Secondary Logo

Journal Logo

Analgesia: Research Report

Experimental Heat Pain for Detecting Pregnancy-Induced Analgesia in Humans

Carvalho, Brendan MBBCh, FRCA*; Angst, Martin S. MD*; Fuller, Andrea J. MD; Lin, Eric MD; Mathusamy, Anbu D. MD§; Riley, Edward T. MD*

Author Information
doi: 10.1213/01.ane.0000239224.48719.28
  • Free

Some animal studies have suggested that sensitivity to pain decreases during pregnancy (1–3) because of the progressive activation of endogenous pain inhibitory systems. Gintzler (1) demonstrated a progressive elevation of electrical nociceptive thresholds in pregnant rats, most notably 1–2 days before parturition. Similarly, thresholds to nociceptive heat were prominently increased in sows before parturition, but steadily decreased during the first postpartum week (2). Systemic or intrathecal administration of an opioid antagonist abolished pregnancy-induced antinociceptive effects (1,3), suggesting a dominant role for the endogenous opioid system which, in turn, is most likely modulated by increased ovarian sex steroids (4–6).

Although pregnancy-induced antinociceptive effects have been well documented in animals, only few studies have explored this phenomenon in humans and findings are inconsistent. While some authors reported decreased sensitivity to pressure and ice water-induced pain during pregnancy, others were unable to detect such differences in models of pressure or heat pain (7–13). Discrepancies between study findings may reflect various sample sizes, lack of a control group, use of different pain rating scales, and experimental assessments occurring at different time points during and after pregnancy (7–13). Studies reporting negative results either lacked a control group or used an insensitive categorical outcome measure (10,11).

The purpose of this study was to determine sensitivity to heat and ice water-induced pain in pregnant women using continuous effect measures before induction of labor and postpartum and compare those results with matched nonpregnant controls undergoing the same tests twice within a 24-h time interval.


Subjects and Setting

After obtaining IRB approval and written informed consent, we enrolled 20 healthy, term parturients and 19 healthy female volunteers in this prospective study. The study was conducted at Lucile Packard Children's Hospital, Stanford University School of Medicine (Stanford, CA). Parturients and volunteers were enrolled over a 6-mo period in 2003.

The pregnant subjects' inclusion criteria were nulliparous or multiparous ASA physical status class I or II women, aged between 18 and 40 yr with singleton pregnancy of more than 37 completed weeks gestational age presenting for induction of labor. The volunteers' inclusion criteria were healthy, 18 to 40-year-old nonpregnant women. Exclusion criteria were diabetes, any recent analgesic consumption with 48 h of testing, history of chronic analgesic medication use, and/or the onset of spontaneous labor defined as any cervical change with associated contractions.

Study Protocol

Pain threshold (lowest stimulus intensity causing pain) and tolerance (stimulus intensity causing maximum tolerable pain) were determined using models of experimental heat pain and ice water-induced pain (cold pressor pain test). The experimental pain test algorithm was explained to the subjects using a standardized script. All tests were performed in an isolated and quiet room with an ambient temperature comfortable to the study subjects. Testing was conducted on two separate sessions separated by 1–2 days. Pregnant subjects were tested on arrival for induction of labor (prior to any labor pain) and 24–48 h postpartum. Healthy volunteers were tested on two occasions separated by 24 h. The order of testing in all subjects was cold pressor pain followed by heat pain after an interval of approximately 10 min. The cold pressor and the heat pain testing were performed on opposite arms.

Heat Pain Testing

A thermal sensory analyzer was used to determine sensitivity to heat-induced pain (TSA 2001; Medoc Advanced Medical Systems, Minneapolis, MN). A 16 × 16-mm thermode was placed in full contact on the mid-forearm. After equilibration between the skin and the thermode at 35°C, the thermode temperature was increased at a rate of 1°C/s. Subjects were asked to push the button of a hand-held device as soon as their perceived pain became intolerable. This resulted in the immediate cooling of the thermode and registration of the maximum thermode temperature. Two training cycles were performed before three actual measurements were taken. The average of three maximum thermode temperatures was recorded as the heat pain tolerance (HPT). The inter-stimulus interval was 30 s. The maximum thermode temperature was limited to 53°C to prevent tissue damage. Subjects unable to perform consistently during training cycles (repeated measurements of pain tolerance not within 1°C) were excluded from the study.

Cold Pain Test

Subjects were asked to immerse one hand and forearm into a container filled with 12 L of ice water. To maintain testing consistency among subjects, care was taken to assure that the whole palm stayed in full contact with the bottom of the container throughout the immersion. A pump steadily circulated the ice water at a constant temperature of 1–2°C. Cold pain threshold was determined by measuring the time between immersing of the hand and the onset of pain. Cold pain tolerance was determined by measuring the time between immersing and withdrawing the hand due to intolerable pain. Cold pain threshold and tolerance were measured twice, and the average of the cold pain threshold and tolerance were recorded. If the repeated measurements were more than 20% of the first measurements, then a third cold pain threshold and tolerance was determined and the average of all readings was recorded. The inter-stimulus interval was 5 min.

Statistical Analysis

A sample size of 20 subjects per group was chosen, as the aim of the study was to determine whether either of the two explored experimental pain models was sensitive enough to detect differences in pain sensitivity in a relatively small sample of pregnant and nonpregnant women. An experimental pain model able to discriminate pain sensitivity in a reasonably small sample of pregnant and nonpregnant women would be practical and useful for further study of mechanistic aspects of pregnancy-induced analgesia.

Parametric nonpaired tests were used to determine whether the difference in mean was significant between the groups. Paired tests were used to compare the results obtained on the first and second study days within each group. Data were normalized with logarithmic transformation, where necessary. Data are presented as mean ± sd unless otherwise stated. A P value <0.05 was considered statistically significant and Bonferroni's correction was used to account for multiple between-group and within-group comparisons.



Data from all 39 subjects (n = 20 pregnant; n = 19 control) enrolled in this study were collected and analyzed. Five pregnant subjects were lost to follow- up and did not have a second testing. Fourteen of the 15 postpartum subjects undergoing a second test session had received postpartum analgesics (12 subjects received ibuprofen 600 mg, one subject hydrocodone 10 mg, and one subject oxycodone 10 mg). Postpartum analgesics were consumed 12 ± 6 h (mean ± sd) before the second test session.

The demographic and obstetric data of the study subjects are outlined in Table 1. There was a statistically significant variation in age and weight between the study groups. The mean difference in age was 6 yr and the mean difference in weight was 22 kg (Table 1). Forty percent of nonpregnant women were in the follicular phase of their menstrual cycle (Day 1–14) and 60% were in the luteal phase (Day 15–28). The ethnicity of the study subjects was as follows: 51% were Caucasian, 23% were Asian, 18% were Hispanic, and 5% were African Americans. The ethic diversity was similar between study groups. Of the pregnant subjects, 45% were nulliparous and 55% multiparous. Seventeen pregnant subjects underwent normal spontaneous vaginal delivery and three subjects had cesarean delivery. The median time between testing sessions was 24 and 39 h in nonpregnant and pregnant study subjects, respectively.

Table 1
Table 1:
Demographic and Obstetric Data

Heat Pain Testing

HPT was increased in pregnant versus nonpregnant subjects on both study sessions (Fig. 1; P < 0.05, power 0.65). HPT was 49.0 ± 1.2°C vs 50.0 ± 1.0°C on session one, and 49.2 ± 1.2°C vs 50.1 ± 0.7°C (mean ± sd) on session two in nonpregnant versus pregnant subjects. Within each group of subjects (nonpregnant or pregnant), the HPT was not different on study session one and two (Fig. 1).

Figure 1.
Figure 1.:
Heat pain tolerance was significantly increased in pregnant women at term (P) when compared with that in nonpregnant controls (NP). These differences were similar before induction of labor (1) and 1–2 days after delivery (2). Within the pregnant and nonpregnant groups, no changes of heat pain tolerance were observed when comparing results of the first and second study day. Values presented as mean ± sd. *P < 0.05 (pregnant subjects (P) compared with nonpregnant (NP) controls).

Cold Pain Testing

No significant differences were detected between nonpregnant versus pregnant women when comparing the cold pain threshold and tolerance. Cold pain threshold was 9.6 s [6.2–12.0] vs 10.8 s [7.6–12.7] on session one and 8.4 s [6.8–10.3] vs 9.5 s [7.6–12.7] on session two (median and interquartile range) in nonpregnant versus pregnant subjects. Cold pain tolerance was 20.5 s [12.9–27.2] vs 18.6 s [13.0–26.5] on session one and 17.1 s [12.2–31.9] vs 16.4 s [12.5–29.8] on session two in nonpregnant versus pregnant subjects (Fig. 2). Cold pain threshold and tolerance were not significantly different between the first and second test sessions within each group (nonpregnant or pregnant) of subjects.

Figure 2.
Figure 2.:
Cold pain testing tolerance or endurance, i.e., the time until the immersed hand was removed from an ice water tank, was not different between pregnant women at term (P) and nonpregnant controls (NP). Results obtained on the first (1) and second day study (2) were also similar. Values presented as median and interquartile range.


To our knowledge, this is the first study to document decreased sensitivity to experimental heat pain in pregnant women at term when compared with nonpregnant control subjects. Our study findings in humans are consistent with animal data (2). The observed increase in HPT of about 1°C at term seems clinically relevant, as it reflects a similar analgesic effect magnitude as seen after epidural administration of 5 mg morphine (14).

Two previous studies in humans using experimental heat pain reported negative results [Table 2; (10,11)]. One study did not include a nonpregnant control group and assessed pain sensitivity before and 2–3 days after delivery. On the basis of our results and animal data, pain sensitivity may not have returned to nonpregnant levels within 2–3 days postpartum, which may explain the negative results (2,10). A second study included controls, but used a 3-point categorical scale for rating the magnitude of pain (11). A 3-point categorical pain rating scale is of questionable sensitivity for detecting changes or differences in pain intensity, which may explain the negative findings (15,16).

Table 2
Table 2:
A Summary of Important Experimental Pain Model Studies in Humans Examining Pregnancy-Induced Antinociception

While many animal studies have demonstrated pregnancy-induced antinociceptive effects (1–3), human studies provide conflicting results (Table 2). The failure to use a nonpregnant control group and the use of relatively insensitive psychometric outcome tools may explain some of the reported negative findings (10,11). In addition, studies comparing pain sensitivity in pregnant women before and after delivery tended to report negative results when postpartum assessments were made 1–3 days after delivery, but reported positive results when assessments were made weeks after delivery (8–10,12,13). Findings from animal studies suggest that pregnancy-induced analgesia may persist for several days after delivery. Finally, characteristics of particular pain tests may account for some of the conflicting results. Saisto et al. (12) found pregnancy-induced analgesic effects with the cold pain test, when pain ratings were made using a visual analog scale. However, similar to our findings, they were unable to detect such changes when measuring pain endurance, i.e., the time until subjects withdrew their hands from the ice water tank. This suggests that pain endurance may be affected by factors other than the perceived intensity of pain. For example, the willingness to report pain is greater in pregnant women when compared with nonpregnant controls (11). The focused attention on the expectant mother during the peripartum period and her anticipated labor pain experience may encourage this behavior of increased willingness to report pain.

The HPT in our pregnant subjects did not return to nonpregnant levels on postpartum Day 1. As mentioned earlier, this may reflect that pregnancy-induced analgesia persists for several days after delivery. Alternatively, we cannot exclude the possibility that the concomitant use of analgesic medications after delivery (88% of patients) may have prevented detection of a change in HPT. However, the average time between the last analgesic dose and postpartum testing was 12 h and the analgesic drug administered to most patients (86%) was ibuprofen. In addition, the effects of ibuprofen are difficult to detect in models of acute heat or cold pain (17,18).

Although we tried to demographically match our pregnant and nonpregnant subjects, there were statistically significant age differences between the two study groups. Age-related differences in pain perception have been described when comparing 20 vs 60 yr olds (19), however, this difference is not apparent between 20 and 30 yr olds (20). A study in nulliparous pregnant women found that age (35 vs 24 yr) was not an important factor in cold pain sensitivity (20). During sub-analysis of the data, we were unable to detect an age effect on pain sensitivity. We feel that the 6-yr age gap between our pregnant and nonpregnant subjects is statistically, but not clinically, relevant. Although there was mixed parity among our pregnant subjects, we found no differences in pain sensitivity between the nulliparous and multiparous subjects.

In conclusion, the heat pain model was sensitive for detecting pregnancy-induced analgesic effects in humans. Negative findings in the cold pain test emphasize the importance of carefully considering characteristics of particular pain test paradigms, as they may assess not only different aspects of the pain signaling system, but may also be affected by factors not directly related to pain intensity. The experimental heat pain model used in this study may prove useful and practical for further elucidating the mechanisms underlying pregnancy-induced analgesia in humans, as adequate sensitivity is provided in a relatively small sample of patients.


1. Gintzler AR. Endorphin-mediated increases in pain threshold during pregnancy. Science 1980;210:193–5.
2. Jarvis S, McLean KA, Chirnside J, et al. Opioid-mediated changes in nociceptive threshold during pregnancy and parturition in the sow. Pain 1997;72:153–9.
3. Sander HW, Gintzler AR. Spinal cord mediation of the opioid analgesia of pregnancy. Brain Res 1987;408:389–93.
4. Jayaram A, Carp H. Progesterone-mediated potentiation of spinal sufentanil in rats. Anesth Analg 1993;76:745–50.
5. Dawson-Basoa MB, Gintzler AR. 17-β-estradiol and progesterone modulate an intrinsic opioid analgesic system. Brain Res 1993;601:241–5.
6. Dawson-Basoa M, Gintzler AR. Gestational and ovarian sex steroid antinociception: synergy between spinal κ and δ opioid systems. Brain Res 1998;794:61–7.
7. Cogan R, Spinnato JA. Pain and discomfort thresholds in late pregnancy. Pain 1986;27:63–8.
8. Shapira SC, Magora F, Chrubasik S, et al. Assessment of pain threshold and pain tolerance in women in labour and in the early post-partum period by pressure algometry. Eur J Anaesthesiol 1995;12:495–9.
9. Sengupta P, Nielsen M. The effect of labour and epidural analgesia on pain threshold. Anaesthesia 1984;39:982–6.
10. Dunbar AH, Price DD, Newton RA. An assessment of pain responses to thermal stimuli during stages of pregnancy. Pain 1988;35:265–9.
11. Goolkasian P, Rimer BA. Pain reactions in pregnant women. Pain 1984;20:87–95.
12. Saisto T, Kaaja R, Ylikorkala O, Halmesmaki E. Reduced pain tolerance during and after pregnancy in women suffering from fear of labor. Pain 2001;93:123–7.
13. Whipple B, Josimovich JB, Komisaruk BR. Sensory thresholds during the antepartum, intrapartum and postpartum periods. Int J Nurs Stud 1990;27:213–21.
14. Angst MS, Ramaswamy B, Riley ET, Stanski DR. Lumbar epidural morphine in humans and supraspinal analgesia to experimental heat pain. Anesthesiology 2000;92:312–24.
15. Jensen MP, Turner JA, Romano JM. What is the maximum number of levels needed in pain intensity measurement? Pain 1994;58:387–92.
16. Breivik EK, Bjornsson GA, Skovlund E. A comparison of pain rating scales by sampling from clinical trial data. Clin J Pain 2000;16:22–8.
17. Jones SF, McQuay HJ, Moore RA, Hand CW. Morphine and ibuprofen compared using the cold pressor test. Pain 1988;34:117–22.
18. Sycha T, Gustorff B, Lehr S, et al. A simple pain model for the evaluation of analgesic effects of NSAIDs in healthy subjects. Br J Clin Pharmacol 2003;56:165–72.
19. Edwards RR, Fillingim RB, Ness TJ. Age-related differences in endogenous pain modulation: a comparison of diffuse noxious inhibitory controls in healthy older and younger adults. Pain 2003;101:155–65.
20. Hapidou EG, DeCatanzaro D. Responsiveness to laboratory pain in women as a function of age and childbirth pain experience. Pain 1992;48:177–81.
© 2006 International Anesthesia Research Society