Polymorphism in the Interleukin-1 Receptor Antagonist Gene Is Associated with Serum Interleukin-1 Receptor Antagonist Concentrations and Postoperative Opioid Consumption

• Interleukin-1 (IL-1) can contribute to increased pain and enhanced immune function after surgery.

• Genetic variability in the endogenous IL-1 receptor antagonist could alter pain after surgery.

• Differences in the IL-1 receptor antagonist genotype are associated with variations in morphine consumption after surgery.

DURING the past decade, it has become evident that interleukin-1 (IL-1) signaling modulates pain sensitivity in basal and inflammatory conditions.1The IL-1 receptor antagonist (IL-1Ra) is the principal determinant of IL-1β bioactivity within the IL-1 gene cluster, regulating both constitutive and stimulated IL-1α and IL-1β release.2Endogenous IL-1Ra competitively inhibits signal mediation of IL-1β by binding to the functional IL-1R3,4(fig. 1). The polymorphic region within intron 2 of the IL-1Ra gene (IL1RN ) contains a variable number of tandem repeats of 86 base pairs, with IL1RN*1  representing four repeats of the 86 base pairs in tandem, IL1RN*2  having two repeats, and IL1RN*3  containing five repeats.5A previous study2demonstrated that, in healthy human subjects, IL1RN*1  homozygotes release more IL-1β in serum than subjects who carry at least one IL1RN*2  allele, which is consistent with a functional genotype effect in that IL1RN*2  is associated with higher IL-1Ra release. In mice, short- or long-term blockade of IL-1 signaling by IL-1Ra significantly prolonged and potentiated morphine analgesia.6In this study, the polymorphic region within intron 2 of IL1RN  was genotyped in patients undergoing nephrectomies to test the hypotheses that genetic polymorphisms of IL1RN  are related to serum IL-1Ra concentrations and opioid consumption in the postoperative period. Patients with IL1RN*2  carrier status should require fewer postoperative pain medications than patients with other genotypes.

After obtaining University of Miami, Miami, FL, Institutional Review Board approval and informed consent, patients with American Society of Anesthesiologists physical status I to III who were scheduled to undergo an elective nephrectomy were enrolled. Patients provided medical histories and demographic information, including height, weight, age, sex, ethnicity, and smoking history.

Patients scheduled to undergo a laparoscopic or open nephrectomy, using only general anesthesia, were eligible. Patients were excluded if they had taken nonsteroidal antiinflammatory drugs within 7 days of the surgical procedure; were currently menstruating females; had a known significant inflammatory process (such as lupus or rheumatoid arthritis); underwent a major surgical procedure within the past 30 days; were scheduled to receive an epidural for pain control intraoperatively or postoperatively; were not willing to remain in the study for at least 72 h; had taken steroids within the past 30 days; had received any experimental medications within the past 30 days; or possessed allergies to any of the protocol medications.

Patients were premedicated with midazolam, 1–2 mg, in the holding area. Propofol, 2–3 mg/kg; lidocaine, 1 mg/kg; and fentanyl, 3 μg/kg, were used for induction. All patients received cisatracurium, 0.2 mg/kg, for neuromuscular blockade. Fentanyl and cisatracurium were redosed as required. All patients were intubated. Maintenance of anesthesia was achieved with sevoflurane (0–4.0% in air) and oxygen (50–70%), titrated to a bispectral index (Aspect Medical, Boston, MA) of 50, with an acceptable range from 45 to 60. Ondansetron, 4 mg, or granisetron, 0.1 mg, was given approximately 30 min before extubation. Reversal of the neuromuscular blockade was achieved with neostigmine, 0.05 mg/kg, and glycopyrrolate, 0.01 mg/kg. Standard vital signs were monitored. All patients received local infiltration of bupivacaine, 0.5%, at the incision sites. Postoperatively, intravenous morphine was initially administered on request in the postanesthesia care unit and then by patient-controlled anesthesia in all patients for the first 72 h of the postoperative period. A visual analog scale for pain was assessed at 6, 12, 24, 30, 36, 48, and 72 h postoperatively. For the few patients given oral Percocet® 5/325 (oxycodone hydrochloride, 5 mg, and acetaminophen, 325 mg, in each tablet; Endo Pharmaceuticals, Chadds Ford, PA) in addition to morphine, an equivalency factor of oral oxycodone/intravenous morphine of 1:0.333 was used.7Laboratory staff members, responsible for genotyping and IL-1Ra concentration assays, were blinded to the clinical results.

Blood samples were collected in EDTA-containing tubes 1 h before surgery and were stored at −80°C. Genomic DNA was extracted according to the Blood and Body Fluid Spin Protocol (QIAGEN, Valencia, CA). The polymorphic region within intron 2 of IL1RN  was amplified by polymerase chain reaction.5Alleles IL1RN*1 , IL1RN*2 , IL1RN*3 , IL1RN*4 , and IL1RN*5  were previously described as representing four, two, five, three, and six tandem repeats of 86 base pairs, respectively.5 

Serum samples were collected before surgery and 24 h postoperatively. Serum separator tubes were used, and the whole blood samples were allowed to clot for 30 min before centrifugation for 15 min at 1,000g . Serum samples were removed, aliquoted, and stored at −80°C. The concentrations of IL-1Ra in serum samples were measured by enzyme-linked immunosorbent assay using commercially available test kits (Quantikine human IL-1Ra; R&D, Minneapolis, MN) with a minimum detection limit of 31.2 pg/ml. The procedures followed the manufacturer's instructions. All samples and standards were run in duplicate. The IL-1Ra concentration was read via  a photometer (Opsys MR microplate Photometer; DYNEX Technologies, Inc., Chantilly, VA). All values are reported as pg/ml via  computer software (Revelation Quicklin; DYNEX Technologies, Inc.).

The effect of genotype on opioid consumption was modeled using a mixed linear model,8in which the patient was treated as a random effect and genotype and postoperative hour were treated as fixed effects. In addition to genotype and hour, the genotype × hour interaction was included in the model. The correlation structure among repeated observations on patients was modeled as a first-order autoregressive process. Genotype effects were represented by the following three categories: IL1RN *1/*1  or *2/*2  plus *1/*2  or *1/*3  plus *3/*3 , chosen to distinguish carriers of alleles *2  and *3  from *1 /*1  homozygotes. The null hypothesis that genotype or genotype × hour effect equals 0 and, therefore, does not affect opioid consumption was tested using an F statistic. Before analysis, opioid consumption values were square root transformed to better approximate a normal distribution. Point estimates and 95% CIs for genotype effect sizes were obtained by back transformation of fitted values. The effect of genotype on IL-1Ra concentration before and after surgery was tested by linear regression. Genotypes were grouped as described for opioid consumption. The null hypothesis that genotype is not related to IL-1Ra concentration was tested using an F statistic. To assess potential confounding of genotype effects and clinical characteristics, analyses with and without covariates (i.e. , age, sex, smoking, ethnic origin, body mass index [BMI], and surgery type) were performed.

Clinical and demographic continuous data were summarized as mean ± SD, and categorical data were summarized by absolute frequencies or percentages. Tests of the null hypothesis of no difference in means of continuous data were performed by t  test or ANOVA; chi-square tests were used to test for a difference in distribution for categorical data. A general linear model was used for the regression analysis. All tests were two sided at a significance of P < 0.05.

After screening, 115 patients were enrolled, with 19 electively withdrawing for unspecified reasons. For the remaining 96 study subjects, one had missing BMI data, one had a missing pain score, and one had a missing smoking history.

IL1RN  polymorphism frequencies are listed in table 1. All subjects were distributed into three genotype groups: IL1RN*1  homozygotes (n = 54), IL1RN* 2  carriers (IL1RN* 1 /2 +IL1RN*2 /2 ) (n = 33), and IL1RN*3  carriers (IL1RN* 1 /3 +IL1RN*3 /3 ) (n = 9). Clinical background characteristics, including age, sex, ethnicity, smoking history, BMI, surgical type, fentanyl use during the operation, and surgical duration of the three genotype groups, are presented in table 2. No statistically significant differences among genotypes were observed.

Based on a mixed linear model, the genotype showed a statistically significant effect on opioid consumption (F^2,93= 3.95, P = 0.023). The genotype × hour interaction effect was not significant (F^2,477= 1.17, P = 0.311). Adjusting for the covariates of ethnicity, age, sex, BMI, smoking history, and surgical type did not alter these inferences; the genotype effect remained statistically significant (F^2,86= 3.35, P = 0.040), whereas the genotype × hour interaction effect did not remain statistically significant (F^2,472= 1.19, P = 0.306). The 95% CIs for the genotype effect of the IL1RN*1  homozygotes (20.4–30.5 mg for 0–12 h and 13.4–19.7 mg for 12–24 h) and the IL1RN*2  carriers (10.4–20.4 mg for 0–12 h and 7.4–13.9 mg for 12–24 h) only slightly overlapped in the 0- to 12-h period and did not overlap in the 12- to 24-h period. For IL1RN*3  carriers (14.9–38.5 mg for 0–12 h and 10.4–25.0 mg for 24 h), a wide CI because of the small sample size was observed.

After overall significant genotype effects and in the absence of evidence of interaction with time, several post hoc  comparisons of genotype differences at times of clinical relevance are summarized. Cumulative 24-h opioid consumption was compared based on ethnicity, age, sex, BMI, smoking history, surgical type, and IL1RN  genotype (table 3). Smokers had a higher 24-h opioid consumption compared with nonsmokers (P = 0.040). Twenty-four hour opioid consumption among the three IL1RN  genotype groups reached statistical significance (fig. 2, P = 0.003), with patients carrying allele IL1RN*2  consuming 43% (95% CI, 38–48%) less opioid in the first 24 h after surgery than IL1RN*1  homozygous patients (P = 0.003). After including smoking as a covariate, 24-h opioid consumption between IL1RN  genotypes remained statistically significant (P = 0.016). No evidence of interaction between smoking and IL1RN  was noted. No statistically significant differences among the three genotype groups were noted for postoperative pain scores.

Only 58 serum samples were available to measure IL-1Ra concentrations at baseline and at 24 h after surgery. A more than 2-fold increase in IL-1Ra concentrations was noted in all subjects after 24 h (628 [95% CI, 563–693] at baseline and 1,692 [95% CI, 1,368–2,015] at 24 h, P < 0.001). Figure 3presents mean IL-1Ra concentrations at baseline and 24 h postoperatively for three genotype classes. Differences in the serum IL-1Ra concentrations among the three genotype groups at baseline (P = 0.039) and at 24 h (P = 0.003) reached statistical significance. After adjusting for age, sex, ethnicity, BMI, smoking, and surgical approach, using the general linear model, these differences remained statistically significant at 24 h postoperatively (P = 0.003). The 95% CIs for genotype effects of IL1RN*1  homozygotes (792.93–1,818.07 pg/ml) and IL1RN*3  carriers (185.34–1,684.95 pg/ml) either slightly overlapped or did not overlap with the CI of IL1RN*2  carriers (1,690.11–2,975.84 pg/ml). No serum IL-1Ra concentration differences were observed between smokers and nonsmokers at either baseline or 24 h postoperatively. IL-1Ra concentrations were assessed in roughly half of the patients because of limited sample collection. We addressed the potential impact of missing IL-1Ra serum measurements on the relationship with genotype by comparing clinical characteristics of individuals with IL-1Ra serum measurements with those without such measurements in two ways. Table 4summarizes the distributions of each of six covariates (i.e. , age, BMI, origin, sex, surgery, and smoking) for individuals with and without missing IL-1Ra serum observations. None of the chi-square tests for a difference in distribution was significant. We also used logistic regression to model the probability that a serum IL-1Ra measurement is missing as a function of age, BMI, origin, sex, surgery, and smoking. Based on the full data, the model (including covariates) was not statistically significant (chi-square test = 4.8, df = 6, P = 0.569). These results suggest that patients with missing serum IL-1 measurements are not unusual for the clinical variables of age, BMI, origin, sex, surgery, and smoking, either individually or jointly.

We have described an association between IL-1Ra polymorphisms, serum IL-1Ra concentrations, and opioid consumption in the immediate postoperative period for patients undergoing a nephrectomy. IL-1Ra concentrations were significantly higher in carriers of IL1RN*2  than in IL1RN*1  homozygotes at 24 h postoperatively. Correspondingly, there was a significantly higher cumulative 24-h opioid consumption postoperatively for individuals homozygous for the IL1RN*1  allele compared with carriers of the IL1RN*2  allele. These variations were independent of age, sex, ethnicity, BMI, smoking, and surgical approach (open or laparoscopic).

In a previous study, Bessler et al.  9reported that the IL-1Ra genotype distribution was different in the first 24-h postoperative period for patients who consumed a medium dose of morphine (28–38 mg/24 h) compared with either low (lower than 28 mg/24 h) or high (greater than 38 mg/24 h) morphine consumers. However, there was no significant difference in IL-1Ra genotype between the low and high morphine consumer groups. Although results from the study of Bessler et al.  are slightly hard to interpret, our study clearly demonstrated that patients with an IL1RN*1  homozygous genotype had higher 24-h opioid consumption compared with that of IL1RN*2  carriers.

Surgery-associated tissue and peripheral nerve injury leads to a local inflammatory reaction, accompanied by increased concentrations of proinflammatory cytokines (e.g. , IL-1β and IL-6).10Of these cytokines, IL-1 acts as a major initial inducer of a proinflammatory state. Two structurally different forms (IL-1α and IL-1β) exist, both of which bind to the same receptor protein (IL-1R); however, IL-1α is approximately 3,000 times less active than IL-1β.11The IL-1 receptor antagonist (IL-1Ra) binds to this same receptor but does not initiate signal transduction, thus acting as an antagonist to both IL-1α and IL-1β.2In a recent animal study, Wolf et al.  12reported that IL-1 plays a critical role in the development and maintenance of incisional pain.

In our study, a greater than 2-fold increase in IL-1Ra concentrations in all subjects was noted 24 h after surgery. However, the IL-1Ra concentrations, 24 h after surgery, in the IL1RN*2  carrier group were much higher than those in the IL1RN*1  group. Our results demonstrating a serum IL-1Ra difference between the *1  and *2  carrier genotype groups are in agreement with several previous studies2,13,14that illustrate a similar relationship between IL1RN  polymorphisms, IL-1Ra, and IL-1β concentrations. Danis et al.  13and Wilkinson et al.  14reported an association of the IL1RN*2  allele with enhanced in vitro  release of IL-1Ra from human monocytes. This finding was also reported in healthy human carriers of the IL1RN*2  allele, compared with noncarriers15Vamvakopoulos et al.  2further confirmed these results and demonstrated that the IL-1Ra  genotype is the principal determinant of IL-1β bioactivity within the IL-1 gene cluster and that the IL1RN*2  allele was consistently associated with higher IL-1Ra concentrations and lower IL-1β release. IL-1β is involved in the mechanism of allodynia and possibly in the development of postoperative neuropathic and chronic pain.16,17Shavit et al.  18demonstrated that the administration of a neutral dose of IL-1β (i.e. , neither analgesic nor hyperalgesic) abolished morphine analgesia in mice, whereas short- or long-term blockade of IL-1 signaling by various IL-1 blockers, including IL-1Ra, significantly prolonged and potentiated morphine analgesia.6,19 

Our analysis of patients possessing IL1RN*3  is limited because of its low population frequency. However, we observed a numerically higher (but not statistically significant) mean consumption of opioids for each 12-h postoperative period, up to 36 h, when we compared subjects who were IL1RN*3  carriers with those who were IL1RN*2  carriers. Furthermore, we observed that the 95% CIs of serum IL-1Ra concentrations at 24 h postoperatively for both IL1RN*1  homozygotes and IL1RN*3  carriers either slightly overlapped or did not overlap with that of IL1RN*2  carriers. For the three alleles reported herein, it appears that the greater the number of 86 base pair variable number tandem repeats the lower the IL-1Ra concentrations, which appear to be associated with increased opioid consumption. The polymorphism of IL1RN  consists of perfect tandem repeats of a conserved 86–base pair sequence, which is reported to contain putative protein-binding sites, an α-interferon silencer A, a β-interferon silencer B, and a short-term phase response element.4,20These binding sites may influence gene expression, and an exact mechanism is under investigation.

In our study, smoking, an environmental factor, contributed to higher 24-h opioid consumption. This observation is consistent with previous studies.7,21After further analysis, it appears that this effect was significant in the first 12-h postoperative period but not in the second 12-h period. It is unclear what mechanism accounts for the increased opioid requirements noted in smokers.

In reference to pain scores, during the 48-h postoperative period, we observed a trend showing that subjects who were homozygous for IL1RN*1  had higher pain scores compared with those who were IL1RN*2  carriers. These differences reached statistical significance at 6 and 24 h postoperatively (P < 0.05, data not shown). However, when the genotype effect was assessed by adjusting covariants by the general linear model, no statistical significance remained. Pain scores in the postoperative period result from the combined results of tissue damage, anesthesia, postoperative pain, pain medications, and psychologic stress.6Because patients had free access to pain medication, it is believed that they were able to titrate their pain levels to an acceptable range.

Finally, patients undergoing an open nephrectomy were expected to have significantly higher opioid consumption than those undergoing laparoscopic surgery. Although we observed a trend toward higher opioid consumption and higher pain scores in the open surgery group compared with those in the laparoscopy group, these differences did not reach statistical significance (P = 0.181).

Our results are consistent with the hypothesis that IL1RN*1  homozygotes release more IL-1β, increasing inflammatory pain hypersensitivity, and would, therefore, be expected to require higher doses of opioid postoperatively. In contrast, those patients who are carriers of at least one IL1RN*2  allele are expected to release more IL-1Ra, decreasing inflammatory pain hypersensitivity and requiring less opioid. Our study clearly supports these concepts by demonstrating that there was a significantly lower serum IL-1Ra and higher 24-h opioid consumption in individuals homozygous for allele IL1RN*1  compared with individuals who were IL1RN*2  carriers. In addition to smoking, the IL1RN  genotype appears to play a significant role in the opioid consumption variability noted in patients in the postoperative period.

The results of this study suggest that genotype variation at the IL1RN  locus can account for a clinically significant amount of variation among patients in postoperative opioid consumption, with patients carrying allele IL1RN*2  consuming approximately 40% less opioids in the first 24 h after surgery than IL1RN*1  homozygous patients.