FormalPara Key Summary Points

The transition of acute to chronic postoperative pain remains a significant problem and there is ongoing research for adjuvants that could possibly prevent it.

We present the long-term follow-up outcomes (chronic pain/neuropathic pain) of an RCT with primary endpoint the effect of lidocaine and dexmedetomidine on acute postoperative pain after open abdominal gynecological surgery.

Dexmedetomidine significantly reduced pain scores compared to placebo at 3 months, while lidocaine was found superior to placebo in preventing neuropathic pain (Douleur Neuropathique score, DN4 < 4) at 6 months.

Increased NRS scores at 48 led to statistically significant chronic pain NRS scores at 3, 6, and 12 months.

The advanced age of patients was also identified to lead to increased chronic pain NRS scores at 6 months after surgery.

Introduction

The term chronic postoperative pain (CPP) is used to describe the pain that develops after a surgical procedure and persists beyond the healing process i.e., at least 3 months after surgery [1]. CPP is an important clinical problem that may impair patients’ physical function and reduce the quality of their lives; therefore, its early prevention during the perioperative period is crucial [2]. Myomectomy and hysterectomy are common gynecological surgical procedures performed for various indications such as pain, menorrhagia, or dysmenorrhea [3]. According to the literature, the incidence of chronic pain after hysterectomy is reported to be 5–50% [4,5,6], while prospective studies of hysterectomy for benign conditions suggest that up to a quarter of the patients still report pain 1 year after the operation [7, 8].

Even though the transition of acute postoperative to chronic pain has been extensively investigated in the past years and the quality of studies has improved, CPP still remains an unsolved issue [9]. The identification of risk factors such as young age, female sex, psychological disorders, preoperative painful syndromes or severe acute postoperative pain, surgical factors and extent of injury, would help in the reduction of CPP incidence [10, 11].

The anesthetic technique has been proposed as an important factor that could reduce the incidence of CPP [12], and there is ongoing research for adjuvants that could be beneficial. Among other agents, dexmedetomidine, a highly selective α2 adrenoreceptor agonist, and lidocaine, a well-established local anesthetic, seem promising. The existing literature on the effects of dexmedetomidine versus lidocaine on CPP and neuropathic elements is very limited. Dexmedetomidine has shown positive effects on acute postoperative pain and opioid consumption [13, 14] but it has been very little investigated in the clinical setting for CPP [15,16,17]. Intraoperatively, it has been given as an intravenous (iv) loading dose of 0.5–1 μg/kg over 10 min, followed by an iv infusion of 0.2–0.7 μg/kg/h [18]. Lidocaine has been more extensively investigated in the chronic pain clinical setting [19, 20]. It exerts its effects through sodium channel blockade [21], inhibition of G proteins [22], and NMDA receptors [23]. In patients undergoing open abdominal surgeries, lidocaine has been given as a bolus iv dose of 1.5–2 mg/kg prior to induction/incision, followed by an infusion of 1.5–3 mg/kg/h [24,25,26].

We performed a long-term (1-year) follow-up to evaluate the effects of intraoperative iv infusion of dexmedetomidine and lidocaine on the occurrence of chronic pain after abdominal gynecological surgery. The study endpoints were the development of CPP with or without neuropathic elements at 3, 6, and 12 months postoperatively. We also identified possible risk factors for the transition of acute postoperative pain to CPP. These long-term outcomes were secondary endpoints of a randomized, double-blind study investigating the effect of dexmedetomidine and lidocaine on acute postoperative pain and analgesic consumption [27].

Methods

The key features of the original trial [27], including design, setting, eligibility, interventions, outcome measures, and sample size calculation are summarized in Table 1. Patients were contacted at 3, 6, and 12 months postoperatively by telephone for a brief interview to assess any residual pain or uncomfortable sensation. The researcher that conducted the telephone interviews was blinded regarding the intervention group of each patient. First, the patients were asked if they had any residual pain in the surgical area or around the surgical incision. If they answered “yes”, then they were asked to provide further information regarding pain site and pain intensity using a numerical pain scale (NRS) from 0 (no pain) to 10 (the worst pain imaginable). To identify possible neuropathic characteristics of the pain, the participants were asked the ten questions from the Greek version of the DN4 [28]. We adapted the questionnaire for telephone interview (Table 2) according to a previous study [29]. The collection of the chronic pain data was completed in January 2021.

Table 1 Summary of primary trial design, methods, randomization, and outcome measures
Full size table
Table 2 Adapted DN4 questionnaire for telephone interview
Full size table

Statistical Analysis

Means and standard deviations were used to describe all scale measurements such as age, NRS, or the DN4 scale. Categorical variables were described with the use of counts and percentages. Repeated measures general linear models examined the differences, in NRS and DN4 scales, observed across time and between the three study groups, adjusting for the effect of age, surgery duration, and the NRS scores during the first 48 h following the surgery and the Bonferroni correction was applied to adjust for multiple comparisons. To identify parameters that could affect the DN4 scores, a repeated measures model was applied, which included the duration of the surgery, the age of the patients and the NRS cough scores measured at 48 h after surgery. Firth’s penalized logistic regression was used to assess the differences in the odds of having neuropathic pain at 6 months after surgery, depending on the study group [30, 31]. Statistical significance was equal to 0.05 in all cases, including the analyses carried out under the Bonferroni correction. All analyses were conducted with the use of STATISTICA v12.0, except for the penalized logistic regression that was carried out on R.

Results

Data from 74 women (24 patients in the DEX group, 25 in the LIDO group and 25 in the Control group) were included in chronic pain analysis (see attached Flowchart). Seven patients were lost to follow-up as we were unable to contact them on the given telephone number in any of the three occasions (3, 6, or 12 months) after three attempts. Patient’s characteristics did not differ among the three groups (p > 0.05) (Table 3). Also, there was no statistical difference regarding the type of surgery (p = 0.962) among the groups. The surgical team included four senior surgeons with at least 15 years of experience.

Table 3 Patient and operative characteristics of the studied groups (DEX group: dexmedetomidine, LIDO group: lidocaine and CONTROL group: normal saline)
Full size table

At 3 months, the DEX group showed statistically significant lower pain scores compared to CONTROL group (p = 0.018). No other statistically differences were observed regarding pain scores at 3, 6, and 12 months among the groups (Table 4).

Table 4 NRS cough (NRS, numerical rating scale 0–10) in the three study groups (DEX group: dexmedetomidine, LIDO group: lidocaine and CONTROL group: normal saline) at 48 h and NRS scores across the three time points for each intervention group
Full size table

The repeated-measures model that was applied at each time point to identify parameters that could affect the outcomes included the NRS scores (rest and cough), the duration of the surgery, as well as the age of the patients. The analysis showed that the cough NRS scores at 48 h had a statistically significant effect on the chronic pain NRS scores (Table 5). This correlation was revealed by the repeated measures general linear model (GLM) at all three times (Table 5; Fig. 1). Higher values of NRS pain scores at all time periods are therefore expected for higher values of NRS cough pain scores at the 48-h measurement.

Table 5 Repeated measures general linear model (GLM) for the effect of group, NRS 48 cough, duration of surgery, and age at each time point on NRS scores at 3, 6, and 12 months
Full size table
Fig. 1
figure 1

NRS scores compared to 48-h NRS cough scores for all groups at 3, 6, and 12 months

Full size image

A statistically significant relation (p = 0.020) was also found between advanced age of patients and higher values of chronic pain NRS scores at 6 months after surgery. This relationship was borderline nonsignificant at 3 and 12 months with p values equal to 0.079 and 0.092, respectively.

Analysis for the DN4 Scores

The DN4 scores across the three time points for each of the treatment groups are shown in Table 6. Even though the control group appeared to score higher at all periods compared to the other two groups (Fig. 2), these differences were not statistically significant (p > 0.05) (Table 6).

Table 6 DN4 (Douleur Neuropathique score 0–10) in the three study groups (DEX group: dexmedetomidine, LIDO group: lidocaine and CONTROL group: normal saline) at 3, 6, and 12 months
Full size table
Fig. 2
figure 2

DN4 scores at 3, 6, and 12 months for all groups

Full size image

In our study, although many patients reported neuropathic elements, most of them did not fulfil the literature cut-off score of 4 for their pain to be diagnosed as neuropathic [32]. Therefore, at all time points, the percentage of patients suffering from neuropathic pain was relatively small, as shown in Table 7. Fisher's exact test showed a statistically significant difference at 6 months (p = 0.031).

Table 7 Neuropathic pain at 3, 6, and 12 months
Full size table

The comparative bar chart in Fig. 3 shows that more patients are expected to suffer with neuropathic pain in the control group. The penalized logistic regression model also showed a statistically significant difference in the neuropathic pain depending on the treatment group (p = 0.048). This difference was observed between the LIDO (0%) and CONTROL (20%) group (p = 0.02), but not among the other groups [DEX (4.2%) vs. CONTROL (p = 0.106) and DEX vs. LIDO (p = 0.442)].

Fig. 3
figure 3

Neuropathic pain at 6 months for all groups

Full size image

The analysis showed that the 48-h cough NRS scores had a statistically significant effect on the DN4 scores at all three times, thus 3, 6, and 12 months, with p values equal to 0.01, < 0.01, and < 0.01, respectively. Higher values of DN4 at all time periods are therefore expected for higher values of NRS cough scores at 48 h, which are depicted in Table 8.

Table 8 Repeated measures general linear model (GLM) for the effect of group, NRS48 cough, duration of surgery, and age at each time point on DN4 scores at 3, 6, and 12 months
Full size table

Adverse Effects of the Investigated Agents

Intraoperative hypotension and bradycardia were more often observed in the study groups (DEX and LIDO) compared to the CONTROL group, but the differences were not statistically significant (p = 0.357 and p = 0.566, respectively). A patient of the LIDO group developed ventricular ectopic beats, which resolved after discontinuation of the intervention.

No other adverse effects were reported during the postoperative period.

Discussion

Our results showed that perioperative iv infusion of dexmedetomidine comparing to placebo had a beneficial effect on the prevention of CPP at 3 months, while lidocaine infusion prevented the development of neuropathic pain at 6 months after gynecological surgery. To our knowledge, this is the first study comparing intraoperative dexmedetomidine and lidocaine with placebo with regards to CPP development. Our long-term follow-up showed superiority of the studied drugs over placebo and highlighted predisposing factors for development of chronic pain with or without neuropathic elements.

The first finding, that dexmedetomidine at 3 months prevented the development of CPP compared to placebo, is also supported by one study, which investigated perioperative infusion of dexmedetomidine in patients undergoing breast cancer surgery [15]. However, in that study, patients had received much greater doses of the investigated drug (1 μg/kg iv bolus, followed by a continuous infusion of 0.5 μg/kg/h iv till the completion of surgery, and then the dose was tapered to 0.2 μg/kg/h for up to 24 h). The beneficial effect of dexmedetomidine has also been shown in an experimental study, where it attenuated persistent postsurgical pain by upregulating K+–Cl− cotransporter-2 in the spinal dorsal horn in rats [33].

The second finding of our study was that lidocaine proved to be superior to placebo regarding DN4 scores, making it a useful drug for the prevention of postsurgical neuropathic pain. For the assessment of neuropathic elements of pain, we used the DN4 questionnaire, a validated and widely used screening tool for the identification and classification of neuropathic pain [32]. It has been shown to achieve an 83% sensitivity and 90% specificity when compared to clinical diagnosis [34]. A former study had successfully performed the DN4 neuropathic pain questionnaire over the telephone to identify possible neuropathic elements [29]. In this study [29], for the first seven symptom items, the respondents answered “yes” or “no” to whether their pain could be described as burning, painful cold, electric shocks, tingling, pins and needles, numbness, and itching. The patients were also asked if their pain area was sensitive to touch, sensitive to pin prick, and sensitive to light brushing, which resembles the clinical examination. A score of 1 was given for each “yes” answer and 0 for each “no” answer. A score of < 4 suggested that the pain was unlikely to be neuropathic. This important finding might be supported by the fact that lidocaine is already a very useful drug, used as infusion in chronic neuropathic pain states, as extensively analyzed in a recent review [35] and in a metanalysis [36].

Moreover, our results showed that the pain scores (NRS cough) at 48 h had a statistically significant effect on the chronic pain scores at all three times, making acute postoperative pain an indication of transition to chronic pain. Fassoulaki et al. have demonstrated that pain during the immediate postoperative period can predict chronic pain after breast surgery for cancer [37]. However, in this study patients who developed chronic pain had experienced higher pain intensity at rest during the first nine postoperative hours rather than pain at cough at 48 h as in our study. Other studies also support our finding [38, 39].

Pain intensity at 48 h postoperatively was found to be a significant risk factor for the development of neuropathic pain, as higher DN4 scores at 3, 6, and 12 months were reported from patients with higher NRS cough scores at 48 h. This finding has been also identified by other studies that have highlighted the role of acute pain as a risk factor for the development of long-term neuropathic pain [40, 41].

Age may also play an important role in the development of CPP. We found that at 6 months after surgery, older patients suffered more severe pain. This is not in agreement with previous findings, as Martinez et al. have suggested the opposite, i.e., that younger age is a predictive factor for chronic pain [42]. Also, according to Schug and Bruce, younger age seems to be in most studies a relatively consistent demographic risk factor for CPP in adults [11].

Limitations and Strengths

This long-term study on CPP derives from a RCT that was powered to reveal differences in acute postoperative pain; thus, the sample size may be relatively small for chronic pain outcomes. Another possible limitation is that we did not study different doses of the administered drugs. Also, several possible risk factors of CPP, such as gender, preexisting pain, and psychological factors were not investigated in the present study, as it was designed to minimize these confounding factors; the studied population were all females (ASA I and II with no major comorbidities or prior disability), while patients with preexisting pain or analgesic consumption, or those treated for depression and anxiety were excluded according to protocol (Table 1). Additionally, the duration of surgery did not differ among the groups (Table 3) and the patients with surgical complications were excluded from follow-up and analysis. We did not assess other possible predisposing factors for CPP, such as education level and lifestyle.

A strength of this study is that the patients were followed up for quite a long time (12 months) and also that we managed to have a high percentage of patients (74/81 or 91.35%) that completed the follow-up period. This is important since previous similar studies report significant dropout rates, ranging from approximately 34% [43] to 43.6% after 12 months [44].

Conclusions

The present study demonstrated that perioperative infusion of dexmedetomidine or lidocaine may exert a beneficial effect on the development and characteristics of CPP after abdominal gynecological surgery. Dexmedetomidine proved to be superior to placebo regarding the duration and severity of CPP, while lidocaine exhibited a protective effect against neuropathic elements of CPP over placebo. We consider that the findings of this RCT could provide the foundation for future studies.