Abstract
We determined the frequencies of antibodies to Anaplasma phagocytophila, the agent of human granulocytic ehrlichiosis (HGE), in different groups of adults and children from Westchester County, New York. The groups included 159 adult blood donors and 215 children who were seronegative for Borrelia burgdorferi antibodies, 118 adult patients and 57 children who were seropositive for B. burgdorferi antibodies, and 42 adult patients with culture-confirmed erythema migrans. Eighteen (11.3%) of the blood donors and 11 (5.1%) of the B. burgdorferi-seronegative children were found to have A. phagocytophila antibodies by indirect immunofluorescent-antibody assay (IFA). Nine of 42 (21.4%) patients with culture-confirmed erythema migrans tested at the baseline visit, 42 of 118 (35.6%) adults, and 3 of 57 (5.3%) children whose sera were reactive for B. burgdorferi antibodies also tested positive for A. phagocytophila antibodies. The geometric mean titer ranged from 219 to 315 for all groups, and the differences in titers among the groups were not statistically significant. Only one-third of the healthy blood donors reactive by IFA were confirmed to be positive by immunoblotting. We conclude that a significant proportion of adults and children without clinical evidence of HGE will test positive for A. phagocytophila antibodies when the conventional cutoff titer of 80 is used in the IFA. This information must be considered in interpretation of test results.
Human granulocytic ehrlichiosis (HGE) is an emerging tick-borne disease first described in 1994 in the midwestern United States (5, 7). The etiologic agent is an ehrlichial species closely related to Ehrlichia phagocytophila. It is often referred to as the HGE agent and was recently named Anaplasma phagocytophila (11). Since the original report, increasing numbers of cases of infection with this agent have been diagnosed in many of the same geographic areas where Borrelia burgdorferi infection is endemic (1, 5, 7, 8). Both infections are transmitted by the same Ixodes tick vectors (12).
Antibody testing by an indirect immunofluorescent-antibody assay (IFA) is the test most commonly used in the laboratory to confirm a diagnosis of HGE. We and others have shown that approximately 90% of acutely infected patients are seropositive by serologic testing during the acute or convalescent phase (≤30 days postinfection) (3, 7).
An antibody titer of ≥64 or ≥80 by IFA is considered a positive result (10). However, few studies exist on the rate of background seropositivity in areas of endemicity, and it is unclear if this cutoff is the most appropriate threshold for seropositivity. Indeed, evidence exists that this titer is too low. A study by Bakken et al. (6) of 475 individuals from Wisconsin without evidence of active tick-borne infection or a prior diagnosis of HGE found a seropositivity rate of 14.9%, and the patients usually had titers ≤320. While such low titers occur at surprisingly high frequencies in individuals in the general population, in our experience patients acutely infected and confirmed to have HGE by recovery of the agent from blood have distinctly higher titers (greater than or equal to 640 in 95.2% of patients) (3). The purposes of the present study were to determine the background seropositivity rate in another area of endemicity and to attempt to determine at what age seropositivity occurs and to what extent any observed seropositivity represents subclinical or resolved A. phagocytophila infection versus antibody reactivity due to other causes.
Coded and unlinked sera frozen at −70°C were tested for antibodies to A. phagocytophila. The sera were obtained from five groups of individuals, as follows. (i) One hundred fifty-nine serum samples were collected from healthy blood donors from Westchester County, New York. These serum samples were randomly selected from the approximately 5,000 such serum samples collected during 1989 and 1990 that were known to be seronegative for antibodies to B. burgdorferi. (ii) A convenience sample consisted of sera from 118 adult patients (age, >19 years) whose sera were randomly selected from among those sera which tested positive for antibodies to B. burgdorferi by enzyme immunoassay and immunoblotting at the Westchester Medical Center from 1990 to 2000. The clinical histories of the donors of these samples were not available. (iii) A second convenience sample consisted of sera from 42 adult patients (age, >19 years) with erythema migrans who were diagnosed with B. burgdorferi infection at the Westchester Medical Center from 1991 to 2000 on the basis of recovery of the spirochete from culture of a skin biopsy sample. The sera tested were baseline samples. (iv) A third convenience sample consisted of sera from 57 pediatric patients (age, ≤19 years) who had tested positive for antibodies to B. burgdorferi at the Westchester Medical Center from 1991 to 1999. (v) The final sample consisted of sera from 215 pediatric patients from Westchester County collected from 1996 to 2000 for endocrinologic testing. These sera were selected in order to obtain representative samples from children in different age groups (<1 to 19 years).
Serologic testing.
Antibodies to B. burgdorferi were measured by the recommended two-step approach. The sera were first assayed by an immunoglobulin G (IgG) and IgM enzyme-linked immunosorbent assay (ELISA; Wampole Laboratories, Cranbury, N.J.), and those that were reactive were subjected to separate IgG and IgM immunoblotting assays (MarDx Diagnostics, Inc., Carlsbad, Calif.) as described previously (2).
An IFA was used to detect antibodies to A. phagocytophila in assays with a local human isolate (isolate NY-13), as described previously (3). Briefly, slides were prepared for IFA when >90% of the HL-60 cells used in the tests were infected with A. phagocytophila NY-13, as detected by Wright staining. Suspensions of infected cells were applied to each well of 12-well Teflon-coated slides (Cell Line; Erie Scientific Co., Portsmouth, N.H.), air dried, and fixed in acetone. The sera were tested at an initial dilution of 1:80, and bound antibodies were detected after incubation with fluorescein isothiocyanate-labeled goat anti-human IgG, IgM, and IgA conjugate (Kirkegaard & Perry Laboratories, Gaithersburg, Md.) at a dilution of 1:50 as a secondary antibody. Those sera that were reactive at the initial dilution of 1:80 were tested after serial dilution up to 1:2,560 to determine the titer. Sera with titers ≥2,560 were given a value of 2,560 for calculations of geometric mean titers (GMTs).
Antigen was purified from isolate NY-13 by a density gradient method and was used to prepare immunoblot strips for detection of antibodies to A. phagocytophila, as described previously (4). Sera were tested at 1:100 dilutions in separate assays for IgG and IgM antibodies. The criterion for positivity was the presence of a distinct band(s) in the area of 42 to 44 kDa.
Statistical methods.
Categorical variables were compared by Fisher's two-tailed exact test. IFA titers were compared by the Kruskal-Wallis one-way analysis of variance on ranks. A P value of ≤0.05 was considered significant.
A. phagocytophila antibodies in sera from healthy blood donors.
Sera from 159 healthy blood donors who reside in an area of Westchester County where Lyme disease is endemic and which were previously tested and found to be nonreactive for antibodies to B. burgdorferi were retested. Eight serum samples (5.0%) were reactive for antibodies to B. burgdorferi by ELISA, but none of these eight serum samples was positive for borrelial antibodies by immunoblotting for IgM or IgG. Eighteen (11.3%) serum samples tested positive for antibodies to A. phagocytophila by IFA (Table 1). The titers of antibodies to the HGE agent ranged from 80 to 640, with a median of 160 and a GMT of 226. These 18 serum samples were also tested by separate immunoblotting assays for IgM and IgG antibodies to A. phagocytophila. On the basis of the reactivities to the 42- to 44-kDa immunodominant antigen, six serum samples (33.3%) were reactive: two serum samples were positive for only IgM antibodies, three serum samples were positive for only IgG antibodies, and one serum sample was positive for both IgM and IgG antibodies (Table 2). Seventeen additional serum samples randomly selected from blood donors that tested negative for antibodies to A. phagocytophila by IFA were similarly tested by immunoblotting; all were nonreactive (P = 0.045 for the comparison of 5 of 18 serum samples versus 0 of 17 serum samples).
TABLE 1.
Seroprevalence of antibodies to A. phagocytophila by IFA in different populations from Westchester County, New York
| Group | No. of serum samples positive/no. tested (%) | IFA titer
|
||
|---|---|---|---|---|
| Range | Median | GMTa | ||
| Healthy blood donors | 18/159 (11.3) | 80-640 | 160 | 226 |
| Pediatric B. burgdorferi-negative subjects | 11/215 (5.1) | 80-640 | 160 | 219 |
| Pediatric B. burgdorferi-positive subjects | 3/57 (5.3) | 160-320 | 320 | 254 |
| Adult patients with erythma migransb | 9/42 (21.4) | 80-≥2,560 | 160 | 242 |
| Adult B. burgdorferi-positive subjects | 42/118 (35.6) | 80-≥2,560 | 320 | 315 |
Differences in titer among groups were not statistically significant (P = 0.695 by Kruskal-Wallis analysis of variance on ranks).
Sera were collected at the baseline, prior to antimicrobial treatment.
TABLE 2.
Immunoblotting results for sera from healthy blood donors and children being tested for endocrinologic disorders positive for A. phagocytophila by IFA
| Population | No. of serum samples IFA positive | No. of serum samples reactive for A. phagocytophila by immunoblotting
|
|||
|---|---|---|---|---|---|
| IgM positive, IgG negative | IgM positive, IgG positive | IgM negative, IgG positive | IgM negative, IgG negative | ||
| Adult blood donors | 18 | 2 | 1 | 3 | 12 |
| Children | 11 | 4 | 0 | 4 | 3 |
Both groups were seronegative for antibodies to B. burgdorferi by two-stage testing.
Seroreactivity to A. phagocytophila in sera of children testing negative for B. burgdorferi antibodies.
Among 215 children (114 males and 100 females plus 1 child whose gender was unknown) who were being tested for an endocrinologic disorder, 17 (7.9%) were reactive for antibodies to B. burgdorferi by ELISA, but none of these 17 tested positive for borrelial antibodies by immunoblotting for IgM or IgG. Eleven of the 215 (5.1%) samples tested positive for antibodies to A. phagocytophila by IFA (Table 1); the 11 samples were from 4 of the 114 (3.5%) males and 7 of the 100 (7%) females (P = 0.35). Of these 11 samples, 8 (72.7%) also tested positive for HGE by immunoblotting (Table 2). The titers of those serum samples that tested positive by IFA ranged from 160 to 640, with a median of 160 and a GMT of 219. The difference in the rates of seroprevalence of A. phagocytophila antibodies determined by IFA between the healthy adult blood donors and this group of pediatric patients was significant (P = 0.03). The rate of antibody reactivity to A. phagocytophila was greatest for the group of children ages 5 to 9 years (12.5%) and was significantly higher for that age group than for the group of children ages <1 to 4 years (1.7%) (P = 0.04) (Table 3).
TABLE 3.
Frequency of A. phagocytophila antibodies in children by age and presence of B. burgdorferi antibodies
| Age group (yr) | No. of serum samples positive/no. tested (%)
|
|
|---|---|---|
| B. burgdorferi negativea | B. burgdorferi positive | |
| <4 | 1/160 (1.7) | 1/2 (50.0) |
| 5-9 | 6/48 (12.5) | 0/16 (0) |
| 10-14 | 3/65 (4.6) | 1/31 (3.2) |
| 15-19 | 1/42 (2.4) | 1/8 (12.5) |
| Total | 11/215 (5.1) | 3/57 (5.3) |
Sera were from children evaluated for endocrinologic disorders.
A. phagocytophila antibodies in sera of children and adults seropositive for antibodies to B. burgdorferi and in sera of adults with culture-confirmed erythema migrans.
Significantly higher rates of seropositivity for A. phagocytophila, however, were observed among adults (42 of 118 [35.6%] versus 18 of 159 [11.3%]; P < 0.001) but not among children (3 of 57 [5.3%] versus 11 of 215 [5.1%]; P = 1.0) seropositive for B. burgdorferi by two-stage testing (Table 1). The IFA titers of the adults positive for B. burgdorferi ranged between 80 and ≥2,560, with a median titer of 320 and a GMT of 315 (Table 1). Comparison of the titers of the positive samples from blood donors and the B. burgdorferi-positive groups was, however, not significant (P = 0.695). In addition, among the 42 baseline serum samples collected between 1991 and 2000 from adults with culture-confirmed erythema migrans, 9 (21.4%) were seropositive for antibody to A. phagocytophila (Table 1), including 5 of 22 (22.7%) patients diagnosed with erythema migrans in 1991 and 4 of 20 (20.0%) patients diagnosed with erythema migrans between 1996 and 2000 (P = 1.0). Among the groups of sera from children and adults reactive for antibodies to B. burgdorferi, the rate of IFA reactivity to A. phagocytophila was slightly higher in the groups of sera containing IgM antibodies to B. burgdorferi (23 of 85; 27.1%) than in those testing positive for IgG only (22 of 90; 24.4%) (P = 0.73) (Table 4).
TABLE 4.
A. phagocytophila IFA reactivity in sera of patients with serologic evidence of prior B. burgdorferi infection
| B. burgdorferi antibody reactivity by immunoblotting | No. of serum samples positive for A. phagocytophila antibodies by IFA/no. tested (%)
|
|
|---|---|---|
| Adults | Children | |
| IgM positive, IgG negative | 9/16 (56.3) | 0/15 (0) |
| IgM positive, IgG positive | 11/30 (36.7) | 3/24 (12.5) |
| IgM negative, IgG positive | 22/72 (30.6) | 0/18 (0) |
The results presented here extend and confirm the report of Bakken et al. (6), who found an A. phagocytophila seropositivity rate of 14.9% among 475 individuals without evidence of HGE or a prior clinical history of HGE when tested by IFA by using a cutoff titer for seropositivity of 80.
Our findings showed that 11.3% of 159 healthy adult blood donors from Westchester County who were seronegative for antibodies to B. burgdorferi were seropositive for A. phagocytophila (antibody titers, ≥80) (Table 1). The explanation for this seropositivity is unclear. Unexplained seropositivity for A. phagocytophila was also found in 11 of 215 (5.1%) pediatric patients whose sera had been submitted for endocrinologic testing, but at a significantly lower rate (P = 0.03). Bakken et al. (6) tested sera from only eight children, but they also demonstrated that the donors whose sera were positive were significantly older than those whose sera were negative. In our study, among adults who were seropositive for antibodies to B. burgdorferi, 35.6% were seropositive for antibodies to A. phagocytophila (Table 1). The difference in the rate of seropositivity between this group and the healthy blood donors was significant (P < 0.001). A similar seroprevalence (21.4%) was observed among adults with culture-confirmed erythema migrans whose sera were obtained at the time of diagnosis (before antibiotic therapy). In contrast, the rate of A. phagocytophila seropositivity (5.3%) among 57 children who tested seropositive for antibodies to B. burgdorferi was almost identical (5.1%) to the rate among children without borrelial seropositivity. Similar to the findings of De Martino et al. (S. J. De Martino, J. A. Carlyon, and E. Fikrig, Letter, N. Engl. J. Med. 345:150-151, 2001), adults with IgM antibodies to B. burgdorferi had a higher frequency of antibody reactivity to A. phagocytophila (20 of 46; 43.5%) compared with the frequency of antibody reactivity for those with only IgG reactivity to B. burgdorferi (22 of 72; 30.6%), but this difference was not significant (P = 0.17).
We are aware of only one prior study that assessed seropositivity for A. phagocytophila by an IFA in a pediatric population. In that study, conducted in Slovenia, children were tested at the time of their routine medical examinations. When reactivity was considered a titer of ≥64, 12 of 95 (12.6%) children tested positive, including 2 of 22 (9.1%) children 5 to 9 years of age, 7 of 31 (22.6%) children 10 to 14 years of age, and 3 of 42 (7.1%) children 15 to 19 years of age (9).
The high rates of background seropositivity for antibodies to A. phagocytophila found in both New York State and Wisconsin suggest that a single low-positive titer should be interpreted with caution in patients suspected of having recent HGE. Raising the threshold titer to 640 is supported by our findings. If this titer were considered the breakpoint, the seroprevalence rates would fall to 3.1% for our healthy blood donors and to 2.9% for the Wisconsin cohort. In contrast, in our laboratory, among 24 patients with culture-confirmed HGE, 95.2% would have still been considered seropositive if this cutoff had been used with an acute- or convalescent-phase serum sample (3).
Although it has been speculated that the high background rate of seropositivity for antibodies to A. phagocytophila indicates that a substantial proportion of the population has had subclinical HGE, this seems to be an unlikely explanation for several reasons. First, testing of sera from our seropositive healthy blood donors by immunoblotting demonstrated a pattern suggestive of prior A. phagocytophila infection in only 6 of 18 (33.3%) samples (Table 2). It should be emphasized, however, that the accuracy of immunoblotting in distinguishing true-positive from false-positive serologic reactivity to A. phagocytophila remains to be established. Second, the seroprevalence rate was substantial among the serum samples from the pediatric population tested, yet unlike adults, children with seroreactivity to B. burgdorferi were not more likely to be seropositive for antibodies to A. phagocytophila than those without reactivity to B. burgdorferi. In addition, the pattern of reactivity did not show an increase with the age of the child. Both observations would be surprising if subclinical infection were the only explanation for seropositivity.
The background seropositivity might represent cross-reactivity to common bacterial antigens or autoantibodies reactive with the HL-60 cells in which A. phagocytophila is cultivated, but no data are available to substantiate these hypotheses.
In summary, by using a cutoff titer of 80 for IFA, approximately 11 to 15% of healthy adults from Wisconsin or Westchester County and 5% of children from Westchester County will test positive for antibodies to A. phagocytophila. This seroreactivity likely does not indicate the occurrence of a prior subclinical case of HGE for some, if not most, of these individuals. Regardless of whether this seroreactivity represents the occurrence of a prior subclinical case of HGE, it will certainly negatively affect the positive predictive values of test results for patients undergoing serologic evaluation because of clinical symptoms. Individual laboratories performing IFA for antibodies to A. phagocytophila need to determine the most suitable cutoff titer for positivity by testing well-characterized serum samples from both healthy individuals residing in areas where HGE is endemic and patients with well-established acute HGE.
Acknowledgments
We thank Kim Barber for the preparations of A. phagocytophila immunoblots used in this study, Louis Rosenfeld for assisting with statistical methods, and Eleanor Bramesco for helping with the preparation of the manuscript.
This study was supported by grant 27019 from the Westchester County Department of Health (to M.E.A.-R.).
REFERENCES
- 1.Aguero-Rosenfeld, M. E., H. W. Horowitz, G. P. Wormser, D. F. McKenna, J. Nowakowski, J. Munoz, and J. S. Dumler. 1996. Human granulocytic ehrlichiosis (HGE): a case series from a single medical center in New York State. Ann. Intern. Med. 125:904-908. [DOI] [PubMed] [Google Scholar]
- 2.Aguero-Rosenfeld, M. E., J. Nowakowski, S. Bittker, D. Cooper, R. B. Nadelman, and G. P. Wormser. 1996. Evolution of the serologic response to Borrelia burgdorferi in treated patients with culture-confirmed erythema migrans. J. Clin. Microbiol. 34:1-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Aguero-Rosenfeld, M. E., F. Kalantarpour, M. Baluch, H. W. Horowitz, D. F. McKenna, J. T. Raffalli, T.-C. Hsieh, J. Wu, J. S. Dumler, and G. P. Wormser. 2000. Serology of culture-confirmed cases of human granulocytic ehrlichiosis. J. Clin. Microbiol. 38:635-638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Asanovich, K. M., J. S. Bakken, J. E. Madigan, M. Aguero-Rosenfeld, G. P. Wormser, and J. S. Dumler. 1997. Antigenic diversity of granulocytic Ehrlichia isolates from humans in Wisconsin and New York and a horse in California. J. Infect. Dis. 176:1029-1034. [DOI] [PubMed] [Google Scholar]
- 5.Bakken, J. S., J. S. Dumler, S.-M. Chen, M. R. Eckman, L. L. Van Etta, and D. H. Walker. 1994. Human granulocytic ehrlichiosis in the upper Midwest United States. A new species emerging? JAMA 272:212-218. [PubMed] [Google Scholar]
- 6.Bakken, J. S., P. Goellner, M. Van Etten, D. Z. Boyle, O. L. Swonger, S. Mattson, J. Krueth, R. L. Tilden, K. Asanovich, J. Walls, and J. S. Dumler. 1998. Seroprevalence of human granulocytic ehrlichiosis among permanent residents of northwestern Wisconsin. Clin. Infect. Dis. 27:1491-1496. [DOI] [PubMed] [Google Scholar]
- 7.Bakken, J. S., J. Krueth, C. Wilson-Nordskog, R. L. Tilden, K. Asanovich, and J. S. Dumler. 1996. Clinical and laboratory characteristics of human granulocytic ehrlichiosis. JAMA 275:199-205. [PubMed] [Google Scholar]
- 8.Centers for Disease Control and Prevention. 2001. Summary of notifiable diseases—United States, 1999. Morb. Mortal. Wkly. Rep. 48:5. [Google Scholar]
- 9.Cizman, M., T. Avsic-Zupanc, M. Petrovec, E. Ruzic-Sabljic, and M. Pokorn. 2000. Seroprevalence of ehrlichiosis, Lyme borreliosis and tick-borne encephalitis infections in children and young adults in Slovenia. Wien Klin. Wochenschr. 112:842-845. [PubMed] [Google Scholar]
- 10.Dumler, J. S., K. M. Asanovich, J. S. Bakken, P. Richter, R. Kimsey, and J. E. Madigan. 1995. Serologic cross-reaction among Ehrlichia equi, Ehrlichia phagocytophila, and human granulocytic ehrlichia. J. Clin. Microbiol. 33:1098-1103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dumler, J. S., A. F. Barbet, C. P. J. Bekker, G. A. Dasch, G. H. Palmer, S. C. Ray, Y. Rikihisa, and F. R. Rurangirwa. 2001. Reorganization of genera in the families Rickettsiaceae and Anaplasmataceae in the order Rickettsiales: unification of some species of Ehrlichia with Anaplasma, Cowdria with Ehrlichia and Ehrlichia with Neorickettsia, designation of six new species combinations and designation of Ehrlichia equi and “HGE agent” as subjective synonyms of Ehrlichia phagocytophila. Int. J. E vol. Microbiol. 51:2145-2165. [DOI] [PubMed] [Google Scholar]
- 12.Pancholi, P., C. P. Kolbert, P. D. Mitchell, K. D. Reed, J. S. Dumler, J. S. Bakken, S. R. Telford III, and D. H. Persing. 1995. Ixodes dammini as a potential vector of human granulocytic ehrlichiosis. J. Infect. Dis. 172:1007-1012. [DOI] [PubMed] [Google Scholar]