Abstract

A major gastroenteritis outbreak among >400,000 residents of Milwaukee, Wisconsin, in April 1993 was attributed to Cryptosporidium parvum oocysts in drinking water. Plasma specimens obtained from children (6 months to 12 years old) for routine blood lead level surveillance March–May 1993 were assayed by ELISA for levels of IgG antibody against the immunodominant Triton-17 and 27-kDa C. parvum antigens. Over a 5-week period, the seroprevalence for antibodies to the 2 antigens increased from 15% to 82% and from 17% to 87%, respectively, in samples from children living in southern ZIP code areas (n=218), whereas smaller increases (20% to 43% and 22% to 46%, respectively) were noted among samples from children living in northern ZIP code areas (n=335; P<.0001). The results demonstrate that C. parvum infection was much more widespread than previously appreciated and confirm that infection was associated with residence in the area served by the southern water treatment plant

Cryptosporidium parvum is a parasitic protozoan that causes diarrhea in both immunocompetent and immunocompromised persons [1]. In addition to diarrhea, infection is associated with abdominal cramps, fever, nausea, dehydration, and weight loss. C. parvum infection can lead to premature death in immunocompromised persons, especially AIDS patients, because the illness is not self-limited in these patients, as it is in immunocompetent persons [1]. Infection can be acquired from infected persons or animals and from food- and waterborne routes of transmission [2–9]

In April 1993, in Milwaukee, Wisconsin, a major gastroenteritis outbreak was attributed to the presence of C. parvum oocysts in water supplied by the Howard Avenue Purification Plant, a water treatment plant that primarily served the southern region of the city. As determined on the basis of a random digit–dialing telephone survey of the greater Milwaukee area, ∼400,000 Milwaukee residents (26% of the population) were ill and met the cryptosporidiosis case definition (watery diarrhea) [10]. The magnitude of the outbreak in Milwaukee established C. parvum as a significant threat to the US water supply and emphasized the need for better drinking water treatment practices and for the establishment of surveillance programs for waterborne parasites [11]

The conclusion that C. parvum was the cause of the outbreak in Milwaukee was reached on the basis of (1) a strong epidemiologic link between drinking water and diarrheal illness, (2) demonstration of C. parvum in the stools of >700 ill persons, (3) the absence of demonstrable infections with other pathogens, and (4) detection of C. parvum in ice made from water collected during the outbreak period [10]. However, the relatively nonspecific nature of the clinical case definition and the finding of oocysts in only 29% of stool specimens submitted for C. parvum testing has raised questions in the minds of some investigators about the magnitude of the estimate of the number of infected persons. The existence of banked plasma specimens, which were obtained from Milwaukee children in 1993 for routine blood lead level determination, provided us with an opportunity to conduct a retrospective analysis of the Milwaukee outbreak, using recently developed serologic assays for Cryptosporidium infections

Methods

Study populationA major gastroenteritis outbreak among >400,000 residents of Milwaukee in April 1993 was attributed to C. parvum oocysts in drinking water. During the same time period (25 March through 2 June 1993), blood samples were obtained from children in Milwaukee for routine surveillance of blood lead levels. Residual blood lead specimens (5–20-μL volumes) were retrieved for serologic analysis at the time of the outbreak investigation. The plasma specimens were centrifuged to remove particulate material and were stored at −80°C until the time of assay. Personal identifiers were not associated with the specimens; however, age, sex, date of collection, and ZIP codes of residences were available

A total of 1484 plasma samples was available from children ⩽12 years old who lived in the greater Milwaukee area. The samples were grouped by ZIP code of residence and by week of collection. To increase the available numbers of samples in the first and last sets, we added samples collected on 25 and 26 March to the week 1 set (29 March to 2 April), and samples from 31 May to 2 June were added to the week 9 set (24–28 May). Samples from weeks 1 (25 March to 2 April), 3 (12–16 April), 5 (26–30 April), 7 (10–14 May), and 9 (24 May to 2 June) from 4 contiguous southern ZIP code areas (n=218) and from 6 contiguous northern ZIP code areas (n=335) were chosen for ELISA analysis. The number of samples collected in each time period is indicated in figure 1A. Samples from the southern and northern ZIP code areas were assayed separately in the approximate order of the date of collection

Figure 1

Prevalence of positive antibody responses to the Triton-17 (A) and recombinant 27-kDa (B) Cryptosporidium parvum antigens among children residing in southern and northern Milwaukee. The nos. of available specimens in each subset are indicated in parentheses on the respective bars in A. Significant differences (P<.05) in the prevalence of positive antibody responses to the Triton-17 and 27-kDa C. parvum antigens between children from northern and southern Milwaukee were observed from the week of 26 April through the week of 24 May (including 31 May to 2 June) 1993

Figure 1

Prevalence of positive antibody responses to the Triton-17 (A) and recombinant 27-kDa (B) Cryptosporidium parvum antigens among children residing in southern and northern Milwaukee. The nos. of available specimens in each subset are indicated in parentheses on the respective bars in A. Significant differences (P<.05) in the prevalence of positive antibody responses to the Triton-17 and 27-kDa C. parvum antigens between children from northern and southern Milwaukee were observed from the week of 26 April through the week of 24 May (including 31 May to 2 June) 1993

An additional 120 residual blood lead samples obtained from children (median age, 3 years; range, 6 months to 5 years) between 8 March and 8 April 1999 were available for antibody testing. As with the 1993 samples, the 1999 samples were not associated with personal identifiers, but age, sex, date of collection, and ZIP codes of residences were available. The samples were subdivided into residents of the southern (n=60), middle (n=30), and northern (n=30) ZIP code areas before ELISA analysis, but they were not divided by date of collection

Laboratory analysisELISA antibody responses to the 17- and 27-kDa antigens were tested, as described elsewhere [12], using a partially purified preparation of a native antigen preparation enriched for a 17-kDa antigen (Triton-17 antigen) and a recombinant form of 27-kDa antigen. In brief, each antigen was diluted in 0.1 M sodium bicarbonate buffer (pH 9.6), was added to 96-well, flat-bottom microtiter immunoassay plates (Immulon 2; Dynex Technologies), and was incubated overnight at 4°C. Triton-17 and 27-kDa antigens were used at 14 and 17 ng per well, respectively. After blocking for 1 h at 4°C with 100 μL of 0.3% Tween-20 in buffer containing 0.85% NaCl and 10 mM Na2HPO4 (PBS; pH 7.2), plasma specimens (diluted 1:50), 3 positive and 4 negative controls (diluted 1:50), and an 8-point standard curve (2-fold serial dilutions from 1:50 to 1:6400 of a strong positive specimen) were loaded onto the plate in duplicate wells (50 μL/well). Tween-20 (0.05%) in PBS was used for all washes and dilutions

The samples were incubated in the wells for 2 h at room temperature and then were incubated for 1 h at room temperature with biotin-conjugated monoclonal anti–human IgG (50 μL/well of a 1:1000 dilution; clone HP6017; Zymed Laboratories) and with alkaline phosphatase–labeled streptavidin (50 μL/well of a 1:500 dilution; Life Technologies) with washing steps in between. Finally, p-nitrophenyl phosphate substrate (50 μL/well; Sigma) in 3 mM MgCl2–10% diethanolamine (pH 10) was added. The plates were read when the optical density of the 1:50 dilution of the standard curve control serum reached an absorbance of 1.5 at 405 nm, and this point on the standard curve was assigned a value of 6400 arbitrary units (AU). AU values for unknown samples were derived from the 8-point standard curve, using a 4-parameter fit, and were expressed per microliter of serum. Mean values were calculated from the duplicate wells. Samples were reassayed if the SD was >20% of the mean value from the duplicate wells. The positive thresholds for an antibody response to the Triton-17 antigen (>57 AU) and for a response to the 27-kDa antigen (>160 AU) were based on the mean+3 SD of the values of the 4 negative controls. These controls were shown elsewhere to be negative for anti–Cryptosporidium antibody by immunoblot assay. In previous studies, the sensitivity and specificity of these assays were >90% relative to the immunoblot [12, 13], and the results were reproducible from day to day and from one laboratory to another [13]

Statistical analysisThe χ2 test was used to assess the significance of differences in antibody response to Triton-17 and 27-kDa antigens between children who resided in ZIP code areas primarily served by southern and northern Milwaukee water treatment plants during the 9 weeks of sample collection. Antibody responses were compared by use of the Kruskal-Wallis test. Statistical significance was defined as P<.05. Statistical analysis was done with Epi Info (version 6.04a; CDC) and SAS for Windows (version 6.12; SAS Institute)

Results

Plasma specimens from 553 children were assayed for antibody against the Triton-17 and 27-kDa C. parvum antigens. The children’s ages ranged from 6 months to 12 years (median, 3 years); 46.6% were ⩽2 years old, 46.8% were 3–5 years old, and 6.6% were 6–12 years old. Samples were examined from children residing in ZIP code areas primarily served by the southern water treatment plant (n=218, representing 6 ZIP codes) and northern water treatment plant (n=335, representing 4 ZIP codes). Of the specimens from southern Milwaukee, 62.8% came from children who resided in one particular ZIP code (53204) area. Likewise, 34.8% and 30.7% of specimens from northern Milwaukee were from children residing in areas covered by ZIP codes 53206 and 53212, respectively

The earliest plasma samples that were available to us (beginning 25 March 1993) corresponded to the beginning of the period during which increases were observed in the number of cases of laboratory-confirmed cryptosporidiosis (figure 2) and clinically defined cases of watery diarrhea (as determined by random digit–dialing telephone surveys) [10]. Figure 2 shows the reported date of illness onset for all laboratory-confirmed cases of cryptosporidiosis (both adults and children; n=270) in the greater Milwaukee area between 16 March and 15 May 1993

Figure 2

Onset date of illness for all laboratory-confirmed cases of cryptosporidiosis (adults and children; n=270) from the greater Milwaukee area between 16 March and 15 May 1993. Bars on the X-axis indicate the dates of collection of the residual blood lead specimens used in this study: week 1, 29 March to 2 April; week 3, 12–16 April; week 5, 26–30 April; and week 7, 10–14 May. Specimens collected on 25 and 26 March were combined with those collected between 29 March and 2 April (week 1). No laboratory data were available for the week 9 period of collection (24–29 May, including May 31 to June 2)

Figure 2

Onset date of illness for all laboratory-confirmed cases of cryptosporidiosis (adults and children; n=270) from the greater Milwaukee area between 16 March and 15 May 1993. Bars on the X-axis indicate the dates of collection of the residual blood lead specimens used in this study: week 1, 29 March to 2 April; week 3, 12–16 April; week 5, 26–30 April; and week 7, 10–14 May. Specimens collected on 25 and 26 March were combined with those collected between 29 March and 2 April (week 1). No laboratory data were available for the week 9 period of collection (24–29 May, including May 31 to June 2)

In this study, samples collected during weeks 1 (29 March to 2 April, including 25 and 26 March), 3, 5, 7, and 9 (24–29 May, including 31 May to 2 June) of the outbreak were analyzed by ELISA. For the period between 25 March and 2 April, 15% and 17% of children served by the southern water treatment plant (n=46) and 20% and 22% of children served by the northern water treatment plant (n=98) had positive IgG antibody responses against the Triton-17 and 27-kDa C. parvum antigens, respectively (figure 1). No significant difference in the prevalence of IgG antibody responses between children residing in areas served by the 2 water treatment plants was observed during this period for either the Triton-17 antigen (P=.61) or the 27-kDa antigen (P=.63). Similarly, the median Triton-17 and 27-kDa antigen ELISA responses for the first time period were below the respective cutoff levels in both the north and south (figure 3)

Figure 3

Comparison of median IgG antibody reactivity to the Triton-17 (A) and 27-kDa (B) Cryptosporidium parvum antigens for children residing in southern and northern Milwaukee. Nos. of samples are the same as those shown in figure 1A. Significant differences (P<.05, Kruskal-Wallis test) in median IgG antibody responses between children residing in southern and northern Milwaukee were observed from the week of 26 April through the week of 24 May (including 31 May to 2 June) 1993

Figure 3

Comparison of median IgG antibody reactivity to the Triton-17 (A) and 27-kDa (B) Cryptosporidium parvum antigens for children residing in southern and northern Milwaukee. Nos. of samples are the same as those shown in figure 1A. Significant differences (P<.05, Kruskal-Wallis test) in median IgG antibody responses between children residing in southern and northern Milwaukee were observed from the week of 26 April through the week of 24 May (including 31 May to 2 June) 1993

To determine whether the age of the child was related to the antibody response, the ELISA responses in the first time period (25 March to 2 April) from the north and south were combined and were grouped by donor age, in years (n=144; median age, 3 years; range, 7 months to 6 years). No significant age-related differences were detected in the prevalence of positive responses (P=.59 and P=.92) or in the median antibody responses (Kruskal-Wallis test; P=.123 and P=.755) for either the Triton-17 or 27-kDa antigen, respectively. The median antibody responses for each age category were <32 AU and <88 AU for the Triton-17 and 27-kDa antigen ELISAs, respectively. The only 6-year-old child in the set had responses of 52 AU (Triton-17 antigen) and 127 AU (27-kDa antigen)

The prevalence of IgG antibody responses to the Triton-17 and 27-kDa antigens increased rapidly after the first week. By week 5 of serum collection (beginning 26 April), 82% and 87% of children served by the southern water treatment plant and 43% and 46% of children served by the northern water treatment plant had IgG antibody responses to the Triton-17 and 27-kDa C. parvum antigens, respectively (figure 1). The prevalence of antibody responsiveness in both the south and the north remained at higher-than-baseline levels through week 9 of serum collection. The prevalence of positive antibody responses in children served by the southern water treatment plant was significantly greater than that of children residing in the north (P<.05) during weeks 5–9 of serum collection

For both the Triton-17 and 27-kDa C. parvum antigens, the median antibody response for children served by the southern water treatment plant peaked during week 5 of serum collection (week beginning 26 April; figure 3); however, the median antibody responses of children residing in the north peaked during week 7 of serum collection (week beginning 10 May). As with antibody prevalence, median antibody responses of children residing in the southern part of the city were greater than those of children residing in the north for plasma samples collected from week 5 (beginning 26 April) through week 9 (beginning 24 May) for both antigens. To rule out the influence of antibody prevalence on these comparisons, we restricted the analysis to children with positive antibody responses. Among antibody-positive children, those in southern Milwaukee had significantly higher median responses (P<.05) than those in northern Milwaukee to both the Triton-17 and 27-kDa antigens during weeks 5–9 of sample collection (table 1)

Table 1

Median antibody response to the Triton-17 and 27-kDa Cryptosporidium parvum antigens among antibody-positive children in 1993 in Milwaukee

Table 1

Median antibody response to the Triton-17 and 27-kDa Cryptosporidium parvum antigens among antibody-positive children in 1993 in Milwaukee

To determine the current prevalence of antibody-positive children in Milwaukee, we used Triton-17 and 27-kDa antigen ELISAs to test residual blood lead specimens (n=120) that were collected in 1999. Samples from children <6 years old were chosen for this work, so that any children who may have been infected with C. parvum during the 1993 outbreak would be excluded. Among the children, 7% and 6% were positive for antibodies to the Triton-17 and 27-kDa antigens, respectively (data not shown). No significant differences in prevalence were observed between residents of the northern, middle, and southern ZIP code areas for either the Triton-17 or 27-kDa responses (P=1.0 and P=.43, respectively)

Discussion

We assessed acquisition of C. parvum infection among children who resided in Milwaukee during the 1993 waterborne cryptosporidiosis outbreak, using newly developed serologic assays that measure levels of IgG antibodies against the immunodominant 27- and 17-kDa sporozoite C. parvum surface antigens [12]. Serologic test results supported the conclusion that C. parvum was the cause of the outbreak of diarrheal illness in Milwaukee. Antibody reactivity to C. parvum antigens was significantly greater among children in southern Milwaukee than among children in northern Milwaukee, which adds to the evidence that the southern water plant was the primary source of the oocysts

Over the period that plasma samples were collected, Cryptosporidium seroprevalence among children served by the southern water treatment plant increased from 17% to 87% for the recombinant 27-kDa antigen and from 15% to 82% for the Triton-17 antigen, which is enriched for the native 17-kDa antigen (figure 1). These increases in IgG antibody prevalence and median antibody level (figure 3) paralleled the epidemic curve of illness onset shown in figure 2, with a lag time of ∼3 weeks (∼4 weeks after the likely time of exposure). This result is consistent with the amount of time required to mount an IgG antibody response [14]. The fact that increased antibody responses were observed to 2 distinct C. parvum antigens over this time period decreases the likelihood that changes in antibody reactivity were a result of cross-reactivity to other agents (e.g., Microsporidia species, Giardia species, etc.)

Evidence from animal studies suggests that antibody responses to C. parvum antigens require inoculation with viable oocysts [15]. Previous studies using human volunteers have shown that serologic responses to the 17-kDa and 27-kDa antigens develop after both symptomatic and asymptomatic C. parvum infection [14]. Thus, serologic responses to C. parvum in children probably reflect infection and not merely exposure to nonviable oocysts. On the basis of this conclusion, children had greater levels of infection with C. parvum than those predicted from the random digit–dialing survey of diarrheal illness [10]. However, the attack rates determined from the telephone survey may not be directly comparable to the rates of seropositivity found in this study, because, although the telephone survey included residents from all the ZIP code areas in the region, our study concentrated on children from only a few ZIP code areas in the northern and southern Milwaukee regions. Thus, our results may not be fully representative of their respective regions. Despite this potential limitation, the rates of serologic positivity in both regions were substantially greater than the attack rates from the telephone survey, even when the attack rates were weighted for the ZIP codes used in our study (W. R. Mac Kenzie, unpublished data)

From a comparison of baseline seroprevalence values from the first week of sample collection with peak seroprevalence, we estimate that children served by the southern and northern water treatment plants had, at a minimum, C. parvum infection rates of 70% and 37%, respectively, during this outbreak. If the outbreak in Milwaukee began in early March, as is suggested by the occurrence of sporadic cases of diarrheal illness earlier in the month, the seroprevalence for the period beginning 25 March was not a true baseline, and these numbers probably underestimate the rate of infections due to the outbreak. Unfortunately, no samples predating the 1993 outbreak are available, and it may be impossible to accurately determine the pre-outbreak baseline seroprevalence in Milwaukee. To demonstrate that the assays perform well in the absence of an outbreak, a baseline seroprevalence was determined for children, using plasma samples more recently collected from Milwaukee (1999). However, we would suggest that a direct comparison between these results and those of the 25 March to 2 April 1993 time period is not valid, because Milwaukee instituted significant changes to improve the overall quality of the water supply in the aftermath of the 1993 outbreak

No serum samples from adult populations were available for measurement of C. parvum–specific responses for the time period represented by the blood lead specimens from children. Nonetheless, given the higher attack rate for diarrheal illness among adults in the greater Milwaukee area (28% for persons ⩾20 years old), compared with that among children (19% for children <10 years old) [10], it seems likely that the antibody prevalence among adults would be comparable in magnitude to that among the children. In fact, Frost et al. [16] recently demonstrated that, in the weeks after the 1996 outbreak of cryptosporidiosis in Collingwood, Ontario, Canada, 69% and 88% of the adult residents who were tested by Western blot assay were positive for antibodies to the 17- and 27-kDa C. parvum antigens, respectively [16]. Both the fraction of positive residents and the mean antibody responses to these 2 markers were higher in residents of Collingwood than in residents of Toronto, where no outbreak occurred. These serologic responses were detected despite the fact that only 16% of the non–nursing home adult population had a laboratory-confirmed infection. Thus, as noted by Mac Kenzie et al. [10], the total number of infected persons in the Milwaukee population probably was higher than the initial estimate of 400,000 people

Differences between published rates of infection from epidemiologic studies (based on measures of illness such as watery diarrhea) and antibody prevalence probably reflect the range of clinical manifestations of infection (from watery diarrhea to asymptomatic infection) in the population. These approaches provide complementary public health information. Case definitions based on illness are better for quantifying the degree of morbidity associated with C. parvum. In contrast, serologic studies provide better measures of infection

Unfortunately, we do not know whether children in our study experienced symptoms of cryptosporidiosis. Although we have no reason to believe that these children experienced symptoms at a higher rate than the rate determined by telephone surveys, we cannot make strong inferences regarding the rate of asymptomatic infection. Nonetheless, the rate of diarrheal illness was lower in children than in adults. Perhaps this disparity is related to the smaller number of organisms ingested by children, because development of symptoms was related to the oocyst dose in studies of Cryptosporidium-infected adult volunteers [17]. Alternatively, the absence of symptoms in children may reflect some degree of prior exposure to infection, as suggested by Morris et al. [18]. Moss et al. [14] suggested that the presence of preexisting antibody responses may be associated with protection from illness. In the absence of specimens collected before 25 March 1993, our data do not address this possibility directly. We postulate, however, that prior outbreaks would be reflected by an increase in C. parvum seroprevalence with age at baseline (period beginning 25 March). Although our sample numbers were small, we saw no age-related differences in antibody prevalence or level in samples collected during the first week, as we would have expected if prior exposures had occurred

There were several potential limitations to our study. Because our study was based on plasma specimens collected for blood lead level surveillance in March through May 1993, only a single specimen was available from each child. Therefore, we were unable to document seroconversion to Cryptosporidium per se. However, we have no reason to believe that there were any systematic differences in the way that specimens were collected or analyzed over the course of the study that could explain the temporal variation in antibody prevalence and level. In addition, we cannot control for the possibility that children residing in northern Milwaukee were exposed to infection in southern Milwaukee or vice versa. Likewise, because pipes in northern and southern Milwaukee are interconnected, children residing in northern Milwaukee may have had some level of exposure to water coming from the southern water treatment plant while they were at home. However, both the movements of children and of water between northern and southern Milwaukee would tend to diminish the magnitude of the differences in seroprevalence and antibody level that we did observe. Finally, as noted above, there were no serum specimens from adults available to us, so we were unable to document the level of infection in adults over the same time period and, thus, were unable to estimate the overall infection level in the community

Despite these limitations, our results provide very strong evidence that C. parvum was the cause of the 1993 Milwaukee outbreak. These results emphasize the utility of serologic assays for C. parvum for epidemiologic studies and surveillance. Further investigations are needed to have a better understanding of the relationship between seroprevalence at the population level and community susceptibility to outbreaks of symptomatic disease

References

1
Juranek
DD
Cryptosporidiosis: sources of infection and guidelines for prevention
Clin Infect Dis
 , vol. 
21(Suppl
  
1):S57–S61
2
Fayer
R
Cryptosporidium and Cryptosporidiosis
Boca Raton, FL: CRC Press
 
3
Millard
PS
Gensheimer
KF
Addiss
DG
, et al.  . 
An outbreak of cryptosporidiosis from fresh-pressed apple cider
JAMA
 , vol. 
272
 (pg. 
1592
-
6
(erratum: JAMA 1995; 273:776)
4
Hayes
EB
Matte
TD
O’Brien TR
Brien
TR
, et al.  . 
Large community outbreak of cryptosporidiosis due to contamination of a filtered public water supply
N Engl J Med
 , vol. 
320
 (pg. 
1372
-
6
)
5
et al.  . 
An outbreak of cryptosporidiosis linked to a foodhandler
J Infect Dis
 , vol. 
181
 (pg. 
695
-
700
)
6
Dworkin
MS
Goldman
DP
Wells
TG
Kobayashi
JM
Herwaldt
BL
Cryptosporidiosis in Washington State: an outbreak associated with well water
J Infect Dis
 , vol. 
174
 (pg. 
1372
-
6
)
7
Bridgman
SA
Robertson
RM
Syed
Q
Speed
N
Andrews
N
Outbreak of cryptosporidiosis associated with a disinfected groundwater supply
Epidemiol Infect
 , vol. 
115
 (pg. 
555
-
66
)
8
Atherton
F
Newman
CP
Casemore
DP
An outbreak of waterborne cryptosporidiosis associated with a public water supply in the UK
Epidemiol Infect
 , vol. 
115
 (pg. 
123
-
31
)
9
Willocks
L
Crampin
A
Milne
L
, et al.  . 
A large outbreak of cryptosporidiosis associated with a public water supply from a deep chalk borehole
Commun Dis Public Health
 , vol. 
1
 (pg. 
239
-
43
)
10
Mac Kenzie
WR
Schell
WL
Blair
KA
, et al.  . 
Massive outbreak of waterborne Cryptosporidium infection in Milwaukee, Wisconsin: recurrence of illness and risk of secondary transmission
Clin Infect Dis
 , vol. 
21
 (pg. 
57
-
62
)
11
Centers for Disease Control and Prevention
Working Group on Waterborne Cryptosporidiosis
Cryptosporidium and water: a public health handbook
Atlanta: Centers for Disease Control and Prevention
 
12
Priest
JW
Kwon
JP
Moss
DM
, et al.  . 
Detection by enzyme immunoassay of serum immunoglobulin G antibodies that recognize specific Cryptosporidium parvum antigens
J Clin Microbiol
 , vol. 
37
 (pg. 
1385
-
92
)
13
Priest
JW
Li
A
Khan
M
, et al.  . 
Enzyme immunoassay detection of antigen-specific immunoglobulin G antibodies in longitudinal serum samples from human cryptosporidiosis patients
Clin Diag Lab Immunol
 , vol. 
 8
 (pg. 
415
-
23
)
14
Moss
DM
Chappell
CL
Okhuysen
PC
, et al.  . 
The antibody response to 27-, 17-, and 15-kDa Cryptosporidium antigens following experimental infections in humans
J Infect Dis
 , vol. 
178
 (pg. 
827
-
33
)
15
Reperant
JM
Naciri
M
Chardes
T
Bout
DT
Immunological characterization of a 17-kDa antigen from Cryptosporidium parvum recognized early by mucosal IgA antibodies
FEMS Microbiol Lett
 , vol. 
78
 (pg. 
7
-
14
)
16
Frost
FJ
Muller
T
Craun
GF
, et al.  . 
Serological analysis of a cryptosporidiosis epidemic
Int J Epidemiol
 , vol. 
29
 (pg. 
376
-
9
)
17
DuPont
HL
Chappell
CL
Sterling
CR
Okhuysen
PC
Rose
JB
The infectivity of Cryptosporidium parvum in healthy volunteers
N Engl J Med
 , vol. 
332
 (pg. 
855
-
9
)
18
Morris
RD
Naumova
EN
Griffiths
JK
Did Milwaukee experience waterborne cryptosporidiosis before the large documented outbreak in 1993?
Epidemiology
 , vol. 
9
 (pg. 
264
-
70
)
Presented in part: annual meeting of the American Society of Tropical Medicine and Hygiene, Washington, DC, December 1999 (poster 1047).
Human experimentation guidelines of the US Department of Health and Human Services were followed in the conduct of the clinical research, and the study was exempted from review by the institutional review board at the Centers for Disease Control and Prevention.

Author notes

a
Present affiliations: Atlanta VA Medical Center, Decatur, Georgia (A.C.M.); Georgia Division of Public Health, Atlanta (W.R.M.); National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta (M.R.H.).