Testing for influenza was not only symptom-driven but was also triggered by elevated C-reactive protein (CRP), which is assessed daily as part of routine care for all patients treated at MIRT. A nasopharyngeal swab was performed on all patients suspected of having influenza-like illness (ILI) or an elevated CRP (>10 mg/L) on or after July 25, 2013. Influenza-like illness was defined as subjective fever or measured temperature of ≥37.7°C (≥100°F) with cough or sore throat, consistent with CDC and the Council of State and Territorial Epidemiologists (CSTE) surveillance definition for ILI. Testing was subsequently expanded to include employees and family members with ILI. Any patients in other areas of the hospital who had ILI were also subject to testing. All specimens were analyzed by the UAMS Department of Pathology′s molecular diagnostics laboratory by using the eSensor Respiratory Viral Panel (RVP) (GenMark Dx, Carlsbad, CA). This assay is a qualitative nucleic acid multiplex diagnostic test that provides for the simultaneous identification of multiple respiratory viral pathogens, including the following: influenza A virus, influenza A virus (H1) seasonal subtype, influenza A virus (H3) seasonal subtype, influenza A virus 2009 pandemic (H1N1) subtype, influenza B virus, respiratory syncytial virus subtype A or subtype B, parainfluenza virus 1, parainfluenza virus 2, parainfluenza virus 3, human metapneumovirus, human rhinovirus, adenovirus species B/E, and adenovirus species C. Because of the unusual timing of influenza, samples were repeated to confirm positivity by PCR both at UAMS and the Arkansas Department of Health (ADH). Information regarding demographics, baseline clinical parameters, treatment modalities, comorbidities, and symptoms were collected by using a standardized interview form and clinical chart abstraction. This investigation was reviewed by CDC for human subjects protection and deemed to be public health practice.

Frequencies of the collected data were calculated, and graphic representations demonstrating the outbreak were created by using Microsoft Excel 2010 (Microsoft Corp., Redmond, WA) and Epi Info 7 (CDC, Atlanta, GA).

With the nonseasonal identification of influenza aiding early recognition of the outbreak, interventions were implemented to identify infected patients and limit spread to others. These interventions led to control of the outbreak within 14 days after identification of the first patient, with no spread to patients or staff in other clinics or areas of the hospital. No outbreak-attributable deaths occurred.

Institutional and local resources were enlisted early on, and such measures as patient cohorting, enhanced cleaning of waiting areas and transport shuttles, provision of personal protective equipment, and increased signage and supplies to encourage hand hygiene were instituted by the infection control team on the day the outbreak was discovered. Surgical masks were provided to all patients entering the center for the duration of the outbreak. After identification of the outbreak, all physicians on the medical campus were alerted to remain vigilant for ILI and consider testing for respiratory viruses. Social distancing was facilitated by rescheduling all nonemergency transplantations and chemotherapy. Furthermore, satellite clinics were established for patients receiving maintenance therapy beginning on the day of the outbreak, continuing until no new cases had been reported for 1 month and existing cases had 2 documented negative RVPs ≥1 week apart. Patients who tested positive were treated with oseltamivir, whereas RVP PCR testing was done weekly until test negative. Given that this outbreak occurred midsummer, influenza vaccine and institutional oseltamivir stocks were limited; therefore, ADH provided oseltamivir from the state stockpile for all exposed family and staff. The center′s molecular laboratory increased the number of RVP PCR runs from once daily to twice daily, including weekends, to enable rapid identification of cases. Arkansas Department of Health provided viral subtyping, and CDC assessed for antiviral resistance as well as susceptibility to the existing and forthcoming vaccine. The CDC also facilitated a partial distribution of the 2013–2014 influenza vaccines order to UAMS ahead of schedule by the manufacturer. Given the limited supply of vaccine available, personnel at high risk were identified, and vaccine was rapidly provided on location to all staff with close contact to MIRT patients on the bone marrow transplant unit, the outpatient transplant center, and other ancillary departments (eg, apheresis laboratories). Vaccination was encouraged and also made available for free by ADH for family members and caregivers of MIRT patients.

Multiple factors likely facilitated the rapid containment and favorable outcomes observed. Early recognition of the outbreak was possible because of vigilance and surveillance provided by the infection control team. Respiratory viral illnesses are not uncommon among the center′s patient population; however, the typical viruses identified during the summer months are parainfluenza, respiratory syncytial virus, and metapneumovirus. Influenza prevention measures among the general population might not be applicable to such populations as cancer patients, and our report highlights some of the elements clinicians need to take into account when developing protocols for prevention among such patients. Cancer patients are encouraged to get influenza vaccinations yearly, and vaccination has been demonstrated to provide adequate protection even for patients undergoing conventional cancer chemotherapy [5, 6]. However, investigators have reported reduced vaccine efficacy among specific cancer patient populations, particularly patients receiving rituximab (an anti-CD20 antibody) [7]. This drug targets B cells [8], a crucial component of the humoral immune system upon which vaccine strategies depend to generate an anamnestic response. Multiple myeloma results from an abnormal proliferation of B cells [17]; thus, myeloma therapy often targets these cells that happen to be the cornerstone of vaccine strategies. As expected, vaccine efficacy has been reported to be low among MM patients [9]. Furthermore, although CDC/Infectious Diseases Society of America recommends vaccination of HSCT recipients 6 months after their transplantation [6, 10], for patients such as those undergoing tandem transplantations for treatment of the underlying cancer, the treatment plan might not permit timely vaccine administration because of the need for multiple and frequent cycles of chemotherapy. Therefore, treating physicians on occasions resort to chemoprophylaxis, albeit not favored, as a prevention strategy during the influenza season [11]. Chemoprophylaxis-dependent strategies for influenza prevention depend on early recognition and timely implementation of vital interventions to be effective. However, vigilance for influenza during the summer months might be lower than during the typical influenza season.

Another infection control challenge that cancer patients face is optimal environmental engineering for outpatient care. Stem cell transplantation historically has been performed in closed bone marrow transplant units with restricted access. These units also typically have HEPA filtration and laminar airflow [12]; however, with increasing use of less immunosuppressive conditioning and the shift from bone marrow to HSCT, the need for such units, especially for autologous stem cell recipients, has reduced. Stem cell transplantation is frequently being performed in the outpatient setting [13]. As a result of high pedestrian traffic in outpatient areas, social distancing is difficult to achieve, and, consequently, a high potential exists for droplet transmission. In this center, the inpatient unit houses 30 single-patient rooms and a maximum of 32 staff per shift in a 23 812-square-foot area, in contrast with the outpatient center that serves 120–150 patients daily and is staffed by 30–38 staff per shift in a 21 992-square-foot area. Thus, patient density, coupled with ongoing immunosuppression, creates ideal conditions for an outbreak such as this to occur. Evidence suggesting the potential for airborne transmission of viruses, such as influenza [1], raises questions as to whether ambulatory transplant centers are prone to airborne transmission in this manner. The lack of documented cases among patients outside the transplant center suggests that immunosuppression and outpatient clinic characteristics might have played a key role in this outbreak.

Of note, although a majority of the patients involved in the outbreak had normal white blood cell counts and absolute neutrophil counts >1000 cells/mm³, these patients were in all likelihood somewhat immunosuppressed because of recent receipt of corticosteroids, stem cell transplantations, ongoing myeloma treatment, and the high prevalence of low lymphocyte counts. T-cell recovery can take as long as 12 months among patients who receive HSCT [14] or considerably longer among patients receiving multiple transplantations. Persistent immunosuppression might also explain why typical influenza symptoms (eg, fever) were not common among these patients [15]. Relying solely on clinical symptoms or ILI case definitions that have been proposed by CDC [16], in this patient population would have impaired case detection because this strategy identified only 23% of cases. In contrast, CRP-guided screening was able to identify 68% of cases. A limitation of this investigation was that most of the patients tested were those receiving treatment at the ambulatory stem cell center where CRP is performed daily; consequently, cases outside of this setting might have been missed. Second, the unique nature of the patient population also might mean certain measures undertaken to contain the outbreak would not be feasible or generalizable to other patient populations.