Comparing perceived heating effect and intensity of therapeutic ultrasound between breast tissue and calf muscle in lactating women: An observational study

Background Inammatory conditions of the lactating breast (ICLB) can cause cessation of breastfeeding and treatment for these conditions should be optimised. Therapeutic ultrasound is the most utilised physiotherapy treatment for ICLB. However, there is no evidence that guides physiotherapists on the most appropriate therapeutic ultrasound parameters to treat ICLB, with current ultrasound treatment protocols based on evidence using musculoskeletal tissue. Therefore, it is necessary to determine whether ultrasound interacts with breast and musculoskeletal tissue differently. This study aimed to determine the difference between breast and calf muscle tissue for perception of heating from therapeutic ultrasound; and the therapeutic ultrasound intensity required to perceive warmth in healthy lactating breast tissue. Methods Fifty adult lactating women who had exclusively breastfed infants aged 6 months or fewer, were recruited from a private physiotherapy clinic in Queensland, Australia. Consent was obtained to participate in this repeated measures within subject descriptive observational study. Therapeutic ultrasound was sequentially applied to the breast and calf; in a randomly assigned order; using a continuous duty cycle of 1 MHz, 1Wcm 2 intensity for an initial period of 5 minutes. Intensity was then increased by 0.1Wcm 2 every 30 seconds. The intensity at which participants reported rst perceived warmth and then most tolerable warmth for the breast and the calf was recorded. Results First warmth was perceived at lower intensities in the breast (Z=-3.637, p < 0.001), but there was no difference between locations for most tolerable warmth (Z=-1.165, p = 0.244). Factors associated with intensity for perception of rst warmth were antidepressant use (β 0.369 95% condence interval [95%CI: 0.103–0.635], p = 0.007) and location (β 0.286 [95%CI: 0.055–0.516], p = 0.015). Body mass index (BMI) was the only factor associated with intensity for most tolerable warmth (β 0.024 [95%CI: 0.004–0.044], p = 0.017). The range of intensities


Abstract
Background In ammatory conditions of the lactating breast (ICLB) can cause cessation of breastfeeding and treatment for these conditions should be optimised. Therapeutic ultrasound is the most utilised physiotherapy treatment for ICLB. However, there is no evidence that guides physiotherapists on the most appropriate therapeutic ultrasound parameters to treat ICLB, with current ultrasound treatment protocols based on evidence using musculoskeletal tissue. Therefore, it is necessary to determine whether ultrasound interacts with breast and musculoskeletal tissue differently. This study aimed to determine the difference between breast and calf muscle tissue for perception of heating from therapeutic ultrasound; and the therapeutic ultrasound intensity required to perceive warmth in healthy lactating breast tissue.
Methods Fifty adult lactating women who had exclusively breastfed infants aged 6 months or fewer, were recruited from a private physiotherapy clinic in Queensland, Australia. Consent was obtained to participate in this repeated measures within subject descriptive observational study. Therapeutic ultrasound was sequentially applied to the breast and calf; in a randomly assigned order; using a continuous duty cycle of 1 MHz, 1Wcm 2 intensity for an initial period of 5 minutes. Intensity was then increased by 0.1Wcm 2 every 30 seconds. The intensity at which participants reported rst perceived warmth and then most tolerable warmth for the breast and the calf was recorded.
Conclusions As perception of breast and calf warmth was different, therapeutic ultrasound parameters used to treat musculoskeletal conditions may not be appropriate for ICLB. Antidepressant use and BMI in uenced heat perception, highlighting the need to screen for these factors prior to therapeutic ultrasound intervention.

Background
One in every three mothers cease breastfeeding due to pain or infection associated with in ammatory conditions of the lactating breast (ICLB) [1]. The World Health Organization (WHO) recommends that all infants are exclusively breastfed for the rst six months of life [2]. Breastfeeding helps to create a strong emotional and psychological bond between mother and baby and assists optimal development, growth and health [3,4]. There is compelling evidence to suggest that lactating and breastfeeding for mother and baby are associated with increased health bene ts [5]. Mothers who have lactated have a reduced risk of pre-menopausal breast cancer, ovarian cancer, cardiovascular disease, type ΙΙ diabetes and metabolic syndrome in comparison to those that have not [5]. Health bene ts for breastfed infants include a reduced incidence of infectious morbidity, reduced risk of childhood obesity, type Ι and ΙΙ diabetes, leukaemia and sudden infant death syndrome [5].
Conditions considered an ICLB may include engorgement, blocked milk ducts and mastitis [6]. The cost of ICLB to the Australian and United Kingdom healthcare systems could be as high as $120 million [7] and £26.8 million [8] respectively to treat infants with illnesses associated with early cessation of breastfeeding. In addition to the estimated economic impact, ICLB has negative emotional and physical impacts on mothers with young babies. Symptoms associated with ICLB can be disabling, with mothers suffering with mastitis stating "I have never felt worse" [9 p746] and "I felt as if I might die" [9 p746].
Physical symptoms of ICLB may include both local symptoms (breast pain, tenderness, redness and heat) and systemic symptoms (fever, hot sweats, cold shivers, lethargy, fatigue, nausea and vomiting) [9,10]. By facilitating breastfeeding through appropriate treatments of ICLB, health care costs and the emotional and physical impacts of ICLB could be reduced [11].
The most commonly reported physiotherapy treatment for ICLB is therapeutic ultrasound (TUS) [12]. Two of three studies [13][14][15] have demonstrated positive results regarding the e cacy of TUS for treatment of ICLB. A case series (n = 25) found that 92% of mothers reported resolution of symptoms associated with blocked ducts following TUS [14]. A cohort study (n = 30) indicated that TUS, patient education and manual therapy provided statistically signi cant decreases in pain, and increased participant breastfeeding e cacy for blocked ducts [13]. Conversely, a randomised controlled trial (n = 109) found no difference between TUS and placebo ultrasound for breast engorgement [15]. Although two of these three studies indicate that TUS may be an effective treatment for ICLB, the studies provide little guidance regarding optimal TUS parameters, as a range of TUS parameters were utilised. Consequently, a range of TUS parameters are used to treat ICLB clinically [12], and it is unknown how lactating women respond to differing TUS intensities. Clinicians need further information to guide their parameter selection, in particular intensity, for the treatment of ICLB.
In contrast, studies have shown that TUS is effective for treatment of skeletal muscle tissue injury [16][17][18]. Recent research, including a systematic review [16,17] found that TUS intensities of less than 1.5Wcm 2 reduced pain and healing time in tendon, muscle and ligament in ammatory injuries. However, it is not known if TUS parameters for in ammatory injuries in skeletal muscle are applicable to ICLB, due to the differences in tissue physiology. Whilst skeletal muscle has a high protein content, the lactating breast contains highly vascularised lymph, glandular tissue and milk substrates [19].
There have been no studies comparing the use of TUS on breast tissue to skeletal muscle tissue. Understanding how TUS interacts with different tissue types will help clinicians and policy makers to determine if results from muscular studies can be extrapolated to breast tissue. This evidence may ultimately aid in improving ICLB treatment, providing bene ts to mothers by alleviating the substantial impact on their physical [9,10] and mental health [20,21]. This promotes long term bene ts of uninterrupted breastfeeding as recommended by the WHO [2].
The dual aims of this current study were to compare the response of breast tissue with skeletal muscle for the perception of heating from TUS whilst accounting for confounding factors; and to determine the TUS intensity for initial perception of warmth and the most tolerable warmth in healthy breast tissue of lactating women.

Study Design
A repeated measures descriptive observational study design was implemented in accordance with the 'STROBE' checklist [22].

Participants and setting
Fifty women were recruited between June and August 2019 in Mackay, Queensland, using yers placed at general practitioner and obstetrician clinics, social media and word of mouth. The research was conducted at a private physiotherapy clinic. All participants were screened for eligibility via telephone. The ow of participants through the study is demonstrated in Fig. 1.
Eligible participants had exclusively breastfed for six months or fewer, were aged 18 or older and spoke English. Exclusion criteria included history of ICLB since the birth of their most recent baby, history of breast or calf cancer, past surgery to the breast or calf, or an active skin condition (e.g. eczema) on the breast or calf (Fig. 1).

Outcome measures
The primary outcome measure was intensity (Wcm 2 ) as measured on the TUS machine output display (Chattanooga, model: Intelect Advanced 2772MC). The intensity was recorded when the participant rst reported warmth and when the participant reported the most tolerable level of warmth. To ensure accuracy of the intensity value, the TUS machine was commercially calibrated, held within the same room and used only for the purpose of the study. All measurements were recorded by the physiotherapist in the same location.
Breast and calf skin temperature were recorded via infrared thermometer (Omron, model: MC-720) immediately prior to the application of TUS. The calf was selected for comparison, as this site is commonly used in musculoskeletal TUS studies [23]. In previous studies, inserted thermocouples recorded deep tissue temperature, which is not appropriate for lactating breast tissue due to increased risk of infective mastitis [24]. Thermometer values were checked for consistency prior to testing.
Participant age, height (tape measure), weight (calibrated digital scale, Omron: HN-286), number and age of participants' children and history of ICLB during previous lactation was collected. Data on the use of medications, such as antidepressants was collected on all participants, although for 23 participants this was collected retrospectively.

Intervention
Eligible participants were instructed to breastfeed their baby or express breast milk, within one hour prior to testing. The breast with the most recent milk removal was used for testing [25].
Two hours prior to testing, ultrasound gel was placed in the testing room. Ambient temperature was xed at 24°C, which standardised environment and gel temperature. A standardised script was used to ensure consistency across participants, who also completed a demographic survey using Qualtrics software (Qualtrics, Provo, UT), and provided their consent to participate. Clothing covering the ipsilateral calf was removed and one layer of towelling covered the calf and breast. Participants acclimatised for 30 minutes to allow skin temperature to equilibrate. An online computer program [26] was used to block randomise the order in which participants received breast or calf testing. Non-trial staff prepared sealed opaque envelopes to conceal allocation of order.
Participants were positioned so that the breast rested in a symmetrical conical shape. A hydrogel breast disc was placed over the nipple (Fig. 2), to contain milk ow in case of milk ejection during TUS [25] and prevent gel contacting the nipple.
Participants read a standard warnings and safety summary sheet prior to testing and kept as a guide throughout [27]. Thermal sensitivity testing was conducted on both locations to ensure the participant was not at risk of burns [27]. Only participants who passed this test had TUS applied.
The inferior lateral region of the breast was tested as it has higher levels of glandular tissue and is more likely to be affected by an ICLB [28]. For the calf, TUS was applied posterior to the bula, in line with the apex of the calf muscles. A template equivalent to twice the area of the TUS transducer head [29] was used to delineate the treatment area (Fig. 3). The transducer head temperature was recorded prior to the breast and calf applications and indicated that no residual heat remained in the transducer head.
After applying 10 ml of gel, the TUS was applied using overlapping linear strokes, with consistent light pressure at a rate of 55 bpm measured via metronome, using a 5 cm 2 soundhead. The ultrasound machine was initially set to continuous duty cycle, head warming off, 1 MHz [30][31][32] at 1Wcm 2 intensity for ve minutes. This starting intensity was chosen to ensure testing protocol was completed within one hour in order to minimise participant burden. Intensity was increased incrementally by 0.1Wcm 2 every 30 seconds. The physiotherapist checked participants comfort and level of warmth throughout the application. During testing, participants were blinded to intensity values. The physiotherapist recorded the intensity at which the participant rst reported warmth and again when she reported highest tolerable level of warmth. Testing ceased when the participant reported most tolerable warmth or 10 minutes after the maximum machine intensity of 2.5Wcm 2 was reached, even if the participant had not yet reported their most tolerable warmth.
Data were checked for outliers and transcription errors. Normality was examined via histograms, Q/Q plots and measures of central tendency. Descriptive summaries were based on frequency distributions for categorical data, and means and standard deviations or medians, interquartile ranges and ranges for continuous data, depending on normality.
Unadjusted comparisons between locations (breast and calf) for skin temperatures and TUS intensities at which rst warmth and most tolerable warmth were perceived, were conducted using non-parametric Wilcoxon signed rank tests.
Data was treated as repeated measures in order to account for within participant correlation between the two locations. Multilevel mixed effects Tobit (censored) regression models with random subject effects were used as the ultrasound intensity range had oor and ceiling effects (1.0Wcm 2 and 2.5Wcm 2 respectively). Patient factors including skin temperature, prior breastfeeding, antidepressant use and BMI that may have in uenced the differences in TUS intensities at which rst warmth and most tolerable warmth were perceived between locations were identi ed in univariable models. The potential confounding factors that were signi cant (p < 0.10) were subsequently entered into multivariable models along with breast or calf location. Residuals were examined graphically to assess model t. Results were summarised using Beta coe cients and 95% con dence intervals.
Data analysis was blinded, hypothesis tests were 2-sided, and P values of < .05 were considered statistically signi cant. Stata version 16.0 (StataCorp LLC College Station, TX) was used for data analysis.

Results
Fifty women participated in the study; one participant experienced anxiety and testing was terminated and her data removed from the data set (Fig. 1). Most participants had two children and had previously lactated ( Table 1). All 49 participants passed thermal sensitivity testing and prior to testing, skin temperature of the breast was warmer than the calf ( Table 2). Seven and six participants did not achieve most tolerable warmth in the breast and calf respectively, as the TUS machine was limited to 2.5Wcm 2 . Testing did not elicit milk ow from the nipple and no adverse events occurred.  Participants perceived rst warmth at a lower intensity in the breast as compared to the calf, while perception of most tolerable warmth was not different between locations (Table 2).
With respect to rst perception of warmth, the variables that were independently associated with intensity were location, skin temperature, antidepressant use and order of testing (Table 3). Univariate analysis demonstrated that higher initial skin temperature and the site tested subsequently, were associated with a lower intensity for the perception of rst warmth. Antidepressant use was independently associated with a higher intensity for rst warmth.
After adjustment for skin temperature, antidepressant use and order of testing, only location and antidepressant use remained signi cantly associated with ultrasound intensity (Table 3). Regarding the most tolerable perceived warmth, body mass index (BMI) was the only variable independently associated with the TUS intensity ( Table 4). Order of testing and BMI were entered into the multivariable model with location. Only BMI remained associated with ultrasound intensity after adjustment (Table 4). Table 4 Analysis for perception of most tolerable warmth (n = 49) With respect to the second aim there was a wide variety of intensities that participants reported for rst and most tolerable warmth in the breast (Table 2).

Discussion
Participants perceived rst warmth at a lower intensity in the breast compared to the calf, but there was no difference between locations for most tolerable warmth. After adjusting for confounding factors, these relationships remained. There was a wide range of intensities reported by participants for rst perceived and most tolerable warmth.
In relation to the primary aim, differences found between locations for rst perceived warmth, may be due to greater tissue heterogeneity and acoustic impedance in breast tissue or psychological and environmental factors [34]. Greater heterogeneity potentially causes greater de ection and absorption of US waves, with subsequent increased super cial heating, and earlier perception of rst warmth [35]. Additionally, the difference between the breast and calf may also be due to higher density of cutaneous thermoreceptors in the breast, resulting in an increased sensitivity to the heating [36]. These ndings may indicate that choosing TUS parameters for the treatment of ICLB based on TUS parameters used previously in musculoskeletal studies may not be appropriate.
No difference in intensity for most tolerable warmth was found between breast and calf, which could be explained by the upper limit in ultrasound machine intensity. As more than a quarter of participants did not achieve most tolerable warmth, this limits the conclusions that can be drawn from the most tolerable warmth data. A difference between the breast and the calf may have been evident, if ultrasound machines capable of intensities greater than 2.5Wcm 2 were available.
In relation to the secondary aim, the perceived range of intensities observed by participants for both rst warmth and most tolerable warmth were representative of the entire range of intensities tested from 1.0-2.5Wcm 2 ( Table 2). This re ects the intensities utilised in clinical practice of 0.8-2.4Wcm 2 [12]. As participants perceived heating over a wide range, clinicians may need to continue to provide individualised intensities when applying TUS to the breast, rather than a predetermined set intensity level.
Antidepressant use was associated with a higher intensity for rst perceived warmth, indicating that participants who use antidepressants required higher intensities of TUS. Antidepressants supress the body's thermoregulatory system causing decreased peripheral receptor sensitivity [37]. This could explain why antidepressants were perhaps not a signi cant factor for the perception of most tolerable warmth; therefore these women may need greater intensities for clinical e cacy. When physiotherapists are using TUS to treat women who use antidepressants, additional care is required, as their perception of heating is lower than women who do not take antidepressants.
Participants with a higher BMI were able to tolerate higher intensities of TUS. This could be due to higher fat deposition in the breast compared to the calf [38] resulting in greater dispersion and reduced absorption of ultrasound. Similar to antidepressants, the altered perceived heating effects in those with a high BMI, highlights the need for physiotherapists to monitor this group of patients more carefully.
Considering these ndings, physiotherapists should consider screening for antidepressant use and calculating BMI in all patients requiring TUS treatment.
The TUS machine utilised in this study is representative of current machines available in Australia, delivering a maximum intensity of 2.5Wcm 2 . Physiotherapists in the eastern states of Australia aim to achieve thermal effects of TUS to treat ICLB [12]. However, machines limited to 2.5Wcm 2 may not provide the desired perceivable heating effects. Historically, machines were able to deliver a maximum intensity of 3Wcm 2 but this was lowered to 2.5 Wcm 2 following reported heat injuries in musculoskeletal tissue [39]. The current data indicates that the manufactured machine intensity level could be raised to allow physiotherapists to treat ICLB with their desired intensity level. This would facilitate research to determine what range of TUS intensities is effective for treatment of ICLB.

Limitations and strengths
This study was a perception-based study as opposed to a study looking at breast tissue temperature.
Temperature perception can be affected by physiological, psychological and environmental factors, [34] and participants may perceive heating effects of TUS differently. Perception may be in uenced by participants' pre-conceived ideas that the breast would be more sensitive than the calf [34]. There have been no studies investigating if perception of warmth accurately re ects change in tissue temperature. Therefore, limited conclusions regarding tissue temperature or heating effects within the tissue can be made. This study was conducted on participants with healthy lactating breast tissue; hence, caution should be applied when extrapolating ndings of this study to women with ICLB.
A strength of this study is that it is the rst observational study to provide a comparison of perceived heating effects of TUS between lactating breast tissue and skeletal muscle and to empirically determine average intensities that lactating women would perceive heating. Thus, it provides an initial indication that research ndings from musculoskeletal studies on the use of TUS may not be extrapolated to the breast.

Future research
Future studies need to explore alternative ways to assess breast tissue temperature accurately without inserting thermocouples into the breast [24]. Research is needed to determine whether perception of warmth is closely correlated with change in tissue temperature. Research is also needed to determine whether thermal or non-thermal effects of TUS are more effective for ICLB and what parameters should be used.

Conclusion
This research suggests that the breast and calf do not respond to TUS in the same way, as warmth from TUS was perceived at lower intensities in the breast compared to the calf in lactating women. Therapeutic ultrasound parameters examined in musculoskeletal research and used clinically, may not be appropriate for ICLB.
The wide range of intensities at which participants perceived warmth was in uenced by antidepressant use and BMI. Whilst skin testing prior to indicating TUS treatment is necessary for all patients, physiotherapists need to speci cally screen for anti-depressant use and BMI in their clinical interview.
Furthermore, not all participants in this study felt most tolerable warmth at the maximal machine intensity. As a result, manufacturers should consider producing TUS machines with higher intensities, which will enable research to fully investigate TUS treatments for in ammatory conditions of the lactating breast. Availability of data and materials

Abbreviations
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request Flow of participants and methods