Laboratory Abnormalities in Former Blood Donors Seropositive for Human T-Lymphotropic Virus Types 1 and 2. A Prospective Analysis

Transcription

1 Laboratory Abnormalities in Former Blood Donors Seropositive for Human T-Lymphotropic Virus Types 1 and 2 A Prospective Analysis Simone A. Glynn, MD; Edward L. Murphy, MD; David J. Wright, PhD; Ronald A. Sacher, MD; Joy Fridey, MD; George B. Schreiber, ScD; for the Retrovirus Epidemiology Donor Study Context. The human T-lymphotropic viruses types 1 and 2 (HTLV-1 and HTLV-2) are highly prevalent among injection drug users in the United States. However, the clinical course of infection has not been well characterized. Objective. To understand HTLV-1 and HTLV-2 associated laboratory abnormalities, which may provide insights into their underlying pathophysiology. Design. Cohort study. Setting. Five US blood centers. Participants. A total of 133 HTLV-1 and 332 HTLV-2 seropositive former blood donors and 717 HTLV-seronegative donors followed up prospectively since Main Outcome Measures. Selected serum chemistry tests and complete blood cell counts were analyzed at enrollment and approximately 2 years later in participants. Repeated-measures analyses were conducted to evaluate the effect of HTLV infection on laboratory measures. Results. Compared with seronegative subjects, HTLV- 1 seropositive subjects had 13% higher creatine kinase (P.02) and slightly elevated lactate dehydrogenase (P ) levels at follow-up. The HTLV-2 seropositive participants had 11% higher absolute lymphocyte counts than seronegative subjects (P.0001). Infection with HTLV-2 also appeared to be associated with slightly higher hemoglobin levels (P ) and hematocrit (P ) and with lower albumin levels (P.01). Conclusions. These results further our understanding of the biological mechanisms underlying HTLV and suggest that HTLV-associated laboratory changes are unlikely to alter clinical evaluation or counseling of otherwise healthy HTLV-infected subjects. (Arch Pathol Lab Med. 2000;124: ) Human T-lymphotropic virus type 1 (HTLV-1) is associated with the development of HTLV-associated myelopathy 1 3 and adult T-cell leukemia (ATL). 1 Likewise, HTLV-associated myelopathy can occur in HTLV-2 infected individuals, who may also be predisposed to infections such as pneumonia and urinary tract infections. 4 These retroviruses can be transmitted through breastfeeding, sexually, by transfusion of blood products, or through the use of contaminated needles. The latter route of transmission seems especially important for HTLV-2, which is highly prevalent in injection drug users. 5 The clinical course of HTLV infection is not well understood. Knowledge of the laboratory test result abnormalities associated with HTLV-1 and HTLV-2 infections can help the clinician better evaluate and counsel patients and can enhance our understanding of the pathophysiology Accepted for publication September 8, From Westat Inc, Rockville, Md (Drs Glynn, Wright, and Schreiber); Department of Laboratory Medicine, University of California, San Francisco (Dr Murphy); Department of Laboratory Medicine, Georgetown University Hospital, Washington, DC (Dr Sacher); and Blood Bank of San Bernardino and Riverside Counties, San Bernardino, Calif (Dr Fridey). Reprints: Simone A. Glynn, MD, MPH, Westat Inc, WB 280, 1650 Research Blvd, Rockville, MD ( GlynnS1@westat.com). associated with these retroviral infections. However, findings from epidemiologic studies that have evaluated the association between HTLV and hematologic or clinical chemistry abnormalities have not been consistent. For example, lymphocyte counts have been reported as increased, 6,7 unchanged, 8 or decreased, 9 and HTLV-1 has been associated with lower 9,10 or conversely higher erythrocyte counts 6 compared with HTLV-negative subjects. It is also unclear whether the elevations in calcium, lactate dehydrogenase (LDH), and creatine kinase (CK) levels observed in subjects with ATL or polymyositis occur, albeit in an attenuated form, in otherwise healthy HTLV-1 or HTLV-2 carriers. Thus, further prospective studies are needed to clarify whether HTLV infection causes significant clinical chemistry, hematologic, or other laboratory changes. To evaluate the effect of HTLV infection on clinical laboratory test measures, we analyzed test results obtained from a cohort of human immunodeficiency virus negative HTLV-1 and HTLV-2 infected and HTLV-negative blood donors whose health status has been prospectively followed up since 1991 as part of the National Heart, Lung, and Blood Institute s Retrovirus Epidemiology Donor Study (REDS). 4 In a previous report, 7 we outlined the clinical hematologic and chemistry characteristics of HTLV- 550 Arch Pathol Lab Med Vol 124, April 2000 Laboratory Abnormalities in HTLV Infection Glynn et al

2 seropositive and HTLV-seronegative donors at study entry. This analysis was, however, limited by its cross-sectional nature and has now been extended to include an evaluation of prospectively obtained laboratory data (collected approximately 2 years after initial laboratory evaluation). MATERIALS AND METHODS Enrollment in the REDS HTLV cohort has been described in detail elsewhere. 4 Blood donors from 5 US blood centers were selected within strata defined by HTLV status, age, sex, race/ ethnicity, type of donation, and center. We chose all HTLV-positive donors and randomly selected seronegative donors within each strata. Complete blood cell counts, sedimentation rate, and selected serum chemistry tests, including calcium, CK, LDH, and alkaline phosphatase measurements, were performed by standard automated analyzers in certified, licensed clinical laboratories contracted by each of the 5 US blood centers involved in the recruitment and follow-up of the cohort. All laboratories fulfilled the quality assurance requirements of licensure. Selected laboratory tests were conducted at study entry and again approximately 2 years later. Laboratories were blinded to HTLV status. The HTLV cohort protocol was approved by the institutional review board at each blood center. Informed consent was obtained from all subjects following full explanation of the nature of the HTLV cohort study. For this exploratory examination of possible changes in laboratory measures associated with HTLV infection, we performed repeated-measures analyses with each laboratory test as the outcome variable and with HTLV status (HTLV-1, HTLV-2, HTLV negative) as the primary independent variable of interest. 11 This technique allowed us to evaluate the association between HTLV infection and laboratory tests at baseline and at follow-up; we were also able to assess the effect of time on test results. To accommodate model assumptions, we used logarithmic transformations of CK and lymphocyte count and square root transformation of the excess in LDH values higher than 300 U/L. The sedimentation rate and absolute eosinophil count could not be analyzed as continuous variables using the method described herein because of poor model fit. These variables were, therefore, categorized (high versus normal using the reference ranges supplied by each laboratory for sedimentation rate and a cutoff of /L for eosinophil count) and their respective association with HTLV status assessed using proc CATMOD. 12 Age, race/ethnicity, sex, education level, income, alcohol consumption, smoking, injection drug use, donation type, center, and follow-up time were considered potential confounders and entered in each model as indicator variables representing predefined groups. Using a backward selection process, variables significantly (P.05) associated with the outcome under consideration were kept in the final adjusted models. Center was considered a potential confounder, since each center contracted with their own laboratory for test analyses. Thus, the center variable described potential differences among laboratories and was retained in each final model regardless of statistical significance to ensure that the resulting adjusted association between HTLV infection and laboratory measure was not attributable to differences among laboratories. We also evaluated whether the effect of HTLV status on each laboratory test was similar at both visits by testing whether the interaction term between HTLV status and initial-versus-follow-up visit was statistically significant. Each repeated-measures model produces parameter estimates ( ) as part of its output. These parameters estimate the magnitude of the association between each independent variable (HTLV status being the one of main interest) and the outcome (a laboratory measure). If the effect of HTLV on a laboratory measure is similar at the initial and follow-up visits, the interaction term of HTLV status and initial-versus-follow-up visit is not significant (P.05), and a single parameter estimate is obtained to describe the association between HTLV and the laboratory measure. For example, the adjusted repeated-measures model that evaluated the effect of HTLV-2 on white blood cell count ( 10 9 /L) revealed a nonsignificant time HTLV-2 interaction term (P.11) and a of.260. Stated differently, HTLV- 2 seropositive donors had, on average, 260 ( /L) more white blood cells per microliter than seronegative donors; this relative increase in white blood cells among HTLV-2 donors was present both at the initial visit and approximately 2 years later. Interpretation of parameter estimates obtained in a model where the outcome has been logarithmically transformed is slightly different. In the adjusted repeated-measures model for lymphocyte count, where we used logarithmic transformation, we obtained a of.103 for HTLV-2. This can be interpreted as an 11% increase (e ) in absolute lymphocyte count among HTLV-2 donors compared with HTLV-negative donors at both initial and follow-up visits. A significant initial-versus-follow-up visit interaction term (P.05) implies that the association between HTLV status and the laboratory measure of interest is not of the same magnitude or is in opposite direction at initial and follow-up visits and that these differences are unlikely to be explained by chance alone. In this case, 2 parameter estimates need to be derived to describe the magnitude (on average) of the effect of HTLV on the laboratory outcome: one for the initial visit and one for the follow-up visit. RESULTS At least one laboratory test was performed in 133 HTLV- 1, 332 HTLV-2, and 717 HTLV-seronegative donors at initial and follow-up visit. There were no instances of HTLV- 1 and HTLV-2 coinfection. Eleven percent of the 1334 donors initially enrolled (with at least one blood test result at baseline) did not return for a follow-up blood test; these 152 subjects (78 HTLV negative, 20 HTLV-1, 54 HTLV-2) were more likely to be black, to be less educated, and to have a history of injection drug use (P.05). As shown in Table 1, mean laboratory test results obtained at visit 1 were similar within HTLV-type category for subjects with 2 visits and those who did not return (t test P.05). HTLV-1 and HTLV-2 subjects were more likely to have a lower level of education than seronegative subjects (P.05). Furthermore, HTLV-2 subjects were more likely to have a lower income and to report previous injection drug use, smoking, and heavy alcohol consumption than seronegative subjects (P.05). The mean laboratory measures for visits 1 and 2 are presented in Table 2. Most subjects laboratory test results were within normal limits at both the initial and follow-up visit (ie, within the reference ranges provided by each laboratory). Table 3 shows the single adjusted parameter estimates obtained to describe the impact of HTLV infection on a laboratory test if the effect of HTLV on that laboratory measure was similar at baseline and at followup. Conversely, we show in Table 4 the 2 adjusted parameter estimates (one for each visit) derived to characterize those associations that differed at initial and follow-up visits; namely, HTLV-1/platelet count, HTLV-1/CK, HTLV-1/ albumin, HTLV-2/mean corpuscular volume, HTLV-2/ platelet count, HTLV-2/CK, HTLV-2/LDH, and HTLV-2/ calcium. The following discussion concentrates on results obtained after adjustment for demographic and other variables. Although all adjustment factors were significantly associated with the outcome variable (or laboratory measure), adjustment did not appear to significantly alter results except for HTLV-2, where the hemoglobin and hematocrit parameter estimates increased approximately twofold (from 0.97 to 1.51 for hemoglobin and from to for hematocrit), and the white blood cell count estimate decreased by 50% with adjustment (from to 0.260). Arch Pathol Lab Med Vol 124, April 2000 Laboratory Abnormalities in HTLV Infection Glynn et al 551

3 Table 1. Laboratory Tests White blood cell count, 10 9 /L Lymphocyte count, 10 9 /L Neutrophil count, 10 9 /L Hemoglobin, g/l Hematocrit Mean corpuscular volume, fl Platelet count, 10 9 /L Creatine kinase, U/L Alkaline phosphatase, U/L Lactic dehydrogenase, U/L Albumin, g/l Calcium, mmol/l Mean Laboratory Measures at Visit 1 by Human T-Lymphotropic Virus (HTLV) Status Comparing Lost to Follow-up to Remaining Cohort Subjects Cohort, (n* 717) 6.46 (0.07) 2.09 (0) 3.74 (0.05) (0.5) (0.21) (2.46) (3.79) (0.88) (1.34) 43.4 (0.1) 2.36 (0.005) Lost to Follow-up, (n 78) P 6.46 (0.20) 2.16 (0.08) 3.64 (0.15) (1.5) 0.42 (0.005) (0.73) (8.38) (14.32) (3.07) (4.63) 43.5 (0.3) 2.35 (0.01) * n represents the number of subjects with at least 1 test available and will vary by test. t tests were conducted. Cohort, (n 133) 6.49 (0.17) 2.15 (0.06) 3.79 (0.14) (1.2) 0.41 (0.003) (0.54) (6.39) (8.31) (3.07) (3.90) 43.4 (0.3) 2.34 (0.01) HTLV-1 Lost to Follow-up, (n 20) P 6.58 (0.38) 2.35 (0.13) 3.70 (0.31) (3.8) 0.41 (0.01) (1.52) (16.26) (13.61) (7.90) (13.70) 42.0 (0.6) 2.33 (0.02) Compared with seronegative donors, HTLV-1 infected subjects were more likely to have higher LDH levels at both initial and follow-up visit (average increase of 6.3 U/ L [this estimate strictly applies only to subjects with LDH values 300 U/L], P, Table 3). Similar results were obtained after removal of 1 HTLV-1 seropositive subject diagnosed as having ATL. Although not significantly associated with HTLV-1 infection at enrollment (P.21), CK was 13% higher in HTLV-1 seropositive than in seronegative donors at follow-up (P.02, Table 4). The HTLV- 1 donors also tended to have higher lymphocyte counts, but the 2% difference observed was small and not significant (P.48, Table 3). The previously reported higher platelet counts (on average) observed at enrollment 7 among HTLV-1 infected subjects (average difference of / L, P.003) did not persist at follow-up (average difference of 4000/ L, P.55, Figure 1). The HTLV-1 infected subjects were found, on average, to experience a large unexplained decline in platelet counts between baseline and follow-up visits (average difference of / L, P.0001). The HTLV-2 seropositive donors tended to have, on average, 260 more white blood cells per microliter (P ) and an 11% increase in absolute lymphocyte count (P.0001) compared with seronegative donors (Table 3 and Figure 2). Also, HTLV-2 infection was associated with average increases of 1.5 g/l and in hemoglobin and hematocrit, respectively (P, Table 3), and with a slight decrease in albumin level (0.5 g/l) compared with seronegative donors (P.01, Table 3). The effect of HTLV- 2 on these laboratory measures did not vary with time. In contrast, whereas HTLV-2 was not associated with changes in LDH at the enrollment visit (P.48), LDH appeared to be significantly higher in HTLV-2 subjects compared with seronegative donors at follow-up (average difference of 5.1 U/L [this estimate strictly applies only to subjects with LDH values 300 U/L] Table 4, P ). We had previously reported that HTLV-2 infected subjects had, on average, higher platelet counts (19 000/ L excess, P.0001), higher mean corpuscular volume (1.1-fL excess, P.005), and lower serum calcium levels (0-mmol/L decrease, P.0001) than seronegative donors at baseline. 7 These differences did not persist at the follow-up visit: 4000/ L increased platelet count (P.36, Figure 1), fl increase for mean corpuscular volume (P.26, Table 4), and mmol/L increase in calcium (P.79, Table 4). Our analyses did not show any significant association between HTLV infection and other test results. COMMENT This study suggests that HTLV infection may be associated with minor changes in laboratory test results even in the absence of diseases such as ATL or polymyositis. Some laboratory changes observed at baseline became more pronounced with time, whereas others did not persist. Most laboratory differences were small and would be unlikely to change clinical evaluation or the counseling or monitoring of otherwise healthy HTLV-infected individuals. However, these changes are of scientific interest in that they may reflect the underlying pathophysiology associated with HTLV infection. The increased lymphocyte count and correspondingly higher total white blood cell count observed among HTLV- 2 infected subjects may be secondary to HTLV-induced upward-regulation of interleukin 2 and its receptor, with subsequent lymphocytic proliferation. 13 The ability to mount a proper immune response against certain pathogens may be consequently diminished, as suggested by findings of an increased prevalence and incidence of infectious diseases among HTLV-2 seropositive subjects in this and other studies. 4,14,15 The HTLV-1 donors also tended to have higher lymphocyte counts, although the difference was not significant (2%; 95% confidence interval, 3% to 7%). We previously reported that HTLV-1 subjects had, on average, 6% more lymphocytes than seronegative subjects (P ; 95% confidence interval, 1% to 11%), a result based on analysis of only baseline laboratory data. 7 Our present results are consistent with this prior cross-sectional analysis (the confidence intervals overlap) and suggest that lymphocyte counts are slightly elevated, if at all, among HTLV-1 subjects. The HTLV-1 and HTLV-2 seropositive subjects appeared to have higher platelet counts than seronegative subjects at study entry but not 2 years later. We previously postulated that HTLV-mediated transactivation of the tax gene may have induced cytokine-mediated platelet proliferation. However, it is unlikely that such a phenomenon, 552 Arch Pathol Lab Med Vol 124, April 2000 Laboratory Abnormalities in HTLV Infection Glynn et al

4 Cohort, (n 332) 6.90 (0.11) 2.38 (0.05) 3.88 (0.09) (0.8) (0.31) (4.27) (6.47) (1.28) (2.48) 43.2 (0.2) Table 1. HTLV-2 Extended Lost to Follow-up, (n 54) P 6.88 (0.28) 2.47 (0.14) 3.77 (0.20) (2.1) 0.42 (0.006) (0.78) (10.01) (9.33) (3.36) (5.24) 42.7 (0.4) 2.33 (0.02) even if cyclical in nature, could explain our findings. These results may, however, reflect minor changes in laboratory techniques, because the observed decrease in platelet counts was small and occurred in both HTLV-positive and HTLV-negative subjects. Continued follow-up and additional collection of prospectively obtained laboratory information may help clarify the nature of the association between platelet counts and HTLV infection. Both HTLV-1 and HTLV-2 infected donors tended to have higher levels of LDH than seronegative subjects (at baseline and follow-up for HTLV-1, solely at follow-up for HTLV-2). LDH is a marker of increased leukocyte turnover 16 and may also correlate with levels of C-terminal parathyroid-related hormone that have been found increased in otherwise healthy HTLV-1 carriers. 17 The association between HTLV-2 and lower albumin levels was observed at both baseline and follow-up visits and may be consistent with the lower socioeconomic status of HTLV-2 subjects, although we attempted to adjust for socioeconomic status in the multivariable model. The significance of the slight alterations in hemoglobin and hematocrit observed in HTLV-2 subjects is unclear. Finally, additional data will be needed to assess whether the increase in CK levels observed in HTLV-1 seropositive subjects during follow-up represents a trend toward subclinical neuromuscular disease in these subjects. The strengths of this study reside in its prospective nature and the ability to evaluate changes in laboratory tests Table 3. Adjusted Mean Differences Between Human T-Lymphotropic Virus (HTLV)-1 and HTLV-Negative or HTLV-2 and HTLV-Negative Subjects for Laboratory Measures With No Temporal Effect* Laboratory Tests White blood cell count Lymphocyte count Neutrophil count Hemoglobin Hematocrit Mean corpuscular volume Alkaline phosphatase Lactic dehydrogenase Albumin Calcium HTLV-1 Versus P HTLV-2 Versus P * Models also contained the following variables: for white blood cell count, age race, sex, alcohol, smoking (pack-years), income, and center; for lymphocyte count, race, sex, alcohol, smoking (pack-years), income, and center; for neutrophil count, age, race, sex, center, donation type, income, smoking (pack-years), and alcohol; for hemoglobin, race, sex, donation type, center, and injection drug use; for hematocrit, race, sex, center, injection drug use, and income; for alkaline phosphatase, age, race, center, income, and follow-up time; for lactic dehydrogenase, age, race, center, and injection drug use; for albumin, age, race, sex, and center; for mean corpuscular volume, age, race, alcohol, center, and income; and for calcium, age, race, center, and smoking (pack-years). The effect of HTLV on laboratory measures is similar at both visits. Natural logarithmic transformation used. The effect of HTLV-1 on albumin and the effect of HTLV-2 on lactic dehydrogenase, mean corpuscular volume, and calcium were different at the baseline and follow-up visits, and a single parameter estimate would not properly describe the association between HTLV-1 or HTLV- 2 and these laboratory measures (see Table 4). Square root transformation of the excess in lactic dehydrogenase values greater than 300 U/L. through time using repeated-measures analyses. Loss to follow-up in this cohort was small, although extrapolation of these results could be potentially limited by the 50% initial enrollment rate among all HTLV-seropositive blood donors. We may also have underestimated the prevalence of some laboratory abnormalities, since subjects with overt health problems or anemia could have been excluded (regardless of their HTLV status) by the predonation health screening questionnaire administered to all blood donors before donating. Moreover, because HTLV in the United States is mostly acquired though sexual transmission or Table 2. Laboratory Tests White blood cell count, 10 9 /L Lymphocyte count, 10 9 /L Neutrophil count, 10 9 /L Hemoglobin, g/l Hematocrit Mean corpuscular volume, fl Platelet count, 10 9 /L Creatine kinase, U/L Alkaline phosphatase, U/L Lactic dehydrogenase, U/L Albumin, g/l Calcium, mmol/l Mean Laboratory Measures by Visit and Human T-Lymphotropic Virus (HTLV) Status Visit 1, 6.46 (0.07) 2.09 (0) 3.74 (0.05) (0.5) (0.21) (2.46) (3.79) (0.88) (1.34) 43.4 (0.1) 2.36 (0.005) Visit 2, 6.48 (0.07) 2 (0) 3.79 (0.05) (0.5) (0.21) (2.49) (3.52) (1.00) (1.46) 44.3 (0.1) Visit 1, 6.49 (0.17) 2.15 (0.06) 3.79 (0.14) (1.2) 0.41 (0.003) (0.54) (6.39) (8.31) (3.07) (3.90) 43.4 (0.3) 2.34 (0.01) Arch Pathol Lab Med Vol 124, April 2000 Laboratory Abnormalities in HTLV Infection Glynn et al 553 HTLV-1 Visit 2, 6.41 (0.17) 2.02 (0.06) 3.74 (0.13) (1.2) 0.41 (0.003) (0.54) (5.49) (11.69) (2.30) (4.01) 43.2 (0.3) 2.33 (0.01) Visit 1, 6.90 (0.11) 2.38 (0.05) 3.88 (0.09) (0.8) (0.31) (4.27) HTLV (6.47) (1.28) (2.48) 43.2 (0.2) Visit 2, 7.09 (0.13) 2.38 (0.05) 4.01 (0.10) (0.8) 0.42 (0.002) (0.32) (4.22) (4.97) (1.39) (3.20) 43.6 (0.2)

5 Table 4. Mean corpuscular volume Platelet count Creatine kinase Lactic dehydrogenase Albumin Calcium Adjusted Mean Differences Between Human T-Lymphotropic Virus (HTLV)-1 and HTLV-Negative or HTLV-2 and HTLV-Negative Subjects for Laboratory Measures With a Temporal Effect* Adjusted for HTLV-1 Versus HTLV-Negative Visit 1 Visit 2 P Adjusted for HTLV-2 Versus HTLV-Negative Visit 1 Visit 2 P * Models also contained the following variables: for platelet count, sex and center; for creatine kinase, race, sex, donation type, center and smoking; for mean corpuscular volume, age, race, alcohol, center, and income; for calcium, age, race, center, and smoking; for lactic dehydrogenase, age, race, center, and injection drug use; for albumin, age, race, sex, and center. The effect of HTLV on laboratory measures differs between visits. In contrast to the P values listed in the text, which test if the laboratory measures are similar among HTLV-1 (or HTLV-2) and HTLV-negative subjects, the P values presented herein test whether the effect of HTLV infection is the same on laboratory measures in visit 1 and visit 2. The effect of HTLV-1 on lactic dehydrogenase, mean corpuscular volume, and calcium and the effect of HTLV-2 on albumin are similar at both visits (see Table 3). Natural logarithmic transformation used. Square root transformation of the excess in lactic dehydrogenase values greater than 300 U/L. Figure 1. Platelet count by human T-lymphotropic virus (HTLV) status and follow-up time. Counts represent directly standardized counts, where the standardization population is the Retrovirus Epidemiology Donor Study population. injection drug use (rather than perinatally), cohort participants might still be in an early phase of infection with insufficient time for clinical changes to have become manifest. In addition, although all models were adjusted for center as a surrogate for potential differences among laboratories, we did not evaluate whether laboratory techniques changed over time, considering the short follow-up period and the additional complexity of a model needed to evaluate such a change (an interaction term between center and initial-versus-follow-up visit would need to be tested). Only stratification variables and potentially important confounders were considered for adjustment, but we cannot exclude the possibilities of underadjustment for unknown confounders; adjustment did not affect results, except for the associations between HTLV-2 and hemoglobin level or hematocrit. Finally, considering the multiple comparisons performed in this exploratory analysis, results with significance levels of.01 or more need to be interpreted with caution and should be confirmed by further studies. In conclusion, otherwise healthy HTLV-positive subjects do not appear to have clinically significant changes in routinely performed hematologic and clinical chemistry tests. Figure 2. Absolute lymphocyte count by human T-lymphotropic virus status and followup time. Directly standardized measures are presented, where the standardization population is the Retrovirus Epidemiology Donor Study population. Natural logarithmic transformation of lymphocyte counts was performed for the repeated-measures models. 554 Arch Pathol Lab Med Vol 124, April 2000 Laboratory Abnormalities in HTLV Infection Glynn et al

6 The HTLV-2 infected donors tended to have higher lymphocyte counts than seronegative donors. The association between platelet counts and HTLV-1 or HTLV-2 infection that was present at baseline did not persist on further follow-up of the cohort. These results exemplify the hypothesis-generating nature of cross-sectional studies and the importance of longitudinal measurements when trying to evaluate the effect of an infectious agent on clinical laboratory outcomes. Continued follow-up of the cohort with additional collection of laboratory data may further our understanding of the clinical course of HTLV infection, associated pathologic processes, and related clinical hematologic and chemistry changes. This work was supported by contracts NO1-HB (superseded by 47114), , ,-97080, , and from the National Heart, Lung, and Blood Institute. We thank Drew Hayes at Westat Inc for his assistance with computer programming. The Retrovirus Epidemiology Donor Study (REDS Study Group) is currently the responsibility of the following persons: Blood Centers: American Red Cross Blood Services Greater Chesapeake and Potomac Region: A. E. Williams (Holland Laboratory), C. C. Nass; American Red Cross Blood Services Southeastern Michigan Region: M. Higgins, J. Campbell; American Red Cross Blood Services Southern California Region: G. Garratty, S. Hutching; Blood Centers of the Pacific-Irwin Centers: E. L. Murphy (UCSF), M. P. Busch; Oklahoma Blood Institute: R. O. Gilcher, J. W. Smith; Medical Coordinating Center: Westat Inc: G. B. Schreiber, R. A. Thomson; National Heart, Lung, and Blood Institute, National Institutes of Health: G. J. Nemo; Steering Committee Chairman: T. F. Zuck (Hoxworth Blood Center). References 1. Hollsberg P, Hafler DA. Pathogenesis of diseases induced by human lymphotropic virus type I infection. N Engl J Med. 1993;328: Murphy EL, Fridey J, Smith JW, et al, and the REDS investigators. HTLVassociated myelopathy in a cohort of HTLV-I and HTLV-II infected blood donors. Neurology. 1997;48: Lehky TJ, Flerlage N, Katz D, et al. Human T cell lymphotropic virus type II-associated myelopathy: clinical and immunologic profiles. Ann Neurol. 1996; 40: Murphy EL, Glynn S, Fridey J, et al. Increased prevalence of infectious diseases and other adverse outcomes in HTLV infected blood donors. J Infect Dis. 1997;176: Khabbaz RF, Onorato IM, Cannon RO, et al. Seroprevalence of HTLV-I and HTLV-II among intravenous drug users and persons in clinics for sexually transmitted diseases. N Engl J Med. 1992;326: Welles SL, Tachibana N, Orav EJ, et al. Changes in hematologic parameters among Japanese HTLV-I carriers. J Acquir Immune Defic Syndr Hum Retrovirol. 1994;7: Murphy EL, Glynn S, Watanabe K, et al. Laboratory test differences associated with HTLV-I and HTLV-II infection. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;17: Prince HE. American blood donors seropositive for human T-lymphotropic virus types I/II exhibit normal lymphocyte subsets. Transfusion. 1990;30: Ho GYF, Nelson K, Nomura A, et al. Markers of health status in an HTLV- I positive cohort. Am J Epidemiol. 1992;136: Murphy EL, Wilks RJ, Morgan OS, et al. Health effects of human T lymphotropic virus type I (HTLV-I) in a Jamaican cohort. Int J Epidemiol. 1996;25: SAS Institute Inc. SAS/STAT Software: Changes and Enhancements, Release Cary, NC: SAS Institute Inc; 1992: SAS Technical Report P SAS Institute Inc. SAS/STAT User s Guide, Version 6, Vols 1 and 2. 4th ed. Cary, NC: SAS Institute Inc; Lal RB, Hjelle B, Rudolph D. Spontaneous proliferation of HTLV-II infected peripheral blood lymphocytes: HLA-DR-driven, IL-2-dependent response. Microbiol Immunol. 1992;36: Guroff RM, Weiss SH, Giron JA, et al. Prevalence of antibodies to HTLV- I,-II and III in intravenous drug abusers from an AIDS endemic region. JAMA. 1986;255: Modahl LE, Young KC, Varney KF, et al. Are HTLV-II-seropositive injection drug users at increased risk of bacterial pneumonia, abcess and lymphadenopathy? J Acquir Immune Defic Syndr Hum Retrovirol. 1997;16: Pan L, Beverley PCL, Isaacson PG. Lactate dehydrogenase (LDH) isoenzymes and proliferative activity of lymphoid cells- an immunocytochemical study. Clin Exp Immunol. 1991;86: Yamaguchi K, Kiyokawa T, Watanabe T, et al. Increased serum levels of C- terminal parathyroid hormone-related protein in different diseases associated with HTLV-1 infection. Leukemia. 1994;8: Arch Pathol Lab Med Vol 124, April 2000 Laboratory Abnormalities in HTLV Infection Glynn et al 555