Anaphylaxis following quadrivalent human papillomavirus vaccination

Methods

Clinical evaluation

Following an initial review, we contacted patients suspected to have had anaphylaxis, or their guardians, for written consent to conduct a telephone interview and to access their medical records (ambulance and hospital). Trained staff used a standardized, piloted questionnaire developed for this study for telephone interviews with the patient, the patient's guardian and the immunization provider who witnessed or managed the reaction. The patient and guardian questionnaire included demographic information, past medical history (especially of atopic disease and allergies), presenting symptoms, treatment and outcome, and immunization history. The provider questionnaire collected details of the preimmunization assessment of the patient, history of latex exposure, symptoms reported at the time of the event, clinical signs, treatment and overall clinical impression.

We invited all patients with suspected anaphylaxis to attend The Children's Hospital at Westmead for allergy testing at least 6 weeks after the episode. The allergy testing protocol included skin-prick testing for a panel of common antigens (cat, rye grass pollen, alternaria, house dust mite), vaccine components (baker's yeast [Saccharomyces cerevisiae], polysorbate 80) and the HPV vaccines Gardasil and Cervarix (GlaxoSmithKline). We performed intradermal testing (0.02 mL per injection) with HPV vaccines and polysorbate 80 (dilutions 1:100 and 1:10). The vaccine adjuvant (proprietary Merck amorphous aluminum hydroxyphosphate sulfate) was not available for inclusion in the protocol.

Following investigation, we classified cases using the Brighton case definition of anaphylaxis and the level of diagnostic certainty2 (Box 1, Box 2). The Brighton definition refers to the level of diagnostic certainty but not the degree of severity of the case (i.e., a level 1 classification is associated with the greatest diagnostic certainty of anaphylaxis). For the purpose of applying the definition, we considered a rash that occurred at any site not contiguous with the injection site to be “generalized.”

• Download figure
• Open in new tab
• Download powerpoint

Box 1.

• Download figure
• Open in new tab
• Download powerpoint

Box 2.

We based our causality assessment on the timing of the onset of symptoms, according to the World Health Organization classification (i.e., certain/very likely, probable, possible, unlikely, unrelated),4 and whether or not an alternative trigger was present.

Results

In the 2007 school-based program, there were 269 680 doses of HPV vaccine administered. This consisted of dose 1 (n = 95 006, 83% coverage), dose 2 (n = 91 289, 80% coverage) and dose 3 (n = 83 845, 74% coverage). Of 155 adverse events reported after HPV vaccination, 110 were for vaccinations in the school program (41 per 100 000 doses). In the general practice program, there were 347 568 doses distributed, but the number of doses administered is not known. We initially classified 12 cases as requiring further investigation for suspected anaphylaxis. However, 1 patient did not provide consent for a telephone interview or medical record review, and thus was not further classified. Among the 12 cases, there was no geographic clustering or vaccine batch number implicated.

We reviewed all data for the 11 patients with suspected anaphylaxis, and we classified 8 (73%) as fulfilling the Brighton case definition (Table 1). Of these, 1 case was level 1, and 7 were level 2. Because the onset of symptoms occurred within 30 minutes of vaccination and because there was no alternate explanation, we determined that these cases were certain or very likely to have been caused by the vaccine. We considered and rejected the possibility that these cases could have been vasovagal episodes or somatic conversion disorder misdiagnosed as anaphylaxis. Of the 8 cases, 6 had generalized urticaria, 1 had angioedema and 1 had no dermatological signs but had stridor and hypoxia. Of the 3 cases not classified as anaphylaxis, 2 were classified as generalized allergic reactions on the basis of failure to progress, and 1 was considered unlikely to be a case of anaphylaxis because of the absence of objective signs and safe receipt of a subsequent dose. No alternate diagnosis could be suggested based on the available evidence. This person had been receiving high-dose steroids for Crohn disease at the time of the reaction.

Of the 8 confirmed cases, 7 were reported from the school-based program. The median age was 16.5 (min–max 15–25) years. Six cases were reported after dose 1 of the vaccine, and 2 cases were reported after dose 2. Neither of the latter 2 patients had reported a reaction to their first dose of the HPV vaccine. Four patients had a history of allergic reactions and atopic disease (1 had asthma, 3 had eczema or urticaria, 3 had allergic rhinitis). All had been fully vaccinated according to the recommendations for their age group, and anaphylaxis had not developed with previous vaccinations.

Four cases of anaphylaxis (one level 1 case, three level 2 cases) received allergy testing. Of these, 2 had a positive reaction to skin-prick testing for house dust mites. All had negative reactions to skin-prick testing for baker's yeast, and to skin-prick testing and intradermal testing for Gardasil, Cervarix and polysorbate 80.

Based on the 7 confirmed cases of anaphylaxis from the school-based program (269 680 doses administered), the estimated rate of anaphylaxis following HPV vaccination was 2.6 per 100 000 doses administered (95% CI 1.0–5.3 per 100 000). Five cases occurred after dose 1 of the vaccine (95 006 doses administered), giving a rate of anaphylaxis of 5.3 per 100 000 people vaccinated (95% CI 1.7–12.3). Two cases occurred after the second dose (91 289 doses administered), giving a rate of anaphylaxis of 2.2 per 100 000 people vaccinated (95 % CI 0.3–7.9).

In comparison, the 2003 school-based meningococcal C vaccination program had 1 identified case of anaphylaxis (anaphylaxis rate 0.1 per 100 000 doses administered, 95% CI 0.003–0.7). A 13-year-old boy, who was coadministered the hepatitis B vaccine, immediately developed tachycardia, diaphoresis, chest pain, an itchy blocked throat, dysphagia, vomiting, agitation and a blotchy rash on his neck. He required 2 doses of adrenaline. On chart review performed for this study, 2 additional cases (girls aged 11 and 12 years) would have been assessed for suspected anaphylaxis in the current investigation, and they may have met the Brighton definition for anaphylaxis if full clinical information were available. One girl had complained of tongue swelling, throat tightness, and chest and abdominal pain 5 minutes after vaccination, and the other reported difficulty breathing and a rash on her neck and leg and at the injection site 1.5 hours after vaccination. If we include these cases, the anaphylaxis rate is 0.4 cases per 100 000 doses (95% CI 0.1–1.1).

Rates of anaphylaxis reported from other immunization programs are given in Table 2 and Appendix 1 (www.cmaj.ca/cgi/content/full/179/6/525/DC2).

Interpretation

The anaphylaxis rate in the 2007 school-based HPV vaccination program was 2.6 per 100 000 doses (95% CI 1.0–5.3 per 100 000 doses). This is higher than reported for other vaccines (Table 2). The rate of anaphylaxis after HPV vaccination is comparable only to the rate following administration of vaccines containing bovine gelatin in Japan (Table 2), which led to removal of the gelatin from the vaccine.17 The lower limit of the confidence interval for the rate of anaphylaxis in the New South Wales HPV vaccination program overlaps with the upper limit of the confidence interval for the Canadian hepatitis B vaccination program and UK measles–rubella vaccination program (Table 2). Although the point estimate for the rate of anaphylaxis is higher than in previous programs, the estimated rate is still an order of magnitude less common than the World Health Organization categorization of adverse events after immunization which are “very rare” (< 1 in 10 000).4

Unpublished safety data from the HPV vaccine manufacturer indicate that no vaccine safety signal for allergy or anaphylaxis was identified in the vaccine trials.18 However, it is well recognized that prelicensure trials are underpowered to detect anaphylaxis.2 Hypersensitivity reactions to HPV vaccine were the fifth most common adverse event reported to the Vaccine Adverse Events Reporting System in 2007,19 and skin rash within 48 hours accounted for 26% of the reports of adverse events after HPV vaccination in New South Wales in 2007.

Reasons for the increased rate of anaphylaxis after HPV vaccine in the New South Wales school-based program are not clear but could be due to a number of factors including possible allergenicity of components of the HPV vaccine, enhanced passive adverse event surveillance and use of the Brighton definition of anaphylaxis. Another potential factor is that, from midadolescence onward, women have higher rates of anaphylaxis than men.20^,21 An English study found that the risk of hospital admission for anaphylaxis is considerably higher among women of child-bearing age than in the rest of the population (rate ratio 1.38 95% CI 1.27–1.50).21 The incidence of anaphylaxis in Australia also appears to be increasing.20 A recent analysis by Siegrist and colleagues22 anticipated the identification of vaccine safety signals for allergic and autoimmune diseases following population implementation of HPV vaccination because medical presentations for allergy-related conditions in young women are common. It is therefore important when assessing causality to note that all of our cases occurred within minutes of vaccination.

The school-based program in New South Wales might be expected to have high sensitivity for adverse events occurring soon after vaccination because experienced nursing teams provide both vaccine and postvaccination observation. However, the same arrangements were in place for the recent school-based vaccination program delivering meningococcal C vaccines to a similar age group in this region and no such signal was detected. Present definitions of anaphylaxis reflect the understanding that it has a variable spectrum of severity and presenting symptoms. This is an important issue when comparing rates with previous published reports, which are likely to be restricted to the most severe cases only. However, when we applied the definitions used in other recent studies of anaphylaxis following immunization to our cases, they all satisfied the case definition for probable anaphylaxis.11^,13^,23 We found the Brighton classification useful, but we considered it inappropriate for screening because routine surveillance reports of adverse events following vaccination usually lack the necessary level of clinical detail.

Pre-existing allergen sensitization is biologically plausible for several components of the quadrivalent HPV vaccine. Natural HPV infection does not cause viremia; however, HPV virus-like particles are highly immunogenic when injected.24 It is unclear whether some people may produce an IgE response to injected HPV antigens. One small trial injected participants (n = 30) with HPV type 11 virus-like particles and found no increase in IgE levels;25 however, data from a study of rhesus macaques (n = 10) suggests it is possible.26 Because the age group targeted by the HPV vaccination program was 15 years and older, many women may have already been exposed to HPV, because the median age of first sexual intercourse in Australia is 16 years.27 There is potential for residual amounts of yeast proteins to be in the HPV vaccine. Although hypersensitivity to yeast is a documented phenomenon,28 there have been infrequent reports of anaphylaxis or hypersensitivity reactions following the receipt of yeast-containing vaccines.29^–32

Quadrivalent HPV vaccine contains polysorbate 80 as a stabilizer.33 Polysorbates are commonly used as solubilizers, stabilizers and emulsifiers in cosmetics and medical preparations. Polysorbate 80 could be a potential trigger of nonallergic (“anaphylactoid”) anaphylaxis. There have been a number of reports of anaphylaxis or generalized hypersensitivity reactions occurring following the receipt of drugs that have implicated polysorbate 80 as the cause.34^–37 Aluminum adjuvants are not a documented cause of anaphylaxis.38

The most common event that may be confused with anaphylaxis during an acute episode is syncope. Syncope has also been reported following HPV vaccination,19 but its onset is typically more rapid, with a person becoming unconscious in seconds.39 The absence of urticaria, pruritus, angioedema or upper respiratory obstruction are helpful in differentiating syncope from anaphylaxis, as are the presence of pallor instead of flushing and nausea without abdominal pain.40

All of the people with anaphylaxis in the New South Wales HPV vaccination program recovered completely. Most did so rapidly after administration of adrenaline and were asymptomatic by the time they reached the emergency department (20–60 minutes later). In such cases, primary care staff may need to rely on witness and patient history to obtain a complete clinical picture.

In July 2007, the package insert for the quadrivalent HPV vaccine was updated to reflect reports of anaphylaxis following HPV vaccination, and the product information has been updated internationally by the manufacturer. Following confirmation of the vaccine safety signal, all general practitioners in New South Wales were reminded via fax to be prepared for the occurrence of anaphylaxis following HPV vaccination and to have appropriate protocols and equipment available. Consent forms used in the school-based program include the small but real risk of anaphylactic reaction. In December 2007, the Therapeutic Goods Administration placed advice on its website alerting providers to the occurrence of anaphylaxis following HPV vaccination.41 In 2008, the New South Wales school-based HPV vaccination program is targeted to girls aged 12–15 years. As of May 2008, 181 235 doses of HPV vaccine have been administered and 78 adverse events following vaccination have been reported. Twenty-two cases of “allergic reaction” have been reported, but no reports of anaphylaxis have been received.