The index case-patient was an 87-year-old woman with congestive heart failure who was referred from a disaster medical assistance team and admitted to our hospital on April 6. She reported leg edema starting at the end of February and a cough beginning in early March. On March 11, the day of the earthquake and resulting tsunami, she spent a night in a shrine near her home, which had been completely destroyed. She had also lost all of her daily medications. She stayed at a disaster shelter with family members and other evacuees on March 12–15 (3 days) and then moved to her daughter’s home. She reported that she experienced fever, headache, general malaise, appetite loss, “terrible” cough, sputum, and dyspnea on April 4; on April 6, she visited the disaster medical assistance team.

Chest radiograph and computed tomography scan were performed, and results revealed extensive infiltrative shadows with air bronchograms in the right lung; results were normal for the left lung. Pneumonia was diagnosed, and an empirical therapy of intravenous antimicrobial drugs was initiated, pending further evaluation for pulmonary tuberculosis. Sputum smear testing showed acid-fast bacilli (AFB) graded Gaffky 2 (grade 1+ on the World Health Organization scale) (4). Mycobacterium tuberculosis complex infection was identified by PCR. The patient was placed in a negative-pressure single room and treated with a 3-drug regimen of isoniazid, rifampin, and ethambutol. Sputum culture yielded M. tuberculosis susceptible to the drugs administered. The patient improved and was discharged on June 17 after sputum smears were repeatedly negative for AFB. She continued to receive directly observed treatment at our outpatient clinic and completed 9 months of treatment.

The shelter at which the patient stayed after the earthquake was small (60 m²), and ≈50 evacuees stayed there at the time she was there. Ventilation was poor because the weather was cold and windows were not opened. Mask supply was insufficient, and most persons did not wear masks or did not wear them properly. Obtaining information on persons who had contact with the index case-patient was difficult because many evacuees were exhausted from stress or had moved to secondary shelters or relatives’ home by the time we visited the shelter. In cooperation with a manager of the shelter and local government, a contact investigation for the index case-patient patient was conducted during June–August 2011 (2–4 months after the last possible exposure) to identify LTBI; a total of 62 contact persons were found. Three contacts (an infant and children <7 years of age) had a tuberculin skin test (TST), and the remaining 57 contacts (persons >7 years of age) underwent whole-blood interferon-γ release assay (IGRA) testing by using QuantiFERON-TB Gold In-Tube (Cellestis, Chadstone, Victoria, Australia), as described (5).

Two of 3 contacts tested by TST were positive for tuberculin purified protein derivative (PPD); 9 of 57 contacts tested by IGRA were positive (Table). A 2-month-old infant, a family member of the index case-patient, had been initially tested in May and was PPD negative but received isoniazid therapy for LTBI because the infant had close and frequent contact with the index case-patient. Two months after the 6 months of treatment was completed (8 months after treatment was initiated), results of a repeat TST on this infant were positive.

For PPD- or IGRA-positive contacts, medical examinations and chest radiographs were performed, but no results characteristic of pulmonary TB were found. After a physician explained risks and benefits of prophylaxis to the contacts (or parents), 8 contacts (7 evacuees and 1 family member) received prophylactic treatment for LTBI; the other contacts received follow-up chest radiography. One year after the earthquake, no active TB cases had been observed among the contacts.

We detected LTBI among evacuees who were exposed to a patient with active TB at a shelter after the 2011 Great East Japan Earthquake. Refugees and populations displaced after natural disasters are particularly vulnerable to TB in developing countries because crowded living conditions and poor nutritional status can facilitate the development and the transmission of TB (6,7). However, as seen in this report and others (8), displaced populations in industrialized countries may also be vulnerable to communicable diseases after natural disasters. Persons in shelters who have TB or suspected TB should be transferred to a medical facility as soon as the illness is detected because isolation and respiratory protection for airborne diseases such as TB is very difficult to implement in shelters (9).

In our case, some physicians and nurses who saw the index patient did not consider the diagnosis of TB or could not perform AFB testing, which may have delayed time to TB diagnosis and resulted in TB spread in the shelter. Health care personnel working in disaster relief must suspect and rapidly diagnose TB and then conduct contact investigations in collaboration with local public health departments. Signs and symptoms and clinical characteristics of TB in the elderly may be atypical (10), so continuous medical education and training are needed to maintain the competence of health care personnel to prevent, diagnose, and treat TB in the elderly.

In this investigation, TST or IGRA results were positive (indicating LTBI) in 11 (18.3%) of 60 contacts; 10 (20%) of 50 evacuees at the shelter had positive results. A previous study in Japan found that the IGRA-positive rate was 7.1% for those 40–69 years of age in the general population (11). Japan is considered a middle-burden TB country and has a bacille Calmette-Guérin vaccination program; for these reasons, IGRA is more useful than TST in LTBI screening and contact tracing.

Adherence rates for LTBI treatment were low (19%) at a TB clinic in New Orleans after Hurricane Katrina (8). This finding indicates a challenging environment for TB control activities after a natural disaster and suggests an increased risk for transmission because of migration and overcrowding. Disrupted health care services, poor access to TB control programs, and difficulty in patient management may lead to poor treatment adherence, which could result in the emergence of drug-resistant TB strains. However, all (8/8, 100%) patients with LTBI who initiated treatment in our investigation adhered to and completed the regimen. Our success in postdisaster TB control measures at this shelter can be attributed, in part, to the efforts of public health nurses in providing education and directly observed treatment despite limited resources and poor health care access in the affected area. Medical institutions and public health departments should work to cooperatively and collaboratively assist shelter in implementing TB care and control activities for evacuees after natural disasters.