Validity of an ion selective electrode for measuring human milk sodium and potassium ion levels at point-of-care in lactating mothers with inflammatory breast conditions

Abstract

Background

Elevated human milk sodium (Na+) levels and Na+ to potassium (K+) ratios are commonly used in research to define breast inflammation in lactating mothers. Portable Na+ and K+ ion selective electrode probes (ISEPs) allow for immediate point-of-care testing by clinicians, potentially accelerating diagnosis, and treatment of inflammatory conditions of the lactating breast (ICLB). We determined validity of ISEPs compared to the reference method inductively coupled plasma – optical emission spectrometry (ICP-OES) in mothers with ICLB, and acceptability of ISEP point-of-care testing of mothers tested with clinical ICLB symptoms.

Methods

Human milk samples were collected from 43 mothers with ICLB, between December 2021 and September 2022. Mothers were recruited from private physiotherapy practices, a public women’s hospital, and the local community, in Melbourne, Australia. Human milk Na+, K+ and Na+:K+ ratio levels were tested at point-of-care (ISEPs), and later, in the laboratory (ICP-OES). Adjusted limits of agreement with Bland-Altman plots compared the two measurement methods, with rank linear mixed effects models establishing their relationship. Mother’s ISEP acceptability was assessed via two survey questions (0–10 numerical rating scale (NRS) and open text response) and analysed via descriptive statistics and thematic analysis.

Results

Adjusted limits of agreement (lower limit mean (95% CI) to upper limit mean (95% CI)) between the two measurement methods were: -6.12 (-7.75, -4.49) to 6.12 (4.49, 7.75) mM for Na+; 7.37 (5.82, 9.47) to 25.6 (23.5, 27.7) mM for K+; and -0.82 (-0.85, -0.79) to 0.80 (0.77, 0.82) mM for Na+:K+ ratio. For Na+:K+ ratio, 100% of values fell within the limits of agreement. For Na+ and Na+:K+ ratio, ISEPs and ICP-OES shared a substantial amount of variability (Na+: conditional R² = 0.87, p < 0.05; Na+:K+ ratio: conditional R² = 0.94, p < 0.05). Mother’s acceptance of ISEP testing was high with a median (Q1, Q3) NRS score of 10 (10, 10). The most frequent theme was that the testing was ‘quick and easy, and unproblematic’ (n = 30).

Conclusions

Point-of-care human milk ISEP Na+:K+ ratio measurement in mothers with ICLB is valid and is rated as a highly acceptable clinical assessment tool by mothers with ICLB.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13006-025-00764-y.

Background

Inflammatory conditions of the lactating breast (ICLB), such as mastitis, can affect more than 25% of lactating mothers, most commonly in the early postpartum period [1]. These conditions are a reported reason for ceasing breastfeeding early and can have considerable impact on the health and wellbeing of mothers and babies [2, 3]. Prompt diagnosis is important to support ongoing breastfeeding and reduce the risk of recurrences, breast abscess or hospitalisation [3]. Health care professionals, for example, doctors, international board-certified lactation consultants, midwives and physiotherapists, manage mothers with ICLB internationally [3–6].

Health care professionals managing mothers with ICLB currently lack validated outcome measures to guide timely and effective clinical decision making [7, 8]. Partial validation of an ICLB-specific patient-reported outcome measure has been completed [9]; however, an objective tool that assists clinicians at point-of-care to measure severity of inflammatory breast symptoms is required. Elevated human milk sodium (Na+) concentration and Na+ to potassium (Na+:K+) ratio are used as markers of breast inflammation in addition to delayed secretory activation, where increases reflect incomplete closure of the mammary epithelial cell tight junctions early postpartum [10, 11]. During inflammation, increases in permeability of the mammary epithelium also allows Na+ to move into the milk thereby increasing concentration within the milk [12]. Changes in the Na+:K+ ratio appear to be driven by the increase in Na+ [11], with K+ concentrations shown to stay relatively stable in breast inflammation [10]. However, use of the Na+:K+ ratio is thought to be more accurate than Na+ alone, as it adjusts for fluctuations in the aqueous and lipid portions of the milk [13, 14], and therefore Na+ and K+ are often reported together. Testing for the specific changes of Na+ and Na+:K+ ratio levels in human milk composition, is therefore a potentially useful clinical tool [15] for detection and management of clinical ICLB. After initiation of lactation, thresholds of Na+ > 16 mmol/L [11] and/ or Na+:K+ ratio > 0.6 [16], > 0.8 [14] or > 1.0 [11, 17] can indicate ICLB.

Traditionally, measurement of Na+ and K+ in human milk has been performed in specialised laboratories requiring costly equipment and trained personnel [13], delaying clinical results. This delay has impeded the usefulness of Na+ and K+ measurement for ICLB indication and management, as mothers with ICLB are often acutely unwell and require rapid detection and management to prevent deterioration. On the other hand, seemingly soft signs of inflammation are often not reported yet measurement of high Na+ may escalate preventative measures for progression of potential inflammation. Recently, portable handheld ion selective electrode probes (ISEPs) have been shown to accurately detect Na+ and K+ concentrations in human milk of mothers without ICLB, in the laboratory, compared to the reference laboratory measurement technique of inductively coupled plasma-optical emission spectrometry (ICP-OES) [13, 15]. Further, milk testing using handheld ISEPs was found to be able to detect the onset of breast inflammation in mothers of preterm infants, during their neonatal unit stay [11]. Validity of ISEPs to measure Na+ and K+ at point-of-care (i.e., bedside, by a clinician) in the population of full-term mothers residing in the community with ICLB has yet to be tested.

The primary objective of this study was to determine the validity of ISEPs as a point-of-care instrument at the bedside, as compared to the reference method laboratory technique of ICP-OES in mothers with ICLB. The secondary objective was to assess the acceptability of ISEP human milk testing at point-of-care in mothers with ICLB who received treatment.

Methods

Validation of the ISEPs was performed, comparing the Na+, K+ and Na+:K+ ratio concentrations measured at point-of-care against the laboratory technique ICP-OES, in human milk samples collected from mothers with ICLB. Ethical approval for this study was given by Mercy Health Human Research Ethics Committee (2021-007). This study was carried out according to The Code of Ethics of the World Medical Association (Declaration of Helsinki) and reported according to the guidelines for reporting reliability and agreement studies (GRRAS) [18].

Participants

Participants were lactating mothers with ICLB, defined as at least one local breast inflammatory symptom (e.g., pain, redness, heat, swelling) and at least one accompanying systemic symptom (e.g., fever, chills, aching, lethargy, headache) this ICLB episode [3]. Mothers were more than seven days postpartum to exclude those with physiological engorgement, which is regarded as a distinct clinical entity, usually occurring between days three and five post birth, with the onset of copious milk production (secretory activation) [3]. Eligible mothers were recruited from Melbourne, Victoria as part of an ongoing larger randomised pilot trial assessing the feasibility and limited effectiveness of different therapeutic ultrasound treatment parameters for ICLB (protocol trial registration at ACTRN12621000118808). Mothers could have been in the treatment or sham treatment arm of the trial and were encouraged to contact the principal investigator of the trial as soon as their symptoms developed or as soon as they heard about the trial which could have been at any point along the course of their ICLB. Mothers were recruited from private physiotherapy practices, a public tertiary women’s hospital and directly from the local community in Melbourne, Victoria, Australia, and gave written informed consent. All participant data was anonymised to maintain participant privacy.

A-priori power calculations, based on human milk Na+ thresholds for mastitis of > 12 mmol/L [19] and > 16 mmol/L [20], indicated 12 mmol/L required the higher number of mothers for a repeated measures within factor model with one group, three human milk samples and an effect size f = 0.2 (equivalent to partial eta squared = 0.04), which was 41 participants (G*Power V3.1.9.4) [21].

Demographic variables were collected from participants via secure electronic REDCap survey [22, 23], prior to sample collection. Variables included: (1) participant background (maternal age, postcode, education level, pre pregnancy body mass index (BMI), health conditions and smoking whilst breastfeeding); (2) lactation history (prior breast/ nipple conditions, other children, secretory activation, current medication); (3) birth and infant (mode of delivery, antibiotic use during labour and infant age, sex and oral variations including tongue, lip or buccal ties, high palate, small mouth, receding chin); and (4) current ICLB (affected breast, symptom onset, antibiotic use, prior ICLB, breastfeeding difficulties and nipple treatments).

Point-of-care measurement (ISEPs)

Three human milk samples of up to 5 mL were collected from the mother’s affected breast on day 1, 3 and 10, between December 2021 and September 2022. Day one represented the first day of the trial, and not necessarily the first day of symptom onset. Day three represented the third and final day of trial treatment and day 10, the follow-up assessment only session of the trial. Samples were not taken on consecutive days to allow for the delay in changes to Na+ and K+ levels seen with breast inflammation [11]. A visual flow of the measurement process is detailed in Fig. 1. Hand hygiene, Chlorohexidine wipe to nipple and surrounding breast area pre-expression, and use of sterile/ sanitised test tubes, bottles and syringes for sample collection and measurement was ensured to mitigate the risk of environmental sodium contamination. Pre-measurement, 2-point calibration of the ISEPs (LAQUAtwin Na: B-722 and K: B-731; HORIBA Ltd, Japan) was performed as per the manufacturer’s instructions, followed by a repeat measurement for quality assurance using a known standard solution (HORIBA Ltd, Japan). The coefficient of variation (CV) for ISEPs, as observed in similar studies, has ranged from 2 to 4.1% for Na+ and 2 to 3.9% for K+, both acceptable values (< 10%) [13, 15]. Six samples from n = 5 mothers could not be tested immediately at point-of-care, due to COVID-19 restrictions or participant geographical location (distanced from the principal investigator). These samples were frozen immediately post expression for later analysis, which occurred a median (Q1, Q3) of 9.5 (5.5, 13.5) days post collection (range: one to 15 days). Sample collection and point-of-care measurement were performed at the private physiotherapy practices, public women’s hospital and participants homes by author E.H., a women’s health physiotherapist with no prior ISEP experience, who was trained to use the ISEPs by author C.T.L. Training comprised a 30-minute online video meeting. Once the a priori sample size was achieved, all samples were flown on dry ice, maintaining lower than minus 20-degree temperature, to University of Western Australia’s Human Lactation Laboratory (Physical Containment Level 2) [24] and stored at minus 80 °C until laboratory analysis.

Fig. 1.

Flow diagram of the measurement process. ICLB, inflammatory conditions of the lactating breast; ISEPs, ion selective electrode probes; ICP-OES, inductively coupled plasma-optical emission spectrometry; Na+, sodium ion; K+, potassium ion

Laboratory measurement (ICP-OES)

All samples were analysed as one batch, except for two individual samples which required re-analysis due to ICP-OES (Agilent 5110, Agilent Technologies, United States) machine error. Quality control measures, comprising machine calibration with known mineral standards, quality control sample and National Institute of Standards Technology standard reference material (skim milk), ensured consistency across the two ICP-OES analyses batches. The laboratory measurement (see Fig. 1) was performed primarily by author E.H., who had no prior ICP-OES experience and received training and assistance from author C.T.L. At the time of testing, the author (C.T.L) who collated the automatically generated ICP-OES measurements was blinded to the ISEP measurements. Due to the automatic nature of the machinery and batch analysis, E.H. could not influence the results.

To determine assay reliability as described in Esquerra-Zwiers et al. (2022) [15], five rounds of spike and recovery assays were conducted, with each round comprising three tubes. One tube contained a known standard solution and milk, the second tube contained water and milk, and the third tube contained a known standard solution and water. The ICP-OES was able to recover 94.0% for Na+ and 103.0% for K+, within the acceptable range of 80–120% [25]. The CV for Na+ and K+ were 2.5% and 2.3%, respectively, less than the acceptable value of 20% [13].

Participant rating of acceptability

Participants completed two questions about the acceptability of the ISEP testing at point-of-care, as part of the final follow-up session of the clinical trial: (1) “How acceptable did you find giving a milk sample for testing with the ion selective electrode/probe?” and (2) “Please describe why you scored this.” Question responses comprised a Numerical Rating Scale (NRS) slide bar from 0 = not acceptable to 10 = very acceptable, and an open text response box, respectively. A pilot study comprising five lactating mothers and/ or women’s health physiotherapists was conducted to ensure the length and functionality of the questionnaire, and clarity and nature of the questions, were acceptable. The mothers and/ or physiotherapists were known to the primary researcher (EH) and selected due to their personal and/ or clinical experience with lactation and ICLB. One minor change related to the functionality of the questionnaire was made based on email feedback received from the mothers/ women’s health physiotherapists.

Data analyses

Data was analysed using IBM SPSS statistics (Version 29), R 4.2.2 [26] and RStudio 2022.07.2 build 576 [27]. Descriptive statistics comprised number (%) and either mean (± standard deviation (SD)) or median (quartile (Q)1, Q3), depending on normal or non-normal distribution respectively. Normality was assessed via visual inspection of histograms, performed by author E.H, and cross-checked by author L.M. for credibility.

To compare the Na+, K+ or Na+:K+ ratio concentrations obtained from the two different methods, the limits of agreement and the precision of these limits were calculated [28], with Bland-Altman plots. A mixed effects model was used to calculate the adjusted mean difference between the two measuring methods, accounting for within-subject correlation and time-specific proportional bias, given up to three measurements were available for each participant (T1, T3 and T4). The acceptable adjusted limits of agreements were defined as 95% of the samples falling within ± 2 SDs from the mean adjusted difference. Separate rank linear mixed effects models were used to establish the relationship between Na+, K+ or Na+:K+ ratio concentrations as measured by the two methods [29]. The models again accounted for individual variability by including a random intercept for each participant to address potential clustering or repeated measures within individuals. Regression equations were reported in the following format: y = mx+b, where y = ICP-OES and x = ISEP. Statistical significance was determined at p < 0.05.

For the participant rating of acceptability data, both quantitative and qualitative analysis was undertaken. Specifically, descriptive statistics, number (%) and median (Q1, Q3), were used for the 11-point NRS responses and thematic analysis for the open text responses, which involved both inductive (theme generation) and deductive (positive and negative response identification) analysis. Thematic analysis was performed by author E.H. and cross-checked by author L.M. to improve the credibility of themes. Both authors were experienced in women’s health qualitative research.

Results

Participants

In total, 128 human milk samples were collected from 43 participants (Fig. 1). There was one individual missing sample from one participant (at day 10). Participant demographic characteristics are presented in Table 1. Mothers had a mean (± SD) age of 35 (± 4) years, and infants a median (interquartile range) age of 2 (5) months. The majority lived in socio-economically advantaged areas, had completed a bachelor’s degree or above (n = 34, 79%), did not smoke while breastfeeding (n = 43, 100%), and had no other children (n = 23, 54%).

Table 1.

Participant demographic characteristics (n = 43)

Characteristics variable	N (%) or mean (± SD)a
Maternal background
Age (years)b	34.6 (± 4)
Postal area SEIFAb, c	Median 9 (7, 10)
Pre pregnancy BMI	23.9 (± 3.2)
Health conditions:
- Anxiety	6 (14%)
- Depression	2 (5%)
- Fertility issues requiring assisted reproduction	4 (9%)
- Thyroid disorder	3 (7%)
- PCOSd	4 (9%)
Lactation history
Prior breast/ nipple conditions:
- Nipple piercing	2 (5%)
- Benign breast lump/s or cyst/s	3 (7%)
- Dense breast tissue	1 (2%)
- Nil prior conditions	37 (86%)
Other children:
- 0	23 (54%)
- 1	13 (30%)
- 2	5 (12%)
- 3	2 (5%)
Secretory activation:
- Within 24 h	3 (7%)
-  24 to 48 h	13 (30%)
- 48 to 72 h	12 (28%)
-  3 to 4 days	6 (14%)
-  4 to 5 days	6 (14%)
- More than 5 days	3 (7%)
Engorgement with secretory activation:
- Yes	21 (48%)
- No	19 (43%)
- Don’t know/ unsure	3 (7%)
Current medications:
- Vitamin/ mineral supplements	23 (54%)
- Iron supplements	12 (28%)
- Anti-depressant or anxiety medications	6 (14%)
Birth and infant characteristics
Mode of delivery:
- Natural vaginal delivery	18 (42%)
- Interventional vaginal deliverye	13 (30%)
- Emergency caesarean	4 (9%)
- Elective caesarean	8 (19%)
Antibiotics during labour	15 (35%)
Infant sex female	23 (54%)
Infant age (days)	Median 66 (34, 188); min: 8 days, max: 20 months
Current oral variationsf	11 (25.6%)
Current ICLB
Affected (or worst) breast = right	22 (51%)
Symptom onset (days)	Median 2 (1, 4)
Current antibiotic use for ICLB	17 (40%)
Prior ICLB this lactation	15 (35%)
Breastfeeding difficulties:
- Sore nipples	19 (44%)
- Nipple injury	1 (2%)
- White spot	4 (9%)
- Attachment difficulties	10 (23%)
- Nipple shield use	8 (19%)
- Low milk supply	3 (7%)
- Oversupply	9 (21%)
Current nipple treatments	0 (0%)

^aUnless otherwise stated. ^b1 missing data point (n = 42; age – year of birth not given, SEIFA – postcode didn’t exist in 2016); ^cSEIFA = Australian Bureau of Statistics Socio-Economic Indexes for Areas (1 = Most disadvantaged, 10 = Most advantaged); ^dPCOS = Polycystic ovarian syndrome; ^ee.g. induction, vacuum, forceps; ^fincludes: tongue tie/s (n = 5), lip tie/s (n = 2), high palate (n = 1), small mouth (n = 2), receding chin (n = 1)

Na+, K+ and Na+:K+ ratio measurements (ISEPs and ICP-OES)

Concentrations of Na+, K+ and Na+:K+ ratio measured with ISEPs and ICP-OES, and their mean differences, are described in Table 2, along with their limits of agreement and precision of the estimated limits of agreement. The data was not normally distributed.

Table 2.

Ion concentrations as measured by ICP-OES and ISEPs (n = 128) and limits of agreement between the ICP-OES and ISEP measurements

Ion (n = 128)	Concentrations		Mean adjusted difference	Adjusted Limits of Agreement
	ISEP	ICP-OES	Mean (95% CI)	Lower	Upper
	Median (Q1, Q3) (min – max)	Median (Q1, Q3) (min – max)		Mean (95% CI)	Mean (95% CI)
Na+ (mM)	7.0 (4.8, 10.8) (2.4–69.6)	10.7 (8.0, 18.3) (4.7–76.6)	-0.001 (-0.94, 0.94)	-6.12 (-7.75, -4.49)	6.12 (4.49, 7.75)
K+ (mM)	13.8 (12.3, 15.4) (8.2–21.0)	21.2 (18.4, 25.2) (8.9–42.6)	16.5 (15.3, 17.7)	7.37 (5.82, 9.47)	25.6 (23.5, 27.7)
Na+:K+ ratio	0.5 (0.4, 0.7) (0.2–8.5)	0.5 (0.4, 0.7) (0.2–5.3)	-0.01 (-0.02, 0.003)	-0.82 (-0.85, -0.79)	0.80 (0.77, 0.82)

Na+, Sodium ion; K+, Potassium ion; ICP-OES, inductively coupled plasma-optical emission spectrometry; ISEP, ion selective electrode probe; Q1, quartile 1; Q3, quartile 3; CI, confidence interval.

For Na+, 88% (112/128) of samples fell within 2 SDs of the adjusted mean difference (Fig. 2a). Visually, Bland-Altman plot displayed a diagonal trend for Na+, with the higher the average measure, the greater the difference between the measuring methods (Fig. 2a). Timepoint 1 accounted for 64% (9/14) of the samples outside of 2 SDs of the mean difference (Fig. 2a). For the regression, the marginal R² was 0.77, indicating the variance explained by the fixed effect (ISEP measurement), while the conditional R² was 0.87, reflecting the total variance explained including random effects (Fig. 3a). The overall regression was significant (p < 0.05) (Fig. 3a). The fitted regression model was: Na+ICPOES = 0.863*Na+ISEP+8.941.

Fig. 2.

Bland-Altman plot and bias for: (a) Sodium ion (Na+) measurement, showing 16 points outside 2 Standard Deviation (SD); (b) Potassium ion (K+) measurement, showing 23 points outside 2 SD; and (c) Na+:K+ ratio measurement, showing 0 points outside 2 SD. Red dotted line = mean adjusted difference. Blue dotted lines = 2 SD from mean adjusted difference. Shaded areas = 95% confidence intervals. Na, sodium ion; K, potassium ion; NaK, sodium : potassium ratio; ICP-OES, inductively coupled plasma-optical emission spectrometry; ISEP, ion selective electrode probe

Fig. 3.

Regression plots for: (a) Sodium ion (Na+) regression: coefficient of determination R² (marginal) = 0.77, R² (conditional) = 0.87, p < 0.05; (b) Potassium ion (K+) regression: R² (marginal) = 0.06, R² (conditional) = 0.61, p < 0.001; and (c) Na+:K+ ratio regression: R² (marginal) = 0.94, R² (conditional) = 0.94, p < 0.05. Solid blue line = regression line. Shaded area = 95% confidence interval. Na, sodium ion; K, potassium ion; Na:K, sodium : potassium ratio; ISEP, ion selective electrode probe; ICP-OES, inductively coupled plasma-optical emission spectrometry

For K+, 82% (105/128) of samples fell within 2 SDs of the adjusted mean difference (Fig. 2b). Visually, Bland-Altman plot for K+ displayed a diagonal trend, with the higher the average measure, the greater the difference between the measuring methods (Fig. 2b). For the regression, the marginal R² was 0.06, while the conditional R² was 0.61, p < 0.001 (Fig. 3b). The fitted regression model was: K+ICPOES = 0.249*K+ISEP+48.662.

For Na+:K+ ratio, 100% (128/128) of samples fell within 2 SDs of the adjusted mean difference (Fig. 2c). For the regression, the marginal R² was 0.94, while the conditional R² was 0.94, p < 0.05 (Fig. 3c), with little deviation observed from the regression line, especially at higher scores (Fig. 3c). The fitted regression model was: Na+K+ICPOES = 0.972*Na+K+ISEP+1.812.

Acceptability

Most participants scored 10, “very acceptable”, for the ISEP point-of-care testing (n = 35/43, 81%). All scores were ≥ 8 (median 10 (10, 10)). Qualitatively, four themes were generated. The most frequent theme was theme 1: The testing was easy and quick with no issues or problems (n = 30), followed by theme 2: The testing is important to advance medical understanding (n = 13): “It was genuinely quite a quick procedure, and it was good to have those markers to compare over the week of the trial. Mentally it was good to know that the inflammation was decreasing” [P6]. The remaining two themes comprised theme 3: Small milk sample required (n = 4) and contrastingly, theme 4: Difficulty with pumping or providing a milk sample from affected breast (n = 3): “Initially it was difficult to express from the affected breast and [E.H.] was very patient with me and gave me the time I needed plus offering to use an electric pump as I didn’t have one” [P15]. Overall, the open text responses were largely positive (n = 39/43, 91%). Of the negative responses (n = 4), three constituted theme 4 above, pertaining to difficulty providing a sample of up to 5 mL. The themes and supporting quotes are outlined in Table 3. Full set of open text responses are provided in Additional file 1.

Table 3.

Thematic responses of participant acceptability of ISEPs point-of-care testing

Themes and supporting quotes
Theme 1: Easy, quick with no issues or problems ( n  = 30)
It was quick and easy. (P43)
No issues with providing the milk sample…. (P7)
Theme 2: Important to advance medical understanding ( n = 13)
I found it very informative to see the different levels of salt on day 1 and 3 as this gave me a good indication of if my infection was improving or getting worse (P13)
If breastmilk properties change when suffering from mastitis, it’s a good way to track it. (P16)
Theme 3: Small sample ( n = 4)
…such a small amount of milk (P10)
… It was a very small amount needed for the sample…. (P16)
Theme 4: Difficulty with pumping or providing a sample from affected breast ( n = 3)
I have trouble with pumping (P8)
It was challenging the first time as I didn’t remember to massage first to stimulate the flow. The subsequent times it was fine. (P5)

ISEPs, ion selective electrode probes

Discussion

This study successfully validated the accuracy of ISEPs at point-of-care with ICP-OES in full-term mothers with ICLB. Importantly an allied health professional (physiotherapist) performed the testing, in a clinical population, in a community setting. Overall, our Bland-Altman and linear regression findings showed that point-of-care human milk Na+:K+ ratio ISEP testing is acceptable and valid in mothers with ICLB. The strongest relationship between the two measuring methods in this population of mothers was seen for the Na+:K+ ratio. Thus, these findings indicate that as a clinical tool, the ISEP demonstrates superior validity for measuring Na+:K+ ratio in this population of mothers with ICLB, compared to Na+ or K+ only. Importantly acceptability of the ISEP at point-of-care was extremely high. Mothers reported the testing to be a very acceptable, a quick easy process, with informative results. Altogether, these findings provide important evidence for the clinical use of ISEPs in mothers with ICLB.

For Na+, there was moderate variation between the ISEP and ICP-OES measurements, less so at higher concentrations (Fig. 3a). However, there appeared to be a systematic bias, with ICP-OES measurements higher than ISEP (Table 2). Previous studies have shown no differences between ISEP and ICP-OES measurement of Na+ in human milk [13, 15]. This may be due to a difference in the handling of our samples which involved the ISEP tested milk (0.25–0.5 mL) being returned to the original sterile test tube housing any remaining sample. Given Na+ is ubiquitous, additional handling may have led to an increase in Na+, with the inherent sensitivity of ICP-OES being able to detect any small but significant increases in Na+ concentration. Thus, we suggest the milk used on the ISEP be discarded post testing and not re-tested in the laboratory. The bias for Na+ appeared to become more prominent in mid to low concentrations or potentially the more well the mother with ICLB, or with developing symptomology or resolution (Fig. 3a). This difference implies that the linear regression equation cannot be consistently applied, to negate the systematic bias. On this basis, our recommendation is not to use Na+ only in the clinic. Further testing of the accuracy of the Na+ ISEP in this population is warranted, to include more samples with low to moderate levels of Na+.

For K+, the relationship between methods was not as strong as Na+ (Fig. 3b), which appeared to be at all values. This is consistent with two other studies [13, 15] and may be due to the design of the K+ ion electrode membrane, which differs from the Na+ ion electrode membrane, and makes it susceptible to interference in measurement from the milk matrix, particularly high lipid content [13, 15]. The Na+ ISE membrane is embedded with sodium ionophores, whereas the K+ ISE membrane has valinomycin as the ionophore. Whilst highly selective for K+, valinomycin is more lipophilic, meaning lipid particles in human milk may interfere with the K+ membrane, physically obstructing the membrane surface / molecular pores and reducing the accuracy of K+ measurement [30, 31].

The ISEPs shared 94% of variability with the ICP-OES measurement for Na+:K+ ratio in our study (R² = 0.94), which was similar to findings from a previous study (strongly correlated, r = 0.98, therefore R² = 0.96) of in-laboratory tested samples donated from breastfeeding women at 10 days postpartum [15]. A ratio is considered more sensitive than individual absolute values, as it can account for/ normalise variation that can occur in the individual absolute values, by comparing one value relative to the other. This helps to mitigate the impact of biases, such as fluctuations in the aqueous and lipid portions of human milk or potential addition of Na+ in our samples. Given the strong relationship between ISEPs and ICP-OES measurement for Na+:K+ ratio, this provides confidence for clinicians using point-of-care Na+:K+ ratio testing in mothers with ICLB. Our range of Na+:K+ ratio levels and collection timepoints also indicate point-of-care Na+:K+ ratio testing is likely valid for a range of ICLB symptom severities/ pathophysiological change, and for the levels used for indicating mastitis (> 0.6 [16], > 0.8 [14] or > 1.0 [11, 17]). It could also prove useful over the course/ progression of ICLB to monitor treatment response and ensure recovery. This, however, warrants further research, specifically, assessing the ability of the ISEPs to detect change in Na+:K+ ratio over the entire clinical time course of ICLB and validation of the ISEP against other ICLB outcome measures, for example a patient reported outcome measure on the severity of symptomology.

For clinicians managing mothers with ICLB, our findings suggest that ISEP point-of-care measurement of Na+:K+ ratio levels are valid and thus may be useful for ICLB detection, monitoring of resolution and identifying when ICLB is not resolving. The high level of acceptability amongst mothers with ICLB on the point-of-care testing strongly supports the use of this tool in clinical practice, and its potential impact to empower mothers in their recovery. Clinicians require an objective outcome measure for ICLB, and ISEPs would fill this current gap in clinical practice.

A limitation of this study were the outliers in the dataset that could be due to potential contamination. For the six samples that could not be tested at point-of-care, freezing prior to ISEP analysis is unlikely to have impacted ion measurement according to previous publications [10, 32, 33]. The strengths of the study include a wide variation in Na+ and K+ levels, stage of lactation and clinical course, reflective of clinical patient presentations, enhancing clinical utility of this valid outcome measure.

The ISEP to measure Na+:K+ ratio has high utility for multi-disciplinary clinicians managing mothers with ICLB, broadening its scope of use across the different health professionals involved in managing ICLB. This may extend beyond utility in ICLB, with the potential for ISEP measurement to have high clinical utility for tracking lactation progress and diagnosing other health concerns in lactation, providing impetus for future research.

Conclusions

Compared to ICP-OES, ISEP is a valid and acceptable point-of-care clinical tool for testing human milk Na+:K+ ratio levels, a marker of inflammation, in mothers with ICLB. Mothers report it to be an easy, quick, informative, and very acceptable measure. Measurement via ISEP has clinical potential in providing objective measures for ICLB, hastening detection and recovery by facilitating prompt, effective clinical decision making for mothers with ICLB. Future research in larger more diverse clinical settings and populations over the entire course of ICLB would be of value to determine the utility of ISEPs in monitoring treatment response.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

E.H.: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Data curation, Writing – Original Draft, Visualisation, Funding acquisition.C.T.L.: Methodology, Software, Formal analysis, Resources, Data curation, Writing – Original Draft, Visualisation, Supervision, Funding acquisition. L.M.: Conceptualization, Validation, Formal analysis, Writing – Review & Editing, Project administration, Supervision. A.M.: Conceptualization, Methodology, Writing – Review & Editing, Supervision. D.G.: Conceptualization, Methodology, Validation, Resources, Writing – Review & Editing, Supervision, Project administration, Funding acquisition.All authors read and approved the final manuscript.

Funding

This work was supported by Australian Government Research Training Program [author E.H. received a scholarship]; Mercy Health [Small Grants 2021/22 and 2022/2023]; and Medela AG [authors C.T.L and D.G. received a salary from an unrestricted research grant]. The funding sources were not involved in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Ethics approval for this study was granted by Mercy Health Human Research Ethics Committee (2021-007). All participants gave written informed consent. The research was completed in accordance with the Declaration of Helsinki as revised in 2013.

Consent for publication

Not applicable.

Competing interests

Emma Heron reports financial support was provided by Australian Government Research Training Program and Mercy Health. Donna Geddes and Ching Tat Lai report financial support was provided by Medela AG. Emma Heron reports equipment was provided by Medela AG. Donna Geddes reports a relationship with Medela AG that includes: past board membership. The remaining authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The funding sources were not involved in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.