Although contaminated water sources [5-7] and infected animals [8-10] with subsequent spread to humans have been identified as sources of infection, reports of daycare center outbreaks [11-14], multiple family infections [15-17], and sequential infections in hospitalized patients and personnel [18, 19] suggest that the main source of transmission is person to person [20-23]. Only one study [24], in Sri Lanka, however, has attempted to prospectively clarify the epidemiology of this infection in families. Although the number of household contacts in that study was small (n = 23), the investigators identified infection by examination of direct stool smears in 10 (43%) of the participants.

Because we have previously reported a high incidence of Cryptosporidium infection among infants in slums in Fortaleza, the capital of Ceara in northeastern Brazil [25], and in a nearby town [4], we chose this community to study the infectivity of Cryptosporidium in families living under crowded urban conditions.

Methods

Top Methods Results Discussion Author & Article Info References

Collection of Epidemiologic Data

During an ongoing study of the epidemiology of Cryptosporidium infection in Goncalves Dias, an urban slum community in Fortaleza, Brazil, we sequentially enrolled households in which a child was identified with Cryptosporidium infection. After a detailed explanation of the study, informed consent was obtained in Portuguese for all participating household members. The study design and participation of humans were approved by the Johns Hopkins Joint Committee on Clinical Investigation, by similar committees at the University of Virginia, and by the Universidade Federal do Ceara in Fortaleza. During the 6-week surveillance period for each household, all diarrheal illnesses were recorded and, when possible, a diarrheal stool sample was obtained. In addition, the study design was to collect stool samples from each household member at the time of diagnosis of the index case and then again 2 and 6 weeks later, and to obtain sera at 0 and 6 weeks after the index diagnosis. In addition, nine household contacts had a third serum sample collected approximately 12 months after the initial collection.

Stool and Serum Collection

Stool samples were collected by household members and placed either into disposable plastic cups, with subsequent transfer to 10% formalin by field personnel, or were placed directly into the formalin-containing tubes. All initial stool samples from index children were quickly brought to the laboratory for subsequent culture for bacterial enteropathogens. Blood was collected by trained personnel using sterile 23-gauge butterfly needles and 10-mL syringes and was placed into 7-mL serum separator tubes. Blood samples were then centrifuged for 10 minutes at 1200 g, and the resulting sera were divided into 300-µL aliquots and were frozen at –70°C until assayed for IgM and IgG, as described below.

Stool Evaluation

Index stool samples were plated onto appropriate media (MacConkey agar, xylose-lysine-deoxycholate agar, thiosulfate-citrate-bile-sucrose agar, trypticase soy agar with 5% defibrinated sheep blood and ampicillin, and trypticase soy agar with 5% defibrinated sheep blood with Campylobacter jejuni supplement) to isolate enteric pathogens and were examined by iodine stain of fresh stool for the presence of ova and parasites. Examinations for classes of pathogenic Escherichia coli [26, 27] and rotavirus are not yet complete. Stool specimens were concentrated using formalin-ethyl acetate [28] at a centrifuge speed of 800 g (Fecal Parasite Concentrator; Evergreen Scientific, Los Angeles, California), and two slides were prepared from the resulting pellet. Slides were then stained using modified acid-fast [29] and auramine [30] stains. Acid-fast slides were read by examining the entire area within a 10-mm etched circle at 1000 x using oil immersion; auramine slides were read at 400 x using a fluorescence microscope. Slides were deemed positive if typical oocysts, 4 to 6 microns in diameter (as measured by an eyepiece micrometer), were visible on acid-fast and auramine-stained slides. Stools from family members of the index cases were evaluated only for the presence of Cryptosporidium by the above methods.

Serologic Evaluation

The enzyme-linked immunosorbent assay used was modified from Ungar and colleagues [31]. Cryptosporidium parvum oocysts, from the AUCp-1 isolate that has infected calves and humans and is maintained by periodic passage in calves (provided by Dr. Ron Fayer at the United States Department of Agriculture Research Center, Beltsville, Maryland [32]), were purified from the stool of experimentally infected calves using cesium chloride gradient centrifugation and were then lysed by freeze-thawing and sonication. Polystyrene plates (96 well; Nunc-Immuno Plate; Nunc Inc., Naperville, Illinois) were coated with approximately 1.25 x 10⁵ disrupted oocysts suspended in 50 µL of carbonate buffer, were incubated at room temperature for 2 hours, and were left overnight at 4 °C. Plates were washed 3 times with phosphate-buffered saline with Tween 20 and were blocked overnight with 1% bovine serum albumin in phosphate-buffered saline. After a second wash, 50 µL of each serum sample (diluted 1:32 with 1% bovine serum albumin in phosphate-buffered saline with Tween 20) was added to each well and incubated for 1 hour at 37 °C. After washing, the plates were incubated with 50 µL of a 1:1000 dilution of alkaline phosphatase-conjugated goat anti-human IgG (H and L chain) or IgM (Kirkegaard & Perry Laboratories Inc.; Gaithersberg, Maryland) for 1 hour at 37 °C and were developed with p-nitrophenyl phosphate (Sigma Diagnostics; St. Louis, Missouri) at room temperature for 30 minutes. The optical densities of individual wells were read at 405 nm with an enzyme-linked immunosorbent assay reader (Titertek Multiscan; Flow Laboratories Inc., Chantilly, Virginia).

Specificity of the measurements of anti-Cryptosporidium IgM was established by three methods. First, 25 serum samples positive for anti-Cryptosporidium IgM were retested on a 96-well polystyrene plate not coated with the Cryptosporidium oocyst antigen preparation. Using these conditions, IgM was not bound nonspecifically. Second, 11 serum samples were preabsorbed with the Cryptosporidium antigen preparation and then tested for the presence of anti-Cryptosporidium IgM. In 10 serum samples, anti-Cryptosporidium IgM was not detected after antigen preabsorption; in one serum sample, the optical density reading was decreased by 90% by antigen preabsorption (the initial specific optical density reading of 0.402 was decreased to 0.036 after antigen preabsorption). Third, serum samples from 37 newborns were tested for anti-Cryptosporidium IgM. As expected, none were positive.

Three to six serum samples from North American children 3 to 13 years of age, presumably with a low risk for previous infection with Cryptosporidium, were used on each plate as negative controls. One serum sample from a child with known Cryptosporidium infection was also used on each plate as a positive control. All samples were run in duplicate, and the optical density readings from the two wells were averaged. A sample was considered positive if the average optical density reading was greater than the mean + 2 standard deviations of the reading of the negative controls.

Seroconversions were defined as either a simultaneous increase from negative IgM and IgG in the initial serum sample to positive IgM and IgG at the 6-week sample or an early IgM positive sample with subsequent IgG seroconversion at 6 weeks. Persons who were positive for IgM and IgG in both specimens were defined as having possible seroconversions, because it was unclear whether or not they already had anti-Cryptosporidium IgG before the production of IgM. Persons with evidence of IgG in either or both serum samples without measurable IgM were defined as having evidence of previous infection with Cryptosporidium. Persons without measurable IgG or IgM were classified as having negative serologic results. Persons with any serologic pattern not fitting one of the above categories were classified as having indeterminate serologic results.

Results

Top Methods Results Discussion Author & Article Info References

Description of Study Population

During the study period from December 1990 to April 1992, 31 households were enrolled. Twenty-nine of 31 households were in the primary study community named Goncalves Dias, a community of 1914 inhabitants occupying 405 dwellings in a urban slum (favela) in Fortaleza. The other two households were located in nearby urban slum communities.

The median household size among these 31 families was 7 (range, 3 to 17). The median age of the 31 index cases was 11 months (range, 5 to 28 months), whereas the median age of the 192 family contacts was 21 years (range, 3 months to 81 years). Sixty-four (including index cases) of the 223 total household members (29%) were 5 years or younger.

Study Compliance

During the study period, 33 children were identified with Cryptosporidium infection in the primary study community, and 29 (88%) were enrolled in the study. Four households were not enrolled in the study because of a delay of more than 2 weeks in identifying Cryptosporidium infection. Of the 223 total household members, 218 (98%) adults and children were enrolled. Of these, 154 (71%) submitted all 3 stool specimens, 46 (21%) submitted 2 specimens, 13 (6%) submitted 1 specimen, and 5 (2%) submitted no stool specimens. Two serum collections were done on 167 of 218 (77%) persons participating in the study, although 30 (14%) had only 1 serum collection, and 21 (10%) had no serum collections. A total data set (3 stools and 2 serum specimens) was obtained from 124 of 218 (57%) household members enrolled in the study.

Clinical Characteristics of Index Cases

Twenty-nine of the 31 index cases (94%) had diarrhea at the time of identification of Cryptosporidium infection. Diagnosis of cryptosporidiosis was made a median of 4 days after the onset of diarrhea. Table 1 shows the clinical characteristics of the index cases of Cryptosporidium. The median duration of index diarrhea was lengthy (13 days), although the range (1 to 84 days) was highly variable. Of the 29 index cases with diarrhea, 13 (45%) met the World Health Organization criteria for persistent diarrhea. Concurrent infection with a second enteric pathogen was identified in only one child with Giardia lamblia. The duration of oocyst shedding after the onset of illness was also variable, but most (83%) had at least one other positive stool test result for Cryptosporidium. Despite the many children with protracted diarrhea, however, only 1 child (3%) continued to shed oocysts throughout the entire 6-week study period. These data may be an underestimate of the duration of oocyst shedding because only a single stool specimen was examined at each time point.

Identification of Patients with Secondary Cryptosporidium

In 12 (39%) of 31 households, at least 1 secondary case of Cryptosporidium infection was identified using stool examination. Of the 182 household members (excluding index cases) with at least one stool examination, 17 (9%) were identified with Cryptosporidium infection. Of the 33 nonindex household contacts 5 years or younger, 10 (30%) were diagnosed with Cryptosporidium infection. In contrast, only 7 of 149 (5%) household contacts older than 5 years of age were diagnosed with Cryptosporidium infection using stool examination (P < 0.001). Table 2 shows the comparison of secondary Cryptosporidium cases with Cryptosporidium-negative diarrheal illnesses identified in household members during the study. Nearly half (8 of 17; 47%) of secondary Cryptosporidium cases identified using stool examination were associated with an episode of diarrhea. Although the age range of symptomatic and asymptomatic persons was similar, the median age of persons with Cryptosporidium infection and diarrhea was younger (29 months compared with 7 years), and a higher percentage were children 5 years or younger. Symptomatic secondary cases were identified a median of 5 days after the index diagnosis (range, 0 to 13 days), whereas asymptomatic infections were identified at a median of 15 days (range, 0 to 41 days) (P 0.02).

Of the 32 persons with one or more diarrheal episodes, 27 persons (or episodes) were sampled and 8 of 27 (30%) were positive for Cryptosporidium. As shown in Table 2, the median age for the 19 persons with diarrheal episodes who were sampled and found to be negative for Cryptosporidium was 7 years, with a range of 3 months to 69 years. The median onset of these negative diarrheal episodes after the onset of index diarrhea was 17 days, in contrast with the diarrheal episodes positive for Cryptosporidium, which occurred at a median of 5 days (P 0.004). The duration of episodes of cryptosporidial and noncryptosporidial diarrhea was similar; the median duration was 3 days for each.

Serologic Analysis

Of the 31 index children in the study, 19 (61%) had at least two serum specimens available for analysis. Of these 19 children, 18 had symptomatic Cryptosporidium infection and 1 had asymptomatic Cryptosporidium infection. Fifty-three percent of children (n = 10) seroconverted in association with documented Cryptosporidium infection using stool examination. Despite the young age of the index population (median age, 11 months), 5 (26%) developed symptomatic cryptosporidiosis with serologic evidence (IgG) of previous Cryptosporidium infection. Only one child had two serum specimens that were negative for IgM or IgG or both. In three children, the serologic test results were classified as indeterminate. Interestingly, when children with one or more serum specimens available for evaluation were analyzed, 21 of 23 (91%) had measurable levels of anti-Cryptosporidium IgM or IgG antibodies or both.

Of the 187 household contacts in the study, 148 (79%) had paired sera of sufficient volume for evaluation of anti-Cryptosporidium IgM and IgG. Table 3 shows the results from the paired serologic evaluation of household contacts with negative stool examinations for Cryptosporidium compared with those who had symptomatic and asymptomatic infection as diagnosed using stool examination. These data show that 13 of 134 household contacts with negative stool examinations (10%) seroconverted compared with 3 of 7 (43%) and 1 of 7 (14%) in the symptomatic and asymptomatic cryptosporidiosis categories, respectively. If the seroconversion and possible seroconversion categories were combined, 11 of 14 (79%) with positive stool examinations for Cryptosporidium oocysts and 56 of 134 (42%) with negative stool examinations for Cryptosporidium oocysts seroconverted, which was consistent with greater antigenic stimulation in those household contacts with positive stool test results. In 43 of 134 (32%) of those household contacts with negative stool test results, there was serologic evidence of previous infection with Cryptosporidium compared with 1 of 7 (14%) for persons with either symptomatic or asymptomatic cryptosporidiosis. The 26 persons with indeterminate serologic test results had IgM or IgG or both in at least one serum sample consistent with past or recent exposure to Cryptosporidium. Ten contacts (8%) with negative stool test results had no evidence of anti-Cryptosporidium IgM or IgG, whereas no persons with positive stool test results had negative serologic results.

Overall, 202 family members (index cases and family contacts) in this study had at least one serum specimen analyzed for the presence of IgM and IgG antibodies to Cryptosporidium. Of these persons, only 11 (5.4%) had no serologic evidence of infection (previous or recent) with Cryptosporidium.

Discussion

Top Methods Results Discussion Author & Article Info References

Cryptosporidium continues to be a threat not only among immunocompromised persons but also among young children in both developed and developing countries. Recent work in the United States has suggested an increasing prevalence of Cryptosporidium infection among daycare attendees [33]. Further, a water-borne outbreak involving an estimated 370 000 persons in Milwaukee, Wisconsin, in April 1993, showed that this parasite is a continuing public health threat, even in the developed world [7]. However, the person-to-person transmissibility of this organism has remained unclear.

These data suggest a high rate of person-to-person transmission of Cryptosporidium based on two sets of criteria: stool examination and serologic test results. Alternatively, these data indicate a high rate of concomitant infections because of a common source of exposure to Cryptosporidium oocysts. Regardless of the source of the oocysts, these data indicate that Cryptosporidium is highly infective. Overall, 18 (58%) of the 31 households in the study had at least one secondary case of cryptosporidiosis identified by either a positive stool test result or by a documented seroconversion (using the most stringent criteria). Nineteen percent of family contacts had documented Cryptosporidium infection using these same criteria. In addition, these data also show one of the highest rates of seropositivity (94.6%) to Cryptosporidium ever reported, indicating that the endemicity of this infection in this community is consistent with our previous observation [25] that 18% of tested water sources and 10% of animals are positive for Cryptosporidium in this community.

Despite the high rate of secondary transmission identified in this study, diarrhea did not occur in most of the secondary cases (8 of 30 [27%]) and was present overall in only 8 of 182 (4%) household members. Further, recognized diarrheal illnesses caused most likely by other enteric pathogens (n = 19) (see Table 2) exceeded diarrheal illnesses identified as cryptosporidiosis [n = 8] Table 2 in household members. Thus, in this community where cryptosporidiosis is highly endemic, transmission of Cryptosporidium infection occurred most often without substantial clinical disease. One possible explanation for this finding is that Cryptosporidium infection is so endemic in this area that many household members have preexisting protective immunity. This is also suggested by our data showing a younger median age of household members with symptomatic cryptosporidiosis compared with those who have asymptomatic cryptosporidiosis (Table 2). However, the clinical impact of secondary transmission of Cryptosporidium infection deserves further study in communities where the infection is not highly endemic, such as occurred in the recent Milwaukee water-borne outbreak [7]. In communities where this parasite infection is not endemic, we hypothesize that diarrheal illness because of secondary transmission of Cryptosporidium infection may be more common.

This study also showed that Cryptosporidium infection was associated with persistent diarrhea in nearly half of all index cases. The onset of diarrhea in symptomatic secondary cases was temporally related to the index cases. Asymptomatic secondary cases, however, occurred significantly later. Although the age ranges of symptomatic and asymptomatic patients with Cryptosporidium infection were similar, the median age was lower in the symptomatic group. The asymptomatic nature of the infections and the older age of this group may indicate the presence of at least partially protective immunity among the older, asymptomatic group.

The serologic patterns in this community showed that a large percentage of persons in this study had evidence of anti-Cryptosporidium IgM and IgG in early and late serum specimens. These data suggest that the secondary infection rate may be far higher than this conservative estimate. Because of delays in the collection of the initial serum samples (median, 3 days; range, 0 to 20 days), a de novo IgG response may already have been generated. Alternatively, these persons may have reinfection with an antigenically distinct strain of Cryptosporidium and, therefore, are generating an IgM response superimposed on an already positive IgG response. The other possibility is that IgM responses are prolonged after infection with Cryptosporidium. Previous serologic studies [34] of Cryptosporidium infection have reported prolonged IgM responses, sometimes without subsequent IgG seroconversion. In fact, persistent IgM responses at 6 weeks were measured in 86% of our study population with two serum samples. In addition, in 9 household members with a third serum specimen obtained 12 months after study entry, 7 (78%) showed persistent IgM seropositivity. Prolonged IgM responses have also been observed after several viral infections, including rubella [35] and hepatitis A [36]. In addition, other parasites, such as Trypanosoma brucei and Trypanosoma cruzi, have been shown to grossly increase IgM levels secondary to polyclonal B-cell expansion [37]. This latter possibility is unlikely in our study population because the specificity of the IgM measured by the Cryptosporidium parvum antigen preparation was determined by several methods.

Even using stringent criteria of analysis, 19% of persons in households with an identified case of cryptosporidiosis also became infected as shown by either a positive stool examination or by a documented seroconversion. This percentage is similar to transmission rates for other highly transmissible enteric pathogens. Shigella species infect 20% of persons in households with an identified case, although that study [38] defined a positive case as any family member with a febrile diarrheal illness, whether or not they had documented positive test results from stool culture. Rotavirus has a transmission rate approaching 20% in children admitted to diarrheal wards in Santiago, Chile [39].

Although cryptosporidiosis is typically a self-limited disease in immunocompetent persons, these data show that it contributes substantially to the diarrheal burden of children 5 years old or younger in this poor socioeconomic community. Urban slum communities in Brazil without access to education about the acquired immunodeficiency syndrome (AIDS) or to barrier prophylactic methods are at particularly high risk for infection with HIV—the same type of community that has endemic and highly transmissible infection with Cryptosporidium. In Brazil (Wuhib T, de Queirez T, and Sears CL. Unpublished observations) and in the United States, Cryptosporidium is the leading parasitic cause of chronic diarrheal disease in patients with AIDS [40-42]. The high rate of transmission of this parasite shown in the present study serves to underscore the highly infective nature of this organism and the difficulty in preventing its spread. In a country where persistent diarrhea remains a serious threat to the health of young children [43-45] and the rate of HIV seropositivity is increasing sharply, these data are cause for great concern.

Presented in part at the annual meeting of The American Society of Tropical Medicine and Hygiene in Seattle, Washington, 17 November 1992.