Sepsis Attributed to Bacterial Contamination of Platelets Associated with a Potential Common Source — Multiple States, 2018

Investigation and Results

California. On May 4, a male patient with acute lymphoblastic leukemia (patient A) received pathogen-reduced apheresis platelets at hospital A in California (Figure). Within minutes of completing the transfusion, he briefly experienced rigors, followed 2 hours later by fever and hypotension. He was transferred to the intensive care unit for management of septic shock and recovered fully. Posttransfusion patient blood cultures (obtained 2 hours after vancomycin administration) grew only ACBC. Gram stain of the implicated platelet bag residual revealed gram-positive cocci in pairs or chains; culture of the platelet bag residual grew ACBC and S. saprophyticus. The implicated platelet unit was one of two platelet units manufactured from a single apheresis donation collected 5 days earlier in California. Pathogen inactivation was performed 13.5 hours after collection. Hospital A located the second platelet unit (which had not been transfused), quarantined it, and notified the blood supplier. Gram stain and culture of this platelet unit were negative. Samples obtained from the donor’s skin were culture-negative for ACBC and S. saprophyticus. Environmental samples obtained weeks later from the platelet collection facility and hospital A yielded no relevant organisms; however, sampled areas had been cleaned in the interim.

Utah. On May 10, a male patient with cirrhosis and thrombocytopenia (patient B) received a platelet transfusion at hospital B to prevent bleeding before a procedure (Figure). One hour after transfusion began, patient B complained of chills, and the transfusion was terminated. Two hours after transfusion, he became febrile, hypotensive, and tachypneic, and antibiotics were started; he died of septic shock 2 days later. ACBC was isolated by culture from platelet bag residuals and posttransfusion blood samples from the patient. The platelet supplier was notified, and a second platelet unit manufactured from the same apheresis donation, which had not been transfused, was recalled. The platelet supplier performed primary aerobic culture of the implicated donation for bacterial contamination 24 hours after collection in Utah; the primary culture remained negative after 5 days. The implicated platelet unit was transfused 5 days after collection. Samples obtained on May 24 from the donor’s urine, perianal area, and multiple skin sites screened negative for ACBC colonization. Samples obtained from platelet agitators at the platelet manufacturing facility (May 23) and hospital B (June 7) yielded ACBC isolates.

Connecticut and Massachusetts. On October 4, at hospital C, two male patients with acute myeloid leukemia (patients C and D) each received a platelet unit manufactured from a common apheresis donation (Figure). Within two hours of transfusion, both patients became hypotensive and febrile. Both were transferred to the intensive care unit, and both recovered. ACBC and S. saprophyticus were isolated by culture from posttransfusion blood samples from both patients and from both platelet bag residuals. Hospital C notified the platelet supplier. The implicated apheresis platelet donation had been collected in Massachusetts 4 days before transfusion and processed in Connecticut. Twenty-four hours after collection, the platelet supplier performed primary aerobic and anaerobic culture for bacterial contamination. Within 5 hours before transfusion, hospital C screened both platelet units with a rapid bacterial detection device; all tests were negative. No ACBC or S. saprophyticus isolates were identified among environmental swabs collected at platelet supplier facilities in Connecticut (November 15) and Massachusetts (November 16); S. saprophyticus was isolated from one platelet agitator at hospital C on November 13.

Multistate investigation. On July 17, notices were issued through CDC’s Epi-X and the Infectious Diseases Society of America’s Emerging Infections Network to identify additional cases of sepsis caused by Acinetobacter infections with onset of symptoms within 24 hours after platelet transfusion. Three cases were reported from two states (North Carolina [patients E and F] and Michigan [patient G]).

Traceback investigation revealed that the three platelet donations implicated in the California, Utah, and Connecticut septic transfusion reactions (i.e., sepsis attributed to transfusion) were from different donors. The donors in California, Utah, and Massachusetts had no known epidemiologic links to one another and no symptoms suggesting bacteremia or illness; all were indefinitely deferred. All three apheresis donations were collected in platelet additive solution using apheresis cell separator machines and collection sets from the same manufacturer; two of three collection sets were from a single lot.

CDC performed whole genome sequencing on collected ACBC and S. saprophyticus isolates (Table) using standard methods (1). ACBC organisms were isolated by culture from posttransfusion blood samples from patients A, B, C, and D; all four associated transfused platelet bag residuals; and environmental samples from hospital B and the platelet supplier in Utah. Fourteen ACBC isolates from these sources were highly related (differing by 0–32 single nucleotide polymorphisms [SNPs] across a 95.6% core genome) (Supplementary Figure 1, https://stacks.cdc.gov/view/cdc/78727) and appear to represent a novel ACBC taxon (only 90% match to Acinetobacter seifertii by average nucleotide identity). In contrast, ACBC isolates from cases in North Carolina and Michigan were not closely related to isolates from cases in California, Utah, and Connecticut by whole genome sequencing (differing by 13,398–14,289 SNPs across a 30.5% core genome).

S. saprophyticus was isolated by culture from posttransfusion blood samples from patients C and D; transfused platelet bag residuals from patients A, C, and D; and an environmental sample from hospital C (Table). Whole genome sequencing analysis revealed two clusters of S. saprophyticus isolates. One cluster consisted of S. saprophyticus isolates from patient C’s blood, patient C’s platelet bag residual, and an environmental swab from hospital C (differing by 0–37 SNPs across a 94.9% core genome) (Supplementary Figure 2, https://stacks.cdc.gov/view/cdc/78728); the second cluster consisted of isolates from patient D’s blood and from patient D’s and patient A’s platelet bag residuals (difference of 1–27 SNPs across a 94.9% core genome).

Discussion

Transfusion of platelets is more likely to result in sepsis than is transfusion of other blood products; data derived from primary cultures have indicated that approximately one in every 5,000 platelet collections is contaminated with bacteria (2). ACBC is not frequently reported as a contaminant of platelets (3). ACBC consists of gram-negative bacilli that commonly occur in wet environments and are opportunistic pathogens; ACBC organisms are resistant to desiccation, persist on environmental surfaces, and avidly adhere to plastics (4). Conversely, coagulase-negative Staphylococcus spp. are among the most common bacterial contaminants of platelets (3,5). However, S. saprophyticus might be less likely to contaminate platelets than other coagulase-negative Staphylococcus spp. because it typically resides in the gastrointestinal and urinary tracts rather than on the skin (6).

Whole genome sequencing analysis indicated an unidentified potential common source of bacterial contamination among the four cases of septic transfusion reactions reported here. Investigation into the contamination source continues. Although skin microflora and donor bacteremia are the most frequent sources of bacterial contamination (7), a cluster of septic transfusion reactions attributed to contamination of blood collection bags during manufacturing or packaging was reported in 1993 (8).

Food and Drug Administration regulations state that blood establishments and transfusion services must assure adequate control of the risk for bacterial contamination of platelets.* Most U.S. blood suppliers fulfill this requirement by performing a primary culture of platelet donations before transfusion (2). Because the risk for platelet transfusion–associated sepsis has persisted despite implementation of primary cultures, additional bacterial mitigation strategies have been implemented, including pathogen-inactivation technology, rapid bacterial detection devices, and alternative culture strategies (2). This report underscores the possibility that sepsis resulting from bacterial contamination of platelets can occur even with application of bacterial contamination mitigation strategies.

The consequences of septic transfusion reactions are often severe morbidity or mortality. In the cluster reported here, one of four patients died, and three recovered only after receiving intensive care. Even with implementation of bacterial contamination mitigation strategies, clinicians should continue to monitor recipients for sepsis after platelet transfusions and immediately report adverse reactions to the platelet supplier and hemovigilance systems.