Epidemiology of Haemophilus ducreyi Infections

Infections are at their lowest level worldwide, but nongenital cutaneous infections have increased.

H aemophilus ducreyi, a fastidious gram-negative bacterium, is the causative agent of chancroid, a genital ulcer disease (GUD). The organism is usually spread during sexual intercourse through microabrasions, and the disease usually manifests as multiple painful superficial ulcers associated with inguinal lymphadenitis (1). As a result of the painful nature of the lesions, patients usually seek immediate treatment, and asymptomatic carriage is therefore uncommon (2). In addition to causing GUD, H. ducreyi has been found in several recent studies to be a major cause of chronic skin ulceration in children from developing countries (3)(4)(5)(6).
The global epidemiology of chancroid is poorly documented, and it is not included in World Health Organization estimates of the global incidence of curable sexually transmitted infections (STIs). There are some key challenges in interpreting data on the epidemiology of H. ducreyi as a causative agent of GUD. First, genital herpes cases are easily misdiagnosed as chancroid on clinical examination. Thus, reports based only on clinical diagnosis can be erroneous. Second, laboratory culture is technically difficult, and the highly sensitive and specific nucleic acid amplification tests, such as PCR, are rarely available outside national reference laboratories or specialized STI research settings, which makes it difficult to confirm clinical diagnoses.
Determination of the true global incidence of chancroid is made more difficult by widespread adoption of syndromic management for bacterial GUD (i.e., treatment with antimicrobial drugs effective against syphilis and chancroid) without microbiological confirmation in many countries. Therefore, countries often report only the total number of GUD cases. In addition, identification of GUD etiology is rarely conducted in resource-poor countries to validate syndromic management for which chancroid could also be common.
Earlier studies of tropical skin ulcers did not generally test for H. ducreyi, with the exception of a small number of case reports (7)(8)(9)(10)(11). There are major limitations in describing the prevalence of causative agents in tropical skin lesions that typically occur in children in rural areas where there is no access to laboratory facilities. Pathogens such as Fusobacterium fusiforme, Staphylococcus aureus, and Streptococcus pyogenes have been reported from Gram staining of exudative material collected from tropical ulcers (12). However, cultures or PCR testing for definitive identification of fastidious pathogens involved has not been traditionally conducted. The purpose of this study was to improve our understanding of the epidemiology of H. ducreyi infection through a systematic review of published data on the proportion of genital and skin ulcers caused by this bacterium.

Search Strategy and Selection Criteria
A systematic review was conducted to identify all relevant studies that examined the etiology of GUD and nongenital skin ulcers involving H. ducreyi. We searched the National Library of Medicine through PubMed for "H. ducreyi," "chancroid," "genital ulcer," OR "skin ulceration" AND "proportion" OR "prevalence." The search was limited to studies published during January 1, 1980-December 31, 2014. In addition, we searched references of identified articles and other databases for other articles, and we reviewed abstracts, titles, and selected studies potentially containing information on chancroid epidemiology. We contacted researchers who were working with H. ducreyi to identify unpublished literature for inclusion. No language restrictions were set for searches.
The decision tree for inclusion or exclusion of articles is shown in Figure 1. We included studies if the proportion of etiologic agents in genital ulcers and nongenital skin ulcers, including H. ducreyi, was confirmed by laboratory techniques. Clinical diagnosis of chancroid is often based on the appearance of the ulcer, which is characteristically painful, purulent, and deep with ragged, undermined edges ( Figure 2). However, because the appearance of these ulcers is similar to ulcers caused by other bacteria, clinical diagnosis can be nonspecific or insensitive and often requires laboratory confirmation (1). In addition, microscopy identification of typical morphologic features and serologic detection lack sensitivity and specificity (13,14). Thus, we only considered the following diagnostic methods as providing acceptable evidence of H. ducreyi infection: 1) isolation and identification by culture; or 2) PCR/real-time PCR.

Data Extraction and Synthesis
For all qualifying studies, extracted data included study country, year of study, diagnostic test used for confirmation, total number of H. ducreyi-positive cases, and sample size. Descriptive analyses of extracted data were conducted, and the number of H. ducreyi-confirmed cases was divided by the total number of cases to calculate the proportion of cases caused by H. ducreyi. Studies qualifying for data extraction were grouped into 2 categories: studies conducted before 2000 and studies after 2000. This date separates studies before and after widespread implementation of syndromic management of GUD. Study sites were also plotted by geographic region. No quantitative metaanalysis was undertaken.

Results
We identified 277 records in which we found 46 articles describing 49 studies on GUD that met our inclusion criteria (Tables 1, 2; online Technical Appendix, http:// wwwnc.cdc.gov/EID/article/22/1/15-0425-Techapp1. pdf). All identified studies were based on cohorts of patients attending STI clinics, including 3 studies that enrolled only commercial sex workers. The age group for all cases was adults >18 years of age, except for 3 studies in Zambia, South Africa, and China, which included patients >16 years of age, and 1 study in Madagascar, which included patients >14 years of age. A total of 9 published studies and 2 unpublished reports that described nongenital skin ulcers caused by H. ducreyi were also included in our systematic review.  Laboratory confirmation of chancroid by PCR or culture was reported in 33 (67%) and 16 (32%) of the 49 studies, respectively. Of 16 studies that used culture, 7 (43%) used Mueller-Hinton agar with a nutritional supplement (e.g., IsoVitalex; Becton Dickinson, Franklin Lakes, NJ, USA), 1% used hemoglobin, and 5 (31%) used chocolate agar-based media; the remaining studies used other culture media. Five (31%) of 16 studies incubated agar plates at low temperatures (33°C-35°C), and 2 (12%) incubated plates at 36°C. Remaining articles did not specify incubating temperature.
Different PCR primer targets were used to amplify DNA sequences, including the 16S rRNA gene, the gro-EL gene, and the hemolysin gene. In addition to herpes s implex virus (HSV) PCR, 23 studies used a multiplex PCR that could simultaneously detect the 3 major causes of GUD (H. ducreyi, Treponema pallidum, and HSV types 1 and 2) (15). Studies encompassed 33 countries: 17 in Africa, 4 in Southeast Asia, 3 in Europe, 2 in the Middle East, 3 in South America, and 2 in the Caribbean, 1 in the United States, and 1 in Australia.

Incidence of Chancroid
Of 49 studies on chancroid analyzed, 35 were published during 1980-1999 (Table 1)   During 1980-1999, the proportion of genital ulcers caused by H. ducreyi in these studies ranged from 0.0% in Thailand and China to 68.9% in South Africa (Table 1). Eleven (31.4%) studies reported high proportions (>40%) of cases of infection with H. ducreyi. All of these studies were conducted in countries in Africa (Côte d'Ivoire, Gambia, Kenya, Lesotho, Senegal, South Africa, and Swaziland). Slightly lower proportions (20%-40% of cases) were observed in 15 (42%) studies: 10 in countries in Africa, 2 in the United States during localized outbreaks, 1 in Jamaica, 1 in the Dominican Republic, and 1 in India.
Only a few countries reported low proportions (<10%) of genital ulcers infected with H. ducreyi, including Singapore (8.3%), Peru (5%), Greece (4.6%), the Netherlands (0.9%), United States (3.1%), and Saudi Arabia (2.1%). The study in Saudi Arabia was conducted during 1995-1999; a total of 27,490 patients were examined for STIs. Chancroid was diagnosed by culture and was reported as the least common STI during this survey. The only studies that reported no cases of chancroid were conducted in Thailand in 1996 and China in 1999; both studies used multiplex PCR for detection of GUD cases.
During 2000-2014, the proportion of H. ducreyi infections was low (<10%) in all studies analyzed, except for 1 study in Malawi (15%) ( Table 2). Studies in 5 countries (Kenya, Namibia, Zambia, Brazil, and Australia) did not report any cases of infection with H. ducreyi. Other studies reporting proportions of infections <10% were conducted in Botswana, Mozambique, South Africa, Uganda, Pakistan, and France. No reports were found for studies in North America, Southeast Asia, or the Caribbean.

Nongenital Skin Infections with H. ducreyi
During 1988-2010, several case reports described 4 children and 4 adults with nonsexually transmitted infections with H. ducreyi that manifested as lower leg lesions but no genital lesions. The reported case-patients were travelers who had been to Fiji (7), Samoa (8), Vanuatu (9), or Papua   New Guinea (10) ( Table 3). Outside the south Pacific region, a 5-year-old refugee from Sudan who had lower leg ulceration was also given a diagnosis of infection with H. ducreyi (11). A cohort study conducted in Papua New Guinea in 2014 showed evidence that H. ducreyi is a major cause of chronic skin ulceration; H. ducreyi DNA was identified by PCR in 60.0% of skin lesions in children (3). Similar studies in other areas reported laboratory-confirmed skin ulcers in children caused by H. ducreyi in Papua New Guinea (6), Solomon Islands (4), Vanuatu (C.Y. Chen, pers. comm.), and Ghana (5) ( Table 3).

Discussion
Our review confirmed 2 major findings. First, reduction in the proportion of genital ulcers caused by H. ducreyi has been sustained for the past decade and a half. Second, there is increasing evidence that H. ducreyi is a common and newly recognized causative agent of chronic skin ulceration in children from developing countries.
In the 1990s, the global prevalence of chancroid was estimated to be 7 million (17). Chancroid was one of the most prevalent GUDs, particularly in resource-poor countries in Africa, Asia, Latin America, and the Caribbean (1; reference 51 in online Technical Appendix). Recommendations to introduce syndromic management for treatment of GUD caused by bacteria were published by the World Health Organization in 1991 and fully implemented by 2000 (reference 61 in online Technical Appendix). Since that time, global incidence of GUDs, particularly chancroid, has decreased substantially, and genital herpes viruses (HSV-1 and HSV-2) have become the predominant cause of GUD (reference 53 in online Technical Appendix). Currently in Europe and the United States, chancroid is restricted to rare sporadic cases. Transmission of H. ducreyi remains ongoing in only a few countries that have limited access to health services (2,6). showed that a high proportion of laboratory-confirmed skin ulcers were caused by H. ducreyi. Nearly half of the 690 enrolled patients with ulcers in these 6 studies had H. ducreyi detectable by PCR, whereas other bacteria, such as T. pallidum subsp. pertenue, the causative agent of yaws, were detected in 25% of patients.
These cases of infection with H. ducreyi confirmed by molecular analysis suggest that clinicians should be more aware of this newly recognized bacterium in skin ulcers of persons in tropical areas. In the context of new efforts to eradicate yaws, mass treatment with azithromycin in 6 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 22, No. 1, January 2016 Papua New Guinea reduced the absolute prevalence of ulcers not caused by yaws, which were mainly caused by H. ducreyi, from 2.7% to 0.6% (prevalence ratio 0.23, 95% CI 0.18-0.29) at 12 months after treatment (6). However, persistence of H. ducreyi at low levels after mass treatment in Papua New Guinea (3) and Ghana (5) suggest that 1 round of mass treatment might not be successful in eradicating H. ducreyi skin ulcers. Our review has several limitations. First, the increase in HSV-related GUD as a result of immunosuppression by HIV infection would result in a decrease in the proportion of chancroid among all GUD case-patients. Second, the lack of sequential studies performed in similar clinical settings at multiple time points precludes an optimal interpretation of the apparent decrease. Third, results might be affected by poor-quality data from many developing countries and might be inflated by publication bias. Fourth, PCR is more sensitive than culture. Therefore, increasing diagnostic yield might have partially masked the scale of the decrease in H. ducreyi as a cause of GUD.
In summary, we observed a quantitative and sustained reduction in cases of chancroid as a result of antimicrobial drug syndromic management and major social changes. In addition, data from several research groups indicate that H. ducreyi can cause nongenital skin lesions in persons residing in different regions. Further studies of this newly described pathogen skin disease association are required, and appropriate policies are needed that include the routine practice of managing tropical skin ulcers. Ms. González-Beiras is a predoctoral fellow at Instituto de Higiene e Medicina Tropical, Lisbon, Portugal. Her primary research interests are strategies for elimination of neglected tropical diseases.