Lymphedema as a Meteoropathic Disease Based on a Retrospective Study in a Single Institute: a Preliminary Report

Since cellulitis is one of the most important factors for the prognosis of lymphedema, the prevention and prediction of cellulitis are considered to be critical in controlling lymphedema. We hypothesized that patients with lymphedema might show meteoropathy, as abdominal aortic aneurysm ruptures are considered to be inuenced by climatic conditions, and the lymphatic system is one of the circulation systems. Thus, we aimed to determine if the onset of cellulitis in limbs affected by lymphedema is related to climatic conditions. We reviewed the clinical records of patients with lymphedema admitted for cellulitis at our institute between January 2007 and December 2017. We identied 40 patients, 25 of whom lived in the same area. We examined the association between the number of patient admissions according to season and meteorological data obtained from the Japan Meteorological Agency database. Thirteen of 25 patients were admitted in summer, whereas only one patient was admitted in winter. Both higher temperature and lower atmospheric pressure around the day of admission were associated with the occurrence of cellulitis. Lymphedema may be regarded as a meteoropathic disease, as climatic conditions were shown to be associated with cellulitis in limbs affected by lymphedema.


Introduction
Most of the activities of living organisms are considerably affected by changes in the climate. Global surface temperatures have been rising for the past 50 years. From 2007 to 2017, the global temperatures increased by 0.26 ºC, while the levels of atmospheric carbon dioxide (CO 2 ), one of the greenhouse gases, increased from 382.7 ppm to 407.5 ppm (https://climate.nasa.gov/vitalsigns/global-temperature/). Cellulitis is well known as one of the most important factors for the prognosis of both primary and secondary lymphedema [1] . Patients with lymphedema with cellulitis often relapse, leading to subcutaneous tissue damage, including lymph vessel damage, followed by scar formation and brosis, which accelerate the deterioration in lymphedema. Accordingly, the prevention and prediction of cellulitis are both considered important in controlling lymphedema [2] .
Furthermore, disorders, such as migraine, pain associated with past trauma, and some infectious diseases are considered to be in uenced by climatic conditions [3][4][5] . Meteoropathy refers to diseases with an occurrence deeply associated with meteorological conditions [3,4] . Even abdominal aortic aneurysm rupture, a severely critical disease, has been reported to be related to changes in atmospheric pressure [6,7] . We hypothesized that lymphedema might also show meteoropathy, as the lymphatic system is one of the circulation systems. In the present study, we reviewed the admission records of patients in our institute to determine if the onset of cellulitis in limbs affected by lymphedema is related to climatic conditions, such as atmospheric pressure and temperature.

Methods
The participants comprised patients who were admitted to the Department of Plastic and Reconstructive Surgery at Yokohama City University Hospital from January 2007 to December 2017 for cellulitis in a limb affected by lymphedema. The computerized medical record system was used to identify candidates diagnosed with "lymphedema," whose reason for admission was cellulitis. Among the identi ed patients, we selected the largest group of patients who indicated the same city (Yokohama City) as their place of residence on the day of admission for a detailed analysis in order to control for geographical conditions and climate factors, such as temperature and atmospheric pressure. The admission date was classi ed into the following periods based on the four seasons: spring (March to May), summer (June to August), autumn (September to November), and winter (December to February) [6,7] .
Meteorological data were obtained from the Japan Meteorological Agency database (https://www.data.jma.go.jp/obd/stats/etrn/index.php). The 15-day centralized moving averages (CMAs) of the temperature and atmospheric pressure were calculated from 15 days before admission to the day before admission (Supplementary Fig. S1). The data were compiled using Microsoft 365 Excel (Microsoft Japan Co., Ltd., Tokyo, Japan).

Statistical analysis
Data normality was evaluated using the Shapiro-Wilk test. For non-normally distributed data, and in cases in which the number of samples did not exceed 5, the Kruskal-Wallis test was used for statistical analysis. For normally distributed data, the t-test was used. Statistical analysis was performed using R for Windows (x64 3.1.3, https://cran.r-project.org/bin/windows/base/). A p-value <0.05 was considered to be statistically signi cant.

Results
In total, 40 patients diagnosed with lymphedema were admitted to our hospital during the 11-year study period; among these, 25 lived in the same city (Yokohama City). None of the patients were admitted two or more times during the study period. Most of the patients were regarded as having cancer-related lymphedema, and none of the patients had lipolymphedema (Table 1, Fig. 1, Supplementary Table S1, Supplementary Fig. S2). Most of the patients were classi ed as having clinical stage 2 according to the scale of the International Society of Lymphology [8] . All of the 40 patients, except for two patients, owned garments or elastic bandages and two of these patients lived in Yokohama City.  . S3). Among the 25 patients from Yokohama City, 13 were admitted from June to August and 12 were admitted from September to May. In terms of seasons, 6 patients were admitted in spring, 13 in summer, 5 in autumn, and 1 in winter (Fig. 1).
The mean temperature and atmospheric pressure during the observation period are shown in Figure 2.
The atmospheric pressure showed a downward slope from spring to summer and an upward slope from summer to winter. The 15-day CMAs of the temperature and atmospheric pressure from 14 days before admission to the admission day showed seasonal differences among the three seasons from spring to autumn ( Table 2, Fig. 3). The mean change in temperature from 4 days before admission to the day before admission was 0.36 ºC in spring, 0.12 ºC in summer, 0.30 ºC in autumn, and 0.10 ºC in winter. The amount of change in temperature around the admission day was relatively large in spring and autumn compared to that in summer and winter. Similar ndings were observed for the atmospheric pressure. The mean change in atmospheric pressure from 4 days before admission to the day before admission was 0.55 hPa in spring, 0.39 hPa in summer, 0.63 hPa in autumn, and 0.47 hPa in winter (Table 3, Fig. 4).
Although the variation in atmospheric pressure before the admission day was lower in summer than in spring and autumn, the seasonal differences failed to reach statistical signi cance (p=0.147, Fig. 4).

Discussion
We conducted the present study based on previous literature suggesting the relevance of low atmospheric pressure to abdominal aortic aneurysm ruptures [6,7] . We applied a similar approach to the lymphatic system and its disorder, lymphedema, and hypothesized a relationship between climatic conditions and the occurrence of cellulitis in lymphedema, which is of relevance considering the unusual weather patterns in recent years, such as global warming. As a result, we found that cellulitis in patients with limb lymphedema is meteoropathic in nature, as in previous studies, including a systematic review [9,10] . Accordingly, patients with limb lymphedema may need to take special care to avoid developing cellulitis in summer or in seasons with lower atmospheric pressure.
In the present study, conducted in Japan, the number of admitted patients was about two times higher in summer than in other seasons ( Supplementary Fig. S2). Moreover, when the analysis was limited to patients living in Yokohama City (removing the in uence of geographical environment factors), the same tendency was observed; most of the 25 patients were admitted during the months and days with higher temperature and lower atmospheric pressure. Thus, severe cellulitis may be more apt to occur in patients with limb lymphedema on days with higher temperature or lower atmospheric pressure (e.g. summer in Japan).
In addition, the average change in the temperature or atmospheric pressure around the time of admission was higher in spring and autumn than in summer, at least when considering the period from 4 days before admission to the day before admission. However, because the present study utilized 15-day CMA data, the development of cellulitis might have been in uenced by climatic conditions during the prior 2 weeks.
Based on the abovementioned study results, there seems to be a seasonal occurrence of cellulitis in patients with limb lymphedema. A higher temperature or lower atmospheric pressure conceivably contributes to the development of cellulitis. Moreover, cellulitis in patients with limb lymphedema might be caused by a continuous higher temperature, rather than by daily and random changes in weather conditions.
Furthermore, if limb lymphedema is categorized as a meteoropathic disease, limb lymphedema conditions may be controlled by medication, as with migraine, rheumatoid arthritis, and bronchial asthma [11][12][13] . In fact, some previous studies have suggested that ketoprofen might be effective in secondary lymphedema [14][15][16] . Based on histopathology, the end stage of lymphedema shows brosis and fat deposition in the subcutaneous tissue [17][18][19][20] . Thus, lymphedema can be regarded as a type of chronic in ammatory disease that is meteoropathic in nature [14,21] .
The limitations of the present study are as follows. First, this study was a retrospective single-center study with only 25 cases. Second, this study did not include patients with cellulitis who were not admitted during the study period. Moreover, we could not evaluate cases with limb lymphedema that had not progressed to cellulitis during the study period. Third, the reasons for admission were varied; the only unanimous nding of the admitted patients was severe redness at the affected site. Thus, the evaluation criteria of admission were obscure. Fourth, data regarding the compliance of the patients in using elastic garments or bandages from the medical records was of poor quality. Given the high temperature in summer in Japan, many patients with limb lymphedema are apt to neglect the garments although they understand the importance of the garments in managing lymphedema. Furthermore, a previous study in Japan indicated that wearing garments in summer is a risk factor of cellulitis for lower limb lymphedema [9] . Additionally, the 15-day CMAs of the temperature and atmospheric pressure might not entirely represent the climatic conditions of the 2 weeks before admission; however, 15 days was de ned and hypothesized as a valid number of days by the authors. Prospective cohort and multi-institutional joint research with uni ed inclusion criteria for admission and consideration of the patient background, such as age, sex, and site of affected limbs, is desirable for the further elucidation of the meteoropathic nature of lymphedema.

Conclusions
Lymphedema may be regarded as a meteoropathic disease, as climatic conditions were shown to be associated with cellulitis in limbs affected by lymphedema. Severe cellulitis may be apt to occur in patients with limb lymphedema on days with higher temperature or lower atmospheric pressure.
Accordingly, patients with limb lymphedema may need to take extra precautions, or even medication, to avoid developing cellulitis in summer or in seasons with lower atmospheric pressure. Mean temperature and atmospheric pressure during the observation period.

Declarations
a. Monthly data. The atmospheric pressure gradually decreased from spring to summer and the temperature simultaneously increased. In contrast, as the temperatures fell from summer to winter, the atmospheric pressure increased.
b. Seasonal data. The difference in atmospheric pressure between autumn and winter was not as obvious as that in temperature.

Figure 3
Box-and-whisker plots of the meteorological data around the admission day for the 25 residents of Yokohama City, shown according to season. Since winter had a sample size of only one, statistical