During August 2010–December 2012, we recruited all patients who had clinically suspected M. ulcerans infection and had attended a clinic in the Buruli ulcer–endemic Asante Akim North District in Ghana. Age- and sex-matched household contacts of patients were also asked to participate; all study participants were >5 years of age. The study protocol was approved by the ethics review committee of the School of Medical Sciences, Kwame Nkrumah University of Science and Technology (CHRPE/91/10).

Figure 1

Figure 1. Mansonella perstans nematode in peripheral blood mononuclear cells from Buruli ulcer patient in GhanaCells were stained with Giemsa (original magnification ×1,000)Mperstans nematodes can be distinguished from Loa loa and Wuchereria bancrofti...

Whole blood samples were taken at baseline, at week 6, and at week 12 from 66 patients in whom the diagnosis of Buruli ulcer disease had been confirmed by PCR for the IS2404 repeat sequence specific for M. ulcerans (8); samples were also obtained from 20 household contacts at the same intervals. The samples were heparinized, and peripheral blood mononuclear cells were separated from 10-mL samples. Filarial infection was confirmed on a blood film stained with Giemsa and Delafield hematoxylin and examined for microfilariae at ×10 and ×40 magnification (the Knott technique; 10). M. perstans nematodes were distinguished from L. loa and W. bancrofti nematodes by their small size and the absence of a sheath (Figure 1).

Patients in whom M. ulcerans infection was found were treated with 10 mg/kg oral rifampin and 15 mg/kg intramuscular streptomycin, administered daily at village health posts under direct observation for 8 weeks (RS8 treatment). The patients were followed up every 2 weeks in the clinic and monitored for complete healing or recurrence of skin lesions. We compared the proportion of household controls versus the proportion of Buruli ulcer patients infected with M. perstans nematodes and the time to complete healing of M. ulcerans lesions in co-infected versus monoinfected patients. Categorical variables such as sex, clinical form of M. ulcerans lesion, and category of M. ulcerans lesion were compared by using the Fisher exact test, and cumulative healing was compared by using the log-rank test.

We found all forms of M. ulcerans disease among the group of patients; proportions of each type and category are shown in the Table. Of 66 patients with M. ulcerans disease, 15 (22.7%) were co-infected with M. perstans nematodes, whereas 4 (13%) of 30 household controls had M. perstans infection (p = 0.4 by Fisher exact test). Three patients in the co-infected group and none in the M. ulcerans–monoinfected group reported pruritus. No other clinical signs of M. perstans infection were found.

Figure 2

Figure 2. Survival analysis curve of cumulative healing for patients with Mycobacterium ulcerans infection who were co-infected with Mansonella perstans nematodes compared with those who had Mulcerans monoinfection, Ghana, August 2010–December 2012No difference...

All 66 patients completed RS8 treatment, but 9 were lost to follow-up during the 12-month follow-up period. Buruli ulcer lesions healed completely in 14 co-infected patients by 58 weeks (median 20 weeks, 95% CI 14.6–30.2) and in 43 monoinfected patients by 50 weeks (median 21 weeks, 95% CI 16.7–25.5). We found no difference in cumulative time to healing for co-infected versus monoinfected patients (p>0.05 by log-rank test) (Figure 2). Buruli ulcer patients who had M. perstans nematodes co-infection were treated with doxycycline (200 mg) and ivermectin (150 μg/kg) daily for 6 weeks, starting during the second to fourth week of RS8 treatment. Viable microfilariae were still visible in peripheral blood mononuclear cell cultures from all co-infected patients after ivermectin and doxycycline treatment, but pruritus subsided in the 3 patients who had reported it.

We found co-infection with M. perstans in 23% of Buruli ulcer patients in a disease-endemic district in Ghana, but this prevalence was not significantly difference from prevalence among household contacts who served as controls (13%). As with Buruli ulcer, M. perstans filariasis is predominantly found in rural populations and infection begins in childhood; the highest infection rates are found in children 10–14 years of age (11), similar to those for children at highest risk for M. ulcerans infection. M. perstans infection occurs in Ghana and was seen in the Volta region of Ghana around Hohoe during the 1990s, but its prevalence is unknown (12), and no information is available about the average number of worms per infection. In Uganda, prevalence of M. perstans infection has been found to range from 0.4% to 50% (13).

M. perstans nematodes are transmitted by the bites of Culicoides midges, but it is not known whether M. perstans–infected midges can be co-infected with M. ulcerans. In a guinea pig model, skin penetration was shown to be a requirement for establishment of M. ulcerans disease (14), and it has been postulated that mosquito bites cause M. ulcerans disease in Australia (6). These organisms might share a common route of transmission, but our findings in this small study do not support this concept.

Our findings suggest that M. perstans nematodes are common in rural Ghana and coincidentally infect patients with M. ulcerans disease, necessitating the consideration of these organisms in the management plan of Buruli ulcer patients. Although often asymptomatic, M. perstans infection may cause eosinophilia, subcutaneous swellings, aches, pains, and skin rashes in a considerable proportion of patients (9). Because filarial nematodes are known to polarize the host immune responses from T-helper type 1 cells needed for protection against mycobacterial infections, toward humoral and T helper type-2 mediated immunity, we plan to undertake a study to investigate this interaction.