Background
Lymphedema can be clinically diagnosed based on the swelling of limbs. However, definitive diagnosis of lymphedema is difficult, because most would suggest that lymphatic dysfunction imaged by lymphoscintigraphy or indocyanine green lymphography is required. The clinical diagnosis of lymphedema includes the observation that the bilateral difference in limb circumference is 2 cm or more [
1], the difference of pre and post operation in volume of the limb is more than 200 ml [
2], or the bilateral difference in volume of the limb change is 10% [
3]. Because of differences in diagnostic criteria in these measurement methods, the definitive diagnosis is difficult. In addition, ultrasound, computed tomography, and magnetic resonance imaging are used to diagnose lymphedema. The ultrasound can measure volumetric and structural changes in the dermis, subcutaneous layer, and muscle, but information on the truncal anatomy of the lymphatic system can not be confirmed [
4]. The computed tomography can detect thickening of the skin and subcutaneous compartment, and thickened perimuscular aponeurosis [
4]. The magnetic resonance imaging can distinguish lymphedema, lipedema, and phlebedema, and can confirm the circumferential measurement edema, thickened dermis, and increased subcutaneous compartment [
4,
5].
The arm circumference measurement is a commonly utilized clinical diagnosis method [
6]. The circumference of both arms at 15 cm below the acromion process and the olecranon process is measured, and the circumference values of the affected arm and unaffected arm are compared [
6,
7]. However, with this method there is no standardized reference point and low sensitivity. The lack of evidence-based diagnostic criteria to define lymphedema has presented tremendous challenges in terms of diagnosis. Therefore, defining criteria for the early detection and treatment of lymphedema is important [
8].
Recently, several researchers have used bioelectrical impedance analysis (BIA) to diagnose lymphedema [
9‐
14]. This method is highly sensitive, can be used as a basis to establish standardized criteria, and can be used to measure extracellular space [
9,
15,
16]. Bioelectrical impedance predicts body composition using differences in electric conductivity upon sending a minute current through the human body [
17,
18]. In several studies, the single frequency bioelectrical impedance analysis (SFBIA) of the two arms obtained using bioelectrical impedance measurements was expressed as the ratio of the values of the operated and non-operated arms [
10,
19]. However, this method has not been validated as a diagnostic tool. It is therefore necessary to study this method further to establish it as an efficient diagnostic means.
The purpose of this study is to determine diagnostic accuracy of bioelectrical impedance as a diagnostic method based on the presence of lymphedema compared with circumference measurements. In addition, the aim is to identify risk factors to help prevent lymphedema for breast cancer survivors.
PIBW
PIBW = actual weight / ideal body weight*× 100.
*ideal body weight = (height in meters)2xideal BMI#.
#ideal BMI = female; 21 kg/m2, male; 20 kg/m2.
SFBIA ratio
SFBIA ratio = unaffected SFBIA / affected SFBIA.
Medical record collection
Clinicopathological information was obtained from the medical records of the participating women. Clinicopathological variables included surgery type (sentinel lymph node biopsy (SLNB) or axillary lymph node dissection (ALND)), number of dissected lymph nodes, and postoperative therapy such as chemotherapy, radiotherapy, hormone therapy, and target therapy.
Statistical analysis
The variables used in this study were anthropometric values, the SFBIA value, the intake of nutrients, activity level, healthy functional foods, surgical methods, the number of removed lymph nodes, and treatment methods. The results are described as mean and standard deviation. The relationship between variables in the non-BCRL and BCRL groups was analyzed. The chi-squared test or Fishers exact test and independent sample t-test or Mann-Whitney U test were used for the analyses. The reported p-values are two-sided and were considered statistically significant at 0.05 or less. All data analyses were performed using IBM SPSS Statistics version 23.0 for Windows (IBM Corp., Armonk, NY, USA).
Discussion
Early prevention and detection of postoperative lymphedema complications in breast cancer patients is important for quality of life. Our data provided evidence to support the use of the SFBIA ratio by BIA in lymphedema in breast cancer survivors. In addition, right axillary surgery was suggested to be a risk factor associated with lymphedema. Lymphedema was more common in patients with the right axillary procedure than those with the left. This may be due to the fact that right handed people are more common and have more axillary activity on the right side. Additionally, the number of dissected lymph nodes [
22‐
25], obesity [
23,
24,
26,
27], and surgery type [
22,
24,
25] were risk factors for lymphedema. It was concordant with previous study [
22‐
27].
In this study, BCRL was determined based on more than a 2 cm difference in the circumference, as measured 15 cm below the olecranon, or acromion process, of the arm not affected by the operation relative to the circumference of the arm on the same side as the operation [
7,
28,
29]. Limitations include potential errors in the measurements and the fact that diagnosis is not possible until clinical symptoms are seen. More precise techniques including ultrasound, computed tomography, magnetic resonance imaging, lymphoscintigraphy and other volumetric measurement can enhance the diagnosis of lymphedema. Further studies that evaluate the comparison these techniques and BIA are needed.
BIA is designed to measure edema as an extremely small electrical current passes through extracellular fluid [
9]. This technique distinguishes extracellular fluid from total limb volume [
9,
30]. In the presence of lymphedema, the SFBIA ratio [
10] is related to the accumulation of extracellular fluid [
31]. Our results show that the SFBIA ratio is larger at 5 kHz than at 1 kHz. BIA had good performance in terms of specificity (95.15%) and negative predictive value (96.08%). A diagnostic tool with a high specificity is more useful for ‘judging’ a disease when a person is positive and the negative predictive value can be used that the probability of not having disease given a negative diagnosis [
32]. These values indicate that BIA can be used as a method of monitoring and diagnosing lymphedema. Previous studies have reported that early surveillance for risk of lymphedema using bioimpedance spectroscopy with early intervention with compression garments can reduce the incidence of more advanced lymphedema [
11,
12]. Our cut-off value of the SFBIA ratio for diagnosing lymphedema should be validated in further studies. The usefulness of early detection through the SFBIA ratio is necessary to be evaluated as well.
Well-known risk factors for lymphedema include the surgery type [
22,
24,
25] and the number of dissected lymph nodes [
22‐
25]. Similarly, our study also demonstrated that the higher the number of lymph nodes removed in the ALND subjects, the higher the incidence of lymphedema. However, these were not observed for the SLNB because the number of lymph nodes removed is too small to affect the risk of lymphedema in these subjects. Overall, on average our SLNB subjects had less than 5 lymph nodes removed whereas the ALND had 16 to 20 lymph nodes removed. Thus, surgery type associated with high lymph node removal is likely to increase the risk of lymphedema as reported previously [
22‐
25].
Obesity-related indicators in breast cancer patients increase the risk of developing lymphedema complications [
26,
33,
34]. Our data investigated the relationship between the incidence of lymphedema and the variables related to anthropometric measurements and type of surgery. The BMI and PIBW of the subjects were significantly higher in the presence of lymphedema. In particular, body fat percentage, BMI, and PIBW were significantly different in patients who underwent SLNB. These findings suggest that the occurrence of lymphedema is associated with obesity and that patients who undergo SLNB procedures should pay attention to maintaining normal weight. There was no significant difference of obesity-related indicators between BCRL and non-BCRL in the patients with ALND. Our results were different from the studies conducted on Westerners. The association between BMI and lymphedema volume in patients with ALND was observed [
35,
36]. The difference between ours and previous studies may be due to difference of study population, a race, culture, lifestyle, and dietary differences between Westerners and Asians. Further research is needed to understand the factors behind these differences.
The lymphedema of our subjects was confirmed by the difference in limb circumference of 2 cm of both sides. This confirmed the lymphedema of the arm, but not the lymphedema that appeared at other sites. Therefore, the diagnosis of more detailed lymphedema will establish an accurate standard of the SFBIA ratio. Nevertheless, our study confirmed the cut-off value of the SFBIA ratio for the determination of lymphedema through 228 subjects and confirmed the sensitivity and specificity. Our findings have shown the possibility of SFBIA ratio as a useful tool for the diagnosis and management of lymphedema in breast cancer survivors. In addition, we found that there was a significant correlation between lymphedema and obesity in patients who underwent SLNB, but not in patients who underwent ALND.
Conclusion
The SFBIA ratio obtained using BIA can be an alternative method for monitoring and/or diagnosing BCRL. The BIA had a sensitivity of 63.64% and a specificity of 95.15% in predicting BCRL. In addition, number of dissected lymph nodes, operation site, surgery type, obesity, and the SFBIA ratio are significantly associated with the occurrence of lymphedema.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.