Pneumococcal vaccine has been available since 1977, when 14-valent polysaccharide vaccine (which was replaced by 23-valent polysaccharide vaccine in 1983) was first licensed. The Advisory Committee on Immunization Practices currently recommends pneumococcal vaccine for high-risk patients, such as persons who have had splenectomy and those with Hodgkin disease, and suggests that re-immunization be done every 6 years [7]. However, vaccine efficacy data in these populations have been limited, and poor responses to pure polysaccharide vaccines have been documented after treatment [8-15]. The Advisory Committee also recommends immunization with HIB-conjugate and 4-valent meningococcal vaccines, even though data on the efficacy of these vaccines for patients with Hodgkin disease who have completed treatment are not available [7, 16].

We examined the immunogenicity of polysaccharide-conjugate and unconjugated polysaccharide vaccines in patients treated for Hodgkin disease. Because the previous studies that have documented poor responses focused on patients immunized soon after treatment, we studied patients whose disease was diagnosed at least 2 years earlier. We evaluated antibody responses to HIB-conjugate and 4-valent meningococcal vaccines and compared licensed 23-valent pneumococcal polysaccharide vaccine with an investigational 7-valent pneumococcal-conjugate vaccine.

Methods

Top Methods Results Discussion Author & Article Info References

Patient Population and Immunization Schema

Patients with Hodgkin disease who had been treated at the Joint Center for Radiation Therapy in Boston since 1969 were eligible for enrollment. Patients were excluded if they had been diagnosed with Hodgkin disease within 2 years of the study, had had relapse, or had developed second tumors. All patients had received radiation (subtotal nodal radiation or total nodal radiation), chemotherapy, or combined radiation and chemotherapy.

Immunization histories were obtained from patients by questionnaires. If patients were unsure of their immunization history, medical records were reviewed or primary care providers were contacted to verify immunization. Patients who had not received a pneumococcal vaccine after completion of treatment and within 6 years of the current study were randomly assigned to receive 23-valent pneumococcal vaccine or 7-valent pneumococcal-conjugate vaccine in a permuted block fashion by time since diagnosis (2 to 4 years, more than 4 to 6 years, or more than 6 years) and type of treatment received (subtotal nodal radiation or other therapy that included total nodal radiation, chemotherapy, or combined radiation and chemotherapy). Allocation of patients to 23-valent pneumococcal or 7-valent pneumococcal-conjugate vaccine was done by central randomization, and patients were blinded to their assignments. Vaccine doses were prepared by a pharmacist and distributed to study personnel with a label that encoded patient name and lot number only. After completion of the study, patients were told which pneumococcal vaccine they had received. All patients received HIB-conjugate vaccine and 4-valent polysaccharide meningococcal vaccine. Simultaneous administration of polysaccharide vaccines was previously shown not to interfere with antibody responses [17]. Because the HIB-conjugate and pneumococcal-conjugate vaccines contain different protein carriers, no interference with responses was expected. Our study was approved by the Institutional Review Board of the Dana-Farber Cancer Institute, and written informed consent was obtained from each patient.

All patients enrolled in the study received 1) 4-valent meningococcal polysaccharide vaccine, which contained 25 µg each of serogroups A, C, Y, and W-135 polysaccharide [Menomune, Connaught, Swiftwater, Pennsylvania], and 2) HIB-conjugate vaccine, which contained 10 µg of oligosaccharide [HibTITER, Lederle-Praxis Biologics, Pearl River, New York]. These were mixed in one syringe and administered intramuscularly. Patients assigned to receive pneumococcal vaccine received either 1) 23-valent pneumococcal polysaccharide vaccine, which contained 25 µg each of capsular polysaccharide types 1, 2, 3, 4, 5, 6B, 7F, 8, 9N, 9V, 10A, 11A, 12F, 14, 15B, 17F, 18C, 19F, 19A, 20, 22F, 23F, and 33F [Pneumovax 23, Merck, West Point, Pennsylvania], or 2) an investigational 7-valent pneumococcal-conjugate vaccine, which contained 1 µg each of capsular polysaccharide types 4, 9V, 14, 18C, 19F, and 23F, combined with 2.5 µg of type 6B individually conjugated to the outer membrane protein complex of Neisseria meningitidis in a 0.5-mL, intramuscular dose (Merck). All vaccines were donated by the manufacturers. Patients were asked to record the occurrence of fever and local reactions at 8, 24, and 48 hours after immunization on a standardized questionnaire. Assessment of local reactions ranged from a minimum score of 0 (no redness, swelling, or pain at any time point) to a maximum score of 18 (more than an inch of redness and swelling with severe pain at all time points). Serum specimens were obtained before and 3 to 6 weeks after immunization, and they were stored at 70 °C until assayed.

Patient Sample

We identified 604 eligible patients who had been treated for Hodgkin disease at the Joint Center for Radiation Therapy between 1 January 1969 and 1 September 1991. We successfully contacted 235 patients, of whom 144 agreed to participate (252 patients did not respond to our letter, and correct addresses were not available for 117 patients). The mean age of all participants was 38 years (range, 13 to 72 years); 50% were females; the median time since diagnosis was 9 years (range, 2 to 24 years); and 130 participants (90%) had had splenectomy as part of their staging evaluation. Eighty-five of 144 patients had received pneumococcal immunization before receiving treatment for Hodgkin disease. Thirteen patients treated since 1985 (when HIB vaccine was first licensed) had received HIB vaccine before treatment.

All 144 participants were immunized with HIB-conjugate vaccine and 4-valent polysaccharide meningococcal vaccine. Sixteen had received a pneumococcal vaccine after treatment for Hodgkin disease and within 6 years of the current study; these persons were not re-immunized. One hundred twenty-eight patients were randomly assigned to receive pneumococcal immunization: Seventy received 7-valent outer membrane protein complex pneumococcal-conjugate vaccine, and 58 received 23-valent pneumococcal polysaccharide vaccine. The clinical characteristics of these two pneumococcal immunization groups are shown in Table 1. Randomization done using permuted blocks resulted in unequal group sizes.

Antibody Assays

Total Anti-HIB Antibody

We measured total binding anti-HIB capsular antibody concentrations by radioimmunoassay according to a standard Food and Drug Administration (FDA) protocol using tritiated polyribosylribitol phosphate. The radioimmunoassay was standardized by using the Center for Biologics Research and Review standard serum pool with an assigned value of 70 µg/mL.

Immunoglobulin G Anti-HIB Antibody

We measured IgG anti-HIB capsular polysaccharide concentrations by using enzyme-linked immunosorbent assay (ELISA). We used HIB oligosaccharide coupled to human albumin followed by goat anti-human IgG alkaline phosphatase conjugate (Caltag Laboratories, South San Francisco, California) to detect IgG anti-HIB antibody. The ELISA was standardized by using a reference plasma pool with an assigned IgG anti-HIB antibody value of 27.2 µg/mL.

Immunoglobulin G Antimeningococcal Group A Antibody

We measured IgG antimeningococcal group A antibody concentrations with ELISA by using a modification of a previously published method [18]. Flat-bottom microtiter plates (Immulon 2, Dynatech Laboratories, Chantilly, Virginia) were coated with equal volumes of methylated human serum albumin and a solution of N. meningitidis group A polysaccharide. The ELISA was standardized by using reference serum (Centers for Disease Control and Prevention, 1992) with an assigned IgG antimeningococcal group A antibody value of 91.8 µg/mL.

C-Polysaccharide Absorbed IgG Antipneumococcal Antibody

We measured IgG antibody to pneumococcal polysaccharide capsular antigens by C-polysaccharide absorbed ELISA modifying a previously described method [19]. We used goat antihuman IgG alkaline phosphatase conjugate (Caltag Laboratories) followed by amplification substrate and amplifier (Gibco BRL, Life Technologies, Gaithersburg, Maryland). A reference plasma pool (PSAB90, FDA, Bethesda, Maryland) was standardized on the basis of a reference pool (89SF, FDA). Control serum specimens for the assays measuring antibody responses to 23-valent pneumococcal vaccine and 4-valent meningococcal vaccine were taken from 19 healthy immunized adults (20 to 60 years of age) in a previous study. Controls and patients received different lots (from the same manufacturers) of 23-valent pneumococcal and 4-valent meningococcal vaccines. Control serum specimens for the assays measuring antibody responses to 7-valent pneumococcal-conjugate vaccine were from 21 healthy adults (18 to 46 years of age) immunized with the same lot of vaccine as the patients with Hodgkin disease. No controls were in both control groups. All serum specimens from patients and controls were measured in the same assays to reduce interassay variability.

Statistical Analysis

Antibody concentrations were transformed into logarithms for statistical calculations. Antibody concentrations below the assay limit of sensitivity were assigned values of one half the lower limit. Comparisons of the geometric means of antibody concentrations were done using the Mann-Whitney rank-sum test. Paired analyses were done using the Wilcoxon signed-rank test. The McNemar test was used to compare the percentage of patients who had anti-HIB antibody concentrations in the protective range before and after HIB immunization. Multiple stepwise linear regression analysis was used to assess the effect of age, sex, age at diagnosis, type of treatment, splenectomy, time since diagnosis, type of pneumococcal vaccine received, and previous HIB or pneumococcal immunizations on antigen-specific antibody concentrations. Partial correlation coefficients reported are calculated with all significant variables included in the regression model.

Results

Top Methods Results Discussion Author & Article Info References

Antibody Responses to HIB-Conjugate Vaccine

The geometric mean total anti-HIB antibody concentration as measured by radioimmunoassay increased from 1.79 µg/mL before to 54.1 µg/mL after immunization in the 144 patients (P = 0.001). Anti-HIB IgG antibody concentrations were also measured by ELISA, and the geometric mean of these concentrations increased from 2.96 µg/mL before to 28.6 µg/mL after immunization (P = 0.001) (Table 2). There was a strong direct linear correlation between the ELISA and the radioimmunoassay anti-HIB antibody determinations; the correlation coefficients were 0.73 (P = 0.001) for values before immunization and 0.70 (P = 0.001) for values after immunization. Radioimmunoassay measures total anti-HIB antibody concentrations (for example, of IgG, IgM, and IgA), and thus its results are higher than the IgG anti-HIB antibody concentrations measured by ELISA. We examined the effect of time since diagnosis on antibody responses, because more recently treated patients might be expected to have poorer responses to the vaccine. Time since diagnosis did not correlate with responses to the vaccine (r = 0.019; P = 0.82). Patients who were only 2 years from time of diagnosis responded well to HIB-conjugate vaccine (Figure 1).

View larger version (22K): [in this window] [in a new window]   Figure 1. Total anti-Haemophilus influenzae type b antibody concentrations in patients with Hodgkin disease, displayed by time since diagnosis. Top. Before immunization. Bottom. After immunization. Closed circles represent patients who had received HIB vaccine before diagnosis; the solid lines represent the estimated long-term protective level (1.00 µg/mL); and the dashed lines represent the estimated short-term protective level (0.150 µg/mL), as measured by radioimmunoassay.

We also determined the number of patients who achieved "protective" anti-HIB antibody concentrations after immunization. The percentage of persons with protective total anti-HIB antibody concentrations was determined by using the estimated long-term protective concentration of 1.00 µg/mL or more as measured by radioimmunoassay [20]. This percentage increased significantly: Sixty-two percent of patients had protective concentrations before immunization and 99% had them after immunization (P < 0.001) (Table 2). Eighty-eight (62%) of the 143 patients who had antibody measurements done both before and after immunization had protective anti-HIB antibody concentrations before immunization and maintained protective concentrations after immunization. Of the 55 patients who were unprotected before immunization, 53 had anti-HIB antibody concentrations in the protective range after immunization (96%; 95% CI, 87% to 100%). After immunization, only 2 persons had antibody concentrations of less than 1.00 µg/mL (0.72 µg/mL and 0.96 µg/mL, respectively); their antibody concentrations before immunization had been 0.39 µg/mL and 0.88 µg/mL, respectively.

Antibody Responses to Unconjugated Pneumococcal and Meningococcal Polysaccharide Vaccines

Fifty-eight patients with Hodgkin disease were assigned to receive 23-valent pneumococcal vaccine, and their individual antibody responses to seven serotypes (4, 6B, 9V, 14, 18C, 19F, and 23F) were determined and compared with the responses of 19 healthy, immunized controls. In addition, the mean value of these seven measurements was calculated for each individual person, and geometric means for the groups were compared.

Before immunization, the geometric mean antibody concentrations of patients with Hodgkin disease for the seven serotypes measured were similar to those of controls, except for serotype 18C (Table 3). After immunization, patients with Hodgkin disease had a 2.1-fold increase in mean IgG antipneumococcal antibody concentration to 9.15 µg/mL; this was similar to the 2.2-fold increase to 10.7 µg/mL seen in the control group (Table 3). A significantly lower geometric mean antibody concentration in patients with Hodgkin disease after immunization was seen only for serotype 18C.

All patients received 4-valent meningococcal polysaccharide vaccine, and we measured their responses to meningococcal group A polysaccharide. The geometric mean antimeningococcal group A antibody concentration of patients did not differ significantly from that of 19 healthy controls either before or after immunization (Table 3).

Antibody Responses to Pneumococcal-Conjugate Vaccine

The antibody responses of the 70 patients with Hodgkin disease who were immunized with 7-valent outer membrane protein complex pneumococcal-conjugate vaccine were compared with those of 20 healthy controls, who were immunized with the same lot of vaccine (Table 4). Patients with Hodgkin disease had a significantly lower response to a single dose of the pneumococcal-conjugate vaccine; their geometric mean IgG antipneumococcal mean antibody concentration after immunization was 4.95 µg/mL compared with 7.87 µg/mL for healthy controls (P < 0.05). Lower antibody responses were also documented for the seven individual serotypes measured, with post-immunization IgG concentrations of antipneumococcal antibody to serotypes 4, 18C, and 19F reaching statistical significance. As shown in Table 4, the impaired response to a single dose of the 7-valent outer membrane protein complex vaccine was also reflected in the geometric mean antibody increases: Patients had significantly lower antibody increases for mean pneumococcal antibody and for each of the seven individual pneumococcal antibody concentrations measured. The geometric mean IgG antipneumococcal antibody concentrations of patients and controls did not differ significantly before immunization.

Antibody Responses to 23-Valent Polysaccharide Vaccine and 7-Valent Pneumococcal-Conjugate Vaccine

Antibody concentrations after immunization were significantly lower in patients immunized with 7-valent vaccine; their geometric mean IgG antipneumococcal mean antibody concentration was 4.95 µg/mL compared with 9.15 µg/mL in the group immunized with 23-valent pneumococcal vaccine (P = 0.0005). Significantly lower antibody concentrations after immunization for each of the seven serotypes measured were also seen in the 7-valent group than in the 23-valent group (P < 0.05 for types 6B, 14, 18C, 19F, and 23F; P < 0.01 for types 4 and 9V). Geometric mean antipneumococcal antibody concentrations before immunization were not significantly different in the two patient groups for either the IgG mean pneumococcal and individual serotype antibody concentrations (Table 3 and Table 4).

Factors Associated with Antibody Concentrations

We evaluated clinical and demographic variables for their effect on antibody concentrations both before and after immunization. The following factors were entered into a multiple stepwise linear regression analysis: age at diagnosis; type of treatment; sex; splenectomy; time since diagnosis; previous HIB or pneumococcal immunization; current age; and type of pneumococcal vaccine received. Type of treatment was coded as either subtotal nodal radiation only or as more intensive therapy that included total nodal radiation, chemotherapy, or combined radiation and chemotherapy.

The only variable significantly associated with higher pneumococcal antibody concentrations after immunization was the type of pneumococcal vaccine received (partial correlation coefficient, 0.344). Time since diagnosis was not significantly correlated with response to pneumococcal vaccine. Before immunization, less intensive treatment (partial correlation coefficient, 0.297), female sex (partial correlation coefficient, 0.212), and less time since diagnosis (partial correlation coefficient, 0.196) were independently associated with higher mean pneumococcal antibody concentrations as well as with higher antibody concentrations for three of the seven serotypes measured. For total anti-HIB antibody concentrations, no variables were significantly associated with antibody concentrations before immunization. After immunization, younger age at diagnosis (or younger age at the time of our study) was independently significantly associated with higher total anti-HIB antibody concentrations (partial correlation coefficient, 0.286). Finally, none of the variables examined was significantly associated with higher antimeningococcal group A antibody concentrations before or after immunization.

Adverse Reactions to Immunizations

One hundred twenty-nine participants recorded their temperatures, and 19 (15%) reported temperatures higher than 37.2 °C (99 °F). Local reactions at the injection site or sites were also evaluated; 133 patients provided complete responses about these reactions on the questionnaire. All patients received the HIB-conjugate and meningococcal vaccine mixture in the right arm, and those randomly assigned to receive pneumococcal immunization received pneumococcal vaccine in the left arm. The severity of local reactions did not differ significantly between the right and left arms; the mean scores for local reactions were 2.9 in the right arm (range, 0 to 12; CI, 2.5 to 3.3) and 3.8 in the left arm (range, 0 to 17; CI, 3.1 to 4.4). In addition, the severity of local reactions to pneumococcal vaccine reported for the 7-valent and 23-valent pneumococcal vaccine groups was not significantly different; the mean scores were 3.2 (range, 0 to 13; CI, 2.6 to 3.9) and 4.4 (range, 0 to 17; CI, 3.1 to 5.7), respectively.

Discussion

Top Methods Results Discussion Author & Article Info References

One dose of HIB-conjugate vaccine (HibTITER) was highly immunogenic in patients with Hodgkin disease; it resulted in a 30-fold increase in total anti-HIB antibody as measured by radioimmunoassay. The geometric mean antibody concentration after immunization was 54.1 µg/mL; ninety-nine percent of persons had antibody concentrations of 1.00 µg/mL or more after immunization compared with 62% of persons before immunization (P < 0.001). The radioimmunoassay antibody measurement of 54.1 µg/mL is substantially higher than the value of 18.9 µg/mL observed after a primary series of HibTITER immunizations in infants [21]. We predict, therefore, that one dose of HIB-conjugate vaccine will protect patients with Hodgkin disease if administered at least 2 years after diagnosis.

Antibody responses to unconjugated 23-valent pneumococcal vaccine and 4-valent meningococcal vaccine were also examined. The average of each individual person's responses to the seven serotypes was calculated, and the geometric means of the averages for the groups were compared. Surprisingly, the mean pneumococcal concentration for the seven serotypes measured for patients with Hodgkin disease did not differ significantly from that of healthy controls (9.15 µg/mL and 10.7 µg/mL, respectively). We could not assess the number of patients with protective concentrations of pneumococcal antibody, because no data have defined protective IgG concentrations. Antibody responses to meningococcal group A polysaccharide were also normal. These results conflict with previous reports, which stated that treated patients with Hodgkin disease responded poorly to polysaccharides [11-15]. However, these previous studies focused on patients immunized within months of completing treatment, whereas we required patients to have been diagnosed at least 2 years before immunization. We speculate, therefore, that patients with Hodgkin disease regain the capacity to respond to polysaccharides within a few years of completing treatment. One limitation of our study, however, is that only 144 of 604 eligible patients were enrolled, and these persons may not have been representative of all patients with Hodgkin disease.

In contrast to the adequate response patients had to the standard 23-valent pneumococcal vaccine, patients assigned to receive the 7-valent outer membrane protein complex pneumococcal-conjugate vaccine had lower responses. The geometric mean pneumococcal antibody concentration to the seven serotypes contained in each vaccine was 4.95 µg/mL in patients who received the 7-valent outer membrane protein complex vaccine and 9.15 µg/mL for the recipients of the standard 23-valent vaccine (P = 0.0005). In addition, the 7-valent vaccine recipients had significantly lower antibody concentrations to each of the seven serotypes measured. The impaired response of patients with Hodgkin disease to the outer membrane protein complex pneumococcal-conjugate vaccine contrasts with their response to the HIB-conjugate vaccine. One explanation for the lower response is that the outer membrane protein complex carrier had not previously been "seen" by this patient sample. In contrast, the patients had all previously been immunized with the HibTITER carrier, diphtheria toxoid, and thus were primed to respond. Current studies examining the immunogenicity of pneumococcal-conjugate vaccine in young children have used a three-dose primary series, and higher antibody responses have been seen after the third dose [22]. Thus, additional doses of 7-valent outer membrane protein complex vaccine might be more immunogenic than unconjugated pneumococcal polysaccharide vaccine in patients with Hodgkin disease. A second explanation for poor response to pneumococcal conjugate vaccine may relate to the lower polysaccharide content in 7-valent outer membrane protein complex vaccine compared with 23-valent vaccine. The 7-valent outer membrane protein complex vaccine contains 1 to 2.5 µg of the individual serotypes conjugated to the outer membrane protein complex carrier; the 23-valent vaccine contains 25 µg of each serotype. The response of adults to the 7-valent outer membrane protein complex vaccine may be limited by the reduced quantity of polysaccharide antigens in the vaccine.

We assessed the influence of several factors on antibody responses in our study sample of patients previously treated for Hodgkin disease. After immunization, the only variable significantly associated with higher pneumococcal antibody concentrations was receipt of 23-valent pneumococcal vaccine instead of 7-valent outer membrane protein complex pneumococcal-conjugate vaccine. Three other factors were independently associated with higher pneumococcal antibody concentrations before immunization: less time since diagnosis, treatment with subtotal nodal radiation only, and female sex. Previous studies [11, 23] support the finding that less intensive treatment (sub-total nodal radiation only) is associated with higher antibody concentrations after treatment. The association of higher antibody concentrations with shorter time since diagnosis may be explained by the waning of antibody concentrations induced by pneumococcal immunization given before treatment. The correlation between female sex and higher antibody concentrations before immunization was unexpected and is probably a spurious finding: We know of no studies associating female sex with better responses to polysaccharide vaccine.

The only variable significantly associated with higher anti-HIB antibody concentrations after immunization was younger age at diagnosis (or younger age at the time of our study). We hypothesize that this association may be related to the carrier effect of the HIB polysaccharide-conjugate vaccine. Younger persons would be more likely to have recently received diphtheria toxoid vaccine as part of their routine childhood immunizations. More recent priming to the carrier may explain the association of increased responses to the conjugate vaccine [24]. The fact that younger age predicted higher anti-HIB antibody concentrations but not pneumococcal or meningococcal antibody concentrations supports this hypothesis.

Because 85 of our 144 patients had received pneumococcal immunization before treatment, we wanted to determine the occurrence of serious infection in our study population after the diagnosis of Hodgkin disease. This population is, of course, a selected one: We have no information about the incidence of serious infection in the 460 eligible patients who did not participate in the study, some of whom may have died of overwhelming sepsis. We found that 4 patients (2.8%) had microbiologically documented pneumococcal or H. influenzae sepsis. Two of the three persons with pneumococcal sepsis had received pneumococcal vaccine before receiving treatment for their disease. One of these cases represents a definite vaccine failure: The pneumococcus was available and serotyped as 23F, which is a vaccine serotype. These infectious episodes occurred 5, 10, and 14 years after treatment, respectively, and none of the patients involved had been immunized after completing treatment. The episode of presumed HIB sepsis (the organism was not typed) occurred 4 years after treatment had been completed, and the patient had never received HIB vaccine. Although these four patients had all had splenectomy, the risk for disease in patients with Hodgkin disease who have not had splenectomy is probably also increased [4]. An increased incidence of pneumococcal and HIB sepsis has been documented in patients treated with chemotherapy for other hematologic malignancies [25, 26]. During our study, we learned of the recent death of a nonsplenectomized 35-year-old patient treated for Hodgkin disease at our institution who had been cured many years previously. The patient had never been immunized and died of overwhelming pneumococcal sepsis.

Many patients treated for Hodgkin disease before 1977 have never received polysaccharide immunizations. Furthermore, many patients immunized before receiving treatment for Hodgkin disease have not been re-immunized with pneumococcal vaccine after treatment. The current 23-valent pneumococcal vaccine may have limitations in certain populations [8, 9], but our study suggests that responses in treated patients with Hodgkin disease are similar to those of healthy controls. Although we did not evaluate vaccine efficacy, we suggest that all patients with Hodgkin disease, whether or not they have had splenectomy, should initially be immunized at least 7 to 10 days before treatment of Hodgkin disease with HIB-conjugate, meningococcal, and pneumococcal vaccines as previously recommended [11, 12, 27]. Patients should be re-immunized with these three vaccines as early as 2 years after the completion of treatment. Pneumococcal immunization (and probably meningococcal immunization) should then be repeated every 6 years, because antibody concentrations decrease over time [7]. Future studies should address the need for booster immunization with HIB-conjugate vaccine.

Dr. George: University of North Carolina School of Medicine, Department of Medicine, Chapel Hill, NC 27599.

Drs. Tarbell and Shamberger: Children's Hospital, 300 Longwood Avenue, Boston, MA 02115.

Dr. Mauch: Joint Center for Radiation Therapy, Department of Radiation Oncology, 50 Binney Street, Boston, MA 02115.

Drs. Diller, Neuberg and Ambrosino and Ms. Phillips: Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115.

Dr. Anderson: Division of Infectious Diseases and Immunology, Saint Louis University Health Sciences Center, 3635 Vista Avenue, FDT-8N, St. Louis, MO 63110-0250.

Ms. Kinsella: The School of Medicine of the University of Connecticut Health Center, Farmington, CT 06030.