In this study, the cold pressor test (CPT) was used to test a model of the effects of acute pain on 10 HIV+ and 10 HIV- adults. Participants were exposed to the CPT for a maximum of 5 minutes. Blood samples were collected immediately before, immediately after, and 1 hour after the CPT. Variables included immune measures (CD4+, CD8+, and CD 16+56+ lymphocyte number, CD4+CD8+ lymphocyte ratio and NK cell cytotoxicity), cardiovascular reactivity (heart rate, systolic and diastolic blood pressure), anxiety, perceived pain intensity and perceived self-efficacy. Effects of pain were generally consistent across HIV+ and HIV- groups, with no between-group differences across time in immune responses, state anxiety and diastolic blood pressure. Within-subjects differences across time averaged over both groups were significant for NK cell cytotoxicity, CD8+ and CD 16+56+ lymphocyte numbers, anxiety and heart rate. Significant nonlinear trends were observed for CD 16+56+ lymphocyte numbers, NK cell cytotoxicity and state anxiety in both groups and for heart rate in the HIV+ group only. Perceived pain intensity was significantly associated with state anxiety (r = .65), systolic (r = -.56) and diastolic (-.52) blood pressure and CD4+ lymphocyte number (r = .48). Heart rate and trait anxiety were significantly associated with all immune variables. Associations were positive for CD4+ lymphocyte number and inverse for all other immune measures. Associations between perceived self-efficacy and both perceived pain intensity and anxiety were inverse, as predicted, but not significant. Overall, the direction and strength of observed relationships provided some support for the theoretical model on which the study was based. Generally, responses to acute pain were consistent and did not differ by HIV status.
There is compelling evidence from laboratory studies in animal models that pain has an effect on immune responses. In these studies, pain resulted in consistent down regulation of the immune system, increased metastases of experimental tumors and reduced survival (Ben-Eliyahu, Yirmiya, Liebeskind, Taylor, & Gale, 1991; Keller, Weiss, Schleifer, Miller, & Stein, 1981; Lewis et al., 1983; Page, Ben-Eliyahu, Yirmaya, & Liebeskind, 1993; Shavit et al., 1986; 1987). The majority of human laboratory studies have examined the effects of psychological rather than physiological stressors and support their immunological effects (Schultz & Schultz, 1992). The few human laboratory studies of the effects of painful stressors have provided preliminary evidence that immune responses to pain do occur (see below). In addition, they demonstrate that these immune responses may be mediated by a number of psychological and physiological variables.
The immunomodulatory effects of pain may have as yet unrecognized consequences in clinical populations in terms of disease severity and survival. In addition, acute pain may increase susceptibility to viral infection in otherwise healthy individuals (Cohen & Williamson, 1991). Before potential longterm effects of pain-related immune responses can be addressed, however, the specific immunological effects and their mediators must be characterized. Therefore, the main objectives of this study, using a psychoneuroimmunology model, were: (a) to examine the effects of a cold pressor test (CPT) pain model on immunity in healthy and HIV positive men and women, and (b) to identify mediators of immune responses to acute pain. Immune measures included CD4+ and CD8+ and CD 16+56+ (natural killer) lymphocyte numbers, CD4+CD8+ ratio and natural killer cell cytotoxicity (NKCC). Mediators included sympathetic nervous system activation, perceived self-efficacy to withstand CPT pain, and perceived pain intensity.
STRESS AND IMMUNITY
Increasingly, human studies have documented immune responses to naturally occurring or laboratory induced physiological and psychological stressors. Immune markers most consistently influenced by pain and other stressors include natural killer cell activity (NKCC), lymphocyte proliferation in response to the plant mitogens phytohemagglutinin (PHA) and concanavalin A (ConA), and alterations in the number of lymphocyte subsets, including CD4+, CD8+ and CD16+CD56+ lymphocytes (Bachen et al., 1992; Bachen, Manuck, Cohen, & Muldoon, 1995; Brosschot et al., 1992; Cacioppo, 1994; Girgis, Shea, & Husband, 1988; Knapp et al., 1992; Manuck, Cohen, Rabin, Muldoon, & Bachen, 1991; Marsland, Manuck, Fazzari, & Stewart, 1995; Naliboff et al., 1991; Naliboof, Solomon, Gilmore, & Benton, 1995; Sieber et al., 1992; Uchino, Cacioppo, Malarkey, & Glaser, 1995; Weisse et al., 1990; Zakowski, McAllister, Deal, & Baum, 1992). Studies in clinical populations, those most likely to be exposed to acute procedural or disease-related pain have not been reported. The immune effects of the stress of acute pain in these populations have not been characterized and may ultimately be relevant to disease outcomes. Whether the effects of acute stress on immunity hold in an HIV+ population is not known. In a recent study using the cold pressor test in persons with HIV, subjects demonstrated blunted autonomic responses, with slower increase and an earlier peak in epinephrine (Kumar, Morgan, Szapocznik, & Eisdorfer, 1991). Therefore, based on catecholaminergic responses, immunological effects of acute pain could be expected to differ in this population.
Recently, preliminary evidence from three controlled studies supports immune responses to pain in healthy human subjects. In these studies, subjects exposed to electric shock, loud noise or venipuncture demonstrated significantly reduced immune responses, including NKCC and CD4+ lymphocyte number (Girgis, Shea, & Husband, 1988; Sieber et al., 1992), and mitogen response (Weisse et al., 1990). These alterations were related to perceived intensity and controllability of pain and level of anxiety. There were inconsistencies in findings, particularly when considering these studies in combination with those that examined psychological stressors and immunity. Four out of six studies reported increases in CD8+ T-lymphocyte numbers. The two studies reporting no increase, however, introduced the variable of controllability of the stressor (Sieber et al., 1992; Weisse et al., 1990). Therefore, lack of effects may have been due to the moderating influence of controllability.
In studies exploring immediate post-stressor immune responses, CD4+ Tlymphocytes were observed to decrease in two studies but remained unchanged in five others. In those studies where significant effects were reported, suppression was related to anxiety and cardiovascular sympathetic response as measured by systolic and diastolic blood pressure. These potential mediating variables were not measured in other studies.
CD 16+ CD56+ T-lymphocytes were found to increase in number in three studies, but Sieber and associates (1992) reported no significant change. Here again the variable of controllability was introduced. Finally, NKCC, a functional measure of immunity, was observed to increase in two studies, decrease in two others, and remain unchanged in another two studies. Inconsistencies in findings across studies may be due to several factors, including the kinetics and mechanisms of stress-related immune changes. First, immune parameters affected by physiological versus psychological stressors may differ. Second, although most studies compared baseline measures with only one post-stress measure, several post-stress measures may be necessary to accurately characterize biphasic immune responses. Third, indicators of sympathetic activation, a mediator of immune change, were not consistently measured or controlled across studies. These include anxiety and cardiovascular reactivity. And finally, controllability of the stressor, which can be measured as perceived selfefficacy, was not considered in most studies.
Mechanisms of Stress-Related Immune Modification
One mechanism postulated for the immunosuppressive effects of acute stressors is activation of the sympathetic nervous system. Immune changes associated with sympathetic activation and increased catecholamine production have been well characterized (Bachen et al., 1992; 1995; Crary et al., 1983; Landmann et al., 1984; Manuck et al., 1991; Marsland et al., 1995; Naliboff et al., 1991; 1995; Uchino et al., 1995). These include increases in NK cell number and NKCC, increased number of CD8+ lymphocytes and reduced CD4+CD8+ lymphocyte ratio. In one study of acute psychological stress, where no changes in sympathetic response (as measured by catecholamine levels) were observed, there were also no changes in immunity (Caudell & Gallucci, 1995).
Recent observations of up-regulation of lymphocyte (3-adrenoreceptors in response to stress support the role of catecholamines in post-stress immune alterations, particularly lymphocytosis. In addition, direct administration of adrenaline resulted in increased numbers of NK cells and CD8+ cells, while CD4+ lymphocyte numbers either were reduced or no changes were observed, (Kappel et al., 1991; Schledowski et al., 1993, 1996). In other experiments confirming these observations, the administration of p-adreneoreceptor antagonists, and, more specifically, P2 adreneoreceptor resulted, in complete inhibition of stress-induced lymphocytosis (Murray et al., 1992; Schedlowski et al., 1996).
Two indicators of sympathetic activation are anxiety, and cardiovascular reactivity which includes heart rate, systolic and diastolic blood pressure. Subjects with high sympathetic reactivity to stress as measured by heart rate, and systolic and diastolic blood pressure reported more pain during CPT (Peckerman et al., 1990). In other studies, high reactivity was associated with elevations in CD8+ and NK cell number (Knapp et al., 1992; Manuck et al., 1991).
Sympathetic response and its subsequent effects on immunity may also be influenced by the intensity of pain. In animal models, the degree of suppression of mitogen responses to PHA and ConA was related to the intensity and number of shocks delivered (Keller et al., 1981). In human subjects, higher perceived intensity of pain resulted in greater suppression of mitogen response to ConA (Weisse et al., 1990).
Anxiety is postulated to increase perceived pain intensity through its influence on sympathetic response and cardiovascular reactivity. Anxiety was significantly and negatively associated with lymphocyte response to ConA and PHA and positively related to NK cell and total lymphocyte number (Knapp et al., 1992; Naliboff et al., 1991). Positive correlations were reported between state anxiety and perceived pain intensity, particularly when anxiety was specific to the pain stimulus (Ahles, Cassens, & Stalling, 1987; Dougher, Goldstein, & Leight, 1987; Malow, 1981; Malow, West, & Sutker, 1987).
Perceived Self-Efficacy and Controllability
Perceived self-efficacy may also moderate perceived pain intensity. Perceived self-efficacy indicates a person's belief in his or her ability to exert control over a stressor and acts as a central control mechanism activating stress-induced analgesia (Bandura, O'Leary, Taylor, Gauthier, & Gossard, 1987; Litt, 1988). Low perceived self-efficacy was associated with autonomic activation and increases in plasma catecholamines (Wiedenfield, O'Leary, Bandura, & Brown, 1990). High self-efficacy was associated with decreased autonomic arousal, increased cold pressor pain tolerance and lower perceived pain (Bandura et al., 1987; Litt, 1988).
Perceived controllability of the stressor appears to play a role in its effect on immunity in both animal and human models. In animal models, suppressed mitogen responses and NKCC were reported in response to uncontrollable but not controllable shock (Laudenslager, Ryan, Drugan, Hyson, & Maier, 1983). Effects of controllability on immune responses to pain were not consistent in two human studies that measured this variable. In one study, NKCC was suppressed only in those subjects who perceived the stressor as uncontrollable (Sieber et al., 1992). Conversely, Weisse and associates (1990) found that those who could control the delivery of painful stimuli had suppressed ConA and PHA responses. Weisse also reported that subjects exposed to controllable shock had significantly higher scores on their perceived intensity of the shock, with a strong negative relationship observed between perceived intensity and suppressed immune response. This may help explain the lower immune responses in those receiving controllable shock. Inconsistent results may be related to the strength of sympathetic response. Subjects able to control painful electric shocks may have had increased anxiety and stronger sympathetic responses related to their efforts to terminate the shock with button presses.
Pain and Infectious Illness
The clinical effects of pain-induced immune modulation in humans are unclear; however, studies support a relationship between stress and self-report of symptoms of upper respiratory infection 3 to 4 days post-stress (Cohen & Williamson, 1991). No laboratory studies were found that examined the potential clinical relevance of immune responses to pain in terms of infectious illness. An examination of the ultimate effects of pain on clinical outcomes is therefore essential. Figure 1 describes hypothesized relationships examined in the study reported here.
Measurement of mediating dependent variables was based on the theoretical model that guided the study. This model was based on review of the literature and the hypothesized relationships among the pain stimulus, perceived selfefficacy, sympathetic nervous system response, perceived pain intensity, and immune responses. Components of the sympathetic response include cardiovascular reactivity and anxiety. Other potential mediators of immune responses to pain, such as cortisol and stress-induced analgesia, were not included for economy of the model since they were not measured. The direction of hypothesized relationships is indicated in Figure 1.
Based on this model, and observed effects of pain in animal studies and effects of pain and psychological stress on immunity in human studies, the following hypotheses were posited:
1. CD4+, CD8+ and CD 16+56+ lymphocyte numbers and NKCC will decrease immediately post exposure to acute pain.
2. Indicators of cardiovascular reactivity (heart rate, blood pressure and anxiety) will increase immediately post exposure to acute pain.
3. CD4+, CD8+ and CD 16+56+ lymphocyte numbers and NKCC will return to baseline at 1 hour post exposure to acute pain.
4. Indicators of cardiovascular reactivity (heart rate, blood pressure and anxiety) will return to baseline at one hour post exposure to acute pain.
5. Changes in immune responses and cardiovascular reactivity will differ between HIV+ and HIV- subjects.
Design and Sample
A within-subjects repeated measures design was used to compare the effects of acute pain in persons with and without HIV at 3 time points. A convenience sample of 20 participants (10 HIV+ and 10 HIV-) was recruited via flyers posted in a large urban university and in HIV clinics in the same urban area. Those interested in participation contacted the researcher via telephone for additional information. Potential participants self-identified as HIV negative or positive. Determination of eligibility was made and those eligible were scheduled into the study. Exclusion criteria were based on potential confounding factors that could influence immune status or responses to acute pain. These included self-report of acute or chronic pain, pregnancy, lactation, hypertension, impaired circulation, allergy to cold, admission of or clinical signs of drug or alcohol abuse, and any medical condition (other than HIV) affecting immune status. Ten HIV negative participants were drawn from a sample of hospital employees, and 10 HIV positive participants were drawn from a sample of patients of the outpatient infectious disease clinic. Subjects were paid $50 for participation in the study. The study protocol was reviewed and approved by the institutional review board of the medical center where the study was conducted.
There were no between-group differences with regard to age, gender, and education. The HIV positive group, however, had a significantly lower average income, with 60% earning below $10,000, while only 20% of the HIV negative group was in this income range. Seventy percent of this group was either unemployed or worked only part-time compared to 90% working full-time in the HIV negative group. Finally, the ethnic distribution of the HIV positive group differed in that 50% were either African American or Latino/a compared to only 10% minority representation in the HIV negative group. These differences could potentially affect responses to an acute pain stressor. Therefore, analyses were conducted separately for HIV+ and HIV-subjects.
Indicators of cardiovascular reactivity included systolic and diastolic blood pressure and heart rate. Each was measured at 1-minute intervals in triplicate at each time point, and the mean of the measures was used.
Immune assays were conducted on fresh peripheral blood mononuclear cells (PBMCs). Blood was stored at room temperature until assays were performed. Peripheral blood was drawn at each of three data points. To reduce potential variance in assay results, the same lot numbers of consumables, reagents and biologicals and were used for all assays.
Natural Killer Cell Cytotoxicity (NKCC). NKCC was measured by the number of lytic units of 51 chromium (51Cr) released by labeled K562 target cells following incubation with subject lymphocytes (Rose, DeMacario, Fahey, Friedman, & Penn, 1992). Ten milliliters of whole blood were collected from each subject aseptically by venipuncture into a vacutainer containing sodium heparin (Becton & Dickinson, Research Triangle Park, NC). Within 1 hour of collection, mononuclear cells were separated from blood samples on Ficoll-Hypaque (Pharmacia, NJ). Cells were collected and washed twice with phosphate-buffered saline. Isolated cells were more than 95% viable by trypan blue exclusion. Cells were resuspended in RPMI with 10% fetal calf serum, 1 % L-glutamine at a concentration of 4x 106 cells per milliliter. Cells were incubated in a U-bottom microtiter plate with 51Cr-labeled K562 cells at effector-target cell ratios of 40:1, 20:1,10:1, and 5:1 in triplicate. The plate was centrifuged at 200g for 5 minutes both before and after 3 hours of incubation at 37°C and 5% CO2. Spontaneous release and total release were obtained from targets incubated with NK media alone, and medium containing 5% Triton X-100 detergent, respectively. Following incubation, supernatant was harvested from each microtiter plate well using SCS harvesting frames and filters (Skatron Instruments, Sterling, VA). Radioactivity in the supernatant was determined by gamma counter and the percentage of 51Cr released was calculated for percent of natural killer cell cytotoxicity. Data generated from NK assays were analyzed using the Von Krogh equation as described by Pross, Baines, Rubin, Schragge, and Patterson (1981) and expressed as lytic units/107 cells where one LU is the number of effector cells required for 25% lysis.
Lymphocyte Subset Phenotyping. CD4+, CD8+ and CD 16+56+ lymphocyte subsets were analyzed using single or dual color direct immunofluorescence. Five milliliters of whole blood were collected from each subject by venipuncture in a vacutainer containing EDTA (Becton & Dickinson, Research Triangle Park, NC). A complete blood count and differential were performed. Whole blood lysis and labeling with the appropriate monoclonal antibody were followed by flow cytometry. The percentage of positively stained cells for each cell surface antigen was obtained from a count of 5,000 cells. Estimates of absolute numbers of CD4+, CD8+ and CD16+CD56+ lymphocytes were determined from the complete blood count and differential (Rose et al., 1992).
State-Trait Anxiety Inventory (STAI)
Sympathetic nervous system response was also measured with indicators of state and trait anxiety. The State-Trait Anxiety Inventory consists of two 20-item questionnaires with statements to describe how the person "generally feels" (trait anxiety) or statements "which describe your present feelings" (state anxiety) (Spielberger, Gorusch & Lushene, 1970). A range of answers from 1 to 4 indicate "almost never" to "almost always" reactions. An example of a question is "I feel anxious." Extensive standardization and validation of this inventory have been conducted over several decades with healthy and clinical populations (Spielberger et al., 1983). In this study, the STAI-TR was measured once at baseline. Cronbach's alpha reliability for this scale was .92. The STAI-S, which measures situation-specific or state anxiety, was measured at four time points. Reliabilities for the scale at each time point ranged from .82 to .93.
Perceived Self-Efficacy Scale
Perceived self-efficacy, which indicated the individual's belief in the ability to exert control over pain, was measured with this scale, which includes 10 100 mm visual analog scales representing increasing lengths of exposure to the CPT. Anchors for the scales were "Absolutely certain that I can tolerate the cold for X time" and "Absolutely certain I can not tolerate the cold for X time." The eight items ranged from 15 seconds to 5 minutes in 15 second increments for the first minute, then 1 minute increments thereafter. A total self-efficacy score was calculated by summing the strength judgments for all the items (Bandura et al., 1987; Litt, 1988; Mogil, Sternberg, & Liebeskind, 1993). Cronbach's alpha for this scale was .90.
Perceived Pain Threshold
Short-Form McGill Pain Questionnaire (SF-MPQ). This questionnaire was used to measure perceived pain intensity. The SF-MPQ includes two pain intensity scales:- a 100 mm VAS with the anchors "No pain" and "Worst possible pain," and a Present Pain Intensity Scale with a range of 1 to 5 items indicating "no pain" to "excruciating pain" (Turk & Melzack, 1992). Reliability was supported by a correlation coefficient of .69 between the VAS and Present Pain Intensity Scale. Pain Threshold and Pain Tolerance, as described below, were also indicators of perceived pain intensity.
Pain Threshold. To measure pain threshold, or the point at which the person first begins to experience pain, participants were instructed "when you just begin to feel pain, say the word 'pain' out loud." The number of seconds from time of immersion to this point was noted.
Pain Tolerance. To measure pain tolerance, the length of time until the person experienced maximum tolerable pain, participants were told "you may withdraw your hand at any time if the pain becomes more than you can tolerate." Number of seconds from time of immersion to time of withdrawal was noted.
Two days before their scheduled participation, participants were contacted and asked to refrain from ingesting alcohol, nonprescription medications and drugs, and from engaging in strenuous exercise during the 24 hours preceding their participation in the study. They were asked to call and reschedule their appointment if they became ill. They were also instructed to take nothing by mouth except water on the morning of the study. All study sessions began between 8 A.M. and 9 A.M. to control for the effects of circadian rhythm on lymphocyte trafficking. When the participant arrived, the procedure was explained, questions answered and informed consent obtained.
Figure 2 summarizes the data collection procedures. Following completion of informed consent, Time 1 measures of state and trait anxiety, cardiovascular response and a demographic questionnaire were collected. Next, to control for effects of venipuncture pain, a topical eutectic mixture of local anesthetic (EMLA) cream was used on venipuncture sites (Girgis et al., 1988). Dermal anesthesia with EMLA is achieved in 1 hour and persists for 2 hours. It was applied 1 hour before the first venipuncture. After a wait of 60 minutes during which neutral reading materials were provided, Time 2 measures were taken. These included state anxiety, perceived self-efficacy and measures of cardiovascular reactivity. In addition, 15 milliliters of blood were collected for measures of immunity: CD4+, CD8+ and CD16+56+ lymphocyte numbers, CD4+:CD8+ lymphocyte ratio, and NKCC.
The cold pressor test then began. The apparatus consisted of a bucket filled with 1/3 crushed ice and 2/3 cold water stirred to maintain a constant temperature of 1°-2°C. In order to insure consistency in exposure to the stressor, participants were seated with their nondominant hand submerged, palm down on the bottom of the bucket, 4 inches above the wrist. The cold pressor was maintained for a maximum of 300 seconds, based on evidence that the cold pressor pain ceiling is reached in approximately 5 minutes (Bandura et al., 1987). Pain threshold was measured during the CPT.
At the time of hand withdrawal, Time 3, pain tolerance, cardiovascular reactivity, state anxiety and perceived pain intensity were measured, and 15 milliliters of blood were collected for immune assays. Participants were then given neutral materials to read for 60 minutes, after which Time 4 measures were taken. These included the same measures as those at Time 2, except for perceived pain intensity. One week following exposure to the CPT, at Time 5, participants were interviewed by telephone to determine the incidence of symptoms of upper-respiratory infection.
Means and standard deviations were computed for state anxiety and cardiovascular reactivity at baseline (Time 1), immediately pre-CPT (Time 2) immediately post-CPT (Time 3), and 60 minutes post-CPT (Time 4). Means and standard deviations were computed for immune measures at Time 2, Time 3, and Time 4. Repeated measures analysis of variance was used to test main effects of group differences, changes across time points, and interactions between group and time. HIV status at two levels (HIV+ and HIV-) was the between-subjects factor. The within-subjects factor was time with four levels for state anxiety and cardiovascular reactivity, and time with three levels for all immune measures. Nine sets of dependent measures were evaluated: CD4+, CD8+ and CD 16+56+ lymphocyte numbers, CD4+CD8+ ratio, NKCC, state anxiety, heart rate, systolic and diastolic blood pressure. If main effects of time or interaction effects were significant, data were plotted and trend analyses with polynomial contrasts were conducted separately for HIV+ and HIV- participants. Consideration of the risks and benefits of Type I versus Type II error led to the judgment that the possibility of falsely rejecting the null hypothesis was a lesser risk than that of falsely accepting it; therefore, a two-tailed apriori alpha of .10 was selected for these statistical tests. Bivariate relationships hypothesized in the model were explored with Pearson correlation coefficients.
Table 1 provides descriptive data for outcome variables at each study time point. As would be expected in comparing HIV+ and HIV- participants, groups differed at baseline on all immune measures. For this reason, all analyses included HIV status as a between-subjects variable (see Table 2).
Relationship of Immune Status to Sympathetic Nervous and Immune Responses
Interaction effects were not significant for any immune measure, indicating that trends over time in immune responses to acute pain did not differ for HIV+ and HIV- participants (see Table 2). Similarly, changes over time in state anxiety and diastolic blood pressure did not differ by HIV status. Group by time interactions were observed for heart rate and systolic blood pressure, indicating that the HIV+ and HIV- groups differed in these responses over time.
Within-subjects differences across time averaged over both groups were significant only for CD8+ and CD 16+56+ lymphocyte numbers, NKCC, state anxiety and heart rate (see Table 2). Data were plotted for those variables with significant effects over time. Significant nonlinear trends were observed for CD 16+56+ lymphocyte numbers, NK cell cytotoxicity and state anxiety in both groups, and for heart rate in the HIV+ group only. Trends for immune variables across the three time points demonstrated an initial increase from baseline to immediately post-CPT, and a subsequent decline to baseline values or lower at 1 hour post-CPT. These trends were the same in both HIV+ and HIV- groups. The most significant change in immune measures occurred in NKCC in both groups, with a nearly 100% increase in NKCC in immediate response to acute pain. Separate analyses for trends, as described below, supported these observed differences.
Trend analysis with polynomial contrasts was conducted separately for HIV+ and HIV-subjects (see Table 3). Analyses revealed significant quadratic trends for CD 16+56+ lymphocyte number and NKCC, and significant cubic trends for state anxiety in both HIV+ and HIV- participants. Trends for heart HIVsubjects.
Correlation coefficients were examined for variables assumed to have a linear relationship in the model for the study (see Table 4). Values observed at the immediate post-CPT time point were used unless otherwise specified.
Additional Model Testing
The relationship between self-efficacy and perceived pain intensity was inverse, as predicted, but was not significant (r = -.27). The associations between self-efficacy and pain threshold (r = .22) and tolerance (r = . 13) were neither significant nor in the predicted direction. The relationships between perceived self-efficacy and sympathetic responses were weak and not in the predicted direction. Post hoc power calculations revealed that power to detect significant relationships ranged from .13 to .43 for immune variables, and .22 for perceived pain intensity, .16 for pain threshold and .08 for pain tolerance.
Sympathetic Responses and Pain Intensity
The relationships between perceived pain intensity and measures of sympathetic response immediately following CPT were moderate and significant for systolic blood pressure (r = .56), diastolic blood pressure (r = .52), and anxiety (r = .65), and weak and not significant for heart rate. Relationships between cardiovascular responses and perceived pain intensity were inverse, which does not support the model's assumption that physiological responses to CPT are directly related to perceived pain intensity. Meanwhile, the relationship between perceived pain intensity and state anxiety (r = .65, p < .001) was positive and significant, indicating that higher levels of pain intensity were related to higher levels of anxiety. Similarly, the relationship between state, anxiety immediately pre-CPT and pain tolerance was inverse, strong and significant (r = -.73, p <.001).
Sympathetic Responses and Immunity
Relationships between immediate post-CPT heart rate and immune measures were strong and significant. Relationships between immediate post-CPT anxiety and immunity were not significant; however, trait anxiety was strongly associated with post-CPT immune changes.
Pain Experience and Immunity
Finally, the relationships between perceived pain intensity and immune measures immediately post-CPT were positive and weak to moderate, and significant only for CD4+ lymphocyte count (r = .48, p < .05). The direction of these relationships indicates that all measures of cellular immunity were initially elevated in response to acute pain. Post hoc power calculations indicated that power to detect significant relationships between perceived pain intensity was .26 for CD8+ lymphocyte number, .09 for NK cell number, .23 for CD4+CD8+ ratioand.41 for NKCC.
Pain tolerance was significantly associated with CD4+, CD8+, and CD 16+56+ lymphocyte number and CD4+CD8+ ratio, but not NK cell cytotoxicity, a functional measure. This indicates a possible effect of length of immersion on lymphocyte trafficking.
Perceived pain intensity and pain threshold did not differ by HIV status, Pain threshold, the time at which the participant first reported feeling pain, ranged from 8 seconds to 300 seconds. One participant reported that he did not feel pain, "just some discomfort." Mean threshold to first pain report was 46 seconds (S.E.M. ±20). Median pain threshold was 15 seconds. Pain tolerance, or total length of time of ice water immersion, did differ by HIV status F (2, 18) = 8.64, p = < .01. Lower tolerance was observed in the HIV+ group, with a mean immersion time of 114 seconds (S.E.M. ±34), and a median immersion time of 69 seconds. This compared to a mean immersion time of 251 seconds (S.E.M. ±32) and median immersion time of 300 seconds in the HIV- group. Tolerance and threshold were positively related (r = .34) but not significantly associated with each other. Pain threshold and pain tolerance were significantly related to perceived pain intensity. Both were negative associations, indicating that high perceived pain intensity was associated with a shorter time to reporting of the first sensation of pain and a shorter time to withdrawal from the ice water.
Incidence of Illness
Between 7 and 8 days following study participation, 16 study participants were contacted by telephone and questioned about the incidence of symptoms of upper-respiratory infection during the days following the study. Four participants (3 HIV+ and 1 HIV-) could not be reached. Two of the participants contacted (14%) reported symptoms. One complained of a "sore throat" and the other a "runny nose." Both were males in the HIV- group.
This study used the cold pressor test (CPT) to test a model of the effects of acute pain on immunity in HIV+ and HIV- participants. Relationships between physiological and psychological variables were explored. Variables included immune measures (CD4+, CD8+, and CD 16+56+ lymphocyte numbers, CD4+CD8+ lymphocyte ratio and NK cell cytotoxicity), cardiovascular reactivity (heart rate, systolic and diastolic blood pressure), anxiety, perceived pain intensity and perceived self-efficacy. Significant differences in CD8+ and CD 16+56+ lymphocyte numbers and NKCC were observed in response to CPT. These responses did not differ by HIV status. The observation that acute pain affects both lymphocyte trafficking and function supports its effects on both surveillance and functional aspects of the immune system. Redistribution and functional alteration of lymphocytes in response to acute pain may have a significant effect on the immune system's ability to respond to the challenge of trauma, injury, cellular mutation, or invasion by pathogens, and thus on eventual health consequences.
State anxiety and cardiovascular reactivity were measured at four time points. Immune variables were measured at three time points. Changes across time included significant quadratic trends for CD 16+56+ numbers and NKCC. Similar initial elevations, immediately post-CPT, occurred in other immune variables but these were not significant. The observed immediate potentiation of immune responses in response to acute pain may indicate a protective physiological response to threat of injury. Increases in lymphocyte populations provide cells for nonspecific immune responses and increased production of cytokines that will mobilize additional immune reactions. It is interesting to note that immune changes observed in response to acute pain were similar to those observed by other researchers in response to acute psychological stress (Bachen et al., 1992, 1995; Brosschot et al., 1992; Marsland et al., 1995; Naliboff et al., 1991; 1995; Uchino et al., 1995). This suggests that mediators of these early immune responses, for example, sympathetic activation, are nonspecific for physical or psychological threat.
At 1 hour post-CPT, lymphocyte numbers returned to baseline. Although there was a decrease, however, NKCC remained elevated above baseline levels. In future studies, measurement of immune responses for longer time periods will permit exploration of the duration of enhancement of functional immune responses. In addition, it is necessary to determine whether lymphocyte numbers continue to regress to below baseline levels or remain stable beyond 1 hour after exposure to acute pain. One study that examined NKCC responses to stress reported decreases at 24 and 72 hours post stress (Sieber etal., 1992).
Unlike immune measures, heart rate and systolic blood pressure did differ over time by HIV status. These differences appear to be due to higher baseline values in HIV+ participants, which leveled off by the pre-CPT time point and remained at similar levels. These elevations at the time when participants initially arrived at the study site support the need for a period of neutral activity prior to collection of measures which can be influenced by sympathetic activation.
Despite Kumar and associates' (1991) report of blunted sympathetic responses in HIV infection, there were no between-group differences noted in immune responses to acute pain. Although a decrease in autonomic reactivity in HIV+ subjects was observed by Kumar et al., there appears to be a sufficient sympathetic response to acute pain to mediate similar immune alterations in HIV- and HIV+ subjects. It is interesting to note that similar changes in immunity occurred despite a total ice water immersion time of approximately 2 minutes in HIV-' subjects compared to nearly 5 minutes in HIV+ subjects. Kumar et al. reported that peak sympathetic response to CPT occurred at 0 to 2 minutes in those who were HIV+ and at 2 to 4 minutes in HIV- subjects, both following a 2-minute CPT. These times are consistent with the differences in immersion time in the two groups in this study. It may be that reported earlier peaks in HIV+ subjects resulted in decreased pain tolerance. In addition, since blood collection occurred immediately post-CPT, the blood drawn for both groups would coincide with the timing of peak sympathetic response, resulting in similar immune changes in response to sympathetic activation.
The change in NKCC was the most dramatic immune response to acute pain, with an immediate doubling of activity, and sustained increase. Observations that NK cell function remained elevated above baseline for 1-hour post acute pain indicates a need to determine the pattern of change in this immune measure for a longer period following exposure to the stressor. Significant time differences in pulse and state anxiety levels support the effects of acute pain on sympathetic activation. The fact that only anxiety was significantly associated with perceived pain intensity, however, suggests the use of this psychological indicator in studies of acute pain.
Only two indicators of sympathetic response, heart rate and state anxiety, varied significantly over time. Variations in anxiety mimicked those seen in immune measures, with sharp elevations immediately post-CPT followed by a return to baseline. There were no changes observed over time in blood pressure, and, in addition, systolic and diastolic blood pressure were not associated with any post-CPT immune measures. This differed from results of other studies linking cardiovascular reactivity with sympathetic response and immune changes (Knapp et al., 1992; Manuck et al., 1991; Peckerman et al., 1990). In these studies, systolic and diastolic blood pressure as well as anxiety and heart rate were associated with immune changes. In two of those studies, however, high and low sympathetic reactors were separated, and cardiovascular responses were observed only in high reactors. Future studies should include a sufficient sample size to explore the effects of high versus low sympathetic reactivity. It is also possible that in this study a ceiling effect was observed. Participants were aware that the study involved exposure to a painful stimulus and may have begun the study with high baseline systolic and diastolic blood pressures. The strong and significant relationships between trait anxiety and immune measures suggest that this relatively stable personality trait may be a better predictor of immune response to acute pain than the transitory emotional state.
In further exploring relationships in the proposed model, the associations between perceived self-efficacy and pain intensity, and self-efficacy and pain tolerance were not significant. These findings differ from those observed by several researchers in both laboratory and clinical pain populations (Bandura et al., 1987; Litt, 1988; Manning & Wright, 1983; Shoor & Holman, 1984).
The relationship between post-CPT state anxiety, an indicator of sympathetic activation, and pain intensity was positive and significant. Anxiety pre- CPT was also associated with pain tolerance. These observations support the proposed relationships between a psychological indicator of sympathetic response and perceived pain intensity and tolerance. They are also similar to observations in other studies (Ahles et al., 1987; Dougher et al., 1987; Knapp etal., 1992;Malow, 1981; Malow et al., 1987; Naliboff et al., 1991).
Systolic and diastolic blood pressure were both negatively and significantly associated with pain intensity. Contrary to the proposed model, this suggests that high cardiovascular reactivity was associated with lower perception of pain. Heart rate was not associated with pain intensity. The proposed relationships between perceived self-efficacy and sympathetic responses also were not supported for either physiological indicators (systolic and diastolic blood pressure and heart rate) or psychological indicators (state anxiety) of sympathetic activation.
All post-CPT immune measures were moderately to strongly associated with heart rate at the same time point. Heart rate may be the most responsive surrogate marker of sympathetic activation when exploring physiological responses to acute pain. Pain tolerance was associated only with enumerative measures of immunity, including CD4+ and CD 16+56+ numbers and CD4+CD8+ ratio.
It must be noted that the sample size in this study, coupled with the number of variables, provided limited power to detect significant relationships. Larger sample sizes in future studies would be needed to determine whether or not lack of significant findings was the result of insufficient power to detect a relationship. For this same reason, significant results reported should be interpreted cautiously. This study adds information about the potential effects of acute pain on immunity. Findings did not differ between a healthy and clinical population and are surprisingly similar to those in studies that examined the effects of acute psychological stressors.
Observed relationships provided some support for the theoretical model for the study. Ideally, however, the study would be replicated with a larger sample adequate for path analysis and further explication of the model. The clinical implications of the effects of acute pain on immunity are particularly relevant to populations with derangements of the immune system, including autoimmune or immmunosuppressive diseases. Further exploration of the kinetics of immune effects is warranted to determine the relevance of these responses to clinical populations.
Ahles, T. A., Cassens, H. L., & Stalling, R. B. (1987). Private body consciousness, anxiety and the perception of pain. Journal of Behavior Therapy and Experimental Psychiatry, 18, 215-222.
Bachen, E. A., Manuck, S. B., Marsland, A. L., Cohen, S., Malkoff, S., Muldoon, M. F., & Rabin, B. S. (1992). Lymphocyte subset and cellular immune responses to a brief experimental stressor. Psychosomatic Medicine, 54, 673-679.
Bachen, E. A., Manuck, S. B., Cohen, S., & Muldoon, M. F. (1995). Adrenergic blockade ameliorates cellular immune responses to mental stress in humans. Psychosomatic Medicine, 57, 366-372.
Bandura, A., O'Leary, A., Taylor, C. B., Gauthier, J., & Gossard, D. (1987). Perceived self-efficacy and pain control: Opiod and nonopiod mechanisms. Journal of Personality and Social Psychology, 53, 563-571.
Ben-Eliyahu, S., Yirmiya, R., Liebeskind, J. C., Taylor, A. N., & Gale, R. P. (1991). Stress increases metastatic spread of a mammary tumor in rate: Evidence for mediation by the immune system. Brain, Behavior and Immunity, 5(2), 193-205.
Brosschot, J. F., Benschop, R. J., Godaert, G. L. R., De Smet, M. B. M., Olff, M., Heijnen, C. J., & Ballieux, R. E. (1992). Effects of experimental psychological stress on distribution and function of peripheral blood cells. Psychosomatic Medicine, 54, 394-406.
Cacioppo, J. T. (1994). Social neuroscience: Autonomic, neuroendocrine and immune responses to stress. Psychophysiology, 31, 113-128.
Caudell, K. A., & Gallucci, B. B. (1995). Neuroendocrine and immunological responses of women to stress. Western Journal of Nursing Research, 17, 672-692.
Cohen, S. & Williamson, G. M. (1991). Stress and infectious disease in humans. Psychological Bulletin, 109(1), 5-24.
Crary, B., Hauser, S. L., Borysenko, M., Kutz, I., Hoban, C., Ault, K. A., Weiner, H. L., & Benson, H. (1983b). Epinephrine-induced changes in the distribution of lymphocyte subsets in the peripheral blood of humans. Journal of Immunology, 131, 1178-1181.
Dougher, M. J., Goldstein, D., & Leight, K. A. (1987). Induced anxiety and pain. Journal of Anxiety Disorders, 1, 259-264.
Girgis, A., Shea, J., & Husband, A. (1988). Immune and psychological responses to acute venipuncture stress. Medical Science Research, 16, 351-352.
Kappel, M., Tvede, N., Galbo, H., Haahr, P. M., Kjaer, M., Linstow, M., Klarlund, M., & Pedersen, B. K. (1991). Evidence that the effect of physical exercise on NK cell activity is mediated by epinephrine. Journal of Applied Physiology, 70, 2530- 2534.
Keller, S. E., Weiss, J. M., Schleifer, S. J., Miller, N. E., & Stein, M. (1981). Suppression of immunity by stress: Effect of a graded series of stressors on lymphocyte stimulation in the rat. Science, 213, 1397-1400.
Knapp, P. H., Levy, E. M., Giorgi, R. G., Black, P. H., Fox, B. H., & Heeren, T. C. (1992). Short-term immunological effects of induced emotion. Psychosomatic Medicine, 54, 133-148.
Kumar, M., Morgan, R., Szapocznik, J., & Eisdorfer, C. (1991). Norepinephrine response in early HIV infection. Journal of Acquired Immune Deficiency Syndromes, 4, 782-786.
Landmann, R. M. A., Muller, F. B., Perini, C., Wesp, M., Erne, P., & Buhler, F. R. (1984). Changes of immunoregulatory cells induced by psychological and physical stress: Relationship to plasma catecholamines. Clinical Experimental Immunology, 58, 127-135.
Laudenslager, M. L., Ryan, S. M., Drugan, R. C., Hyson, R. L., & Maier, S. F. (1983). Coping and immunosuppression: Inescapable but not escapable shock suppresses lymphocyte proliferation. Science, 221, 568-570.
Lewis, J. W., Shavit, Y., Terman, G. W., Nelson, L. R., Gale, R. P., & Liebeskind, J. C. (1983). Apparent involvement of opiod peptides in stress-induced enhancement of tumor growth. Peptides, 4, 635-638.
Litt, M. D. (1988). Self-efficacy and perceived control: Cognitive mediators of pain tolerance. Journal of Personality and Social Psychology, 54(1), 149-160.
Malow, R. M. (1981). The effects of induced anxiety on pain perception: A signal detection analysis. Pain, 11, 397-405.
Malow, R. M., West, J. A., & Sutker, P. B. (1987). A sensory decision theory analysis of anxiety and pain in chronic drug abusers. Journal of Abnormal Psychology, 96, 184-189.
Manning, M. M., & Wright. (1983). Self-efficacy expectancies, outcome expectancies and the persistence of pain control in childbirth. Journal of Personality and Social Psychology, 45,421-431.
Manuck, S. B., Cohen, S., Rabin, B. S., Muldoon, M. F., & Bachen, E. A. (1991). Individual differences in cellular immune response to stress. Physiological Science, 2(2), 111-115.
Marsland, A. L., Manuck, S. B., Fazzari, T. V., & Stewart, C. J. (1995). Stability of individual differences in cellular immune responses to acute psychological stress. Psychosomatic Medicine, 57(3), 295-298.
Mogil, J. S., Sternberg, W. F., & Liebeskind, J. C. (1993). Studies of pain, stress and immunity. In C. R. Chapman & K. M. Foley (Eds.), Current and emerging issues in cancer pain: Research and practice (pp. 3147). New York: Raven Press.
Murray, D. R., Irwin, M., Rearden, C. A., Ziegler, M., Motulsky, H., & Maisel, A. S. (1992). Sympathetic and immune interactions during dynamic exercise. Mediation via a |32-adrenergic-dependent mechanism. Circulation, 86, 203-213.
Naliboff, B. D., Benton, D., Solomon, G. F., Morley, J. E., Fahey, J. L., Bloom, E. T., Makinodan,T.,&Gilmore, S. L. (1991). Immunological changes in young and old adults during brief laboratory stress. Psychosomatic Medicine, 53(2), 121-132.
Naliboff, B. D., Solomon, G. F., Gilmore, S. L., & Benton, D. (1995). The effects of the opiate antagonist naloxone on measures of cellular immunity during rest and brief psychological stress. Journal of Psychosomatic Research, 39, 345-359. Page, G., Ben-Eliyahu, S., Yirmaya, R., & Liebeskind, J. C. (1993). Morphine attenuates surgery-induced enhancement of metastatic colonization in rats. Pain, 54, 21-28.
Peckerman, A., Saab, P., McCabe, P., Skyler, J., Winter, R., & Schneiderman, N. (1990). Cold pressor pain and cardiac reactivity during the forehead cold stimulation procedure. Paper presented at the 11th annual meeting of the Society for Behavioral Medicine, Chicago, IL.
Pross, H. F., Baines, M. G., Rubin, P., Schragge, P., & Patterson, M. S. (1981). Spontaneous lymphocyte-mediated cytotxicity against tumor target cells. The quantitation of natural killer cell activity. Journal of Clinical Immunology, 7(1), 51-63.
Rose, N. R., DeMacario, E. C., Fahey, J. L., Friedman, H., & Penn, G. M. (1992). Manual of clinical laboratory immunology. Washington, DC: American Society for Microbiology.
Schedlowski, M., Falk, A., Rohne, A., Wagner, T. O. F., Jacobs, R., Tewes, U., & Schmidt, R. E. (1993). Catecholamines induce alterations of distribution and activity in human NK cells. Journal of Clinical Immunology, 13, 344-351.
Schedlowski, M., Hosch, W., Oberbeck, R., Benschop, R. J., Jacobs, R., Raab, H. R., & Schmidt, R. E. (1996). Catecholamines modulate human natural killer (NK) cell circulation and function via spleen-independent p2-adrenergic mechanisms. Journal of Clinical Immunology, 756(1), 93-99.
Schultz, K. H., & Schultz, H. (1992). Overview of psychoneuroimmunological stress and intervention studies in humans with emphasis on the uses of immunological parameters. Psycho-oncology, 1, 51-70.
Shavit, Y., Terman, G. W., Lewis, J. W., Zane, C., Gale, R. P., & Liebeskind, J. C. (1986). Effects of footshock stress and morphine on natural killer lymphocytes in rats: Studies of tolerance and cross tolerance. Brain Research, 372, 382-385.
Shavit, Y., Martin, F. C., Yirmiya, R., Ben-Eliyahu, S., Terman, G. W., Weiner, H., Gale, R. P., & Liebeskind, J. C. (1987). Effects of a single administration of morphine or footshock stress on natural killer cell cytotoxicity. Brain, Behavior and Immunity, 1, 318-328.
Shoor, S. M., & Holman, H. R. (1984). Development of an instrument to explore psychological mediators of outcome in chronic arthritis. Transactions of the Association of American Physicians, 97, 325-331.
Sieber, W. J., Rodin, J., Larson, L., Ortega, S., Cummings, N., Levy, S., Whiteside, T., & Heberman, R. (1992). Modulation of natural killer cell activity by exposure to uncontrollable stress. Brain, Behavior and Immunity, 6(1), 141-156.
Spielberger, C. D., Gorusch, R. I., & Lushene, R. (1983). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.
Turk, D. C, & Melzack, R. (1992). Handbook of pain assessment. New York: Guilford Press.
Uchino, B. N., Cacioppo, J. T., Malarkey, W., & Glaser, R. (1995). Individual differences in cardiac sympathetic control predict endocrine and immune responses to acute psychological stress. Journal of Personality and Social Psychology, 69, 736-743.
Weisse, C. S., Pato, C. N., McAllister, C. G., Littman, R., Breier, A., Paul, S. M., & Baum, A. (1990). Differential effects of controllable and incontrollable acute stress on lymphocyte proliferation and leukocyte percentages in humans. Brain Behavior and Immunity, 4, 339-351.
Wiedenfield, S., O'Leary, A., Bandura, A., & Brown, S. (1990). Impact of perceived self-efficacy in coping with stressors on components of the immune system. Journal of Personality and Social Psychology, 59, 1082-1094.
Zakowski, S. G., McAllister, C. G., Deal, M., & Baum, A. (1992). Stress, reactivity and immune function in healthy men. Health Psychology, 11, 223-232.
Acknowledgment. This research was supported by a postdoctoral fellowship from the National Institute of Nursing Research (No. T32 NR07052) and by an Alumni Research grant from the Frances Payne Bolton School of Nursing, Case Western Reserve University
Lucille Sanzero Eller, PhD, RN
Offprints. Requests for offprints should be directed to Lucille Sanzero Eller, PhD, RN, Rutgers University, College of Nursing, 180 University Avenue, Newark, NJ, 07102.…