Gestational Diabetes and Intraoperative Tubal Sterilization are Risk Factors for Pain After Cesarean Delivery: A Prospective Observational Study.

Background: Postcesarean delivery pain leads to several adverse maternal outcomes. The primary objective of this study was to determine the incidence of moderate-to-severe pain after the use of spinal morphine for cesarean delivery. The secondary aim was to identify the factors inuencing the moderate-to-severe pain. Methods: This prospective observational study was conducted at a single university hospital. The inclusion criteria were a patient age of ≥ 18 years, and undergoing elective cesarean delivery under spinal anesthesia with intrathecal morphine (200 mcg). Moderate-to-severe pain was dened as a numerical-rating-scale score of more than 3 within 24 hours postoperatively. Patients’ demographic characteristics, preoperative obstetric data, and intraoperative and postoperative data were collected and analyzed. Results: In all, 660 patients were enrolled. As 16 were subsequently removed because they met the study withdrawal criteria, data relating to 644 patients were analyzed. The incidence of moderate-to-severe pain during the rst postoperative day was 451/644 patients (70.03%; 95% condence interval [CI], 66.38%– 73.44%). The median pain score (interquartile range [IQR]) was 5 (3–6), with 176/644 (27.33%) of patients requiring a rescue analgesic. A multivariate analysis revealed that two factors were associated with moderate-to-severe pain on the rst postoperative day: gestational diabetes (adjusted OR [AOR], 1.849; 95% CI, 1.068–3.203; p-value = 0.028), and intraoperative tubal sterilization (AOR, 1.533; 95% CI, 1.060–2.218; p = 0.023). A signicantly higher number of patients experienced moderate-to-severe pain on postoperative Day 1 (451/644 [70.03%]) than Day 2 (348/644 [54.19%]; p < 0.001). The median pain score (IQR) on postoperative Day 2 was 4 (3–5), which was less than that for Day

Conclusions: A high incidence of moderate-to-severe postoperative pain was found after a single dose of spinal morphine for cesarean delivery. Gestational diabetes and intraoperative tubal sterilization were the independent factors correlated with the pain.
Trial registration: Clinicaltrial.gov registration number NCT03205813 Background Cesarean delivery is a common procedure, and its rate has increased continually to unprecedented levels worldwide [1]. Postcesarean delivery pain leads to an unpleasant maternal experience, delays functional recovery, interferes with breastfeeding, and increases the rate of postpartum depression [2][3][4]. Moreover, severe postcesarean delivery pain contributes to persistent postsurgical pain [3,5,6]. The provision of optimum postcesarean delivery pain control is therefore essential. Even though intrathecal morphine produces bothersome side effects (for example, pruritus, nausea, and vomiting), it is generally used as a postcesarean delivery analgesic due to its good e cacy and long analgesic duration [7,8]. A systematic review reported that prolonged postcesarean delivery pain relief was obtained with a variety of intrathecal morphine doses, with a median of 27 hours (range, 11-29 hours) before the rst analgesic requirement [9]. However, some patients suffer from a high intensity of pain during the rst postoperative day. Borges et al. found that a substantial proportion of patients (78.4%) experienced moderate-to-severe pain (numerical pain scale ≥ 5) after spinal anesthesia with intrathecal morphine for cesarean delivery, and the associated factor was preoperative anxiety [10]. Another study revealed that patients' responses to suprathreshold thermal stimuli as well as their degree of preoperative anxiety were predictors for postcesarean section pain [11].
Moreover, surgical factors can affect postoperative pain levels. Eken et al. revealed that closure of both the parietal and visceral peritonea increased postoperative pain after cesarean delivery [12]. The qualitative systematic review found that the degree of postoperative pain and analgesic consumption were affected by the presence of preoperative pain; patient anxiety; patient age; and surgery type (such as open abdominal, thoracic, or orthopedic) [13]. We therefore postulated that surgical factors-such as skin incision type, or a cesarean section with a tubal sterilization or appendectomy-may in uence postoperative pain. Additionally, other factors that may affect postoperative pain were examined: patient education level, the presence of diabetes or hypertension, and anesthetic factors (such as intraoperative inadequate block level and pain).
This study set out to determine the incidence of moderate-to-severe postoperative pain after the administration of spinal morphine for cesarean delivery, as well as the predictive factors related to that pain.

Methods
This prospective observational study was conducted at a single university hospital and was approved by the Siriraj Institutional Review Board, Faculty of Medicine Siriraj Hospital, Bangkok, Thailand (protocol number, 164/2560 [EC2]; approval number, Si 248/2017). The study was registered at www.clinicaltrials.gov (NCT03205813, date of registration 02/07/2017). The data collection was conducted from July 2017 to August 2019. The inclusion criteria were a patient age of ≥ 18 years; an American Society of Anesthesiologists physical status I or II; term pregnancy; elective cesarean delivery under spinal anesthesia with intrathecal morphine; and a proper understanding of the numerical rating scale of pain. Patients were not enrolled if they did not understand Thai; they had a complicated pregnancy for which general anesthesia may be required during the cesarean delivery; or the use of spinal anesthesia with intrathecal morphine was contraindicated. Patients were also excluded if they received either general anesthesia or a peripheral nerve blockade after the spinal anesthesia.
Patients were evaluated the day before surgery by an anesthesiologist. The following preoperative data were recorded: demographic characteristics; education level; history of previous normal labor or cesarean delivery; and comorbidities (including gestational diabetes and hypertension). Written informed consent was also obtained from the participants. On the morning of the surgery, they were premedicated with oral ranitidine (150 mg) and metoclopramide (10 mg), as per standard practice. In the operating theatre, the patients received intravenous uid and were monitored in the standard manner. The spinal anesthesia was administered in the lateral position at L2-3 or L3-4 with a 26-or 27-gauge Quincke spinal needle (Becton Dickinson, Franklin Lakes, NJ, USA) or a 25-gauge Whitacre spinal needle (Becton Dickinson), depending on the decision of the anesthesiologist. The intrathecal medication given to all patients was 0.5% heavy bupivacaine (10-12 mg) with morphine (200 mcg). Details were collected of the intraoperative data (anesthesia level testing by cold sensation, and the number of uterotonic agents used) and intraoperative pain. Intraoperative pain was de ned as the usage of an intravenous opioid analgesic medication or ketamine while the patients were undergoing surgery. Records were also made of the surgical data: skin-incision type (midline or Pfannenstiel incision); the surgeon (trainees, such as residents; or graduated obstetricians, like obstetric fellows or consultants); peritoneal closure or nonclosure; and concurrent operations (for instance, tubal sterilization or appendectomy). All patients were transferred to the recovery room and observed for 1-2 hours. Subsequently, in the maternity ward, research assistants or research nurses visited the patients at 24 h and 48 h, postoperatively, to assess their pain level, nausea or vomiting symptoms, need for pain medication, and satisfaction. Patients rated their pain using an 11-point numerical rating scale, with 0 signifying "no pain", and 10 representing "the worst possible pain". Those patients scoring 4-6 were considered to have moderate pain, whereas those with scores of 7 to 10 were deemed to be experiencing severe pain [14]. If required, the rescue analgesic employed was either intravenous pethidine (20 mg) or tramadol (25 mg). Other analgesic medicationssuch as paracetamol and nonsteroidal anti-in ammatory drugs (NSAIDs)-were administered at the discretion of each obstetrician. Nausea and vomiting were graded with the following 4-point system: 0, no symptoms; 1, nausea; 2, vomiting 1-2 times; and 3, vomiting > 2 times. The patients were requested to rate their satisfaction with their perioperative pain management 24 hours postoperatively, using a scale of 0 to 100 (with 100 denoting "totally satis ed").

Statistical analysis
The sample size calculation was performed using the results of an earlier survey at our institute. It had revealed an incidence of 55% for moderate-to-severe pain during the rst 24 hours following cesarean delivery (unpublished data). To obtain a con dence level (1-α) of 95%, and with an allowable error of 0.05, 381 patients were determined to be required. With regard to the secondary objective, we anticipated that 20 patient and surgical factors might be associated with moderate-to-severe pain. The patient factors comprised age, BMI, educational level, gestational diabetes, gestational hypertension, primigravida, multiple pregnancy, and history of vaginal or cesarean delivery. The surgical factors were surgical incision type; type of surgeon performing the operation (trainee or graduated obstetrician); closure of both the parietal and visceral peritonea; a concurrent fallopian tubal sterilization or appendectomy; use of more than one type of uterotonic agent; intraoperative pain requiring a rescue analgesic; anesthetic level lower than the 4th thoracic dermatomal level (T4); neonatal birthweight (with twin pregnancies, the summation of the two neonatal birthweights was used for analysis purposes); and postoperative use of paracetamol or NSAIDs.
The sample size estimation was based on the recommendation of statisticians for multiple logistic regression analyses, namely, that the number of patients recruited be ten times the number of risk factors. Given that, approximately 400 patients with moderate-to-severe pain were required. A sample size of 600 was therefore deemed adequate to achieve the two study objectives. A total of 660 patients were enrolled to compensate for any subject loss during the study.
All analyses were performed using PASW Statistics for Windows (version 18.0; SPSS Inc., Chicago, Ill., USA). Categorical data are presented as number and proportion, expressed as a percentage. The decreasing of number of patients experiencing pain from postoperative Day 1 to Day 2 was identi ed using McNemar's Test. The Chi-squared test was used to analyze categorical data. Continuous data are reported as mean ± standard deviation, or as median with interquartile range. For the univariate analysis, the Chi-squared test or Fisher's exact test was used to examine the risk factors. Risk factors with a pvalue of < 0.20 from the univariate analysis were entered into the multivariate analysis of moderate-tosevere postoperative pain using forward stepwise logistic regression. Crude and adjusted odds ratio (OR) and 95% con dence interval (CI) are used to present the data. P-values < 0.05 were considered statistically signi cant. An independent t-test was used to compare the means of the satisfaction scores.

Results
A total of 660 patients were enrolled. Twelve were subsequently withdrawn because they received general anesthesia (four underwent a hysterectomy after the cesarean delivery, while the remaining eight failed to achieve an adequate level of spinal block). A further 4 patients were excluded as they received a quadratus lumborum block upon completion of the operation. Therefore, 644 patients entered the complete analysis (Fig. 1); their demographic and obstetric data are listed in Table 1. One patient had a history of chronic pain and received pain relieving medication preoperatively. The number of patients reporting a pain score of ≥ 4 on postoperative Day 1 and Day 2 were 451/644 and 349/644 patients, representing 70.03% (95% CI, 66.38-73.44%) and 54.19% (95% CI, 50.33-58.0%), respectively. Although there was a high proportion of patients with a pain score of ≥ 4 on the rst postoperative day, the number requiring rescue pain medication was only 176/644 (27.33%). The percentage of patients experiencing moderate-to-severe pain decreased signi cantly from Day 1 to Day 2 (p < 0.001). There was a corresponding fall in the average pain score, with a statistically signi cant decline from 4.9 ± 2.3 (mean ± standard deviation) for the rst postoperative day to 3.8 ± 1.8 for the second day (p < 0.001; Table 2). The factors that had been assumed to be associated with postcesarean delivery pain are described and analyzed in Table 3. The authors used the cut point of a neonatal birthweight of > 4,000 gm as the de nition of fetal macrosomia [15]. In the case of twin pregnancies, we used the summation of the two fetal birthweights for analysis purposes. The univariate analysis revealed that two factors were related: gestational diabetes and intraoperative tubal sterilization. The multiple logistic regression was then performed using the associated factors from the univariate analysis with a p-value of < 0.2, namely, maternal age ≥ 35, gestational diabetes, patients with a history of normal labor, patients with a history of previous cesarean delivery, patients with an intraoperative anesthetic level below the 4th thoracic dermatome, residents as the surgeon performing operation, and intraoperative tubal ligation. The multivariate analysis revealed two factors were independently related to postcesarean delivery pain: gestational diabetes (adjusted OR, 1.849; 95% CI, 1.068-3.203; p = 0.028) and intraoperative tubal ligation (adjusted OR, 1.533; 95% CI, 1.060-2.218; p = 0.023). The patients' satisfaction toward their pain management is presented in Table 4. The results indicated that there was a high level of satisfaction with the postoperative pain management, with the mean satisfaction score being 93. On the rst postoperative day, 584 patients (90.7%) had no symptoms of either nausea or vomiting. Of the other 60 patients, 30 (4.7%) experienced nausea but not vomiting; 16 (2.5%) vomited but did not need treatment; and the remaining 14 (2.2%) vomited > 2 times and required treatment.

Discussion
A considerable proportion of the patients-approximately 70%-experienced moderate-to-severe pain on the rst postoperative day following cesarean delivery involving a single shot of spinal anesthesia with morphine (200 mcg). Likewise, Borges et al. revealed a large percentage of patients experienced moderate-to-severe pain after spinal anesthesia with intrathecal morphine for cesarean delivery (78.4%) [10]. Our study de ned moderate-to-severe pain as a numerical rating scale pain score of 3 or more. This was different from that used by Borges and colleagues: their cut point for moderate pain was more than 4 [10]. A dissimilarity in the intrathecal morphine quantity explains the higher incidence of pain in the work by Borges et al. Our study used a higher dose of 200 mcg for every patient, which was the usual amount at our institute, Borges and colleagues used a lower average spinal morphine dose of 86.5 ± 12.2 mcg [10]. A meta-analysis found that the mean time to the rst analgesic requirement with a high dose of spinal morphine was 4.49 hours longer than that for a low dose. The range for the high dose (> 100 mcg) was 13.8 to 39.5 hours, compared with 9.7-26.6 hours for the low dose (50-100 mcg). [8]. However, the meta-analysis also found that the means of the pain score and the supplemental morphine consumption of the high-and low-dose groups were not different [8]. It is interesting to note that our study also revealed that-despite the high incidence of moderate-to-severe postoperative pain on Day 1-only a relatively small percentage of patients (27.3%) required the rescue analgesic.
A systematic review identi ed that a younger patient age was a factor associated with postoperative pain [13]. However, that review also found that there were some con icting reports about the correlation between patient age and postoperative pain. Moreover, some of the studies examined by the review had used sample sizes which were too small to detect any correlation. In the case of our study, patient age demonstrated no signi cant relationship with pain. However, this nding is explained by the narrow age range of our study cohort (mean, 33.6 ± 5 years). In addition, whereas numerous studies have shown that preoperative pain was a factor related to postoperative pain [2, 10, 13], our study revealed an extremely low incidence. As there was only one patient who had a history of chronic pain before the cesarean section, the preoperative pain factor was not included in our analysis.
Gestational diabetes was one of the two independent factors that our study found to be associated with pain after cesarean delivery. Karci et al. found that diabetic patients undergoing a total abdominal hysterectomy had higher pain scores and required larger amounts of additional morphine postoperatively than nondiabetic patients [16]. In addition, Kim et al. reported a signi cant correlation between preoperative HbA1c (plasma glycosylated hemoglobin) levels and opioid consumption in the rst 48 hours following elective open nephrectomy [17]. Moreover, another recent study found that gestational diabetic patients required a higher dosage and frequency of an opioid analgesic six hours after a cesarean section than patients without diabetes [18]. Those gestational diabetic patients also reported a higher mean visual analog pain score for the 24-hour period after their cesarean delivery. The reason for the heightened perception of pain and consequential elevated usage of opioids by those gestational diabetic patients is assumed to be changes in the pharmacokinetics and pharmacodynamics of opioids when administered to diabetics. A study on animals demonstrated that a hyperglycemic state decreased the analgesic effect of morphine [19]. Several studies have reported that diabetic patients have higher pain scores and tend to have a higher opioid consumption. The studies suggested that these patients have different pain perception due to the decrease in the analgesic effect of opioids resulting from a direct antagonistic effect on opioid receptors [16][17][18]. Consistent with that, the present work identi ed that almost 80% (71/89) of the gestational diabetic patients experienced moderate-to-severe pain after their operation. Gestational diabetic mothers have an elevated risk of developing fetal macrosomia [15]. We assumed that a neonatal birthweight equal or more than 4,000 grams might affect pain levels after cesarean delivery as a result of the surrounding tissue trauma experienced during the operation. Nevertheless, our study did not discover any differences in the maternal pain scores for this factor (OR 1.300; 95% CI 0.555-2.298; p = 0.529).
There are two components of postcesarean delivery pain. The rst relates to the somatic pain derived from the skin incision together with the visceral pain arising from the peritoneum; the second component involves the cramping that arises from uterine involution after delivery [2,12,20,21]. We assumed the surgical incision type or the skill level of the surgeon performing the cesarean delivery (for instance, a trainee versus a graduated obstetrician) would in uence the degree of skin and soft tissue trauma and hence the postoperative pain level. However, no difference was found for the parameters of skin incision or surgeon type.
Other surgical factors related to postcesarean delivery were studied. Eken et al. revealed that patients receiving closure of both the parietal and visceral peritonea reported higher visual analog pain scores than those only receiving a parietal peritoneal closure [12]. Tissue trauma and ischemia from suturing may explain the higher postoperative pain [12]. However, our study showed a small number of patients (37/644; 5.7%) only received a parietal peritoneal closure, and no difference was found in the pain scores for this factor. On the other hand, we identi ed that tubal sterilization after cesarean delivery was associated with postoperative pain. Three quarters of the patients receiving tubal sterilization (167/221; 75.6%) reported having moderate-to-severe postoperative pain. Pain from tubal sterilization is believed to be derived from the prostaglandins released from the traumatized fallopian tubes [22]. As well, pain arises from ischemia or necrosis of the ligation site of the tubes [23]. These reasons explain why the tubal sterilization had a high likelihood of generating postoperative pain.
We assumed that anesthetic factors would affect postcesarean delivery pain. Even though the level of sensory blockade below T4 tended to be related with pain after the operation, there was no statistical signi cance (OR 1.618; 95% CI 0.984-2.661; p = 0.056). We could not nd a correlation between patients experiencing intraoperative pain and postoperative pain as only a small number of patients (16) required intravenous opioid medications in the operating theatre.
Although many studies have concluded that uterine involution causes pain after delivery [20,21], we did not nd that the usage of more than one uterotonic agent promoting additional uterine contraction affected pain. A systematic review showed that the NSAIDs used were better than a placebo and paracetamol for the relief of pain arising from uterine cramping [20]. Furthermore, NSAIDs have been reported to have an opioid-sparing effect after cesarean delivery [21]. Nevertheless, our study failed to demonstrate the effects of NSAIDs on postcesarean section pain.
The early identi cation of the predictors in patients at risk of experiencing postoperative pain will allow more effective interventions and better management. Carvalho et al. suggested the personalized management of pain [24]. That is, patients would be able to make individual choices for their pain protocol by selecting one of three doses of spinal morphine: low (50 mcg), medium (100 mcg), and high (200 mcg) [24]. However, higher doses of intrathecal morphine, as well as of several other analgesic medications, have more possible side effects, like nausea and vomiting, itchiness, and the risk of sedation. Therefore, the plausible complications should be advised to patients.
The authors recommend two areas of improvement in the use of analgesia for postcesarean deliveries at our institute. Firstly, a pain protocol should be developed for every patient. It should include the use of multimodal analgesics, such as paracetamol and NSAIDs. Secondly, the analgesics should be supplemented with adjuvant medications for patients at increased risk for postoperative pain (for example, due to high preoperative anxiety, a preexisting chronic pain condition, gestational diabetes, or receiving tubal sterilization intraoperatively).
The main limitation of this study was that data were not collected on the blood sugar levels that may affect pain after surgery. Therefore, we were not able to nd a direct correlation between hyperglycemia and the pain score ratings of the patients. Additionally, the side effects of the use of the high-dose spinal opioid (such as sedation and pruritus) were not recorded. It is acknowledged that the usage of 200 mcg of spinal morphine causes a higher incidence of pruritus than a lower dose (< 100 mcg). Future research should be carried out on postoperative cesarean section pain medication requirements for diabetic patients.
In conclusion, at 70%, the incidence of moderate-to-severe pain after cesarean delivery was high. The independent factors were gestational diabetes and intraoperative tubal sterilization. We therefore suggest that it is essential to identify the risk of developing a higher pain intensity after surgery and to provide adequate postoperative pain control.