- Email: [email protected]
Central and peripheral temperature changes in sheep following ovariectomy
Maturitas 46 (2003) 231 /238 www.elsevier.com/locate/maturitas
Central and peripheral temperature changes in sheep following ovariectomy J.M. MacLeay a,*, E. Lehmer b, R.M. Enns c, C. Mallinckrodt d, H.U. Bryant d, A.S. Turner a a
Department of Clinical Sciences, Colorado State University, Ft. Collins, CO 80523, USA b Department of Biology, Colorado State University, Ft. Collins, CO 80523, USA c Department of Animal Sciences, Colorado State University, Ft. Collins, CO 80523, USA d Lilly Research Laboratories, Lilly Corporate Center, Indianapolis, IN 46285, USA
Received 15 November 2002; received in revised form 8 April 2003; accepted 17 April 2003
Abstract Objectives: To determine if ovariectomized ewes undergo periodic body temperature rises (hot flashes) similar to women at menopause. Methods: Eighteen mature ewes were assigned to ovariectomy (OVX), ovariectomy/17ßestradiol implant (OVXE) or Sham. Electronic temperature loggers placed subcutaneously over the carotid artery and within the abdomen (core) and subcutaneously in the thigh and axilla (peripheral) were programmed to record body temperatures every 2.5 min for 59 days. Circadian rhythm changes were avoided by dividing readings into 1 h intervals. Hot flashes were defined as a 0.2 or 0.4 8C increase over the minimum temperature recorded for a 1 h interval for each sheep. Results: Logger placement did not reflect core and peripheral temperatures. The carotid and abdominal sites were most useful. The percentage of readings considered HF at the carotid site was 63% OVX, 54% OVXE and 37% Sham (P B/0.001), and at the abdominal site were 32% OVX, 15% OVE and 17% Sham (P B/0.001). When only the first 7 days after ovariectomy were analyzed, the percentage of readings considered to be HF at the carotid site was 75% OVX, 63% OVXE, and 49% Sham (P B/0.001), and at the abdominal site was 35% OVX, 15% OVXE and 17% Sham (P B/0.001). Conclusions: Ovariectomy in the ewe does illicit changes in body temperature compared with control ewes, which may be interpreted as HF. However, shifts in the circadian rhythm were not apparent. Estradiol treatment led to milder and less frequent HF. Periodic HF in species other than rats have heretofore not been reported. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Menopause; Hot flashes; Climacteric; Animal model
1. Introduction Hot flashes affect 50/80% of perimenopausal women and are the most characteristic symptom
associated with the climacteric [1]. Hot flashes in women are often preceded by an aura. The aura has typically been assumed to coincide with a transient downward resetting of the thermoregulatory centers of the hypothalamus. However, based on evidence provided by Freedman et al.,
* Corresponding author. 0378-5122/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0378-5122(03)00196-8
232
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
the idea that a hot flash is triggered by a downward resetting of the hypothalamic set point has been disputed. Instead, it is apparent that the thermoneutral zone is small or nonexistent in symptomatic women. In addition, radiotelemetry measurement of core body temperatures has identified that core body temperature elevations may precede the majority of hot flashes in women and may serve as a trigger of the body to cool the body core [2,3]. The body physiologically responds by attempting to cool the body core through sweating and peripheral vasodilation. This is interpreted by the woman as a feeling of heat and then chills as the body core cools. Peripheral body temperature as measured by finger temperature initially rises during the vasodilatory phase and subsequently body core temperature falls at the terminus of the hot flash as a result. The body core temperature then slowly returns to normal, presumably with resetting of the central thermoregulatory centers. A surge in luteinizing hormone (LH) may or may not accompany the hot flash. The accepted animal model for the hot flash associated with the female climacteric is either the ovariectomized or the morphine-dependent ovariectomized rat. During opiate withdrawal with the morphine antagonist naloxone, tail skin temperature rises coincident with a small drop in core temperature. The rise in skin temperature is also accompanied by a surge in LH and transient tachycardia. Chronic estrogen treatment suppresses the rise in tail skin temperature and associated hormonal changes. This model has been used to evaluate the estrogenic and antiestrogenic activities of partial estrogen receptor agonists and antagonists [4]. Others have measured tail skin temperature in female rats after ovariectomy (without opiate addiction) and found that they too experience transient rises in temperature, albeit less drastic ones, and have used this model to compare soy isoflavones (phytoestrogens) with oral micronized estradiol [1]. We have used the aged ovariectomized ewe as a model for a variety of perimenopausal conditions including skeletal and mandibular bone loss as well as bone and cardiovascular responses to selective estrogen receptor modulators (SERMs) [5 /9]. To further understand the pathophysiology
of temperature changes associated with menopause, we surgically implanted four electronic temperature sensitive data loggers into ovariectomized (OVX), OVX/estradiol (OVXE) treated and sham ewes. Loggers were placed in two general regions, the ‘core’ and the ‘periphery’.
2. Methods 2.1. Animals This study utilized 18 skeletally mature Rambouillet X Columbian ewes. Studies were conducted between August and October 2000. The sheep were randomly assigned, by computer, using the Unix utility, to three groups of six to make up the ovariectomized (OVX), ovariectomy/17b-estradiol implant (OVXE) and Sham and groups. 2.2. Reproductive hormones Blood was obtained via jugular venapuncture 2 weeks before ovariectomy, at ovariectomy and approximately every 2 weeks thereafter. Serum was isolated and frozen at 0 8C until analyzed. Serum LH, estradiol and progesterone were measured for each time point using a radio immunoassay that has been previously described [10,11]. 2.3. Ovariectomy Under general anesthesia, bilateral oophorectomy (OVX, 12 animals) or sham procedure (visual identification of the ovaries in 6 animals) was performed via ventral midline laparotomy. In six OVX sheep, an estradiol implant in the form of a capsule was placed under the skin in the inguinal region (OVXE). The capsule was prepared from silastic tubing, 0.335 cm inside diameter /0.465 cm outside diameter, (Dow Corning, Midland MI) and filled with pure crystalline 17b-estradiol. The length of the capsule was 5.0 cm. Implants maintained a fixed serum estradiol concentration of approximately 2 pg/ml, similar to that during the luteal phase of the estrus cycle [12,13].
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
2.4. Temperature loggers Four temperature sensitive data loggers (18.2 g, 3 cm diameter; Stow AwayTM TidbiT Onset, Pocasset, MA) were surgically implanted into each ewe. Loggers were placed in two general regions, the ‘‘core’’ and the ‘‘periphery’’. Core loggers were attached to the peritoneal surface of the rumen or abdominal wall and subcutaneously near the carotid artery in the mid-cervical region. Peripheral loggers were placed subcutaneously in the region of the axilla and proximal inner-thigh. Data loggers were programmed to record body temperatures to the nearest 0.01 8C every 2.5 min for 59 days and were synchronized at the beginning of the study. At the end of this period, loggers were removed at necropsy and data from the loggers was downloaded to a desktop computer using BOX CAR PRO† software (version 3.6, Onset). Proper calibration of the loggers to the nearest 0.00 8C was verified before implantation and after removal. This procedure involved placing the loggers in a small chamber of known and constant temperature and comparing the temperature recorded by the loggers with the ambient temperature of the chamber, recorded with a mercury thermometer. Calibration was also verified by placing loggers into a water bath of known temperature. All loggers showed accurate calibration within 0.05 8C prior to implantation and after removal. 2.5. Analysis of data All temperature loggers remained in the animals for at least 62 days. Because the actual start and finish dates varied among animals, statistical analysis was conducted only on data recorded between 10 October and 5 December 2000 (57 days). A second analysis utilizing only the first 7 days post ovariectomy was also performed to determine if temperature fluctuations were greater in the immediate post ovariectomy period. The percentage of time intervals in which hot flashes occurred was summarized by treatment group. To avoid complications that may be introduced due to body temperature fluctuations
233
related to normal circadian rhythms, temperature data was grouped by 1 h time intervals for a total of 24 intervals/day. For each sheep maximum and minimum body temperature was noted for each hour long interval. Two definitions of a hot flash were then evaluated. The first was an overall increase over minimum body temperature during the hour long time interval of 0.2 8C. The second alternative evaluated was an increase over minimum body temperature of 0.4 8C. Using these definitions of a hot flash, each time period for each sheep was coded as a 0 (no hot flash) or a 1 (hot flash). Circadian rhythms between groups were evaluated by comparing the overall mean temperature for each interval. All hypotheses were tested using two-sided level of significance of 0.05. Any temperature observation below 35 8C was coded as unknown. Preliminary analyses were conducted to evaluate normality assumptions and to evaluate validity of observations. Results suggested that temperature data were normally distributed. Two data sets were evaluated, the first included 57 days of data and the second only the first 7 days post ovariectomy. The effects of treatment on the incidence of hot flashes was assessed using Fisher’s Exact Test. The significance and magnitude of treatment group differences in body temperatures by time of day was assessed using a generalized least squares, restricted maximum likelihood-based fixed-effects model via PROC MIXED in SAS [14]. Analyses included the categorical, fixed effect of treatment group, time of day as 1 h increments, and the treatment /time of day interaction.
3. Results All animals recovered from surgeries uneventfully and no incisional complications were noted. Temperature loggers were not noted to cause any local irritation or inflammation. Interestingly, some loggers that were sutured to the peritoneal surface of the rumen sloughed into the rumenal lumen and were recovered in the reticulum. Some rumenal loggers were not recovered and it was assumed that these had passed through the intestinal tract and were lost. Loggers sutured to the
234
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
peritoneum along the abdominal wall were recovered without incident. Lost or sloughed loggers were not included in the data set. Results for LH, estradiol and progesterone are summarized in Fig. 1a/c, Table 1. LH increased markedly in the face of ovariectomy and was ameliorated by estrogen supplementation as was expected. Estradiol concentrations were greatest in the estrogen supplemented group and were lowest in the ovariectomized group. Progesterone levels
Table 1 Overall mean serum reproductive hormone concentrations including data from day 14 through day 56 LH
SHAM OVX OVXE
Estradiola
Progesterone
Mean
S.D.
Mean
S.D.
Mean
S.D.
0.66 2.06 0.69
1.24 1.80 0.68
3.04 2.80 4.09
2.31 1.69 3.72
0.71 0.56 0.33
0.75 0.98 0.68
Mean of six data points/sheep and six sheep per treatment group. Sham, control sheep; OVX, ovariectomized; OVXE, ovariectomized sheep supplemented with subcutaneous estradiol. a Estradiol outliers omitted (three points in OVX group).
were greatest in the SHAM group and were lowest in the estrogen supplemented group.
3.1. Body temperature
Fig. 1. Graphical representation of the mean serum LH, Estradiol and Progesterone concentration for each group of sheep (n /6 sheep/treatment group). Treatment groups are ovariectomy (OVX), ovariectomy with estradiol implant (OVXE) and sham. Day 0 is the day of surgery.
Correlations between temperatures from various locations were moderate in magnitude. Given the large sample sizes, all correlations were significantly different from zero. The strongest correlation was between axilla and carotid temperature (r /0.56) and the weakest correlation was between abdominal temperature and thigh temperature (r /0.38). Therefore, we could not conclude that any particular site was more representative of peripheral temperature than core or vice versa. Details of results are given in Table 2 and a review of tests of significance for differences between all pairs of treatment groups for both definitions of a hot flash are presented in Table 3. When temperature data from all time points was used and a hot flash was based on either a 0.2 or a 0.4 8C increase in minimum temperature within any given interval for each sheep, the OVX treatment group exhibited the greatest percentage of hot flashes most of the time. Of all temperature logger placement sites, the carotid and intraabdominal sites appeared to be the most useful and the most consistent. For the carotid site, the OVX group had the greatest percentage of hot flashes followed by OVXE and subsequently by the SHAM treatment groups. For the abdominal site the OVX group had the greatest percentage of
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
235
Table 2 Percentage of readings that may be considered a hot flash for different areas of temperature logger placement for the three treatment groups (OVX, ovariectomy, OVXE, ovariectomy plus estradiol replacement, SHAM, control; n , six sheep/group) Differential 0.2 8C, 57 days of readings Carotid N OVX OVXE SHAM
%
Axilla
N
N
%
8301 87.5 3562 37.6 7004 86.2 1402 17.2 4502 66.5 1238 18.3 Differential 0.2 8C, 7 days of Carotid
OVX OVXE SHAM
Abdomen
N 1022 818 545
% 86.9 81.2 64.9
Thigh %
8143 85.9 6948 85.5 5922 87.4 readings
Abdomen
Axilla
N 413 149 143
N 1036 853 696
% 35.1 14.8 17.0
Differential 0.4 8C, 57 days of readings Carotid
% 88.1 84.6 82.9
Axilla
N
N
N
%
N
7948 5846 5393
83.8 71.9 79.6
5992 63.2 3011 31.7 4424 54.4 1234 15.2 2537 37.4 1136 16.8 Differential 0.4 8C, 7 days of
Thigh N 929 745 768
%
Abdomen
Carotid % 79.0 73.9 91.4
N 881 639 409
% 74.9 63.4 48.7
%
%
7479 78.9 6198 76.2 4878 72.0 readings
Abdomen
Axilla
N 413 149 143
N 987 826 637
% 35.1 14.8 17.0
Thigh N
%
6518 5508 4802
68.7 67.7 70.9
Thigh % 83.9 81.9 75.8
N 855 729 698
% 72.7 72.3 83.1
Comparisons between groups for each body site were made using Fisher’s Exact Test P value two-tail. All values were significant at P B/0.001. Each sheep had a temperature logger at each location. All loggers registered body temperature every 2.5 min for 57 days. Temperatures were analyzed by time of day as 1 h intervals. An increase in body temperature of either 0.2 or 0.4 8C over the minimum temperature recorded for each interval for each sheep was considered a hot flash and the total number of readings (N ) and the percentage of readings (%) meeting this criteria is reported in the table below.
hot flashes followed by similar rates of hot flashes within the OVXE and SHAM groups. Regardless of the differential used, ovariectomized sheep had the highest percentage of readings considered hot flashes from loggers placed at the
carotid and abdominal sites. While there was statistical significance between groups at the thigh and axillary sites, the percentage of readings fulfilling the criteria for a hot flash did not appear to be discriminating. Therefore, we concluded that
Table 3 A review of tests of significance for differences between all pairs of treatment groups and for each logger placement site for both definitions of a hot flash and for 57 days and the first 7 days post surgery Differential 0.2 8C, 57 days of readings OVX vs. OVXE Carotid Abdomen Axilla Thigh
Carotid Abdomen Axilla Thigh
OVX vs. SHAM
Differential 0.4 8C, 57 days of readings OVXE vs. SHAM
OVX vs. OVXE
OVX vs. SHAM
OVXE vs. SHAM
B/0.001 B/0.001 B/0.001 B/0.001 0.020 B/0.001 0.005 B/0.001 Differential 0.2 8C, 7 days of readings
B/0.001 0.200 0.311 B/0.001
B/0.001 B/0.001 B/0.001 B/0.001 0.230 B/0.001 0.848 B/0.001 Differential 0.4 8C, 7 days of readings
B/0.001 0.200 0.002 B/0.001
OVX vs. OVXE
OVX vs. SHAM
OVXE vs. SHAM
OVX vs. OVXE
OVX vs. SHAM
OVXE vs. SHAM
B/0.001 B/0.001 0.020 0.005
B/0.001 B/0.001 B/0.001 B/0.001
B/0.001 0.200 0.311 B/0.001
B/0.001 B/0.001 0.230 0.848
B/0.001 B/0.001 B/0.001 B/0.001
B/0.001 0.200 0.002 B/0.001
OVX, ovariectomy; OVXE, ovariectomy plus estradiol replacement; SHAM, control; n , six sheep/group. Comparisons between groups for each body site were made using Fisher’s Exact Test P value two-tail. P values are given in the chart.
236
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
based on both definitions of a hot flash and analysis of data from the entire time period and data from only the first week of collection, the most promising temperature locations for identifying a hot flash were the carotid and abdomen. In addition, the first 7 days of readings appeared to be highly discriminating between groups. Over the course of a day temperature readings at specific body locations were generally higher from 15:00 to 18:00 h. Additionally, the OVXE group tended to have overall higher temperatures but the circadian rhythm appeared similar between groups. Fig. 2.
4. Discussion Given the very large data set acquired from the temperature loggers many differences were found to be significant. However, emphasis should be placed on those differences that are considered biologically important. Therefore, we propose use of the intra-abdominal and carotid sites for temperature measurement as the most useful and most discriminating for future studies. In addition,
analysis of both the entire 57 day data set and the 7 day data set appeared to be discriminating between groups. Therefore, this model may prove useful in both short and long term studies as sheep in this study demonstrated changes in temperature in the immediate post ovariectomy period similar to that seen in women after surgical menopause [15]. The placement of temperature loggers at the carotid, abdominal, thigh and axilla was intended to determine if core and peripheral temperatures could be obtained and compared. However, correlations between placement sites did not support that true measures of peripheral and core temperatures were being made. The carotid site and the abdominal site appeared to be most useful with the abdominal site likely being most representative of core body temperature. Further studies will be needed to determine if a true and accurate peripheral or skin temperature log can be made using this animal model. However, as core body temperature rises have been documented in women experiencing hot flashes [3], it is possible that this model may be useful despite lack of skin temperature monitoring.
Fig. 2. Graphical representation of the mean hourly temperature for all data points over the 57 day trial period for temperature loggers placed over the carotid artery, for each hourly increment throughout a 24 h period starting at midnight. Six sheep/group, OVX, ovariectomy, OVXE, ovariectomy/Estradiol supplementation, SHAM, control.
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
Circadian rhythms between groups did not appear different or shifted. However, at multiple time points differences between groups were significant. The ovariectomized sheep supplemented with estrogen had a higher overall mean temperature than the other groups during many intervals. An explanation for this finding is not forthcoming. Certainly, further analysis of the significance of the circadian rhythm and the effect of estrogen supplementation, ovariectomy and time of day will need to be investigated in greater depth in future studies. The convenience of the commonly used rat hot flash model cannot be disputed. In the rat model, elevation of tail skin temperature ranges from 2.8 8C in the OVX model without opiates [1] to 6 8C in the opiate-addicted model [4]. Certainly in the proposed ewe model reported here, more subtle changes in temperature were noted in comparison to the rat model. In addition, more sophisticated equipment to capture body temperature and modeling of data was necessary. Neverthe-less, introduction of a large animal model to study body temperature changes in the face of estrogen depletion may prove to be an invaluable tool for researchers. We performed this experiment when the sheep were in the estrous period of the year, where circulating levels of estradiol were highest. We hypothesized that with these levels of estradiol, estrogen-deficient animals would exhibit subtle changes in subcutaneous and core temperature monitored by the temperature loggers. Future studies conducted during other times of year may answer questions about whether day length plays a role in the manifestation of body temperature changes related to estrogen depletion in this species. In this animal model, serum LH levels provided a good index of seasonal changes in reproductive neuroendocrine activity [16]. Typically, LH is high during the breeding season and then decreases more than 20-fold to undetectable levels as animals become anestrus. Although, it is unlikely that LH is responsible for the temperature changes, future studies using this model may also be directed at more frequent monitoring of the temperature in concert with LH concentration
237
rhythms to determine the relationship between these variables.
Acknowledgements The authors would like to that Lilly Research Laboratories for funding this study.
References [1] Pan Y, Anthony M, Binns M, Clarkson T. A comparison of oral micronized estradiol with soy phytoestrogen effects on tail skin temperatures of ovariectomized rats. Menopause 2001;8(3):171 /4. [2] Freedman RR. Physiology of hot flashes. Am J Hum Biol 2001;13(4):453 /64. [3] Freedman RR, Woodward S. Core body temperature during menopausal hot flushes. Fertil Steril 1996;65(6):1141 /4. [4] Merchenthaler I, Funkhouser J, Carver J, Lundeen S, Ghosh K, Winneker R. The effect of estrogens and antiestrogens in a rat model for hot flush. Maturitas 1998;30:307 /16. [5] MacLeay J, Les C, Toth C, Wheeler D, Enns M, Turner A. Effect of a diet low in cation-anion balance on bone mineral density in mature ovariectomized ewes. J Vet Int Med 2002;16(3):338. [6] Turner A. The sheep as a model for osteoporosis in humans. Vet Journal 2002;163:1 /8. [7] Turner A, Hannan M, Bouxsein M, Kaufman L, Mallinckrodt C, Bryant H. Effects of Raloxifene (LY139481) on bone mineral density and bone strength in aged, ovariectomized ewes. In: Proc Orthop Res Soc, Anaheim, CA, 1999. [8] Gaynor J, Monnet E, Selzman C, Parker D, Kaufman L, Bryant H. The effect of raloxifene on coronary arteries in aged ovariectomized ewes. J Vet Pharmacol Therap 2000;23:175 /9. [9] Turner A, Alvis M, Mallinckrodt C, Bryant H. Doseresponse effects of estradiol on bone mineral density in ovariectomized ewes. Bone 1995;17S:421S /7S. [10] Niswender G. Influence of site of conjugation on specificity of antibodies to progesterone. Steroids 1973;22:413 /24. [11] Niswender G, Reichert L, Midgley A, Nalbandov A. Radioimmunoassay for ovine and bovine luteinizing hormone. Endocrinology 1969;84:1166 /73. [12] Karsch F, Legen S, Ryan K, Foster D. Importance of estradiol and progesterone in regulating LH secretion and estrus behavior during the sheep estrus cycle. Biol Reprod 1980;23:404 /13. [13] Karsch F, Dierschke D, Weick R. Positive and negative feedback control by estrogen of luteinizing hormone
238
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
secretion in the rhesus monkey. Endocrinology 1997;92:799 /804. [14] SAS procedures guide, version 8. 1999, Cary, NC: SAS Institute, Inc. [15] Tulandi T, Lal S. Menopausal hot flush. Obstet Gynecol Surv 1985;40(9):553 /63.
[16] Thrun L, Dahl G, Evans N, Karsch F. A critical period for thyroid hormone action on seasonal changes in reproductive neuroendocrine function in the ewe. Endocrinology 1997;138:3402 /9.
Central and peripheral temperature changes in sheep following ovariectomy J.M. MacLeay a,*, E. Lehmer b, R.M. Enns c, C. Mallinckrodt d, H.U. Bryant d, A.S. Turner a a
Department of Clinical Sciences, Colorado State University, Ft. Collins, CO 80523, USA b Department of Biology, Colorado State University, Ft. Collins, CO 80523, USA c Department of Animal Sciences, Colorado State University, Ft. Collins, CO 80523, USA d Lilly Research Laboratories, Lilly Corporate Center, Indianapolis, IN 46285, USA
Received 15 November 2002; received in revised form 8 April 2003; accepted 17 April 2003
Abstract Objectives: To determine if ovariectomized ewes undergo periodic body temperature rises (hot flashes) similar to women at menopause. Methods: Eighteen mature ewes were assigned to ovariectomy (OVX), ovariectomy/17ßestradiol implant (OVXE) or Sham. Electronic temperature loggers placed subcutaneously over the carotid artery and within the abdomen (core) and subcutaneously in the thigh and axilla (peripheral) were programmed to record body temperatures every 2.5 min for 59 days. Circadian rhythm changes were avoided by dividing readings into 1 h intervals. Hot flashes were defined as a 0.2 or 0.4 8C increase over the minimum temperature recorded for a 1 h interval for each sheep. Results: Logger placement did not reflect core and peripheral temperatures. The carotid and abdominal sites were most useful. The percentage of readings considered HF at the carotid site was 63% OVX, 54% OVXE and 37% Sham (P B/0.001), and at the abdominal site were 32% OVX, 15% OVE and 17% Sham (P B/0.001). When only the first 7 days after ovariectomy were analyzed, the percentage of readings considered to be HF at the carotid site was 75% OVX, 63% OVXE, and 49% Sham (P B/0.001), and at the abdominal site was 35% OVX, 15% OVXE and 17% Sham (P B/0.001). Conclusions: Ovariectomy in the ewe does illicit changes in body temperature compared with control ewes, which may be interpreted as HF. However, shifts in the circadian rhythm were not apparent. Estradiol treatment led to milder and less frequent HF. Periodic HF in species other than rats have heretofore not been reported. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Menopause; Hot flashes; Climacteric; Animal model
1. Introduction Hot flashes affect 50/80% of perimenopausal women and are the most characteristic symptom
associated with the climacteric [1]. Hot flashes in women are often preceded by an aura. The aura has typically been assumed to coincide with a transient downward resetting of the thermoregulatory centers of the hypothalamus. However, based on evidence provided by Freedman et al.,
* Corresponding author. 0378-5122/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0378-5122(03)00196-8
232
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
the idea that a hot flash is triggered by a downward resetting of the hypothalamic set point has been disputed. Instead, it is apparent that the thermoneutral zone is small or nonexistent in symptomatic women. In addition, radiotelemetry measurement of core body temperatures has identified that core body temperature elevations may precede the majority of hot flashes in women and may serve as a trigger of the body to cool the body core [2,3]. The body physiologically responds by attempting to cool the body core through sweating and peripheral vasodilation. This is interpreted by the woman as a feeling of heat and then chills as the body core cools. Peripheral body temperature as measured by finger temperature initially rises during the vasodilatory phase and subsequently body core temperature falls at the terminus of the hot flash as a result. The body core temperature then slowly returns to normal, presumably with resetting of the central thermoregulatory centers. A surge in luteinizing hormone (LH) may or may not accompany the hot flash. The accepted animal model for the hot flash associated with the female climacteric is either the ovariectomized or the morphine-dependent ovariectomized rat. During opiate withdrawal with the morphine antagonist naloxone, tail skin temperature rises coincident with a small drop in core temperature. The rise in skin temperature is also accompanied by a surge in LH and transient tachycardia. Chronic estrogen treatment suppresses the rise in tail skin temperature and associated hormonal changes. This model has been used to evaluate the estrogenic and antiestrogenic activities of partial estrogen receptor agonists and antagonists [4]. Others have measured tail skin temperature in female rats after ovariectomy (without opiate addiction) and found that they too experience transient rises in temperature, albeit less drastic ones, and have used this model to compare soy isoflavones (phytoestrogens) with oral micronized estradiol [1]. We have used the aged ovariectomized ewe as a model for a variety of perimenopausal conditions including skeletal and mandibular bone loss as well as bone and cardiovascular responses to selective estrogen receptor modulators (SERMs) [5 /9]. To further understand the pathophysiology
of temperature changes associated with menopause, we surgically implanted four electronic temperature sensitive data loggers into ovariectomized (OVX), OVX/estradiol (OVXE) treated and sham ewes. Loggers were placed in two general regions, the ‘core’ and the ‘periphery’.
2. Methods 2.1. Animals This study utilized 18 skeletally mature Rambouillet X Columbian ewes. Studies were conducted between August and October 2000. The sheep were randomly assigned, by computer, using the Unix utility, to three groups of six to make up the ovariectomized (OVX), ovariectomy/17b-estradiol implant (OVXE) and Sham and groups. 2.2. Reproductive hormones Blood was obtained via jugular venapuncture 2 weeks before ovariectomy, at ovariectomy and approximately every 2 weeks thereafter. Serum was isolated and frozen at 0 8C until analyzed. Serum LH, estradiol and progesterone were measured for each time point using a radio immunoassay that has been previously described [10,11]. 2.3. Ovariectomy Under general anesthesia, bilateral oophorectomy (OVX, 12 animals) or sham procedure (visual identification of the ovaries in 6 animals) was performed via ventral midline laparotomy. In six OVX sheep, an estradiol implant in the form of a capsule was placed under the skin in the inguinal region (OVXE). The capsule was prepared from silastic tubing, 0.335 cm inside diameter /0.465 cm outside diameter, (Dow Corning, Midland MI) and filled with pure crystalline 17b-estradiol. The length of the capsule was 5.0 cm. Implants maintained a fixed serum estradiol concentration of approximately 2 pg/ml, similar to that during the luteal phase of the estrus cycle [12,13].
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
2.4. Temperature loggers Four temperature sensitive data loggers (18.2 g, 3 cm diameter; Stow AwayTM TidbiT Onset, Pocasset, MA) were surgically implanted into each ewe. Loggers were placed in two general regions, the ‘‘core’’ and the ‘‘periphery’’. Core loggers were attached to the peritoneal surface of the rumen or abdominal wall and subcutaneously near the carotid artery in the mid-cervical region. Peripheral loggers were placed subcutaneously in the region of the axilla and proximal inner-thigh. Data loggers were programmed to record body temperatures to the nearest 0.01 8C every 2.5 min for 59 days and were synchronized at the beginning of the study. At the end of this period, loggers were removed at necropsy and data from the loggers was downloaded to a desktop computer using BOX CAR PRO† software (version 3.6, Onset). Proper calibration of the loggers to the nearest 0.00 8C was verified before implantation and after removal. This procedure involved placing the loggers in a small chamber of known and constant temperature and comparing the temperature recorded by the loggers with the ambient temperature of the chamber, recorded with a mercury thermometer. Calibration was also verified by placing loggers into a water bath of known temperature. All loggers showed accurate calibration within 0.05 8C prior to implantation and after removal. 2.5. Analysis of data All temperature loggers remained in the animals for at least 62 days. Because the actual start and finish dates varied among animals, statistical analysis was conducted only on data recorded between 10 October and 5 December 2000 (57 days). A second analysis utilizing only the first 7 days post ovariectomy was also performed to determine if temperature fluctuations were greater in the immediate post ovariectomy period. The percentage of time intervals in which hot flashes occurred was summarized by treatment group. To avoid complications that may be introduced due to body temperature fluctuations
233
related to normal circadian rhythms, temperature data was grouped by 1 h time intervals for a total of 24 intervals/day. For each sheep maximum and minimum body temperature was noted for each hour long interval. Two definitions of a hot flash were then evaluated. The first was an overall increase over minimum body temperature during the hour long time interval of 0.2 8C. The second alternative evaluated was an increase over minimum body temperature of 0.4 8C. Using these definitions of a hot flash, each time period for each sheep was coded as a 0 (no hot flash) or a 1 (hot flash). Circadian rhythms between groups were evaluated by comparing the overall mean temperature for each interval. All hypotheses were tested using two-sided level of significance of 0.05. Any temperature observation below 35 8C was coded as unknown. Preliminary analyses were conducted to evaluate normality assumptions and to evaluate validity of observations. Results suggested that temperature data were normally distributed. Two data sets were evaluated, the first included 57 days of data and the second only the first 7 days post ovariectomy. The effects of treatment on the incidence of hot flashes was assessed using Fisher’s Exact Test. The significance and magnitude of treatment group differences in body temperatures by time of day was assessed using a generalized least squares, restricted maximum likelihood-based fixed-effects model via PROC MIXED in SAS [14]. Analyses included the categorical, fixed effect of treatment group, time of day as 1 h increments, and the treatment /time of day interaction.
3. Results All animals recovered from surgeries uneventfully and no incisional complications were noted. Temperature loggers were not noted to cause any local irritation or inflammation. Interestingly, some loggers that were sutured to the peritoneal surface of the rumen sloughed into the rumenal lumen and were recovered in the reticulum. Some rumenal loggers were not recovered and it was assumed that these had passed through the intestinal tract and were lost. Loggers sutured to the
234
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
peritoneum along the abdominal wall were recovered without incident. Lost or sloughed loggers were not included in the data set. Results for LH, estradiol and progesterone are summarized in Fig. 1a/c, Table 1. LH increased markedly in the face of ovariectomy and was ameliorated by estrogen supplementation as was expected. Estradiol concentrations were greatest in the estrogen supplemented group and were lowest in the ovariectomized group. Progesterone levels
Table 1 Overall mean serum reproductive hormone concentrations including data from day 14 through day 56 LH
SHAM OVX OVXE
Estradiola
Progesterone
Mean
S.D.
Mean
S.D.
Mean
S.D.
0.66 2.06 0.69
1.24 1.80 0.68
3.04 2.80 4.09
2.31 1.69 3.72
0.71 0.56 0.33
0.75 0.98 0.68
Mean of six data points/sheep and six sheep per treatment group. Sham, control sheep; OVX, ovariectomized; OVXE, ovariectomized sheep supplemented with subcutaneous estradiol. a Estradiol outliers omitted (three points in OVX group).
were greatest in the SHAM group and were lowest in the estrogen supplemented group.
3.1. Body temperature
Fig. 1. Graphical representation of the mean serum LH, Estradiol and Progesterone concentration for each group of sheep (n /6 sheep/treatment group). Treatment groups are ovariectomy (OVX), ovariectomy with estradiol implant (OVXE) and sham. Day 0 is the day of surgery.
Correlations between temperatures from various locations were moderate in magnitude. Given the large sample sizes, all correlations were significantly different from zero. The strongest correlation was between axilla and carotid temperature (r /0.56) and the weakest correlation was between abdominal temperature and thigh temperature (r /0.38). Therefore, we could not conclude that any particular site was more representative of peripheral temperature than core or vice versa. Details of results are given in Table 2 and a review of tests of significance for differences between all pairs of treatment groups for both definitions of a hot flash are presented in Table 3. When temperature data from all time points was used and a hot flash was based on either a 0.2 or a 0.4 8C increase in minimum temperature within any given interval for each sheep, the OVX treatment group exhibited the greatest percentage of hot flashes most of the time. Of all temperature logger placement sites, the carotid and intraabdominal sites appeared to be the most useful and the most consistent. For the carotid site, the OVX group had the greatest percentage of hot flashes followed by OVXE and subsequently by the SHAM treatment groups. For the abdominal site the OVX group had the greatest percentage of
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
235
Table 2 Percentage of readings that may be considered a hot flash for different areas of temperature logger placement for the three treatment groups (OVX, ovariectomy, OVXE, ovariectomy plus estradiol replacement, SHAM, control; n , six sheep/group) Differential 0.2 8C, 57 days of readings Carotid N OVX OVXE SHAM
%
Axilla
N
N
%
8301 87.5 3562 37.6 7004 86.2 1402 17.2 4502 66.5 1238 18.3 Differential 0.2 8C, 7 days of Carotid
OVX OVXE SHAM
Abdomen
N 1022 818 545
% 86.9 81.2 64.9
Thigh %
8143 85.9 6948 85.5 5922 87.4 readings
Abdomen
Axilla
N 413 149 143
N 1036 853 696
% 35.1 14.8 17.0
Differential 0.4 8C, 57 days of readings Carotid
% 88.1 84.6 82.9
Axilla
N
N
N
%
N
7948 5846 5393
83.8 71.9 79.6
5992 63.2 3011 31.7 4424 54.4 1234 15.2 2537 37.4 1136 16.8 Differential 0.4 8C, 7 days of
Thigh N 929 745 768
%
Abdomen
Carotid % 79.0 73.9 91.4
N 881 639 409
% 74.9 63.4 48.7
%
%
7479 78.9 6198 76.2 4878 72.0 readings
Abdomen
Axilla
N 413 149 143
N 987 826 637
% 35.1 14.8 17.0
Thigh N
%
6518 5508 4802
68.7 67.7 70.9
Thigh % 83.9 81.9 75.8
N 855 729 698
% 72.7 72.3 83.1
Comparisons between groups for each body site were made using Fisher’s Exact Test P value two-tail. All values were significant at P B/0.001. Each sheep had a temperature logger at each location. All loggers registered body temperature every 2.5 min for 57 days. Temperatures were analyzed by time of day as 1 h intervals. An increase in body temperature of either 0.2 or 0.4 8C over the minimum temperature recorded for each interval for each sheep was considered a hot flash and the total number of readings (N ) and the percentage of readings (%) meeting this criteria is reported in the table below.
hot flashes followed by similar rates of hot flashes within the OVXE and SHAM groups. Regardless of the differential used, ovariectomized sheep had the highest percentage of readings considered hot flashes from loggers placed at the
carotid and abdominal sites. While there was statistical significance between groups at the thigh and axillary sites, the percentage of readings fulfilling the criteria for a hot flash did not appear to be discriminating. Therefore, we concluded that
Table 3 A review of tests of significance for differences between all pairs of treatment groups and for each logger placement site for both definitions of a hot flash and for 57 days and the first 7 days post surgery Differential 0.2 8C, 57 days of readings OVX vs. OVXE Carotid Abdomen Axilla Thigh
Carotid Abdomen Axilla Thigh
OVX vs. SHAM
Differential 0.4 8C, 57 days of readings OVXE vs. SHAM
OVX vs. OVXE
OVX vs. SHAM
OVXE vs. SHAM
B/0.001 B/0.001 B/0.001 B/0.001 0.020 B/0.001 0.005 B/0.001 Differential 0.2 8C, 7 days of readings
B/0.001 0.200 0.311 B/0.001
B/0.001 B/0.001 B/0.001 B/0.001 0.230 B/0.001 0.848 B/0.001 Differential 0.4 8C, 7 days of readings
B/0.001 0.200 0.002 B/0.001
OVX vs. OVXE
OVX vs. SHAM
OVXE vs. SHAM
OVX vs. OVXE
OVX vs. SHAM
OVXE vs. SHAM
B/0.001 B/0.001 0.020 0.005
B/0.001 B/0.001 B/0.001 B/0.001
B/0.001 0.200 0.311 B/0.001
B/0.001 B/0.001 0.230 0.848
B/0.001 B/0.001 B/0.001 B/0.001
B/0.001 0.200 0.002 B/0.001
OVX, ovariectomy; OVXE, ovariectomy plus estradiol replacement; SHAM, control; n , six sheep/group. Comparisons between groups for each body site were made using Fisher’s Exact Test P value two-tail. P values are given in the chart.
236
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
based on both definitions of a hot flash and analysis of data from the entire time period and data from only the first week of collection, the most promising temperature locations for identifying a hot flash were the carotid and abdomen. In addition, the first 7 days of readings appeared to be highly discriminating between groups. Over the course of a day temperature readings at specific body locations were generally higher from 15:00 to 18:00 h. Additionally, the OVXE group tended to have overall higher temperatures but the circadian rhythm appeared similar between groups. Fig. 2.
4. Discussion Given the very large data set acquired from the temperature loggers many differences were found to be significant. However, emphasis should be placed on those differences that are considered biologically important. Therefore, we propose use of the intra-abdominal and carotid sites for temperature measurement as the most useful and most discriminating for future studies. In addition,
analysis of both the entire 57 day data set and the 7 day data set appeared to be discriminating between groups. Therefore, this model may prove useful in both short and long term studies as sheep in this study demonstrated changes in temperature in the immediate post ovariectomy period similar to that seen in women after surgical menopause [15]. The placement of temperature loggers at the carotid, abdominal, thigh and axilla was intended to determine if core and peripheral temperatures could be obtained and compared. However, correlations between placement sites did not support that true measures of peripheral and core temperatures were being made. The carotid site and the abdominal site appeared to be most useful with the abdominal site likely being most representative of core body temperature. Further studies will be needed to determine if a true and accurate peripheral or skin temperature log can be made using this animal model. However, as core body temperature rises have been documented in women experiencing hot flashes [3], it is possible that this model may be useful despite lack of skin temperature monitoring.
Fig. 2. Graphical representation of the mean hourly temperature for all data points over the 57 day trial period for temperature loggers placed over the carotid artery, for each hourly increment throughout a 24 h period starting at midnight. Six sheep/group, OVX, ovariectomy, OVXE, ovariectomy/Estradiol supplementation, SHAM, control.
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
Circadian rhythms between groups did not appear different or shifted. However, at multiple time points differences between groups were significant. The ovariectomized sheep supplemented with estrogen had a higher overall mean temperature than the other groups during many intervals. An explanation for this finding is not forthcoming. Certainly, further analysis of the significance of the circadian rhythm and the effect of estrogen supplementation, ovariectomy and time of day will need to be investigated in greater depth in future studies. The convenience of the commonly used rat hot flash model cannot be disputed. In the rat model, elevation of tail skin temperature ranges from 2.8 8C in the OVX model without opiates [1] to 6 8C in the opiate-addicted model [4]. Certainly in the proposed ewe model reported here, more subtle changes in temperature were noted in comparison to the rat model. In addition, more sophisticated equipment to capture body temperature and modeling of data was necessary. Neverthe-less, introduction of a large animal model to study body temperature changes in the face of estrogen depletion may prove to be an invaluable tool for researchers. We performed this experiment when the sheep were in the estrous period of the year, where circulating levels of estradiol were highest. We hypothesized that with these levels of estradiol, estrogen-deficient animals would exhibit subtle changes in subcutaneous and core temperature monitored by the temperature loggers. Future studies conducted during other times of year may answer questions about whether day length plays a role in the manifestation of body temperature changes related to estrogen depletion in this species. In this animal model, serum LH levels provided a good index of seasonal changes in reproductive neuroendocrine activity [16]. Typically, LH is high during the breeding season and then decreases more than 20-fold to undetectable levels as animals become anestrus. Although, it is unlikely that LH is responsible for the temperature changes, future studies using this model may also be directed at more frequent monitoring of the temperature in concert with LH concentration
237
rhythms to determine the relationship between these variables.
Acknowledgements The authors would like to that Lilly Research Laboratories for funding this study.
References [1] Pan Y, Anthony M, Binns M, Clarkson T. A comparison of oral micronized estradiol with soy phytoestrogen effects on tail skin temperatures of ovariectomized rats. Menopause 2001;8(3):171 /4. [2] Freedman RR. Physiology of hot flashes. Am J Hum Biol 2001;13(4):453 /64. [3] Freedman RR, Woodward S. Core body temperature during menopausal hot flushes. Fertil Steril 1996;65(6):1141 /4. [4] Merchenthaler I, Funkhouser J, Carver J, Lundeen S, Ghosh K, Winneker R. The effect of estrogens and antiestrogens in a rat model for hot flush. Maturitas 1998;30:307 /16. [5] MacLeay J, Les C, Toth C, Wheeler D, Enns M, Turner A. Effect of a diet low in cation-anion balance on bone mineral density in mature ovariectomized ewes. J Vet Int Med 2002;16(3):338. [6] Turner A. The sheep as a model for osteoporosis in humans. Vet Journal 2002;163:1 /8. [7] Turner A, Hannan M, Bouxsein M, Kaufman L, Mallinckrodt C, Bryant H. Effects of Raloxifene (LY139481) on bone mineral density and bone strength in aged, ovariectomized ewes. In: Proc Orthop Res Soc, Anaheim, CA, 1999. [8] Gaynor J, Monnet E, Selzman C, Parker D, Kaufman L, Bryant H. The effect of raloxifene on coronary arteries in aged ovariectomized ewes. J Vet Pharmacol Therap 2000;23:175 /9. [9] Turner A, Alvis M, Mallinckrodt C, Bryant H. Doseresponse effects of estradiol on bone mineral density in ovariectomized ewes. Bone 1995;17S:421S /7S. [10] Niswender G. Influence of site of conjugation on specificity of antibodies to progesterone. Steroids 1973;22:413 /24. [11] Niswender G, Reichert L, Midgley A, Nalbandov A. Radioimmunoassay for ovine and bovine luteinizing hormone. Endocrinology 1969;84:1166 /73. [12] Karsch F, Legen S, Ryan K, Foster D. Importance of estradiol and progesterone in regulating LH secretion and estrus behavior during the sheep estrus cycle. Biol Reprod 1980;23:404 /13. [13] Karsch F, Dierschke D, Weick R. Positive and negative feedback control by estrogen of luteinizing hormone
238
J.M. MacLeay et al. / Maturitas 46 (2003) 231 /238
secretion in the rhesus monkey. Endocrinology 1997;92:799 /804. [14] SAS procedures guide, version 8. 1999, Cary, NC: SAS Institute, Inc. [15] Tulandi T, Lal S. Menopausal hot flush. Obstet Gynecol Surv 1985;40(9):553 /63.
[16] Thrun L, Dahl G, Evans N, Karsch F. A critical period for thyroid hormone action on seasonal changes in reproductive neuroendocrine function in the ewe. Endocrinology 1997;138:3402 /9.