Cryptosporidium is an important opportunistic path- have low CD4+ cell counts, preventing infections de-
ogen affecting HIV-infected persons, and it has been pends on avoiding exposure to the parasite and main-
associated with chronic diarrhea [1, 2], decreased qual- taining immune competence. In developed countries,
ity of life, and shortened survival in HIV-positive pa- access to highly active antiretroviral therapy (HAART)
tients [3, 4]. Because no antiparasitic agent is effective has reduced the morbidity from cryptosporidiosis [3–
against cryptosporidiosis in patients with AIDS who 5]. Nonetheless, infections with Cryptosporidium spe-
cies are still a major threat to patients with AIDS who
do not have access to HAART, especially in developing
Received 10 October 2006; accepted 15 March 2007; electronically published countries [6, 7].
13 July 2007.
Potential conflicts of interest: none reported. As with other opportunistic protozoa, there is limited
Financial support: Opportunistic Infections Working Group of the Centers for knowledge about the transmission dynamics and path-
Disease Control and Prevention (CDC); RG-ER Fund; National Institute for Allergy
and Infectious Disease, National Institutes of Health (projects 5P01AI051976-04 ogenicity of the different species and subtypes of Cryp-
and 5R21AI059661-02 to R.H.G. and V.A.C.); Division of Parasitic Diseases, CDC tosporidium species. One study in the United Kingdom
(Research Participation Program appointment to S.C. administered by the Oak
Ridge Institute for Science and Education through an interagency agreement
reported that infections with C. hominis in immuno-
between the US Department of Energy and the CDC). competent persons were associated with extraintestinal
The findings and conclusions in this article are those of the authors and do
not necessarily represent the views of the Centers for Disease Control and
sequelae [8]. A small study of cryptosporidiosis in HIV-
Prevention. infected patients in Tanzania demonstrated potential
Reprints or correspondence: Lihua Xiao, Div. of Parasitic Diseases, Centers for
clinical differences between 15 persons infected with C.
Disease Control and Prevention, 4770 Buford Hwy. NE, MS-F12, Atlanta, GA 30341
(lxiao@cdc.gov). hominis and 6 infected with C. parvum [9]. Recently,
The Journal of Infectious Diseases 2007; 196:684–91 a study in Brazil showed that children infected with C.
2007 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2007/19605-0006$15.00
hominis or C. parvum had reductions in their anthro-
DOI: 10.1086/519842 pometric measurements, but long-term effects were
Subjects, Episodes of
Parameter no. diarrhea, no. (%) OR (95% CI) P
Model 1: risk of diarrhea by microscopy
Cryptosporidium speciesa 230 87 (38) 1.5 (1.1–2.0) .019
b
NOTE. All models were controlled for source of patients, CD4+ cell counts, and infections with Isospora belli, Cyclospora
cayetanensis, or Enterocytozoon bieneusi.
a
The model excluded participants with acute diarrhea.
b
Statistically significant at a p .05.
c
Statistically significant at a p .0071 (Bonferroni’s correction).
children [13]. Altogether, these findings are in agreement with role in the presentation of clinical symptoms. Our data showed
the suggestion that socioeconomic and geographic differences that, among genotypes, C. parvum was probably the most path-
affect the distribution of Cryptosporidium genotypes and sub- ogenic because of its significant association with chronic di-
type families of C. hominis and C. parvum [23]. arrhea and vomiting, followed by C. canis and C. felis, which
It has been frequently reported that not all HIV-infected had significant associations with diarrhea. Contrarily, C. me-
persons with cryptosporidiosis have diarrhea [26–30]. These leagridis appeared to be the less pathogenic, given that it was
differences in clinical manifestations were previously attributed not associated with the symptoms evaluated in this study.
to the immune status of the person, and persons with CD4+ We also observed differences in clinical manifestations among
cell counts !180 cells/mm3 were most likely to have chronic subtype families of C. hominis: persons infected with subtype
cryptosporidiosis [31], as well as other opportunistic infections family Id were more likely to have an increased risk for diarrhea,
[32]. However, our results suggested that genetic differences in whereas infections with subtype family Ib were marginally as-
the parasite, such as genotypes or subtype families, also play a sociated with diarrhea. By contrast, subtype family Ia was not
Subjects,
Parameter no. (%) OR (95% CI) P
Model 7: risk of vomiting based on microscopy
Cryptosporidiumspecies 188 (44) 1.5 (1.1–2.1) .009a
No Cryptosporidium 1718 (29) Referent
Model 8: risk of vomiting by genotypes
C. hominis 117 (40) 1.3 (0.9–1.9) .214
b
C. parvum 15 (80) 7.2 (2.0–25.8) .003
C. meleagridis 15 (33) 1.0 (0.3–3.1) .941
C. felis or C. canis 12 (42) 1.3 (0.4–4.4) .622
No Cryptosporidium 1718 (29) Referent
Model 9: risk of vomiting by subtype families
of C. hominis
Subtype families
Ia 30 (27) 0.7 (0.3–1.7) .468
Ib 37 (49) 1.8 (0.9–3.6) .077
NOTE. All models were controlled for source of patients, CD4+ cell counts, and infections with
Isospora belli, Cyclospora cayetanensis, or Enterocytozoon bieneusi. Not all patients provided data
about vomiting. Data from variables not having significant associations were not presented.
a
Statistically significant at a p .05.
b
Statistically significant at a p .0071 (Bonferroni’s correction).
associated with diarrhea. These findings are in disagreement symptoms previously associated with cryptosporidiosis, such as
with those of a previous small-scale study of hospitalized HIV- fever, acid reflux, and muscle and joint pain [36] or extraintestinal
infected South African children with diarrhea and cryptospo- sequelae [8], and frequently reported in infections with other
ridiosis, in which infections with all subtype families had similar pathogens were not corroborated by our findings.
symptoms [33]. Overall, our findings suggest that infections with C. hominis
Our study also found that infections with C. meleagridis were subtype family Id, C. parvum, C. canis, or C. felis can severely
not associated with any of clinical manifestations, which is in affect HIV-infected persons, because chronic diarrhea can lead
contrast to the results of a previous Portuguese study in which to wasting syndrome and eventually death [2, 34, 35]. Because
a few HIV-infected persons with C. meleagridis had diarrhea, infections with C. parvum subtype family IIc were also asso-
despite having been prescribed antiretroviral therapies. We have ciated with vomiting, HIV-infected patients also infected with
corroborated their finding that persons infected with C. me- this subtype may be at higher risk for severe complications and
leagridis excreted fewer parasites than those infected with other can be considered for receiving more aggressive antiretroviral
Cryptosporidium species [37]. This apparent contradiction may therapies.
be due to the endemicity of C. meleagridis in Peru [11, 18], The analyses of risk factors identified that persons who had
where it seems to be more prevalent than in most other geo- contacts with children !5 years of age were more likely to be
graphical locations [37–39], and to the small number of cases infected with C. hominis subtype family Ie. This finding is in
in the Portuguese study. Because of the cross-sectional nature concordance with the anthroponotic nature of C. hominis. Our
of our study, we can only hypothesize that the lack of associated study did not categorically identify other infection risk factors,
clinical manifestations could be a consequence of prior infec- and this was probably due to a multitude of reasons, including
tions and subsequent amelioration of symptoms [40, 41] or of the cross-sectional nature of our study design to evaluate
the genetic uniqueness of C. meleagridis in Peru. chronic infections, the use of a very stringent significance level
It is also conceivable that persons in the study might have in our analyses (Bonferroni-adjusted a levels), and the fact that
been infected with other bacterial or viral pathogens. Neverthe- our study was conducted in an area of endemicity, where ex-
less, there is no evidence suggesting that Cryptosporidium-in- posure to Cryptosporidium may occur frequently and through
fected persons were more likely to have other enteropathogens multiple routes.
that might have caused diarrhea or vomiting. Furthermore, other Although not statistically significant after the Bonferroni cor-