Potential Health Risks of Indoor Particulate Matter Heavy Metals in Resource-Constrained Settings of South Africa

Abstract

Evidence suggests that pollutants like particulate matter and heavy metals significantly impact health. This study investigated the determinants and sources of indoor PM_2.5 heavy metals and assessed the health risk to children in Durban. Thirty households of mothers participating in the mother and child in the environment (MACE) birth cohort study were assessed for PM_2.5 heavy metals. Multivariate linear regression was used to identify the determinants of PM_2.5 heavy metals, while Pearson correlation and principal component analysis identified their sources. The health risk for children was assessed using measured metal concentrations. Proximity to industry, wall type, house age, and presence of windows increased some PM_2.5 heavy metals, while cross ventilation reduced indoor PM_2.5 As and Cu levels. Pearson correlation and principal component analysis indicated household, traffic, industrial, and natural sources as potential contributors. The health index was greater than 1, and cancer risk values for PM_2.5 As and Pb exceeded acceptable levels. The findings highlight the toxicity of indoor air due to heavy metals and the potential for adverse health outcomes in children. To protect vulnerable groups, it is essential to prevent exposure to high-risk metals.

1. Introduction

Exposure to indoor air pollution is a major source of morbidity and mortality globally [1,2,3]. Indoor air pollution is linked to a higher incidence of respiratory and cardiovascular diseases [4,5,6], adverse birth outcomes, chronic obstructive pulmonary diseases, and other conditions [2,7,8,9,10,11,12,13].

Particulate matter, a complex mixture of solid and/or liquid particles, is one of the main pollutants found indoors. Indoor PM_2.5 can be generated through cooking, indoor combustion activities (i.e., burning candles, incense, etc.), and certain outdoor sources. Outdoor sources include fossil fuel combustion, diesel truck exhaust fumes, and vehicle emissions, construction, industrial activities, and vehicle component wearing (e.g., brakes, tires, paint) [14,15,16,17,18]. These PM_2.5 emissions also contain various toxic heavy metals, including arsenic (As), lead (Pb), and cadmium (Cd) [19,20].

Small particles (less than 2.5 μm in diameter or smaller) can penetrate the lungs, and some may even get into the bloodstream [14]. The toxicity of the heavy metals in PM_2.5 depends on the type of metal and the levels of exposure. For example, metals such as copper (Cu) and zinc (Zn) are vital for life and yet have the potential to cause harmful effects at very high levels of exposure. As, Cd, chromium (Cr), nickel (Ni), and Pb are considered toxic pollutants due to their carcinogenic and mutagenic properties. Exposure to even low levels of these metals can affect the brain, kidneys, and heart and even result in death [21,22]. The main routes of exposure to heavy metals are oral ingestion, dermal contact, and inhalation. Children are at increased risk, as exposure may also occur through putting contaminated fingers, toys, and other objects into the mouth, and a condition called pica, in which children eat non-food items, such as soil.

The contamination of the environment by heavy metals in relation to the background values (i.e., amount of chemicals that occur naturally in a specific environment) has been assessed by many studies using pollution indices (i.e., enrichment factor, contamination factor, pollution load index, and geo accumulation index) [20,23]. Gruszecka-Kosowska [23] assessed the impact of heavy metals bound with PM on the level of contamination by calculating pollution indices for a range of heavy metals. The enrichment indices indicated that the study area was highly contaminated with Cd and Zn, considerably contaminated with As, Pb, and tin (Sn), and moderately contaminated with Cu [23]. Similar findings were noted in a study carried out in Japan, where the study area was reported to be contaminated with PM_2.5-bound As, Pb, Cu, Ni, and Zn [20].

The direct risks of heavy metals are significant, especially among children. In order to determine the potential risk of children developing adverse health outcomes, the risk of carcinogenic and non-carcinogenic effects through skin contact, inhalation, and ingestion can be considered. Different studies assessing the effect of PM_2.5-bound heavy metals on health have reported a cause for concern. In Iran, the lifetime cancer risk values for Cd in both children (1.2 × 10⁻⁴) and adult (4.8 × 10⁻⁴) groups indicated elevated risk of cancer development in humans [19]. In China, the health index and carcinogenic risk values ranging from 2.0 × 10⁻⁴ to 5.0 × 10⁻² and 1.14 × 10⁻⁸ to 2.61 × 10⁻⁶, respectively [24], were reported. In Nigeria, a health risk index above 1 was also reported among children exposed to heavy metals in the respirable fraction of PM [25].

This study formed part of the mother and child in the environment (MACE) birth cohort, which investigates the impact of pollution on children’s health in Durban. The current study aims to (i) describe the severity of indoor PM_2.5 heavy metal pollution, (ii) identify the predictors and common sources of indoor PM_2.5 heavy metals, and (iii) evaluate the health risks posed by indoor PM_2.5 heavy metals to the children living in Durban.

2. Materials and Methods

2.1. Study Area, Population, and Selection of Homes

The study was conducted at eThekwini Municipality, which is found in the South African province of KwaZulu-Natal. The eThekwini municipality is located on the east coast of South Africa and experiences a tropical climate, characterized by hot summers and mild winters [26,27]. The municipality has a population of 3.9 million, of whom 34% are children [26,27]. This study was part of the MACE birth cohort project, which explores the effects of indoor and outdoor air pollution on pregnancy outcomes and childhood health in the eThekwini Municipality. Residential suburbs situated in the south (industrialized area) and north (less industrialized area) of Durban are part of the study. The major sources of pollution in the south of Durban include oil refineries, Sapref, Mondi, Illovo, and Hullet, and chemical plants [28]. Other sources of air pollution in this area include traffic emissions from the major highways and household combustion activities [15,16]. Homes situated in four suburbs, namely, Wentworth, Bluff, Merebank, and Clairwood, were assessed. In the north of Durban, air pollution originates mainly from traffic and household combustion activities [28], and the residential suburbs such as Newlands East, Newlands West, and KwaMashu were assessed in this area. The eThekwini region of South Africa is a coastal metropolis. It is characterized by hilly terrain that is dissected by rivers such as Umgeni, Umlazi, and Mbokodweni. The ground adjacent to the sea is made up of ancient red sand dunes, which are extremely steep in some places. The geology of eThekwini is also characterized by sandstones, quartz arenites, subarkoses, and arkoses [29].

Thirty homes of MACE participants were chosen based on the presence of an adult member at home. All participants gave written consent, and the study received approval from the Biomedical Research Ethics Committee at the University of KwaZulu-Natal (Reference Number: BE572/17). The geographical location and sampling points of the study area are shown in Figure 1.

2.2. Sample Collection and Analysis

In each home (n = 30), a living room was selected for sampling PM_2.5. When a living room was not available, other frequently used rooms, such as the bedroom or cooking area, were selected. Airmetrics MiniVol samplers (Airmetrics, Eugene, OR, USA) equipped with Teflon membrane filters (Pall Corporation, Ann Arbor, MI, USA) were used to collect indoor PM_2.5 samples over a 24 h period. The particle size separation was achieved using Air Metrics MiniVol samplers, which are equipped with size-selective inlets to segregate particles by aerodynamic diameter. The measurement period was between April to May (warm season) and June to July (cold season) in 2017 and 2018. A total of sixty samples were collected during the warm season and fifty-eight samples during the cold season, with two measurements taken in each home per season, spaced one week apart. All the filters were analyzed at the University of North-West. All sample handling and processing following receipt of samples were performed in a controlled laboratory with an average temperature of 22 ± 2 °C and a relative humidity of 40 ± 2% [30]. Samples were weighed using a six-digit electronic microbalance (XP26 DeltaRange Microbalance: Mettler-Toledo AG, Greifensee, CH), which has an internal anti-static source and was treated with an ionizing blower to eliminate static electricity. Weights of 1 g and 20 g were used to calibrate the microbalance before each weighing session, adhering to the US-EPA Quality Assurance Guidance Method Compendium for PM_2.5 Mass Weighing Laboratory Standard Operating Procedures.

To identify the PM_2.5 sample chemical components, all the filters were analyzed for Al, As, Br, Cd, Cr, Cu, Ni, Fe, Mg, Mn, Pb, Sr, Ti, V and Zn using the PANalytical AxiosmaX wavelength dispersive X-ray fluorescence spectrometer. The spectrometer was controlled by an external PC using the SuperQ software (Malvern PANalytical Ltd., Malvern, UK, 2010). MICROMATTER™ calibration standards were utilized for calibration. These standards are National Institute of Standards and Technology (NIST) traceable reference materials and have a Nucleopore^® polycarbonate aerosol membrane mounted in a 25 mm ring. Each element was calibrated using two standards: a lighter one ranging from 3–8 μg.cm⁻² and a heavier one ranging from 40–60 μg.cm⁻².

2.3. Walkthrough Indoor Assessment

In this study, we utilized a walkthrough questionnaire that had been adapted from a previously validated questionnaire used in Detroit and South Africa [15,16,31]. The instrument included both observations made by the investigator and questions posed to the participant. The walkthrough questionnaire had been previously employed to develop a model for predicting indoor PM_2.5 levels in homes where measurements were not taken (n = 500) [15,16], and was demonstrated to be effective for predicting indoor PM_2.5 levels, especially in resource-constrained settings. In this study, we employed the walkthrough questionnaire to identify the determinants of PM_2.5-bound heavy metals. It gathered information on housing characteristics such as dwelling type (including wall, floor, and roof type, house age, presence of windows and cross ventilation), combustion-related activities like burning incense and candles, smoking habits, and details about pollutant sources around the home. Cross ventilation denotes the natural airflow that allows wind to enter from one side of a room and exit through the other. Data collection included the assessment of all rooms in each home.

2.4. Data Analysis

We used Microsoft Excel to clean and check data for quality. The data was subsequently exported to Stata IC version 17 (StataCorp, College Station, TX, USA) for further analysis. We summarized the data using mean (SD), median (range), scatterplots, and histograms to summarize the levels of heavy metals. We assessed contamination levels using the mean values of indoor PM_2.5 heavy metals. Three indicators for assessing the level of contamination were used in this study (Equations (1)–(3)), and aluminum (Al) was used as a reference element for the crust [32]. The enrichment factor was assessed using Equation (1). Where C_x sample is the value of the metal levels measured in the current study, and C_x background refers to the background value. Values of 0.5 ≤ EF ≤ 1.5 represent natural sources, and values of EF > 1.5 represent human activities [33]. Contamination factor (Equation (2)) was measured by dividing the indoor PM_2.5-bound metal levels measured in the current study (Cn), and the background levels (Bn) emanating from the crust. Values of CF < 1; 1≤ CF < 3; 3 ≤ CF ≤6 and CF > 6 suggest low pollution, moderate pollution, considerable pollution, and very high pollution, respectively. The contamination factors for all metals measured using Equation (2) were used to assess pollution load index (PLI) (Equation (3)) [34].

$$EF={\displaystyle \frac{Cx\left(measured\right)}{Cx\left(background\right)}}$$

(1)

$$CF={\displaystyle \frac{Cn}{Bn}}$$

(2)

PLI = (CF₁ × CF₂ × CF₃… × CFn)^1/n

(3)

We used multivariable linear regression to determine the predictors of indoor PM_2.5 heavy metals identified through the walkthrough questionnaire. In all our models, non-significant variables were eliminated using forward stepwise regression with a significance threshold of p < 0.05. Pearson correlation was employed to evaluate the relationship between the different indoor PM_2.5-bound heavy metals, and we also applied principal component analysis to extract underlying patterns. Prior to applying PCA, we conducted the Bartlett’s test and Kaiser–Meyer–Olkin (KMO) measure to test the appropriateness of PCA [35]. The varimax rotation technique was used to determine the relationship between the extracted components. This method facilitates the interpretation of the components. To determine which components to retain, we applied the eigenvalue criterion. This method ensures that components that explain a substantial amount of variance in the data are considered. Components with eigenvalues of 1 or greater were retained.

To evaluate the potential health risk for children, we assessed average daily intake (ADI) through ingestion (Equation (4)), inhalation (Equation (5)), and dermal contact (Equation (6)). Exposure was evaluated based on the guidelines provided by the USEPA [36]. The following equations were used to assess ADI in milligrams per kilogram per day (mg/kg per day).

Table 1. Exposure parameters for calculating children’s health risk [36].

Parameter
Ingestion rate (mg/day)	IngR	200
Exposure rate (365 days/year)	EF	365
Duration of exposure (days)	ED	6
Childs body weight (kg)	BW	15
Mean duration for carcinogens (days)	AT	25,550
Mean duration for non-carcinogens (days)	AT	2190
Inhalation rate (mg/cm2)	Irair	10
Dermal surface area (cm2)	SA	2100
Soil adherence factor (mg/cm2)	AF	0.2
Dermal absorption factor	ABS	0.1
Dermal exposure ratio	FE	0.61
Particulate emission factor (m3/kg)	PEF	1,300,000,000
Conversion factor (kg/mg−1)	CF	0.000001

shows the exposure parameters used for calculating health risks for children.

$$ADI(ingestion)={\displaystyle \frac{C\times IngR\times EF\times ED}{BW\times AT}}$$

(4)

$$ADI(inhalation)={\displaystyle \frac{Cs\times InR\times EF\times ED}{BW\times AT\times PEF}}$$

(5)

$$ADI(dermalcontact)={\displaystyle \frac{Cs\times InR\times EF\times ECs\times SA\times FE\times AF\times ABS\times EF\times ED\times CFD}{BW\times AT}}$$

(6)

To evaluate the potential non-carcinogenic risk of heavy metals to children, we assessed the hazard quotient (HQ) using the ratio of ADI from the three exposure pathways measured in this study divided by the reference levels. The reference levels of ADI for ingestion, inhalation, and dermal contact were taken from USEPA. For ingestion, the ADI reference levels were as follows (in mg/kg per day): 3.0 × 10⁻⁴ (As), 5.0 × 10⁻⁴ (Cd), 3 × 10⁻⁴ (Cr), 4.0 × 10⁻² (Cu), 7.0 × 10⁻³ (Fe), 1.4 × 10⁻³ (Mn), 20.0 × 10⁻³ (Ni), 3.6 × 10⁻³ (Pb), 3.0 × 10⁻¹ (Zn) [36,37]. The ADI reference levels for inhalation were as follows: 3.0 × 10⁻⁴ (As), 10.0 × 10⁻⁶ (Cd), 1.0 × 10⁻⁴ (Cr), 4.0 × 10⁻³ (Cu), 2.2 × 10⁻⁴ (Fe), 50.0 × 10⁻⁶ (Mn), 90.0 × 10⁻⁶ (Ni), 1.4 × 10⁻³ (Pb) and 3.0 × 10⁻¹ (Zn) [36,37]. The ADI reference levels for ingestion were as follows: 3.0 × 10⁻⁴ (As), 1.0 × 10⁻⁶ (Cd), 6.0 × 10⁻⁶ (Cr), 12.0 × 10⁻³ (Cu), 7.0 × 10⁻² (Fe), 1.8 × 10⁻³ (Mn), 5.4 × 10⁻³ (Ni), 1.4 × 10⁻³ (Pb), and 7.5 × 10⁻² (Zn) [37,38].

In addition, we also evaluated the incremental probability of children developing cancer throughout their lifetime by measuring the health index (HI). HI is measured using the sum of HQ of the three exposure pathways. HI less than 1 means low risk, and HI more than 1 indicates high risk. The cancer risk associated with lifetime exposure (LCR) was assessed only for As and Pb because data was not available for other metals. The cancer slope factor (CSF) values used in this study were 1.5 × 10⁰ for As and 8.5 × 10⁻³ (ingesting), 4.2 × 10⁻² (inhalation), and 4.2 × 10⁻² (dermal contact) for Pb [36]. Carcinogenic risks with values below 1 × 10^–6 were considered negligible, while those above 1 × 10^–4 were deemed harmful to humans, in line with USEPA guidelines. [36]. QGIS version 3.38.1 was used to produce the distribution of pollution load index levels in the study area, while Google Earth Pro version 7.1.8 was utilized for visual illustration.

3. Results

3.1. Summary Statistics of Heavy Metal Levels

Out of the 30 homes assessed, a total of 118 samples were obtained. Descriptive statistical data for the levels of heavy metal in PM filters are shown in

Table 2. Heavy metal concentrations in PM samples.

ng/m3	Al *	As	Br	Cd	Cr	Cu	Ni	Fe	Mg	Mn	Pb	Sr	Ti	V	Zn
Min	18	15	3	38	4	48	4	18	29	1	49	1.8	1	1	78
Max	836	315	123	261	641	859	163	1456	1169	39	854	72	37	46	1596
Mean (SD)	214 (167)	271 (52)	37 (26)	140 (76)	197 (164)	792 (142)	98 (32)	513 (361)	384 (212)	16 (9)	723 (138)	22 (21)	17 (9)	12 (9)	1235 (247)
Median	166	277	32	131	179	810	92	467	367	15	715	12	15	10	1246
%>Sample mean	-	60%%	43%	50%	43%	90%	43%	43%	47%	43%	63%	41%	40%	40%	57%
%>background value	-	77%	0%	50%	40%	97%	18%	87%	90%	0%	97%	0%	0%	0%	100%
WHO annual limits [39]		6		5			20				500

* Aluminum (Al) was used as a reference element for the crust.

. The mean (SD) levels for As, Br, Cd, Cr, Cu, Fe, Mg, Mn, Ni, Pb, Sr, Ti, V, and Zn were 271 ng/m³ (±52), 37 (±26), 140 ng/m³ (±76), 197 ng/m³ (±164), 792 ng/m³ (±142), 98 (±32), 513 ng/m³ (±361), 384 ng/m³ (±212), 16 ng/m³ (±9), 723 ng/m³ (±138), 22 (±21), 17 ng/m³ (±9),12 ng/m³ (±9), and 1235 (±247) ng/m³ respectively. Levels of As (77%), Cd (50%), Cr (40%), Cu (97%), Ni (18%), Fe (87%), Mg (90%), Pb (97%), and Zn (100%) were above the background value.

3.2. The Use of Contamination Indices to Assess the Level of Contamination in the Homes

The average EF estimated for As, Br, Cr, Cu, Fe, Mg, Mn, Pb, Ni, Sr, Ti, V, and Zn were 1.00, 0.03, 0.20, 0.73, 3.87, 2.37, 1.77, 0.00, 3.40, 0.30, 0.00, 0.00, 0.00, and 5.77, respectively. According to the CF average values, the homes were highly contaminated with Zn (8.20), considerably contaminated with Pb (4.74), Cu (5.21), and Fe (3.26), moderately contaminated with As (1.78), Cr (1.33), and Mg (2.14), and minimally contaminated with Ti (0.10), V (0.08), Mn (0.11), Ni (0.64), Br (0.21), Sr (0.13), and Cd (0.33) (Figure 2).

According to PLI (Figure 3), 28 of the 30 homes were contaminated with heavy metals. The PLI values ranged from 8 × 10⁻² to 6.6 × 10⁶.

3.3. The Relationship between Pollutant Sources and PM_2.5 Heavy Metals

The multivariate linear regression results are shown in

Table 3. Multivariable linear regression models developed using information obtained via the walkthrough indoor questionnaire.

	As		Cr		Cu		Mn		Ni		Pb		Ti		V		Zn
	β	95% CI	β	95% CI	β	95% CI	β	95% CI	β	95% CI	β	95% CI	β	95% CI	β	95% CI	β	95% CI
Constant	5.38	4.87–5.89	4.34	3.27–5.50			3.00	2.14–3.87	4.37	3.75–5.00	6.37	5.90–6.84	2.37	1.60–3.12	2.89	1.98–3.81	6.91	6.44–7.39
Proximity to industry											0.53	−0.01–1.07
Use of incense/ candles													1.12	0.04–2.20
Presence of windows	1.02	−0.01–2.05			1.03	0.05–2.01					1.00	0.02–1.92
Cross Ventilation	−0.91	−1.79–(−0.02)			−0.88	−1.72–(−0.03)
Home heating							−0.91	−1.66– (0.18)							−0.78	−1.56–0.01
Proximity to pollutant generating activities	0.93	0.14–1.72	1.71	−0.02–0.81	0.98	0.23–1.73	1.51	0.17–2.85	1.30	0.32–2.27	0.86	0.13–1.60			1.59	0.17–3.01	0.99	0.24–1.73
Wall type: corrugated iron/wood)	−1.06	−1.62–(−0.49)			−1.07	−1.60–(−0.53)			−1.22	−1.91–(−0.52)	−0.98	−1.50–(−0.46)					−0.99	−1.52–(−0.46)
Age of the house: ≤25 years	−0.33	−0.68–0.01			−0.33	−0.67–(−0.00)					−0.32	−0.64–(−0.00)			−0.75	−1.39–(−0.11)	−0.36	−0.68–(−0.03)

The forward stepwise regression technique was employed to eliminate non-significant variables, using a significance threshold of p < 0.05. The variables that remained significant at a 95% confidence interval (95% CI) are denoted by bolded β coefficients.

. Proximity to pollutant generating activities was associated with an increase in PM_2.5-bound As, Cr, Cu,

Mn, Ni, Pb, V, and Zn. Houses ≤25 years old were linked to a decrease in PM_2.5-bound As, Cu, Pb, and V. Compared to houses built with brick walls, houses built with corrugated iron/wood were linked to a decrease in PM_2.5 As, Cu, Ni, Pb, and Zn. The presence of windows was linked to an increase in PM_2.5 As, Cu, and Pb, while cross ventilation resulted in a significant reduction in As and Cu levels. A significant reduction in PM_2.5 Mn was observed in households that did not use heating energy sources.

3.4. Analysis of Correlations among Heavy Metals in PM_2.5

We observed significant positive correlations at p < 0.001 between As and Fe, As and Ni, As and Pb, Cr and Cu, Cr and Fe, Cr and Ni, Cu and Mn, Cu and Ni, Cu and Pb, Cu and Zn, Fe and Ni, Mg and Mn, Mg and Pb, and Mg and Zn; at p < 0.01 for Cd and Zn, Cu and Mg, Fe and Mn, Mn and Ni, Mn and Ti, Mn and V, Mn and Zn; and at p < 0.05 for Br and V, Cd and Pb, Cr and Mn, Ni and V, and Ti and Zn (

Table 4. Matrix of correlation coefficients among heavy metals.

	As	Br	Cd	Cr	Cu	Fe	Mg	Mn	Ni	Pb	Sr	Ti	V	Zn
As	1
Br	0.29	1
Cd	0.30	0.33	1
Cr	*** 0.08	0.14	0.22	1
Cu	0.92	0.27	0.22	*** 0.30	1
Fe	*** 0.34	0.18	0.26	*** 0.98	0.34	1
Mg	0.31	* 0.20	0.04	0.40	** 0.36	0.49	1
Mn	0.34	* 0.20	0.35	* 0.16	*** 0.38	** 0.20	*** 0.31	1
Ni	*** 0.40	0.23	0.20	*** 0.90	*** 0.38	*** 0.88	0.46	** 0.30	1
Pb	*** 0.92	0.40	* 0.50	0.14	*** 0.91	0.20	*** 0.35	0.42	0.46	1
Sr	0.24	−0.03	0.08	−0.01	0.16	0.01	0.26	0.32	0.11	0.19	1
Ti	0.39	0.10	0.08	−0.00	0.33	0.16	*** 0.38	** 0.09	0.02	0.39	0.29	1
V	0.16	* −0.18	−0.30	0.11	0.24	0.05	−0.08	** 0.29	* 0.24	0.17	−0.02	−0.25	1
Zn	0.83	0.33	** 0.40	0.44	*** 0.89	0.48	*** 0.38	** 0.39	0.68	0.85	0.18	* 0.31	0.25	1

*** indicates correlation is significant at the 0.001 level. ** indicates significance at the 0.01 level. * indicates significance at the 0.05 level.

).

3.5. Identification of the Sources of Heavy Metals in Indoor PM_2.5 Using PCA

Table 5. Principal components analysis for heavy metals in PM_2.5.

Variable	PC1: Household Sources, Traffic Emissions and Industries	PC2: Industries and Natural Sources (i.e., Natural Soil and Marine Aerosols)	PC3: Natural Sources (i.e., Marine Soil and Natural Soil)	PC4: Natural Sources (i.e., Natural Soil)
As	0.488
Br				−0.38
Cd			0.31
Cr		0.569
Cu	0.461
Fe		0.556
Mg			0.351
Mn			0.385	0.310
Ni		0.464
Pb	0.487
Sr			0.666
Ti			0.321	−0.381
V				0.701
Zn	0.419
Eigenvalue	5.5	2.3	1.6	1.2
% of variance	39.6	16.2	11.1	8.0
Cumulative %	39.6	55.8	66.9	75.4

shows the results of a principal component analysis (PCA) conducted to identify common sources of heavy metals in indoor PM_2.5. Four components with eigenvalues greater than 1, accounting for 75.4% of the overall variance, were retained. PC1 had high loadings for As, Cu, Pb, and Zn, explaining 39.6% of the overall variance; PC2 had high loadings for Cr, Fe, and Ni, explaining 16.2% of the overall variance; PC3 had high loadings for Cd, Mg, Mn, Sr, and Ti, explaining 11.1% of the overall variance; and PC4 had high loadings for Br, Ti, and V, explaining 8% of the overall variance. The themes for the 4 components are as follows: PC1 (household characteristics (i.e., lead-based paints, house age, burning incense and candles, etc.), PC2 (industries and natural sources), PC3 (natural sources), and PC4 (natural sources).

3.6. Evaluation of the Health Risk for Children

The HI for all metals was >1, indicating a high risk of adverse health outcomes in children. The carcinogenic risk exceeded the maximum acceptable level (

Table 6. The assessment of the health risk.

	ADIing	ADIinh	ADIderm	HQing	HQinh	HQderm	HI	LCR
As	3.5 × 10–3	1.3 × 10−7	4.4 × 10−4	1.155 × 101	4.4 × 10−4	1.5 × 100	1.3 × 101	2.0 × 101
Br	-	-	-	-	-	-	-	-
Cd	1.8 × 10−3	6.9 × 10−8	2.3 × 10−4	3.6 × 100	6.9 × 10−3	2.3 × 101	2.7 × 101	-
Cr	2.5 × 10−3	9.7 × 10−8	3.2 × 10−4	8.4 × 100	9.7 × 10−4	5.4 × 100	1.4 × 101	-
Cu	1.0 × 10−2	3.9 × 10−7	1.3 × 10−3	2.5 × 10−1	9.7 × 10−5	1.1 × 10−1	3.6 × 10−1	-
Fe	6.7 × 10−3	2.5 × 10−7	8.4 × 10−4	9.4 × 10−1	1.1 × 10−3	1.2 × 10−2	9.5 × 10−1	-
Mg	-	-	-	-	-	-	-	-
Mn	2.1 × 10−4	7.9 × 10−9	2.6 × 10−5	1.5 × 10−1	5.6 × 10−6	1.95 × 10−2	1.6 × 101	-
Ni	1.3 × 10−3	4.8 × 10−8	1.6 × 10−4	6.3 × 10−2	5.4 × 10−4	3.0 × 10−2	9.3 × 10−2	-
Pb	9.2 × 10−3	3.6 × 10−7	1.2 × 10−3	2.6 × 100	2.5 × 10−4	8.5 × 10−1	3.8 × 101	3.2 × 10−1
Sr	-	-	-	-	-	-	-	-
Ti	-	-	-	-	-	-	-	-
V	-	-	-	-	-	-	-	-
Zn	1.6 × 10−2	6.1 × 10−7	2.0 × 10−3	5.3 × 10−2	2.0 × 10−6	2.7 × 10−2	8.0 × 10−2

).

4. Discussion

The study provides evidence of the contribution of heavy metals to indoor PM_2.5 toxicity. Indoor PM_2.5 As, Cd, Pb, and Ni concentrations were above the WHO recommended values for ambient PM_2.5 heavy metals. The pollution indices indicate that contamination does exist, with linear regression analysis, correlation analysis, and principal component analysis pointing to the potential influence of household characteristics, traffic emissions, industrial operations, and natural sources. HI for assessed metals was >1, indicating high risk of adverse health outcomes. The carcinogenic risk exceeded the maximum acceptable level. These findings are in line with studies assessing PM_2.5 heavy metals [19,21,25].

The concentrations of PM_2.5 heavy metals from homes located in the north and south of Durban were within the range reported in studies conducted elsewhere [20,25]. For instance, the average PM_2.5: Cd (140 ng/m³), Cu (792 ng/m³), and Ni (98 ng/m³) concentrations were in the range reported in an iron and steel scrap metal of Nigeria: Cd (270 ng/m³), Cu (100 ng/m³), and Ni (110 ng/m³) [25]. The average Mn (17 ng/m³) concentrations reported in Durban previously [40] were in the same range as that of the current study (16 ng/m³), while the Pb (723 ng/m³) levels reported here were much higher than the latter study: Pb (77 ng/m³). In a study conducted in Iran, the range of the concentration levels of PM_2.5-bound heavy metals were Cd (0–40 ng/m³), Cr (10–50 ng/m³), Fe (410–7590 ng/m³), Ni (20–80 ng/m³), Pb (0–40 ng/m³), and V (0–20 ng/m³) [41]. In the current study, Zn was the most abundant (1235 ng/m³), followed by Cu (792 ng/m³), Pb (723 ng/m³), Fe (513 ng/m³), Mg (384 ng/m³), and As (271 ng/m³). According to WHO ambient air quality standards for As (6 ng/m³), Pb (500 ng/m³), Ni (6.6 ng/m³), and Cd (5 ng/m³) [39], the communities in this study area are exposed to elevated levels, with a consequential risk for adverse health outcomes.

The pollution indices are effective tools for communicating risk. Using the two categories of enrichment, the PM_2.5 heavy metals that were especially concerning were Cu, Fe, Mg, Pb, and Zn, suggesting contribution by anthropogenic sources. In this study, the south of Durban is recognized as the most industrialized area in Durban and is recognized for its petroleum, pulp, and paper manufacturing industries, with some of the communities situated as close as 100 m from the industries [42]. This is further compounded by pollution from the major highways. Studies conducted in the south of Durban report implications for public health, animal health, as well as the environment [40,42,43]. The highest CF values were observed in the south of Durban: As (2.17), Br (0.23), Cd (0.48), Cu (6.49), Cr (1.99), Fe (4.71), Mg (2.54), Mn (0.15), Ni (0.86), Pb (5.92), Sr (0.15), Ti (0.11), V (0.13), and Zn (10.57), compared to the north of Durban: As (1.39), Br (0.19), Cd (0.20), Cu (3.93), Cr (0.67), Fe (1.80), Mg (1.73), Mn (0.06), Ni (0.42), Pb (3.56), Sr (0.11), Ti (0.08), V (0.02), and Zn (5.82). Notably, PLI for 28 homes (15 in the south and 13 in the north) was >1, indicating a significant decline in the indoor air quality of the homes. Though there is limited literature on indoor PM_2.5 heavy metals, the contamination of the environment by heavy metals is evident in other studies conducted in polluted urban settings [44]. In a study conducted by MalAmiri et al. [45], the most abundant heavy metals were Pb, Ni, Cr, and Zn [45]. In another study conducted by Asvad et al. [46], the pollution indices indicated moderate contamination from heavy metals like As, Cd, and, to a lesser extent, Zn and Pb in indoor samples. The enrichment factor values in this study were in the range reported in another study that assessed contamination on indoor dust samples: As (5.76 vs. 5.43), Pb (3.41 vs. 3.61), and Zn (5.76 vs. 4.77) [46].

Certain meteorological factors, including precipitation and wind speed, may influence the levels of particles in the atmosphere. For example, frequent rainfall may reduce the concentrations of PM in the atmosphere [47]. Previously, precipitation was reported to result in a reduction in indoor PM_2.5 levels in a study conducted in Durban [16]. In the current study, except for PM_2.5 Ni, Fe, Mn, and Sr, the levels of heavy metals were higher in the colder season as opposed to the warmer season, though this difference was minimal. Indoor household characteristics (wall material, house age, window presence, cross ventilation), occupant activities (burning incense or candles), and outdoor sources (proximity to industrial operations and other pollution-generating activities) emerged as significant predictors of indoor PM_2.5 heavy metals in our multivariate linear regression analysis. Pb is a significant component of industrial emissions [25]. In our linear regression analysis, proximity to an industry was linked to higher levels of PM_2.5 Pb levels. Heavy metals are intentionally used in paint formulations for various purposes, including color development and protection of surfaces from corrosion [48,49]. In this study, households with brick walls had significantly higher levels of As, Cu, Ni, Pb, and Zn.

To gain a clearer understanding of the relationship between the 14 heavy metals, and further identify the sources, we conducted Pearson correlation and PCA. The results were in agreement with the linear regression analysis results. PC1 had high loadings for As, Cu, Pb, and Zn, explaining 39.6% of the total variance. As, Cu, Pb, and Zn are known to result mainly from traffic emissions, combustion of fossil fuels, industrial emissions, building construction (e.g., use of lead-based paints, asbestos roofing, wood preservatives containing As), and resuspension of road dust [20]. The wear and tear on vehicles can also release Cu and Zn [50]. PC2 had high loadings for Cr, Fe, and Ni, explaining 16.2% of the total variance. Cr has both natural and human-made sources. Both Cr and Ni were highly correlated (r = 0.90, p < 0.001), indicating similarity in source. These two metals are constituents of traffic emissions as a result of exhaust emissions [51,52]. PC3 had high loadings for Cd, Mg, Mn, Sr, and Ti, explaining 11.1% of the total variance. According to EF, Cd, Mn, Sr, and Ti had very low values, pointing to natural sources [53]. PC4 had high loadings for Br, Ti, and V, explaining 8% of the total variance. This component also had very low EF values, also pointing to natural sources. According to studies conducted in the study area, the coastal location makes marine aerosols and beach sediments (which contain elements such as Mg and V) highly relevant.

Compared with studies conducted in Iran, China, and South Africa [19,40,41,53], there are indications that the homes of the north and south of Durban are exposed to somewhat elevated levels of heavy metals as a result of certain household features, traffic emissions, industrial activities, and natural sources.

Heavy metals (As, Cd, Cr, Cu, Fe, Mn, Ni, Pb, and Zn) were considered for assessing the health risk as per the USEPA guidelines [36]. Ingestion was the primary route of heavy metal exposure among children. These findings are consistent with studies conducted previously [54,55]. When assessing health risks for children, the HI for As, Cd, Cr, and Pb were >1, indicating that children living in these homes may face significant health risks due to heavy metal concentrations in the respirable fraction of particulate matter. According to a study conducted by Behrooz et al. [56], both children and adults experienced higher non-carcinogenic and carcinogenic risks through the inhalation pathway, however, for the carcinogenic risk, the ingestion route was an important pathway [56]. In the current study, the ingestion route and dermal contact were the major exposure pathways. The HI values in this study were As (13.03), Cd (26.52), Cr (13.77), Cu (0.36), Ni (0.10), Pb (37.57), Mn (0.16), Zn (0.08), and Fe (0.95). The HI values for As are in the range reported in a study conducted in Iran [56], which reported a value of 10.1. However, the HI values for Mn (1.04) and Ni (1.04) were significantly lower than the values reported in the current study. In line with a study conducted in a polluted urban area [57], the carcinogenic risk for As was found to be above the safe limit of 1 × 10⁻⁴. The same was observed for Pb.

In conjunction with research carried out in other parts of South Africa, our study indicates the importance of implementing and, where applicable, developing environmental policies related to the indoor and outdoor environment.

The study had some limitations. For example, the walkthrough indoor assessment used in this study was for assessing the determinants of particulate matter, not heavy metals. It is possible that there were unmeasured sources of heavy metals. However, through using Pearson correlation and PCA analysis, we were able to identify the common sources of the measured and unmeasured sources of heavy metals. Some exposure variables were not used in our multivariate analysis because of the small sample size. For example, with over 90% of homes using electricity for cooking, heating, and lighting, it was not possible to compare different energy sources. In addition, due to the small sample size, the study lacks the capability of quantitatively estimating the contributions of sources using more advanced models such as positive matrix factorization and long-range transport models. Another limitation is that our study only focused on the PM_2.5 size fraction, and that not all PM_2.5 heavy metals were assessed. The toxicity of particulate matter is influenced by factors beyond heavy metal content, including the overall particle size distribution and the presence of other toxic substances. Future research should address these gaps by analyzing content across a broader spectrum of particle sizes, including total dust, to offer a more thorough understanding of the overall toxicity.

5. Conclusions

This study demonstrates evidence of exposure to indoor PM_2.5 heavy metals in people living in low socio-economic communities of Durban. The levels of indoor PM_2.5 Cd, Ni, and Pb were above the WHO ambient air quality standards. The use of multivariate analysis suggests that household characteristics, traffic emissions, industries, and natural sources play a significant role. The carcinogenic and non-carcinogenic risks indicated that children are at an increased risk of developing adverse health outcomes resulting from exposure to indoor PM_2.5 heavy metals. Among the exposure pathways assessed, it was found that the ingestion pathway contributed the most to the cancer risk associated with heavy metal exposure in this study. This study highlights the need to strengthen environmental policies related to indoor air quality in resource-constrained settings of South Africa. Currently, in South Africa, human settlements are situated near sources of pollution, highlighting a missed opportunity to apply healthier urban planning practices, and there are no policies related to indoor air quality. This research contributes to the current body of evidence concerning exposure to particulate matter-bound heavy metals in communities situated near sources of pollution. Future research should focus on assessing long-term exposure and its correlation with health outcomes, especially among vulnerable groups. This will help protect vulnerable groups from exposure to potentially harmful elements and their related health risks. We hope that this study can inform and strengthen the establishment and enforcement of policies that are centered around human and environmental health.