Transfusion-Associated Babesiosis in the United States : A Description of Cases

Measurements: Distributions of ascertained cases according to Babesia species and period and state of transfusion. Results: 159 transfusion-associated B. microti cases were included; donors were implicated for 136 (86%). The case patients’ median age was 65 years (range, 1 to 94 years). Most cases were associated with red blood cell components; 4 were linked to whole blood–derived platelets. Cases occurred in all 4 seasons and in 22 (of 31) years, but 77% (122 cases) occurred during 2000–2009. Cases occurred in 19 states, but 87% (138 cases) were in the 7 main B. microti–endemic states. In addition, 3 B. duncani cases were documented in western states.

tic, clinical, and epidemiologic aspects of transfusionassociated and other Babesia cases. In addition to CDC records (such as records of test results, consultations, and case investigations), data sources for this endeavor included health departments, blood collection and transfusion services, other health professionals, and published materials and abstracts. The data available via health departments varied by jurisdiction and period; babesiosis was not a reportable disease in all states and was not nationally notifiable until January 2011. Although data were not systematically collected, some health departments, including those in babesiosis-endemic states, have routinely notified the CDC of potential transfusion cases and have submitted Babesia surveillance data to the CDC. Despite the inherent limitations of passive surveillance, collaborative relationships with public health and other pertinent agencies facilitated case ascertainment and data collection. We compiled and compared information obtained from multiple sources to maximize the quantity and quality of data and to minimize double counting.

Case Criteria and Classification
For these analyses, we established selection and classification criteria for transfusion-associated Babesia cases.  Table 1)

Linked to a donor (n = 118)
By type of case (cluster vs. not; index vs. not) and by class of index case (definite, probable, or possible). This figure, in conjunction with Table 1, provides perspective about the criteria for and the tallies of cases, donors, and donations. The 159 B. microti cases include 141 index cases and 18 nonindex, cluster cases. Each index case was associated with a different donor, whether implicated (n ϭ 118) or virtual (n ϭ 23; see Methods section).
The 61 index cases classified as definite include the index cases for the 12 multicase clusters ( Table 1), which encompass 18 additional cases, for a total of 79 cases. The 3 B. duncani cases are not included in the figure.

Context
Babesiosis, a parasitic infection transmitted through tick bites, can also be acquired via blood transfusion and may result in life-threatening disease. There is no U.S. Food and Drug Administration-licensed test to screen blood donors for Babesia infection.

Contribution
The risk for transfusion-associated Babesia infection may be increasing. Cases have occurred year-round and have been seen in states where Babesia species are not endemic.

Caution
Although the cases ascribed to transfusion undoubtedly represent a fraction of those that occurred, some tickborne cases inadvertently might have been included.

Implication
Improvements in the prevention and detection of transfusion-associated babesiosis are urgently needed.

-The Editors
Original Research Transfusion-Associated Babesiosis Our minimal case criteria included receipt of 1 or more cellular blood components during 1979 -2009, posttransfusion laboratory evidence of Babesia infection detected by 2010, and no reported evidence that another route of transmission (for example, tickborne or perinatal) was more likely than transfusion. We also required that linked (implicated) donors have laboratory evidence of infection.
We excluded potential transfusion cases if all pertinent do-  (5); the other 29 were linked to RBC components. Among infants with available data, the smallest transfused volume was approximately 8 mL. In 2 multidonation clusters (J and K), case patients were identified in 2 states. In cluster J, both donations were in Maine, by a donor probably exposed in Massachusetts; in cluster K, a Wisconsin resident also donated in Florida. Five of 12 implicated donors had parasitologically confirmed infection, on the basis of testing an original unit segment (B, C, D, and G) or subsequent specimens (H); the donor linked to cluster H still had demonstrable parasitemia, by PCR analyses, 4 mo after the fourth donation, 10 mo after exposure (5). For cluster A's donor, a segment was available but results of PCR analyses were negative. † "Seropositive" is noted only for the 4 nonindex cases that were not parasitologically confirmed: The reciprocal antibody titers ranged from 256 to 1024 in B. microti indirect fluorescent antibody testing. ‡ For recipients other than case patients, "tested negative" denotes seronegativity, at a minimum. § 18 cases (13 in infants and 5 in adults); 2-3 cases per cluster. 12 cases (all in adults); 1 case per donation; 2-4 donations per cluster.  † Because a lower proportion of patients with index vs. nonindex cases were younger than 1 y (P ϭ 0.001), the age distributions for index vs. nonindex patients were significantly different (P ϭ 0.009), but not if the age comparison was limited to adults (P ϭ 0.3). ‡ See Methods section and Figure 2. The "eastern state" category consists of Delaware, Florida, Indiana, Maryland, New Hampshire, North Carolina, Ohio, Pennsylvania, and Virginia. The "not an eastern state" category consists of California, Texas, and Washington. § If both were known and were different, the earlier month was specified. Data for the kidney donor (see text) were not included in analyses of month of diagnosis or interval to diagnosis. See Figure 3 regarding index patients. Among nonindex patients (Table 1), the interval to diagnosis depended on host factors, type of recipient (corecipient vs. other), and various aspects of the investigations. Although most of the ascertained nonindex patients who were adults reportedly were asymptomatic, clinical information in such regards typically was anecdotal or unspecified. In some investigations, other recipients could not be tested because they had already died.  (7) 1 (4)  9 (15) 5 (9)  5 (22)  13 (21) 9 (16)  4 (17)  0 0 ¶ Index cases were known or presumed to be parasitologically confirmed, with the exception of 2 cases classified as probable transfusion cases: the case in the kidney donor (see text) and a case diagnosed in retrospect, after recovery (30). ** The data constitute minimum numbers of case patients. Among the 12 known to have undergone splenectomy during the peritransfusion period, the contexts were trauma (n ϭ 8) or abdominal surgery for other reasons (n ϭ 4). The cases in the 3 patients known to have undergone posttransfusion splenectomy include 1 definite case (the index case of cluster L [25]; Table 1) and 2 probable cases, including the first described transfusion case (14).

Original Research Transfusion-Associated Babesiosis
† † Three received a kidney (living related [31], living unrelated, or cadaveric), 1 received a heart (29), and 1 underwent bilateral lung transplantation. ‡ ‡ Although outcome data were unavailable for some patients, we assumed that no other case patients died in the short term. The patients known to have died include 2 cluster-associated infants whose gestational ages were 23 and 24 wk, 2 (of 5) patients aged Ն90 y, and 6 (of 32) patients known to have undergone surgical splenectomy. § § In at least 4 case investigations, more than 1 donor had laboratory evidence of infection, typically 1 of whom was the most plausible on the basis of laboratory or epidemiologic data. However, the possibility of receipt of more than 1 contaminated unit could not be excluded. nors tested negative. If multiple cases were linked to the same donor, we defined the interrelated cases as a cluster, the first identified case as the index case (1 per donor), and the other cases in the cluster as nonindex cases (Figure 1 and Table 1). To facilitate bookkeeping, we defined all cases that were not cluster-associated as index cases (1 per donor).

Original Research
In general, index cases were parasitologically confirmed ( Table 2) (30,31); their detection prompted a transfusion investigation; and the linked donors and nonindex cases, if any, that were identified had parasitologic or serologic evidence of infection. We defined parasitologic evidence as detection of Babesia parasites (on blood smear or by animal inoculation) or Babesia DNA (by a molecular method). Serologic evidence of B. microti infection required positive results either by indirect fluorescent antibody (IFA) testing for total immunoglobulin or IgG or by immunoblot for IgG.
Index Babesia cases that fulfilled the selection criteria were classified as definite, probable, or possible transfusionassociated cases (Figure 1). If no donor was implicated among the subset of pertinent donors who could be tested, an index case was defined as a possible case, even if transfusion was the only known risk factor for infection. All index cases that were linked to a donor were classified as definite or probable cases. An index case was defined as a definite (vs. probable) transfusion case if at least 1 of the following additional criteria was fulfilled: 1) Transfusion was the only known or plausible risk factor for infection (for example, there was no history of residence or travel in babesiosis-endemic areas); 2) a multicase cluster was identified, with at least 1 nonindex case besides the index case ( Table 1); 3) the linked donor's infection was parasitologically confirmed by testing an extant segment from the original blood unit; or 4) other donor evidence indicated active infection at the time of donation (for example, a polymerase chain reaction [PCR]-positive specimen that reflected the donor's status at donation).

Data Analysis
We conducted univariate analyses for descriptive purposes by using Epi Info, version 3.5.1 (CDC, Atlanta, Georgia), and SAS software, version 9.2 (SAS Institute, Cary, North Carolina). Proportions were compared by using the chi-square test, or if expected cell counts were less than 5, the Fisher exact test. The Wilcoxon 2-sample test was used to compare the ranked distributions of ordinal variables. Statistical significance was defined as a 2-sided P value less than 0.05.
Unless otherwise specified, we stratified cases by period and state of transfusion (Table 2 and Figure 2) (32). We refer to 7 states with well-established foci of zoonotic transmission as "B. microti-endemic states": 5 states in the Northeast (Connecticut, Massachusetts, New Jersey, New York, and Rhode Island) and 2 in the upper Midwest (Minnesota and Wisconsin) (1, 12). The distinction be-tween these and other states (for example, in Figure 2) is not meant to imply that tickborne transmission occurs throughout these 7 states, that it occurred in all 7 states throughout 1979 -2009, or that these are the only states in which it did or does occur. Of note, during case selection and classification, we considered the evolving focality of tickborne transmission within and among states.

Role of the Funding Source
The study received no external funding.

General Perspective and Summary
For the period of 1979 -2009, we included 162 transfusion-associated cases: 159 B. microti cases and 3 B. duncani cases, which are described separately. The 159 B. microti cases include 12 multicase clusters encompassing 30 cases: 12 index cases (1 per cluster) and 18 nonindex cases (5, 8, 9, 20 -25) (Figure 1 and Table 1). In total, 141 B. microti cases were defined as index cases: the 12 clusterassociated index cases and 129 additional cases ( Table 2). Overall, 122 (87%) of the index cases (138 total cases [87%]) were associated with transfusions in the 7 main B. microti-endemic states (Figure 2 and Table 2), although not necessarily in areas of endemicity. The other 19 index cases (13%) generally were attributable to interstate movements of donors or blood components (Figure 2). Various scenarios are exemplified by the 4 cases not in eastern states ( Table 2), 2 of which were attributable to donor travels: A Rhode Island resident donated while training in Washington (26), and a Texas resident donated in that state after spending the summer in Massachusetts (6). In contrast, the other case in Texas and the case in California were linked to donations in New Jersey and Maine (27), respectively. Local distributions of components collected in New Jersey also accounted for 2 cases in Pennsylvania (8, 37) and 1 in Delaware (18). Table 2 summarizes selected characteristics of the cases, stratified by type of case (index vs. nonindex) and by class of index case (definite, probable, or possible). Table 1 provides additional perspective on the cluster-associated cases, which necessitated distinguishing between index and nonindex cases. Overall, the case patients had a median age of 65 years; 32% were either very old (33 were in the ninth or tenth decade of life) or very young (18 were infants, 13

Case Characteristics
Original Research Transfusion-Associated Babesiosis of whom were cluster-associated). The 19 patients with hereditary blood disorders account for 73% of the 26 patients in the age range of 4 to 43 years. These 19 patients include 11 with sickle cell disease (8,9,28), 7 with thalassemia major (35,43), and 1 with Diamond-Blackfan anemia (18); they account for at least 9 of the 32 patients known to have undergone surgical splenectomy. Three elderly patients with hematologic disorders underwent posttransfusion splenectomy (32, 134, or 215 days later), and their Babesia cases were diagnosed thereafter (Figure 3 and Table 2). For 2 of these patients, parasites were noted during retrospective review of presplenectomy blood smears, a finding that refocused the investigations on earlier transfusions and donors than on those initially targeted.
Five patients with transfusion cases had been transplanted with solid organs within the previous 3 months ( Table 2). In addition, indirect evidence suggests that a kidney donor who received multiple transfusions the day he died served as a conduit of Babesia parasites from 1 of his blood donors to both of his kidney recipients, who developed parasitologically confirmed infection (40). No B. microti antibodies were detected by IFA testing of archived pretransplantation serum from the kidney recipients or of pretransfusion serum from the kidney donor ( Table  2). However, postdonation specimens from 1 of his blood donors were seropositive (24).
The median interval from transfusion to onset of clinical manifestations was 37 days (range, 11 to 176 days) among 84 index patients with available data (Figure 3). Although babesiosis generally is considered a febrile illness, 13 (of 105) index patients were afebrile (9, 26, 32), including at least 4 adults who had cancer or were receiving immunosuppressive therapy. The median interval from symptom onset to diagnosis of index cases was 6 days (range, 0 to 54 days; n ϭ 84). Babesiosis often was diagnosed incidentally, in some instances during routine outpatient evaluations (6), during hospitalizations for unrelated reasons, or after the patient had recovered (30) or died (data not shown). Typically, Babesia parasites were an unexpected finding when a blood smear was examined, usually in the context of a complete blood count with a manual differential (9). When intraerythrocytic ring forms were noted, malaria was the first diagnostic consideration for more than 20 index patients, at least 14 of whom were initially treated for malaria.
The minimum all-cause mortality rate among index patients was 19% (6 -9, 18, 19, 32-34, 40, 44) (Table 2); Figure 3 provides various intervals to death. Some patients had a bleak prognosis even without the potential compounding effects of babesiosis. The 27 index patients known to have died include the kidney donor described earlier, whose posttrauma death on the day of transfusion clearly was unrelated to babesiosis. For other patients with available data, there was a spectrum of likelihood that   (8). Information on the donors linked to the other 15 index cases is provided in the text or the footnotes below for 7 and 8 cases, respectively. ‡ The donor was exposed in Massachusetts (32). § The 4 index cases in Maryland and Virginia were linked to donations in these states. The linked donors either were or could have been exposed in the Northeast. || The cases in Ohio (n ϭ 2) and Indiana (n ϭ 1) were linked to donations in Indiana (n ϭ 2) and Ohio (n ϭ 1) by donors exposed in B. microti-endemic states. babesiosis had a causal or contributory role (6, 7); causes of death often were presumptive or unclear (data not shown).

Blood Donors and Components
A linked donor with laboratory evidence of B. microti infection was identified for 118 index cases (84%), which encompass 136 total cases (86%) (Figure 1). Among the 117 linked donors whose B. microti IFA test results were known, the median reciprocal antibody titer was 256 (range, 64 to 4096; interquartile range, 256 to 1024). Twenty-four donors (20%) had parasitologically confirmed infection ( Table 2). The 20 donors with positive PCR results include 12 (71%) of 17 for whom blood retained from the original donation was tested compared with 8 (14%) of 56 for whom only postdonation specimens were available (P Ͻ 0.001). The median age of the 80 donors with available data was 49 years (range, 17 to 72 years); 18 donors (23%) were at least 60 years of age. Although clinical information typically was anecdotal or unspecified, some donors had pre-or postdonation symptoms or anemia of potential relevance (5, 24 -27). For example, the donor who had 4 consecutive donations linked to transmission (cluster H in Table 1) had been temporarily deferred because he was anemic when he first attempted to donate after exposure (5).
Among the 151 cases for which the type of blood component was determined, 4 cases were linked to whole blood-derived platelets (4,5,14) and 147 were associated with red blood cells (RBCs). The median age of liquidstored RBCs at the time of transfusion was 16 days (range, 4 to 40 days; n ϭ 106); 4 case patients received RBCs that were 35 to 40 days old. At least 4 patients received frozendeglycerolized (vs. liquid-stored) RBCs (18,35,43). Many patients received leukoreduced RBCs (data not shown); at least 10 received irradiated RBCs.

Babesia duncani Cases
The 3 documented B. duncani cases were linked to RBC transfusions in Washington (in 1994 [16]) and California (in 2000 [45] and 2008). In each instance, the case patient and implicated donor lived in the same state and had parasitologically confirmed infection. The case patients include a preterm infant (45), a 59-year-old man with a hemoglobinopathy (Bloch EM, Herwaldt BL, Leiby DA, et al. Unpublished data), and a 76-year-old man with a myelodysplastic syndrome who underwent cardiac surgery (16).

DISCUSSION
Babesiosis is an uncommon but potentially lifethreatening complication of transfusion that has been increasingly recognized since the first described U.S. case in 1979. Donor-screening practices do not yet include routine testing for evidence of Babesia infection. In this context, prompt detection, treatment, investigation, and reporting of Babesia cases are essential. Babesiosis should be included in the differential diagnosis of unexplained posttransfusion hemolytic anemia, with or without fever, regardless of the season or U.S. region. To enhance the abil- The data are limited to the subsets of the 141 index patients for whom particular intervals were relevant and were known or estimable (for example, the incubation period was unclear for some patients who had comorbid conditions or altered mental status). Each box represents the interquartile range (IQR), the internal vertical line indicates the median, the whiskers show the minimum and maximum, and the dots indicate the outliers with the longest intervals (Ն75% quartile plus 1.5 ϫ IQR). The 21 total dots-5 (6%), 6 (9%), 8 (7%), and 2 (10%) for the first, second, third, and fourth intervals from the top, respectively-are accounted for by 8 case patients, all of whom were linked to a donor. The farthest outliers include a patient with sickle cell disease who received hematopoietic progenitor cells from a sibling with sickle trait and became symptomatic approximately 6 months after the implicated peritransplantation transfusion (28) and 2 of 3 patients who underwent posttransfusion splenectomy (see text). The second interval from the top reflects the posttransfusion hospitalization during which babesiosis was explicitly diagnosed, for patients who had been discharged at least once in the interim or had been transfused as outpatients but were hospitalized thereafter. The fourth interval includes data for 21 of 27 index patients known to have died, including the kidney donor who died on the day he was transfused (see text). This interval was greater than 90 days for 2 immunocompromised patients whose intervals from diagnosis to death were less than 60 days. The patient who underwent splenectomy 215 days after transfusion died 280 days after transfusion; the patient's lymphoma also relapsed. For patients with available data, the median interval from symptom onset to death was 10 days (range, 2 to 51 days; n ϭ 18) and the median interval from diagnosis to death was 7 days (range, 0 to 55 days; n ϭ 22).
Original Research Transfusion-Associated Babesiosis ity of public health authorities to detect, monitor, and prevent transfusion and tickborne cases, babesiosis has been designated a nationally notifiable condition, effective January 2011; as such, cases reported to health departments are notifiable to the CDC. For the 31-year period of 1979 -2009, we included 159 B. microti transfusion-associated cases, which were dispersed in time (all 4 seasons and 22 years) and place (19 states). Protracted parasitemia in some infected donors (5,6,16,17), donor travels to and from areas of endemicity (6,26), and distributions or shipments of blood components account for the potential for year-round transmission anywhere in the country. Donor travel also accounted for the 1 reported transfusion-associated case of babesiosis in Canada, which was linked to a Canadian donor infected during a camping trip in Massachusetts (46). The majority (87%) of the 159 identified U.S. cases occurred in the 7 main B. microti-endemic states, which probably reflects higher risk and greater awareness. The annual case counts fluctuated, both overall and by locale (data not shown); the limited available risk estimates for transfusion transmission also have varied in time and place (2,3,8,9,30,39). Even so, that the majority (77%) of these 159 cases occurred during 2000 -2009 is noteworthy, regardless of whether some of the aggregate increase reflects improved recognition and reporting. In comparison, for the period of 1979 -2009, the CDC's National Malaria Surveillance System tallied 49 cases of transfusion-associated malaria, only 5 of which occurred during 2000 -2009 (Arguin P. Personal communication). Babesia microti has become the most frequently reported transfusion-transmitted parasite in the United States (2,3). In general, public health reports of tickborne Babesia cases also have increased in aggregate, with temporal and spatial fluctuations (CDC. Unpublished data); a national surveillance definition was first implemented in January 2011.
In addition to the 159 B. microti cases, we included 3 B. duncani cases in western states (16,45), for a total of 162 transfusion-associated cases. The B. duncani cases, like those caused by B. microti, were in patients who ranged from preterm to elderly and who had comorbid conditions. That infection with B. duncani-and with other U.S. zoonotic Babesia agents described since the 1990s (47, 48)-is not detected by serologic or molecular assays for B. microti has implications for diagnostic testing, transfusion investigations, and potential future donor screening.
As expected, almost all cases for which the type of component was determined were associated with RBC transfusions. Red blood cell components of all storage ages, including greater than 5 weeks, were associated with transmission, as were components that had been leukoreduced, irradiated, or frozen. Although we did not conduct risk analyses, our findings underscore that Babesia parasites can survive blood bank procedures and storage conditions for RBC components. The 4 identified cases linked to whole blood-derived platelets span from 1979 (the first described transfusion case) to 2000 and presumably were attributable to residual RBCs or to extracellular parasites in the platelet units (4,5,14,49). These 4 cases-and the cases in infants transfused with small RBC aliquots-underscore that small inocula can suffice to cause infection. However, even a segment from an implicated unit may test negative by PCR: The small volumes tested do not approximate the volumes transfused (1,2).
Some of the demographic and other characteristics of the case patients reflect those of transfused patients in general (2, 4) but may have particular importance in the context of babesiosis. For example, advanced age is a risk factor for severe babesiosis, even in otherwise healthy persons; transfusion recipients often have comorbid conditions that can increase their vulnerability to the compounding effects of babesiosis and interrelated complications (such as multiorgan dysfunction and death) (6,7,18,19,33,34). On the other hand, even some of the adult index patients were afebrile, including several patients receiving immunosuppressive therapies that may affect the host response to infection. Although most index cases with available data were diagnosed within 2 months of transfusion, a noteworthy minority of cases were diagnosed months later, such as in the context of posttransfusion splenectomy (Figure 3). These points not only have clinical relevance but also may affect transfusion investigations and case counts: The likelihood that transfusion transmission is considered and is investigated successfully may be lower for cases with longer intervals from the pertinent transfusion to symptom onset or diagnosis.
The 162 transfusion-associated cases we enumerated undoubtedly represent a fraction of those that occurred. The extent to which cases were not detected, investigated, or reported (to the CDC, to other public health authorities, or in publications) is unknown, both in general and with respect to periods, regions, and various case characteristics and outcomes. As underscored by the incidental diagnosis of Babesia infection, even severe cases in babesiosis-endemic regions can be missed or misdiagnosed, not just cases that are asymptomatic or mild or that occur in other U.S. regions. Even if a case is diagnosed, a transfusion investigation might not be considered, conducted, completed, or conclusive. The cases we included that were not linked to a donor (Figure 1 and Table 2) highlight the challenges associated with contacting all pertinent donors and obtaining posttransfusion specimens for testing; segments from the original donations typically are not still available. Our tallies probably constitute undercounts even of documented transfusion cases (for example, those that did not come to our attention or did not meet our selection criteria) but inadvertently might include some tickborne cases. As with all surveillance, case ascertainment, selection, and classification depended on the completeness and accuracy of the available data. Our findings underscore the year-round vulnerability of the U.S. blood supply-especially, but not only, in and near babesiosis-endemic areas. They also highlight the importance of multiagency collaborative efforts to detect, investigate, and document transfusion cases; to assess the risks for transfusion transmission; and, thereby, to inform the scope of prevention measures. In 2009, the Transfusion-Transmitted Diseases Committee of AABB (formerly, the American Association of Blood Banks) categorized babesiosis in the highest risk level for blood safety to be prioritized for intervention (50). Donors with subclinical infection are not identified by existing measures (such as temporary deferral of persons with systemic symptoms, fever, or anemia), no Babesia assay for screening donors has been approved by the FDA, and pathogen reduction techniques for RBCs or platelets are not available in the United States (1, 2, 50). The FDA's Blood Products Advisory Committee that was convened on 26 July 2010 supported the concept of regional donor testing for Babesia (51). The increasing recognition of transfusion cases strengthens the impetus for screening strategies that mitigate the transmission risk (1-3, 50, 51), including testing approaches implemented under FDA-approved protocols (1, 3, 51) and longer-term strategies with development of a highthroughput Babesia screening assay.