Learning to Appreciate Our Differences. Part 2
These findings deserve attention, in part because of the use of a second population for independent validation, and in part because the SNP associations are biologically plausible: the products of these genes are involved in host defense against poxviruses. Although these results already suggest further experiments to understand the mechanisms that underlie these associations, and although they already suggest potential clinical applications, it may be useful to step back and consider a variety of different paths forward.
Our genome sequence serves as a blueprint for the “system” and contains a wealth of information about vulnerabilities, strengths, and potential, most of which still remains beyond our current ability to interpret. However, the technology and methodology used in genomewide SNP assessments have matured quickly, offering an opportunity for a more agnostic, “unsupervised” approach, rather than promoting reliance on prior suspicions or assumptions with the use of targeted assays. This technology and methodology includes next‐generation sequencing platforms, mass spectroscopy, allele‐specific polymerase chain reaction, single‐nucleotide primer extension, oligonucleotide ligation techniques, high‐density oligonucleotide microarrays, and combinations of the aforementioned approaches. So‐called genomewide association (GWA) studies are increasingly common in the biomedical literature, and they have revealed previously unsuspected links between genetic loci and disease. These studies will soon provide new clues about individuals who have an elevated risk for vaccine‐associated AEs. However, GWA studies also create important new needs, including well‐characterized host populations, large numbers of cases and controls, methods for distinguishing between true‐positive and false‐positive associations, and approaches for untangling polygenic traits or epistatic effects (gene‐gene interactions). As useful as GWA studies may be for revealing host vulnerabilities and risks, other approaches will also have an important place in this area of investigation. This is because our genome is dynamic: genetic and epigenetic structural modifications, changing levels of expression, and other aspects of regulatory control suggest approaches that will add to the value of primary sequence data.
Profiles of genomic response based on mRNA abundance patterns and, more recently, on abundance patterns of microRNAs have provided novel diagnostic and prognostic information about patients with cancer and, in a fewer number of studies, patients with acute inflammatory disorders and infection. Although these genomic patterns may be used to classify patients on the basis of clinical outcome and identify hosts in whom a protective immune response has or will occur, these patterns display significant temporal and spatial dynamic properties, especially during the acute phase after exposure or perturbation. As a result, it may be necessary to measure responses from specific anatomical compartments and/or at specific time points, to draw clinically useful conclusions. More‐traditional approaches for profiling human susceptibilities to infection and immunologic AEs have focused on functional aspects of the immune system, such as the lymphocyte activation state, epitope‐specific lymphocytes, secreted cytokine levels, and antibody reactivity profiles. Although useful, these approaches restrict our attention to a narrow subset of human physiological responses. Because our history of prior and current microbial exposures plays a significant role in determining how we respond to a new encounter, it is possible that profiling of the human indigenous microbiota will contribute to a more effective risk assessment for vaccine‐ and pathogen‐associated adverse outcomes.
As the complexity and dimensionality of host genetic, genomic, and immune response profiles expand, so will the challenges of validating putative predictors, diagnostics, and biomarkers and understanding the mechanisms behind these profiles. The solutions will include large, prospective, replicated cohorts; standardized specimen collection with clinical metadata; reconsideration of criteria for assessing causal relationships; and focused experimental investigation. Reif et al. highlight several of these needs. Importantly, the results of these efforts will promote public health and strengthen strategies for prevention of infectious diseases.
Learning to Appreciate Our Differences
Despite overall levels of genetic similarity and shared physiological characteristics across the species, humans vary quite a bit in their individual responses to biological stress and noxious stimuli. Sometimes variant responses are dramatic and clinically important. This is certainly the case with respect to the responses of humans to microbial pathogens and their antigens. The consequences of infection by a pathogen in a host population vary from benign to catastrophic, as a complex function of the host genetics, the history of prior exposures and experiences, the current state of immune activation (both local and systemic), and, probably, a variety of other factors, including host nutritional status and the composition and structure of the indigenous microbiota. Although typical responses to vaccines are favorable, occasional responses are pathologic and costly to the host. If these aberrant responses could be predicted and understood mechanistically (the first element does not necessarily require the second), we would be able to reduce the number and/or severity of vaccine adverse events (AEs), improve vaccine design, and create personalized strategies for eliciting immune protection. In this issue of the Journal, Reif et al. address this goal and provide an opportunity to discuss challenges and possible solutions.
Reif et al. collected host genetic sequence data from 2 independent studies of the smallpox vaccine (Aventis Pasteur) in vaccinia virus–naive adults. Of 85 vaccinated subjects included in the first study, 16 developed a systemic AE (fever, lymphadenopathy, or generalized rash). Twenty‐four of the 46 vaccinated subjects included in the second study developed 1 of these 3 systemic AEs.
The investigators obtained data on 1442 single‐nucleotide polymorphisms (SNPs) located in or near 386 genes, as drawn from the National Cancer Institute Cancer Genome Anatomy Project SNP500Cancer Database, which includes genes or specific genetic variants associated with signaling pathways, immune response, and oncogenesis. These SNPs were assayed using a highly parallel genotyping technology based on allele‐specific primer extension, ligation, amplification, and hybridization to bead‐based oligonucleotide arrays (Illumina). Only the 36 SNPs (linked to 26 genes) that were found to have an AE‐associated P value .05 in the first study were assessed in the second study. Of these 36 SNPs, 3 were found to have an AE‐associated P value .05. One SNP is located in the 5,10‐methylenetetrahydrofolate reductase (MTHFR) gene, and 2 SNPs are located in the interferon regulatory factor–1 (IRF1) gene. In addition, 3 SNPs in the interleukin‐4 (IL4) gene were significantly associated with AEs in the first study but did not quite achieve statistical significance in the second study.
Herd Immunity
Thus, the risk‐based approach has not resulted in control of epidemic influenza, nor has it been effective in preventing serious morbidity and mortality.
Herd immunity (protection) is defined as vaccinating one group to reduce the exposure of another. More precisely, it describes a type of immunity that occurs when vaccination of a part of the population (the herd) provides protection to unimmunized individuals. The theory is that, for diseases passed from person to person, it is more difficult to maintain the infection when large numbers are immune. The more individuals who are immune, the lower the chance that a susceptible person will come into contact with an infectious person. Herd immunity is not achieved when an effective vaccine is not available or when vaccines are rejected by a segment of the population. Epidemics of pertussis, measles, and mumps and a resurgence of cases of Haemophilus influenza type b infection have appeared in regions in which immunization has been refused.
Herd immunity has been demonstrated for viral diseases (eg, rubella, measles, and mumps), as well as for diseases caused by bacteria (eg, pertussis, H. influenza type b, and Streptococcus pneumoniae). In the case of rubella, widespread immunization in the United States has resulted in a reduction of the frequency of congenital rubella syndrome from 823 cases per year to 1. By comparison, in the >50% of countries in which rubella immunization is not routinely administered, >100,000 cases of congenital rubella syndrome occur.
Considerable evidence indicates that herd immunity is operative in the control of influenza as well. In Tecumseh, Michigan, >85% of 3159 schoolchildren were given TIV over 4 days and compared to a similar population in the neighboring community of Adrian, where vaccine was not administered. Three times more influenza‐like illness occurred among people of all ages in Adrian than in Tecumseh, demonstrating that immunizing schoolchildren in a community significantly protects the population at large in that community. In 1962, Japanese authorities mandated that all schoolchildren 5–15 years old receive TIV. That practice rapidly and significantly decreased the number of excess deaths attributable to pneumonia and influenza, predominantly among elderly persons. It was estimated that up to 49,000 deaths were prevented annually and that 1 death was prevented by immunizing 420 schoolchildren. In 1986, parents were allowed to refuse vaccination, and the excess death rate rose. In the Temple‐Belton area of Texas, schoolchildren in 2 counties were immunized with LAIV and compared with unimmunized children in 3 counties for the incidence of MAARI among adults >35 years old; even with a vaccine uptake rate of only 20%–25%, indirect protection of 8%–18% of the adults studied occurred. In Russia, Rudenko et al gave either LAIV or TIV to 35%–65% of schoolchildren 7–14 years old. Respiratory illness rates among adults on staff and unimmunized children were inversely related to coverage rates in the immunized children, suggesting that herd immunity occurred. In a day care setting in San Diego, 149 children 24–60 months old and their families received either TIV or hepatitis A vaccine; despite the <45% efficacy of the seasonal TIV used, there was an 80% reduction in the frequency of febrile respiratory illness and a 72% decrease in absenteeism for the school‐age contacts. In another Russian study, 1 dose of TIV was given to school‐age children 3–17 years old (coverage, 57%–72%) in one community (immunized) and none was given to another (comparison). Both communities contained >400,000 people. In the comparison community, the rate of influenza‐like illness in adults >60 years old was 3.4 times higher. In the Maryland SchoolMist Study, LAIV was administered to 40% of a targeted elementary school, and 2 other nonimmunized schools served as controls. Significantly fewer child and adult medical visits for episodes of febrile respiratory illness, fewer over‐the‐counter medications purchased, and fewer days of absenteeism were noted among both children and adult members of families in which children were immunized. In 2000, Ontario, Canada, initiated a universal influenza immunization program for individuals >6 months old. Other provinces used targeted vaccination, and Ontario’s vaccination rate increased 20%, compared with 11% in other provinces. Mortality rates for influenza and influenza‐associated health care use decreased significantly more in Ontario than in the other provinces. Finally, Weycker et al, using a stochastic epidemic model of influenza transmission, clinical illness, and economic costs, estimated the benefits of routinely vaccinating children 6 months to 18 years old against influenza. They predicted that immunizing 20% of this population would result in a reduction in the number of influenza cases by 46% and that 80% coverage would reduce the number of cases by 91%. Similar concurrent reductions were estimated to occur for influenza‐related mortality and economic costs. Thus, significant evidence exists for the operative workings of herd immunity for influenza.
A Paradigm for the Control of Influenza
In this issue of the Journal, Glezen et al have produced the latest in a series of descriptions of an innovative approach to the control of influenza. Dr Glezen and colleagues at Baylor College of Medicine have long been advocates of the concept that the epidemiology of influenza is the most important element in the control of this disease.
The study by Glezen and colleagues investigates the effectiveness of immunizing 47.5% of elementary‐school children 5–11 years old in 25 public schools and 3 parochial schools in which 1 dose of live attenuated influenza vaccine (LAIV) was administered (intervention group), relative to a comparable community in which vaccine (LAIV or trivalent inactivated vaccine [TIV]) was administered in an off‐protocol manner. A significant degree of herd protection (reduction in the number of medically attended acute respiratory illness [MAARI]) was seen in all age groups except 12–17‐year‐olds, with risk ratios for all age groups >18 years old being lower even than that for the target group of 5–11‐year‐olds. This occurred despite there being an excess of persons >75 years old in the intervention community. Furthermore, LAIV was 1.7 times more effective in the prevention of proven influenza virus infection in the target group than among those who received TIV and was 6 times more effective than among those who did not receive vaccine.
Preschool and school‐age children are the major disseminators of influenza. They have the highest age‐specific attack rates for influenza, are less likely to observe cough and sneeze precautions, and are in close proximity to each other and family members. Furthermore, they excrete influenza A virus longer before becoming ill (6 days vs 1 day) and after illness appears (14 days vs 4.5 days), compared with adults. They are centrifugal spreaders to family members, other children, and individuals in the community, including those at high risk.
Serious morbidity from influenza is increasing, and the negative effect on human life and the economy ($87 billion per year) is considerable. Previous efforts to control influenza have concentrated on immunizing populations—for example, elderly persons, those medically at risk, and young children. This approach has fallen far short of the targets determined by the Healthy People 2010 initiative. One of the at‐risk groups that has been demonstrated to be at particularly high risk of H1N1/09 infection—pregnant women—has been underimmunized historically. In one study conducted during the 2007–2008 influenza season, only 24% of pregnant women were immunized, the most common reason given being “MD did not mention”. Historically, the group at highest risk is individuals >65 years old, who frequently exhibit immune senescence. Simonsen et al reported no significant effect of influenza vaccine on seasonal mortality among elderly persons. Furthermore, despite an increase in the vaccination rate among individuals >65 years old between 1989 and 1997, mortality and hospitalization rates continued to increase, suggesting that administering vaccine to this group is not very effective. This notion is supported by the work of Jackson et al, who observed that those who take the vaccine are in better health and are more mobile; therefore, they are more able to receive the vaccine than their sicker counterparts. This has led to an overestimation of the effectiveness of influenza vaccine in this highest‐risk age group and may, in part, account for the observation of increasing morbidity and mortality among elderly persons in the face of increasing vaccine use.
Rescue of Severely Immunocompromised HIV‐Positive Persons. Part 8
As demonstrated in this and a previous outbreak investigation, environmental sampling is useful for confirming a varicella case or outbreak, particularly in situations in which lesions are no longer present or clinical specimens would be difficult to collect and environmental specimen collection and testing are feasible. We were able to detect VZV DNA in environmental samples from case patients’ bedrooms and belongings and found that it could still be detected in the environment several months after rash onset. Other outbreak investigations have also found that VZV may possibly be spread through airborne transmission, and VZV has been detected from throat and air filter samples of herpes zoster and varicella case patients in a hospital setting. Collection of airborne particles using aerosol samplers, which has been used for detection of influenza and respiratory syncytial virus, may be a potential method for providing additional information on airborne transmission of VZV. Detection of VZV DNA in the environment should be interpreted with care, because it could result from viral shedding that occurred remotely and because detection of VZV DNA in the environment may not necessarily indicate the presence of viable infectious virus. Because clinical specimens are still the optimal method for confirmation of a varicella case and limitations in interpreting results, we do not recommend environment specimens as the routine source for laboratory testing.
This outbreak demonstrated that adults who have lived in residential settings for most of their lives are potentially susceptible to varicella disease. It is important for residential facilities to screen all current and potential residents and staff for varicella immunity prior to admission or employment and to vaccinate those who are susceptible to help prevent disease in this setting. Susceptible residents or staff who are not screened prior to admission or employment should be vaccinated within 5 days of exposure to VZV, although vaccination is recommended even after this period because vaccination will provide protection for future exposures. Staff in these facilities should remain alert for herpes zoster, as well as varicella, and implement appropriate infection control measures in a timely fashion to prevent VZV transmission to susceptible residents and staff. Laboratory testing plays an important role in determining susceptibility to varicella in adults; it can be used to confirm diagnoses in an outbreak so that adequate control measures can be implemented, and it can be used to identify vaccine adverse events.
Rescue of Severely Immunocompromised HIV‐Positive Persons. Part 7
In this outbreak, it is likely that a single unrecognized herpes zoster case resulted in 3 generations of disease transmission in the facility and community. Herpes zoster typically occurs in older populations, although zoster can occur in younger persons with an estimated annual rate of 1–2/1000 persons for 20–40‐year‐olds. To prevent VZV transmission from herpes zoster cases, contact precautions should be followed. In this outbreak, proper infection control measures were not implemented because the herpes zoster case was retrospectively identified after the rash had resolved. It is important for staff in residential facilities to consider herpes zoster as a diagnosis for unilateral rashes and implement control measures as appropriate to prevent VZV transmission from these cases. This recommendation is also important for school settings and other residential facilities, such as long‐term care facilities, prisons, hospitals, army barracks, and shelters, in which there is a higher risk of exposure if VZV is introduced in this type of setting due to the constant close contact of students or residents.
Healthcare providers in residential facilities should be screened for immunity to varicella and other vaccine‐preventable diseases prior to employment. Birth before 1980 should not be considered evidence of immunity to varicella for US‐born healthcare staff, since it is important that they have confirmed immunity to varicella. For healthcare staff who are not born in the United States, it is also important that they are screened for immunity to varicella regardless of when they were born because the epidemiology of varicella may differ in other countries. Ensuring immunity among healthcare providers ensures protection for both them and the residents they care for, who may not have immunity to these diseases. For other staff in residential facilities, a requirement for evidence of immunity to varicella and other vaccine‐preventable diseases can be considered depending on their level of contact with residents and the prevalence of contraindications preventing vaccination of susceptible residents.
Laboratory testing of clinical and environmental samples are important tools for investigation and control of outbreaks. Varicella and herpes zoster can easily be mistaken for other rash illnesses. Pain, a characteristic commonly associated with herpes zoster, may be less prevalent in younger adults or difficult to ascertain in persons who are nonverbal. PCR of skin lesion specimens is the preferred method for laboratory confirmation of varicella cases while skin lesions are still present. Serology testing requires invasive blood collection procedures and is less sensitive for establishing a diagnosis. Laboratory testing also plays a critical role in identifying vaccine‐associated adverse events. Through genotyping, we were able to identify a resident with a vaccine‐associated rash during this outbreak.
Rescue of Severely Immunocompromised HIV‐Positive Persons. Part 6
This varicella outbreak in a facility for adults with intellectual and physical disabilities highlights several important aspects of varicella prevention and control. Residents affected by this outbreak had been living in residential settings for most of their lives. This presumably resulted in greater social isolation and fewer opportunities for exposure to varicella in childhood. Although current guidelines from the Advisory Committee on Immunization Practices state that birth before 1980 is evidence of varicella immunity for the general population, this may not be predictive of immunity for individuals who have lived in residential facilities since childhood. As a result of the outbreak, the Connecticut Department of Developmental Services amended its varicella guidelines for similar institutional settings to no longer accept birth before 1980 as evidence of varicella immunity for residents. Because of challenges in obtaining complete medical and vaccination histories and collecting serum specimens, it can be difficult to determine varicella immunity among residents of adult residential facilities. Thus, for residents for whom it is difficult to document a history of varicella disease or vaccination and challenging to obtain a serology specimen to assess immunity, the most efficient approach may be to screen all current and potential residents for evidence of immunity to varicella and to vaccinate susceptible individuals, even those born before 1980, prior to admission to the facility.
Varicella outbreaks have been described in other residential facilities for adults, including long‐term care facilities, hospitals, and prisons, but few have been described in residential settings for people with intellectual and physical disabilities. Other than young age at admission to a residential facility, we did not find any individual level risk factors for varicella among residents in this outbreak. Because most adults have naturally acquired immunity to varicella, varicella disease in adult settings typically does not spread extensively. Although the overall attack rate in this outbreak was high (16%) compared with that for other reported outbreaks among adults (0.2%–3.6%), disease presentation was not particularly severe. Adults, however, often have more severe disease with increased rates of mortality when they develop varicella. A varicella outbreak among adults with learning disabilities, the majority of whom have lived most of their lives in a residential facility in the Netherlands, resulted in a varicella‐related death.
Varicella is highly infectious; secondary attack rates in susceptible household contacts might reach 90%. Recommended control measures, such as airborne respiratory isolation measures or isolation of case patients to their own room, can be extremely difficult and expensive for residential facilities to implement. Case patients in this outbreak could not be effectively isolated alone in their rooms because they required 24‐h supervision for their personal safety. Due to their level of disability, residents did not have the capacity to follow basic infection control practices.
Rescue of Severely Immunocompromised HIV‐Positive Persons. Part 5
Outbreak control measures. Case patients were kept in their bedrooms or on their residential floor until their lesions scabbed over, although bedroom doors remained open at all times to allow staff to monitor residents. Non‐case residents living in apartments with case patients were not allowed to leave their apartments for 14 days following rash onset in the last case. Due to the limited number of rooms, facility A did not isolate cases from their non–case roommates. There were 2 rooms where both roommates became cases, although in both instances, rash onset dates were within 3 days of each other, indicating that transmission did not occur between roommates. Facility A recommended that all staff (healthcare and non‐healthcare) check their varicella immunity status with their private healthcare provider and undergo vaccination if susceptible; however, evidence of immunity was not required to continue working.
Facility A vaccinated 55 (93%) of 59 residents who had not developed a varicella‐like rash from 30 December 2008 through 6 January 2009 because information on history of varicella disease was incomplete on medical records and could not be accurately obtained from guardians. In addition, it was challenging to obtain a serology specimen from all residents to assess susceptibility. A second dose of varicella vaccine was given 28 days later to 50 of these residents. On 17 January 2009, a maculopapular vesicular rash with Canadian pharmacy no rx
Environmental testing. Of the 71 environmental samples collected in January 2009, we detected wild‐type (WT) VZV DNA from 9 (82%) of the 11 case patients’ beds and belongings, from the bedroom floor and wheelchair of the possible herpes zoster case patient, and from the common areas of building 1. No VZV DNA was detected in samples collected from building 2, where there were no cases. An additional 25 environmental specimens were collected in March 2009: VZV DNA remained detectable in samples from the bedrooms of 6 (75%) of 8 varicella case patients and from the common area of building 1. In addition, Oka‐ (vaccine) strain VZV was detected in environmental samples collected in January and March 2009 from the bedroom of the resident with vaccine‐associated rash.
Rescue of Severely Immunocompromised HIV‐Positive Persons. Part 4
Case investigation. From 4 December 2008 through 7 January 2009, 11 of the 70 residents of facility A had varicella rash onset, for an overall attack rate (AR) of 15.7%. Varicella diagnoses were laboratory confirmed for 3 case patients (2 who were positive for varicella IgM and 1 with VZV DNA detected in a skin lesion). Case patients ranged in age from 32 to 49 years (median, 39 years). All case patients resided in buildings 1 and 3 (8 cases in building 1 and 3 cases in building 3; AR in buildings 1 and 3 were 33% and 13%, respectively) and attended 1 of 2 off‐site day programs using 1 of 2 transportation vehicles. In addition to day programs, residents interacted with each other at occasional facility‐wide social events. Five case patients had 50 lesions, and 1 had >500 lesions. The median duration of rash was 9 days, and fever was present in 90% of case patients. Complications did not develop in any case patients, but 1 patient died 10 days after rash onset of causes unrelated to varicella. Two case patients had a documented history of varicella disease.
Varicella was identified among 2 facility A staff (a nurse and a caregiver, aged 28 and 33 years, respectively). Both developed a rash with >250 lesions on 22 December 2008. Neither staff member reported a history of varicella vaccination or disease; one was born outside the United States. One additional case was identified in a severely disabled nonfacility participant of one of the day programs attended by 7 of the 11 infected residents. This case patient had no history of varicella vaccination or disease and developed a rash with 250–500 lesions on 18 December 2008.
During the medical records review conducted in January, we identified a possible case of herpes zoster in the 41‐year old roommate of the first varicella case patient. The roommate developed a localized vesicular rash on his right arm on 17 November 2008 (18 days before the start of the varicella outbreak) that lasted 8 days. He had a documented history of varicella disease in 1985 and a positive VZV IgG test result from a blood sample obtained on 10 December 2008. His lesions were thought to be fungal in origin at the time of rash. Because his lesions were not diagnosed as herpes zoster at the time of presentation, his rash was not kept covered.
Analysis. There were no statistically significant differences among case patients and non‐case residents by age, sex, or race and/or ethnicity. The mean age when they were first admitted to a residential facility was younger for case patients than non‐case residents (9.5 vs 15.0 years; ). A larger proportion of case patients than non‐case residents required full or some physical assistance, and proportionally more of them attended day program A. Non‐case residents were more likely to attend other day programs. Only 1 case patient was on an immunosuppressive medical regimen (methotrexate and tumor necrosis factor α inhibitor) for treatment of rheumatoid arthritis; there were no statistically significant differences between cases and non–case residents in the prevalence of medical conditions associated with immunosuppression.
Rescue of Severely Immunocompromised HIV‐Positive Persons. Part 3
Varicella disease in adults is uncommon in the United States because most adults have naturally acquired immunity. National seroprevalence data from 1988 to 1994 showed that 95% of adults 20 years old were immune to varicella‐zoster virus (VZV). In addition, adults comprised only 2%–4% of varicella cases in outbreaks reported from a varicella surveillance site during 1995–2005.
In December 2008, the Connecticut Department of Public Health was notified of a varicella outbreak in a residential facility for adults with intellectual disabilities (facility A) operated by the Connecticut Department of Developmental Services. The Connecticut Department of Public Health subsequently undertook an investigation to describe the outbreak and identify challenges in case management and outbreak control in this setting.
Case investigation. Facility A employed 145 staff and housed 70 residents with various levels of intellectual and physical disabilities in apartments in 3 buildings. Each apartment had 3 bedrooms with 2 beds in each bedroom. Staff included nursing, direct care, physical and occupational therapists, psychologists, cleaning staff, and clerical staff. A varicella case was defined as a generalized maculopapular rash (with or without vesicles and without another apparent cause) occurring between 1 November 2008 and 1 February 2009 in a resident or employee in facility A. Cases were identified by facility medical staff and/or chart review.
Information on all residents was abstracted from medical charts and facility admission histories using a standardized form and from interviews with facility caregivers. Staff case patients were interviewed by using a standardized case investigation form.
Laboratory testing. Serum samples were tested for VZV‐specific immunoglobulin M (IgM) and immunoglobulin G (IgG) by using an in‐house Centers for Disease Control and Prevention assay as described elsewhere and an enzyme immunoassay at a commercial laboratory. VZV DNA isolation from skin lesions by polymerase chain reaction (PCR) and genotyping were performed as described elsewhere. To identify VZV in the environment, sterile polyester swabs moistened with phosphate buffered saline were used to collect samples from various surfaces ( cm2 sample area) in residents’ rooms and common areas. Samples were collected from residential building 2, where no cases were identified, as control samples. Environmental samples were collected during the outbreak investigation and 2 months after rash onset in the last case. VZV detection and genotyping of the environmental samples by PCR was done as reported elsewhere. Laboratory staff was blinded as to whether specimens were case or control samples.
Definitions and statistical analysis. The attack rate was calculated as the proportion of cases among the residents of facility A. Residents’ degrees of intellectual and physical disabilities were categorized by an intelligence quotient score (moderate, 35–55; severe, 20–40; or profound, <20) and by functional ability (requiring full physical assistance, some physical assistance, or verbal and/or visual and/or psychosocial prompts). We used a Student t test for continuous variables and Pearson χ2 or Fisher exact test for categorical variables to analyze data. A significant association was defined as one with a 2‐sided P value of <.05. Approval from an institutional review board was not required because this investigation was conducted as part of a public health response.