*2.3. Questionnaire*

A questionnaire was used to collect data about the socio-demographic characteristics of participating lactating mothers. The questionnaire was adopted from di fferent studies [14,16–20]. It consisted of eight subsections collected from di fferent sources, addressing general information about the lactating mothers and the new-borns, dietary habits of lactating mothers, smoking status, geographic characteristics, pesticides spraying, medical history and supplements intake as well as the drinking water intake.

The questionnaire was translated into Arabic to ensure accuracy of the translation, back translation from Arabic to English was conducted by another translator. Both English versions were the same. The Arabic version of the questionnaire was pilot-tested on a sample of 30 mothers. The questionnaire was administered twice within a period of two weeks between each sitting. To determine test–retest reliability, a paired *t*-test was performed to compare mean scores at T1 (before) and T2 (after two weeks). Pearson's correlation was calculated between T1 and T2, with a consistency interclass correlation (ICC) [21]. The values of ICC were considered excellent if >0.75 [22].

#### *2.4. Sample Collection*

During December 2018 and February 2019, 50 nursing mothers were contacted but only 45 agreed to answer the survey and participated in our study. The 45 nursing mothers that were visited for the survey were contacted again and only 40 of them agreed to donate breast milk by triplicate over three weeks.

Breast milk collection was done in the morning an hour after the previous breastfeeding. Samples (25–40 mL) were collected using a sterile, single-use electrical pump (Chico, Mod.06836, China). A polypropylene pump container was sterilized with 70% alcohol solution to prevent cross contamination between samples. Each sample was placed into a separate 100 mL dark glass (to preserve the samples from light), numbered then transported on ice (at 4 ◦C) to the laboratory, where they were transferred into urine cups, numbered and stored at −20 ◦C until analysis.

#### *2.5. Chemicals and Reagents*

The reagents used were of analytical grade. The antibiotic standards belonged to three di fferent families: sulphonamides (sulfamethazine), tetracyclines (tetracycline, oxytetracycline) and beta-lactam antibiotics (ampicillin); in addition, 161 kinds of pesticide were purchased from Sigma-Aldrich (St. Louis, MO, USA). The standards used present high purity grades (>99%). Individual stock solutions were prepared at 1000 μg/<sup>L</sup> in acetonitrile and stored at −20 ◦C.

The working standard solutions of a concentration of 10 μg/<sup>L</sup> each were prepared as dilution of the stock solution in water/methanol (50:50, *v*/*v*). The working standard solutions were stored at −20 ◦C. HPLC-grade water, HPLC-grade acetonitrile, and magnesium sulphate (MgSO4) were also supplied by Sigma-Aldrich.

#### *2.6. Sample Extraction for Antibiotics Residues Analysis*

Using liquid chromatography–mass spectrometry, a multi-class method was developed for identifying and quantifying three antibiotics belonging to two di fferent chemical classes (tetracyclines and beta-lactams). Afterwards, this method was optimized for the detection of antibiotics in breast milk.

100 μL from 1 % formic acid was added to a 2 mL of breast milk then vortexed for 60 s and kept in the dark for 10 min. Then 500 μL of EDTA (0.5%) was added and shaken vigorously for 1 min. Next, 8 mL TCA (5%) was added then shaken for 10 min before subjecting the tubes to centrifugation at 5000 rpm at 4 ◦C for 10 min. The supernatant obtained was then subjected to filtration through a 0.45 μm polyvinylidene fluoride (PVDF) filter for further liquid chromatography tandem mass spectroscopy (LC-MS/MS) analysis.

#### *2.7. Sample Extraction for Pesticide Residues Analysis: QuEChERS Extraction*

To ensure a quick and easy sample treatment, a modified QuEChERS (quick, easy, cheap, e ffective, rugged and safe) approach, originally developed as a powerful sample preparation, was used to analyse hundreds of pesticides in di fferent kinds of food.

Acetonitrile (10 mL) was added to 10 mL of breast milk and shaken vigorously before adding magnesium sulphate MgSO4 (4 g), sodium chloride NaCl (1 g), sodium dibasic citrate (1 g) and sodium tribasic citrate (0.5 g) and then shaken for 3 min. The tubes were then subjected to centrifugation at 5000 rpm at 4 ◦C for 5 min.

The aqueous supernatant (5 mL) was mixed with dispersive Enhanced Matrix Removal (EMR) dissolved in 5 mL water then shaken vigorously for 5 min. Another centrifugation was conducted at 5000 rpm at 4 ◦C for 5 min.

The aqueous supernatant (5 mL) obtained was transferred in a new Polish EMR Lipid (MgSO4 and NaCl tube, then shaken vigorously for 3 min before being centrifuged at 5000 rpm at 4 ◦C for 5 min. Afterwards, a filtration was made through a 0.22 μm polyvinylidene fluoride (PVDF) filter for further LC-MS/MS and GC/MSMS analysis.

#### *2.8. LC-MS*/*MS Equipment*

LC-MS/MS analyses was conducted using an Agilent 6430 LC/MS (Agilent technologies, Santa Clara, USA) for antibiotic and pesticide residues detection, and LC-MS/MS analyses were performed on LC-NexeraX2 Shimadzu 8045 LC/MS (Kyoto, Japan). The mass spectrometer was operated with an electron spray ionisation (ESI) in multiple reaction monitoring (MRM) mode, ionspray voltage 4 kV, nitrogen for desolvation and dried gas 11 L/min.

The quantification of the three antibiotics families and 161 kinds of pesticide in 120 breast milk samples was done through measuring peak areas in the multiple reaction monitoring chromatogram, and comparing them with the relevant matrix-matched calibration curves. A calibration curve ranging between 1 μg/<sup>L</sup> and 500 μg/<sup>L</sup> was carried out to verify linearity.

The performance of the analytical method was evaluated by checking the identification criteria for the presence of two transitions at the same retention time, the signal to noise ratio ≥ 10, the relative retention time of the analyte within a tolerance of 2.5% and the relative ion intensities ratio within a tolerance defined by the EU commission decision 2002/657/EC. The MRM parameters are shown in Table 1.


**Table 1.** Multiple reactions monitoring (MRM) acquisition condition for each antibiotic used.

#### *2.9. LC*/*MS*/*MS Parameters*

The separation of the pesticide residues was performed using a C18 analytical column (zorbax 2.1 mm inner diameter × 150 mm length, 3.5 μm particle size; Agilent technologies, Santa Clara, CA, USA) and a C18 analytical column (Shim-pack GIST 2.1 mm inner diameter × 100 mm length, 3 μm particle size; Japan) for antibiotic residues separation. The separation of pesticides was accomplished at 40 ◦C. The flow rate and injection volume were 0.4 mL/min and 10 μL, respectively. The mobile phases used were (A) 5 mM ammonium acetate and 0.1% formic acid in water, and (B) 5 mM ammonium acetate in methanol. The gradient elution program started with 100% A for 5 min, then decreased to 50% for 13 min, then to 0% for 2 min and returned back to the initial conditions for 5 min. The final run time of the method was 29 min.

The separation of sulfamides, tetracyclines and beta-lactam was accomplished at 40 ◦C. The flow rate and injection volume were 0.3 mL/min and 10 μL, respectively. The mobile phases used were (A) 5 mM ammonium acetate and 0.1% formic acid in water, and (B) 5 mM ammonium acetate in methanol. The gradient elution program was as follow: B (40%) (2 min), B (90%) (9 min), B (10%) (9 min and 1 s), the final run time of the method was 13 min.

#### *2.10. GC*/*MSMS Parameters*

GC-MS/MS analyses were carried out with an Agilent 7890A GC equipped with 7693 Agilent auto-sampler and 7000C triple quadrupole GC/MS system. The separation was performed on a HP-5MS Agilent technologies (Santa Clara, CA, USA), 0.25 mm × 30 m, 0.25 μm, and helium (purity 99.996%) was used as a carrier gas at a constant pressure of 11 psi. The inlet temperature was set at 250 ◦C; the mode of inlet was splitless; the injection volume was 1 μL. The column temperature program was as follows: the initial temperature was maintained at 70 ◦C for 1 min, increased to 160 ◦C at a rate of 50 ◦C/min, raised to 200 ◦C at 2 ◦C/min, and then at 16 ◦C/min up to 280 ◦C, and held there for 7.2 min. The total run time was 35 min. The mass spectrometer was operated with an electron impact (EI) source in multiple reaction monitoring (MRM) mode. The electron energy was 70 eV, and the source, transfer line and quadrupole temperatures were set at 280 ◦C and 150 ◦C, respectively. In order to prevent instrument damage, the solvent delay was set at 4.5 min.

## *2.11. Recovery Test*

In-house validation was performed by fortifying the blank matrix at a level of 50 μg/L. The extraction was performed by the methods described in the Section 2.6 and 2.7. The spiked and blank samples were then analysed by LCMS/MS. Recovery was calculated by comparing the analysed concentrations with spiked concentrations.

#### **3. Results and Discussion**

#### *3.1. Validation of the Arabic Version of the Questionnaire*

The results of test–retest reliability are presented in Table 2. The paired *t*-test showed that the mean scores of the Arabic questionnaire did not vary significantly between T1 and T2. A statistically significant inter-item correlation, *p* < 0.05, was noted between T1 and T2. An excellent consistency was also noted, with an ICC varying between 0.998 and 1.00 [22].


**Table 2.** Mean and standard deviations, Pearson's correlation coefficients and intraclass correlation for the scores of the Arabic and Arabic versions (*n* = 30).

#### *3.2. Survey Results*

The results of the questionnaire filled by 40 Syrian refugee nursing mothers is presented in Table 3. All the participants had an age below 30 years. The number of years of residency in North Lebanon camps for all lactating mothers was between one year and five years. All the mothers (100%) had an educational degree below secondary and they were not employed.

While assessing the smoking habits, it was shown that 100% of nursing mothers had never smoked before or after pregnancy since Syrian culture inhibits women from smoking, but they were randomly exposed to second hand smoking from their husbands. Lactating mothers were also asked about their supplement intake, 100% indicated that they were taking vitamins and iron supplements during their pregnancy period, however they stopped consuming these supplements after their child's birth in order to save money.


**Table 3.** Survey results of participating mothers.

**Table 3.** *Cont*.



**Table 3.** *Cont*.

When questioned on their dietary habits, it was clearly noticed that 100% of the participants never consumed fish and sea food due to poverty. The consumption of meat and poultry was also negligible for the same reason. With respect to their location of residence, 100% were living near a waste disposal site and cultivation activity since the camp was located in an agricultural area.

#### *3.3. Assessment of Antibiotic Residues in Breast Milk*

#### 3.3.1. LC-MS/MS Method Performance for Antibiotic Residues

An in-house method was verified as per the criteria specified in EU commission decision 2002/675/EC to quantify antibiotics residues. The validation parameters were fixed by spiking blank powdered milk at a level of 50 ng/g. The measured parameters were specificity, linear range, repeatability, reproducibility, accuracy and limit of quantification (LOQ).

The antibiotic residue chromatograms of the reference standards used as well as the calibration curves are shown in Figure 1. The calibration curves were created from six matrix-calibration standards which were injected in each batch in the range of 1 to 50 ng/g. The calibration curves showed good linearity, characterized by a high correlation coefficient (*R*<sup>2</sup> > 0.99).

The precision of the method was determined using the spiked standards at the 50 μg/<sup>L</sup> level. The results for repeatability ranged from 5 and 18.8% (Table 4). The limit of quantification is considered as the lowest quantified level with S/N ≥ 10 in presence of the two transitions at the same retention time. The LOQ was calculated to be 2 μg/kg for all tested antibiotics. The mean recoveries of the

residues for the spiked samples ranged between 97% and 108% (Table 4). These values presented high mean recoveries within the acceptable range (80%–120%) except for ampicillin, which presented a low mean recovery, although it was still higher than 50%. These values were within the acceptable ranges (50%–120%) recommended by AOAC, 2002 [23].

**Figure 1.** Calibration curves for (**a**) oxytetracycline, (**b**) tetracycline, (**c**) ampicillin and (**d**) sulfamethazine 1, 5, 10, 20, 30, 40 and 50 μg/L.


**Table 4.** Results of in-house verification of the LC-MS/MS method for the antibiotics considered in this study, with the standard deviation (STDEV), relative standard deviation of repeatability (RSD) and limits of quantification (LOQ).

3.3.2. Occurrence of Antibiotic Residues in Breast Milk Samples

The method developed was applied to the determination of four antibiotics from three di fferent chemical classes (sulphonamides, tetracyclines, and beta-lactams) in 120 breast milk samples collected from 40 lactating mothers residing in Lebanon's camp. In order to validate the results, an internal quality control was carried out on every batch of samples. Moreover, the retention time, quantification and confirmation transitions and relative ion intensities of the detected ion of the human milk samples were compared to those of the corresponding calibration standards in the same batch to identify the detected analytes.

While assessing the antibiotic residues in the breast milk samples (Table 5), the results showed that only 3.33% of the samples were contaminated with antibiotic residues. It is to be noted that none of the samples contained more than one antibiotic residues. The antibiotic residue contaminating the four samples was oxytetracycline. Of the samples, 96.67% presented a total absence of antibiotic residues. This result is well correlated with the survey, that shows that all the mothers did not receive antibiotics during pregnancy nor lactation period. Additionally, the survey results showed that 100% of the nursing mothers do not consume meat and chicken, that are known to be sources of antibiotic residues, due to the misuse of antibiotics in farms [10,24].


**Table 5.** Occurrence of sulphonamides, tetracyclines, and β-lactams in the 120 breast milk samples.

\* mean value of four contaminated samples.

3.3.3. Mean Concentrations of Antibiotic Residues for the Di fferent Families in Breast Milk Samples

Table 5 represents the summary of multi-antibiotic residues occurrence in 120 human milk samples. The assessment of the sulphonamides, beta-lactam families and tetracyclines showed that three antibiotic residues were not detected in all milk samples, i.e., tetracyclines, sulfamethazine and ampicillin. However, as mentioned before, oxytetracycline was detected in four samples (3.3%). These four samples belong to three mothers; one of them reported one antibiotic residue concentration between the LOQ (= 2 μg/L) and LOD (= LOQ/3 = 2/3 = 0.66μg/L) followed by a second concentration higher than the LOQ in the week after. The other two samples exceeded the LOQ.

Hence, the percentage of positive samples (>LOQ = 2 μg/L) is 2.5%, since the fourth sample presented a mean of 0.64 μg/<sup>L</sup> which is lower than LOQ and LOD.

The mother that reported the highest value of oxytetracycline in her breast milk (6 μg/L) was the only mother that consumed milk on a daily basis. The mean value of oxytetracycline (5.04 μg/L) is 2.5 times higher than the LOQ (2 μg/L). The low oxytetracycline concentration in only three breast milk samples is derived from the nursing mother's food since they did not use any kind of antibiotics after pregnancy. However antibiotic residues can be found in dairy products since antibiotics are usually used in farms of dairy cattle for the treatment of diseases such as mastitis [25]. Probably the source of oxytetracycline is from nursing mothers' dairy product intake because our survey findings indicate the contaminated breast milk samples belong to 15% of mothers consuming dairy products twice per week. The presence of oxytetracycline in dairy products, could be due to the misuse of this antibiotic in farms and lack of abidance to the recommended withdrawal times [26].

#### *3.4. Assessment of Pesticide Residues in Breast Milk*

#### 3.4.1. LC-MS/MS Method Performance for Pesticide Residuess

The pesticide residue chromatograms of the reference standards used as well as the calibration curves are shown in Figure 2. The calibration curves were created from nine matrix-calibration standards which were injected in each batch in the range from 1 to 500 ng/mL. The calibration curves showed good linearity, characterized by a high correlation coe fficient ( *R*<sup>2</sup> > 0.99).

**Figure 2.** LC standard calibration curves for (**a**) lufeneron, (**b**) methamidophos and (**c**) chlorpyriphos at 5, 10, 20, 30, 50, 100, 200, 500 μg/L.

The precision of the method was determined using two spiked levels of 0.05 and 0.1 mg/kg. The limit of quantification was considered as the lowest quantified level with S/N ≥ 3 in presence of the two transitions at the same retention time. The LOQ was calculated to be 5 μg/kg for all tested pesticides. The mean recoveries of the residues for the spiked samples was between the acceptable range (80%–120%).

#### 3.4.2. Occurrence of Pesticide Residues in Breast Milk Samples

While assessing the pesticide residues in the breast milk samples, the results showed that only 5% of the samples were contaminated with pesticide residues. The samples 95% were non-contaminated. It is to be noted that none of the samples contained more than one pesticide residue.

#### 3.4.3. Mean Concentrations of Pesticide Residues in Breast Milk Samples

Table 6 represents the summary of multi-pesticide residues occurrence in 120 human milk samples. The assessment of the 161 pesticides showed that only three pesticide residues were detected in six samples, lufeneron, methamidophos and chlorpyriphos. Hence, the percentage of positive samples (>LOQ = 5 μg/L) was 4.16%. Only four samples were contaminated with lufeneron, the fifth sample was contaminated with methamidophos and the final sample was contaminated with chlorpyriphos with a concentration of 12.32 μg/L. The recovery tests results were at one level (10 μg/L) for methamidophos (86.5%), chlorpyriphos (103.6%) and lufeneron (105.2%).


**Table 6.** Occurrence in the 120 breast milk samples of lufeneron, methamidophos and chlorpyriphos.

\* mean values of six contaminated samples.

While assessing the pesticide residues in breast milk, the percentage of positive samples was only 5% in which methamidophos presented the highest concentration (13.1927 μg/L) in one sample only. The only mean value that exceeded the LOQ (5 μg/L) was reported for lufeneron pesticide (5.874 μg/L). The present results can be explained by nursing mother exposure to pesticides from fruit and vegetable intake [27], since the percentage of mothers consuming fruit and vegetables twice per week were 55% and 35% respectively, to the contrary of other studies conducted in several countries where the percentage of pesticide contamination in breast milk is high [14].
