Epidemiological aspects of mycetoma from a retrospective study of 264 cases in West Bengal
Soma Bandyopadhyay 112/B Ananda Palit Road, Kolkata 7000 14, India. Fax: (91) 033 244 8950, E-mail: email@example.com
Between 1981 and 2000, 264 cases of mycetoma were diagnosed clinically and microbiologically at Calcutta School of Tropical Medicine. Retrospective analysis of the records revealed that the ratio of actinomycetomas and eumycetomas was 197 : 67; the male to female ratio was 183 : 81. Ninety-four cases occurred in the 1980s and 170 in 1990s, with significantly more infections of Actinomadura spp. (P < 0.01) and fewer with Nocardia caviae (P < 0.01) during the last decade. Pricking was the most common injury associated with eumycetomas (P < 0.01). A total of 196 infections were in exposed body parts and 68 in covered areas. The localization of mycetomas differed significantly (P < 0.01) according to sex, incidence of actinomycetomas or eumycetomas, and obvious history of trauma. Exposed area cases were more common among agricultural workers (P < 0.01), while covered area mycetomas were almost always actinomycetomas with a remarkably lower incidence of N. caviae, A. madurae and Madurella grisea infections. The peak age of onset was between 16 and 25 years. The delay of diagnosis for the 80th percentile of cases was around 6 years for cases caused by N. brasiliensis and Streptomyces spp.; 8 years for N. caviae and N. asteroides; and 10 years for M. grisea and Actinomadura spp. From the history of trauma in 130 patients, the 80th percentile incubation period (IP) was calculated for N. brasiliensis, N. caviae and N. asteroides as 3 years; for Actinomadura spp. 7 years and for M. grisea 9 years. The minimum IP for all organisms was around 3 months.
Mycetoma is endemic around the Tropic of Cancer, including India, Sudan, Senegal, Somalia, Mexico, Venezuela and other countries (Mahgoub & Murray 1973). Although the disease was first reported from India in the middle of the nineteenth century, reliable epidemiological information remains scarce because of the rarity of the disease, its slow and unremarkable clinical course, multifarious causal agents, prolonged incubation period (IP) and frequently, the lack of a reliable history. Some information is available from studies in different parts of India (Pankajalakshmi & Taralakshmi 1984; Joshi et al. 1987; Chakrabarti & Singh 1998).
This subcutaneous mycosis essentially requires a predisposing trauma to create a point of entry for the causal saprophytic micro-organisms. The interval between the incidence of trauma and the appearance of overt disease can thus be taken as the IP for the isolate, provided the patients can recall the exact incidence of trauma, however minor it may have been. As the disease usually occurs in injury-prone parts of body months or years after the relevant trauma, the patient's recall may not be reliable. Similarly, the exact onset of the disease may remain indeterminate, as patients often do not present until noticeable discomfort, swelling or discharge appear. Therefore, it is difficult to gather reliable data on the average delay of diagnosis of individual causal agents in any single centre. Other questions, such as whether incidence of infection of a particular causal agent is increasing or decreasing with time and changing geo-ecological conditions, or whether the infection rate of causal agents varies with area of localization or nature of trauma, can only be answered by proper analysis of a large data set. With these objectives in mind, we conducted a retrospective study.
Materials and methods
The study included 264 mycetomas which were clinically and microbiologically confirmed in the Medical Mycology Department, Calcutta School of Tropical Medicine, between 1981 and 2000. The records included general information such as date of diagnosis, age, sex, occupation, residence, initial site of lesion and duration of the disease. Any relevant history of trauma of the affected part in the recent or distant past was especially enquired about and the time interval of stated trauma and appearance of first symptom were also recorded, if available.
If clinically suspected, the diagnosis of all cases was established by microscopic demonstration of grain morphology and identification of isolates from expressed discharge collected by opening the impending sinuses. As much clinical material was available it was processed in multiple sets in modified neutral Sabouraud's Dextrose Agar medium (Kwon-Chung & Bennett 1992) with or without cycloheximide and/or chloramphenicol and incubated at 37 °C for 1–6 weeks. Where grains were detected by direct microscopic examination of a KOH smear but culture showed no growth, attempts were repeated withholding specific therapy. In some cases we found grains by histopathological rather than direct microscopic examination. In most cases the size, shape, colour and consistency of grains permitted identification of causal agents. Finally isolates were characterized by studying colony morphology, fungal morphology and their conidiogenesis. Actinomycetes were identified by decomposition of casein, xanthine, tyrosine and urea (Mishra et al. 1980).
We broadly classified the initial sites of lesions into two categories, i.e. `exposed' and `covered' areas of the body. The covered area included chest, back, abdomen, shoulder and neck, buttock and perineum and the upper third of thighs and arms. Exposed areas included the lower two-thirds of both extremities, scalp and face. Any difference of epidemiological parameters in these two body areas were tested for statistical significance by chi-square test.
Injuries were classified in five categories, of which pointed injuries such as pricking had special importance in the epidemiology of mycetoma (Basset et al. 1965). Others included wide open injuries (cuts or operative injuries), shallow surface injuries (friction, scratch or abrasion) and injuries with intact surface, e.g. impact of hard objects. Inapparent injuries were those which went unnoticed, or were too insignificant to recall.
According to nature of occupation, cases were divided into agricultural workers, other manual and sedentary workers. Agricultural workers had special epidemiological importance (Abbott 1956). Statistical significance for values of different subgroups were determined by chi-square test.
Based on the available case history, the IP of the 80th percentile was taken as the highest range of IP of any mycetoma agent. Values higher than that seemed to be less reliable because of the wide dispersion of data. The delay of diagnosis of a particular mycetoma agent was expressed as the duration of illness before diagnosis of their 50th and 80th percentile cases. The age of onset of the disease was calculated after deduction of duration of illness from the age on reporting. Most of the epidemiological reports ignored this error factor.
Ninety-four of the 264 mycetomas were detected between 1981 and 1990, the remainder between 1991 and 2000 (Table 1). The initial sites of involvement were on exposed body parts in 196 cases and in covered areas in 68 cases. A total of 197 cases were actinomycetomas and 67 were eumycetomas. The distribution of causal agents was as follows: Nocardia brasiliensis (n=75), N. caviae (now called N. otitidiscaviarum) (n=38), N. asteroides (n=52), Streptomyces spp. (n=9), Actinomadura spp. (n=23), Madurella grisea (n=35), M. mycetomatis (n=8), Pseudallescheria boydii (n=5), Acremonium spp. (n=6), Pyrenochaeta romeroi (n=6), Exophiala jeanselmei (n=1) and unidentified eumycetoma agents (n=7). Among actinomycetomas, the incidence of N. caviae infections decreased significantly (P < 0.01) and that of Actinomadura infections increased significantly (P < 0.01) during the 1990s. The incidences of N. caviae, Actinomadura spp. and M. grisea infections in exposed areas were at least nine times as frequent as covered area cases. Localization of the cases was as follows: exposed parts of lower extremities (n=186), exposed parts of upper extremities (n=8), scalp (n=2), abdomen (n=3), chest (n=11), back (n=14), neck and shoulder (n=10), buttocks (n=5), upper thigh (n=12) and upper arm (n=13).
Distribution of mycetoma agents over covered and exposed areas of the body in two decades
There were 183 (69.3%) male and 81 (30.7%) female patients (Table 2). The earliest age of onset was 6; the latest, 77. At onset, most patients (114) were aged 16–25, with a median of 25 years. Of the exposed area cases, 57 were agricultural workers, 36 other manual labourers and 103 sedentary workers. Five agricultural labourers had covered area mycetoma, as did 15 other manual and 48 sedentary workers. Only 130 patients could provide a probable history of trauma preceding the disease. For actinomycetomas the incidence of pricking was 20 cases; for wide open injuries, 27; for shallow surface injuries, 17; for blunt injuries, 19 and for inapparent injuries, 114. Eighteen eumycetomas were caused by prickings, 12 wide open injuries, 8 shallow surface injuries, 9 blunt injuries and 20 inapparent trauma. The incidence of eumycetomas associated with pricking was significantly higher (P < 0.01) than with other causes. Pricking injuries comprised of puncture with a thorn, wooden spicule or plant fibre, nail or needle. Of 39 wide-open injuries, 19 were operative wounds including nine abscess drainages, seven surgical excisions of a swelling, cyst or corn and three cases of skin grafting.
Incidence of mycetoma in relation to age of onset and sex
The risk of males to contract mycetomas at exposed sites was almost three times higher than that of females; covered area infections were almost equally distributed between the sexes (Table 3). The difference (P < 0.01) might be the result of their occupational risks. Exposed area cases were more prone (P < 0.01) to mycetomas. In exposed areas the incidence of actinomycetomas was almost double the incidence of eumycetomas, whereas almost all (65 of 68) covered area cases were actinomycetomas except three cases caused by M. mycetomatis which occurred after wide open injuries that could be considered as transiently exposed parts. The incidence of mycetomas in covered areas was about 11 times less than that of exposed area cases among farm workers, and about half for other occupational groups (P < 0.01). This might be the result of high occupational risk of trauma and soil contact in agricultural workers.
Risk factor-related differences of mycetoma incidences in exposed and covered parts of the body
The delay of diagnosis (Table 4) for 20th percentile cases of Nocardia and Streptomyces spp. was up to 1 and 2 years for Actinomadura spp. and M. grisea infections. The 50th and 80th percentile cases of N. brasiliensis infections were diagnosed after 2–6 years, N. caviae after 4–8 years, N. asteroides after 3–8 years, Streptomyces spp. after 3–6 years, Actinomadura spp. and M. grisea infections after 5–10 years. The average IP or 50th percentile of reported IP (Table 5) for N. brasiliensis, N. asteroides and Actinomadura spp. was 1 years and for N. caviae, Streptomyces spp. and M. grisea, 2 years. The shortest IP for all mycetoma agents was about 3 months; the longest, if accepted as the 80th percentile for practical purposes, was 3 year for all three Nocardia spp., 5 years for Actinomadura spp. and 9 years for M. grisea.
Delay of diagnosis of common mycetoma agents
Reported incubation periods of common mycetoma agents, in years
We retrospectively compared the overall prevalence of mycetoma and the prevalence of infective agents in West Bengal and surrounding areas of eastern India between the 1980s and 1990s. Overall, the prevalence of the disease increased, nearly six-fold in the case of Actinomadura infections, although the prevalence of N. caviae infections decreased by a factor of three. This may be because of the rapid geo-ecological changes in eastern India during the last 2–3 decades after the green revolution, when agricultural practices changed, resulting in more intense use of groundwater. A similar retrospective study in an area of low mycetoma prevalence in the hills of north-west India (Chakrabarti & Singh 1998) found no marked change in the prevalence.
Farmers are at greater occupational risk of injuries that permit infection with saprophytic agents. But trauma is not the only factor for the outcome of the disease; viability of the organisms and local wound conditions can also modify the outcome (Mackinnon 1962). As the nature of trauma in covered areas of the body is usually minor, only those organisms which can colonize short-term shallow wounds are likely to infect these areas. The chance of soil contact with these injuries is also smaller. The conditions related to the precipitating trauma, e.g. associated bacterial flora (Gonzalez & Gonzalez-Mendoza 1960) and the temperature of the anatomical region are of great importance for the disease outcome. As these factors are expected, to some extent, to be different for covered and exposed areas, the prevalence of aetiological agents of mycetoma would differ correspondingly (Maiti & Haldar 1998). Such risk factor-related differences of mycetoma incidences are statistically well supported in our study.
The term `incubation period' for each of mycetoma agent is difficult to apply because the exact time of infection can only be arbitrarily recorded from the available history of a small number of cases. The past history of relevant minor injuries more than 5 years back should be considered with circumspection. Hence, for all practical purposes, the IP of the 80th percentile is taken as the highest IP range of any causal agent. The stated IP for the remaining cases are dispersed over a wide range of time up to 10 years and not taken into account. The shorter IP indicated from experimental animal models (Gonzalez & Sandoval 1960) probably represents the time required for a part of events occurring from time of entry of organism to appearance of symptoms in human diseases. In case of human mycetomas, usually scanty spores of pathogens can enter by traumatic implantations and remain quiescent for several months or years before forming a clumpy mass after tissue adaptation. But in an experimental set up, clumpy inoculum is placed in the target sites. Probably for that reason the disease evolves more quickly.
In a review of 333 cases, the variation of the duration of illness before confirmation of the disease was 3 months to 20 years, although most cases were detected within 2–3 years (Mahgoub & Murray 1973). This depends upon patients' awareness and availability of diagnostic facilities. In our study most cases were diagnosed within 3–4 years, indicating the need for better diagnostic facilities for mycotic diseases in tropical developing countries like India.
The authors are indebted to the Director, Calcutta School of Tropical Medicine, for his kind permission to use the institutional records.