¹ Department of Civil Engineering, Faculty of Technology, Makerere University ² Department of Veterinary Public Health and Preventive Medicine, Faculty of Veterinary Medicine, Makerere University

Correspondence to: Dr Francis Ejobi, Department of Veterinary Public Health and Preventive Medicine, Faculty of Veterinary Medicine, Makerere University, P.O. Box 7062, Kampala Uganda, Tel: 256 77 492236, E-mail: ejobi@vetmed.mak.ac.ug

Code Number: hs05004

ABSTRACT

Background: In the sub-urban areas of Kampala city, springs are a major source of water for domestic use. Though spring water is considered to be aesthetically acceptable for domestic use, presence of poorly designed pit latrines, poor solid waste management as well as poor and inadequate spring protection, may lead to contamination of spring water with pathogenic bacteria.
Objectives: The objectives of the study were to examine the bacteriological quality of water from ten springs in Katwe and Kisenyi parishes of Kampala, and to identify and quantify risks for spring water contamination with faecal bacteria.
Methods: A cross-sectional sanitary risk assessment using a standardised format was carried out in ten randomly selected springs in the parishes of Katwe and Kisenyi parishes in Kampala. A total of 80 samples of water from these springs were collected from December 2001 to March 2002. The samples were analysed for indicators of faecal contamination: total coliforms, faecal coliforms and faecal streptococci. Physico-chemical parameters were measured.
Results: Aggregate qualitative sanitary risk scores ranged from medium to high. The total coliform counts in 90% of the samples exceeded the WHO guideline for drinking water. All the samples had faecal coliform counts above the WHO guideline. A strong correlation (r2= 887) was observed between the median faecal coliform counts and the sanitary risk score. Sixty percent of the samples had nitrate levels above the WHO recommended limit. There was no correlation between the levels of chlorides and nitrates and levels of indicators of faecal bacterial contamination. Conclusions: The sanitary risk assessment score is a reliable tool for predicting the likely levels of bacterial contamination of spring water. Water from the ten protected springs studied is unsuitable for drinking without treatment.
African Health Sciences 2005; 5(1): 14-20

INTRODUCTION

In Uganda, like other developing countries, sanitation and water supply are often inadequate. As a result, many low-income communities in these countries rely on untreated groundwater for drinking and other domestic use¹. In Kampala City, only 55% of the population is supplied with piped water, out of which only 9% is connected to the sewerage system².

In the sub-urban areas of Kampala City, springs are a major source of water for domestic use. This is because of the availability of springs in these areas, and also due to the lack of piped water.Though spring water is considered to be aesthetically acceptable for domestic use, presence of poorly designed pit latrines, poor wastewater management, poor solid waste management as well as poor and inadequate spring protection, may lead to contamination of water from the springs with pathogenic bacteria. In addition, drinking water may be contaminated with nitrates. The health effects of presence of nitrates in water are linked to methemohemoglobinemia³. Children between the ages of 12 to 14 years drinking water containing greater than 105 mg/l of nitrate have been reported to exhibit delayed reactions to light and sound stimuli³.

Previous studies conducted in and around Kampala City examined largely the bacterial load of spring water without attempting to identify and quantify the potential risk factors contributing to the contamination. In the case of the parishes of Katwe and Kisenyi, the risk factors for spring water contamination were largely unknown. The objectives of the present study therefore were, firstly, to examine the bacteriological quality of water from selected springs in these parishes, and secondly, to identify and quantify risks for spring water contamination with bacteria indicator organisms.

On-site measurement of water temperature, pH and electrical conductivity

The water temperature and pH were measured using aWTWÒ microprocessor pH/temperature meter. The meter was calibrated with pH 4 and 7 using standard buffer solutions according to manufacturer’s instructions (WTW, Vienna, Austria). The electrode was rinsed with distilled water between samples. Electrical conductivity wasmeasured using a WTWÒ microprocessor conductivity meter calibrated at 25°C.

Laboratory analyses

The laboratory analyses were carried out in the Public Health and Environmental Engineering Laboratory of the Department of Civil Engineering, Faculty of Technology, Makerere University. The bacteriological parameters analysed were total coliforms, faecal coliforms and faecal streptococci, while the physico-chemical parameters measured were turbidity, nitrate and chloride levels.

Enumeration of total coliforms

Total coliforms were isolated and enumerated using the membrane filtration technique and growth on membrane lauryl sulphate broth (MLSB). The sample was filtered and the membrane was then placed on an absorbent pad soaked in MLSB. The plates were then incubated for one hour at room temperature, to aid resuscitation, and incubated for a further 23 hrs at 37°C. All yellow colonies extending on the membrane were counted with the aid of a magnifying lens and recorded as presumptive total coliforms. For confirmation, a representative number of characteristic yellow colonies were sub-cultured into tubes of Brilliant Green Bile (BGB) broth and incubated at 37°C for 48 hours. Gas production in BGB broth confirmed the presence of coliforms. The results were then expressed as the number of colonies in 100 ml of original sample as described by APHA⁵.

Enumeration of faecal coliforms

Faecal coliforms were also enumerated using the membrane filtration technique. The plates were incubated at 44°C. Sub-cultures for confirmation were incubated at 44°C for 24 hours. Yellow colonies of various sizes extending on the membrane were counted with the aid of a magnifying lens and taken to be thermotolerant coliforms (presumptive). Gas production in BGB broth confirmed the presence of thermotolerant coliforms.

Enumeration of faecal streptococci

Faecal streptococci were isolated and enumerated by the membrane filtration technique. The filters were placed on membrane Enterococcus agar (Slanetz and Bartley agar, Oxoid Ltd), and plates were then inverted. The agar plates were pre-incubated at 37°C for 4 hours to aid bacterial resuscitation. The plates were then incubated at 44°C for a further 44 hours. After incubation all red, maroon and pink colonies that were smooth and convex were counted and recorded as faecal streptococci. The number of colonies in 100 ml of original sample was then computed.

Measurement of turbidity, nitrate and chloride levels

Turbidity was measured using the HACH DR/2010Ò Spectrophotometer. Twenty five millilitres of a wellmixed sample were measured into a clean sample cell. Another sample cell was filled with distilled water. The intensity of light scattered and absorbed by the sample was compared to that measured for standard Formazin suspensions and was read at a wavelength of 860 nm.

Nitrate levels were measured using the cadmium reduction method. Twenty-five millilitres of sample were transferred into a sample cell, and another sample cell was filled with an equal amount of distilled water. The contents of one Nitraver 5 Reagent Powder Pillow were added to each sample cell. The sample cells were stoppered and vigorously shaken for one minute. They were then left to stand for five minutes to allow development of the colour. The concentration in mg/l was measured against the blank (distilled water) at a wavelength of 500 nmusing the HACH DR/2010Ò spectrophotometer.

Chloride levels were measured using the mercuric thiocyanate method. The sample cell was filled with 25 ml of sample and another with equal amount distilled water. Two millilitres of mercuric thiocyanate solution were added to each cell and swirled. One millilitre of ferric ion solution was pipetted into each sample cell and again swirled. After 2 minutes, chloride concentration of the sample in mg/l was measured against the blank andread at wavelength 455 nm using HACH DR/2010Ò spectrophotometer.

Table 2 presents the levels of the physico-chemical parameters measured. The nitrate levels ranged from 21 to 145 mg/l. The highest mean nitrate levels of 140 mg/l was recorded at Kasule spring. Sixty percent (52 out of 80) of the samples had nitrate levels above the WHO maximum permissible level of 50 mg/l. The chloride levels ranged from 6 to 79 mg/l. The chloride levels in all samples were below the WHO maximum permissible level of 250 mg/l. The electrical conductivity ranged from 95 to 705 ìS/cm. All samples had turbidity levels within WHO permissible level of less than 5 FTU. The pH ranged from 4.4 to 6.7. Only 5% (4 out of 80) of the samples had a pH within the WHO recommended range of 6.5 to 8.5. There was little variation in temperature readings, with a range of 23.6 to 26.4°C.

Table 2: Physico-chemical quality of water from springs in Katwe and Kisenyi springs, Kampala City

Bacteriological levels

Table 3 presents the levels of bacterial indicator organisms in water from the springs studied. All the springs were contaminated with faecal indicator bacteria. There was a wide variation in the microbial levels recorded in the springs. The total coliform counts ranged from 3 to 80,000 cfu/100 ml. The total coliform counts in 90% of the samples exceeded the WHO guideline for drinking water. The faecal coliform counts ranged from 1 to 60,080 cfu/100 ml. All the samples had faecal coliform counts above the WHO guideline for drinking water. The faecal streptococci levels ranged from 0 to 3,000 cfu/100 ml. Kakajo spring had the highest median levels for all bacterial indicator organisms of faecal contamination considered in the study. On the other, Kasule spring had the least median contamination levels. There was no significant difference (p=0.503) between the median total coliform counts of spring water from Katwe and Kisenyi parishes.

Table 3: Levels of indicators of faecal bacterial contamination of spring water in Katwe and Kisenyi parishes, Kampala City

Correlation of bacterial counts and the levels of chlorides and nitrates

The Pearson correlation coefficient of median total coliform counts with the nitrate levels was 0.293, and with the levels of chlorides was 0.193. The Pearson correlation coefficient of median faecal coliform counts with the levels of nitrates was 0.275, and with the levels of chlorides was 0.176. The Pearson correlation coefficient of median faecal streptococci counts with the levels of nitrates was 0.323, and with the levels of chlorides was 0.143.

Correlation of bacterial counts and the sanitary risk score

The Pearson correlation coefficient of median total coliform counts with the sanitary risk score was 0.887, of median faecal coliform with sanitary score was also 0.887, and of median faecal streptococci with sanitary score was 0.795.

DISCUSSION

Surveillance of water quality to ensure microbiological and chemical safety is a vital public health function especially in developing countries⁶. The WHO recommended guidelines for bacteriological quality of drinking water is less than 1 faecal coliform per 100 ml, and 10 total coliforms per 100 ml⁷. In this study, the bacteriological quality of water from protected springs in Katwe and Kisenyi parishes was poor. The median bacteria counts recorded in this study were quite comparable with earlier findings from other areas of Kampala City reported by Taylor and Howard⁸ and Barret et al⁹. The presence of the indicator bacteria suggests sewage contamination of these sources. Humans drinking such water untreated are at a risk of waterborne diseases.

In the present study, there was a weak correlation between nitrate levels and levels of bacterial indicator organisms of faecal contamination. In some springs, high nitrate levels were found even when low or no faecal coliform or faecal streptococci were found. Similar observations have been reported for groundwater sources in Iganga district in eastern Uganda⁹. This may be explained by the differences in hydro-geological regimes and likely contaminant entry point. Residence times in the shallow groundwater systems is likely to be short, and groundwater nitrate will be controlled by the degree of faecal loading and on mineralisation and nitrification of faecal nitrogen. The level of nitrates is controlled by rate of mineralisation of organic nitrogenous compounds and subsequent nitrification of ammonia, and possible localised denitrification⁹. The weak correlation observed could also be due to high dilution of faecal matter by rainwater and surface runoff or selective removal of nitrates and chlorides during transport in the soil¹⁴.

The low pH observed in some springs could have been as a consequence of carbon dioxide saturation in the groundwater¹⁷. Nshekanabo and Wozei¹⁸ found the pH of spring water in Katanga, a suburb North of Kampala city, to be acidic.

We conclude that sanitary risk assessment score is a reliable tool for predicting the likely levels of bacterial contamination of spring water, and that water from the ten protected springs studied is unsuitable for drinking without treatment.

ACKNOWLEDGEMENTS

We are grateful to Makerere University for the financial support. We thank all the local authorities and residents of Katwe and Kisenyi parishes for their co-operation.