Background. Superinfection is an adverse event following smallpox vaccination. The clinical presentation is similar to that of a large normal vaccine reaction or “robust take,” and the frequency is unknown.
Methods. We retrospectively reviewed all reported severe local reactions consistent with superinfection among United States civilian smallpox vaccinees from January 2003 through January 2004. We applied a standard case definition and estimated the frequency of superinfection following smallpox vaccination.
Results. We identified 48 reported cases for further review among 36,043 smallpox vaccinees. Two (4%) of the 48 reported cases met the case definition for superinfection; neither of the patients had a pathogenic organism isolated from their infection site. Both were treated with antibiotics and resolved their infection. Of the 46 cases determined not to be superinfection, 41 (89%) were temporally consistent with a large normal vaccine reaction. Thirty (75%) of 40 reported case patients for whom data were available received antibiotic therapy.
Conclusions. Superinfection following smallpox vaccination is rare. Most of the reported superinfection cases were probably large normal smallpox vaccine reactions. Educating providers about the normal response to smallpox vaccine may decrease the overdiagnosis of superinfection and the unnecessary use of antimicrobials.
In June 2001, the Advisory Committee on Immunization Practices (ACIP) made recommendations for using smallpox vaccine to protect persons working with orthopoxviruses and to prepare for a possible terrorist attack . On 24 January 2003, following the events of 11 September 2001, the Department of Health and Human Services authorized voluntary smallpox vaccination of health care workers who may be called upon to monitor or treat persons exposed to smallpox, as well as members of smallpox response teams identified by state or local health departments . The military began vaccination in December 2002 as part of the national program of preparedness against biological attack .
The Centers for Disease Control and Prevention (CDC), in coordination with state and local health departments, administered the civilian bioterrorism preparedness program and established surveillance for adverse events following vaccination. The CDC received a number of reports of cellulitis or superinfection following smallpox vaccination administered as part of the program. Previously, bacterial superinfection has rarely been reported following smallpox vaccination .
Superinfection is clinically difficult to distinguish from a large normal vaccination reaction or “robust take.” A robust take is cellulitis caused by the replicating vaccinia virus. The signs and symptoms of a robust take are similar to those of cellulitis caused by bacteria or viruses other than vaccinia virus .
A joint Smallpox Vaccine Safety Work Group of the ACIP and Armed Forces Epidemiological Board, in collaboration with the CDC, received expert opinion from smallpox eradication experts, ophthalmologists, dermatologists, and infectious disease specialists to draft surveillance case definitions for select adverse events following smallpox vaccination. This group agreed on a case definition for superinfection in March 2004 (table 1). We summarize the results of our review of cases of superinfection reported to the CDC from January 2003 through January 2004.
Design, case ascertainment, and source of reports. We retrospectively reviewed all reports of superinfection following smallpox vaccination identified during January 2003 through January 2004. We reviewed all reports of local smallpox vaccine reactions. All reports that included signs and symptoms consistent with superinfection or a local inflammatory reaction (erythema, warmth, swelling, fluctuance, discharge, or pain) at or near the vaccination site were selected for further investigation. Reports came from 3 sources: the Vaccine Adverse Event Reporting System (VAERS), a national passive surveillance system established in 1990 that receives reports of suspected adverse events associated with US-licensed vaccines and is jointly administered by the CDC and the US Food and Drug Administration ; calls received by the CDC Clinician Information Line, a toll-free number staffed 24 h/day with trained nurses who provide information to clinicians and state health department personnel about smallpox vaccine and common adverse events ; and calls received directly by the CDC. Reports originated from vaccine recipients, vaccine providers, or state smallpox vaccination programs. We excluded persons vaccinated through the military vaccination program and referred them to the military smallpox vaccination program. Civilian laboratory workers who received smallpox vaccine but were not part of the civilian smallpox vaccination preparedness and response program were included.
Data collection. We developed a questionnaire to collect information about reported superinfection following smallpox vaccination. A study physician contacted and interviewed the vaccine provider or the person reporting the possible superinfection. The state adverse event coordinator often assisted in data collection and was informed of the outcome of each report. Data collected included demographic information, number of previous smallpox vaccinations, clinical details of the local reaction and associated systemic symptoms, laboratory results, treatments received, and clinical outcome (table 2). We kept all data confidential.
Case definition. We used the case definition (table 1) to classify superinfection reports. The temporal criteria in the case definition play an important role in distinguishing a superinfection from a robust take. A confirmed case of superinfection requires laboratory or radiographic evidence to support the clinical diagnosis. We reviewed each case and either classified it as suspected or confirmed superinfection or discarded it.
Data management and analysis. We entered the data into a database (Access, version 10.0; Microsoft) and completed descriptive analyses. We used the pre-event vaccination system (a registry of smallpox vaccine recipients)  to determine the number of vaccinees for the denominator of rates and the demographic distribution among all vaccinees. An additional 662 vaccinated civilian laboratory workers who were not part of the smallpox vaccination preparedness and response program were included in the analysis.
We identified 48 reports of superinfection among 36,043 civilian vaccine recipients during the study period (rate, 1.3 reports of superinfection per 1000 vaccine recipients). Most vaccinations were administered and most reports of superinfection were received during February–May 2003. A second peak in reported superinfection occurred during October and November, when few vaccinations were being administered (figure 1).
Compared with all vaccinees, patients with reported superinfection were younger (mean age, 37 years vs. 47 years) and more likely to be first-time vaccinees (42% vs. 24%). The sex distribution was similar among those with reported superinfection and all vaccinees (63% vs. 64%, respectively). The onset or peak day of symptoms for the adverse event was distributed around day 8 following vaccination (figure 2). Two people were hospitalized, one for chronic folliculitis unrelated to smallpox vaccination and the other for chest pain. There were no deaths or permanent disability reported.
Bacterial cultures for 5 (13%) of 40 reported cases of superinfection for whom these data were available yielded no pathogenic organisms. Nevertheless, 30 (75%) of the 40 case patients received an oral or parenteral antimicrobial agent.
The rate of reported superinfection was 5 times higher (20 of 8711 vaccine recipients, or 2.3 reports per 1000 vaccine recipients) among first-time vaccinees than among revaccinees (12 of 26,400 vaccine recipients, or 0.45 reports per 1000 vaccine recipients) (P < .001, by the χ2 test). Revaccinees experienced the onset of symptoms earlier (median onset day, 7) than did first-time vaccinees (median onset day, 9) (P < .01, by Wilcoxon rank sum test) (figure 3).
Only 2 reported superinfections (4%) met the case definition for superinfection. Both were classified as suspected cases because supporting laboratory or radiographic data were not available to confirm the clinical diagnosis for either patient. Of the 46 cases that did not meet the case definition, 41 patients (89%) had peak or onset of symptoms 6–12 days following vaccination. The remaining 5 cases (11%) did not meet other diagnostic criteria to be classified as superinfection; 4 did not meet the clinical criteria and 1 was diagnosed as folliculitis, unrelated to the smallpox vaccination.
Case reports for suspected superinfection. On 13 March 2003, a 54-year-old female revaccinee who had a history of successfully treated right-sided breast cancer and right arm melanoma received smallpox vaccination. Because of her medical history, her provider elected to vaccinate her on the left thigh. Four days after her vaccination, she had erythema, induration, and warmth at her vaccination site, with inguinal adenopathy and malaise. She did not have fever or chills. No laboratory or radiological studies were done. She was treated with cephalexin (500 mg q.i.d.) for 5 days, and her symptoms resolved.
On 25 November 2003, a 42-year-old male revaccinee received smallpox vaccination on the left deltoid. Two days after vaccination, he experienced pain and warmth at the vaccination site, with erythema and edema 4–6 cm in diameter. No laboratory or diagnostic studies were done. He was treated with cephalexin (500 mg q.i.d.) for 10 days and completely recovered.
Only 2 (4%) of the 48 cases of reported superinfection met the case definition of suspected superinfection. Our findings confirm previous studies suggesting that superinfection following smallpox vaccination is rare [4, 8]. Most of the cases reported were probably robust takes rather than superinfections. This highlights the difficulty in differentiating robust takes from superinfection caused by pathogenic bacteria or other viruses. We found that diagnostic tests were rarely done, and antimicrobials were frequently prescribed empirically.
In the 1963 US national survey by Neff et al. , only 2 secondary bacterial infections at the vaccination site were reported among 433 smallpox vaccine complications of 14 million vaccinations. Although low, our rate of 55 superinfections per million vaccinees (2 superinfections among 36,043 vaccinees) is substantially higher than the rate reported during 1963 (0.14 superinfections per million vaccinated) . Our rate is also higher than the rate reported by the military: 13 superinfections per million vaccinees (6 cases among >450,000 vaccinees) ; however, the military rate reflects only cases diagnosed as cellulitis and treated with intravenous antibiotics .
Why might we have detected a higher rate of superinfection than was previously reported in the United States? First, there is a greater public awareness of vaccine safety . The smallpox vaccination program, specifically, generated considerable publicity and controversy [10, 11]. Media coverage and educational and screening programs might have heightened the awareness of adverse events and increased reporting. Second, differences between case ascertainment and surveillance might have contributed to the higher rates that we found. The civilian smallpox vaccination program relied on passive reporting of adverse events, as in the past, but there are many improvements in the way adverse events are reported now, compared with the 1960s. The reporting mechanisms for adverse events following immunization in general and smallpox vaccination in particular are much more sophisticated than they were previously. VAERS  accepts reports by mail or online, allowing for simplified reporting of any adverse event following vaccination by anyone, including providers and vaccine recipients. Smallpox vaccine adverse event coordinators in every state must identify and report adverse events following smallpox vaccination. The Clinician Information Line, with a toll-free telephone number, also facilitated adverse event reporting following smallpox vaccination. Combined, these mechanisms resulted in “stimulated” passive reporting for adverse events following smallpox vaccination, which may be more sensitive than historical methods. Third, the age distribution of the vaccinee population in 2003 was older than that of the population in the 1960s. During the 1963 campaign, only 3.8 million (27%) of the 14 million people vaccinated were adults aged ⩾20 . The 2003 program vaccinated only adults, who were largely health care workers and might be more likely to recognize and report adverse events. Finally, an occlusive dressing was used to prevent contact transmission during the 2003 civilian campaign, and these dressings were changed under medical supervision. This practice may have contributed to an increase in superinfections and their recognition . Occlusive dressings were not widely used for smallpox vaccination during the 1960s.
It is difficult to differentiate a superinfection from a robust take, which is vaccinial cellulitis and is clinically indistinguishable from a pathogenic superinfection, without the benefit of a culture. A study examining complications of smallpox vaccination during 1968 did not differentiate severe reactions following vaccination from superinfection . In general, distinguishing viral from bacterial infection, particularly in the case of cellulitis, is clinically difficult. For example, a viral exudative tonsillitis cannot usually be distinguished from a bacterial tonsillitis by means of clinical examination alone. However, there are simple and quick diagnostic tests available for tonsillitis [14, 15]. This is not the case for determining viral versus bacterial cellulitis, because culture samples are difficult to obtain and may be helpful in only 25% of the cases . Expected signs and symptoms following smallpox vaccination, as in most cases of cellulitis, include lymphangitis, regional lymphadenopathy, and considerable edema and erythema in ∼7% of vaccinees [17, 18]. A robust take typically occurs 8–12 days after vaccination, improves within 72 h after peak symptoms, and does not progress clinically. It usually has an area of redness of >7.5 cm in diameter with swelling, warmth, and pain at the vaccine site [5, 19]. Robust takes have been reported in up to 16% of first-time vaccinees in a small sample . Systemic symptoms of fever, malaise, and myalgia are also part of the normal response to smallpox vaccination [16–18]. In the absence of diagnostic studies, the temporal onset of symptoms might be helpful in distinguishing a normal robust take from a superinfection. The case definition gives clinical and temporal criteria for the diagnosis of suspected superinfection. These are most helpful in classifying cases as less likely to be superinfections if the symptoms peak during the 6- to 12-day interval after vaccination. We are not aware of any data with which to calculate the sensitivity or specificity of these clinical or temporal criteria.
Adverse reactions to smallpox vaccination are more common among first-time vaccinees than others [4, 8, 18, 20]. For this reason, the ACIP and the Healthcare Infection Control Practices Advisory Committee recommended that first-stage vaccination for bioterrorism preparedness should preferentially include previously vaccinated health care personnel . Although 3 times as many revaccinees as first-time vaccinees participated in the preparedness program, we found that twice as many possible superinfections were reported among first-time vaccinees, a rate 5 times that among the revaccinees. Among our cases of reported possible superinfection, reaction to smallpox vaccination occurred earlier in revaccinees than in first-time vaccinees. These observations are consistent with robust takes and support our conclusion that most of the cases reported were probably robust takes rather than superinfections.
The difficulty in differentiating superinfection from a normal smallpox vaccine reaction makes our findings subject to misclassification bias. Another limitation may be reporting bias. Some sites may have overreported robust takes. Three cases reported from a single location during October 2003 were robust takes but were reported as adverse events because of the “intensity” of the local reaction. Conversely, passive surveillance can also result in underreporting of true cases of superinfection.
Our review was retrospective and therefore did not allow for “real-time” evaluation of reported cases, so recall bias might be introduced. This can be problematic, because the case definition relies heavily on temporal criteria; a case with onset of symptoms 5 days after vaccination would be included, whereas a similar case with an onset 6 days after vaccination would be discarded.
None of our cases were confirmed by isolation of a pathogen. Nevertheless, most patients were treated with antibiotics. Educating providers about the clinical spectrum of reactions can decrease the number of unnecessary antibiotics prescribed. We recommend careful and serial observation, or “watchful waiting,” as an alternative to antimicrobial therapy for patients with symptoms consistent with cellulitis, especially if the onset of symptoms is between day 6 and day 12 after vaccination. Supportive wound care and symptomatic treatment should be sufficient .
This study confirms that superinfection following smallpox vaccination is rare. Educating health care providers about the normal response to smallpox vaccination could decrease the overdiagnosis of superinfection and the unnecessary use of antimicrobials. A bioterrorist attack with smallpox (variola virus), resulting in a national smallpox vaccination campaign, would have a minor burden of superinfection.
We gratefully acknowledge the case patients, health care providers, state and local health department officials, VAERS reporters, State Adverse Event Coordinators, and CDC staff assigned to the Smallpox Vaccine Adverse Event Monitoring and Response Activity, for their assistance in identification, evaluation, report, and follow-up of possible cases of superinfection. For assistance in the development of the superinfection case definition and case classification, we thank Gina Mootrey, Katrin S. Kohl, and Louisa Chapman (CDC); the joint ACIP—Armed Forces Epidemiological Board smallpox vaccine safety working group; and especially John Neff (University of Washington School of Medicine and Children's Hospital and Regional Medical Center, Seattle) and Timothy G. Berger and Toby Maurer (University of California, San Francisco). For their contributions to data analysis we appreciate the efforts of James Baggs and Roseanne English (CDC).
Potential conflict of interest. All authors: No conflict.