Background

Despite the advance of antiretroviral therapy (ART), diarrhea is still a common problem of HIV infection and contributes to the reduced life quality and survival of HIV patients [1, 2]. It is estimated that diarrhea occurs in roughly 90% HIV/AIDS patients in developing countries, and 30–60% in developed countries [3]. Opportunistic pathogens that cause diarrhea in HIV-infected people include protozoa, fungi, viruses, and bacteria [4]. Several protozoan species belonging to Cryptosporidium, microsporidia and Isospora, are among the most common causative pathogens responsible for significant morbidity and mortality in HIV patients [5].

With a worldwide distribution of Cryptosporidium, C. parvum and C. hominis are the most common species detected in humans, though other species, including C. meleagridis, C. felis and C. canis, have also been reported [6]. Despite the use of ART in many countries of the world, the infection rates of Cryptosporidium in HIV patients are still high, accounting for up to a third of diarrhea cases in HIV patients [7].

Microsporidia are obligate intracellular eukaryotic pathogens, which are phylogenetically related to fungi, and have been considered as opportunistic infections in both developed and developing countries, especially in HIV patients with a CD4 cell count below 100 cells/μl [8]. Of the 15 species of microsporidia that infect humans, Enterocytozoon bieneusi and Encephalitozoon intestinalis can cause gastrointestinal diseases, with E. bieneusi being the more commonly identified species in HIV-infected people [9].

Isospora belli is the only species of the genus Isospora, and is frequently found in HIV-infected people of tropical and subtropical regions, accounting for up to 20% of diarrhea cases in AIDS patients [7]. The species can cause acute self-limiting diarrhea in immunocompetent individuals, but in severely immunocompromised patients, this parasite can cause severe chronic diarrhea which may result in a wasting syndrome, or even the death of AIDS patients [10].

The opportunistic parasites Cryptosporidium spp., microsporidians and Isospora spp. develop in enterocytes, and are excreted via feces and transmitted through the fecal-oral route via ingestion of contaminated water or food, or direct contact with infected animals or humans [11]. HIV-infected people are more likely to develop abrupt, severe, and explosive diarrhea when infected with opportunistic protozoa than immunocompetent individuals. Millions of people are affected by the morbidity caused by these parasites, as there was an estimated 36.7 million people living with HIV in 2015 worldwide [12]. Since there is no reliable or well-defined treatment for the protozoan infections in immunocompromised patients [1], understanding their epidemiology is central in formulating effective control strategies against cryptosporidiosis, microsporidiosis, and isosporiasis in these populations. We undertook a systematic review and meta-analysis to evaluate the worldwide prevalence of Cryptosporidium, microsporidia and Isospora infection in people with HIV.

Methods

Search strategy

We searched PubMed, ScienceDirect, Google Scholar, Embase, Chinese Web of Knowledge, Wanfang, and Chongqing VIP databases for studies reporting Cryptosporidium, microsporidia, or Isospora infection in HIV-infected people from inception to 31 December 2016. The databases were searched using the term “Cryptosporidium”, “cryptosporidiosis”, “microsporidia”, “microsporidiosis”, “Isospora” or “isosporiasis” cross-referenced with “HIV”, “immunodeficiency”, “acquired immune deficiency syndrome”, or “AIDS”, without language restriction. We did our analyses according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [13] (see PRISMA checklist in Additional file 1: Table S1).

Selection criteria

The included studies were required to investigate HIV-infected people and needed to have data that allowed us to calculate the prevalence of Cryptosporidium, microsporidia, and Isospora infection. We excluded studies if they were reviews, animal studies, or repeated studies; if there were no raw data; if the sample size was less than 20; or if the diagnostic methods of parasite infection were unclear.

Two independent reviewers (LZ and SL) carefully examined all titles and abstracts identified in the search, and assessed the full text considered potentially relevant. Any disagreements were resolved by discussion with other two authors (Z-DW and H-HL).

Data analysis

Two reviewers (Z-DW and SL) extracted the information about the first author, publication year, country of the study, numbers of HIV-infected people and Cryptosporidium, microsporidia, or Isospora co-infected people, diagnostic methods, study design, and demographic characteristics from each eligible study, and reached a consensus after discussing any controversial finding.

We assessed the quality of the included publications on the basis of criteria derived from the Grading of Recommendations Assessment, Development and Evaluation method [14]. We used a scoring approach to grade quality. Studies were given one point each if they had probability sampling, larger sample sizes of more than 200, and repeated detection. Up to four points could be assigned to each study. We regarded publications with a total score of three or four points to be of high quality, whereas two points represented moderate quality and scores of one or zero represented low quality.

We did a meta-analysis by a random-effects model or fixed-effects model to calculate the pooled prevalence of Cryptosporidium, microsporidia, or Isospora infection using Stata version 12.

The heterogeneity between studies was evaluated using Cochran’s Q and the I2-statistic, which presents the percentage of variation between studies. Due to high heterogeneity (I2 > 50%, P < 0.1), random effects models were used for summary statistics. A potential source of heterogeneity was investigated by subgroup analysis and meta-regression analysis. We examined factors both individually and in multiple-variable models to determine the possible factors that caused heterogeneity in our study. The factors included geographical region by comparison of sub-Sahara Africa with other regions, income level by comparison of low-income countries with others, and patients with diarrhea by comparison of patients with diarrhea with others. We also evaluated the effect of selected studies on the pooled prevalence by excluding single studies sequentially. A study was considered to have no influence if the pooled estimate without it was within the 95% confidence limits of the overall prevalence [15].

Results

Our research identified 2785 records. After initial screening and removal of duplicates, 193 papers were reviewed in full. Of these, 51 articles did not include sufficient data that were required or conform to the criteria, 13 were unavailable for full text, five had duplicate samples, and two included the sample size of less than 20. After an updated search, nine papers were included and we had 131 articles for quality assessment and meta-analysis (Fig. 1).

Fig. 1
figure 1

Flowchart of the study selection process

Full size image

According to our criteria, 51 publications were of high quality with a score of three or four, 48 had a score of two indicating moderate quality, and the remaining 32 were of low quality with a score of zero or one (Tables 1, 2 and 3).

Table 1 Included studies of Cryptosporidium infection in people with HIV listed in order of year published
Full size table
Table 2 Included studies of microsporidia infection in people with HIV listed in order of year published
Full size table
Table 3 Included studies of Isospora infection in people with HIV listed in order of year published
Full size table

One hundred and six studies assessed Cryptosporidium infection in HIV-infected people (Fig. 1, Table 1), including a total of 43,218 HIV-infected patients. These studies were done in 36 countries (Fig. 2), including five countries of western and central Europe and North America, 15 of sub-Saharan Africa, four of Latin America and the Caribbean, two of eastern Europe and central Asia, nine of Asia and the Pacific, and one of Middle East and North Africa. Of these identified studies, 16 were done in low-income countries, 76 were in middle-income countries, and 14 were in high-income countries (Fig. 2). Ninety-eight papers were written in English, and eight in Chinese [16,17,18,19,20,21,22,23].

Fig. 2
figure 2

Map of Cryptosporidium infection in HIV-infected people worldwide. Pooled percentage prevalence and 95% CI are shown for each country

Full size image

The prevalence of Cryptosporidium infection ranged between 0 and 78.1% (Fig. 3). Meta-analysis by random-effect model showed that the estimated pooled prevalence of Cryptosporidium infection in people with HIV infection was 14.0% (3283/43,218; 95% CI: 13.0–15.0%) overall, 21.1% (1105/5315; 95% CI: 16.1–21.1%) in sub-Saharan Africa, 7.3% (1042/28,283; 95% CI: 5.4–9.2%) in western and central Europe and North America, 12.6% (896/7529; 95% CI: 10.5–14.7%) in Asia and the Pacific, 13.0% (121/1272; 95% CI: 7.3–18.7%) in Latin America and the Caribbean, 4.7% (43/702; 95% CI: 0.5–8.8%) in the Middle East and North Africa, and 60.1% (76/119; 95% CI: 24.1–96.1%) in eastern Europe and central Asia. Only four studies were done in Middle East and North Africa, and two in eastern Europe and central Asia, where the prevalence of Cryptosporidium infection in HIV-infected people was very poorly recorded.

Fig. 3
figure 3

Random-effect meta-analysis of Cryptosporidium infection in HIV-infected people

Full size image

With a substantial heterogeneity (I2 = 97.6%, P < 0.0001; Table 4), meta-regression analyses showed that geographical distribution (P = 0.039) and patients with diarrhea (P = 0.009) might be sources of heterogeneity, whereas we detected no significant differences in income levels (P = 0.328). Subgroup analysis showed the pooled prevalence of Cryptosporidium infection in HIV-infected people was significantly lower in western and central Europe and North America than in sub-Saharan Africa (OR 0.73, 95% CI: 0.54–0.99, P = 0.044), and higher in patients with diarrhea (OR 1.21, 95% CI: 1.00–1.46, P = 0.047).

Table 4 Pooled prevalence of Cryptosporidium infection in HIV-infected patients
Full size table

Forty-seven studies reported prevalence of microsporidia (Fig. 1, Table 2), including a total of 18,006 HIV-infected people tested for microsporidia infection. The included studies were conducted in 23 countries (Fig. 4), including 11 countries of sub-Saharan Africa, four of western and central Europe and North America, three of Asia and the Pacific, three of Latin America and the Caribbean, one each of Middle East and North Africa and eastern Europe and central Asia. Of the identified studies, 9 were done in low-income countries, 30 were in middle-income countries, and 9 were in high-income countries (Fig. 4). Forty-five papers were written in English, one each in Chinese and French [17, 24].

Fig. 4
figure 4

Map of microsporidia infection in HIV-infected people worldwide. Pooled percentage prevalence and 95% CI are shown for each country

Full size image

The prevalence of microsporidia infection ranged between 0.7–81.3% (Additional file 2: Figure S1). Meta-analysis by random-effect model indicated that the estimated pooled prevalence of microsporidia infection in people with HIV infection was 11.8% (1090/18,006; 95% CI: 10.1–13.4%) overall, 15.4% (425/3834; 95% CI: 11.1–19.7%) in sub-Saharan Africa, 14.4% (277/8089; 95% CI: 7.8–21.1%) in western and central Europe and North America, 11.7% (251/2791; 95% CI: 8.2–15.1%) in Asia and the Pacific, 5.6% (123/2890; 95% CI: 1.9–9.3%) in Latin America and the Caribbean, 2.2% (8/356; 95% CI: 0.7–3.8%) in the Middle East and North Africa, and 13.0% (6/46; 95% CI: 3.3–22.8%) in eastern Europe and central Asia. Only three studies were done in Latin America and the Caribbean, one each in Middle East and North Africa, and in eastern Europe and central Asia. The prevalence of microsporidia infection in these regions should be interpreted with caution.

Due to the substantial heterogeneity (I2 = 96.7%, P < 0.0001; Table 5), meta-regression analyses indicated that the income level (P = 0.024) and patients with diarrhea (P = 0.004) might be sources of heterogeneity, whereas we detected no significant differences in geographical distribution (P = 0.323). Subgroup analysis showed the pooled prevalence of microsporidia infection in HIV-infected people was significantly higher in low-income countries than in middle-income countries (OR 1.58, 95% CI: 1.08–2.31, P = 0.018), and higher in patients with diarrhea than the control (OR 1.54, 95% CI: 1.14–2.07, P = 0.005).

Table 5 Pooled prevalence of microsporidia infection in HIV-infected patients
Full size table

Fifty-eight studies tested 105,922 HIV-infected patients for Isospora infection (Fig. 1, Table 3). The selected studies were done in 20 countries (Fig. 5), including three countries of western and central Europe and North America, eight of sub-Saharan Africa, five of Asia and the Pacific, three of Latin America and the Caribbean, and one of Middle East and North Africa. No studies were found from eastern Europe and central Asia. Of the identified studies, 12 were done in low-income countries, 39 were in middle-income countries, and seven were in high-income countries (Fig. 5). All the included papers were written in English.

Fig. 5
figure 5

Map of Isospora infection in HIV-infected people worldwide. Pooled percentage prevalence and 95% CI are shown for each country

Full size image

The prevalence of Isospora infection ranged between 0.2–26.9% (Additional file 3: Figure S2). Meta-analysis by random-effect model showed that the estimated pooled prevalence of Isospora infection in people with HIV infection was 2.5% (788/105,922; 95% CI: 2.1–2.9%) overall, 6.1% (232/3995; 95% CI: 4.5–7.7%) in sub-Saharan Africa, 0.5% (324/97,721; 95% CI: 0.2–0.8%) in western and central Europe and North America, 7.1% (215/2848; 95% CI: 5.2–9.0%) in Asia and the Pacific, 1.0% (13/922; 95% CI: 0.3–1.7%) in Latin America and the Caribbean, 0.8% (4/436; 95% CI: 0–2.0%) in the Middle East and North Africa. However, few data were available from Latin America, Middle East and North Africa. Only two studies were conducted in Middle East and North Africa, five were done in Latin America and the Caribbean, showing a poor record of Isospora infection in these regions.

With a substantial heterogeneity (I2 = 89.8%, P < 0.0001; Table 6), meta-regression analyses showed that patients with diarrhea might be sources of heterogeneity (P = 0.005), whereas we detected no significant differences in region distribution (P = 0.143) and income levels (P = 0.806). Subgroup analysis showed that the pooled prevalence of Isospora infection in HIV-infected people was significantly lower in central Europe and North America than in sub-Saharan Africa (OR 0.40, 95% CI: 0.27–0.59) and in Asia and the Pacific (OR 0.37, 95% CI: 0.26–0.54). Additionally, it was significantly higher in low-income countries (OR 1.94, 95% CI: 1.24–3.04, P = 0.005) and middle-income countries (OR 2.08, 95% CI: 1.41–3.07, P < 0.0001) than in high-income countries. We also found that patients with diarrhea had a higher prevalence of Isospora infection (OR 1.53, 95% CI: 1.14–2.06, P = 0.005).

Table 6 Pooled prevalence of Isospora infection in HIV-infected patients
Full size table

We determined the effect of selected studies on the pooled prevalence by excluding single studies sequentially, and found no significant effect of study quality on prevalence of Cryptosporidium and microsporidia infection in HIV-infected people (all P > 0.05), but there was significant effect of study quality on the prevalence of Isospora infection (P = 0.033 and 0.043).

When we excluded the studies by Sorvillo et al. [25], Guiguet et al. [26], and Lagrange-Xelot et al. [27], the pooled prevalence of Isospora infection in HIV-infected people was increased from 2.5% (95% CI: 2.1–2.9%) to 3.0% (95% CI: 2.5–3.5%), 3.3% (95% CI: 2.8–3.8%), and 3.0% (95% CI: 2.5–3.4%), respectively. These findings indicated that the pooled prevalence of Isospora infection in HIV-infected people was substantially influenced by the three studies, and adjusted to 5.0% (469/8570; 95% CI: 4.1–5.9%) by excluding these studies (Additional file 4: Figure S3).

Discussion

Our aim was to estimate the worldwide prevalence of opportunistic intestinal protozoa in people with HIV, showing that Cryptosporidium and microsporidia are the main intestinal protozoa in HIV-infected people, followed by Isospora; their prevalences are usually high in sub-Saharan Africa and in patients with diarrhea, and low in high-income countries. Because of the large proportion of low-income countries and the large number of people with HIV [12], sub-Saharan Africa has a very high burden of Cryptosporidium, microsporidia and Isospora infection, reinforcing the importance of routine testing for opportunistic intestinal protozoa in all HIV-infected people. To our knowledge, this is the first systematic review and meta-analysis of the global prevalence of Cryptosporidium, microsporidia and Isospora infection in HIV-infected people.

Our findings corroborate evidence for a high prevalence of Cryptosporidium, microsporidia and Isospora infection in Africa and a low prevalence in Europe. In HIV-infected people, a high prevalence has been reported in Uganda (73.6%) [28], Nigeria (79.0%) [29], and South Africa (75.6%) [30] for Cryptosporidium infection; in Zimbabwe (50.9%) [31] and Uganda (76.9%) [28] for microsporidia infection; and in DR Congo (19.6%) [32] and Nigeria (24.3%) [33] for Isospora infection. In contrast, a low prevalence has been shown in France (1.3%) [27], USA (5.4%) [34] and Portugal (7.7%) [35] for Cryptosporidium infection; in France (0.8%) [27] and USA (5.9%) [36] for microsporidia infection; and in France (0.8%) [37] and Italy (0.6%) [38] for Isospora infection.

The incidence of opportunistic intestinal protozoa infection varies, relying on sanitation facilities, drinking contaminated water, animal exposure, CD4 T cell count, ART, diagnostic methods [39, 40]. Thus, the prevalence of infection may vary substantially, even within a country or among different populations of the same region. For example, in the USA, the prevalence of Cryptosporidium infection is 3.8% in Los Angeles [41], 5.4% in San Francisco [34] and 10.8% in Houston [42]. Large differences of Isospora infection have also been reported in India, with a prevalence of 3.4% in Coimbatore [43], 8.0% in Pune [44] and 25.0% in New Delhi [45]. There are significant differences between different countries for Cryptosporidium (0–78.1%), microsporidia (0.7–81.3%) and Isospora (0.2–26.9%) infection in HIV-infected people. However, limited country-level surveys of Cryptosporidium, microsporidia and Isospora infection have been undertaken, making it difficult to compare the infections between regions or populations.

The majority of the studies had additional data on opportunistic intestinal protozoa. Due to the variability of data quality and reporting consistency, we only extracted and analyzed the data on diarrhea, and demonstrated it was related to Cryptosporidium (OR: 1.21, 95% CI: 1.01–1.46, P = 0.047), microsporidia (OR 1.53, 95% CI: 1.13–2.07, P = 0.007) and Isospora (OR 1.53, 95% CI: 1.14–2.06, P = 0.005) infection in HIV-infected people in comparison with their controls. Moreover, there were some case-control studies that investigated opportunistic intestinal protozoa infection in people with HIV with and without diarrhea. We analyzed the association of diarrhea with Cryptosporidium, microsporidia and Isospora infection in HIV-infected people. The estimated pooled random effects ORs of Cryptosporidium, microsporidia and Isospora infection in HIV people with diarrhea compared with their controls were 4.09 (95% CI: 2.32–7.20), 4.72 (95% CI: 3.47–6.42), and 4.93 (95% CI: 3.33–7.29), respectively (Additional files 5, 6 and 7: Figures S4, S5 and S6). These findings show that diarrhea is associated with opportunistic intestinal protozoa infection in HIV people. However, other factors seem to increase the likelihood of infection with opportunistic intestinal protozoa, including CD4 T-lymphocyte counts of less than 100 cells/μl [46], ingestion of contaminated drinking water or food [47], exposure to infected pets or animals [48] and unsafe homosexual activity [49].

There are a few limitations of the present meta-analysis, which may affect the results. First, many relevant studies were identified through our literature search, but not all data were available; there is a possibility that some qualified data were missed. Secondly, the majority of the studies were of moderate or low quality, as most of the data resulted from the conventional microscopic diagnostic techniques; these have a sensitivity which is inferior to polymerase chain reaction, ELISA and direct fluorescent-antibody tests. Additionally, most studies examined a single stool specimen, potentially leading to a false negative result. This means that the reported prevalence was possibly underestimated. Thirdly, the included studies were concentrated in Asia (n = 50), sub-Saharan Africa (n = 45), and western and central Europe and North America (n = 17), Latin America and the Caribbean (n = 12), with few studies from Middle East and North Africa (n = 5), and eastern Europe and central Asia (n = 2), and the study quality was variable, emphasizing the need for more robust surveillance of Cryptosporidium, microsporidia and Isospora infection in HIV-infected people in these regions. Fourthly, different species and genotypes of Cryptosporidium and microsporidia may cause different clinical manifestations in HIV-infected people [40, 50]. However, we did not analyze their distribution characteristics as the microscopic diagnostic techniques in most of the selected studies could not identify the species within the genus Cryptosporidium and microsporidians.

To explain the specific causes of heterogeneity, we did univariate meta-regression analyses on various sources including geographical distribution, income level, and patients with diarrhea, and found different main causes of heterogeneity for the three opportunistic protozoa. These may come from geographical distribution (P = 0.039) and patients with diarrhea (P = 0.009) for Crytosporidium infection, from income level (P = 0.024) and patients with diarrhea (P = 0.004) for microsporidia infection, and from patients with diarrhea (P = 0.005) for Isospora infection. Other potential causes of heterogeneity may include publication year, sample size, and detection methods. Unfortunately, we did not analyze them, as there were not enough data available.

Moreover, we did dummy variable analysis on geographical distribution, income level, and patients with diarrhea. The countries in sub-Saharan Africa had a higher prevalence of Cryptosporidium and Isospora infection in HIV-infected patients than those in western and central Europe and North America, and the low-income countries had a higher prevalence of microsporidia and Isospora infection than the middle or high-income countries. These findings support an association between parasite infection and the income level of countries, which could be due to the fact that people in high-income countries have access to safe water and sanitation facilities, which are responsible for the reduced odds of parasite infection.

Conclusions

The results of our global meta-analysis show a heavy burden of Cryptosporidium, microsporidia and Isospora infection in HIV-infected people, especially in low-income countries and sub-Saharan Africa. Thus, routine screening of opportunistic intestinal protozoa should be done, particularly for those who have CD4 T-lymphocyte count less than 100 cells/μl, and early treatment should be administered. This should include a combination of antibiotics of azithromycin, paramomycin and nitazoxanide for Cryptosporidium infection, albendazole for microsporidia infection, and trimethoprim-sulfamethoxazole for Isospora infection [51, 52]. However, antibiotics alone may not necessarily reduce the symptoms associated with opportunistic intestinal protozoa infection [7, 51]. More importantly, it is obligatory to reconstruct the immune system by ART. Additional preventive measures should also emphasize the environmental and personal hygiene, along with the quality of drinking water [47].