Cansino, Catherine MD, MPH; Edelman, Alison MD, MPH; Burke, Anne MD, MPH; Jamshidi, Roxanne MD, MPH
Vacuum aspiration for induced abortion is one of the most common surgical procedures in the United States. In the United States, local anesthesia is routinely used for first-trimester surgical abortions.1 A survey of almost 2,300 women having abortions during local anesthesia showed that 78% of women reported “moderate” or “severe” pain.2,3 Despite widespread use, pain regimens consisting of preoperative nonsteroidal antiinflammatory drug (NSAID) and paracervical block with local anesthesia may not be sufficient.4 Many clinics use adjunctive medications to improve pain control, including oral and intravenous narcotics, antiinflammatory agents, and anxiolytics. However, it is unclear whether supplemental medication results in a clinically significant decrease in patient discomfort.5
Procedure-related pain from surgical abortion arises from both the cervix and the uterus. Cervical pain results from mechanical stretching of the cervical os, whereas uterine pain is caused by curettage of the uterine wall and involution of the uterus. The innervation of the cervix is thought to be derived from two distinct regions: Frankenhauser’s nerve plexuses and the infundibulopelvic ligaments. Although the exact mechanism of action of the paracervical block is unknown, paracervical anesthesia is thought to block pain conduction via Frankenhauser’s plexus, causing an infiltrative effect that aids in decreasing pain associated with dilation.6–8 The persistence of deep pain despite paracervical block may be partially explained by the lack of effect on the infundibulopelvic ligaments’ nerve plexus.9
The exact mechanism of action of NSAIDs on the uterus is unclear. The decrease in patients’ pain perception may be related to either a decrease in uterine activity or dampening of the pain fiber response. Preexisting factors such as history of dysmenorrhea, preoperative depression and anxiety, fearfulness, expected moderate or extreme bleeding, younger patient age, and later gestational age may predispose women to be more sensitive to uterine activity and report greater pain scores.2,8,10,11 In a study of women undergoing first-trimester surgical abortions, women who received preoperative naproxen reported lower pain scores than women who did not receive preoperative naproxen.12 Another study showed that ibuprofen given preoperatively improved postoperative but not intraoperative pain.11
Ketorolac is a potent NSAID with time to peak plasma concentration of 2.9 minutes for a 30-mg intravenous dose and 33 minutes for a 60-mg intramuscular dose.13 The most clinically important adverse effects of ketorolac parallel other NSAIDs, including gastrointestinal, renal, and hematologic effects.14 A large retrospective postmarketing surveillance study showed that the overall risk of gastrointestinal or operative site bleeding related to parenteral ketorolac was only slightly higher compared with opioids.15
Ketorolac has been more recently administered as a local infiltration to improve postoperative analgesia. A systematic review found that intraarticular NSAID administration had a statistically significant positive effect compared with systemic administration.16 Although studies on the effect of wound infiltration with NSAIDs compared with systemic administration show mixed results, most studies comparing local infiltration with placebo show a favorable effect in favor of local infiltration.17
Use and effectiveness of ketorolac when administered as a paracervical block in combination with lidocaine are not widely documented. Our study investigated the difference in perceived pain from first-trimester surgical abortion comparing preoperative ibuprofen and paracervical block with lidocaine alone to placebo and paracervical block of combined ketorolac and lidocaine. We hypothesized that paracervical block with combined ketorolac and lidocaine would provide improved pain control compared with lidocaine alone.
MATERIALS AND METHODS
We conducted a double-masked, placebo-controlled, randomized clinical trial from February 2008 to May 2009 at four clinic sites affiliated with Johns Hopkins University in Baltimore, Maryland (Johns Hopkins Bayview Medical Center, Whole Woman’s Health, and Planned Parenthood of Maryland), and Oregon Health and Science University in collaboration with Planned Parenthood Columbia-Willamette in Portland, Oregon. The institutional review boards at both universities and Planned Parenthood Federation of America approved the study protocol.
We recruited women seeking elective surgical abortion of an ultrasonography-confirmed intrauterine pregnancy with an estimated gestational age not exceeding 76 days (10 6/7 weeks) from the first day of the preceding menstrual cycle. The preoperative evaluation was consistent with the clinic protocol and included pregnancy options counseling and sonographic assessment of gestational age. Women were recruited only after a voluntary decision to terminate the pregnancy was made. Exclusion criteria included 1) incomplete abortion or early pregnancy failure, 2) required or requested sedation, 3) untreated acute cervicitis or pelvic inflammatory disease, 4) contraindications to lidocaine, 5) allergic reaction or sensitivity to lorazepam or NSAIDs, 6) long-term NSAID use, 7) history of gastritis or gastric ulcer, 8) acute renal failure or chronic renal disease, 9) chronic liver disease, 10) history of bleeding diathesis, 11) long-term narcotic use, and 12) current or past history of illegal drug use (excluding marijuana).
All patients who met eligibility criteria were sequentially recruited by research coordinators at the study sites. All participants provided informed written consent for study participation. Sociodemographic and medical information was collected before the procedure. Study participants were randomly assigned into two treatment groups: 1) preoperative oral ibuprofen 600 mg and paracervical block with 18 mL of 1% lidocaine combined with 2 mL of saline (control group), or 2) preoperative oral placebo and paracervical block with 18 mL of 1% lidocaine combined with 2 mL (30 mg) of ketorolac (treatment group). Study (oral ibuprofen 600 mg or placebo) and standard (sublingual lorazepam 2 mg) medications were given approximately 90 minutes before the procedure. The surgical procedure was standardized at both institutions. After speculum insertion, 2 mL of the prepared paracervical block mixture was placed at the tenaculum site intracervically. After tenaculum placement, a paracervical block was administered using all of the remaining 18 mL of the mixture at four and eight o’clock at the cervicovaginal reflection. Cervical dilation was performed using mechanical dilators. Suction aspiration was performed with the manual vacuum aspirator using flexible cannulae. Providers attempted to use a cannula that was consistent with the participant’s gestational age of pregnancy. Sharp curettage was not performed.
Participants were allocated to study groups based on a predetermined computer-generated blocked randomization sequence created by the research pharmacies at Bayview Medical Center and Oregon Health and Science University. Because the medications used at Bayview Medical Center, Whole Woman’s Health, and Planned Parenthood of Maryland were obtained from the Johns Hopkins Bayview Research Pharmacy, treatment assignments of participants recruited from these sites were based on one randomization list. The Johns Hopkins University randomization list was created based on computer-generated 1:1 treatment allocation with random block sizes of 4. Oregon Health and Science University had its own randomization list based on computer-generated 1:1 treatment allocation with random block sizes of 10. Sequentially numbered, opaque, sealed envelopes contained the randomization information. A unique study identification number for each participant was marked on all data collection forms and the assigned analgesic mixture. The randomization code was broken only after all women completed study participation, including follow-up data collection.
All study participants, research coordinators, and providers were masked with respect to treatment assignment. The Johns Hopkins Bayview Research Pharmacy and a research coordinator at Oregon Health and Science University who was not involved with participant recruitment prepared the appropriate analgesic mixture and allocated the randomized treatment assignment for each participant at each of the two study sites, respectively. A drug compatibility study was conducted before recruitment of the first participant to ensure that both paracervical block mixtures appeared similar.
Participants were asked to record their pain level at various times during the procedure using a 100-mm visual analog scale (VAS) with anchors 0 mm=no pain and 100 mm=worst imaginable (time points: expected pain, after speculum insertion, during paracervical block administration, immediately after cervical dilation, immediately after speculum removal, and 30 minutes postoperatively). Expected procedure-related pain was assessed at the time of administration of the oral medications, approximately 90 minutes before the procedure. Total satisfaction with pain control was also recorded (VAS anchors: 0 mm=not satisfied, 100 mm=very satisfied). A fenestrated piece of paper was used to cover the scales that had been marked such that responses would be independent from one another.
A follow-up clinic visit or phone interview was conducted 2–4 weeks after the initial study visit. Participants were asked to report any side effects or complications related to their abortion procedure. If participants did not return for their scheduled follow-up clinic visit, the research coordinator attempted to contact the participants at least twice via telephone to obtain this information.
Prior studies using a 100-mm VAS suggest that a 20-mm change, either increase or decrease, is associated with clinically significant change in symptoms.17–20 We therefore defined effectiveness of treatment as a decrease (improvement) of 20 mm or greater on the 100-mm VAS compared with reported pain scores in the control group, or a relative difference of 20% between study groups. From prior experience, we estimated that the standard deviation (SD) of the VAS in abortion-related pain was 20 mm, assuming a normal distribution.2 To achieve 90% power at a significance level of 0.05, a sample size of 22 participants per group would be required. This was increased to 25 participants per group (total n=50) to allow for dropout. Statistical analyses were performed using Stata 9.0 (Stata Corporation, College Station, TX; 2005) statistical software.
The sociodemographic profiles of the study groups were compared using descriptive statistics (χ2 and Fisher exact tests for categorical data and Student t tests for continuous data). The primary exposure variable was the NSAID used for the procedure (oral ibuprofen or paracervical ketorolac). Because pharmacokinetic data on ketorolac report a short peak plasma concentration, we theorized that direct administration as a paracervical block would cause an immediate effect. Therefore, the primary outcome variable was identified as the mean difference in distance (in millimeters) on the 100-mm VAS comparing pain scores after cervical dilation between study groups. Secondary outcome variables included intrapersonal and group differences in VAS measurements of expected pain after speculum insertion, during paracervical block administration, after speculum removal, and postoperatively. Student t tests were used to compare mean differences in pain at each time point between the two groups. The Student t test was also used to compare the difference in expected and actual pain score recorded between groups. Paired t tests were performed to analyze intrapersonal differences in the VAS scores between time points. The effects of potential confounding factors, such as age, gestational age, number of prior vaginal deliveries, number of prior abortions, level of menstrual symptoms, and expected pain level during the procedure, on pain scores were analyzed.
Fifty women were enrolled and 25 participants were randomly assigned to each study group between February 2008 and April 2009 (Fig. 1). Follow-up clinic visits and phone interviews were completed by May 2009. All participants received the entire 20-mL mixture of paracervical block. Appropriate cervical dilation was achieved in all participants. One participant was withdrawn from study participation after the provider suspected a possible uterine perforation during suction aspiration. Surgical abortion was not completed, and the participant underwent an uncomplicated medical abortion at a later date. There were two protocol deviations: one participant underwent electric vacuum aspiration instead of manual vacuum aspiration, and one participant required both manual vacuum aspiration and electric vacuum aspiration to complete the procedure; both were in the control group. All statistical analyses were performed based on intent to treat.
Participants in the study cohort reported a mean age of 26.3 years and a pregnancy with mean gestational age of 55.7 days (7 6/7 weeks of gestation). Sociodemographic characteristics did not differ between study groups (Table 1). Participants in both groups reported similar numbers of prior vaginal deliveries, numbers of prior abortions, and levels of menstrual symptoms. Follow-up rates were similar in both groups (clinic, n=3; phone, n=33).
Participants in each study group reported similar pain scores (52.2 mm in the control group compared with 51.7 mm in the treatment group, P=.94) when asked how much pain they expected to have during the abortion procedure (Table 2). Participants in the treatment group reported significantly less pain after cervical dilation (59.8 compared with 74.8 mm, relative difference between study groups=20.0%, P<.05). This difference persisted after adjusting for age, gestational age, number of prior vaginal deliveries, number of prior abortions, level of menstrual symptoms, and expected pain level during the procedure. There were no differences in pain scores between groups at any other time point (Table 2).
Both groups reported significantly less pain 30 minutes after the procedure as compared with their expected level of procedure-related pain (control group: 21.8 compared with 51.7 mm, P<.01; treatment group: 17.6 compared with 48.8 mm, P<.01). Participants in the treatment group reported similar pain scores after cervical dilation compared with their expected level of pain during the procedure (59.8 compared with 50.3 mm, P=.16). However, participants in the control group reported significantly more pain after cervical dilation compared with their expected procedure-related pain level (74.8 compared with 55.0 mm, P<.01). There were no differences in the expected level of procedure-related pain and perceived pain after cervical dilation based on parity or number of prior abortions, regardless of study group. No participant required additional analgesia either intraoperatively or postoperatively.
Participants reported similar satisfaction scores with pain control during the procedure between study groups (Table 3). Overall rates of reported symptoms and complications were rare and did not differ between the groups (Table 3). Minor adverse events included one participant who was diagnosed with urinary tract infection (control group) and one participant who was diagnosed with bacterial vaginosis (treatment group). There were two serious adverse events, including one participant with an immediate postoperative hemorrhage (400 mL) most likely due to her known leiomyoma (treatment group). She was treated conservatively as an outpatient and did not require blood transfusion. Her postoperative course was uncomplicated. Another participant underwent an uncomplicated repeat vacuum aspiration procedure on postoperative day 3 (treatment group). An association between satisfaction and reported symptoms (such as fever, chills, vomiting, and heavy bleeding/clots) and complications (both minor and serious adverse events) could not be adequately assessed because overall rates of these events were low.
Pain control for first-trimester surgical abortion should be safe, effective, and affordable. Regimens that include preoperative oral NSAIDs and intraoperative paracervical block with a local anesthetic are widely used but have been shown to provide incomplete pain relief.4,7 Several nongynecologic studies report improved pain control when ketorolac is administered locally.16 Our study demonstrated greater pain reduction associated with cervical dilation among participants who received paracervical block with combined ketorolac and lidocaine compared with those who received an oral NSAID and lidocaine-only paracervical block. These findings support published data reporting improved pain control associated with local infiltration of ketorolac.
The strengths of our study include random treatment assignment and enrollment of a diverse population of women in several practice settings. Masking of treatment assignment among participants, research coordinators, and providers minimizes the influence from knowledge of the pain regimen on reported pain. We also designed our study to accommodate clinic protocols, potentially improving the study’s external validity. For example, we administered preoperative anxiolytic medication. Because many clinics routinely offer this option to women seeking abortion, our results may be more generalizable in many clinic settings across the United States. In addition, it is possible that the provision of an anxiolytic medication mitigated any effects of procedure-related anxiety on pain scores.
Limitations of our study include sources of imprecision and bias. Although our findings support similar reports on the efficacy of local administration of ketorolac, the exact mechanism of action is still unknown. Assessment of intraoperative and postoperative pain control is limited by the use of the VAS. As previously mentioned, related studies using a 100-mm VAS suggest that a 20-mm change is associated with clinically significant change in symptoms.17–20 The difference in cervical dilation pain that we demonstrated (mean difference: −15.0 mm, relative difference between study groups: 20.0%, P<.05) may not accurately represent a true clinical difference. Further, concurrent assessment of intraoperative pain is difficult during surgical abortion. Other factors that influence the experience of pain, such as preoperative depression, were not assessed for potential confounding. Our study also did not include an adequate sample size to assess any interaction between total satisfaction and rates of rare adverse events.
To limit the disruption during surgery and minimize the burden of multiple study-related pain assessments, we also decreased the frequency of pain assessments and particularly excluded assessment during the actual suction aspiration. We thought that capturing these pain levels immediately after the procedure would be accurate. In addition, rates of reported symptoms may be influenced by the interview style of the research coordinators because we collected follow-up information from most participants over phone interviews instead of in-person clinic visits. The low follow-up rate is consistent with the reported follow-up rate of 35–59% in the general population of patients seeking abortion services and related research studies.21,22
Our study suggests improved intraoperative pain control during first-trimester surgical abortion with the use of paracervical block with combined ketorolac and lidocaine. The use of ketorolac, an immediate-acting NSAID, increases clinic efficiency by eliminating wait time for the onset of action of an oral medication. Infiltrative NSAID administration may also decrease gastrointestinal side effects, which may be more pronounced in the first trimester of pregnancy. Our findings also suggest that paracervical block with combined ketorolac and lidocaine may improve pain control associated with cervical dilation performed in other gynecologic procedures, such as hysteroscopy. Such improvement in pain control may increase availability of such procedures in an ambulatory setting. However, studies specifically designed to assess the safety profile of locally administered ketorolac are needed before we can recommend routine use of paracervical block with combined ketorolac and lidocaine.
1.Nichols M, Halvorson-Boyd G, Goldstein R, Gevirtz C, Healow D. Pain management. In: Paul M, Lichtenberg ES, Borgatta L, Grimes DA, Stubblefield PG, Creinin MD, editors. Management of unintended and abnormal pregnancy: comprehensive abortion care. Oxford: Wiley-Blackwell; 2009. p. 91–4.
2.Bélanger E, Melzack R, Lauzon P. Pain of first-trimester abortion: a study of psychosocial and medical predictors. Pain 1989;39:339–50.
3.Smith GM, Stubblefield PG, Chirchirillo L, McCarthy MJ. Pain of first-trimester abortion: its quantification and relations with other variables. Am J Obstet Gynecol 1979;133:489–98.
4.Renner RM, Jensen JT, Nichols MD, Edelman A. Pain control in first trimester surgical abortion. Cochrane Database of Systematic Reviews 2009, Issue 2. Art. No.: CD006712. DOI: 10.1002/14651858.CD006712.pub2.
5.Rawling MJ, Wiebe ER. Pain control in abortion clinics. Int J Gynaecol Obstet 1998;60:293–5.
6.Glantz JC, Shomento S. Comparison of paracervical block techniques during first trimester pregnancy termination. Int J Gynaecol Obstet 2001;72:171–8.
7.Stubblefield PG. Control of pain for women undergoing abortion. Suppl Int J Gynecol Obstet 1989;3:131–40.
8.Wiebe ER. Comparison of the efficacy of different local anesthetics and techniques of local anesthesia in therapeutic abortions. Am J Obstet Gynecol 1992;167:131–4.
9.Smith R, Heltzel J. Interrelation of analgesia and uterine activity in women with primary dysmenorrhea: a preliminary report. J Reprod Med 1991;36:260–4.
10.Singh RH, Ghanem KG, Burke AE, Nichols MD, Rogers K, Blumenthal PD. Predictors and perception of pain in women undergoing first trimester surgical abortion. Contraception 2008;78:155–61.
11.Wiebe ER, Rawling M. Pain control in abortion. Int J Gynaecol Obstet 1995;50:41–6.
12.Suprapto K, Reed S. Naproxen sodium for pain relief in first-trimester abortion. Am J Obstet Gynecol 1984;150:1000–1.
13.Physicians’ desk reference. 56th ed. Montvale (NJ): Medical Economics Company; 2002. p. 3018–22.
14.Hakim-Elahi E, Tovell HM, Burnhill MS. Complications of first-trimester abortions: a report of 170,000 cases. Obstet Gynecol 1999;76:129–35.
15.Gillis JC, Brogden RN. Ketorolac: a reappraisal of its pharmacodynamic and pharmacokinetic properties and therapeutic use in pain management. Drugs 1997;53:139–88.
16.Rømsing J, Møiniche S, Ostergaard D, Dahl JB. Local infiltration with NSAIDs for postoperative analgesia: evidence for a peripheral analgesic action. Acta Anesthesiol Scand 2000;44:672–83.
17.Rowbotham MC. What is a “clinically meaningful” reduction in pain? Pain 2001;94:131–2.
18.Hendey GW, Donner NF, Fuller K. Clinically significant changes in nausea as measured on a visual analog scale. Ann Emerg Med 2005;45:77–81.
19.Jensen MP, Chen C, Brugger AM. Interpretation of visual analog scale ratings and change scores: a reanalysis of two clinical trials of postoperative pain. J Pain 2003;4:407–14.
20.Todd KH, Funk KG, Funk JP, Bonacci R. Clinical significance of reported changes in pain severity. Ann Emerg Med 1996;27:485–9.
21.Paul M, Lackie E, Mitchell C, Rogers A, Fox M. Is pathology examination useful after early surgical abortion? Obstet Gynecol 2002;99:567–71.
22.Westfall JM, Sophocles A, Burggraf H, Ellis S. Manual vacuum aspiration for first-trimester abortion. Arch Fam Med 1998;7:559–62.