Kawasaki disease and vasculitis associated with immunization
Abstract
The etiology and pathogenesis of Kawasaki disease (KD) remain poorly understood. Among the diverse infectious and environmental factors examined to be triggers for, or be associated with, KD, are immunizations. In this report, we first describe the case of a Caucasian girl who presented with classic KD shortly after receiving her routine 4 months vaccination series. Prior published case reports and large epidemiological studies, which explore potential associations between immunization and KD and other vasculitides, are then comprehensively reviewed. The methodologic challenges that complicate analysis in this area are also considered.
Kawasaki disease (KD) is the leading cause of acquired heart disease, and the second most common form of vasculitis, in children in developed nations.1 Despite this, the etiology and pathogenesis underlying KD remain poorly understood.2 Numerous reports and investigations have considered diverse infectious and environmental factors as triggers for, or being associated with, KD, including: bacterial superantigens, numerous viruses, carpet cleaners, proximity to water, trophosporic wind patterns, and others; nevertheless, none have been definitely established as risk factors for KD. Relatedly, cases of KD after various immunizations have been reported, and large epidemiological studies have investigated potential associations between vaccines and KD. In this report, we describe a Caucasian girl who presented with classic KD shortly after receiving her routine 4 month vaccination series. The available literature exploring potential connections between immunizations and KD and other vasculitides, and the methodologic challenges that complicate analysis in this area, are subsequently reviewed.
Briefly, the present patient was first admitted with fever, edema and erythematous–purplish rash of the extremities, 1 week after receiving diphtheria‐tetanus‐acellular pertussis/inactivated polio virus/hepatitis B (DTaP/IPV/HepB; Pediarix®; GlaxoSmithKline, Rixensart, Belgium), Haemophilus influenzae type B (Hib), Prevnar 13 (Wyeth/Pfizer, Philadelphia, PA, USA), and Rotarix (rotavirus [RV]1; GlaxoSmithKline, Rixensart, Belgium) at age 2 months. At that presentation, she had poor perfusion, elevated leukocytes, platelets, and lactate, and underwent conventional infant rule‐out sepsis management. Cerebrospinal fluid panel, urinalysis, and all cultures were normal/negative; the patient quickly improved with supportive care, was suspected to have viral sepsis, and was discharged after 2 days.
The patient remained well through her routine 4 months well child visit. Twelve hours after receiving vaccines (again, Prevnar 13, DTaP/IPV/HepB, and Rotarix) at that visit, she developed a fever. This persisted until presentation on day of illness (DOI) 4, at which time she had a temperature of 39.4°C, heme‐occult positive diarrhea, and an erythematous macular rash on her torso. By DOI 7 she had dry, cracked lips, intermittent non‐purulent conjunctivitis, and edematous feet with red soles. KD was suspected based on standard (American Heart Association) criteria:2 laboratory data were notable for C‐reactive protein 97.1 mg/L, hemoglobin 10.2 g/dL, and albumin 2.8 g/dL; viral respiratory screen, urine and blood cultures were negative; echocardiogram and abdominal ultrasound were normal. With ongoing flares of body rash and feet swelling, and refractory >39°C fever, she was given i.v. immunoglobulin (IVIG) and aspirin on DOI 9. She defervesced during the IVIG infusion, the swelling resolved the next day, and besides a fading reticular rash, she was at her clinical baseline at discharge on DOI 11. Subsequently, she received all expected 6 months vaccines except Rotarix, and received measles‐mumps‐rubella (MMR) and Varicella vaccines at 12 months – all with no issues – and had remained otherwise well through 15 months of age.
Prior case reports have temporally associated KD with various vaccines, including hepatitis B, yellow fever, pneumococcal polysaccharide vaccine (PPSV)23, influenza, and others (Table 1).3-9 The present case has similarities to a case from China in 2012, in which a 20‐month‐child began with fever the night after receiving her second rotavirus vaccine (Lanzhou lamb rotavirus vaccine; LLV), and had broader clinical and laboratory features diagnostic for KD, including coronary artery dilatation, by day 5 of illness.3 Two other cases in infants ≤2 months – an unusually young age for KD, given that <2% of vasculitis cases are seen in ≤3‐month‐olds – who also had refractory fever ≤1 day after vaccination, similarly were suggestive of immunization as being a trigger for KD.3, 4
| First author, year of publication (Location) | Vaccine(s) | Summary | Reference |
|---|---|---|---|
| Miron D, 2003 (Israel) | Hepatitis B | 35‐day‐old boy began with fever 1 day after second hepatitis vaccine, persisted for 14 days prior to KD diagnosis (4/5 clinical features; platelets 909 000 mm3, various ECG and echo abnormalities, including coronary artery aneurysm); temperature and heart rate normalized ≤12 h after IVIG | 5 |
| Schmöeller D, 2009 (Brazil) | Yellow Fever | 12‐year‐old boy developed fever and rash 20 days after yellow fever vaccination, diagnosed with KD 14 days later with refractory fever, conjunctivitis, cervical lymph node enlargement, and right and left coronary artery dilatation | 6 |
| Oka K, 2012 (Japan) | DPT (others?) | 3‐month‐old boy diagnosed with KD 6 days after first DPT vaccination, and again at 8 months of age 5 days after third DPT vaccination, based on clinical criteria; prompt response to IVIG and aspirin on both occasions. | 7 |
| Ibrahim E, 2014 (Turkey) | DTaP‐IPV‐Hib, PCV13 | 2‐month‐old developed fever 1 day after DTaP‐IPV‐Hib and PCV13, had persistent fever and broad clinical and laboratory features of KD by day 6, including platelets 826 000/mm3 and left and right coronary artery dilatations; defervesced ≤12 h after IVIG | 4 |
| Yin S, 2015 (China) | LLV and hepatitis A | 20‐month‐old girl, fever began <24 h after second dose of LLV and first dose of live attenuated hepatitis A vaccines, diagnosed with KD on day 5; left main and right coronary artery abnormalities | 3 |
| Shimada S, 2015 (Japan) | Influenza | 2‐year‐old girl diagnosed with KD 8 days after her second (and 36 days after her first) influenza vaccination with 5/5 clinical criteria; prompt response to IVIG and aspirin | 8 |
| Kraszewska‐Głomba B, 2016 (Poland) | PPSV 23 | 7‐year‐old boy, with family history of KD (sister at age 3) and 2 prior episodes of KS (both at age 4; no vaccination association raised), with third episode of KS that began 3 days after PPSV23 | 9 |
- DPT, diphtheria‐pertussis‐tetanus; DTaP, diphteria‐tetanus‐acellular pertussis; ECG, electrocardiogram; Hib, Haemophilus influenzae type B; IPV, inactivated polio vaccine; IVIG, i.v. immunoglobulin; KD, Kawasaki disease; KS, Kawasaki syndrome; LLV, lamb Luzhon rotavirus; PCV, pneumococcal conjugate vaccine; PPSV, pneumococcal polysaccharide vaccine.
Notably, KD was found at higher rates in those receiving RotaTeq (RV5; Merck, Whitehouse Station, NJ, USA) in the 42 days after vaccination compared with placebo, in phase 3 of clinical trials: unadjusted RR, 4.9 (5/36,150 vs 1/35,536).10 Given concerns raised by these findings and separate case reports, several post‐licensure studies have tracked both intussusception and KD rates closely after rotavirus vaccines were licensed and introduced into national immunization programs.11-14 In an analysis led by an independent group, an increase in KD was reported to the Vaccine Adverse Events Reporting System (VAERS; co‐managed by the US Centers for Disease Control [CDC] and Food and Drug Administration [FDA]), in the first five quarters of post‐marketing of RV5, with 11 cases being associated with use (or co‐use) of this specific vaccine during this time period.11
Subsequent large studies from around the world, however, have not confirmed any definitive association of KD with rotavirus vaccines.12-16 In a comprehensive analysis of VAERS data from 1990 to 2007 (Table 2), KD symptom onset was not temporally clustered at any specific timepoint after RotaTeq receipt; also, rates of KD reported in association with RotaTeq (even at its highest level after the aforementioned phase 3 findings were included in an FDA label revision) were substantially lower than established background rates of disease.12 In independently published data from a phase 3 study of Rotarix (RV1) in three high‐income Asian countries, KD rates were equivalent in vaccine and placebo recipients.13Similarly, in an industry‐led integrated analysis of 28 randomized clinical trials from various global regions, which included >56 000 Rotarix recipients and analyzed adverse events (AE) ≤31 days after vaccination, there was only one case of KD in each of the vaccine and placebo groups.14 Lastly, after both RotaTeq and Rotarix were introduced into the national immunization programs of Germany and Austria, analyses of billing data, and mandatory reporting of post‐vaccine AE (respectively), found no association of KD with either rotavirus vaccine.15, 16
| First author, year of publication | Study type | Vaccine(s)/Database | Summary | Reference |
|---|---|---|---|---|
| US Food and Drug Administration, 2007 | Phase III RCT | RotaTeq | 5 cases of KD in 36 150 RotaTeq recipients ≤30 days after vaccination, vs 1 of 35 536 in placebo recipients. Crude RR, 4.9 (95%CI: 0.6–239.1) | 10 |
| Geier D, 2008 | Retrospective analysis/passive surveillance of vaccine safety | RotaTeq/VAERS | In first five quarters post‐marketing of the vaccine, 11 cases of KD temporally associated with RotaTeq were reported to VAERS. The majority were in the first 7 months of 2007 (in 2003–2005, prior to RotaTeq licensure, there were 3–6 KD reports to VAERS per year) | 11 |
| Hua W, 2009 | Retrospective analysis/passive surveillance of vaccine safety | RotaTeq/VAERS | Comprehensive analysis, including medical record review, of all KD cases reported to VAERS from 1990 to mid‐October 2007. PPR of KD related to RotaTeq and 21 other FDA‐approved vaccines analyzed. Main results/conclusions: No clustering of symptom onset 1 day after any vaccination; for RotaTeq, no temporal clustering for any timepoint after vaccine receipt. Reporting rate of KD with RotaTeq, even at highest level after revision of FDA label with phase III data (ref. 10), at 2.78/100 000, was substantially lower than the background rate of 9–19/100 000 person‐years in <5‐year‐old children. PPR was also elevated for Pediarix, with rates of 0.37 and 2.44 per 100 000 before and after RotaTeq label revision. | 12 |
| Lau Y, 2013. | Phase III RCT | Rotarix (Singapore, Hong Kong, Taiwan) | In pre‐licensure Rotarix study in 3 high‐income Asian countries, equivalent cases of KD in the vaccine group (2/1513) and blinded placebo (2/1512) | 13 |
| Buyse H, 2014 | Integrated analysis of phase II and III RCT | Rotarix (numerous countries and global regions) | GlaxoSmithKline led integrated analysis of safety and reactogenicity of Rotarix from 28 double blinded RCT. 56 562 in vaccine group, 45 512 in placebo group. Solicited AE at 8 days, and unsolicited AE, severe AE, and death, at ≤31 days, assessed. 1 case of KD in both groups: RR, 1.00 (95%CI: 0.01–78.35; P = 1.00). | 14 |
| Uhlig U, 2014. | Population‐based retrospective analysis | RotaTeq and Rotarix (Germany) | Analysis of hospitalization billing data 2006–2012 in Germany in 0–5‐year‐old children. No increase in KD‐related hospital admissions were found after the introduction of the 2 RV vaccines, nor were there differences in KD rates between high‐ and low‐vaccination coverage areas. | 15 |
| Paulke‐Korinek M, 2013 | Passive surveillance of post‐vaccine AE | RotaTeq and Rotarix (Austria) | Reporting of post‐vaccine AE described as mandatory in Austria (but not actively surveyed nor verified). In 2010–2011, no cases of KD (or with suggestive features) were reported to be associated with either RV vaccine. | 16 |
| Center K, 2009 | Phase 4 observational safety study | PCV7; Kaiser Permanente Northern California (USA) | PCV7 recipients 15 March 2000–30 November 2002 (65 927 infants) were compared to (40 223) historical controls who received first dose of Hib from March 1995 to February 1996. Main results: 42 cases of KD in PCV7 vs 17 in controls, for unadjusted RR of 2.02 (95%CI: 1.16–3.63, P = 0.012); controlling for sex, race, age at first dose, length of follow up, and season, the adjusted RR was 1.67 (95%CI: 0.93–3.00, P = 0.0831). Majority of KD cases occurred after third or fourth PCV7 (n = 36), rather than first (n = 2) or second doses (n = 6), at a median of 107 days after the first dose | 17 |
| Tseng H, 2013. | Prospective observational (passive surveillance) | PCV13/VSD | Utilizing electronic heath records from the VSD Project (data from 8 managed care organizations), KD associated with PCV13‐exposed children 1 month–2 years old from April 2010 to January 2012 (immediately after introduction) were compared to historical controls who received PCV7. Adjusted RR (for age, sex, seasons) of KD was 1.97 for PCV13 recipients (95%CI: 0.79–4.94), when adjusted | 18 |
| Abrams J, 2015. | Retrospective analysis (passive surveillance) | VSD, US Database | Utilizing Poisson regression and case‐crossover analysis of 152 verified cases of KD in approx. 1.7 million children <7 years of age in the VSD database in 1996–2006 (before RotaTeq), KD patients had lower rates of vaccination in the 42 days prior to symptom onset (RR, 0.38; 95%CI: 0.20–0.75) than non‐KD patients. | 20 |
| Hall GC, 2016 | Retrospective analysis | UK electronic health records | Between 2008 and 2012, utilizing UK electronic health records of >275 000 children <6 years of age, there were 6 cases of KD ≤28 days after any vaccination stage, for an overall incidence of 12.8/100 000 person years (95%CI: 5.7–28.4), and an incidence after any single immunization stage ranging from 0 to 27.4 (95%CI: 8.8–84.8) per 100 000 person years. | 21 |
- AE, adverse event; Hib, Haemophilus influenzae type B; KD, Kawasaki disease; PPR, proportional reporting ratio; PCV, pneumococcal conjugate vaccine; RCT, randomized controlled trial; RV, rotavirus; VAERS, Vaccine Adverse Events Reporting System; VSD, Vaccine Safety Datalink.
In part due to concerns raised by the RotaTeq phase 3 experience, Prevnar 7 and subsequently Prevnar 13 have had similar monitoring and investigations for associations with KD. In a post‐licensure analysis from Northern California Kaiser Permanente of >65 000 children, although an unadjusted analysis found a statistically significant increase in KD hospitalization incidence in pneumococcal conjugate vaccine (PCV7; Prevnar 7) recipients compared with historical controls, there was no association after adjusting for potential confounding variables (e.g. sex, race, length of follow up etc.); nevertheless, worth noting from that study is that the overall adjusted RR of 1.61 approached statistical significance (95%CI: 0.93–3.00, P = 0.0831), but also that KD episodes occurred at a median >100 days after the first use of PCV7, and typically after the third or fourth dose.17 A later study assessed AE associated with PCV13 immediately after its introduction, through the Vaccine Safety Datalink Project, and found an adjusted RR of 1.97 ≤ 28 days after PCV13 vaccination compared with historical controls who had received PCV7 (95%CI: 0.79–4.94).18 Given this persistent signal, FDA's PRISM program is maintaining ongoing surveillance for KD associated with PCV13.19
Also, retrospective analyses of various vaccine safety databases (VSD) have examined the association of vaccines, other than RV and PCV, with KD and also with other vasculitis diagnoses.12, 20-22 In the previously cited review of VAERS data from 1990 to 2007, the proportional reporting ratio of KD for Pediarix® was elevated compared with other vaccines – including in the time period before the RotaTeq FDA label revision – but again this (highest, approx. 2.44/100 000) was lower than background KD incidence rates.12 In a review of VSD data (prior to RotaTeq introduction) from 1996 to 2006, verified KD patients were less likely than non‐KD controls to have received any vaccine in the 42 days prior, raising the possibility of a transient protective effect (against KD) of immunization.20 Most recently, an analysis of UK electronic health data of <6‐year‐olds from 2008 to 2012 (during which time PCV but not RV were part of the routine schedule), incidence of KD ≤28 days after any immunization stage was 12.8/100 000 person‐years (95%CI: 5.7–28.4), which fell within the wide range of available age‐related background KD rates.21
There are noteworthy limitations, however, to these studies and to the investigation of vaccine associations in general, including that AE capture in safety databases is mainly by voluntary reporting, which may lead to underestimations; that such reports are symptom rather than diagnosis focused; and that vasculitides lack standard diagnostic definitions.22The lack of a standardized KD case definition has been highlighted by the Brighton KD Study Group, who recently proposed criteria for diagnostic certainty levels (definite, complete and incomplete; probable; and possible) to improve the data collection and analysis of KD surveillance and research.23 Last, epidemiologists have long noted that there are social and medical attributes associated with both the avoidance of vaccinations and the increased risk of AE, and that if these confounders are not adequately controlled for in studies, the association of AE with vaccines may be underestimated.24
The current patient's first presentation after vaccination, with widespread rash and edema, was consistent with vasculitis; her second met the criteria for KD, and the prompt response to IVIG corroborated this diagnosis. The rapidity of symptoms after the second round of vaccinations suggested the possibility of antigen sensitization with her previous exposure and an immune‐mediated phenomenon. It is worth noting that the patient's personal contacts all lacked any illnesses around the time of diagnosis, and there were no other KD cases in the present hospital for 4 weeks before or after. Given an absence of issues after the 6 month immunizations when Rotarix was omitted, we suspect that this vaccine was the precipitating agent; nevertheless, an independent or combined effect with a concomitantly received immunization remains possible. In conclusion, vaccines may potentially be associated with vasculitides, including KD, in individual children, and ongoing, systematic surveillance of such events is warranted, particularly for newer immunizations.
No comments:
Post a Comment