatological diseases have even higher risk of Guillain-Barré syndrome. Conclusions In this study of more than 1 million parous women, medical conditions with known immune pathophysiology were associated with the subsequent risk of Guillain-Barré syndrome over 25 years of follow-up. In addition, transfusion and preeclampsia, conditions with a suspected immune component, both increased the risk of Guillain-Barré. The pathophysiology of Guillain-Barré syndrome is poorly understood, and the present focus on triggers such as infection and vaccination is unlikely to lead to significant advances in prevention and management without better knowledge of predisposing factors. Risk factors such as transfusion are potentially avoidable, and a closer understanding of pathways present in diseases with shared immune pathophysiology may provide therapeutic opportunities in the future. Guillain-Barré syndrome is rare, but morbidity and mortality are elevated and new research angles are needed to better characterize risk. Supplementary Data Supplementary data are available at IJE online. Funding This work was supported by the Canadian Institutes of Health Research (MOP-130452); and the Fonds de recherche du Québec-Santé (career grants awarded to Auger 34695 and Quach 26873). Conflict of interest: None. References 1 Yuki N, Hartung H-P. Guillain–Barré syndrome. N Engl J Med  2012; 366: 2294– 304. Google Scholar CrossRef Search ADS PubMed  2 Willison HJ, Jacobs BC, Doorn PA van. Guillain-Barré syndrome. Lancet  2016; 388: 717– 27. Google Scholar CrossRef Search ADS PubMed  3 Cao-Lormeau V-M, Blake A, Mons S, et al.   Guillain-Barré syndrome outbreak associated with Zika virus infection in French Polynesia: a case-control study. Lancet  2016; 387: 1531– 39. Google Scholar CrossRef Search ADS PubMed  4 Goodfellow JA, Willison HJ. Guillain-Barre syndrome: a century of progress. Nat Rev Neurol  2016; 12: 723– 31. Google Scholar CrossRef Search ADS PubMed  5 Yang B, Lian Y, Liu Y, Wu B-Y, Duan R-S. A retrospective analysis of possible triggers of Guillain–Barre syndrome. J Neuroimmunol  2016; 293: 17– 21. Google Scholar CrossRef Search ADS PubMed  6 Gutierrez-Arcelus M, Rich SS, Raychaudhuri S. Autoimmune diseases—connecting risk alleles with molecular traits of the immune system. Nat Rev Genet  2016; 17: 160– 74. Google Scholar CrossRef Search ADS PubMed  7 Davidson A, Diamond B. Autoimmune diseases. N Engl J Med  2001; 345: 340– 50. Google Scholar CrossRef Search ADS PubMed  8 Blum S, McCombe PA. Genetics of Guillain-Barré syndrome (GBS) and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP): current knowledge and future directions. J Peripher Nerv Syst  2014; 19: 88– 103. Google Scholar CrossRef Search ADS PubMed  9 Compston A, Coles A. Multiple sclerosis. Lancet  2008; 372: 1502– 17. Google Scholar CrossRef Search ADS PubMed  10 Lisnevskaia L, Murphy G, Isenberg D. Systemic lupus erythematosus. Lancet  2014; 384: 1878– 88. Google Scholar CrossRef Search ADS PubMed  11 Xin N, Namaka MP, Dou C, Zhang Y. Exploring the role of interleukin-22 in neurological and autoimmune disorders. Int Immunopharmacol  2015; 28: 1076– 83. Google Scholar CrossRef Search ADS PubMed  12 Auger N, Fraser W, Schnitzer M, Leduc L, Healy-Profitós J, Paradis G. Recurrent preeclampsia and subsequent cardiovascular risk. Heart  2017; 103: 235– 43. Google Scholar CrossRef Search ADS PubMed  13 Ministry of Health and Social Services. Cadre Normatif du Système Med-Écho  – Maintenance et exploitation des données pour l’étude de la clientèle hospitalière (Med-Echo System Normative Framework – Maintenance and use of data for the study of hospital clientele.). Quebec: Government of Quebec, 2016. 14 Hauck LJ, White C, Feasby TE, Zochodne DW, Svenson LW, Hill MD. Incidence of Guillain–Barré syndrome in Alberta, Canada: an administrative data study. J Neurol Neurosurg Psychiatry  2008; 79: 318– 20. Google Scholar CrossRef Search ADS PubMed  15 Berg B van, den, Walgaard C, Drenthen J, Fokke C, Jacobs BC, Doorn PA van. Guillain-Barre syndrome: pathogenesis, diagnosis, treatment and prognosis. Nat Rev Neurol  2014; 10: 469– 82. Google Scholar CrossRef Search ADS PubMed  16 Marrack P, Kappler J, Kotzin BL. Autoimmune disease: why and where it occurs. Nat Med  2001; 7: 899– 905. Google Scholar CrossRef Search ADS PubMed  17 Bianchi M, Vaglio S, Pupella S, et al.   Leucoreduction of blood components: an effective way to increase blood safety? Blood Transfus  2016; 14: 214– 27. Google Scholar PubMed  18 Cheng S-B, Sharma S. Preeclampsia and health risks later in life: an immunological link. Semin Immunopathol  2016; 38: 699– 708. Google Scholar CrossRef Search ADS PubMed  19 Chaiworapongsa T, Chaemsaithong P, Yeo L, Romero R. Pre-eclampsia part 1: current understanding of its pathophysiology. Nat Rev Nephrol  2014; 10: 466– 80. Google Scholar CrossRef Search ADS PubMed  20 Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet  2008; 371: 75– 84. Google Scholar CrossRef Search ADS PubMed  21 Howe CJ, Cole SR, Westreich DJ, Greenland S, Napravnik S, Eron JJ. Splines for trend analysis and continuous confounder control. Epidemiology  2011; 22: 874– 75. Google Scholar CrossRef Search ADS PubMed  22 Lau B, Cole S, Gange S. Competing risk regression models for epidemiologic data. Am J Epidemiol  2009; 170: 244– 56. Google Scholar CrossRef Search ADS PubMed  23 George B, Seals S, Aban I. Survival analysis and regression models. J Nucl Cardiol  2014; 21: 686– 94. Google Scholar CrossRef Search ADS PubMed  24 Langer-Gould A, Albers KB, Eeden SVD, Nelson LM. Autoimmune diseases prior to the diagnosis of multiple sclerosis: a population-based case-control study. Mult Scler  2010; 16: 855– 61. Google Scholar CrossRef Search ADS PubMed  25 Shin IS, Baer AN, Kwon HJ, Papadopoulos EJ, Siegel JN. Guillain-Barré and Miller Fisher syndromes occurring with tumor necrosis factor alpha antagonist therapy. Arthritis Rheum  2006; 54: 1429– 34. Google Scholar CrossRef Search ADS PubMed  26 Raghavan M, Marik PE. Anemia, allogenic blood transfusion, and immunomodulation in the critically ill. Chest  2005; 127: 295– 307. Google Scholar CrossRef Search ADS PubMed  27 Lannan KL, Sahler J, Spinelli SL, Phipps RP, Blumberg N. Transfusion immunomodulation – the case for leukoreduced and (perhaps) washed transfusions. Blood Cells Mol Dis  2013; 50: 61– 68. Google Scholar CrossRef Search ADS PubMed  28 LaMarca B, Cornelius DC, Harmon AC, et al.   Identifying immune mechanisms mediating the hypertension during preeclampsia. Am J Physiol Regul Integr Comp Physiol  2016; 311: R1– 9. Google Scholar CrossRef Search ADS PubMed  29 Jørgensen KT, Pedersen BV, Jacobsen S, Biggar RJ, Frisch M. National cohort study of reproductive risk factors for rheumatoid arthritis in Denmark: a role for hyperemesis, gestational hypertension and pre-eclampsia? Ann Rheum Dis  2010; 69: 358– 63. Google Scholar CrossRef Search ADS PubMed  © The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Epidemiology Oxford University Press

Early predictors of Guillain-Barré syndrome in the life course of women

Loading next page...
 
/lp/ou_press/early-predictors-of-guillain-barr-syndrome-in-the-life-course-of-women-KE19DICz5m
Publisher
Oxford University Press
Copyright
© The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association
ISSN
0300-5771
eISSN
1464-3685
D.O.I.
10.1093/ije/dyx181
Publisher site
See Article on Publisher Site

Abstract

Abstract Background We sought to determine if immune disorders early in life were associated with the later risk of Guillain-Barré syndrome, a neurological disorder thought to be infection-related. Methods We conducted a longitudinal cohort study with 16 108 819 person-years of follow-up for a population of 1 108 541 parous women in Quebec, Canada (1989–2014). The outcome was Guillain-Barré syndrome. We identified women with potential risk factors for future Guillain-Barré syndrome, including immune-mediated and rheumatological diseases, cancer, transfusion, surgical procedures and pregnancy-specific disorders. We estimated hazard ratios (HR) and 95% confidence intervals (CI) for the association of risk factors with later onset of Guillain-Barré syndrome, adjusted for personal characteristics of women. Results The overall incidence of Guillain-Barré syndrome was 1.42 per 100 000 person-years. Incidence was higher for women with immune-mediated (8.79 per 100 000 person-years) and rheumatological disorders (9.84 per 100 000 person-years), transfusion (4.41 per 100 000 person-years), and preeclampsia (2.62 per 100 000 person-years). Immune-mediated disorders were associated with six times the risk of Guillain-Barré syndrome (HR 6.57, 95% CI 3.58 to 12.04), rheumatological disorders with seven times the risk (HR 7.23, 95% CI 3.21 to 16.28), transfusion three times the risk (HR 3.58, 95% CI 1.83 to 6.98) and preeclampsia two times the risk (HR 2.01, 95% CI 1.29 to 3.12). Women with other potential risk factors did not have an increased risk of Guillain-Barré syndrome. Conclusions Immune-related conditions that occur early in life are associated with an increased risk of Guillain-Barré syndrome. The pathophysiology of Guillain-Barré syndrome may extend beyond infectious triggers. Autoimmune disease, autoimmune neuropathy, blood transfusion, Guillain-Barré syndrome, preeclampsia, rheumatic diseases Key Messages Pre-existing risk factors for Guillain-Barré syndrome are poorly understood. Using a cohort of 1 million parous women, we examined the association of immune problems early in life, with later risk of Guillain-Barré. Immune-mediated and rheumatological disorders were associated with seven times the risk of Guillain-Barré syndrome after 25 years of follow-up. Blood transfusion and preeclampsia were associated with two to three times the risk of Guillain-Barré. Immune conditions early in life can influence the future risk of Guillain-Barré syndrome beyond infectious triggers. Introduction A growing number of studies indicate that infectious agents are important triggers for Guillain-Barré syndrome,1–3 but there is a paucity of information on other predisposing risk factors. Guillain-Barré syndrome is the most common cause of acute flaccid paralysis worldwide, with an overall incidence of 1–2 cases per 100 000 individuals annually,1,2,4 and the Zika outbreak has recently drawn more attention to this disorder.3 Most patients with Guillain-Barré syndrome make a full recovery, but a proportion suffer permanent disability.2,4 Guillain-Barré syndrome is an autoimmune disease thought to have both environmental and genetic triggers.1 Previous research has focused primarily on environmental exposures such as infectious agents, vaccination or surgical procedures,1–3,5 but the role of preexisting immune disorders as risk factors for Guillain-Barré syndrome has received less attention.2,4 Immune disorders occur in 5% of the population and are more common in women than men.6,7 Although the pathophysiology of immunological diseases varies widely, some of the aetiological pathways recently identified in Guillain-Barré syndrome are present in a range of diseases with immunological features.6,8 Molecular mimicry is critical to the pathogenesis of Guillain-Barré, but is also implicated in multiple sclerosis and type I diabetes mellitus.7 Guillain-Barré syndrome shares genetic properties and characteristics of B cell dysfunction with diseases such as systemic lupus erythematosus and multiple sclerosis.8–10 Biomarker studies show that immune cytokines such as interleukin-22, which is elevated in patients with Guillain-Barré, also play a role in the pathogenesis of myasthenia gravis and rheumatoid arthritis.11 Yet no study has investigated whether immune-related diseases are related to the risk of Guillain-Barré syndrome. Our objective therefore was to test whether medical disorders with immunological features were associated with a greater risk of later onset of Guillain-Barré syndrome. Methods Study population We used a population-based cohort comprising 1 108 541 women who delivered in the province of Quebec, Canada, between 1989 and 2013. The decision to use a cohort of pregnant women was 2-fold. First, the cohort limits the potential for Berkson’s selection bias at cohort inception because hospital delivery in Quebec is independent of disease status. Nearly all women in Quebec deliver in hospital,12 enabling us to construct a representative sample of all parous women in the province. Second, risk of immunological disease is generally higher in women than men,6 allowing us to focus on the population most pertinent to the research question. Women were eligible to enter the cohort if they delivered an infant in any hospital in Quebec between 1989 and 2013, and did not have Guillain-Barré syndrome recorded in their medical history. We followed the women over time for subsequent hospitalizations, with follow-up extending to 31 March 2014. Follow-up ended with hospitalization for Guillain-Barré syndrome (event), death (competing event) or the end of the study (censored). We obtained data on the clinical characteristics of women from the Maintenance and Use of Data for the Study of Hospital Clientele registry, a dataset containing discharge summaries for all hospitalizations in Quebec.13 Discharge summaries are compiled by trained hospital personnel and are validated with rigorous algorithms.13 Each summary includes information on up to 26 diagnoses and 20 procedures, and indicates if admission to an intensive care unit or death occurred during hospitalization. Diagnoses are coded with the 9th and 10th revisions of the International Classification of Diseases (ICD). Procedures are coded with the Canadian Classification of Diagnostic, Therapeutic, and Surgical procedures, and beginning in 2006, the Canadian Classification of Health Interventions. Guillain-Barré syndrome We identified women hospitalized for Guillain-Barré syndrome using the ICD codes 357.0 and G61.0 marked on the hospital discharge summary.14 To corroborate the diagnosis of Guillain-Barré, we determined the proportion of women who were intubated or admitted to an intensive care unit. Approximately 20–30% of individuals with Guillain-Barré syndrome require artificial ventilation and monitoring in an intensive care setting.3 Risk factors We focused on exposures that were common, as Guillain-Barré syndrome has a relatively low incidence and the number of cases was insufficient in women with rare risk factors. We classified potential predictors of Guillain-Barré into two general categories, medical and pregnancy-specific (see Supplementary Table S1, available as Supplementary data at IJE online). We included disorders that had immune and non-immune features in both categories. Medical disorders included immune-mediated diseases (autoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease),6,7,16 rheumatological diseases (systemic lupus erythematosus, rheumatoid arthritis, scleroderma, ankylosing spondylitis),6,7,16 malignant cancer and endometriosis. Rheumatological diseases have an immune component,7 but cancer and endometriosis are less directly linked with immunological pathophysiology. We identified if women had any of these disorders at baseline or during follow-up, and considered them time-fixed as they reflect a phenotype present throughout life. We identified women who received a transfusion any time during follow-up. Evidence is mounting that transfusion has effects on the immune system through immunization and immunomodulation.17 Because previous surgery has been linked to Guillain-Barré,5 we evaluated surgical procedures that commonly occur at delivery or during follow-up, including caesarian section, hysterectomy and cholecystectomy. We analysed transfusion and surgery as time-varying covariates. Pregnancy-specific risk factors included preeclampsia, placental disorders (placenta abruption, accreta, praevia, infarction or retention), gestational diabetes mellitus, preterm birth before 37 weeks of gestation, intrauterine growth restriction, postpartum haemorrhage and multiple pregnancy. Preeclampsia is a hypertensive disorder of pregnancy with a suspected immune component.18,19 Immunological processes are also suspected in preterm birth.20 The remaining pregnancy-specific exposures are less directly linked with immune pathophysiology, and were used as comparison groups. These exposures were analysed as time-fixed covariates. Covariates We considered baseline age (continuous, 12 to 56 years), year of cohort entry (continuous, 1989–2013), socioeconomic deprivation (poorest fifth of the population, no, unknown)12 and total lifetime parity (1, 2, ≥3 deliveries) as potential confounders. To improve adjustment, we transformed age and year into quadratic splines with knots at the 5th, 50th and 95th percentiles.21 Accounting for age as accurately as possible is important, as the incidence of Guillain-Barré syndrome increases with age.2 Data analysis In primary analyses, we calculated the annual incidence of Guillain-Barré syndrome per 100 000 person-years, as well as the cumulative incidence after 25 years of follow-up with death as a competing event.22 We estimated hazard ratios (HR) and 95% confidence intervals (CI) for the risk of Guillain-Barré syndrome using Cox proportional hazards regression, comparing women with and without each risk factor. We used the number of days since cohort entry as the time scale, and censored women who died or never developed Guillain-Barré syndrome by the last day of the study. We analysed exposures as fixed covariates, except for transfusion and surgery which were time-varying. We adjusted all models for age and year at cohort entry, socioeconomic status and total parity. We ensured that the proportional hazards assumption was not violated using log(‐log survival) plots. In secondary analyses, we used Weibull accelerated failure time models to estimate the adjusted median number of years needed for the exposed and unexposed groups to attain a cumulative incidence of 20 cases of Guillain-Barré per 100 000 women.23 In sensitivity analyses, we tested models in which we excluded women who entered the cohort after 2005, to ensure that we had a minimum of 10 years of follow-up for all women. We did so because women who entered the cohort towards the end of the study had significantly shorter follow-up. In addition, use of anti-tumour necrosis factor (TNF) therapies for rheumatological diseases was more likely after 2005. Finally, we examined regression models in which we did not adjust for parity. Women with high fertility may be healthier and less likely to have immunological diseases, making parity a potential mediator rather than confounder. We performed the analysis in SAS v9.3 (SAS Institute Inc., Cary, NC). As the data were de-identified and the study conformed to ethical requirements for research involving humans in Canada, the University of Montreal Hospital Centre’s institutional review board determined that ethical review was not necessary. Results There was a total of 228 cases of Guillain-Barré syndrome during 16 108 819 person-years of follow-up (Table 1). The overall incidence rate of Guillain-Barré in the cohort was 1.42 per 100 000 person-years (95% CI 1.24 to 1.61). Slightly more than a quarter of women with Guillain-Barré syndrome required intensive care unit admission or intubation (28.1%, N = 64) and three died (1.3%). Among medical risk factors, women with rheumatological disease had the highest incidence of Guillain-Barré at 9.84 cases per 100 000 person-years, followed by women with immune-mediated disorders at 8.79 cases per 100 000 person-years. Women who received transfusions had an incidence of Guillain-Barré syndrome of 4.41 per 100 000 person-years. Among pregnancy-specific risk factors, incidence rates were not elevated for any disorder except preeclampsia (2.62 per 100 000 person-years). Table 1 Incidence of Guillain-Barré syndrome according to characteristics of women   No. women  No. person- years  No. Guillain- Barré syndrome  Incidence per 100 000 person-years (95% confidence interval)  Medical disorders   Immune-mediateda  7 632  125 175  11  8.79 (4.87 to 15.87)   Rheumatologicalb  3 578  60 988  6  9.84 (4.42 to 21.90)   Cancer  29 236  544 555  10  1.84 (0.99 to 3.41)   Endometriosis  37 123  737 420  6  0.81 (0.37 to 1.81)  Medical procedures   Transfusion  23 054  204 092  9  4.41 (2.29 to 8.48)   Caesarian section  272 754  3 597 379  51  1.42 (1.08 to 1.87)   Hysterectomy  50 694  515 507  8  1.55 (0.78 to 3.10)   Cholecystectomy  69 299  713 573  11  1.54 (0.85 to 2.78)  Pregnancy-specific   Preeclampsia  64 350  840 607  22  2.62 (1.72 to 3.97)   Placental disordersc  93 442  1 399 364  13  0.93 (0.54 to 1.60)   Gestational diabetes mellitus  73 710  999 006  15  1.50 (0.90 to 2.49)   Preterm birth, <37 weeks  120 281  1 769 983  24  1.36 (0.91 to 2.02)   Intrauterine growth restriction  58 421  816 232  10  1.23 (0.66 to 2.28)   Postpartum haemorrhage  89 909  1 252 527  13  1.04 (0.60 to 1.79)   Multiple pregnancy  25 424  362 029  5  1.38 (0.57 to 3.32)  Covariates   Age at baseline, years            <25  309 235  4 570 103  68  1.49 (1.17 to 1.89)    25–34  686 935  10 031 560  144  1.44 (1.22 to 1.69)    ≥35  112 371  1 507 156  16  1.06 (0.65 to 1.73)   Total parity            1  501 746  6 820 960  120  1.76 (1.47 to 2.10)    2  444 399  6 590 047  86  1.31 (1.06 to 1.61)    ≥3  162 396  2 697 813  22  0.82 (0.54 to 1.24)   Socioeconomic deprivation            Disadvantaged  208 992  2 919 806  41  1.40 (1.03 to 1.91)    No  809 728  11 381 497  167  1.47 (1.26 to 1.71)   Year at baseline            1989–96  497 311  10 752 794  147  1.37 (1.16 to 1.61)    1997–2004  287 942  3 751 490  47  1.25 (0.94 to 1.67)    2005–12  323 288  1 604 535  34  2.12 (1.51 to 2.97)  Total  1 108 541  16 108 819  228  1.42 (1.24 to 1.61)    No. women  No. person- years  No. Guillain- Barré syndrome  Incidence per 100 000 person-years (95% confidence interval)  Medical disorders   Immune-mediateda  7 632  125 175  11  8.79 (4.87 to 15.87)   Rheumatologicalb  3 578  60 988  6  9.84 (4.42 to 21.90)   Cancer  29 236  544 555  10  1.84 (0.99 to 3.41)   Endometriosis  37 123  737 420  6  0.81 (0.37 to 1.81)  Medical procedures   Transfusion  23 054  204 092  9  4.41 (2.29 to 8.48)   Caesarian section  272 754  3 597 379  51  1.42 (1.08 to 1.87)   Hysterectomy  50 694  515 507  8  1.55 (0.78 to 3.10)   Cholecystectomy  69 299  713 573  11  1.54 (0.85 to 2.78)  Pregnancy-specific   Preeclampsia  64 350  840 607  22  2.62 (1.72 to 3.97)   Placental disordersc  93 442  1 399 364  13  0.93 (0.54 to 1.60)   Gestational diabetes mellitus  73 710  999 006  15  1.50 (0.90 to 2.49)   Preterm birth, <37 weeks  120 281  1 769 983  24  1.36 (0.91 to 2.02)   Intrauterine growth restriction  58 421  816 232  10  1.23 (0.66 to 2.28)   Postpartum haemorrhage  89 909  1 252 527  13  1.04 (0.60 to 1.79)   Multiple pregnancy  25 424  362 029  5  1.38 (0.57 to 3.32)  Covariates   Age at baseline, years            <25  309 235  4 570 103  68  1.49 (1.17 to 1.89)    25–34  686 935  10 031 560  144  1.44 (1.22 to 1.69)    ≥35  112 371  1 507 156  16  1.06 (0.65 to 1.73)   Total parity            1  501 746  6 820 960  120  1.76 (1.47 to 2.10)    2  444 399  6 590 047  86  1.31 (1.06 to 1.61)    ≥3  162 396  2 697 813  22  0.82 (0.54 to 1.24)   Socioeconomic deprivation            Disadvantaged  208 992  2 919 806  41  1.40 (1.03 to 1.91)    No  809 728  11 381 497  167  1.47 (1.26 to 1.71)   Year at baseline            1989–96  497 311  10 752 794  147  1.37 (1.16 to 1.61)    1997–2004  287 942  3 751 490  47  1.25 (0.94 to 1.67)    2005–12  323 288  1 604 535  34  2.12 (1.51 to 2.97)  Total  1 108 541  16 108 819  228  1.42 (1.24 to 1.61)  No., number. aAutoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease. bLupus, rheumatoid arthritis, scleroderma, ankylosing spondylitis. cPlacenta: abruption, accreta, praevia, infarction or retention. Women with immune-mediated disease had the highest cumulative incidence of Guillain-Barré syndrome after 25 years (282.4 per 100 000), followed by women with rheumatological disease (208.1 per 100 000) (Table 2). Compared with no disease, immune-mediated disease was associated with six times the risk of Guillain-Barré syndrome (adjusted HR 6.57, 95% CI 3.58 to 12.04), and rheumatological disease with seven times the risk (adjusted HR 7.23, 95% CI 3.21 to 16.28). Transfusion was associated with three times the risk of Guillain-Barré syndrome (adjusted HR 3.58, 95% CI 1.83 to 6.98), but surgical procedures did not increase the risk. No other medical risk factor was associated with a higher risk of Guillain-Barré syndrome. Among transfused women with Guillain-Barré syndrome, there were no cases of rheumatological disease and only one case of immune-mediated disease. Table 2 Association of medical disorders and procedures with risk of Guillain-Barré syndrome   No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Immune-mediatedb   Yes  11  7 632  282.4 (119.4 to 598.2)  6.56 (3.58 to 12.02)  6.57 (3.58 to 12.04)   No  217  1 100 909  29.4 (25.4 to 33.9)  Ref  Ref  Rheumatologicalc   Yes  6  3 578  208.1 (85.0 to 454.9)  7.25 (3.23 to 16.32)  7.23 (3.21 to 16.28)   No  222  1 104 963  30.9 (26.5 to 35.8)  Ref  Ref  Cancer   Yes  10  29 236  45.8 (23.8 to 83.8)  1.35 (0.72 to 2.55)  1.43 (0.75 to 2.71)   No  218  1 079 305  30.7 (26.3 to 35.7)  Ref  Ref  Endometriosis   Yes  6  37 123  19.2 (8.2 to 41.5)  0.59 (0.26 to 1.32)  0.56 (0.25 to 1.27)   No  222  1 071 418  32.1 (27.5 to 37.3)  Ref  Ref  Transfusion   Yes  9  23 054  96.1 (34.7 to 233.4)  3.35 (1.72 to 6.52)  3.58 (1.83 to 6.98)   No  219  1 085 487  30.4 (26.1 to 35.3)  Ref  Ref  Caesarian section   Yes  51  2 72 754  29.9 (21.4 to 41.2)  1.00 (0.73 to 1.37)  1.04 (0.76 to 1.42)   No  177  835 787  31.9 (26.9 to 37.6)  Ref  Ref  Hysterectomy   Yes  8  50 694  34.5 (14.3 to 76.3)  1.33 (0.65 to 2.74)  1.30 (0.63 to 2.67)   No  220  1 057 847  32.2 (27.5 to 37.4)  Ref  Ref  Cholecystectomy   Yes  11  69 299  41.7 (20.1 to 81.2)  1.21 (0.66 to 2.23)  1.25 (0.68 to 2.30)   No  217  1 039 242  31.7 (27.2 to 36.9)  Ref  Ref    No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Immune-mediatedb   Yes  11  7 632  282.4 (119.4 to 598.2)  6.56 (3.58 to 12.02)  6.57 (3.58 to 12.04)   No  217  1 100 909  29.4 (25.4 to 33.9)  Ref  Ref  Rheumatologicalc   Yes  6  3 578  208.1 (85.0 to 454.9)  7.25 (3.23 to 16.32)  7.23 (3.21 to 16.28)   No  222  1 104 963  30.9 (26.5 to 35.8)  Ref  Ref  Cancer   Yes  10  29 236  45.8 (23.8 to 83.8)  1.35 (0.72 to 2.55)  1.43 (0.75 to 2.71)   No  218  1 079 305  30.7 (26.3 to 35.7)  Ref  Ref  Endometriosis   Yes  6  37 123  19.2 (8.2 to 41.5)  0.59 (0.26 to 1.32)  0.56 (0.25 to 1.27)   No  222  1 071 418  32.1 (27.5 to 37.3)  Ref  Ref  Transfusion   Yes  9  23 054  96.1 (34.7 to 233.4)  3.35 (1.72 to 6.52)  3.58 (1.83 to 6.98)   No  219  1 085 487  30.4 (26.1 to 35.3)  Ref  Ref  Caesarian section   Yes  51  2 72 754  29.9 (21.4 to 41.2)  1.00 (0.73 to 1.37)  1.04 (0.76 to 1.42)   No  177  835 787  31.9 (26.9 to 37.6)  Ref  Ref  Hysterectomy   Yes  8  50 694  34.5 (14.3 to 76.3)  1.33 (0.65 to 2.74)  1.30 (0.63 to 2.67)   No  220  1 057 847  32.2 (27.5 to 37.4)  Ref  Ref  Cholecystectomy   Yes  11  69 299  41.7 (20.1 to 81.2)  1.21 (0.66 to 2.23)  1.25 (0.68 to 2.30)   No  217  1 039 242  31.7 (27.2 to 36.9)  Ref  Ref  Ref, reference. aAdjusted for age and year in splines, socioeconomic status and total parity. bAutoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease. cLupus, rheumatoid arthritis, scleroderma, ankylosing spondylitis. Among pregnancy-specific disorders, preeclampsia was associated with the highest cumulative incidence of Guillain-Barré syndrome (55.4 per 100 000 women), with a 2-fold increase in risk (adjusted HR 2.01, 95% CI 1.29 to 3.12) compared with no preeclampsia (Table 3). No other pregnancy risk factor was associated with a greater risk of Guillain-Barré syndrome. Table 3 Association between pregnancy-specific disorders and risk of Guillain-Barré syndrome   No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Preeclampsia   Yes  22  64 350  55.4 (34.7 to 86.1)  1.91 (1.23 to 2.96)  2.01 (1.29 to 3.12)   No  206  1 044 191  30.2 (25.8 to 35.3)  Ref  Ref  Placental disordersb   Yes  13  93 442  18.0 (10.1 to 31.0)  0.63 (0.36 to 1.11)  0.73 (0.42 to 1.28)   No  215  1 015 099  32.8 (28.1 to 38.2)  Ref  Ref  Gestational diabetes mellitus   Yes  15  73 710  41.2 (23.1 to 70.4)  1.05 (0.62 to 1.78)  1.23 (0.73 to 2.09)   No  213  1 034 831  31.0 (26.6 to 36.1)  Ref  Ref  Preterm birth to <37 weeks   Yes  24  120 281  30.1 (19.4 to 45.9)  0.95 (0.62 to 1.45)  1.04 (0.68 to 1.59)   No  204  988 260  31.7 (27.0 to 37.0)  Ref  Ref  Intrauterine growth restriction   Yes  10  58 421  28.8 (14.5 to 54.1)  0.86 (0.45 to 1.61)  0.91 (0.48 to 1.71)   No  218  1 050 120  31.7 (27.2 to 36.9)  Ref  Ref  Postpartum haemorrhage   Yes  13  89 909  22.7 (12.6 to 39.5)  0.71 (0.41 to 1.25)  0.77 (0.44 to 1.35)   No  215  1 018 632  32.3 (27.6 to 37.5)  Ref  Ref  Multiple pregnancy   Yes  5  25 424  46.6 (15.0 to 125.2)  0.97 (0.40 to 2.35)  1.01 (0.42 to 2.46)   No  223  1 083 117  31.3 (26.9 to 36.3)  Ref  Ref    No. Guillain- Barré  No. women  Cumulative incidence at 25 years per 100 000 women (95% confidence interval)  Hazard ratio (95% confidence interval)   Unadjusted  Adjusteda  Preeclampsia   Yes  22  64 350  55.4 (34.7 to 86.1)  1.91 (1.23 to 2.96)  2.01 (1.29 to 3.12)   No  206  1 044 191  30.2 (25.8 to 35.3)  Ref  Ref  Placental disordersb   Yes  13  93 442  18.0 (10.1 to 31.0)  0.63 (0.36 to 1.11)  0.73 (0.42 to 1.28)   No  215  1 015 099  32.8 (28.1 to 38.2)  Ref  Ref  Gestational diabetes mellitus   Yes  15  73 710  41.2 (23.1 to 70.4)  1.05 (0.62 to 1.78)  1.23 (0.73 to 2.09)   No  213  1 034 831  31.0 (26.6 to 36.1)  Ref  Ref  Preterm birth to <37 weeks   Yes  24  120 281  30.1 (19.4 to 45.9)  0.95 (0.62 to 1.45)  1.04 (0.68 to 1.59)   No  204  988 260  31.7 (27.0 to 37.0)  Ref  Ref  Intrauterine growth restriction   Yes  10  58 421  28.8 (14.5 to 54.1)  0.86 (0.45 to 1.61)  0.91 (0.48 to 1.71)   No  218  1 050 120  31.7 (27.2 to 36.9)  Ref  Ref  Postpartum haemorrhage   Yes  13  89 909  22.7 (12.6 to 39.5)  0.71 (0.41 to 1.25)  0.77 (0.44 to 1.35)   No  215  1 018 632  32.3 (27.6 to 37.5)  Ref  Ref  Multiple pregnancy   Yes  5  25 424  46.6 (15.0 to 125.2)  0.97 (0.40 to 2.35)  1.01 (0.42 to 2.46)   No  223  1 083 117  31.3 (26.9 to 36.3)  Ref  Ref  Ref, reference. aAdjusted for age and year in splines, socioeconomic status, and total parity. bPlacenta: abruption, accreta, praevia, infarction or retention. Analyses of the time taken to reach a cumulative incidence of 20 cases of Guillain-Barré syndrome per 100 000 women indicated that immune-mediated or rheumatological disorders led to faster onset of Guillain-Barré syndrome (Table 4). The median estimated time to reach an incidence of 20 Guillain-Barré cases per 100 000 was less than 2 years for immune-mediated or rheumatological disorders, compared with more than 14 years for women without these disorders. Women who were transfused took around 4 years and women with preeclampsia took around 7 years to reach an incidence of 20 cases of Guillain-Barré per 100 000. Table 4 Number of years to reach a cumulative incidence of Guillain-Barré syndrome of 20 per 100 000 women   Median no. years (interquartile range)a   Risk factor present  Risk factor absent  Medical disorders   Immune-mediatedb  1.7 (1.3 to 2.3)  14.5 (10.5 to 20.0)   Rheumatologicalc  1.5 (1.1 to 2.0)  14.2 (10.3 to 19.6)   Cancer  10.2 (8.1 to 13.9)  14.0 (10.1 to 19.3)   Endometriosis  25.0 (20.1 to 34.8)  13.6 (9.9 to 18.9)  Medical procedures   Transfusion  3.5 (2.6 to 5.0)  14.2 (10.3 to 19.7)   Caesarian section  14.1 (10.4 to 19.5)  13.6 (9.8 to 18.6)   Hysterectomy  13.2 (10.6 to 17.9)  13.4 (9.9 to 18.7)   Cholecystectomy  13.3 (10.2 to 18.7)  13.5 (9.7 to 18.5)  Pregnancy-specific   Preeclampsia  6.8 (4.9 to 9.8)  14.8 (10.7 to 20.3)   Placental disordersd  23.0 (17.2 to 31.9)  13.3 (9.6 to 18.1)   Gestational diabetes mellitus  13.2 (9.6 to 18.6)  14.0 (10.1 to 19.2)   Preterm birth, <37 weeks  14.7 (10.8 to 21.0)  13.8 (9.9 to 18.8)   Intrauterine growth restriction  16.2 (11.7 to 23.3)  13.8 (10.0 to 18.9)   Postpartum haemorrhage  19.6 (14.5 to 28.1)  13.5 (9.8 to 18.5)   Multiple pregnancy  14.2 (10.6 to 20.4)  13.9 (10.0 to 19.1)    Median no. years (interquartile range)a   Risk factor present  Risk factor absent  Medical disorders   Immune-mediatedb  1.7 (1.3 to 2.3)  14.5 (10.5 to 20.0)   Rheumatologicalc  1.5 (1.1 to 2.0)  14.2 (10.3 to 19.6)   Cancer  10.2 (8.1 to 13.9)  14.0 (10.1 to 19.3)   Endometriosis  25.0 (20.1 to 34.8)  13.6 (9.9 to 18.9)  Medical procedures   Transfusion  3.5 (2.6 to 5.0)  14.2 (10.3 to 19.7)   Caesarian section  14.1 (10.4 to 19.5)  13.6 (9.8 to 18.6)   Hysterectomy  13.2 (10.6 to 17.9)  13.4 (9.9 to 18.7)   Cholecystectomy  13.3 (10.2 to 18.7)  13.5 (9.7 to 18.5)  Pregnancy-specific   Preeclampsia  6.8 (4.9 to 9.8)  14.8 (10.7 to 20.3)   Placental disordersd  23.0 (17.2 to 31.9)  13.3 (9.6 to 18.1)   Gestational diabetes mellitus  13.2 (9.6 to 18.6)  14.0 (10.1 to 19.2)   Preterm birth, <37 weeks  14.7 (10.8 to 21.0)  13.8 (9.9 to 18.8)   Intrauterine growth restriction  16.2 (11.7 to 23.3)  13.8 (10.0 to 18.9)   Postpartum haemorrhage  19.6 (14.5 to 28.1)  13.5 (9.8 to 18.5)   Multiple pregnancy  14.2 (10.6 to 20.4)  13.9 (10.0 to 19.1)  aMedian number of years to reach a cumulative incidence of Guillain-Barré syndrome of 20 per 100 000 women, adjusted for age and year in splines, socioeconomic status and total parity. Transfusion and surgery are modelled as time-fixed. bAutoimmune haemolytic anaemia, myasthenia gravis, celiac disease, multiple sclerosis, type 1 diabetes mellitus, Hashimoto's thyroiditis, Graves' disease. cLupus, rheumatoid arthritis, scleroderma, ankylosing spondylitis. dPlacenta: abruption, accreta, praevia, infarction or retention. For immune-mediated and rheumatological disorders, the incidence of Guillain-Barré syndrome increased sharply and almost immediately compared with women who did not have these diseases (Figure 1). This was also the case for transfusion and preeclampsia, though the increase was not as abrupt or pronounced. Figure 1 View largeDownload slide Cumulative incidence of Guillain-Barré syndrome in women with immune-mediated disease, rheumatological disease, transfusion and preeclampsia. Solid line, exposure present; dotted line, exposure absent. Figure 1 View largeDownload slide Cumulative incidence of Guillain-Barré syndrome in women with immune-mediated disease, rheumatological disease, transfusion and preeclampsia. Solid line, exposure present; dotted line, exposure absent. In sensitivity analyses, exclusion of women who entered the cohort after 2005 did not change the pattern of results. Similarly, exclusion of parity from the regression models did not affect the findings. Discussion In this study comprising up to 25 years of follow-up for a cohort of 1.1 million parous women, we found an association between immunological medical conditions and the subsequent risk of Guillain-Barré syndrome. Women with disorders such as multiple sclerosis, autoimmune haemolytic anaemia, rheumatoid arthritis and systemic lupus erythematosus had seven times the risk of Guillain-Barré syndrome compared with women who had no immune-mediated or rheumatological disease. In addition, conditions with a suspected immune component, including transfusion and preeclampsia, were associated with two to three times the risk. Thus, past history of existing or possible immune dysfunction may be a novel risk factor for Guillain-Barré syndrome. Risk factors for Guillain-Barré syndrome may not be limited to infectious triggers, as previously thought. Further research is necessary to obtain insight into the underlying pathophysiological processes that connect immunological diseases and infectious triggers with Guillain-Barré syndrome. Guillain-Barré syndrome is an acute paralytic neuropathy typically preceded by infection, most commonly Campylobacter jejuni.2 Guillain-Barré neuropathy is caused by a dysfunctional autoimmune response targeting peripheral nerves and spinal roots.2 The two variants of Guillain-Barré syndrome, including acute inflammatory demyelinating polyneuropathy and acute motor axonal neuropathy,4 both involve an aberrant immune response.2 In North America, the demyelinating subtype predominates.1,4 The immunological mechanisms are not completely understood, but molecular mimicry is thought to be an important component of Guillain-Barré, based on studies of C. jejuni demonstrating that surface polysaccharides of the microbe resemble glycoconjugates of human nerve tissues.2 Cross-reactive antibodies generated by B cells in patients with C. jejuni subsequently lead to an autoimmune response against peripheral nerves.2 Guillain-Barré is currently thought to be primarily humorally mediated, although a T cell-mediated response cannot be ruled out.2,4 Interleukin-22, part of the interleukin-10 cytokine family which is produced by T helper cells, is elevated in cerebrospinal fluid of Guillain-Barré patients.11 Although autoimmune diseases tend to cluster in individuals and families,6,7 few studies have investigated if immune diseases could be risk factors for Guillain-Barré syndrome. In a descriptive analysis of comorbidity in 496 patients with Guillain-Barré syndrome, immune-mediated or rheumatological diseases were not specifically identified, although diabetes mellitus, anaemia and hypothyroidism were frequent comorbidities.14 A case-control study of 5296 patients in California found that individuals with multiple sclerosis were more likely to develop Guillain-Barré syndrome compared with controls.24 A large body of research on shared immunological pathways implicates molecular mimicry in the pathogenesis of Guillain-Barré syndrome,1,2,4 type 1 diabetes,7 and less concretely, multiple sclerosis.7,24 Altered expression of the cytokine interleukin-22 is also common to multiple sclerosis, systemic lupus erythematosus, rheumatoid arthritis and Guillain-Barré.11 Anti-TNF therapy for rheumatological disease has been associated with numerous case reports of Guillain-Barré syndrome.25 Moreover, individuals with autoimmune or rheumatological disease may be more prone to adverse immune reactions to infectious triggers such as C. jejuni. Infectious triggers may thus be mediators of upstream risk factors that predispose to Guillain-Barré syndrome. The association of transfusion with Guillain-Barré syndrome is impressive in light of growing evidence that transfusion has lasting immune effects with persisting immunomodulation.17 Transfusion-related immunomodulation has been associated with adverse clinical outcomes such as infection, acceleration of cancer growth, transfusion-associated graft versus host disease, and mortality.26,27 Mechanisms responsible for such effects are unclear, but leukocyte dysfunction has been incriminated for many transfusion-related adverse events.17,27 The mechanisms by which transfusion-related immunomodulation could increase the risk of Guillain-Barré are unknown; however, both involve altered leukocyte responses. We found no association between surgery and Guillain-Barré, but the evidence for surgery is based on case studies, not cohort data,5 and involvement of autoimmune pathways after surgery has yet to be shown. Our results therefore provide novel evidence that transfusion has implications for subsequent risk of Guillain-Barré syndrome. Among pregnancy-specific disorders, only preeclampsia was associated with Guillain-Barré syndrome. The pathophysiology of preeclampsia involves reduced placental perfusion due to abnormally shallow trophoblast invasion with inadequate vascular remodelling of the uterine spiral arteries.19 Imbalance in T-cell subtypes and autoantibody production is implicated in this process.18,28 Growing evidence suggests that preeclampsia is associated with an increase in pro-inflammatory T-cells,28 as well as production of autoantibodies against the angiotensin II receptor.19,28 Immune dysregulation and inflammation in preeclampsia are believed to persist past pregnancy.18 Some studies indicate that preeclampsia increases the risk of rheumatological diseases.29 Many of the immunological pathways in preeclampsia have yet to be firmly established, but the association we found with later risk of Guillain-Barré, a disorder with established autoimmune pathways, is compelling. Limitations This study covered a long period of follow-up in a large sample of women, providing sufficient power to detect associations with Guillain-Barré syndrome, a rare neurological disorder. Although the findings are novel and the biological rationale is robust, a number of limitations deserve mention. First, we could not audit charts or validate cases of Guillain-Barré using standard diagnostic criteria,4 though the proportion requiring intensive care (28%) was consistent with existing research.2 We did not have information on preceding infections, or whether patients had the demyelinating or axonal subtype of Guillain-Barré. We identified risk factors at baseline for all women, but exposures may nonetheless be misclassified, particularly if women who developed disease were never readmitted. Due to small case numbers, we could not study specific immune-mediated or rheumatological disorders separately. There is also the possibility that cases of Guillain-Barré syndrome were over-detected in women with immune-mediated or rheumatological disorders due to more intense medical surveillance, or because these women were more likely to be hospitalized for milder forms of Guillain-Barré. We could not identify women who left the province during the study or died out of hospital, although the proportion is unlikely to differ between women with and without risk factors. The results are likely generalizable to parous women in other settings, but we do not know if similar associations would be observed in the general population, including men who may have greater risk of Guillain-Barré syndrome.1 More research is needed to determine if nonparous women with infertility due to severe immune-mediated or rheumatological diseases have even higher risk of Guillain-Barré syndrome. Conclusions In this study of more than 1 million parous women, medical conditions with known immune pathophysiology were associated with the subsequent risk of Guillain-Barré syndrome over 25 years of follow-up. In addition, transfusion and preeclampsia, conditions with a suspected immune component, both increased the risk of Guillain-Barré. The pathophysiology of Guillain-Barré syndrome is poorly understood, and the present focus on triggers such as infection and vaccination is unlikely to lead to significant advances in prevention and management without better knowledge of predisposing factors. Risk factors such as transfusion are potentially avoidable, and a closer understanding of pathways present in diseases with shared immune pathophysiology may provide therapeutic opportunities in the future. Guillain-Barré syndrome is rare, but morbidity and mortality are elevated and new research angles are needed to better characterize risk. Supplementary Data Supplementary data are available at IJE online. Funding This work was supported by the Canadian Institutes of Health Research (MOP-130452); and the Fonds de recherche du Québec-Santé (career grants awarded to Auger 34695 and Quach 26873). Conflict of interest: None. References 1 Yuki N, Hartung H-P. Guillain–Barré syndrome. N Engl J Med  2012; 366: 2294– 304. Google Scholar CrossRef Search ADS PubMed  2 Willison HJ, Jacobs BC, Doorn PA van. Guillain-Barré syndrome. Lancet  2016; 388: 717– 27. Google Scholar CrossRef Search ADS PubMed  3 Cao-Lormeau V-M, Blake A, Mons S, et al.   Guillain-Barré syndrome outbreak associated with Zika virus infection in French Polynesia: a case-control study. Lancet  2016; 387: 1531– 39. Google Scholar CrossRef Search ADS PubMed  4 Goodfellow JA, Willison HJ. Guillain-Barre syndrome: a century of progress. Nat Rev Neurol  2016; 12: 723– 31. Google Scholar CrossRef Search ADS PubMed  5 Yang B, Lian Y, Liu Y, Wu B-Y, Duan R-S. A retrospective analysis of possible triggers of Guillain–Barre syndrome. J Neuroimmunol  2016; 293: 17– 21. Google Scholar CrossRef Search ADS PubMed  6 Gutierrez-Arcelus M, Rich SS, Raychaudhuri S. Autoimmune diseases—connecting risk alleles with molecular traits of the immune system. Nat Rev Genet  2016; 17: 160– 74. Google Scholar CrossRef Search ADS PubMed  7 Davidson A, Diamond B. Autoimmune diseases. N Engl J Med  2001; 345: 340– 50. Google Scholar CrossRef Search ADS PubMed  8 Blum S, McCombe PA. Genetics of Guillain-Barré syndrome (GBS) and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP): current knowledge and future directions. J Peripher Nerv Syst  2014; 19: 88– 103. Google Scholar CrossRef Search ADS PubMed  9 Compston A, Coles A. Multiple sclerosis. Lancet  2008; 372: 1502– 17. Google Scholar CrossRef Search ADS PubMed  10 Lisnevskaia L, Murphy G, Isenberg D. Systemic lupus erythematosus. Lancet  2014; 384: 1878– 88. Google Scholar CrossRef Search ADS PubMed  11 Xin N, Namaka MP, Dou C, Zhang Y. Exploring the role of interleukin-22 in neurological and autoimmune disorders. Int Immunopharmacol  2015; 28: 1076– 83. Google Scholar CrossRef Search ADS PubMed  12 Auger N, Fraser W, Schnitzer M, Leduc L, Healy-Profitós J, Paradis G. Recurrent preeclampsia and subsequent cardiovascular risk. Heart  2017; 103: 235– 43. Google Scholar CrossRef Search ADS PubMed  13 Ministry of Health and Social Services. Cadre Normatif du Système Med-Écho  – Maintenance et exploitation des données pour l’étude de la clientèle hospitalière (Med-Echo System Normative Framework – Maintenance and use of data for the study of hospital clientele.). Quebec: Government of Quebec, 2016. 14 Hauck LJ, White C, Feasby TE, Zochodne DW, Svenson LW, Hill MD. Incidence of Guillain–Barré syndrome in Alberta, Canada: an administrative data study. J Neurol Neurosurg Psychiatry  2008; 79: 318– 20. Google Scholar CrossRef Search ADS PubMed  15 Berg B van, den, Walgaard C, Drenthen J, Fokke C, Jacobs BC, Doorn PA van. Guillain-Barre syndrome: pathogenesis, diagnosis, treatment and prognosis. Nat Rev Neurol  2014; 10: 469– 82. Google Scholar CrossRef Search ADS PubMed  16 Marrack P, Kappler J, Kotzin BL. Autoimmune disease: why and where it occurs. Nat Med  2001; 7: 899– 905. Google Scholar CrossRef Search ADS PubMed  17 Bianchi M, Vaglio S, Pupella S, et al.   Leucoreduction of blood components: an effective way to increase blood safety? Blood Transfus  2016; 14: 214– 27. Google Scholar PubMed  18 Cheng S-B, Sharma S. Preeclampsia and health risks later in life: an immunological link. Semin Immunopathol  2016; 38: 699– 708. Google Scholar CrossRef Search ADS PubMed  19 Chaiworapongsa T, Chaemsaithong P, Yeo L, Romero R. Pre-eclampsia part 1: current understanding of its pathophysiology. Nat Rev Nephrol  2014; 10: 466– 80. Google Scholar CrossRef Search ADS PubMed  20 Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet  2008; 371: 75– 84. Google Scholar CrossRef Search ADS PubMed  21 Howe CJ, Cole SR, Westreich DJ, Greenland S, Napravnik S, Eron JJ. Splines for trend analysis and continuous confounder control. Epidemiology  2011; 22: 874– 75. Google Scholar CrossRef Search ADS PubMed  22 Lau B, Cole S, Gange S. Competing risk regression models for epidemiologic data. Am J Epidemiol  2009; 170: 244– 56. Google Scholar CrossRef Search ADS PubMed  23 George B, Seals S, Aban I. Survival analysis and regression models. J Nucl Cardiol  2014; 21: 686– 94. Google Scholar CrossRef Search ADS PubMed  24 Langer-Gould A, Albers KB, Eeden SVD, Nelson LM. Autoimmune diseases prior to the diagnosis of multiple sclerosis: a population-based case-control study. Mult Scler  2010; 16: 855– 61. Google Scholar CrossRef Search ADS PubMed  25 Shin IS, Baer AN, Kwon HJ, Papadopoulos EJ, Siegel JN. Guillain-Barré and Miller Fisher syndromes occurring with tumor necrosis factor alpha antagonist therapy. Arthritis Rheum  2006; 54: 1429– 34. Google Scholar CrossRef Search ADS PubMed  26 Raghavan M, Marik PE. Anemia, allogenic blood transfusion, and immunomodulation in the critically ill. Chest  2005; 127: 295– 307. Google Scholar CrossRef Search ADS PubMed  27 Lannan KL, Sahler J, Spinelli SL, Phipps RP, Blumberg N. Transfusion immunomodulation – the case for leukoreduced and (perhaps) washed transfusions. Blood Cells Mol Dis  2013; 50: 61– 68. Google Scholar CrossRef Search ADS PubMed  28 LaMarca B, Cornelius DC, Harmon AC, et al.   Identifying immune mechanisms mediating the hypertension during preeclampsia. Am J Physiol Regul Integr Comp Physiol  2016; 311: R1– 9. Google Scholar CrossRef Search ADS PubMed  29 Jørgensen KT, Pedersen BV, Jacobsen S, Biggar RJ, Frisch M. National cohort study of reproductive risk factors for rheumatoid arthritis in Denmark: a role for hyperemesis, gestational hypertension and pre-eclampsia? Ann Rheum Dis  2010; 69: 358– 63. Google Scholar CrossRef Search ADS PubMed  © The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association

Journal

International Journal of EpidemiologyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off