INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IN ADDITION TO SUBSERVING important thermoregulatory reflexes, human cutaneous vascular control is subject to modification by other factors, including, in women, reproductive hormone status. Progesterone and estrogen have been shown to alter the thermoregulatory control of skin blood flow in hyperthermia (3, 7, 8, 22). When progesterone and estrogen are elevated, either in the luteal phase of the menstrual cycle (7, 8, 22) or exogenously by oral contraceptives (3), a higher internal temperature (threshold) is required to initiate cutaneous vasodilation during hyperthermia (+0.3 to 0.5°C). The effect of this threshold shift on skin blood flow can be striking considering the sensitivity of skin blood flow to increments in internal temperature (10) and the maximum capacity of the skin for blood flow of 6-8 l/min (18).

The reflex control of the human cutaneous circulation involves two branches of the sympathetic nervous system: the adrenergic vasoconstrictor system and the nonadrenergic cutaneous active vasodilator system. This latter system is not tonically active but is responsible for most of the cutaneous vasodilation observed during hyperthermia in humans (10). We recently found that progesterone and estrogen shifted the control of the cutaneous active vasodilator system to higher internal temperatures (3).

Elevated female reproductive hormones also cause increases in resting body temperature and shift the thermoregulatory control of sweating and heart rate to higher internal temperatures during hyperthermia (3, 5, 7, 22). That the control of skin blood flow, sweating, and heart rate (HR) are similarly shifted suggests a central shift in thermoregulatory control (3). This influence on thermoregulatory control is likely to be predominantly due to progesterone, which causes increases in temperature when administered alone (9, 15, 17). Estrogen alone has a depressant effect on body temperature (9) and has been shown to shift the control of cutaneous vasodilation and sweating to lower body temperatures (1).

Another example of a central resetting of thermoregulatory control toward higher internal temperatures occurs in a fever induced by bacterial infection. After infection, pyrogenic cytokines [e.g., interleukin (IL)-1 and IL-6] are elevated in the circulation, and prostaglandin E₂ (PGE₂) increases in the area of the preoptic/anterior hypothalamus (PO/AH) (20, 25). This elevated PGE₂ is thought to be responsible for the increase in the temperature around which the PO/AH regulates body temperature (19, 20, 25). Nonsteroidal anti-inflammatory drugs such as ibuprofen are antipyretic via their capacity as cyclooxygenase inhibitors, thereby lowering the level of PGE₂ in the PO/AH area (6, 25). It has been suggested that pyrogenic cytokines might also be responsible for the rise in body temperature in the luteal phase of the menstrual cycle (2, 13). In an earlier study, Rothchild and Barnes (17) examined whether salicylate administration would affect the body temperature rise induced by very high doses (5-50 mg) of progesterone injected intramuscularly in men, but the results of these experiments were considered by the authors to be inconclusive.

We hypothesized that the shift in reflex control of cutaneous vasodilation with elevated progesterone and estrogen is a prostaglandin-dependent shift in the overall control of temperature regulation. To address this hypothesis, we conducted heat stress experiments in resting women in conditions of low and high hormone status, using acute ibuprofen treatment to test whether cyclooxygenase inhibition would reverse or reduce this shift.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The protocol for this series of experiments was approved by the Institutional Review Board of the University of Texas Health Science Center at San Antonio. Nine young women (ages 21-32 yr) with normal health status and no history of cardiovascular disease volunteered for the study; each gave written informed consent before participation. All were nonsmokers, did not consume caffeine within 12 h of the experiments, and did not take any pain medication within 7 days of the experiments. Of the nine subjects, six were taking oral contraceptives (OC group) and the other three had regular menstrual cycles (NOC group). The subjects in the OC group were voluntarily taking combination oral contraceptives of the type that includes ethinyl estradiol and a low-dose progestin and that provides a steady level of hormones for 21 days followed by a 7-day placebo or no-pill period (Table 1) (3). Each subject participated in four experiments. For the OC group the experiments were carried out as follows: at the end of the 3rd wk of hormone pills (high hormone status), at the end of the placebo/no pill week (low hormone status), and one experiment in each phase in which ibuprofen was administered. The women in the NOC group were studied in the early follicular (days 2-8) and midluteal (days 19-24) phases of the menstrual cycle, with one experiment in each phase in which ibuprofen was given. For all subjects the order of experiments was randomized. Each subject monitored her sublingual temperature on waking each morning during the time period over which the four experiments were conducted (~8 wk).

For whole body heating, the subject dressed in a tube-lined suit for control of body temperature (27). The suit did not cover the head or the areas of blood flow measurement. Mean skin temperature (_sk) was elevated to 38.5°C for ~30-45 min. After whole body heating, _sk was returned to normal, and maximum cutaneous vascular conductance (CVC) was determined by local warming of the area around the flow probe to 42°C for 30 min (12).

Cutaneous blood flow was measured as laser-Doppler blood flow (LDF) (14) on the ventral forearm. LDF values were divided by mean arterial pressure to give an index of CVC. Sweating rate was measured by capacitance hygrometry at the forearm site where LDF was monitored. _sk was assessed from the weighted average of six thermocouples (upper back, lower back, abdomen, upper chest, thigh, and calf) (27). Internal temperature (T_or) was measured with a sublingual thermocouple. To ensure that sublingual temperature was an accurate index of internal temperature, subjects kept their mouths closed and breathed only through the nose throughout all experiments, and no liquids were consumed at any time during any experiment. Mean arterial pressure was continuously measured with a finger blood pressure cuff (Finapres) placed on the middle finger of the left hand and referred to heart level, and HR was monitored by electrocardiogram.

All studies were conducted at the same time of day, beginning at 8:00 AM, to avoid the complications of circadian variation in body temperature and temperature regulation (22, 23). For those experiments in which ibuprofen was administered, 800 mg were administered orally 90-105 min before heat stress. This dose is twice the recommended antipyretic dose; it is less than one-half that recommended for daily treatment of arthritis (6). Ibuprofen takes 1.5-2 h to reach maximal plasma concentrations and has a plasma half-life of ~2 h (6). This time course, therefore, allowed for maximal plasma concentrations of ibuprofen during heat stress.

For confirmation of menstrual phase in the NOC group, a 5-ml blood sample was taken from an antecubital vein before each experiment. Samples were collected in EDTA tubes and immediately centrifuged at 4°C; plasma was then stored at 20°C until the hormone assays were performed. Plasma progesterone and estradiol levels were measured using commercially available RIA kits (Coat-A-Count, Diagnostic Products, Los Angeles, CA). Each sample was analyzed in duplicate.

Data analysis. Data from OC and NOC groups were analyzed separately. T_or at rest was compared across phase and ibuprofen treatment by two-way ANOVA with repeated measures. CVC data are expressed as percentage of maximum to allow for intra- and intersubject comparisons (12). CVC was analyzed as a function of T_or. The T_or at which cutaneous vasodilation began was identified as the threshold for cutaneous vasodilation. For the determination of this threshold, plots of CVC as a function of time were used by an investigator blinded to the subjects and experimental conditions. The point after which there was a sustained increase over baseline was selected, and the T_or at that time was identified as the threshold. The thresholds for sweating were identified as the T_or at which active sweating was first detected. Thresholds for vasodilation and sweating were compared across phase and treatment by two-way ANOVA with repeated measures. The amount by which the thresholds shifted was compared between control and ibuprofen conditions by paired t-test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Resting internal temperature was increased in the high hormone phase in the OC group and in the luteal phase in the NOC group when measured on waking and at the start of the experiment ("pre-heat stress"; Table 2). This increase in resting temperature was not affected by ibuprofen treatment in either group, as pre-heat stress T_or (measured 90 min after ibuprofen administration but before the onset of heat stress) was not affected by ibuprofen treatment regardless of hormone status (P > 0.5; Table 2).

In the NOC group, menstrual cycle phase was confirmed by measurement of plasma progesterone and estradiol. Plasma progesterone (13.75 ± 1.02 and 0.82 ± 0.08 ng/ml in luteal and follicular phases, respectively, P < 0.01) and estradiol (85.08 ± 18.61 and 31.01 ± 9.59 pg/ml in luteal and follicular phases, respectively, P < 0.01) in the luteal phase were significantly elevated above follicular phase levels.

Whole body heat stress caused an increase in T_or of ~0.5°C, which was sufficient to cause marked cutaneous vasodilation and sweating in all experiments. Figure 1A shows the cutaneous vasodilator response to heating as a function of T_or for a representative subject in the two phases of the oral contraceptive cycle. Figure 1B shows this response for the same subject in the experiments in which ibuprofen was administered. The shift in the cutaneous vasodilator response is still apparent in the high hormone phase. Figure 2 shows sweating as a function of T_or from control experiments and those involving ibuprofen. As with CVC, the shift in the sweating response to higher internal temperatures was unaffected by ibuprofen.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Cutaneous vascular conductance (CVC) as a function of sublingual temperature (Tor) during whole body heating in 2 phases of oral contraceptive use [low (open symbols) and high (filled symbols) hormone phases] for a representative subject. A: control experiments (no ibuprofen). B: experiments after ibuprofen administration. Note shift in threshold for vasodilation to higher internal temperatures with high hormone status; this phase-related shift was not affected by ibuprofen treatment.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Sweating rate (SR) as a function of Tor during body heating in low and high hormone phases for a representative subject (see Fig. 1. legend for explanation of symbols). A: control experiments. B: ibuprofen experiments. As with CVC, shift in sweating response to higher internal temperatures with high hormone status was unaffected by ibuprofen treatment.

For all subjects the T_or thresholds for cutaneous vasodilation and sweating were increased with high hormone status regardless of whether ibuprofen was administered. Figure 3 shows average thresholds for the OC group, and the values for the NOC group are shown in Table 3. In the OC group the threshold for cutaneous vasodilation increased 0.27 ± 0.07°C in control experiments and 0.29 ± 0.08°C in ibuprofen experiments (P > 0.1 for threshold shifts, control vs. ibuprofen). Thresholds for sweating increased 0.22 ± 0.05°C in control and 0.35 ± 0.05°C in ibuprofen experiments (P > 0.1, control vs. ibuprofen). Threshold shifts in the NOC group were also unaffected by ibuprofen treatment (P > 0.1). Overall, the absolute values of the thresholds for cutaneous vasodilation and sweating were unaffected by ibuprofen treatment regardless of hormone status (P > 0.5). The sensitivities of the cutaneous vasodilator and sweating responses with respect to T_or were not significantly affected by hormone status or ibuprofen treatment (Table 4).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Tor thresholds for cutaneous vasodilation (A) and sweating (B) in oral contraceptive group (n = 6, means ± SE). Thresholds for cutaneous vasodilation were significantly increased in high hormone phase (P < 0.001); neither thresholds nor threshold shift was influenced by ibuprofen treatment (P > 0.5). Similarly, sweating thresholds were significantly increased with high hormone status (P < 0.001) but unaffected by ibuprofen. Influence of hormone status and lack of effect of ibuprofen were also true for subjects who were not taking oral contraceptives (Table 2). LH, low hormone phase; LH + I, low hormone phase + ibuprofen; HH, high hormone phase; HH + I, high hormone phase + ibuprofen. * P < 0.001, LH vs. HH.

HR increased linearly with T_or in all experiments, as has been shown previously (3, 28). The slope of the HR-T_or relationship was not altered by hormone status or ibuprofen treatment (37.8 ± 6.8, 31.7 ± 4.9, 43.1 ± 6.3, and 35.7 ± 5.7 beats · min¹ · °C¹ for low hormone, high hormone, low hormone + ibuprofen, and high hormone + ibuprofen, respectively, P > 0.05). As we have shown previously (3), the HR-T_or relationship was shifted to higher internal temperatures in the high hormone phase, such that, for a given T_or, HR was always lower in the high hormone phase. For example, at a T_or of 37°C, the corresponding HR value was lower in high hormone phase (P < 0.05). This shift was not affected by ibuprofen (P > 0.05); values were as follows: 88.8 ± 3.3, 80.8 ± 3.5, 83.2 ± 5.4, and 77.8 ± 4.6 beats/min in low hormone, high hormone, low hormone + ibuprofen, and high hormone + ibuprofen, respectively.

Overall, data from all subjects demonstrated a consistent lack of effect of cyclooxygenase inhibition: there was not one individual example in which ibuprofen caused any diminution of the thermoregulatory control shift in the OC or the NOC group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cutaneous vasodilator responses to hyperthermia have been shown to be altered by elevated progesterone and estrogen (3, 7, 8, 22, 23); however, the mechanisms involved remain incompletely understood. This alteration is part of a shift in the control of temperature regulation to higher internal temperatures (3, 7, 8, 22, 23). It has been proposed that this shift may be similar to a fever that follows bacterial infection (2, 13). Consistent with this idea is a report that plasma IL-1 is increased in the luteal phase of the menstrual cycle (2). However, Gonzalez et al. (4) found no increase in the levels of IL-1, IL-6, or tumor necrosis factor in the luteal phase, and Rogers and Baker (16) reported that the levels of IL-1 and IL-6 did not change with the exogenous progesterone and estrogen in oral contraceptives. The fever that accompanies a bacterial infection is dependent on an increase in PGE₂ in the PO/AH area (19, 20, 25) and is therefore reduced by cyclooxygenase inhibition with nonsteroidal anti-inflammatory drugs such as ibuprofen (6, 25).

In the present study we found no measurable effect of cyclooxygenase inhibition on the shift in the cutaneous vascular response to hyperthermia, indicating that prostaglandins are not involved. The dose administered in the present study was twice the recommended antipyretic dose for adults (6), yet there was no effect on the absolute values of the thresholds for cutaneous vasodilation or on the amount the threshold shifted between phases. Consistent with these threshold data for CVC, pre-heat stress T_or and the T_or thresholds for sweating were similarly increased with high hormone status compared with low hormone status regardless of whether ibuprofen was administered. Also, ibuprofen did not affect the shift in the HR-T_or relationship to higher internal temperatures with high hormone status.

Although many of the observations regarding the thermoregulatory influences of progesterone and estrogen have been made when both hormones are elevated (3, 5, 7, 8, 22), progesterone is probably the dominant hormone in the thermoregulatory control shifts in the luteal phase of the menstrual cycle and the high hormone phase of oral contraceptive use. Progesterone is said to be "thermogenic" in that its administration causes increases in body temperature (9, 15, 17). Estrogen alone appears to shift body temperature and temperature regulation to lower levels (1, 9, 24, 26). Although the influence of progesterone appears to dominate, it is also possible that the hormones act in combination in a way that is different from their separate influences.

In 1952, Rothchild and Barnes (17) examined the influence of cyclooxygenase inhibition (oral salicylate administration) on the rectal temperature increase following intramuscular injections of large doses of progesterone (5-25 mg) in men. Their results varied from no effect of salicylate in some individuals to a partial reversal of the internal temperature increase in some of the subjects given larger doses of progesterone (10 and 25 mg). The authors commented that their results were not conclusive, since the effects of salicylate were not uniform among subjects (17).

A possible central mechanism for the influences of female reproductive hormones on thermoregulatory control involves a direct effect on the temperature-sensitive neurons of the PO/AH. Silva and Boulant (21) demonstrated in rat brain slice preparations that estradiol changes the firing rate of temperature-sensitive neurons in the PO/AH; progesterone can also influence the firing rate of these neurons (11). To the extent that these neurons are ultimately in control of thermoregulation, such direct influences of progesterone and estrogen during the menstrual cycle may contribute to the alterations in thermoregulatory control by these hormones. The cellular mechanism(s) by which the hormones alter firing rate in these neurons remains to be determined; the present results, however, do not support a role for prostaglandins.

In agreement with previous observations from our laboratory and others (3, 8, 22), the present data show no influence of hormone status on the sensitivities of the cutaneous vasodilator or sweating responses as functions of internal temperature during hyperthermia (Table 3). Hessemer and Brück (7) reported an increase in the sensitivity of cutaneous vasodilation in the luteal phase; the reason for the discrepancy between their results and the above studies is unclear.

In summary, the alterations in cutaneous vascular control during hyperthermia as well as the changes in body temperature and in the control of sweating that accompany increases in plasma progesterone and estrogen are not sensitive to cyclooxygenase inhibition by ibuprofen. These alterations therefore appear to be prostaglandin independent. Therefore, although the changes in thermoregulatory control accompanying elevated progesterone and estrogen follow a pattern similar to that observed with the fever following an infection, the mechanisms must be different. Although direct influences of progesterone and estrogen on temperature-sensitive neurons of the PO/AH may be involved, further study is needed to clarify the mechanisms by which these hormones shift thermoregulatory control.