Anti-factor VIII antibodies in brothers with haemophilia A share similar characteristics.

INTRODUCTION: The development of neutralizing antibodies (inhibitors) against coagulation factor VIII (FVIII) is currently the most serious complication for patients with haemophilia A undergoing FVIII replacement therapy. Several genetic factors have been acknowledged as risk factors for inhibitor development. AIM: To analyze the influence of genetic factors on the nature of the humoral immune response to FVIII in eight brother pairs with inhibitors. METHODS: The domain specificity of FVIII-specific IgG was analysed by antibody binding to FVIII fragments and homologue-scanning mutagenesis (HSM). The FVIII-specific IgG subclasses were measured by direct ELISA. RESULTS: Of the 16 patient analysed with both methods, 12 had A2- and 13 had C2-specific IgG. The presence of A1-, A3- or C1-specific IgG was identified in nine of 14 patients analysed by HSM. IgG1, IgG2 and IgG4 subclasses contributed to the anti-FVIII IgG response, and the amount of FVIII-specific IgG1 (r = 0.66) and IgG4 (r = 0.69) correlated significantly with inhibitor titres. Patients with high concentrations of total anti-FVIII IgG (r = 0.69) or high inhibitor titres (r = 0.52) had a high proportion of FVIII-specific IgG4. Statistical analysis revealed trends/evidence that the subclass distribution (P = 0.0847) and domain specificity to HC/LC (P = 0.0883) and A2/C2 (P = 0.0011) of anti-FVIII IgG were more similar in brothers compared to unrelated subjects. CONCLUSION: Overall, our data provide a first hint that anti-FVIII IgG characteristics are comparable among haemophilic brothers with inhibitors. Whether genetic factors also influence the nature of patients' antibodies needs to be confirmed in a larger study population.

F8 haplotype and inhibitor risk: results from the Hemophilia Inhibitor Genetics Study (HIGS) Combined Cohort.

Ancestral background, specifically African descent, confers higher risk for development of inhibitory antibodies to factor VIII (FVIII) in haemophilia A. It has been suggested that differences in the distribution of FVIII gene (F8) haplotypes, and mismatch between endogenous F8 haplotypes and those comprising products used for treatment could contribute to risk. Data from the Hemophilia Inhibitor Genetics Study (HIGS) Combined Cohort were used to determine the association between F8 haplotype 3 (H3) vs. haplotypes 1 and 2 (H1 + H2) and inhibitor risk among individuals of genetically determined African descent. Other variables known to affect inhibitor risk including type of F8 mutation and human leucocyte antigen (HLA) were included in the analysis. A second research question regarding risk related to mismatch in endogenous F8 haplotype and recombinant FVIII products used for treatment was addressed. Haplotype 3 was associated with higher inhibitor risk among those genetically identified (N = 49) as of African ancestry, but the association did not remain significant after adjustment for F8 mutation type and the HLA variables. Among subjects of all racial ancestries enrolled in HIGS who reported early use of recombinant products (N = 223), mismatch in endogenous haplotype and the FVIII proteins constituting the products used did not confer greater risk for inhibitor development. Haplotype 3 was not an independent predictor of inhibitor risk. Furthermore, our findings did not support a higher risk of inhibitors in the presence of a haplotype mismatch between the FVIII molecule infused and that of the individual.

Extended IL10 haplotypes and their association with HIV progression to AIDS.

Interleukin-10 (IL-10) is a pleiotropic cytokine with both immunosuppressive and immunostimulatory functions. Its roles in infections and autoimmunity may have resulted in selective pressures on polymorphisms within the gene, leading to genomic coexistence of several semi-conserved haplotypes involved with diverse pathogen interactions during genomic evolution. Previous studies focused either exclusively on promoter haplotypes or on individual SNPs. We genotyped 21 single nucleotide polymorphisms in the human IL10 gene and examined this variation compared to other mammalian species sequences. Haplotype heterogeneity in human populations is centered around 'classic' 'proximal' promoter polymorphisms: -592, -819 and -1082. High-producing GCC haplotypes are by far the most numerous and diverse group, the intermediate IL-10 producing ACC-inclusive haplotypes seem to be related most closely to the ancestral haplotype, and the ATA-inclusive haplotypes cluster a separate branch with strong bootstrap support. We looked at associations of corresponding haplotypes with HIV progression. A haplotype trend regression confirmed that individuals carrying the low-producing ATA-inclusive haplotypes in European Americans progress to AIDS faster, and most likely explain the role of IL10. Our findings are consistent with the hypothesis that existing polymorphisms in this gene may reflect a balance of historic adaptive responses to autoimmune, infectious and other disease agents.

Value added: increasing the power to assess treatment outcome in joint haemorrhages.

Subject reports of efficacy of treatment of haemophilia-related joint bleeding are by definition subjective, yet are often the primary outcome in studies comparing therapies. Verbal descriptors such as effective, partially effective, poorly effective, not effective are treated as dichotomous or categorical variables in analyses, lowering the statistical power relative to that which might be achieved with a continuous variable. The aims of this study were to examine reports of pain recorded on a 100-mm visual analogue scale (VAS) during the course of joint bleeding; determine whether pain varied by treatment period among pairs reporting discordant outcomes on a verbal scale (one product effective, the other not effective); test whether the two products under study were equivalent with respect to VAS scores; and evaluate their relationship to verbal reports of efficacy. Data from the international, prospective, randomized, crossover FEIBA NovoSeven Comparative study of two bypassing agents used for treatment of 96 bleeding episodes in 48 participants were examined. VAS scores were associated with verbal descriptors of efficacy at every time point, and were equivalent between treatment periods. There were differences in mean scores at time points at which participants rated one treatment effective, the other not effective. As a continuous variable, the VAS score may have more power than a dichotomous variable and when used with verbal descriptions of efficacy can improve the overall accuracy of assessment. This report highlights an important consideration in the selection of outcome measurement that can be generalized to other haemophilia treatment research.

Long-term major joint outcomes in young adults with haemophilia: interim data from the HGDS.

A study of major joint outcomes, specifically range of motion and synovitis, was conducted with data from a subset of adolescents enrolled in the prospective Hemophilia Growth and Development Study (HGDS). Clinical observations were carried out over a 7-year period from 1989 to 1996. A secondary aim was to gain insight into factors that might influence decisions regarding maintaining or discontinuing prophylaxis during early adulthood. Twenty-nine participants (median age 17.4 at entry) were included. Median follow-up was 7 years (range: 4.8-7.7). Range of motion (ROM) and synovitis in six major joints (knees, elbows and ankles), were evaluated by physical examination every 6-12 months. At the baseline observation, 73.6% of joints showed no ROM abnormalities or synovitis, and all joints were normal in 11 patients. Of the 11 participants, 54.5% developed abnormalities and 28.1% of normal joints at baseline became abnormal during the follow-up. Ankles were the most severely affected and had persistent progression during late adolescence and adulthood. Elbows and knees did not show progression after the first few years of the follow-up. The progression of haemophilic arthropathy in adolescents and young adults varies from individual to individual and also in the site of affected joints. In view of this, the decision regarding discontinuation of prophylaxis in patients with haemophilia should be individualized.

Hepatitis C virus infection and neurocognitive function.

Hepatitis C virus (HCV) infection may be associated with neurocognitive deficits. The Hemophilia Growth and Development Study enrolled HIV-infected and HIV-uninfected patients and a group of nonhemophiliac siblings. After controlling for multiple factors, HCV monoinfection was not associated with deficits in adaptive behavior, intelligence, or attention/concentration.

Single nucleotide polymorphisms and haplotypes in the IL10 region associated with HCV clearance.

Hepatitis C virus (HCV) is an infectious blood-borne pathogen that usually persists as a chronic infection. However, approximately 15% of the time, patients can clear the virus, indicating that host differences could be critical in determining the course of HCV infection. The inflammatory response is crucial to resolving or failing to resolve an acute HCV infection. Some previous reports have implicated interleukin 10 (IL10) polymorphisms with successful anti-HCV therapy and natural viral clearance. We tested 54 single nucleotide polymorphisms (SNPs) in the IL10 region (+/-300 kb and 24 within the IL10 gene itself), which contains 13 genes including the IL10 immunomodulatory paralogs IL19, IL20, and IL24, for association with HCV clearance vs persistence. SNPs from two haplotype block regions, one at IL10 and the other from IL19/IL20, were associated with HCV clearance in African Americans (91 clearance cases and 183 chronically infected matched controls; P=0.05-0.002) while with expectation-maximization algorithm-reconstructed haplotypes, these associations remained (P=0.05-0.002). However, no significant associations were detected in European Americans (108 clearance and 245 chronic). Our results indicate that variants of the immunomodulatory IL10 and IL19/IL20 genes may be involved in natural clearance of HCV in the African-American population.

Frequency and causes of hospitalization in HIV-negative children and adolescents with haemophilia A or B and its effect on academic achievement.

The aim of the study is to determine the causes and frequency of hospitalization in HIV-negative boys and adolescents with haemophilia and evaluate their impact on academic achievement. One hundred and twenty-six HIV-negative boys and adolescents were followed prospectively from 1989-96, at 14 comprehensive haemophilia treatment centres. One hundred and fifteen participants with haemophilia A or B were included in the investigation. These participants contributed an average of 57.8 months of follow-up. There were 203 hospitalizations in 65 participants and 50 participants were never hospitalized. Haemarthroses and soft tissue bleeds accounted for 46 and 44 causes of hospitalization. Central line infection was the third most common cause. Participants with inhibitor had the majority of central line infections and hospitalizations. Intracranial haemorrhage resulted in five hospitalizations in two participants. Other causes of bleeding accounted for 22% of hospitalizations. The median number of hospitalizations per year was 0.18. Duration of hospital stay was significantly related to lower spelling scores. Acute and chronic joint problems and soft tissue bleeds still account for the majority of hospitalizations. Positive inhibitor status was associated with higher numbers of hospitalizations and central line infections. Academic achievement was affected, to some degree, by length of hospital stay.

Surveillance of infectious complications associated with central venous access devices in children with haemophilia.

PURPOSE: To analyse the risk factors for infection associated with central venous access device (CVAD) use in children with haemophilia. METHODS: Risk factors for CVAD infection among patients with congenital haemophilia who had had a CVAD implanted at a single institution were evaluated utilizing the following variables: age at CVAD placement, age at end of study, number of days with a CVAD, percentage of lifetime with a CVAD, and history of inhibitor. RESULTS: Fifty-nine patients had a total of 97,936 (median 1768 days per patient) CVAD days in the study period. The median age at CVAD placement was 2.7 years (range 0-14.0). Twenty-six (44%) patients reported CVAD infections during the study period from January 1993 to October 2000. Twenty-four patients had their CVAD replaced, 17 (71%) of whom reported having infections and seven (29%) of whom had a history of inhibitor. The strongest predictor for having any infections was inhibitor status (P=0.16), although none of the risk factors had statistically significant effects. Among the 26 patients reporting infections, 42% had more than one CVAD-related infection. Seven patients had multiple infections involving the same organism. The mean rate of infection was 0.45 per 1000 catheter days, with a 95% confidence interval of 0.33-0.60. Those with a history of inhibitor had an infection rate of 0.66 compared with 0.38 per 1000 catheter days (P=0.09) for those without a history of inhibitor. Patients who were older (greater than the median age of 2.7) at CVAD placement had a lower rate of infection (0.29 vs. 0.65, P<0.01) compared with those < or =2.7 years. Adjustment for inhibitor status had little impact on these results. For the group as a whole, the median time to first infection was 1977 days from CVAD placement. Patients who were older at CVAD placement or study exit had lower relative hazards of infection (P=0.05 and P=0.09 respectively), while those who had inhibitors had a higher but not statistically significant relative hazard of 1.88 (P=0.13). CONCLUSIONS: These data reveal that while considerable numbers of patients develop CVAD-related infection, the interval between catheter placement and infection can be quite long. In addition, the earlier in life a CVAD is placed, the higher the risk of infectious complications, as evidenced by the tendency towards a higher infection rate. Measures to prevent CVAD-related infection might be focused on very young patients who appear to be at higher risk.

Defining the impact of hemophilia: the Academic Achievement in Children with Hemophilia Study.

OBJECTIVES: We characterized a population-based cohort of school-aged children with severe hemophilia with respect to type of treatment, on-demand versus prophylaxis, and frequency of bleeding episodes in the year before enrollment. We also investigated the association between hemophilia-related morbidity, measured by number of bleeding episodes in the year before enrollment, and academic performance after adjustment for other factors known to have an effect on achievement. Finally, we explored the mechanisms for the association between bleeding episodes and academic achievement. STUDY DESIGN: This study was a multicenter investigation of boys 6 to 12 years old with severe factor VIII deficiency (clotting factor level <2%) receiving care in US hemophilia treatment centers. Children with a history of inhibitor, severe developmental disorder, significant psychiatric disorder, or insufficient fluency in English were excluded from the study. On-demand treatment was defined as administration of clotting factor on the occurrence of a bleeding episode. Prophylactic therapy was defined as a course of regular infusions for >2 months with a goal of preventing bleeding episodes. Academic achievement was measured by the Wechsler Individual Achievement Test. Quality of life was measured by the Child Health Questionnaire. Of particular interest was the Physical Summary (PhS) measure of the Child Health Questionnaire. The type of information captured by the PhS includes limitations in physical activity, limitations in the kind or amount of schoolwork or social activities the child engaged in, and presence of pain or discomfort. RESULTS: One hundred thirty-one children were enrolled, a median center recruitment rate of 77%. The mean age of the participants was 9.6 years, and approximately half of the participants had completed less than the fourth grade at the time of enrollment. Sixty-two percent of the children were on prophylaxis at enrollment, and 9% had previously been on prophylaxis but were currently on on-demand therapy. Two groups were defined: ever treated with prophylaxis and never treated with prophylaxis. For those ever treated, treatment duration ranged from 2.7 months to 7.7 years, with one half of the children treated with prophylaxis for >40% of their lifetimes; 29% had always been on on-demand therapy. Children in both treatment groups were similar with respect to age, clotting factor level, parents' education, and IQ. The median number of bleeding episodes experienced in the year before enrollment for the cohort as a whole was 12. The median number of bleeding episodes in children on prophylaxis at enrollment was significantly lower than in children on on-demand therapy (6 vs 25.5). The mean achievement scores were within the average range of academic performance: reading, 100.4; mathematics, 101.6; language, 108.1; writing, 95.4; and total achievement, 102.5. When children were categorized as above or below the study group median by number of bleeding episodes, those who had a low number of bleeding episodes (< or =11) had better total achievement (104.4 vs 100.6) and mathematics (103.6 vs 99.6) than children in the higher bleeding episode category (> or =12) after adjusting for child's IQ and parents' education. Treatment with prophylaxis per se was not associated with better test scores, but children who had been treated on a regimen of long-term prophylaxis (>40% of lifetime) and reported < or =11 bleeding episodes in the year before enrollment had significantly higher scores in total achievement (104.9 vs 100.6), mathematics (105.2 vs 99.6), and reading (104.0 vs 98.6) than all other children reporting > or =12 bleeding episodes in the same time period. Increased school absenteeism and hemophilia-related limitations in physical functioning among children with greater frequency of bleeding episodes were proposed as the mechanisms for lower scores. The number of bleeding episodes was positively correlated with school absenteeism (Spearman correlation = 0.23), and children with more school absences had lower scores in mathematics, reading, and total achievement, even after adjusting for the child's IQ and parents' education. Children with fewer bleeding episodes also had better PhS scores than children in the high bleeding episode category (48.4 vs 41.3). The mean PhS for children in the low bleeding episode group (48.4) was similar to that of the general US population (50), but the mean PhS for children in the higher bleeding episode group was almost a full standard deviation lower than the mean for the general US population. PhS scores were positively related to reading and total achievement scores after adjusting for IQ and parents' education. Of interest and concern was a group of children who were reportedly being treated with prophylaxis during the year before enrollment (N = 18) but whose bleeding events were not optimally suppressed. These children were 3 times as likely (33.3% vs 11.1%) to be receiving < or =2 infusions per week as children on prophylaxis who reported < or =11 bleeding episodes during the same period. A review of the sites of bleeding reported for the 18 children revealed that 12 (66.6%) experienced > or =25% of their bleeding episodes in the same joint. CONCLUSIONS: Each child should have the opportunity to achieve his or her potential. Control of a chronic disorder must include this important goal as well as the more commonly identified medical outcomes. This study has identified an important association between the number of bleeding episodes experienced and academic achievement in a cohort of school-aged children with severe hemophilia. The data support the assertion that therapeutic care programs in this population must not be evaluated only in terms of financial cost to achieve adequate musculoskeletal outcomes. Also significant are the individual and societal benefits of increased academic accomplishments if adequate suppression of hemorrhagic events can be attained. The number of bleeding episodes experienced, regardless of treatment regimen, should be followed to optimize the child's academic outcome.

Response to measles, mumps, and rubella revaccination among HIV-positive and HIV-negative children and adolescents with hemophilia. Hemophilia Growth and Development Study.

The effect of human immunodeficiency virus (HIV) infection on response to measles, mumps, and rubella revaccination in children and adolescents with hemophilia was evaluated. Antibody levels of measles, mumps, and rubella were assayed at baseline and two annual examinations in 207 HIV-positive and 126 HIV-negative hemophiliacs participating in the Hemophilia Growth and Development Study (HGDS). Response to revaccination was analyzed for participants whose antibody levels were below the cut-off at the start of a year-long observation period. Among HIV-positive participants, antibody levels were below cut-off in 52 subjects for measles, in 71 for mumps, and in 96 for rubella. Among HIV-negative participants, antibody levels were low in 23 subjects for measles, in 23 for mumps, and in 31 for rubella. For measles and mumps antigens, revaccination was associated with a significant increase in redraw antibody levels for HIV-negative participants. Although there was an increase in the mean measles titers for revaccinated HIV-positive participants, it was not significant. Revaccination was associated with an increase in rubella antibodies in HIV-positive and HIV-negative participants. Revaccination with measles and mumps was associated with an increase in antibody levels in HIV-negative participants but not in HIV-positive participants. Both HIV-positive and HIV-negative participants responded to rubella revaccination with an increase in antibody levels.

The effect of plasma human immunodeficiency virus RNA and CD4(+) T lymphocytes on growth measurements of hemophilic boys and adolescents.

OBJECTIVE: The investigation examined the associations of plasma human immunodeficiency virus (HIV) RNA and CD4(+) T lymphocytes with height, weight, skeletal maturation, testosterone levels, and height velocity for hemophilic children and adolescents with HIV infection in the Hemophilia Growth and Development Study. STUDY DESIGN: Two hundred seven participants were evaluated over 7 years. RESULTS: A threefold increment in baseline plasma HIV RNA was associated with a 0.98-cm decrease in height and a 1.67-kg decrease in weight; 100-cells/microL decrements in baseline CD4(+) were associated with a 2.51-cm decrease in height and a 3.83-kg decrease in weight. Participants with high plasma HIV RNA (>3125 copies/mL) experienced significant delay in achieving maximum height velocity and lower maximum velocity compared with those with low viral load. The high CD4(+) (>243)/low plasma HIV RNA group had earlier age at maximum height velocity compared with the other 3 groups and higher maximum height velocity compared with the low CD4(+)/high plasma HIV RNA and low CD4(+)/low plasma HIV RNA groups. Decrements in CD4(+) were associated with decreases in bone age and testosterone level. CONCLUSIONS: CD4(+) and HIV RNA were important in predicting growth outcomes.

Incidence of focal white matter lesions in a population of hemophiliac children and their normal siblings. Hemophilia Growth and Development Study.

OBJECTIVE: This analysis was undertaken to evaluate the etiology and sequelae of 2- to 5-mm focal white matter hyperintensities on T2-weighted MR images of some participants enrolled in the Hemophilia Growth and Development Study (HGDS). MATERIALS AND METHODS: The HGDS is a multicenter study of the growth and development, neurological, neuropsychological, and immune functioning of a cohort of children and adolescents, 62% of whom were infected with HIV through the use of clotting factor concentrates, and their non-hemophiliac, non-HIV infected male siblings. The current investigation was conducted with all three groups of HGDS participants: HIV-positive hemophiliacs (n = 207), HIV-negative hemophiliacs (n = 126), and their siblings (n = 47). Magnetic resonance imaging was performed at each center, with a variety of 0.3 to 1.5 T instruments. Standard examinations included 5-mm-thick T1-weighted sagittal and axial images, intermediate, and T2-weighted axial images. A study of abnormalities of the coagulation system known to be associated with thrombotic events was conducted among a subgroup of participants (n = 51) from eight centers. RESULTS: Lesions were not associated with hemophilia-related factors, immune function, hematologic, or neurologic factors. There were no associations between the presence of white matter lesions and defects of coagulation in any of the assays completed. CONCLUSION: The 2- to 5-mm focal white matter hyperintensities on T2-weighted MR images of the brain were incidental findings in our study population.

Changes in hepatitis B serologic titers in HIV+ and HIV- children with haemophilia.

This longitudinal study examines differences in hepatitis B immune titres in children and adolescents with haemophilia to determine if they are dependent on how immunity was acquired (vaccination or natural infection), and whether they are related to the child's HIV status and/or are influenced by HIV disease progression. Serologic titres (HBcAb, HBsAb) and HBsAg were measured prospectively at baseline, and at years 1, 2 and 3 of follow-up in 126 HIV- and 207 HIV+ children and adolescents with haemophilia. Analyses were performed to assess the impact of HIV status on the measured titres, and for HIV+ subjects to examine the association with CD4+ lymphocyte counts and p24 antigen status. The results show that HIV+ children were more likely than HIV- children to lose vaccine-induced immunity as indicated by the loss of HBsAb. There was an increased risk of losing HBsAb with higher CD4+ counts and younger age. Re-immunization was not successful in seven of eight HIV+ children. Two subjects (one HIV+, one HIV-) entered the study HBsAg- but became HBsAg+ over the course of follow-up. Seven HIV+ subjects lost natural immunity as indicated by the loss of HBcAb. The loss of either HBsAb or HBcAb in HIV--subjects was negligible to absent. In conclusion, because of the loss of immunity in HIV+ children the viral safety of factor replacement concentrates for these children is an important consideration. HIV- children rarely lose immunity, therefore frequent measures of HBsAb are not necessary.

HIV-associated immune dysfunction and delayed pubertal development in a cohort of young hemophiliacs. Hemophilia Growth and Development Study.

As part of the Hemophilia Growth and Development Study (HGDS), we investigated the relationship between HIV-associated immune dysfunction and delayed pubertal development in a cohort of 333 boys and adolescents with moderate or severe hemophilia who were between the ages of 6 and 19 years at study entry in 1989. Sixty-two percent of the cohort was infected with HIV in the late 1970s and early 1980s through exposure to contaminated clotting factor concentrates. The cohort was observed during follow-up at 6-month intervals; measurements taken at each follow-up visit included Tanner stage and CD4+ cell count. This analysis of data from the first 4 years of follow-up revealed statistically significant delays in pubertal development associated with increasing levels of immune dysfunction. Our results emphasize the importance of following pubertal development in HIV-infected adolescent boys since delays in maturation may reflect underlying disease progression.

Hemophilia morbidity, cognitive functioning, and academic achievement.

Data from the Hemophilia Growth and Development Study (HGDS) were used to evaluate the association between hemophilia morbidity, measured by abnormalities in coordination and gait (CG), and intellectual ability and academic achievement. The CG abnormalities observed in the HGDS participants (n = 333) were primarily due to hemophilia-related morbidity. Although HGDS participants performed within the average range for age on measures of intellectual ability, there were meaningful differences between CG outcomes at baseline and throughout the 4 years of study. Participants without CG abnormalities consistently achieved higher scores than those with CG abnormalities on Reading, Spelling, and Arithmetic subtests of the Wide Range Achievement Test-Revised. Our findings suggest that lowered achievement is related to the functional severity of hemophilia.

Comparability of absolute/percent CD4+ T-lymphocytes completed locally and centrally. Hemophilia Growth and Development Study.

The Hemophilia Growth and Development Study (HGDS) is a multicenter longitudinal study of 333 male children and adolescents with moderate or severe hemophilia, ranging in age from 6 to 19 at entry. Sixty-two percent of the cohort was infected with human immunodeficiency virus (HIV) in the late 1970s and early 1980s through exposure to contaminated clotting factor concentrates. The HGDS has followed this cohort since 1989. HGDS subjects have blood drawn twice each year for t-lymphocyte subsets, with fresh blood shipped overnight to a central laboratory. T-lymphocyte subsets from the same blood draw are often determined locally as well. To evaluate interlaboratory variation, we examined the comparability of pairs of local and central results for CD4+ absolute counts and percents. Ninety-four pairs of absolute counts and 73 pairs of percent CD4 + results were available. We calculated concordance correlation coefficients, which evaluate the agreement between two readings from the sample by measuring the variation from the 45 degrees line through the origin. Absolute counts were square root transformed. Comparability of the pairs was high for both absolute counts and percents (0.93 and 0.92, respectively). Agreement was high whether we determined the CD4+ counts and percents centrally, using fresh samples received the day after the examination (0.95, 0.95), or from specimens that were frozen upon receipt and batched for later testing (0.90, 0.87). We conclude that when a centrally processed CD4+ result is unavailable because of shipping problems or loss of specimens, a study may reasonably accept a CD4+ result completed locally, if validity checks indicate good comparability. In the HGDS, the data provided by the local laboratories were of comparable quality to those provided by the central laboratories.