Household Transmission of Pandemic (H1N1) 2009, San Antonio, Texas, USA, April–May 2009

Transmission rates were lower than those for seasonal influenza.

The Texas Department of State Health Services provides public health services in counties without local health departments. Within Health Service Region 8, public health services for Bexar County are provided by the San Antonio Metropolitan Health District and for Comal County by the Comal County Health Department.

Case Defi nitions
We defi ned a laboratory-confi rmed case-patient as a resident of Health Service Region 8 who had a positive respiratory specimen showing nucleic acid sequences unique to pandemic (H1N1) 2009; a real-time reverse transcription-PCR (rRT-PCR) assay was used to detect the virus (9). For persons with no laboratory test performed, we assessed whether they had infl uenza-like illness (ILI), defi ned as fever (measured or unmeasured) with either cough or sore throat; or acute respiratory infection (ARI), defi ned as >2 of the following signs or symptoms: fever, cough, sore throat, and rhinorrhea. The index case-patient was defi ned as the household member with the earliest symptom onset date of ARI, ILI, or laboratory-confi rmed pandemic (H1N1) 2009. A secondary case-patient was defi ned as a household member with ARI, ILI, or laboratory-confi rmed pandemic (H1N1) 2009 and symptom onset 1-9 days after symptom onset in the index case-patient. We chose the maximum interval of 9 days because shedding of seasonal infl uenza virus uncommonly lasts >8 days (10) and the median incubation period for seasonal infl uenza is ≈1.4 days (11). Household members were defi ned as persons who lived at the same address as a case-patient who had laboratory-confi rmed pandemic (H1N1) 2009 infection.

Case Finding
During April 10-May 8, 2009, we identifi ed laboratory-confi rmed cases of pandemic (H1N1) 2009 by reviewing 1,167 laboratory records of infl uenza specimens submitted by healthcare providers for rRT-PCR testing by the regional public health laboratory in San Antonio. We also reviewed 1,251 laboratory records of all specimens submitted by military medical treatment facilities in San Antonio. These specimens were tested for infl uenza by rRT-PCR at the Epidemiology Laboratory Service of the Department of Defense Global Infl uenza Surveillance Program at the United States Air Force School of Aerospace Medicine in San Antonio. In addition, we conducted telephone interviews with 540 (67%) of 802 high school students who were reported as absent by their school administrators during April 9-28 in the Texas counties where the fi rst 2 identifi ed case-patients attended school. Respiratory samples were collected from students who reported an acute respiratory illness at the time of interview. Additional casepatients were identifi ed by collecting respiratory samples from nonhousehold contacts of laboratory-confi rmed casepatients (i.e., those who had been within 6 feet of someone with ARI for at least 1 hour during the period 1 day before through 7 days after onset of illness in the contact).

Household Investigations
We interviewed case-patients with laboratory-confi rmed infection and all household members about the occurrence of illness, receipt of infl uenza vaccination in the previous 12 months, and medical history. We asked all persons about their use of antiviral medication and reviewed health department pharmacy records where appropriate to ascertain the type, dosage, and timing of antiviral medication and to defi ne whether antiviral medications were prescribed for treatment or prophylaxis. Respiratory samples were collected from household contacts who had an acute respiratory illness at the time of interview; respiratory samples were collected from all members of 9 households identifi ed early in the investigation, regardless of respiratory symptoms.

Sample Collection and Laboratory Testing
Nasal wash samples were collected from military servicemen and their household family members, and nasopharyngeal swabs were collected from all others. Nasal wash samples were sent to the Epidemiology Laboratory Service at the United States Air Force School of Aerospace Medicine; nasopharyngeal swabs were sent to the regional public health laboratory in San Antonio. We used rRT-PCR to test all respiratory samples for seasonal infl uenza (A/H1 and A/H3 infl uenza viruses). Specimens positive for infl uenza A but negative for seasonal infl uenza by rRT-PCR were sent to the Centers for Disease Control and Preven- tion (CDC) for confi rmatory testing for pandemic (H1N1) 2009 (12).

Statistical Analysis
We calculated the serial interval as the number of days from the onset date of illness in the index case-patient to onset date of illness in the secondary case-patient. Secondary household attack rates were calculated by dividing the number of secondary case-patients (excluding the index case-patient) by the total number of household members (excluding the index case-patient). Secondary case-patients for ILI and ARI attack rates also included laboratory-confi rmed case-patients. We compared characteristics between groups by using the χ 2 test or Fisher exact test for categorical data and the Wilcoxon signed-rank test for continuous variables (13).

Ethics
The collection of information about cases of pandemic (H1N1) 2009 was part of the emergency public health practice response and was not deemed to be research in accordance with the federal human subjects protection regulations (45 Code of Federal Regulations 46.101c and 46.102d) and CDC's Guidelines for Defi ning Public Health Research and Public Health Non-Research. All protocols pertaining to the pandemic were reviewed for protection concerns and the necessity of Institutional Review Board review by the CDC's National Center for Immunization and Respiratory Diseases (NCIRD) Human Subjects Contact and the NCIRD Associate Director of Science.

Results
We identifi ed 110 persons with laboratory-confi rmed pandemic (H1N1) 2009 infection. We were unable to contact 23 (21%) of these persons, and 3 (3%) did not agree to provide further information. Of 84 persons with laboratoryconfi rmed pandemic (H1N1) 2009 infection who provided information, 77 (92%) lived with >1 persons. These 77 households comprised 349 persons; the median household size was 4 persons (range 2-9 persons), including the index case-patient. Seventy fi ve percent of household interviews were conducted >8 days (range 0-24 days) after the onset of infection in the index case-patient.
From household interviews, we identifi ed an additional 47 persons who reported respiratory symptoms or had laboratory evidence of pandemic (H1N1) 2009 infection: 13 persons with laboratory-confi rmed pandemic (H1N1) 2009 infection, 24 persons with ILI, and 10 persons whose illness met the case defi nition for ARI only. We did not classify 15 of these persons as secondary case-patients: 8 persons had the same date of symptom onset as the index case-patient; we could not establish the date of symptom onset for 3 persons; and 4 persons reported illness onset 10-15 days after the index case-patient. In 1 household where 2 persons had ILI, 1 had a nasopharyngeal swab that was positive for pandemic (H1N1) 2009; the other was positive for infl uenza A, but the subtype could not be determined, possibly because of the quality of the sample or because 9 days had elapsed between illness and sample collection, thus decreasing viral load. We considered this person to have laboratoryconfi rmed pandemic (H1N1) 2009 on the basis of an epidemiologic link to another laboratory-confi rmed case. In 2 households where secondary case-pateints were identifi ed, nasal swab samples were obtained from members of all 7 households; 1 person, 14 years of age, who did not report any respiratory symptoms, was positive for pandemic (H1N1) 2009 infection.
Secondary case-patients were found in 24 (31%) of 77 households; 5 had 2 secondary case-patients, and 1 had 3 case-patients (Table 1). Secondary infections appeared most likely to be transmitted between children (12/32, 38%) or children to adults (10/32, 31%) than from adults to children (6/32, 19%) or adults to adults (4/32, 13%) (p = 0.034). The median serial interval for ARI, ILI, and laboratory-confi rmed pandemic (H1N1) 2009 combined was 4 days (range 1-9 days) ( Table 1; online Appendix Figure, www.cdc.gov/EID/content/16/4/631-appF.htm). Antiviral treatment was given to the index case-patient of 23 (72%) of 32 secondary case-patients; in these households, the serial interval was 3 days, compared with 5 days when the index case-patient was not given treatment (p = 0.17). Inclusion of 5 household contacts with illness that occurred 10-15 days after symptom onset of the index case-patient did not alter the median serial interval estimate. The median serial interval also remained unchanged when only members of households interviewed >9 days after the onset of symptoms in the household index case-patient were included. Limiting the estimate of median serial interval to include only persons with ILI or laboratory-confi rmed case-patients reduced the median serial interval to 3 days (range 1-8 days).
The secondary household attack rate was 13% for ARI, 9% for ILI, and 4% for laboratory-confi rmed pandemic (H1N1) 2009 (Table 2). Secondary attack rates were highest in children <5 years of age and were higher in children 5-18 years of age than in adults 19-49 and >50 years of age ( size, nor diagnosis of the index patient (with ARI, ILI, or laboratory-confi rmed pandemic [H1N1] 2009) were predictive of treatment with antiviral medication. The secondary household attack rates for ARI, ILI, and laboratoryconfi rmed pandemic (H1N1) 2009 combined in households where the index case-patient was given antiviral treatment was 12% compared with 16% in other households (p = 0.64) ( Table 3). Antiviral prophylaxis was given to 39% of household contacts (92/235 with data available) (Table  3), and the secondary attack rate of ARI, ILI, and laboratory-confi rmed pandemic (H1N1) 2009 combined was 14% (12/83) in households where the index patient took treatment, compared with 66% (6/9) (p = 0.003) in households where the index patient did not take treatment (Table 3).

Discussion
During an outbreak of pandemic (H1N1) 2009 in the San Antonio, Texas, area, we identifi ed 97 persons with laboratory-confi rmed infection in 77 households. The epidemiologic and clinical features were similar to summary reports from the United States (14,15) and other countries (15,16). Nearly one third of households had secondary case-patients who also had respiratory illness, with a median of 4 days between onset of illness in the index casepatient and household members, a fi nding similar to that for seasonal infl uenza (17).
The secondary attack rate was 4% for laboratoryconfi rmed pandemic (H1N1) 2009, 9% for ILI, and 13% for ARI. In general, these rates are lower than for seasonal infl uenza and lower than anticipated for a pandemic strain, although rates vary from 13% to 30%, depending on infl uenza subtype and year and pandemic period (4,5,(18)(19)(20)(21). The highest proportion of laboratory-confi rmed pandemic (H1N1) 2009 and secondary attack rates occurred in children, a fi nding consistent with the epidemiology of seasonal and pandemic infl uenza, where we know children experience higher rates of illness (4,5,7) and higher secondary attack rates (19). Adults may have some cross-protection against pandemic (H1N1) 2009 from antibodies developed during infections with seasonal infl uenza A virus (H1N1) (22-24).
Four randomized controlled trials of zanamivir and oseltamivir for seasonal infl uenza have shown that these antiviral medications reduce but do not eliminate viral shedding *ARI, acute respiratory infection; ILI, influenza-like illness (fever measured or subjective and cough or sore throat). Ill household members were not included in the calculation of the secondary attack rate if they had the same symptom onset as the index case or if symptom onset was not known. and are effective in preventing disease among household contacts, especially if taken within 48 hours of illness onset in the index case-patient (19,20,25,26). We found that secondary attack rates for all households were lower when the index case-patient received treatment, although this difference was not signifi cant. The role of prophylaxis in the absence of treatment of the index case-patient was difficult to determine; our investigation included only a small number of such persons. Nevertheless, because most index and secondary case-patients received antiviral treatment, household secondary attack rates may have been reduced. Our investigation has several limitations. Because early case fi nding was most intensive among high school children associated with school outbreaks, our cohort may have been biased in favor of households where the index case-patients were children; however, this would not explain a lower secondary attack rate among adult household contacts. We did not assess the role of mild or asymptomatic pandemic (H1N1) 2009 infection because we collected respiratory samples only; serologic assays to detect infl uenza antibodies are the most sensitive method for detecting asymptomatic infection, but virus assays for pandemic (H1N1) 2009 were not available at the time of the investigation. Volunteer challenge studies with seasonal infl uenza viruses have found that up to 30% of infected persons may be asymptomatic and could be identifi ed through serologic testing (10). Because 25% of household interviews were conducted <8 days after onset of illness of the index case-patient, we may have underestimated the secondary attack rate if these households had secondary case-patients with long serial intervals. However, when we restricted our analysis to persons interviewed >8 days after onset of symptoms in the index case-patient, we found no difference in the median serial interval or distribution of attack rates by age. Conversely, household members interviewed >8 days after onset of illness in the index case-patient may have had incomplete recall of acute respiratory infections. Finally, some of the secondary illnesses may have been acquired in the community, leading to overestimate of household secondary attack rates.
We found that pandemic (H1N1) 2009 disproportionately affected children, who in turn posed a risk for secondary household transmission, especially to their caregivers and siblings. The Advisory Committee on Immunization Practices (2009) recommends that children 6-18 years of age and caregivers of infants be included as initial target groups for the new pandemic (H1N1) 2009 vaccine (27), which may reduce household transmission. As pandemic (H1N1) 2009 continues to spread internationally, ongoing investigations are needed to shed further light on transmission dynamics, to monitor epidemiologic changes over time, and to assess the effectiveness of public health interventions.