The Current Risk of Retroviral Infections Transmitted by Transfusion in Patients Who Have Undergone Multiple Transfusions FREE

AFTER THE introduction of blood donor screening for antibodies to human immunodeficiency virus (HIV), the risk of acquiring HIV infection after transfusion has been mainly confined to transfusion of infected blood collected either during the window period (ie, the period between infection and the development of anti-HIV antibodies)^1- 3 or from infected subjects with negative test results using current serologic assays.^4- 5 Recently, this risk has been estimated to be low using a mathematical model based on the duration of the window period and the incidence of infection among repeated donors.^6- 7 However, information on the actual rate of seroconversion in transfusion recipients is limited.

Other retroviruses, including human T-lymphotropic virus (HTLV), can be transmitted through blood transfusion.^8- 9 However, the epidemiological factors of transfusion-associated HTLV infection have not been completely defined. More information is needed, particularly from countries where anti-HTLV screening is not mandatory.

To collect experimental evidence of the risk of blood-borne retroviral infections, in 1989 we started a monitoring survey within the Cooleycare Cooperative Group, a program of centers in charge of approximately two thirds of Italian patients with β thalassemia.¹⁰ Such patients represent an informative cohort, since they undergo regular transfusions every 2 to 4 weeks throughout their lifetime.¹⁰ Moreover, transfusion treatment for homozygous β thalassemia is started early in life, and most patients are children, who have limited chances to acquire HIV infection through routes different from blood transfusion compared with adults. A report¹¹ based on the analysis of the baseline data of our patients was published elsewhere. Following this initial investigation, a prospective longitudinal study continued until the beginning of 1996. In this article, we present the rate of retroviral infections acquired through blood transfusion in this large cohort of patients with β thalassemia during a 6-year follow-up.

In 1989, the centers of the Cooleycare Cooperative Group were invited to participate in a prospective longitudinal study aimed at assessing the risk of blood-borne infections. Thirty-six centers agreed to participate, and all the 1384 patients who were homozygous for β thalassemia and underwent regular transfusion at these centers were enrolled (716 males, 668 females; median age, 16 years; range, 0-45 years). For each patient, a serum sample collected in the period from December 1989 to March 1990 (baseline sample) and a record including demographic data and information on the transfusion regimen were sent to Milan, Italy, where the reference laboratory of the Cooleycare Cooperative Group is located. Additional serum samples of the same patients were collected from the participating centers at 3-year intervals after baseline (1992-1993 and 1995-1996). The number of blood units transfused to the cohort was estimated on the basis of the experimental evidence¹⁰ of a mean annual red blood cell requirement (pure red blood cells with hematocrit equal to 1.0) of 180 and 137 mL/kg in patients who did not undergo splenectomy and those who had undergone splenectomy, respectively. Furthermore, we assumed that a 450-mL unit of blood contains on average 180 mL of red blood cells with a hematocrit of 1.0 and approximately 10% of this amount is lost due to the white blood cell reduction procedure routinely performed for recipients with β thalassemia in Italy.

Patients' serum samples were screened as follows. For HIV, an enzyme immunoassay for the detection of antibodies to HIV-1 and HIV-2 (Murex, Dartford, England) was used. Since 1995, a test able to detect antibodies to HIV-1 subtype O was used (Ortho Diagnostic Systems, Raritan, NJ). Antibodies to HTLV were detected by a particle agglutination assay (Fujirebio, Tokyo, Japan) or enzyme immunoassay (Cambridge Biotech, Ballybrit, Ireland, or Murex). Confirmatory tests were performed on samples repeatedly reactive to anti-HIV antibodies using Western blot analysis (Blot 2.2, Diagnostic Biotechnology Limited, Singapore) and by a line immunoassay for anti-HIV subtype O antibodies (LIA, Innogenetics N.V., Zwijnaarde, Belgium). Reactivities to HIV-2 were further evaluated by a specific Western blot analysis (Diagnostic Biotechnology Limited). For confirmation of anti-HTLV antibodies, Western blot analysis (Blot 2.4, Diagnostic Biotechnology Limited) was used. In the patients who experienced anti-HIV seroconversion, the available serum samples preceding seroconversion were retested using a third-generation screening test (Ortho Diagnostic Systems) and a confirmation assay (Diagnostic Biotechnology Limited).

The incidence of infection was expressed as number of new infections per 10000 person-years. The risk of acquiring infection was computed by the ratio between the number of patients who experienced seroconversion and the total number of red blood cell units transfused to the patient group during the study period (December 1989 to March 1996). The upper boundary of the 95% confidence interval (CI) of incidence and risk was calculated using the exact binomial method. The χ² test was used to compare proportions.

Of the 1384 patients enrolled in 1989-1990, samples from 1073 patients at 31 centers were obtained from December 1992 to March 1993 and from 1001 patients at 26 centers from December 1995 to March 1996. A total of 383 patients were unavailable for follow-up. During 1992-1993, of the 1073 patients evaluated, 311 patients were unavailable for follow-up (66 patients died [in 6 of these patients, the serum samples were positive for anti-HIV antibodies], 35 underwent bone marrow transplantation, 193 had unavailable samples [in 2 of these patients, the serum samples were positive for anti-HIV antibodies], and 17 were transferred to other health care units). During 1995-1996, of the 1001 patients evaluated, 72 were unavailable for follow-up (50 died [in 4 of these patients, the serum samples were positive for anti-HIV antibodies], 20 underwent bone marrow transplantation, and 2 had unavailable samples). The cohort of 1001 patients included 504 males and 497 females, with a median age at enrollment of 15 years (range, 0-45 years).

The prevalence of HIV infection calculated in the baseline study was 40 (2.9%) of 1384. The infection was acquired before 1985 in 36 of the 40 patients. In 4 patients, seroconversion occurred after the implementation of HIV-1 screening in blood donors (in 1987 in 2 patients and in 1988 and 1989 in 1 patient each).¹¹ Ten HIV-infected patients (25%) died during the study period. During the 6-year follow-up, 1 (0.09%) of the 1001 patients acquired HIV infection. This patient was a 20-year-old woman who experienced seroconversion to anti-HIV in 1994. When interviewed by her physician, the woman denied any past exposure to risk factors other than blood transfusion, such as occasional or habitual drug abuse, sexual promiscuity, or engaging in casual sexual relationships. The incidence of HIV infection was 1.7 per 10000 person-years (upper boundary of the 95% CI, 5 per 10000). Since approximately 190000 red blood cell units were administered to the patients of the cohort during the study period, the residual risk of blood-borne HIV infection was 1 in 190000 U (upper boundary of the 95% CI, 1 in 67000) (Table 1). No infection was due to HIV-1 subtype O or HIV-2.

The results in 4 patients (0.3%) were positive for anti-HTLV antibodies in the baseline study. Of these, 3 were infected with HTLV-1 and 1 with HTLV-2. These patients did not show neurologic or clinical symptoms related to viral infection. The pattern of reactivity determined in the results of Western blot analysis remained unchanged during the study period. Because no seroconversions were observed during the period 1989-1996, the risk of blood-borne HTLV infection was estimated to be less than 1 in 190000 red blood cell units. Considering all the units received by the patients of this cohort since early childhood (approximately 640000), the risk was 1 in 160000 (upper boundary of the 95% CI, 1 in 80000) (Table 1).

Recent estimates of the risk of transmitting retroviral infections by the transfusion of screened blood in the United States are reassuring.^6- 7 However, experimental evidence on the actual risk of acquiring blood-borne infections is lacking mainly because large numbers of blood recipients need to be tested to reach reliable conclusions.¹² To overcome this problem, we based our studies on patients who had undergone multiple transfusions, such as patients homozygous for β thalassemia. These patients represent an ideal sentinel population since they regularly receive blood transfusions from early childhood.¹⁰ In addition, the age of the majority of the patients limits the potential bias due to other risk factors for retroviral infection, such as illicit drug use and sexual activities.¹³

At the end of the 6-year study period of this investigation, more than 70% of the patients originally enrolled were still undergoing follow-up. We observed that 8% of the patients had died, accounting for a mortality rate not dissimilar to that reported in other series of Italian patients homozygous for β thalassemia.^14- 15

We found that the prevalence of HIV infection in this series during 1989-1990 was comparable with that observed in similar patient groups evaluated in other European countries.¹⁶ As previously reported, these infections are mainly due to the transfusion of infected blood in the years before blood was screened (ie, before 1985), rather than to a failure of first-generation anti-HIV assays in identifying seropositive donors.¹¹ The observation of 1 case of HIV seroconversion during the study period shows that the current risk of transfusion-associated transmission in Italy is 1 in 190000 blood units (upper boundary of the 95% CI, 1 in 67000). In comparison with the figure reported in a similar study from the United States a few years before (1 in 60000; upper boundary of the 95% CI, 1 in 19000),¹² our data suggest a trend toward a reduction of the risk. It is probable that the results were due to the introduction of more sensitive serologic assays and the adoption of safer policies for donor selection during our study.^3,17- 21 In addition, our analysis showed that the risk of acquiring HIV infection through blood transfusion is slightly higher than that estimated using a mathematical model based on the incidence of infection among repeated blood donors (1 in 493000; upper boundary of 95% CI, 1 in 200000).⁷ This finding could be due to a possible underestimation of the risk using the incidence model, since seroconversion in one-time blood donors could not be included in this model.⁷

A small proportion of patients had HTLV infection at the baseline evaluation. Moreover, we did not observe cases of seroconversion during the follow-up period. Although our patients underwent transfusion with blood unscreened for HTLV antibodies during their lifetime, the risk of HTLV infection appears to be lower than that observed in the United States both before and after the introduction of the screening for blood donors.^12,22- 23 Two possible explanations for this finding are the low frequency of HTLV infection among blood donors in Italy²⁴ and the routine administration of leukocyte-reduced red blood cells to patients homozygous for β thalassemia, which seems to transmit HTLV less efficiently than standard cellular blood components.^25- 26

In conclusion, our study provides experimental evidence that the application of reliable screening procedures for donor selection reduced the transmission of HIV infection through blood transfusion in 1989-1995 to less than 2 cases in 10000 person-years or 1 case per 190000 U of red blood cells transfused. In addition to reassuring transfusion recipients and health care providers, these data can be useful in discussion of the cost-effectiveness of introducing supplemental assays for donor screening.

Accepted for publication October 9, 1997.

Supported in part by a grant from the Italian National Institute of Health, Rome.

The following members of the Cooleycare Cooperative Group in Italy provided samples for this study.

Agrigento: G. Malfitano, MD; Bari: F. Schettini, MD; Bergamo: P. Bellavita, MD; Casarano: E. Corvaglia, MD; Catania: M. Alessi, MD, F. Di Gregorio, MD, C. Magnano, MD; Como: R. Longhi, MD; Genova: A. Rasore-Quartino, MD; Iglesias: A. Biolchini, MD; Locri: D. Costantino, MD; Magenta: S. Calò, MD; Matera: V. Cilla, MD; Messina: A. Meo, MD; Milan: A. Zanella, MD; Monselice: S. Montin, MD; Monza: W. Monguzzi, MD, G. Sciorelli, MD; Naples: C. De Rosa, MD; Olbia: M. Lendini, MD; Oristano: A. Cambosu, MD, A. Carta, MD; Palermo: P. Di Paola, MD; Ragusa: P. Rizzone Favacchio, MD; Rho: F. Betto, MD; Roma: G. Girelli, MD; Rovigo: E. Cichella, MD; San Gavino Monreale: G. Batzella, MD; Sassari: G. Bertrand, MD; Siracusa: A. Mangiagli, MD, D. Gallisai, MD; and Verona: C. Borgna, MD.

Reprints: Daniele Prati, MD, Blood Transfusion and Transplantation Immunology Center, Istituto di Ricovero e Cura a Carattere Scientifico, Ospedale Maggiore, Via Francesco Sforza, 35, 20122 Milan, Italy.