Air and hygiene quality in crowded housing environments – a case study of subdivided units in Hong Kong

Abstract

The purpose of this study is to explore the environmental quality and hygiene in crowded living environments, subdivided units in Hong Kong. Subdivided units are an emerging form of housing environment for the urban poor. It is hypothesised that subdivided unit residents have a higher risk of exposure to poor hygiene conditions but no measurement has ever been taken to test this hypothesis. Twenty questionnaires and environmental assessments were conducted. Dominant bacterial species were identified as Micrococcus luteus and Staphylococcus spp., and the microbial counts were correlated with building, occupants and environmental parameters. Driven by the high bacterial counts and poor hygiene observation, eight subdivided units were selected for endotoxin, glucan and allergen analysis in bed and floor dust. Total airborne bacterial counts and endotoxin and glucan in dust were found at very high levels in some subdivided units, while unexpectedly, the allergen and mould levels were low. In crowded environments the skin bacteria may mislead the environmental and atmospheric bacterial contamination. Outdoor microbial pollution and deteriorated building conditions can be the main source of indoor contamination. ‘Good’ or ‘Excellent’ class of bacterial counts satisfying the Indoor Air Quality Objective does not guarantee a low endotoxin and glucan level.

Keywords

Subdivided units indoor air quality environmental hygiene bioaerosols endotoxin

Introduction

The trend for human populations to flock to cities for employment and cultural amenities, among other things, shows no signs of stopping. Arguably as a result, indoor and outdoor living environments, particularly for the urban poor, are worsening. A number of surveys have reported that about 35,000 to 280,000 households are living in subdivided units (SDUs) in Hong Kong.^1–3 A ‘subdivided unit’ is defined as individual living quarters that have been divided into two or more smaller units for rental.¹ These units are generally perceived to have poor indoor environment quality (IEQ), and residents are worried about their health as a result of exposure to a substandard environment. However, more specifically, no measurement has ever been taken to determine the extent to which a substandard environment in SDUs poses a health risk and the causes of the poor environment.

Housing and health are intricately linked in a variety of ways.^4,5 Indoor air quality (IAQ) measurements for certain chemical (e.g. VOCs and CO₂), physical (e.g. PM₁₀) and biological (e.g. bacteria and fungi) agents are commonly used as indicative parameters for further environmental investigations.⁶ Some environmental and hygiene factors such as noxious odours, cleaning routine, abundance of pests and mis- and non-management of waste around living areas are treated as indicators for environmental hygiene quality and signs of the presence of environmental hazards. Environmental allergens, from cockroaches, rodents and dust mites, endotoxin and glucan are some of the common agents found in low-income inner city areas and are associated with asthma, skin irritations and other respiratory symptoms in children and susceptible adults.^7–9 Dampness and mould contamination is also an acknowledged risk factor for poor health.⁵ Poor ventilation and maintenance of air conditioning systems may degrade occupants’ IAQ and provide breeding grounds for microbial growth.¹⁰ The rise in reports of fleas and pathogen vectors in Hong Kong represents yet another important environmental health concern linked to the urban environment and current lifestyles.¹¹

With a wide range of possibilities that can deteriorate the environmental quality, this study is proposed to collect data and identify the potential sources of contamination. In addition, the current IAQ Objectives used to assess the IAQ in public places in Hong Kong will be examined for use in this housing environment. Housing transformation can create new environmental challenges and health concerns to different groups of city dwellers.

Methods

Environmental assessment and questionnaires

Environmental assessment and questionnaires were designed to collect information about the occupants, building and building conditions, environmental hygiene and cleaning practices and the impression of environmental quality given by residents and our researchers. From July to September 2013 we successfully visited 20 households and interviewed 19 respondents in different districts.

IAQ analysis and dust sampling

With the consent of respondents, we arranged the IAQ measurements and other samplings before the environmental assessment and questionnaires. All the samplings were conducted in the daytime. These units were located in four districts: five in Tsuen Wan (TW); seven in Mong Kok (MK); three in Kwun Tong (KT) and five in Kwai Chung (KC). Each of the household units was identified through the use of location and unit number, e.g. TW01. Indoor and outdoor air samples were taken simultaneously in each SDU in accordance with the HK IAQ Certification Scheme.⁶ The parameters measured comprised the following: temperature (temp), relative humidity (RH), air velocity, carbon dioxide (CO₂), carbon monoxide (CO), total volatile organic compounds (total VOCs), PM₁₀ and airborne bacteria and fungi. Fungal count is not yet included in the IAQ objectives, but an indicative level of 500 CFU/m³, for both ‘Good’ and ‘Excellent’ classes, will be added to the next IAQ review.⁶ Samplers were also placed outside windows to obtain outdoor information. In units without windows or with restricted access, samples were taken in lobbies or building stairwells. Bacterial and fungal samples were collected using SAS SUPER 100 Air sampler with the corresponding agar plate (Bacteria: Tryptic soy agar, Fungi: Potato Dextrose Agar) and incubated at 35℃ for 48 h (Bacteria) and 28℃ for 3–7 days (Fungi). Floor and bed dusts were collected in accordance with ISACC phase II protocol and stored at −20℃ until analysis.¹²

Microbial identification

Other than the total bacterial and fungal counts in the IAQ assessment, bacteria were also counted separately, in accordance with their morphological differences, on agar plates and under the microscope after gram staining. The most abundant types of bacteria were further identified by sequencing the 16S rRNA gene at V3–V4 region using 341 Forward and 805 Reverse Primers on MiSeq System following the Illumina Nextera protocol (Illumina, USA). Fungi were identified at genus level based on the morphological difference.

Endotoxin, glucan and allergen extraction and analysis

Samples were extracted in accordance with the study Prevention of Allergy: Risk factors for Sensitization In children related to Farming and Anthroposophic Lifestyle (PARSIFAL).⁸ Extracted endotoxins were analysed using endpoint chromogenic Limulus amebocyte lysate (LAL) assay. Dust mite (Dermatophagoides pteronissinus), cockroach (Blattella germanica), rat and Aspergillus fumigatus allergens were measured using Der p 1 ELISA kit, Bla g 1 ELISA kit, Bla g 2 ELISA kit, Rat n 1 ELISA kit and Asp f 1 ELISA kit (Indoor Biotechnologies, Inc., USA). (1→3)-β-D-glucans were analysed using the Glucatell® assay kit (Associates of Cape Cod, Inc., USA). Sensitivity and limit of detection of these assays are provided by the manufacturer. A calibration curve was prepared for each sample run. Levels of the contaminants were presented as concentration per gram of dust (unit/g) and load per metre square of the sampling surface (unit/m²).

Statistical analysis

IAQ parameters that were measured in inside and outside environments of the units were presented as median and range; and the percentage of the units meeting the HK IAQ objectives was calculated for each parameter. Microbial parameters were correlated with building, occupant and environmental factors using Spearman’s rank correlation (p < 0.05 or p < 0.01). Microbial community data were fourth root transformed and statistically analysed using Analysis of Similarities (ANOSIM) calculated by PRIMER (v.6) (significant level p < 0.05) in order to determine the differences in the community composition among indoor or outdoor air in each district and total indoor and outdoor air.

Results

Questionnaire and environmental assessment

Table 1 summarises the characteristics of our sample sites. These old buildings are located mainly in deprived areas. Several conditions such as living area per person, ventilation quality, sanitation facilities and internal and external environments are inadequate. Some units have no toilet or kitchen facilities. For the units without a kitchen, some residents have to use the toilet area to prepare food. This practice is likely to increase the risk of food contamination. The general building and surrounding environments are considered to be substandard by the residents. Outside environments consist mainly of trafficked roads, back alleys, waste disposal sites and wet markets. Because these buildings are all low-rise, outside environments may play a direct role in influencing the indoor environment. Over 90% of the SDUs have an air-conditioning system in place, but in order to save money the air conditioning system is mainly operated at night when it is absolutely needed. During our sampling, all the SDUs are naturally ventilated. In so far as public hygiene is concerned, about 74% of the respondents consider the public hygiene as poor. The most frequent hygiene complaints are garbage in public area (53%) and poor ventilation and bad smell (47%). Over 63% of the respondents claimed that they clean their home everyday. All the households with children are more conscious of their cleaning practice and clean their home regularly and thoroughly. Garbage collection points are always located at the corner of stairs, a typical feature of old buildings and garbage containers are usually uncovered. In terms of pest infestation, mosquitoes (reported in 84% households), cockroaches (74%), rats (37%) and fleas (26%) are all reported. In addition, 32% of the households have mould contamination now or in last spring. Our researchers also ranked the living conditions of the SDUs. Four households are scored above 8 and are considered to be at very poor living conditions.

Table 1.

Characteristics of the sampling sites.

Parameters	Characters
Building age and number of SDUs in the flat	• All the buildings are over 30 years old. The median age is > 40 years. The oldest one is 60 years old. These building ages are considered to be old buildings in Hong Kong. • All the flats are at least sub-divided into three SDUs. One flat is divided into 12 units.
Living area per capita	• The average living area per capita in the SDUs is 3.8 m². According to the Hong Kong Housing Authority,^a households in public rental housing with <5.5 m² are categorised as overcrowded households.
Floor level and ceiling height	• The highest floor level is 8th, and the median floor level is 4th. In general, older buildings in Hong Kong have seven floors, and some SDUs are located in the rooftop structure (i.e. 8th floor). • The ceiling height standard in Hong Kong is 2.5 m based on the building regulation. All the SDUs meet this standard except three SDUs with 2.4 m height.
Windows	• About 16% of the SDUs have no window but some have vents to the corridor.
Kitchen	• About 79% of the households have no kitchen in their units. Two households use shared kitchen in the flat, while for others, some cook in the living room, bathroom or washroom.
Washroom	• Two households have no washroom in their units and use shared facility in the flat.
Water leakages and other building conditions	• Over 47% of the households have water leakage problems in their units. The most reported problem of the building is cracks in wall.

SDU: subdivided unit.

Hong Kong Housing Authority, ‘Housing in Figure 2013’ http://www.housingauthority.gov.hk/en/common/pdf/about-us/publications-and-statistics/HIF.pdf.

IAQ in SDUs

Of the IAQ parameters, temperature and relative humidity showed the worst performance; only 20% and 30% of the units met the ‘Good’ Class level, respectively (Table 2). Most of the units had a high temperature and were very humid during our visit, and the data correlated with the outdoor measurements. Most units reached the ‘Good’ Class for CO, total VOCs, PM₁₀ and fungal count levels. One unit (TW02) had particularly high indoor air movement due to the operation of a ceiling fan during our visit. It is likely that this indoor air movement has drawn outdoor air into the unit, with the outdoor air contributing to the high indoor PM₁₀ level (indoor 0.9 mg/m³; outdoor 0.8 mg/m³) and bacterial counts in this unit (>2614 CFU/m³). Although only 60% of the SDUs met the CO₂ ‘Good’ Class level, the SDUs below the ‘Good’ Class were not at extremely high CO₂ level (the majority was at approximately 1100 ppmv, one at 1300 and one at 1600 ppmv). It is worth noting that in some environments, outdoor CO₂ could be high too (up to 1092 ppmv in this study). Outdoor CO₂ levels were high in some units (30% > 600 ppmv outdoors), which could mislead the ventilation quality of indoor air. The high outdoor CO₂ may be indicative of busy traffic and close proximity to building exhausts and other sources of CO₂. During our SDU visits no occupant was cooking, and >50% SDUs had an electric induction cooker and no gas supply. The contribution of CO₂ and CO into the air due to cooking and heating was not significant when compared to conventional flats.

Table 2.

IAQ measured compared to HKIAQ objectives.⁶

Parameters	Indoor median (range); outdoor median (range)	% SDUs met (Good Class) or above
Temp (℃)	26.6 (22.7–29.6); 27.1 (23.5–30.5)	20% (<25.5)
RH (%)	74 (58–93); 73.5 (59–88)	30% (<70)
Air movement (m/s)	0.3 (0–2.15); 0.6 (0.12–0.5)	45% (<0.3)
CO levels (ppmv)	1 (0–2); 0.85 (0–2)	100% (<8.7)
CO₂ levels (ppmv)	879 (471–1600); 559 (378–1092)	60% (<1000)
PM₁₀ levels (mg/m³)	0.05 (0.026–0.9); 0.046 (0.025–0.8)	80%^a (<0.18)
Total VOCs (µg/m³)	0, except one 3749; 0	95% (<600)
Total bacterial counts (CFU/m³)	605 (133–>6535); 215 (60–>6535)	63%^b (<1000)
Total fungal counts (CFU/m³)	179 (50–695); 147 (61–453)	95% (<500)

IAQ: indoor air quality; SDU: subdivided unit.

% Calculated from 15 SDUs.

% Calculated from 19 SDUs.

The indoor total VOCs in one unit (KT03) was high (3749 µg/m³) compared to no detection outdoors. The potential source of VOCs was, however, unknown, e.g. no storage of solvents and no new furniture and decoration were reported but this unit was in fact located above a wood-crafting workshop. For the bacterial data, although approximately 63% of the units met the ‘Good’ Class levels, those not meeting the level reported up to six times higher bacteria than the ‘Good’ Class level (>6535 CFU/m³). The IAQ objectives only consider indoor measurements, but when we compared the indoor and outdoor microbial counts, SDUs had different indoor/outdoor count ratios (I/O), indicating the likely variation of the source of microorganisms. In summary, high temperature, RH and biological contamination from bacteria were of the greatest public health concern in the SDUs.

Microbial community

In total, 81 types of Gram +ve and 115 types of Gram −ve were identified morphologically. Seven species could not be identified and one species was not a bacterium. The most abundant and frequently identified bacteria indoors and outdoors were all Gram +ve bacteria. Using the 16S rRNA sequence method for identification, these dominant bacteria were Micrococcus luteus, Staphylococcus capitis, Staphylococcus cohnii and Staphylococcus epidermidis (Figure 1). The number of bacterial types ranged from 8 (MK07) to 63 (MK01) types, with a median of 30 types. The number of total bacterial counts in Figure 1 is less than in Table 2, because these counts are absolute CFU data without adjusting the statistical data conversion. Twenty-seven fungal species were isolated from the units and the outside environments among the genus Alternaria, Aspergillus, Aureobasidium, Bipolaris, Chaetomium, Curvularia, Mucor, Pythium, Rhizopus, Schizophyllum and Ulocladium (Figure 2). The most abundant and frequently identified fungi in both indoor and outdoor samples were Alternaria spp., Aspergillus spp. and Rhizopus spp.

Figure 1.

Bacterial diversity in indoor and outdoor air in the 20 SDUs.

Figure 2.

Fungal diversity in indoor and outdoor air in the 20 SDUs.

Table 3 shows the correlation between the microbial parameters and the environmental, building and occupant factors. For example, indoor bacterial counts were associated with indoor CO₂ and outdoor bacterial counts. The number of bacterial types indoors was associated with number of occupants and children living in the SDUs. Although all the dominant bacteria are also common human skin flora, only Staphylococcus capitis and Staphyloccocus cohnii showed a strong correlation with indoor CO₂ levels and number of occupants and people at sampling. Micrococcus luteus levels were inversely correlated with the number of households divided from the same flat. Indoor fungal levels were not associated with any factor, while the levels of individual fungi, Alternaria spp., Aspergillus spp. and Rhizopus spp. were associated with some building parameters such as building age and number of SDUs divided from the flat. Outdoor fungal levels were correlated to the outdoor temperature and RH.

Table 3.

Spearman’s correlation between biological parameters and environmental, building and occupant factors.^a

Biological parameters	Significantly correlated with/inversely correlated with		r Coefficient
Indoor bacteria	Occupant/Environment	Indoor CO₂ conc.	0.480^b
	Environment	Outdoor bacteria	0.576^c
Outdoor bacteria	Environment	Indoor bacteria	0.576^c
Bacteria I/O	Occupant/Environment	Indoor CO₂ conc.	0.551^b
Indoor G +ve/G −ve	Environment	Indoor PM ₁₀	−0.606 ^b
Indoor G +ve/G −ve	Environment	Outdoor PM ₁₀	−0.551 ^b
Outdoor G +ve/G −ve	Building	Ceiling height	0.653^c
Indoor bacterial types	Occupant/Environment	Indoor CO₂ conc.	0.581^c
	Occupant	Household size	0.620^c
		Number of kids	0.543^b
		Number of people at sampling	0.488^b
Outdoor bacterial types	Occupant/Environment	Indoor CO₂ conc.	0.481^b
Outdoor bacterial types	Occupant	Household size	0.495^b
Indoor fungi		No significant correlation
Outdoor fungi	Environment	Outdoor temperature	−0.537 ^b
Outdoor fungi	Environment	Outdoor RH	0.455^b
Fungi I/O		No significant correlation	No sig.
Indoor Alternaria	Building	Building age	0.703^c
Indoor Aspergillus	Building	Number of units in flat	−0.634 ^c
Indoor Rhizopus	Building	Number of units in flat	−0.712 ^c
		Odor and ventilation quality	−0.508 ^b
		Hygiene of living Room	−0.513 ^b
		Hygiene of washroom	−0.638 ^c
		Average overall hygiene score ^d	−0.493 ^b
Indoor M. luteus.	Building	Number of units in flat	−0.620 ^c
Indoor S. epidermidis		No significant correlation
Indoor S. capitis	Occupant/Environment	Indoor CO₂ conc.	0.483^b
Indoor S. cohnii.	Occupant/Environment	Indoor CO₂ conc.	0.779^c
	Occupant/Environment	Outdoor air velocity	−0.638 ^c
	Occupant	Household size	0.512^b
	Occupant	Number of people at sampling	0.487^b

I/O = Indoor/outdoor ratio.

Factors correlated: Environment – Inside and outside temperature, RH, CO₂, PM₁₀, air velocity, bacteria, fungi; Building – Living area, living area per capita, building age, floor levels, number of units in flat, ceiling height, window coverage area, hygiene of living room, hygiene of washroom, odor and ventilation quality, average overall hygiene score; Occupants – Household size, number of people at sampling, number of kids under the age of 12, average hours staying at home.

Correlation is significant at the 0.05 level (2-tailed).

Correlation is significant at the 0.01 level (2-tailed).

Average hygiene score = Hygiene score given by our researchers (from 1 to 10; 1: Clean, 5: Acceptable, 10: Very Dirty).

Based on the ANOSIM analysis, the grouped indoor and outdoor microbial communities across all districts were significantly different (p = 0.008). The outdoor bacterial community of TW was markedly different from that of MK (p = 0.002), while MK was different from KC (p = 0.002), MK differed from KT (p = 0.022) and KT differed from KC (p = 0.043). The indoor air bacterial communities among the districts were significantly different between TW and MK (p = 0.028) and TW and KC (p = 0.034). Fungal communities between indoor and outdoor samples and districts were also demonstrably different (p = 0.001). As illustrated in Figure 3, the outdoor fungal diversity was clustered by districts. The closer the sampling locations shown would mean the higher similarity of their fungal diversity.

Figure 3.

Comparison of fungal diversity in different districts.

Level of allergens, endotoxins and glucans in eight units

The poor IAQ and hygiene conditions reported in the questionnaire and the high levels of bacteria in some units strongly verify that biological pollution is a concern in SDUs. A wide range of biological pollutants may be present in this environment (Table 4). In general, indoor dust levels (median floor = 0.087 g/m², range = 0.033–0.22 g/m²; median bed = 0.033 g/m², range = 0.0097–0.10 g/m²) were low, which may relate to the regular cleaning practices and type of furnishings. No household uses carpet or sofas and bedding areas are tidied up for use in the daytime as living space. The median endotoxin concentration on floors was in-between 10,609 and 11,619 EU/g, and on beds was 4993 EU/g. The lowest and the highest levels were shown to be 69 and 36 times different on floors and beds, respectively. This implies a wide range of contamination levels among the SDUs. Endotoxin concentrations on floors were significantly correlated with the outdoor bacterial counts (r = 0.714), while concentrations on beds were negatively correlated with both indoor and outdoor Gram +ve/−ve bacteria ratio (r > −0.88). For endotoxin loads, the median load on floors was between 831 and 885 EU/m² and on beds was 164 EU/m². The difference between the highest and the lowest loads was 46 and 143 times on floors and beds, respectively. The loads on floors were correlated with the indoor bacterial counts (r = 0.97), while the loads on beds were negatively correlated with both indoor and outdoor Gram +ve/−ve bacteria ratio (r >−0.88). For glucan concentrations (Table 4), the difference between the highest and the lowest glucan concentration among the sampled units could be over 20 times on floors and 14 times on beds. Interestingly, the glucan concentrations on floors were negatively correlated with the outdoor Gram +ve/−ve bacteria ratio (r = −0.826), while the concentrations on beds were correlated with the hours respondents spent at home (r = 0.805); the longer they spent time at home, the higher were the glucan concentrations on beds. In terms of the loads, the highest and the lowest load difference could be over 80 times on floors and 11 times on beds. The glucan loads on floors showed no correlation with other parameters, while the loads on beds were negatively correlated with the total bacteria counts inside the SDUs (r = −0.821). Most respondents reported cockroaches, rats and mould in their units. However, no SDUs exhibited detectable allergens, except in MK01: 3.62 U/g of Bla g1.

Table 4.

Levels of indoor and outdoor microbes and endotoxin and glucan in dust samples.

Locations	Indoor bacteria (CFU/m³)	Bacteria In/Out	Indoor fungi (CFU/m³)	Fungi In/Out	Endotoxin floor/bed conc.(EU/g)	Endotoxin floor/bed load (EU/m²)	Glucan floor/bed conc.(µg/g)	Glucan floor/bed load (µg/m²)
TW01	>2158	15.87	104	1.01	11619/10314	885/1072	21.22/7.14	1.62/0.74
TW02	>2614	1	89	1.05	53952/52043	2360/1129	24.05/36.92	1.05/0.80
TW04	363	6.05	148	0.87	3188/60246	507/2280	>41.36/100.11	>6.58/3.79
MK01	>1560	8.21	250	1.17	10609/2840	355/101	12.15/41.62	0.42/1.49
MK06	173	0.82	135	1.08	1398/4142	205/104	>65.27/60.46	>9.58/1.51
MK07	>6535	1	380	2.62	96166/1671	9429/16	3.22/19.83	0.32/0.19
KC01	>6535	16.84	308	3.84	Floor 10559	Floor 2373	Floor 28.45	Floor 6.39
KC04	1643	0.89	168	0.99	13777/4993	831/164	18.72/10.32	1.13/0.34

Discussion

Our findings reveal some of the air quality and hygiene issues in SDUs. Although most IAQ parameters are satisfactory, the high levels of airborne microbial counts and endotoxin and glucan levels in dust are worthy for further discussion. In this section, we first compare the fungal and bacterial levels between SDUs and the conventional homes in Hong Kong because local environments can play an intrinsic role in the environmental microbial level. Only when the microbial levels in SDUs are higher than the conventional homes, we can presume a poorer microbial air quality in SDUs. Afterwards, the health risk of endotoxin and glucan found in the SDUs is discussed by referencing the health outcomes after exposure to these contaminants in other studies. The allergen levels in SDUs are compared to the local data as well. Microbial counts are commonly used in IAQ inspection but SDUs are a unique environment crowed with people, which are one of the main sources of bioaerosols in indoor environments. Therefore, we also discuss the potential limitation of using the current bacterial count method to indicate the microbial air quality. Finally, we examine some of the urban and building management issues hoping to improve the environmental quality in SDUs.

Fungal and bacterial counts in SDUs in comparison with the conventional homes

Airborne fungal and bacterial counts are indicative parameters, which reflect the potential biological contamination such as microbial growth and emissions from a proximate source and the accumulation of microbes due to poor cleaning or inadequate ventilation. Among our tested parameters the bacterial level could be six times higher than the HK IAQ recommended ‘Good’ Class objective. Results from other studies in Hong Kong have shown that 100 to >85% of conventional living rooms had indoor bacterial counts below 1000 CFU/m³. Furthermore, the conventional flats exceeding the level had bacterial counts below the average 1500 CFU/m³.^13,14 The implication here is that not only do more SDUs have poor biological air quality, but the level of contamination in SDUs is also much more severe. In the Lee et al.¹⁴ study, the conventional flats with high indoor bacterial counts also had a similar level of outdoor bacterial counts, thereby designating outdoor air as the source of the indoor air pollution. However, in our study, only three of the seven problematic units were directly associated with an outdoor source, and the remainder were linked to indoor contamination. Most of the SDUs we visited had no separate kitchen area: cooking and food preparation were carried out in the same space as the living room and bedroom. This squeezed layout may also contribute to the high bacterial counts in SDUs.¹⁴ The sampling period of the Lee et al.¹⁴ study also took place from July to October. Outdoor air samples can therefore be reasonably compared between ours and their study; in their study, only one outdoor sample exceeded 1000 CFU/m³ and the rest (five samples) had bacterial counts below 500 CFU/m³, i.e. 17% of the outdoor air had a high bacterial count above 1000 CFU/m³. These values closely approximate to ours, in which 16% of the outdoor air in SDUs exceeds 1000 CFU/m.³ However, our samples (two out of three) have much higher counts of over 2500 CFU/m³, exceeding the upper range measured in their study (about 1500 CFU/m³). In other words, although the frequency of encountering poor environments outside the SDUs is no higher than outside the conventional flats, the environmental quality can be much worse where a problem arises.

SDUs are concentrated in deprived areas so we also investigated whether each district has its own microbial communities. Interestingly, fungal communities are distinctive in each district, while bacterial communities in some districts differ from other districts. Our finding implies that neighborhood outdoor environments contribute to the spatial disparity of indoor biological air quality.¹⁵ Although further studies are needed to assess the health implications of exposure to distinct microbial communities in different districts, our study has highlighted that living in different districts can have different inherent risks of exposure to different types and levels of microbial species. For fungal counts, although humid weather conditions (high temperature and RH), poor ventilation and over 40% of mouldy conditions were reported in the SDUs compared to only 27.5% of visible dampness and mould patches present at conventional homes,¹⁶ fungal counts were mostly at the ‘Good’ Class level.

Health risks of endotoxin and glucan

Unlike the total microbial count, which can only serve as an indicative parameter, endotoxin and glucan are biological hazards that can precipitate different health responses after exposure. Among the SDUs tested, 87.5% of them (seven units) had at least one of the endotoxin measurements (either on bed or floor; concentrations or loads) above the median value found in conventional homes with asthmatic children in Hong Kong.¹⁷ In addition, as in the previous study, all homes had detectable endotoxin levels. In the Leung et al. study,¹⁷ they found that mattress endotoxin levels were associated with the increasing frequency of wheezing episodes (11,300 EU/g associated with 1–3 attacks and 12,300 EU/g with >12 attacks). The implication here is that two SDUs (TW02 and TW04) with >50,000 EU/g can trigger >12 wheezing attacks if asthmatic children live in the units. In the Park et al.¹⁸ study of endotoxin and fungi in floor and chair dust in a water-damaged office building, 33,700 EU/mg of floor dust and 6700 EU/m² of floor area were categorised as high contamination levels. Endotoxin levels in MK07 exceeded both of the aforementioned levels. Park et al.¹⁸ reported a causal relationship of endotoxin and fungal exposure on wheezing, cough attacks and throat irritation. Other factors can impact the indoor endotoxin level such as cigarette smoking, cockroach infestation and cat and dog ownership.¹⁹ Smoking habits of the occupants were not collected in this study, while no household owns a cat or dog.

With regard to indoor glucan concentrations and loads, we found no reference in Hong Kong; instead a study from Taiwan was used for comparison due to the similarity in terms of people and environment with Hong Kong. Half of the housing units had glucan concentrations in bed dust above the geometric mean of 25.5 µg/g and 27.9 µg/g on the sleeping pad and the bamboo side of the mattress, respectively.⁹ In our study some of our respondents also used a bamboo mattress to help reduce heat during sleep in the summer time. According to a review of glucan and respiratory health,⁷ 0.169 µg/m² of glucan on living room floors were associated with peak flow variability, particularly in atopic children from a similar indoor study. All of our SDUs had glucan loads above this value. However, it is noteworthy that respiratory health effects of glucan are not well defined.⁷

Allergens

Different kinds of allergens are commonly detected in inner city homes and are associated with the development of asthma.^20,21 Gruchalla et al.²⁰ reported that housing type influences the levels of cockroach (higher risk in high-rise flats) and dust-mite allergens (higher risk in detached homes). MK01 had 3.62 U/g of Bla g1, surpassing the sensitisation level of 1 U/g. Bla g1 levels over 1 U/g are associated with sensitisation in asthmatic children, and over 2 U/g increases the risk of wheezing in infants.^9,16,22 MK01 had the best hygiene score but occupants also reported cockroaches at home and in common areas. The windows were facing wet markets and restaurants, the most likely source of the cockroaches. The undetectable levels of cockroach allergens in most SDUs may be due to the type of cockroach species present in the area. In our cockroach survey study (data not shown), Supella longipalpa and Periplaneta australasiae are the dominant species found at homes in Hong Kong. Although Bla g 1 and Bla g 2 from Blattella germanica are the common markers for cockroach infestation, these markers are limited to detect other cockroach species. The simplicity of room layout and furnishings may reduce cockroach hiding places. For example, fixed cupboards are uncommon, most of the units have no separate kitchen, and tables and other kitchen utilities are moved around regularly while in use. In addition, tenants usually move frequently in SDUs, e.g. every 1–2 years, which could help to reduce pest infestation in the room due to cleaning. The sighting of cockroaches and rats is likely to be associated with a poor outdoor environment. Kulhankova et al.²³ found that the effect of co-exposure of endotoxins with cockroach allergens was more than additive. The implication here is that in a complex environment like SDUs, where co-exposure to various contaminants may occur, an integrated approach to health assessment should be considered.

In terms of dust mite allergens, Leung et al.¹⁷ found that the median value of Der p 1 in bed dusts in the control and asthmatic homes were about 2 µg/g and 0.61 µg/g, respectively, and 58 to 70% of the samples had detectable levels in Hong Kong. However, none of our samples had detectable levels of Der p 1, which may be due to the regular use of bed space for other purposes. As a result of this practice, favourable conditions for the growth of dust mite are limited. Other than Der p 1, Blo t 5 from Blomia tropicalis has also been measured in house dust samples in Hong Kong.²⁴ In relation to the assessment of risk, other than the environmental exposure level, sensitivity to these allergens is also important in the determination of risk. Yuen et al.²⁵ reported that among the 977 patients with chronic rhinitis in Hong Kong, 63% had positive skin prick test reactions with house dust mite allergens, 23% with cockroach, 14% with cat, 5% with dog, 4% with pollen and 3% with mould. In summary, allergen levels in SDUs were unexpectedly lower than those of conventional homes in Hong Kong.

Limitation of the bacterial count method in IAQ objectives

In 2003 the IAQ Management Group of the HKSAR Government⁶ issued a set of objectives to assess IAQ in offices and public places, but at that time did not apply objectives for residential buildings. Only a few available studies focussing on residential environments in Hong Kong have been published.^13,14,17 Total airborne bacterial count is one of the most common biological parameters used in IAQ inspection. It is also the only parameter used for biological analysis in the IAQ objectives.⁶ It is therefore necessary and timely to carry out investigations of environmental quality and hygiene in SDUs, and through this study to examine and improve the use of airborne bacterial and fungal counts in these crowed living environments. The HKIAQ recommendation states that ‘the exceedance of bacterial count does not necessarily imply health risk but serves as an indicator for further investigation’. In air-conditioned public places and offices, cleaning the air-handling system is one of the most common actions taken if the bacterial count level is poor, under the assumption that the air-handling system is the most probable source of airborne bacteria. However, what should be the appropriate further investigation in SDUs? More importantly, should we eliminate further investigation in the SDU units rated with a ‘Good’ Class microbial level because they are supposedly ‘clean’?

The results show that in crowded environments like SDUs, the environmental and atmospheric bacterial contamination may be masked or over-loaded by the skin flora,²⁶ although skin floras can live in environments other than human skin, such as in soils and by forming biofilms on various surfaces.²⁷ We are aware that when sampling indoor and outdoor environments at the same location, cross sampling between indoor and outdoor air can happen; however, with the close proximity between buildings and flats, bacteria originated from human sources are likely to pass from one place to the others. Therefore, the use of the conventional non-selective agar medium in sampling may actually limit the indicative power of the bacterial counts for other sources of environmental microbial contamination. Although physical and chemical parameters are commonly used in IAQ inspection, total microbial counts are not applied globally to reflect potential microbial contamination because there is no solid evidence linking the total microbial exposure levels to pathogens or hazardous agents, and subsequently, to health effects. More importantly, based on our data we can summarise that only endotoxin loads in floor dusts were correlated with the indoor bacterial counts. It is important to keep in mind that a low microbial count level does not guarantee low bio-contamination. Therefore, on the basis of our research, it is not justifiable to quit further investigation into the unit(s) with a ‘Good’ Class microbial objective. Public health inspectors should be cautious about using microbial counts as indicative parameters for hygiene investigation.

Urban and building management

Traditionally, the management of the domestic environment is the occupant’s responsibility, and environmental health intervention is focused on educating and changing the occupant’s behaviour. However, in this complex environment, where the source of contamination is multiple and usually beyond the control of the tenants, this occupant-centred strategy is not adequate to secure a healthy environment. We assessed that most of the SDUs we visited are considered ‘clean’. SDUs with children have the best hygiene score, while SDUs occupied by retired men have the worst scores. This finding verifies that cleaning is critical to the reduction of endotoxin and glucan levels. However, in TW02, the occupants with children cleaned their floors every day and washed bed sheets every 10 days using water, cleansing detergent or bleaching agents (i.e. they clean more frequently than do other SDUs); they nevertheless still had a high floor and bed endotoxin concentration and load. Considering the outside environment of TW02, the high contamination level may be due to the rubbish and other poor conditions in the outside environment, which continuously release organic dusts into the SDUs. Here lies a complex problem with regard to cleaning the outer environment. In old buildings owners may not be as proactive in building maintenance as the new building owners may be. In some situations, responsibility for maintenance may not be clearly defined because some buildings may not comply with stated building codes and thus be illegal. Finally, there is no recommended guideline or standard of ‘clean’ to follow.

Poor building conditions are likely to be the leading contributor of the highest endotoxin concentration and load. Ventilation infrastructure, e.g. windows and air vents, and their quality and accessibility certainly can play a role in reducing pollutant levels emitted from indoor sources. We found that close proximity between buildings and multiple units subdivided inside the same space, security issues, privacy and unpleasant outdoor environments can restrict the use of ventilation infrastructure and further worsen the indoor condition. A squeezed environment can further restrict airflow and ventilation. For example, in one of the SDUs we visited, three adults live together in the unit. Clothes are hung around the living space (no wardrobe) further blocking sunlight and ventilation from the already limited window space. The overcrowded environment can also increase moisture accumulation in the unit. Most occupants would open windows widely during summer time. The opening of windows would help to remove indoor moisture and reduce condensation. Arguably though, it also reinforces the indoor and outdoor air quality connection, and if outside air quality is poor, it would pose a significant impact on the indoor environment. Poor hygiene conditions in the surrounding environment such as garbage or sewage accumulation in canopies and patios are characteristic of SDUs. For the SDUs with particularly high levels of PM₁₀ and airborne bacterial counts, a nylon shade roof covered with rubbish and clutter was located outside these units. For the unit with a high level of VOCs, although we could not confirm if the lower level wood-crafting workshop is the source of VOCs, it reinforces the problem that surrounding areas can also largely impact on the SDU environment. Various wood, glass and metal workshops, recycling facilities and auto repair shops are closely situated around some SDUs. Other SDUs are located directly above restaurants, e.g. MK01, which can increase the risk of pest infestation. Further studies should also examine these non-residential sources of pollution.

Conclusions

Housing transformation creates new environmental challenges and health concerns. This study explored the environmental quality of some SDUs and justified the need of more health-oriented studies in this housing environment. We hope this information will be useful for different professionals to improve the SDU environment. From our experience, many IAQ inspectors only consider the bacterial count to decide the follow-up action in the hygiene aspect. Although airborne bacterial counts are useful to indicate the environmental quality and hygiene as we perceived, other signs of poor hygiene should not be ignored. Neighbourhood environments, built environments, building conditions and occupant’s behaviour can all contribute to the indoor environmental quality, a more integrated approach can help create a healthy living environment.

Footnotes

Acknowledgements

We thank Tommy Lo at the World Green Organization for organizing the site visits and Maggie Chan and KK Ma at HKBU for their technical support.

Authors’ contribution

Dr Lai and Dr Yu initiated this research and contributed to the design of the study. Dr Lai led her team to conduct the study and prepare the manuscript, while Dr Yu arranged for the site visits. Dr Lee contributed to the microbial diversity analysis. All the authors read, commented and approved the final manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Policy 21 Ltd. Survey on subdivided units in Hong Kong. Transport and Housing Bureau, The Government of the Hong Kong Special Administrative Region, Hong Kong, China; 2013.

Platform of Concerning Subdivided Units and Issue in Hong Kong. Platform of Concerning Subdivided Units and Issue in Hong Kong, Hong Kong, China; 2013.

Transport and Housing Bureau. Long Term Housing Strategy Consultation Document. The Government of the Hong Kong Special Administrative Region, Hong Kong, China; 2013.

Krieger

Higgins

. Housing and health: time again for public health action. Am J Public Health 2002; 92: 758–768.

World Health Organization. WHO guidelines for indoor air quality: dampness and mould, Copenhagen, Denmark: WHO Regional Office for Europe, 2009.

Indoor Air Quality Management Group. A guide on indoor air quality certification scheme for offices and public places. The Government of the Hong Kong Special Administrative Region; Hong Kong, 2003.

Douwes

. (1→3)-β-D-glucans and respiratory health: a review of the scientific evidence. Indoor Air 2005; 15: 160–169.

Schram

Doekes

Boeve

Douwes

Riedler

Ublagger

von Mutius

Budde

Pershagen

Nyberg

Alm

Braun-Fahrländer

Waser

Brunekreef

. PARSIFAL Study Group. Bacterial and fungal components in house dust of farm children, Rudolf Steiner school children and reference children-the PARSIFAL Study. Allergy 2005; 60: 611–618.

Siebers

Chang

Hsieh

Chen

Pierse

Crane

. Indoor allergens and microbial bio-contaminants in homes of asthmatic children in central Taiwan. J Asthma 2009; 46: 745–749.

10.

Leung

Tang

Yang

Chan

Lin

Nielsen

Niu

Qian

Sleigh

Sundell

Wong

Yuen

. Role of ventilation in airborne transmission of infectious agents in the built environment: a multidisciplinary systematic review. Indoor Air 2007; 17: 2–18.

11.

Wong

Leung

Lai

Wong

Chan

Que

Chow

Yuen

Lau

Simon

. Review of vector-borne diseases in Hong Kong. Travel Med Infect Dis 2011; 9: 95–105.

12.

Weiland

Björkstén

Brunekreef

Cookson

von Mutius

Strachan

. International Study of Asthma and Allergies in Childhood Phase II Study Group. Phase II of the international study of asthma and allergies in childhood (ISAAC II): rationale and methods. Eur Respir J 2004; 24: 406–412.

13.

Lee

Chang

Chan

. Indoor and outdoor air quality investigation at six residential buildings in Hong Kong. Environ Int 1999; 25: 489–496.

14.

Lee

. Investigation of indoor air quality at residential homes in Hong Kong: case study. Atmo Environ 2002; 36: 225–237.

15.

Rintala

Pitkaranta

Toivola

Paulin

Nevalainen

. Diversity and seasonal dynamics of bacterial community in indoor environment. BMC Microbiol 2008; 8: 56–68.

16.

Leung

Lam

Chan

Lee

Chan

Pang

Lai

. Indoor environment of residential homes in Hong Kong: relevance to asthma and allergic disease. Clin Exp Allergy 1998; 28: 585–590.

17.

Leung

Wong

Chan

Yung

Wong

Lam

Wong

. Indoor determinants of endotoxin and dust mite exposures in Hong Kong homes with asthmatic children. Int Arch Allergy Immunol 2010; 152: 279–287.

18.

Park

Cox-Ganser

Rao

Kreiss

. Fungal and endotoxin measurements in dust associated with respiratory symptoms in a water-damaged office building. Indoor Air 2006; 16: 192–203.

19.

Hasday

Bascom

Costa

Fitzgerald

Dublin

. Bacterial endotoxin is an active component of cigarette smoke. Chest 1999; 115: 829–835.

20.

Gruchalla

Pongracic

Plaut

Evans

Visness

Walter

Crain

Kattan

Morgan

Steinbach

Stoul

Malindzak

Smartt

Mitchell

. Inner city asthma study: relationships among sensitivity, allergen exposure, and asthma morbidity. J Allergy Clin Immunol 2005; 115: 478–485.

21.

Arshad

. Does exposure to indoor allergens contribute to the development of asthma and allergy? Curr Allergy Asthma Rep 2010; 10: 49–55.

22.

Tranter

. Indoor allergens in settled school dust: a review of findings and significant factors. Clin Exp Allergy 2005; 35: 126–136.

23.

Kulhankova K George

Kline

Snyder

Darling

Field

Thorne

. Early life co-administration of cockroach allergen and endotoxin augments pulmonary and systemic responses. Clin Exp Allergy 2009; 39: 1069–1079.

24.

Sun

Chan

Chik

Wong

Zhong

. House dust mite allergen (Derp1 and Blot5) levels in asthmatics’ home in Hong Kong. Chinese Med Sci J 2004; 19: 185–188.

25.

Yuen

Cheung

Tang

Wong

Cheung

. The skin prick test results of 977 patients suffering from chronic rhinitis in Hong Kong. Hong Kong Med J 2007; 13: 131–136.

26.

Gandara

Mota

Flores

Perez

Green

Gibbs

. Isolation of Staphylococcus aureus and antibiotic-resistant Staphylococcus aureus from residential indoor bioaerosols. Environ Health Persp 2006; 114: 1859–1864.

27.

O’Gara

Humphreys

. Strataphylococcus epidermidis biofilms: importance and implication. J Med Microbiol 2001; 50: 582–587.