| dc.description.abstract | Mycotoxin poisoning in maize has been a frequent occurrence in Kenya with the most
significant outbreak being that of 2004 when 317 patients and 125 deaths were recorded.
Mycotoxin contaminated grains cannot be obviously identified using naked eyes. This project
sought to establish the spread of fumonisins and aflatoxin producing fungi and their
associated mycotoxins in Machakos, Meru and Kitui counties of Kenya. The project further
tested the efficiency of visible and near infra red light maize sorting machine in
decontaminating maize from market samples of Machakos, Meru and Kitui.
Maize samples were collected from 30 markets in diverse agro-ecological zones of Meru,
Machakos and Kitui counties during the 2013 harvest crop. Fusarium and Aspergillus were
isolated from maize samples and mycotoxin producing Fusarium and Aspergillus species
identified based on cultural and morphological characteristics. Fumonisins and aflatoxin were
extracted from the samples and the mycotoxins analysed by enzyme-linked immunosorbent
assay (ELISA). Based on the results of mycotoxins analysis the maize samples were
categorised into high fumonisins, high aflatoxin, medium fumonisins, medium aflatoxin and
low mycotoxins. Nine samples from each category were sorted using near infra-red single
kernel sorting machine into accept and reject fractions. The sorted fractions were subjected to
microbiological and ELISA analysis to determine the population of Fusarium and Aspergillus
species and their respective mycotoxin concentrations. The percentage accept / reject kernels
was correlated to the population of the mycotoxigenic fungal species and level of respective
mycotoxins.
Aspergillus flavus S and L strains and A. parasiticus were found to contaminate the maize
samples in high levels while F. verticillioides, F. proliferatum and F. oxysporum were the
fumonisins-producing fungi isolated. Other Aspergillus species isolated were A. niger, A.
ochraceous, A. candudus and A. tamarii. The mean aflatoxin levels detected in Meru, Kitui
and Machakos were beyond the acceptable limits of 10ng/kg. Kitui had the highest levels of
aflatoxin detected while Machakos had the least. Fumonisins was detected in levels above the
acceptable limits in Meru and though detected in Kitui and Machakos the levels were within
the acceptable limits. The fumonisins levels did not match the population of F. verticillioides
and this could be attributed to differences in environmental conditions as well as the
pathotype.
The near infra-red single kernel sorting machine was effective in removing aflatoxin and
fumonisin-contaminated kernels from the samples, with an accuracy of up to 97.8% for
aflatoxins and 60.8% for fumonisins. The accepted fractions had statistically lower aflatoxin
and fumonisin levels than rejected maize from the same bulk sample while rejecting only 015%
of the sample. Market maize samples with toxin positive samples had reject rates of 0 to
25% and toxin negative samples having reject rates of 0-1%. These rejection rate data suggest
the near infra-red single kernel sorting machine rejects kernels in a dose-dependent manner.
There was a positive relationship between percentage accepted grain fractions and fumonisins
and aflatoxin levels while total Fusarium species did not seem to have any relationship with
the percentage accepted grain.
The study established that there are high risks of exposure to mycotoxin poisoning from
consuming maize available in local market stores and that optical sorting technique is
effective in removing mycotoxin-contaminated grain. The near infra-red single kernel sorting
machine detected aflatoxin and fumonisin mycotoxins consistently but did not consistently
detect the fungal contamination. The optical sorting technology can be further improved into
a high throughput machine and offer potential for commercialization for use by local grain
stores and millers to reduce mycotoxin contamination in maize. | en_US |
