Pain and Immunologic Response to Root Canal Treatment and Subsequent Health Outcomes

Pain and Immunologic Response to Root Canal Treatment and Subsequent HealthOutcomesHENRIETTA L. LOGAN, PHD, SUSAN LUTGENDORF, PHD, H. LESTER KIRCHNER, PHD, ERIC M. RIVERA, DDS, MS, AND

DAVID LUBAROFF, PHD

Objective: This study examined the effects of pain and stress associated with a dental procedure, root canaltreatment (RCT), on natural killer cell cytotoxicity (NKCC) and the subsequent development of symptoms of upperrespiratory illness during the following month. Methods: Patients (N 5 33) were recruited from those scheduled forRCT appointments. Subjects for a non-RCT comparison group (N 5 14) were also recruited from dental clinicpatients. Peripheral blood was drawn by use of an indwelling catheter three times: just before RCT, 30 minutes afterinjection of a local anesthetic, and 30 minutes after RCT (a parallel time course was followed for the comparisongroup.) Blood was assayed for cortisol and NKCC. Subjects completed a health diary in the month after RCT.Results: Patients showed a significant increase in NKCC between baseline and RCT and a significant decrease fromRCT to after RCT, whereas the comparison group did not. The NKCC following the RCT was negatively correlatedwith the pain level during RCT (r 5 20.48, p , .01) and pain levels 2 and 6 hours after RCT (r 5 20.43, p , .05;r 5 20.44 p , .05, respectively). The patient group reported significantly more illness episodes 2 weeks after RCTthan the comparison group (Wilcoxon rank sum 5 4.78, p 5 .03). Discriminant function analysis correctly classified88% of the subjects into the illness category using predictor variables of post-RCT NKCC, stress, and pain levelsduring RCT (F(3,21) 5 8.23, p , .001). Conclusions: Transitory changes in NKCC associated with pain and stressmay be implicated in the development of infectious disease episodes after an acute stressful event. Key words: pain,stress, natural killer cell cytotoxicity, health outcomes, dental procedures.

IV 5 intravenous; NK 5 natural killer cells; NKCC 5natural killer cell cytotoxicity; RCT 5 root canal treat-ment.

INTRODUCTION

Although there is substantial evidence that acuteand chronic stress are associated with changes in im-mune function (1–5), relationships among stress, im-mune changes, and disease outcomes are less clear(6–9). In addition, there are reported associations be-tween pain and measures of immune function (10, 11),but relationships among pain, immune function, andvulnerability to disease are not well established. Inves-tigation of these relationships is particularly relevantfor patients who must undergo painful or stressfulmedical or dental procedures and who may be vulner-able to complications or negative health outcomes.There is also evidence that stress and vulnerability toillness may be linked (12). We previously found thatpatients who reported moderate to severe levels of

stress during the week before RCT reported signifi-cantly more flu and cold symptoms in the subsequentmonth (13). The results of these studies suggest a pro-spective relationship between stress and diseaseoutcome.

Several attempts have been made to establish im-mune and neuroendocrine mediators linking stressand disease outcomes. One of the most extensivelyexamined immune measures associated with stressand illness is the cytotoxic function of natural killercells (NKCC) (14–17). NK cells play an important rolein viral surveillance and in the overall defense againstviral infections (18, 19). It has been proposed that evenlimited transitory changes in the immunologic equilib-ria in which NK activity fluctuates in response topsychological stressors may have negative conse-quences for health (20). Most studies have observeddownregulation of NK activity in patients underchronic stress, but NK activity seems to be temporarilyenhanced during acute stress (21). It is also possiblethat after an extended acute stressor there is a recoveryperiod during which NK activity drops below somecritical level, and this time point may provide a win-dow for disease susceptibility (22).

Acute pain usually results in a suppression of NKCC(11, 23), although one study found results to the con-trary. Greisen et al. (24) reported a slight increase inNKCC in response to painful electric stimulation. Theincrease in NKCC was blocked when local anestheticwas applied before painful stimulation. Unmanagedsurgical pain also leads to suppression of NK cellactivity, which seems to mediate the surgery-inducedenhancement of metastatic colonization of tumor cellsin rats (25). Whether and to what extent such acute

From the Division of Public Health Services and Research (H.L.L.),University of Florida, Gainesville, Florida; Departments of Psychol-ogy (S.L.), Endodontics (E.M.R.), and Urology and Microbiology(D.L.), University of Iowa, Iowa City, Iowa; and Department of Pe-diatrics (H.L.K.), Case Western Reserve University, Cleveland, Ohio.

Address reprint requests to: Henrietta Logan, PhD, Box 100404, 1600SW Archer Rd., D8-39, Public Health Services and Research, Universityof Florida, Gainesville, FL 32610. Email: [email protected]

Received for publication March 16, 2000; revision received Octo-ber 11, 2000.

453Psychosomatic Medicine 63:453–462 (2001)

0033-3174/01/6303-0453Copyright © 2001 by the American Psychosomatic Society

pain–related changes in NK cell activity are of clinicalimportance in humans remain to be established. Somerecent data, however, suggest that short-term immuno-modulation related to pain and stress may be an im-portant predictor of short-term health-related eventssuch as vulnerability to infectious diseases (26).

Taken together these findings suggest that there areimportant relationships between pain- and stress-re-lated immune mechanisms and health outcomes.However, few studies have demonstrated immunechanges and negative health outcomes in a prospec-tive, experimentally controlled study. Because of thisgap in the literature, we prospectively investigated thepotential relationship among pain, immune changes,and the development of symptoms of infectious dis-ease in a clinic setting.

This study used RCT as an experimental model toexamine the mechanisms that link pain, stress,changes in immune function, and disease outcomes.Worldwide, RCT is a well-accepted and beneficialtreatment for diseases of the dental pulp and the se-quelae of the surrounding tissues (27). In the UnitedStates, it is estimated that the majority of adults un-dergo one or more RCTs during their lifetime (28).Although not usually life-threatening, the underlyingdisease state leading to RCT and the treatment itselfinvolve inflammation, nociception, tissue damage,and removal of tissue common to many health prob-lems. Because the RCT model was used to study areal-life stressor performed in relatively healthy peo-ple, it allowed for the investigation of mechanisms thatin more compromised patients (such as those with HIVinfection or cancer) might be masked by the diseaseprocess or its treatment. Hence, the observed changesin systemic immune markers in the RCT model aremore likely to be associated with the pain and stress oftreatment than the underlying disease state.

The disease state that ultimately requires RCT doesnot develop in a predictable pattern, and it may seemto have a sudden onset. This unpredictability can leadto intense stress among some patients. It is also com-mon for patients to perceive of RCT as painful (29) andnot to habituate despite multiple exposures (30). TheRCT model offers several advantages, including exper-imental control in which biological compromise canbe observed.

The objective of this study was to examine the ef-fects of pain and stress associated with a dental pro-cedure (RCT) on NKCC and the subsequent develop-ment of symptoms of upper respiratory illness duringthe following 4 weeks. We hypothesized that amongpatients, the acute stress of RCT would be associatedwith a rise in NKCC during the procedure and with adrop below baseline in the recovery period after the

stressor. We expected to see no change among thecomparison group. Furthermore, we expected to see asignificant negative relationship between pain andNKCC during and after the RCT and that lower NKCCwould be related to the development of infectiousdisease symptoms in the following month.

METHODS

Procedure

Subjects were recruited by phone from a list of patients sched-uled for RCT appointments. Patients were included if they werebetween 21 and 70 years of age. Patients were excluded if theyreported that they had active cancer, were taking oral steroids, werecurrently taking any type of chemotherapy agent or undergoingradiation treatment, or had undergone chemotherapy or radiation inthe last 12 months. Patients who met the criteria were given detailedinformation about the study. To control for diurnal variation ofcortisol and NKCC, only patients whose appointment was scheduledbetween 8:00 and 9:00 AM were included in this study (31). Subjectsfor the comparison group were recruited by phone from dental clinicpatients previously seen in a separate dental research project butwho were not currently scheduled for dental treatment. The sameexclusion criteria were applied to the comparison group. All assess-ments were completed by 10:30 AM. Approximately 25% of thescheduled patients either did not meet the inclusion criteria ordeclined to participate because of time constraints.

Subjects were asked to report 75 minutes before their appoint-ment. On arrival at the Dental Clinical Research Center, patientswere seated comfortably and then completed a research question-naire packet. Next, an IV catheter was inserted into the antecubitalvein of the nondominant arm. Participants viewed a 22-minuteneutral travelog video of the local area. The purpose of the video wasto produce a homogenous environment and waiting period duringwhich blood hormones could stabilize after venipuncture. The firstblood draw occurred 30 minutes after placement of the catheter. Toensure the homogeneity of experimental conditions and to reducemovement artifacts, all subjects were moved by wheelchair to thetreatment area. Comparison subjects came back (by wheelchair) tothe center for clinical studies and viewed a second neutral 60-minute video. The second blood draw occurred 30 minutes afterinjection of local anesthetic for patients receiving RCT and at anequivalent time point for comparison subjects. After the RCT or60-minute video, the patients and comparison subjects were againmoved by wheelchair to the clinical research center. All subjectsviewed a 30-minute video about gardens around the world, again toproduce a homogenous environment before the blood draw. Thethird and final blood draw occurred at the end of this video, afterwhich the IV catheter was removed. Subjects completed the researchinstruments; patients completed the pain questionnaire, and allsubjects answered written questions about the videos they hadwatched. The project nurse explained the health review diary to allsubjects, and for patients the pain chart was explained. The callschedule for the 2-week phone call to assess compliance with thehealth review diary was established. The diaries were collected 1month later. Both patients and comparison subjects were compen-sated $150 for completing the study.

Psychological Measures

Iowa Dental Control Scale. A modified version of the Iowa DentalControl Index was administered as a control variable during the

H. L. LOGAN et al.

454 Psychosomatic Medicine 63:453–462 (2001)

phone interview. Prior research has established this measure as areliable predictor of the level of distress individuals experienceduring dental treatment (32–35).

Perceived Stress Scale. The Perceived Stress Scale is a 14-itemsurvey used to assess the extent to which an individual feels incontrol vs. overwhelmed with current life stressors (36, 37). In thisstudy, the time referent was within the last month. The scale wasadministered before placement of the IV catheter as a controlvariable.

Pain and stress assessment. Baseline pain and stress measuresusing a visual analog scale format were collected from the patients.Two additional measurements of patients’ pain and stress levelswere collected. The first assessed the intensity of the pain, theunpleasantness of the pain, and the level of stress experiencedduring the dental appointment. This questionnaire was completedimmediately after RCT. The second questionnaire assessed painlevels 2 hours and 6 hours after RCT and was completed at home.Both questionnaires used a visual analog format (38). Patients wereinstructed to mark the scale, which was anchored with “no pain atall” and “worst pain (or unpleasantness or stress) imaginable,” ateach of six specific time points.

Health Review

A standardized rating instrument was used to assess the presenceof specific illness symptoms and illness episodes (39). The validityof this scale was previously shown through direct comparison withphysicians’ ratings of patients’ symptoms (40). Using publishedmethodology, symptoms were weighted with a score of 1 to 3, basedon the probability that they were indicative of an infectious illness.Summary scores for each week were computed by adding the as-signed value for each symptom present. Episodes reaching a score of3 or more were counted as an infectious illness episode. Thus, foreach week, an individual either had an illness episode or did not,and a score of 1 or 0 was accordingly assigned. Across the 4 weeksof diary completion the greatest number of episodes possible was 4and the least was 0. A total number of symptom scores across the 4weeks using the weighed value were also computed. The projectnurse called each subject 2 weeks after RCT to ensure compliancewith charting symptoms. At the 1-month follow-up appointment,the health review was collected.

Manipulation Check

All patients and subjects were asked to answer questions aboutthe their level of attention to the videos.

Blood Preparation and Assays

Three blood draws were prepared as follows: For NK assays, 30ml was drawn at each time point into heparinized Vacutainer tubes(Becton-Dickinson, Rutherford, NJ). Samples were kept at room tem-perature and assayed immediately after completion of the RCT or atan equivalent time for comparison subjects. For the cortisol assay, 5ml of blood was collected at each time point in Vacutainer tubestreated with ethylenediaminetetraacetic acid (Becton-Dickinson).These blood samples were immediately spun down, separated, andfrozen at 270°C.

Cytolytic Activity of NK Cells

Thirty milliliters of venous blood was collected in heparinizedVacutainer tubes. Lymphocytes were isolated by centrifugation on a

Histopaque gradient (specific gravity, 1.077). The peripheral bloodlymphocytes were tested for cytolytic activity against the NK-sensi-tive K562 cells and the NK-resistant, lymphokine-activated killercell–sensitive ZKBL target cells in a standard 4-hour 51Cr-releaseassay as reported previously (41). Effector peripheral blood lympho-cytes were dispensed into wells of 96-well microtitration plates intwo-fold dilutions. Target cells labeled with Na2

51CrO4 were addedto each well. The cell mixtures were incubated for 4 hours and thencentrifuged at 500 rpm. Subsequently, 150 ml of the supernatantfluid was removed and counted in a Beckman gamma counter. Lysisof K562, but not ZKBL, was indicative of NK activity of the lympho-cytes. The calculation of lytic units was used as the final measure ofNK activity (42, 43).

Cortisol

Cortisol is commonly measured to assess activation of the neu-roendocrine response. Cortisol, the primary glucocorticoid pro-duced by the adrenal cortex, was assayed by radioimmunoassay(Double Antibody Cortisol RIA, Diagnostic Products Corporation,Los Angeles, CA). The assay procedure is a standard competitiveradioimmunoassay in which 125I-labeled cortisol competes with cor-tisol in the sample for a limited number of binding sites with theantibody. Patient sample concentrations were read from a calibra-tion curve. Approximate sensitivity of the assay is 0.3 mg/dl. Inter-assay coefficients of variation range from 5.8% to 7.8%, and intraas-say coefficients of variation range from 2.4% to 4.5% (44). Expectedvalues for blood samples collected in the morning are 5 to 25 mg/dl(44).

Data Analysis

All distributions were tested for normality, and logarithmictransformations were applied when necessary. Raw (nontrans-formed) values are presented in Figure 1 for illustration purposes,but analyses of NKCC and cortisol were conducted on transformedscores. Illness symptoms and episodes were analyzed using theWilcoxon rank sum test.

To assess whether there were changes over time, differencesbetween groups, and differences between groups at each time point,a repeated-measures analysis was performed for cortisol and NKCC.

Fig. 1. NKCC by group over time. m 5 patients; □ 5 comparisonsubjects.

PAIN, HEALTH, IMMUNE CHANGES


In this study, multiple measures were obtained from each individ-ual. Measures within an individual are correlated, but measuresbetween individuals are independent. A typical linear model wasinappropriate here because independence of observations would beviolated. Therefore, for these analyses, we used a linear model usingrestricted maximum-likelihood estimation and an unstructured vari-ance-covariance structure. This type of analysis is appropriate fordata that are correlated and is performed using methods of mixedlinear models. Tests of individual time points can be obtained fromthese models. Those corresponding test statistics follow a t distribu-tion and are presented. Because observations were obtained for allsubjects at the same measurement times and because of the nonlin-earity of the response profiles, time was treated as discrete in allmodels. Each model controlled for baseline values. All post hocstatistics were adjusted for the number of comparisons.

Pearson correlations were used to examine the relationship be-tween pain, stress, and immunological measures. Discriminant func-tion analysis was performed to determine the predictability of pain,stress, and immune measures between those who had illness epi-sodes during the second week and those who did not. To obtain thediscriminant function, variables entered in the model were derivedfrom previously published research (eg, Kiecolt-Glaser et al., Refs. 1,3, 4, 21) and univariate analyses. t tests were used to comparechronic stress scores (from the Perceived Stress Scale) between thetwo groups. All statistical tests were performed using the SAS ver-sion 7 statistical package for the PC (45).

Three patients had missing NKCC data for the last two timepoints. Blood for two patients was hemolyzed from the second andthird draw, and for one patient there was inadequate blood to assayat two time points. Baseline blood from three patients was unusable.In one case the baseline blood was hemolyzed, and for two patientsthe IV catheter was not successfully placed on the first attempt.

RESULTS

Study Sample

Forty-seven patients and subjects (20 men and 27women) met the selection criteria and participated inthis study. Thirty-three patients (13 men and 20women) received RCT, and 14 subjects (7 men and 7women) formed the comparison group. The averageage was 41 years (range, 21–66 years). There were nosignificant age differences between the comparisongroup and the RCT group, nor were there differences innumber of infectious disease symptoms in the prior 2weeks.

The sample was largely made up of dental subjectswho reported low levels of dental anxiety (32). Ac-cording to our previously published research, 17 of 33patients met the criteria for low dental anxiety (30).There were no statistically significant differences be-tween patients and comparison subjects (p 5 .21) inscores on the Perceived Stress Scale, and scores werewithin previously published ranges (patients: mean 536.9, SD 5 6.6; comparison subjects: mean 5 34.0, SD5 5.9) (37).

Manipulation Check

It was important to confirm that patients and com-parison subjects did not differ in the level of engage-ment while watching the video before RCT. Using a100-point visual analog scale anchored with “didn’twatch intently” and “watched intently,” subjects wereasked to what extent they watched the video. Resultsshowed that there were no significant differences be-tween the two groups on their level of engagement (p. .4). To further verify that there were no systematicdifferences between the two groups in their attentive-ness to the videos, all patients and subjects were askedto answer questions about the videos they watched. Atleast 98% of the questions were correctly answered,with the lowest percentage being 95% and the highestbeing 100% correct, suggesting no significant differ-ences between groups in attention to videos.

NK Cell Activity

Analysis of variance showed that there were nosignificant differences between patients and compari-son subjects at baseline on NKCC (p . .99), suggestingthat the groups were equivalent at the start of thestudy. The repeated-measures analysis showed thatthere was a statistically significant main effect of time(F(2,38) 5 14.91, p , .0001) and an interaction be-tween time and group (F(2,38) 5 8.59, p , .0001).There were no significant differences between the twogroups at any of the three time points. Within-groupanalyses revealed a significant increase in NKCC be-tween baseline and 30 minutes after injection in thepatient group (adjusted p , .0001) and a significantdecrease in NKCC among patients between 30 minutesafter injection and 30 minutes after RCT (adjusted p ,.001). It is noteworthy that 55% of patients were belowtheir baseline levels of NKCC at the third time point(30 minutes after RCT). Among the comparison group,there was a marginally significant increase in NKCCfrom time 1 to time 3 (adjusted p 5 .07) (see Figure 1).

Cortisol

For the entire sample, there was a marginally sig-nificant correlation between cortisol levels at baselineand scores on the Perceived Stress Scale (r 5 0.30) (36)(see Table 1). The repeated-measures analysis showedthere was not a significant group-by-time interactioneffect for cortisol. However, there was a statisticallysignificant main effect of time (F(2,41) 5 36.90, p ,.01) but not group. Specifically, all subjects had signif-icantly lower levels of cortisol at the third blood drawthan at the baseline blood draw (p , .0001).

H. L. LOGAN et al.


Relationships Among Pain, Stress, andPhysiological Measures

Pearson correlations were used to determine rela-tionships among levels of stress, pain, and physiolog-ical measures among patients (see Table 1). Greaterstress at baseline (visual analog scale) was associatedwith higher baseline cortisol levels (r 5 0.47, p , .01).The stress during RCT was significantly correlatedwith pain levels 2 and 6 hours after RCT (r 5 0.79, p ,.01 and r 5 0.45, p , .05, respectively).

Patients reporting higher levels of pain at baselinehad lower NKCC both concurrently (r 5 20.52, p ,.01) and after RCT (r 5 244, p # .05). Patients report-ing greater pain during RCT had lower levels of NKCC(r 5 20.48, p , .01) 30 minutes after RCT. Pain levels2 and 6 hours after RCT were also negatively correlatedwith the final NKCC (r 5 20.43, p , .05 and r 5 20.44,p , .05, respectively).

Illnesses

An important question in this study was whetherthe stress, pain, and immune changes observed as aresult of RCT served as risk factors for the develop-ment of infectious illness symptoms during the subse-quent 4 weeks. To test this question, comparisons weremade between the groups on self-reported illnesssymptoms and episodes.

There was a significant difference between dentalpatients and the comparison group in the number ofself-reported illness symptoms as recorded in the diaryduring the 4-week period after RCT (see Table 2). Ex-amining the total diary illness data, we observed thatthe differences between the two groups occurred pri-marily during the second week after RCT, when thepatient group reported more illness episodes than the

comparison group. There were no significant differ-ences between groups in self-reported illness episodesduring the first, third, or fourth weeks.

Thirteen of the patients developed an illness epi-sode during the second week after RCT, comparedwith only 2 subjects in the comparison group. Giventhis distribution of episodes, further analyses weredone within the patient group to establish potentialpredictor variables of illness episodes at the secondweek.

Patients who had self-reported illness episodes dur-ing the second week had significantly higher painscores 6 hours after RCT (illness group: mean 5 25.2,SD 5 24.1; nonillness group: mean 5 10.1, SD 5 10.9;F(1,32) 5 4.37, p # .05) and more stress during theirdental appointment (illness: mean 5 35.07, SD 5 26.5;nonillness: mean 5 12.7, SD 5 18.8; F(1,32) 5 8.12, p, .05). There was also a trend toward greater pain 2hours after RCT in the illness group (p 5 .06). NKactivity 30 minutes after RCT was significantly loweramong individuals who later reported illness episodesthan those who did not (illness: mean 5 108.7 lyticunits, SD 5 82.30; nonillness: mean 5 209.9 lyticunits, SD 5 176.3; F(1,32) 5 5.34, p , .05). There wasa trend toward higher cortisol levels at the final blooddraw in the illness group as compared with the non-

TABLE 1. Correlations Among Main Study Variables

Cortisol NKCC

Baseline During RCT After RCT Baseline During RCT After RCT

Total sample (N 5 47)a

Chronic stress 0.30* 0.17 0.24 0.09 0.14 20.1Patients only (N 5 33)

Anticipated pain 0.33 0.22 0.46* 0.16 0.30 0.09Anticipated unpleasantness 0.21 0.12 0.42* 0.02 0.18 0.02Pain before RCT (baseline) 20.12 20.15 20.02 20.52* 20.30 20.44*Stress before RCT (baseline) 0.47* 0.27 0.26 0.07 0.14 20.05Pain during RCT 20.14 20.1 0.23 20.41* 20.36 20.48*Unpleasantness during RCT 20.02 20.03 0.33 20.33 20.21 20.37*Stress during RCT 0.22 0.16 0.32 20.18 20.06 20.15Pain 2 h after RCT 20.17 20.22 0.16 20.27 20.15 20.43*

a Equivalent time points for patient and comparison groups.* p # .05.

TABLE 2. Mean Illness Symptoms and Episodes Between DentalPatients and Comparison Subjects

Dental Patients(N 5 33)

Comparison Subjects(N 5 14)

Total illness symptoms (SD)a 35.87 (44.6) 7.14 (12.0)Illness episodes during second

week (SD)b0.4 0.1

a Wilcoxon rank sum, 7.37; p 5 .007.b Wilcoxon rank sum, 4.78; p 5 .03.



illness group (illness: mean 5 14.89, SD 5 7.1; noillness: mean 5 11.15, SD 5 3.9; F(1,41) 5 3.17, p 5.09). It should be noted that a review of the dentalcharts showed that the two groups did not differ onpulpal or periapical diagnosis. For instance, therewere no more cases of irreversible pulpitis among theillness group than among the nonillness group.

Classification of Illness

To further understand the relationship between therisk factors and self-reported illness episodes, dis-criminant function analysis was used to classify pa-tients into groups of those who reported an illnessepisode and those who reported no illness episodeduring the second week after RCT. Based on the uni-variate analysis and variables identified in prior stud-ies (eg, Herbert and Cohen, Ref. 14), the followingvariables were entered into the discriminant function:pain 2 and 6 hours after RCT, pain level immediatelybefore RCT, stress experienced during RCT, and NKCC30 minutes after RCT. Variables that failed to achieve aprobability of ,.15 were eliminated from the model.Subjects who did not have complete data were ex-cluded from the analysis.

The final function included pain 2 hours after RCT,the level of stress reported during RCT, and NKCC 30minutes after RCT. The overall model was statisticallysignificant (F(3,21) 5 8.23, p , .001) and classified allbut three subjects into the correct category (see Table3). The following variables provided significant inde-pendent contributions to the final model: stress expe-rienced during RCT (F(1,21) 5 23.62, p , .001), pain 2hours after RCT (F(1,21) 5 15.93, p , .001), and NKCClevel 30 minutes after RCT (F(1,21) 5 6.84, p , .05).

DISCUSSION

This study used a naturalistic stressor, RCT, to testthe relationships among pain, stress, NKCC, and post-RCT illness vulnerability. The main findings of thisstudy were that the patient group developed signifi-cantly more self-reported illness episodes 2 weeks af-ter the procedure than the comparison group and thatthe predictor variables of NKCC after treatment, stress,and pain correctly classified 88% of the patients into acategory of illness episode or no illness episode. Giventhat millions of similarly stressful and painful medi-cal/dental procedures are done each year, these find-ings are particularly relevant for patients who may bevulnerable to complications or negative healthoutcomes.

In addition, we observed a negative relationshipbetween self-reported pain and NKCC at multiple timepoints, including higher baseline pain and lower con-current NKCC and lower NKCC 30 minutes after RCT.Patients who reported sustained high levels of pain 2and 6 hours after treatment were also those who hadlower NKCC 30 minutes after treatment. This study isone of the first prospective human studies to show thatpain may be associated with suppression of NK cellactivity. In contrast to the statistically significant neg-ative relationships between NKCC and pain, there wasnot a statistically significant relationship betweenNKCC and self-reported stress during RCT.

Effects of RCT on NKCC and Cortisol

As shown in Figure 1, the NKCC response patternbetween and within the groups across the three blooddraws is complex, and there is large variability withinthe groups at any given time point. Prior research hasshown that patients find RCT to be highly stressful andcategorize it along with oral surgery treatment as pro-ducing the highest level of dental stress (46). Patientsalso frequently report that RCT is painful (30). Changesin NKCC in response to stress (21) and pain (25) havebeen previously reported. During treatment, as shownin Figure 1, patients had greater NKCC 30 minutesafter injection than at baseline or after RCT. It is prob-able that the changes in NKCC we observed are a resultof the treatment (RCT). The changes may be due to thestress, due to the dental injection, related to the short-term effect of the local anesthetic, or a combination ofthese. This pattern is consistent with those found byGreisen et al. (24), who observed a slight increase inshort-term NKCC after acute pain stimuli. This in-crease was blocked when local anesthetic was admin-istered before the pain stimuli, suggesting that acutepain produced the increase. On the other hand, in this

TABLE 3. Discriminant Function Analysisa Classifying SubjectsWith and Without Illness 2 Weeks After RCT Using Pain 2 Hours

After RCT, Self-Reported Stress During RCT, and NKCC asPredictor Variables

Week 2

Model Classification

Total Actualby Category

NoIllness

Illnessepisode

N % N %

Actual presence, noillness episode

14 93 1 6.6 15

Actual presence, illnessepisode

2 20 8 80 10

Total classified intocategory by model

16 64 9 36 25

a F(3,21) 5 8.23, p , .001.

H. L. LOGAN et al.


study we observed negative correlations between self-reported pain and NK cell activity at multiple timepoints. This latter finding is consistent with reports ofLiebeskind (11), Kiecolt-Glaser et al. (23), and Page etal. (10) showing immunosuppression in response topain.

The mechanism by which pain may disrupt theequilibrium of the immune system has yet to be fullydescribed. One proposed mechanism is glucocorti-coid-induced immune suppression. However, consis-tent with Page et al. (25), we did not find support for amediational role of cortisol between pain and NKCC.

It is now believed that the adrenergic effects on NKcells consist of two components: first, a large transitoryincrease in NK cells leading to an increase in activityper unit of blood, followed by a suppression of activityat about 50 minutes after the stressor (47). Consistentwith this explanation, we observed an overall patternthat shows an initial rise 30 minutes after dental in-jection followed by a decrease 30 minutes after RCT.The reader is reminded, however, that Shakhar andBen-Eliyahu’s (47) study was an animal study, so gen-eralizations may not apply to humans. On the otherhand, 55% of the patients showed a decrease in NKCCbelow their baseline at this third blood draw, whichwas approximately 11⁄2 hours after the injection. Notonly is this pattern consistent with that predicted byShakhar and Ben-Eliyahu, but this drop may parallelthat reported by Benschop et al. (22), who observedthat the acute rise of NKCC after a stressor was fol-lowed by a sharp decrease. NKCC suppression aftersurgical stress reaches its nadir at 24 hours, suggestingthat following the recovery pattern longer than 30 min-utes after RCT may reveal important patterns of post-stressor NKCC dynamics (48). Whether this suppres-sion has a long-term impact on overall vulnerability todisease remains to be determined, but the illness datafrom this study suggest that it may.

It should be noted that the trend among the compar-ison group toward greater NKCC at the final blooddraw as compared with baseline (p 5 .07) is consistentwith other research showing a rise in NK cell activityafter planned relaxation interventions (49, 50). It isunclear, however, whether passively watching videoswould produce the same effect as progressive musclerelaxation, focused breathing, or imagery on immunefunction, but the possibility cannot be ruled out.

Pain, Stress, and Illness

Patients reported more illness symptoms during theentire follow-up period and more illness episodes dur-ing the second week than comparison subjects. Amongthose who became ill, reported stress level during RCT

and pain levels after RCT were significantly higherthan among those who did not become ill. In addition,NKCC of patients who developed an illness episodewas significantly lower after treatment than amongthose who did not develop an illness. Although admit-tedly speculative, the development of illness 2 weeksafter RCT in this study may point to the diminishedcapability of NK cells among those who had experi-enced the greatest RCT pain. The data we present heresuggest a relationship between acute changes in NKCCand the development of upper respiratory tract infec-tion; however, in the absence of medical documenta-tion of reported illness, this conclusion remains tenta-tive. In addition, collection of a more detailed historyof upper respiratory symptoms in the period before thestudy would have helped to rule out other health-related factors that might have affected illnessvulnerability.

The discriminant function model showed thathigher levels of pain and acute stress, along with lowerNKCC, were associated with the presence of a self-reported illness episode during the second week. It hasbeen previously shown that stress and upper respira-tory illness are associated (51), and there are publisheddata showing a link between pain and suppression ofNKCC (25). This prospective suggests a link in humansamong pain, suppression of NKCC, and the develop-ment of cold and flu symptoms. Readers should, how-ever, interpret the results of the discriminant functioncautiously because of the relatively low subjects-to-variable ratio (52). If these data hold in future investi-gations, they will provide additional evidence thateffective management of pain and stress are importantconsiderations to ensure the well-being of the patientboth immediately after and in the weeks that followmedical interventions.

Limitations

One alternate interpretation of these data is that thedisease state that underlies the need for RCT mayproduce an impairment of the immune system andmay be the major contributing factor to the differencesin NKCC and the observed illness symptoms. Al-though this is plausible, there are several factors thatsuggest otherwise. There were no significant differ-ences in baseline NKCC between patients and compar-ison subjects. The dental diagnoses between those pa-tients who became ill and those who did not were notdifferent. Even if there were differences in the etiologyof the dental disease that we did not identify, no rela-tionship has been reported between bacterial infec-tions of odontogenic origin (ie, periapical endodonticinfections) and systemic disease states other than in-



fections states resulting from the acute uncontrolledspread of the initial bacterial infection (53). Neverthe-less, bacteria have been shown to alter the immunehost response, and we cannot rule out the possibilitythat our results are related to the oral disease underly-ing the need for RCT. In addition, strict asepsis waspracticed in the study clinic. These aseptic proceduresmake it unlikely that patients were exposed to mi-crobes during treatment that could account for an in-crease in infectious disease symptoms. On the otherhand, with the present design we cannot rule out thepossibility that some factor related to the RCT mayhave increased patient’s vulnerability to illness. Inretrospect it would have been useful to include a “non-operated” comparison group. In fact, it is a challengeto find any dental procedure that is not invasive be-cause even cleaning requires “probing” of the soft tis-sue, frequently without the benefit of local anesthetic.This probing often causes bleeding, and the process isdescribed by some patients as being as painful as hav-ing fillings placed (29). Without flow cytometry datafor the number of NK cells, we cannot rule out thepossibility that the alteration in NK cell activity seenin the RCT group may have been somewhat influencedby changes in NK cell numbers. However, as Shakharand Ben-Eliyahu (47) have demonstrated, in vivo b-ad-renergic stimulation suppresses NK activity indepen-dent of the transitory increase in peripheral NK num-bers. Additionally, Z. K. Ballas (University of Iowa,unpublished observations, 1997) reports no significantcorrelation between NK cell number and NK activity(54). Thus, it is unlikely that the alterations we ob-served are purely reducible to changes in cell migra-tion patterns, but that possibility cannot be ruled out.

Second, pain, stress, and illness variables are basedon self-reports. Although illness data were elicited bya nurse and checked carefully with interview probes,these data may reflect a subject’s tendency to overre-port pain, stress, and illness symptoms (55). It is alsopossible that a propensity toward a negative reportbias may be a contributing factor in the low NKCCobserved in such individuals (56, 57). Medical docu-mentation of the illness symptoms would havestrengthened the reported relationships. However, therelationships of the pain and stress reports with thebiological measures in a prospective study providesome objective corroboration of their validity.

SUMMARY

This study examined the effects of an acute stressor(RCT) on NKCC and the subsequent development ofsymptoms of upper respiratory illness during the fol-lowing 4 weeks. Patients showed greater overall illness

symptoms and illness episodes in the second weekafter RCT than comparison subjects. Discriminantfunction analysis showed that a combination of ele-vated stress during RCT coupled with a depressedNKCC and elevated pain after RCT resulted in thecorrect assignment of 88% of the patients into thecorrect illness episode category. This finding suggeststhat even limited transitory changes in immunologicequilibria may be of consequence to health (14). TheRCT model provides an innovative paradigm in whichto pursue mechanisms that may underlie the recoveryperiod following exposure to pain and stress and sub-sequent health consequences.

This work was supported by a research grant fromthe Fetzer Institute, Kalamazoo, Michigan. Specialthanks to John Sheridan for his support in the designof this study and comments on the manuscript. Wealso acknowledge Ann Bossen for help in data collec-tion, Chris Logan for assistance with the laboratoryassays, and Sue-Young Hong and Mike Andrews foraid in training our staff.

REFERENCES

1. Kiecolt-Glaser JK, Malarkey WB, Chee M, Newton T, CacioppoJT, Mao HY, Glaser R. Negative behavior during marital conflictis associated with immunological down-regulation. PsychosomMed 1993;55:395–405.

2. Matthews K, Caggiula AR, McAllister CG, Berga SL, Owens JF,Flory JD, Miller AL. Sympathetic reactivity to acute stress andimmune response in women. Psychosom Med 1995;57:564–71.

3. Kiecolt-Glaser JK, Glaser R, Gravenstein S, Malarkey WB.Chronic stress alters the immune response to influenza virusvaccine in older adults. Proc Natl Acad Sci U S A 1996;93:3043–7.

4. Glaser R, Kiecolt-Glaser JK. Stress-associated immunemodulation: relevance to viral infections and chronic fatiguesyndrome. Am J Med 1998;105:35S–42S.

5. Matthews K, Cagiula AR, McAllister CG, Berga SL, Owns JF,Flory JD, Miller AL. Sympathetic reactivity to acute stress andimmune response in women. Psychosom Med 1995;57:564–71.

6. Levy SM, Fernstrom J, Herberman RB, Whiteside T, Lee J, WardM, Massoudi M. Persistently low natural killer cell activity andcirculating levels of plasma beta endorphin: risk factors forinfectious disease. Life Sci 1991;48:107–16.

7. Ironson G, Friedman A, Klimas N, Antoni M, Fletcher MA,LaPerriere Simoneau J, Schneiderman N. Distress, denial, andlow adherence to behavioral interventions predict faster diseaseprogression in gay men infected with human immunodeficiencyvirus. Int J Behav Med 1994;1:90–105.

8. Cohen S, Doyle WJ, Skoner DP, Rabin RS, Gwaltney JM. Socialties and susceptibility to the common cold. JAMA 1997;277:1940–4.

9. Fawzy FI, Fawzy NW, Hyun CS, Elashoff R, Guthrie D, Fahey JL,Morton DL. Malignant melanoma: effects of an early structuredpsychiatric intervention, coping and affective state on recur-rence and survival 6 years later. Arch Gen Psychiatry 1993;50:681–9.

10. Page GG, Ben-Eliyahu S, Liebeskind JC. The role of LGL/NK

H. L. LOGAN et al.


cells in surgery-induced promotion of metastasis and its atten-uation by morphine. Brain Behav Immun 1994;8:241–50.

11. Liebeskind JC. Pain can kill [editorial]. Pain 1991;44:3–4.12. Cohen S, Tyrrell D, Smith AP. Psychological stress and suscep-

tibility to common cold. N Engl J Med 1991;325:606–12.13. Wilton J, Logan HL. Relation between infectious disease symp-

toms and sequelae following dental treatment. J Dent Res 1997;76:2870.

14. Herbert TB, Cohen S. Stress and immunity in humans: a meta-analytic review. Psychosom Med 1993;55:364–79.

15. Lischinsky S, Greenberg A, Levi N, Zinder O. Experimentalinduction and termination of acute psychological stress in hu-man volunteers: effects on immunological, neuroendocrine, car-diovascular, and psychological parameters. Brain Behav Immun1998;12:34–52.

16. Schedlowski M, Jacobs R, Alker J, Prohl F, Stratmann G, RichterS, Hadicker A, Wagner TO, Schmidt RE, Tewes U. Psychophys-iological, neuroendocrine and cellular immune reactions underpsychological stress. Neuropsychobiology 1993;28:87–90.

17. Brosschot JF, Benschop RJ, Godaert GLR, DeSmet MB, Olff M,Heijnen CJ, Ballius RE. Effects of experimental psychologicalstress on distribution and function of peripheral blood cells.Psychosom Med 1992;54:394–406.

18. Herberman RB, Ortaldo JR. Natural killer cells: their roles indefenses against disease. Science 1981;214:24–30.

19. Easterling BA, Kiecolt-Glaser JK, Glaser R. Psychosocial modu-lation of cytokine-induced natural killer cell activity in olderadults. Psychosom Med 1996;58:264–72.

20. Naliboff BD, Solomon GF, Gilmore SL, Fahey JL, Benton D, PineJ. Rapid changes in cellular immunity following a confronta-tional role-play stressor. Brain Behav Immun 1995;9:207–19.

21. Kiecolt-Glaser JK, Cacioppo JT, Malarkey WB, Glaser R. Acutepsychological stressors and short-term immune changes: what,why, for whom, and to what extent? Psychosom Med 1992;54:680–5.

22. Benschop RJ, Benschop JF, Godaert GLR, DeSmet MB, GreenenR, Olff M, Heijnen CJ, Ballieux RE. Chronic stress affects immu-nologic but not cardiovascular responsiveness to acute psycho-logical stress in humans. Am J Physiol 1994;266:R75–80.

23. Kiecolt-Glaser JK, Page GG, Marucha PT, MacCallum RC, GlaserR. Psychological influences on surgical recovery: perspectivesfrom psychoneuroimmunology Am Psychol 1998;53:1209–18.

24. Greisen J, Hokland M, Grofte T, Hansen PO, Jensen TS, VilstrupH, Tonnesen E. Acute pain induces an instant increase in nat-ural killer cell cytotoxicity in humans and this response isabolished by local anaesthesia. Br J Anaesth 1999;83:234–40.

25. Page GG, McDonald JS, Ben-Eliyahu S. Pre-operative versuspostoperative administration of morphine: impact on the neu-roendocrine, behavioural, and metastatic-enhancing effects ofsurgery. Br J Anaesth 1998;81:216–23.

26. Glaser R, Rabin B, Chesney M, Cohen S, Natelson B. Stress-induced immunomodulation: implications for infectious dis-eases? JAMA 1999;281:2268–70.

27. Ray HA, Trope M. Periapical status of endodontically treatedteeth in relation to the technical quality of the root filling andthe coronal restoration. Int Endod 1995;28:12–8.

28. Buckley M, Spangberg LS. The prevalence and technical qualityof endodontic treatment in an American subpopulation. OralSurg Oral Med Oral Pathol Oral Radiol Endod 1995;79:92–100.

29. Wong M, Lytle WR. A comparison of anxiety levels associatedwith root canal therapy and oral surgery treatment. J Endod1991;17:461–5.

30. Logan H, Baron RS, Kohout F. Sensory focus and procedural

information as therapeutic treatment for dental stress. Psycho-som Med 1995;57:475–84.

31. Kronfol Z, Nair M, Zhang Q, Hill EE, Brown MB. Circadianimmune measures in healthy volunteers: relationship to hypo-thalamic-pituitary-adrenal axis hormones and sympathetic neu-rotransmitters. Psychosom Med 1997;59:42–50.

32. Logan H, Baron R, Keeley K, Law A, Stein S. Desired and feltcontrol as mediators of stress in a dental setting. Health Psychol1991;10:352–9.

33. Law A, Logan H, Baron R. Desire for control, felt control, andstress inoculation training during dental treatment. J Pers SocPsychol 1994;67:926–36.

34. Baron RS, Logan HL. Desired control, felt control and dentalpain: recent finding and remaining issues. Motivation Emotion1993;17:181–204.

35. Baron RS, Logan H, Hoppe S. Emotional and sensory focus asmediators of dental pain among patients differing in desired andfelt dental control. Health Psychol 1993;12:381–9.

36. Cohen S, Kamarck T, Mermelstein R. A global measure of per-ceived stress. J Health Soc Behav 1983;24:385–96.

37. Naquin MR, Gilbert GG. College students’ smoking behavior,perceived stress, and coping styles. J Drug Educ 1996;26:367–76.

38. Huskisson E. Measurement of pain. J Rheumatol 1982;9:768–9.39. Orts K, Sheridan JF, Robinson-Whelen S, Glaser R, Malarkey

WB, Kiecolt-Glaser JK. The reliability and validity of a struc-tured interview for the assessment of infectious illness symp-toms. J Behav Med 1995;18:517–29.

40. LaRue A, Bank L, Jarvik L, Hetland M. Health in old age: how dophysicians’ ratings and self-ratings compare? J Gerontol 1979;34:687–91.

41. Christensen AJ, Edwards DL, Wiebe JS, Benotsch EG, McKelveyL, Andrews M, Lubaroff DM. Effect of verbal self-disclosure onnatural killer cell activity: moderating influence of cynical hos-tility. Psychosom Med 1996;58:150–5.

42. Pross HF, Maroun JA. The standardization of NK cell assays foruse in studies of biological response modifiers. J Immunol Meth-ods 1984;68:235–49.

43. Bryant J, Day R, Whiteside TL, Herberman RB. Calculation oflytic units for the expression of cell-mediated cytotoxicity. J Im-munol Methods 1992;146:91–103.

44. Diagnostic Product Corporation. Cortisol double antibody. LosAngeles: Diagnostic Product Corporation; 1992.

45. SAS Institute. SAS version 7. Cary (NC): SAS Institute Inc; 1997.46. Wong M, Lytle WR. A comparison of anxiety levels associated

with root canal therapy and oral surgery treatment. J Endod1991;17:461–5.

47. Shakhar G, Ben-Eliyahu S. In vivo beta-adrenergic stimulationsuppresses natural killer activity and compromises resistance totumor metastasis in rats. J Immunol 1998;260:3251–8.

48. Beilin B, Shavit Y, Hart J, Mordashov B, Cohn S, Notti I, BesslerH. Effects of anesthesia based on large versus small doses offentanyl on natural killer cell cytotoxicity in the perioperativeperiod. Anesth Analg 1996;82:492–7.

49. Van Rood YR, Bogaards M, Goulmy E, van Houwelingen HC.The effects of stress and relaxation on the in vitro immuneresponse in man: a meta-analytic study. J Behav Med 1993;16:163–81.

50. Ironson G, Field T, Scafidi F, Hashimoto M, Kumar M, Kumar A,Price A, Goncalves A, Burman I, Tetenman C, Patarca R,Fletcher MA. Massage therapy is associated with enhancementof the immune system’s cytotoxic capacity. Int J Neurosci 1996;84:205–17.

51. Cohen S, Line S, Manuck SB, Rabin BS, Heise ER, Kaplan JR.Chronic social stress, social status, and susceptibility to upper



respiratory infections in nonhuman primates. Psychosom Med1997;59:213–21.

52. Hair JF Jr, Anderson RE, Tatham RL, Black RC. Multivariate dataanalysis. 5th ed. Upper Saddle River (NJ): Prentice Hall; 1998.

53. Walton RE, Torabinejad M. Principles and practice of endodon-tics. 2nd ed. Philadelphia: WB Saunders; 1996.

54. Bonavida B, Katz J, Gottlieb M. Mechanism of defective NK cellactivity in patients with acquired immunodeficiency syndrome(AIDS) and AIDS-related complex. I. Defective trigger on NKcells for NKCF production by target cells, and partial restorationby IL 2. J Immunol 1986;137:1157–63.

55. Watson D, Pennebaker JW. Health complaints, stress, anddistress: explore the central role of negative affectivity. PsycholRev 1989;96:234–54.

56. Stone AA, Neale JM, Cox DS, Napoli A, Valdimarsdottir H,Kennedy-Moore E. Daily events are associated with a secretoryimmune response to an oral antigen in men. Health Psychol1994;13:440–6.

57. Valdimarsdottir HB, Bovbjerg DH. Positive and negative mood:association with natural killer cell activity. Psychol Health1997;12:319–27.

H. L. LOGAN et al.


Pain and Immunologic Response to Root Canal Treatment and Subsequent Health Outcomes

Documents

Transcript of Pain and Immunologic Response to Root Canal Treatment and Subsequent Health Outcomes