Faecal contamination of drinking water sources of Dhaka city during the 2004 flood in Bangladesh and use of disinfectants for water treatment


M. Sirajul Islam, Environmental Microbiology Laboratory, Laboratory Sciences Division, ICDDR, B GPO Box-128, Dhaka-1000, Bangladesh. E-mail: sislam@icddrb.org


Aims:  To describe the extent of faecal pollution and point of use water treatment strategy during and after the 2004 flood in Dhaka.

Methods:  A total of 300 water samples were collected from 20 different drinking water sources in Kamalapur, Dhaka city from August 2004 to January 2005. The level of faecal contamination was estimated using measurements of faecal indicator bacteria (total coliforms, faecal coliforms and faecal streptococci) and isolation of Vibrio cholerae was carried out following standard procedures. Total dissolved solids, dissolved oxygen, hardness, chloride and pH were also monitored. The efficacy of four disinfectants including Halotab, Zeoline®-200, alum potash and bleaching powder were tested as point of use water treatment agents. The unacceptable level of contamination of total coliforms (TC), faecal coliforms (FC) and faecal streptococci (FS) ranged from 23·8% to 95·2%, 28·6% to 95·2% and 33·3% to 90·0%, respectively. The isolation rates of V. cholerae O1 and O139 were both 0·33%, and non-O1/non-O139 was 7·0%.

Conclusion:  Water collected during and after floods was contaminated with TC, FC, FS and V. cholerae. Although alum potash, bleaching powder, Halotab and Zeoline®-200 were all effective general disinfectants, Halotab and Zeoline®-200 were superior to bleaching powder and alum potash against FC.

Significance and Impact of the Study:  During and after floods, point of use water treatment could reduce waterborne diseases among flood-affected people.


Bangladesh is a flood-prone country where many people are affected by rising waters and displacement every year. The monsoon or rainy season occurs from June through October, causing minor annual flooding over an area of 26 000 km2 rising up to 52 000 km2 in more severe floods (MPO 1985). Surveys report that the floods of 1954, 1968, 1978 and 1984 affected areas ranging from 35 000 km2 to 52 000 km2 (Khalilur and Bhuiya 2000). The 1998 flood has been described as one of Bangladesh's worst floods in recent times. It covered nearly two-thirds of the country, affecting millions of people (Sabina 2000). Studies have reported an increase in diarrhoeal diseases in the postflood periods (Shears 1988). The flood of 1998 induced diarrhoeal epidemics before the water receded. This was associated with over 400 000 diarrhoea cases and 500 diarrhoea-related deaths (Kunii et al. 2002). In this most recent flood of 2004, 38 out of 64 districts were affected in Bangladesh, and a total of 201 762 cases and 87 diarrhoea-related deaths were recorded between 12 July and 22 August 2004 (Akram and Zamman 2004).

Studies have demonstrated that surface water sources in Bangladesh are heavily contaminated with faecal coliforms (FC) and by various pathogenic bacteria (Islam et al. 1991, 1992a,b, 1993, 1994a, 1995, 2000). Water-related diseases continue to be one of the major health problems globally as well as in Bangladesh. An estimated 4 billion cases of waterborne diarrhoea represented 5·7% of the global disease burden in 2000 (WHO 2002). Bacteriological contamination of surface water in Bangladesh and the high morbidity and mortality owing to waterborne diseases have led to the promotion of the use of hand-pumped and tap water for drinking. The majority (64%) of the urban population and nearly all (93%) of the rural population have access to hand-pumped or piped water (UNICEF 1995). Despite the availability and promotion of the use of such safe water sources, water-related diseases remain an important cause of mortality and morbidity in Bangladesh (Mitra and Associates 1992), and suggest that ingestion of contaminated water is an important mode of pathogen transmission.

Even if disinfection is practised in water supply systems, failure of the disinfection system could result in serious health hazards if contamination occurs. However, improving source water quality alone does not always decrease diarrhoeal disease incidence (Brisco 1978), because the diarrhoeal pathogens may be transmitted through food, vectors or through using contaminated water for various household purposes, e.g. washing utensils, raw-consumed vegetables and the like. In many developing countries, even municipal piped water is unsafe because of inadequately maintained pipes, low pressure, intermittent delivery, lack of chlorination and clandestine connections. For example, Vibrio cholerae was repeatedly isolated from nonchlorinated municipal water systems in Peru, associated with large cholera epidemics (Ries et al. 1992; Swerdlow et al. 1992). In Guayaquil, Ecuador, even central chlorination of the municipal water system was insufficient to maintain adequate free chlorine residuals at peripheral distribution sites, and drinking unboiled municipal water remained a primary cholera infection source (Weber et al. 1994).

An inexpensive treatment procedure at point of use is therefore necessary for people who do not have access to potable water. This study was designed to assess the quality of the drinking water in the flood-affected area of Kamalapur, in Dhaka city and to determine the efficacy of various points of use water disinfectants.

Materials and methods

Description of the study area

Kamalapur is located in the southeastern section of Dhaka near the northern side of the Buriganga River. This is an economically impoverished, densely populated area with unplanned housing and inadequate water and sanitation systems, where the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B) established a surveillance/intervention project in 1998 (Brooks et al. 2005). In the first year, the population was approximately 70 000, covering a 4-km2 area, but grew to approximately 200 000 in the same area by 2004. During the 2004 flood, field research assistants (FRA) from ICDDR,B's field office visited the community to determine how many families had been displaced to camps, the status of those who remained in their homes, the status of the water supply in the community and whether people had access to food and medicine.

The FRAs visited 20 000 households (40%) throughout the community and systematically gathered data. It was estimated that 6% of the population (12 000) had been displaced to camps and were in active need of food. Fifty per cent of the remainder (94 000) had standing water in and around their houses. Given the standing water, it was decided to examine the piped water status in Kamalapur. While all those who were not in the camps had access to food, water supply was of uncertain quality and 8% (16 000) of the population of Kamalapur were in active need of medical attention. A map of the Kamalapur study area and locations of sample collection are shown in Fig. 1.

Figure 1.

 Map of Kamalapur urban surveillance area showing location of sampling sites. Study site (inline image), stratum boundary (inline image), road (inline image), health care (inline image) and community centre/social club (inline image).

Sampling sites and sources

Samples were collected from 20 different defined locations to represent the total flood affected area of Kamalapur (Table 1). The site is divided into seven geographical strata based on roads, bodies of water, open fields and the like. Between two and three sampling sites were selected from each stratum (Fig. 1). A total of 300 water samples were collected on 16 sampling dates from 3 August 2004 to 2 January 2005. Water samples were collected from a variety of sources [e.g., the water and sewerage authority (WASA) taps, hand pumps, overhead tanks and lorries carrying water into the community] that were used as the main supply sources of drinking water for the flood-affected areas (Table 1).

Table 1.   Location and sources of the sampling sites
Site no.Area address/locationStratumType of source
  1. WASA, Water and Sewerage Authority, Dhaka.

1OutfallIWASA hand pump
2Outfall ansar campIWASA Gazi tank
3Outfall workshopITap water
4Telegue colonyIIWASA hand pump
510 city corpo. Staff colonyIIWASA hand pump
6Ashrafuddin house MandaIIITap water
751 RishiparaIIIWASA hand pump
834/3, ManiknagarIIITap water
9WASA colonyIVTap water
10WASA colonyIVTap water
11WASA Lorries waterIVWASA Lorries
1260/9/C/2, DhalpurVTap water
1365/1, BrahmanchironVWASA hand pump
1479/2, Gopibag BazarVIWASA hand pump
1595/C, R K Mission Rd.VIWASA hand pump
1689/2/A/3, R K Mission Rd.VIWASA hand pump
1716/2/A, N. JatrabariVIITap water
1882/C-1, North JatrabariVIIWASA hand pump
1915/1, North JatrabariVIITap water
20No. 8 DhalpurVWASA hand pump

Collection of water samples

Water from different sources was collected following WHO-recommended procedures (WHO 1984). For example, the water tap was first wiped, using a clean cloth. Then the tap was turned on at maximum flow rate and allowed to flow for 2 min. The interior of the tap was sterilized using alcohol and a gas burner. Then 500-ml water samples were aseptically collected in sterile Nalgene plastic bottles. The hand pump was treated similarly. Sampling from the tanks was performed by immersing the sterile bottle into the tank to a depth of about 20 cm with the mouth facing slightly upwards. Two litres of water were collected from each source for performing the chemical and bacteriological examinations. All samples were transported directly to the Environmental Microbiology Laboratory of ICDDR,B in an insulated box filled with cool packs (Johnny Plastic Ice; Pelton Sheperd, Stockton, CA, USA).

Estimation of faecal indicators

Each water sample was tested for total coliforms (TC), FC and faecal streptococci (FS) following procedures described elsewhere (APHA 1998; Islam et al. 2001). In brief, for TC, FC and FS, 100 ml of water samples were filtered through 0·22-μm pore-size membrane filter (Millipore Corp., Bedford, MA, USA), and the filters were placed on membrane faecal coliforms (mFC) and KF-streptococcus agar plates. The mFC plates were incubated at 37°C and 44°C for 18–24 h for TC and FC, respectively. Then the characteristic blue colonies were counted as TC and FC and expressed as colony-forming units (CFU) per 100 ml. The KF-streptococcus agar plates were incubated at 37°C for 48 h and the characteristic light and dark red colonies were counted as FS following standard procedures (APHA 1998).

Culture of Vibrio cholerae

A water sample of 50 ml was inoculated into 25 ml triple strength alkaline peptone water (APW) for enrichment, incubated for 6 h and then plated onto thiosulfate citrate bile salt sucrose (TCBS) agar and taurocholate tellurite gelatin agar (TTGA) plates (Monsur 1961). The inoculated plates were incubated at 37°C for 18–24 h. Vibrio-like colonies were streaked on gelatin agar (GA) and incubated at 37°C to determine the production of gelatinase and then inoculated into Kligler iron agar (KIA) and motility indole urea (MIU) agar media. Presumptive Vibrio isolates were checked for lysine and ornithine decarboxylase, arginine dihydrolase and for fermentation of glucose, mannitol, sucrose, mannose, arabinose, inositol and growth in 0%, 6·5% and 8% NaCl solutions. Serology was performed if indicated. Isolation and identification of V. cholerae were carried out following the procedures described by Islam et al. (1994b).

Measurement of chemical parameters

The total dissolved solids (TDS), dissolved oxygen (DO) and pH were measured using portable meters (HACH Conductivity Meter, Cat. No. 51800–18; HACH Portable Dissolved Oxygen Meter, Cat. No. 51850-18; SensionTM6, CO, USA and Orion Portable pH Meter, Cat. No. 210 A; Orion Research, MA, USA). The hardness and chloride were measured by titrimetric methods following standard procedures (APHA 1998).

Water treatment strategy

Water samples were treated with four different disinfectants, e.g. alum potash, Zeoline®-200 (commercially available sodium-hypochlorite solution; Zeolite India Pvt. Ltd., Kolkata, India), Halotab (Halazone USP, 15 mg; Sonear Laboratories Ltd., Dhaka, Bangladesh) and bleaching powder (calcium hypochlorite). Zeoline®-200 was used at four drops per litre for 5 min. Halotab was used as per the instruction of the manufacturer, i.e. one tablet per 3 l of water for 30 min. Bleaching powder was used at 0·25 spoon (1·25 g) per 20 l for 30 min. Alum, commonly known as alum potash [K2SO4 Al2 (SO4)3.24H2O], was used at 500 mg l−1 for 3 h. Only the heavily contaminated samples were treated with the afore-mentioned disinfectants and plated for TC and FS. Characteristic colonies were counted following the procedures described earlier.


Bacteriological contamination level of different water samples

Table 2 shows that the level of unacceptable drinking water sources ranging from 23·8% to 95·2%, 28·6% to 95·2% and 33·3% to 90·0% for TC, FC and FS, respectively according to the WHO guidelines (WHO 1984). The level of unacceptable supplies owing to TC and FC counts at the onset of the flood was 75% at round # 01 (3 August 2004) but during round # 2 (5 August 2004) at the height of the flood, it increased to 95·2%. The level of unacceptable sources gradually decreased from 95·2% to 83·3%, 65·0% and 60·0% in subsequent rounds (round # 3, 9 August 2004; round # 4, 16 August 2004 and round # 5, 22 August 2004). At their lowest, the levels reached 23·8%, 28·6% and 33·3% in round 13 (29 November 2004) based on TC, FC and FS, respectively.

Table 2.   Trend of unacceptability based on indicator bacterial level in 300 drinking water samples
Date of collectionNo. Of samplesRound no.Unacceptable (%)†
  1. *Not done.

  2. †According to World Health Organization guidelines, the unacceptable level of total coliforms and faecal coliforms is >0 per 100 ml.

03 Aug 200420175·075·0ND*
05 Aug 200421295·295·2ND*
09 Aug 20046383·383·383·3
16 Aug 200421465·065·065·0
22 Aug 200421560·060·060·0
01 Sep 200420657·947·368·4
15 Sep 200420765·065·070·0
23 Sep 200420855·050·090·0
13 Oct 200422963·663·659·1
20 Oct 2004221068·268·268·2
31 Oct 2004181166·750·077·8
11 Nov 2004211242·938·157·1
29 Nov 2004211323·828·633·3
12 Dec 2004181442·928·638·1
27 Dec 2004211547·647·647·6
02 Jan 200581650·050·062·5

Isolation of Vibrio cholerae

Table 3 shows that V. cholerae non-O1/non-O139 were isolated from 10 out of 16 rounds (62·5%). However, culturable V. cholerae O1 and O139 were isolated from rounds 2 and 5, respectively, as a mixed contamination with V. cholerae non-O1/non-O139. Overall, the isolation rate of vibrios was 7·6% of which V. cholerae O1, O139 and non-O1/non-O139 were 0·3%, 0·3% and 7·0%, respectively, of the 300 samples tested.

Table 3.   Isolation of Vibrio cholerae O1, O139 and non-O1/non-O139
Round no.No. of samples testedV. cholerae O1V. cholerae O139V. cholerae non-O1/ non-O139Total vibrios
 3 60033
16 80000
Total (%)300 (100%)1 (0·3%)1 (0·3%)21 (7·0%)23 (7·6%)

Chemical parameters

Table 4 shows that the maximum and minimum values of all the chemical parameters except pH were within the acceptable limit recommended by WHO (1984). The pH values at sites 2, 7, 9, 10, 11, 13, 14 and 16 were 6·47, 6·26, 6·28, 6·19, 6·41, 6·43, 6·39 and 6·41, respectively which were below the acceptable limit according to WHO guidelines (1984).

Table 4.   Range of chemical parameters of drinking water tested
SamplingTDS (mg l−1)DO (mg l−1)Hardness (mg l−1)Chloride (mg l−1)pH
  1. *Not available.

  2. TDS, total dissolved solids; DO, dissolved oxygen.

WHO guidelines (1984)<1000NA*NANA<500NA<250NA8·506·50

Efficacy of different treatment procedures

Of the four disinfectants tested, we found varying results with respect to different bacterial groups (Fig. 2). Zeoline®-200 was found to be the most effective (83·8%) against TC followed by Halotab (81·5%), alum (73·0%) and bleaching powder (64·7%). Halotab (77·1%), Zeoline®-200 (72·6%) and bleaching powder (72·4%) were more effective against FS than alum potash (29·7%). Out of 300 water samples, only the heavily contaminated samples were selected for testing the efficacy of various disinfectants. This includes 65, 68, 70 and 56 samples for Halotab, Zeoline®-200, alum and bleaching powder, respectively.

Figure 2.

 Treatment efficiency of different disinfectants. Total coliforms (bsl00000), faecal streptococci (bsl00001).


The unique finding of this study is the isolation of culturable epidemic strains of V. cholerae O1 and O139. This is the first time in Bangladesh that V. cholerae O1 and O139 have been isolated from a water reservoir which is actively used by the surrounding population. In addition to the isolation of V. cholerae O1 and O139, 7·6% of all the supplied water systems became contaminated with V. cholerae non-O1/non-O139. These findings clearly showed the extent of contamination of drinking water by potentially pathogenic bacteria during the 2004 flood, and the need to treat the water for potability during and after flood periods.

The results also demonstrated that during flood, the drinking water sources in the study area became heavily contaminated. However, the extent of contamination varies according to the stage of the flood. During the initial stages of the flood, the water became predominantly contaminated with faeces, presumably by flood water mixing with raw sewage. As the water started receding, the contamination level fell by 37·3% and 47·9% for TC and FC, respectively from round # 2 to round # 6. It has been observed that outbreaks of communicable diseases normally occur after floods owing to contamination of water and disruption of water purification and sewage-disposal systems (Siddique et al. 1989; Dietz et al. 1990; Aghababian and Teuscher 1992; Toole 1992). During the 2004 flood, a high incidence of diarrhoea owing to enterotoxigenic Escherichia coli (ETEC) and V. cholerae O1 was observed in flood-affected areas of Dhaka (Qadri et al. 2005). The disease pattern was quite similar to that of the 1988 (Siddique et al. 1991) and 1998 floods in Bangladesh (Kunii et al. 2002). Previous studies elsewhere also suggested that several epidemics have been caused by the contamination of drinking water (Patel and Isaacson 1989; Swerdlow et al. 1992; Tuttle et al. 1995).

The increase of diarrhoea during the present flood is likely attributed to the widespread heavy contamination of drinking water sources. These results indicate that the water reservoirs need to be cleaned during and after a flood situation. Furthermore, public health authorities should make the public aware of the potential danger of the public water supply, and encourage in-house treatment of the water before consumption. Specifically, the public should be informed that although the water smells and looks clean, it might contain infectious bacteria like V. cholerae O1 and O139 that can cause cholera or other diarrhoea.

The physico-chemical composition of the supplied water, e.g. total dissolved solids, dissolved oxygen, hardness and chloride were within the acceptable limits recommended by WHO (1984). The pH was the only chemical parameter which was lower than the acceptable limit recommended by WHO (1984). This study demonstrates that during flood, it is not the chemical composition of drinking water that is a concern, but its bacteriological content that poses a public health threat.

Finally, multiple disinfectants were tested on the contaminated water to identify the simplest, easiest and least expensive techniques of point of use water treatment. Four disinfectants were tested. Among the four agents, though Halotab and Zeoline®-200 were more effective than bleaching powder and alum potash, perhaps the most appropriate based on cost, ease of use, familiarity amongst local people and availability is the alum potash. People, especially in rural areas, use it as an antiseptic, as an after-shave, and as a water-purifying agent. Our data indicate that water treatment with alum potash may be sufficient to render the water potable. All the results presented in this study are based on treated water supplied by the municipality but surface water was not tested. Therefore, further study is required in which not only municipality-supplied water but the surface water collected from various open water sources, e.g. ponds, lakes, rivers and canals need to be tested and the efficacy, acceptability and sustainability of the point of use water treatment agents need to be evaluated during and after flood emergencies. Further study examining the potential synergistic effects of inexpensive, locally available agents such as bleach, lime and potash, should be conducted.


This research was funded by the emergency fund for the 2004 flood provided by UNDP, a private donation given by Professor Mitsuaki Nishibuchi, Center for Southeast Asian Studies, Kyoto University, Japan and the Government of Japan. ICDDR,B acknowledges with gratitude the commitment of UNDP, Professor Mitsuaki Nishibuchi and Government of Japan to the centre's research efforts.