Pain control in first-trimester surgical abortion: a systematic review of randomized controlled trials
Regina M. Renner, Jeffrey T. Jensen, Mark D. Nichols and Alison B. Edelman  2010/4/28 14:43:00

Contraception,  2010,  Volume 81,  Issue 5
   
   
Abstract
Background

First-trimester abortions especially cervical dilation and suction aspiration are associated with pain despite various methods of pain control.
Study Design

Following the guidelines for a Cochrane review, we systematically searched for and reviewed randomized controlled trials comparing methods of pain control in first-trimester surgical abortion at less than 14 weeks gestational age using electric or manual suction aspiration. Outcomes included intra- and postoperative pain, side effects, recovery measures and satisfaction.
Results

We included 40 trials with 5131 participants. Because of heterogeneity, we divided studies into seven groups:

Local anesthesia: Data were insufficient to show a clear benefit of a paracervical block (PCB) compared to no PCB. Reported mean pain scores (10-point scale) during dilation and aspiration were improved with carbonated lidocaine [weighted mean difference (WMD), −0.80; 95% confidence interval (CI), −0.89 to −0.71; WMD, −0.96; 95% CI, −1.67 to −0.25], deep injection (WMD, −1.64; 95% CI, −3.21 to −0.08; WMD, 1.00; 95% CI, 1.09 to 0.91), and with adding a 4% intrauterine lidocaine infusion (WMD, −2.0; 95% CI, −3.29 to −0.71; WMD, −2.8; 95% CI, −3.95 to −1.65).

PCB with premedication: Ibuprofen and naproxen resulted in small reduction of intra- and postoperative pain.

Conscious sedation: The addition of conscious intravenous sedation using diazepam and fentanyl to PCB decreased procedural pain.

General anesthesia: Conscious sedation increased intraoperative but decreased postoperative pain compared to general anesthesia (GA) [Peto odds ratio (Peto OR) 14.77 (95%, CI 4.91–44.38) and Peto OR 7.47 (95% CI, 2.2–25.36) for dilation and aspiration, respectively, and WMD −1.00 (95% CI, −1.77 to −0.23) postoperatively). Inhalation anesthetics are associated with increased blood loss (p<0.001).

GA with premedication: The cyclooxygenase (COX)-2 inhibitor etoricoxib; the nonselective COX inhibitors lornoxicam, diclofenac and ketorolac IM; and the opioid nalbuphine improved postoperative pain.

Nonpharmacological intervention: Listening to music decreased procedural pain.

No major complication was observed.
Conclusions

Conscious sedation, GA and some nonpharmacological interventions decreased procedural and postoperative pain, while being safe and satisfactory to patients. Data on the widely used PCB are inadequate to support its use, and it needs to be further studied to determine any benefit.

Keywords: First trimester abortion; Surgical abortion; Pain control; Anesthesia; Randomized controlled trials; Systematic review; Meta-analysis
Article Outline

1. Introduction
2. Materials and methods

2.1. Criteria for eligible trials
2.2. Search methods
2.3. Data extraction
2.4. Quality assessment
2.5. Analysis

3. Results

3.1. Description of included studies

3.1.1. Group 1: local anesthetics, local anesthesiatechnique, premedication and paracervical block (Table 1)
3.1.2. Group 2: PCB with premedication (Table 1)
3.1.3. Group 3: analgesia alone (Table 2)
3.1.4. Group 4: conscious sedation (Table 2)
3.1.5. Group 5: GA (Table 3)
3.1.6. Group 6: GA with premedication (Table 4)
3.1.7. Group 7: nonpharmacological interventions

3.2. Description of excluded studies
3.3. Methodological quality of included studies
3.4. Results of studies

3.4.1. Group 1: local anesthetics (Comparison 1) (Table 1)
3.4.2. Local anesthesia technique (Comparison 2) (Table 1)
3.4.3. Group 2: PCB with premedication (Comparison 3) (Table 1)
3.4.4. Additional outcomes and sub-analysis resultsreported in studies for Comparisons 1, 2 and 3
3.4.5. Group 3: analgesia alone (Comparison 4) (Table 2)
3.4.6. Group 4: conscious sedation (Comparison 5) (Table 2)
3.4.7. Group 5: GA (Comparison 6) (Table 3)

3.4.7.1. Inhalation anesthetics

3.4.7.1.1. Side effects of inhalation anesthetics

3.4.7.2. Sedatives, hypnotics and opiates
3.4.7.3. Side effects of sedatives, hypnotics and opiates
3.4.7.4. PCB added to GA
3.4.7.5. Conscious sedation with GA

3.4.8. Group 6: GA with premedication (Comparison 7) (Table 4)

3.4.8.1. COX 3 inhibitor

3.4.8.1.1. Side effects

3.4.8.2. Opioids

3.4.8.2.1. Side effects

3.4.9. Group 6: nonpharmacological intervention(Comparison 9)

4. Discussion

4.1. Summary of main results
4.2. Overall completeness evidence
4.3. Quality of evidence

5. Conclusion
Acknowledgements
References

1. Introduction
Elective abortions are among the most common outpatient surgical procedures performed on women with an estimated 46 million performed yearly worldwide [1]. Nearly 90% are performed in the first trimester before 13 weeks gestation [2]. A major complication occurs in less than 1 in 100 women and mortality is around 0.7 in 100,000 [3], [4] and [5]. Complications from all forms of anesthesia accounted for 16% of all legally induced abortion mortality in the United States in the years 1988 to 1997, a marked decrease from the 29% attributable to general anesthesia (GA) alone reported for 1983 to 1987 [4] and [6].

 

Anesthesia is important for women undergoing an abortion since most will experience pain with the procedure. Key factors that influence the choice of anesthesia or analgesia include effectiveness, safety, side effects and costs. Other important factors include patient preference, practitioner choice or bias, facility resources and medical indications [7].

Pain perception is a complex phenomenon composed of both physical and psychosocial elements and their interaction, and varies considerably among women [8]. The physical pain associated with abortion most likely originates from the S2 to S4 parasympathetic fibers (the Frankenhäuser plexus) that innervate the cervix and the lower part of the uterine body [9] and [10]. In addition, the fundus and lower part of uterine body are innervated by sympathetic fibers from T10 to L1 via the inferior hypogastric nerve and the ovarian plexus [7].

Additionally, psychological (affective, motivational, interpretive) and social (context, support) features play into pain perception [11]. Increased pain with abortion has been associated with young age, nulliparity, less education, anxiety, depression, “moral problems” (with the procedure), a retroverted uterus and dysmenorrhea [12] and [13]. A history of prior vaginal delivery correlates well with decreased pain [12]. Data on the relationship between pain and gestational age, as well as the amount of cervical dilation performed, have been conflicting [11], [12] and [14].

Due to this complex nature, effective management of abortion-related pain requires a combination of pharmacological and nonpharmacological methods [7]. Pharmacological methods include local anesthetic, nonsteroidal anti-inflammatory medications, narcotics, anxiolytics, sedatives and/or hypnotics. Concerns regarding GA stem from its association with greater costs and personnel and increased morbidity and mortality based on observational data that include analysis of cases until the mid-1980s [7], [15], [16] and [17]. Therefore, GA is less frequently used in the United States [18] and [19] compared to other countries, where abortions occur primarily in hospitals [20].

Nonpharmacological aspects of pain have a considerable impact on pain perception [7]. Active participation in one's own pain management and control over the life situation have been found to be beneficial [12].

Unfortunately, despite these advances, many patients still find surgical abortion extremely uncomfortable; 78–97% report at least moderate procedural pain [8], [12], [14] and [21]. Therefore, optimizing pain control should be a goal in every procedure. Opinions may vary on how much pain reduction is clinically relevant. Strategies designed to reduce abortion-related pain have great public health importance considering the large numbers of women who undergo first-trimester surgical abortions.

This review will examine the existing randomized controlled trials to compare the effect of different methods of pain control during first-trimester surgical abortion on patient-perceived pain, satisfaction, side effects and safety. The review will investigate preemptive as well as intraoperative analgesia, focusing on pharmacological methods administered via mucosal (oral, vaginal, intrauterine, buccal/sublingual), intramuscular or intravenous routes, but also include nonpharmacological methods. The objective of this review is to compare the effect of different methods of pharmacological and nonpharmacological pain control administered prior to or during first-trimester surgical abortion (<14 weeks gestation with electric or manual suction aspiration) on patient perceived pain, satisfaction, side effects and safety.
2. Materials and methods
2.1. Criteria for eligible trials

Randomized controlled trials, of different types of pain control, including placebo-controlled, in any language comparing pharmacological pain control administered via mucosal (oral, vaginal, intrauterine, buccal/sublingual), intramuscular, or intravenous routes or nonpharmacological pain control prior to or during a first-trimester surgical abortion at less than 14 weeks gestational age using electric or manual suction aspiration were reviewed.

The main outcome was patient-reported effectiveness of pain control or perceived pain during and immediately post-abortion using validated scales, for example, visual and verbal analogue scales, categorical or dichotomous assessment. Additional outcomes were adverse effects and side effects (including if the method of pain control causes pain), as well as patient satisfaction.
2.2. Search methods

We identified trials using the Cochrane Fertility Regulation Group search strategy. Without language preference, we searched the Cochrane Central Register of Controlled Trials (fourth quarter, 2007), MEDLINE (1950 to January 2008), EMBASE (1974 to January 2008) and POPLINE (1927 to December 2007) with appropriate keywords. We contacted professionals in the field to seek other trials and searched reference lists of articles retrieved.
2.3. Data extraction

We evaluated titles and abstracts of the search results for eligibility and retrieved full text articles if they possibly met inclusion criteria. In some cases, inclusion was based on additional information received from the study's author.

If the authors specified that they performed terminations in the first trimester, but did not state that the procedure was performed by suction aspiration versus sharp curettage, and the study was performed in a time period and region where suction aspiration was the predominant procedure method, we assumed that the study used suction aspiration as well and, therefore, included the study. If a sharp curettage check was preformed after suction aspiration, we included the study and noted this step in the individual included study table.

Quality assessment and data extraction were done by two independent authors following the Cochrane Handbook [22]. Discrepancies were resolved by consensus.
2.4. Quality assessment

We assessed methodological quality, including the study design, randomization method, group allocation concealment and exclusion after randomization without consideration of their results. The quality score for concealment of allocation per the Cochrane Handbook differentiates between (A) adequate concealment of the allocation, (B) unclear whether adequate concealment of the allocation, (C) inadequate concealment of allocation (D) allocation concealment not used [22]. Trials scoring A, B and C were included in this review. The few studies that scored C were included due to their overall importance to the review.
2.5. Analysis

We found the studies to be included in this review to be extremely heterogeneous and, thus, could not perform a single meta-analysis. However, it was possible to group trials into 7 groups:

Group 1: local anesthesia

Group 2: paracervical block with premedication

Group 3: analgesia per os only

Group 4: conscious sedation

Group 5: GA

Group 6: GA with premedication

Group 7: nonpharmacological interventions

Data were processed using the RevMan (Cochrane Collaboration, Version 5, 2008) software. Peto odds ratios (ORs) using a fixed-effects model with 95% confidence interval (CI) were calculated for all dichotomous outcomes. Weighted mean differences (WMDs) using a random-effects model with 95% CIs were used for continuous outcomes. The data from 11-point visual or verbal pain scales were treated as continuous data to allow comparisons to 10-cm scales.

For the purpose of this review, conscious sedation was defined as a drug-induced depression of consciousness during which patients responded purposefully to verbal commands (spontaneous respiration with no interventions needed to maintain a patent airway). With deep sedation, the ability to independently maintain ventilation may be impaired and patients may require assisted ventilation. GA was defined as drug-induced loss of consciousness with patients not arousable, not even by painful stimulus. Frequently, these patients will require assistance in maintaining an open airway, possibly including positive pressure ventilation. Cardiovascular function may be affected [23].

The primary outcome, pain, was assessed at different time points depending on the type of anesthesia used. With local anesthesia, pain was usually assessed during and sometimes after the procedure. In the sedation and GA groups, pain was assessed postoperatively.

Instruments used to assess pain varied: continuous, dichotomous and categorical. While many studies used an 11-point VAS (visual/verbal analog scale), some used a 100-mm VAS [24], [25] A. Edelman, M.D. Nichols, C. Leclair, S. Astley, K. Shy and J.T. Jensen, Intrauterine lidocaine infusion for pain management in first-trimester abortions, Obstet Gynecol 103 (2004), pp. 1267–1272. View Record in Scopus | Cited By in Scopus (10)[25], [26], [27], [28], [29], [30] and [31]. In order to facilitate comparability, if possible, the 100-mm VAS was converted to an 11-point VAS by dividing the results by 10 [25], [26], [29] and [30].

Within groups, we organized studies by anesthesia technique, substance and, within these subgroups, by no intervention/placebo versus intervention. In most instances, we chose to create subcategories within an outcome for doses and route of administration rather than different times of outcome assessment. Side effects were listed as subcategories when deemed appropriate for better overview.

Cointerventions were heterogeneous as well, and participants were not randomized to them, which may have affected the results or introduced bias. Due to their heterogeneity, we did not include them in our data analysis.

If results were only reported as graphs [31], [32], [33] and [34] or medians [24], [27], [28], [35] and [36] only, we could not extract data for comparisons and only described them. In other studies, percentages were reported and we calculated the number of patients based on the total number of participants per group [15], [37] and [38]. If the mean and the SEM were reported, we calculated the SD using the formula SE×square root of N [39], [40] and [41]. If the mean and the CI were reported, we calculated the SD using the formula square root of N×(upper limit−lower limit)/3.92 for 95% CI [42]. Some studies reported categorical outcomes with three groups. If deemed appropriate, the result groups were dichotomized for outcomes including pain, side effects or satisfaction to allow for analysis [24], [31], [36], [40] and [42]. In case a study compared more than two groups, we selected two groups at a time for comparison in RevMan. We attempted to contact study authors with missing or unclear data.
3. Results
3.1. Description of included studies

Forty studies [13], [15], [24], [25], [26], [27], [28], [29], [30], [31], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61] and [62] met inclusion criteria with a total of 5131 participants. Based on type of pain control, they were divided in the following groups.
3.1.1. Group 1: local anesthetics, local anesthesiatechnique, premedication and paracervical block (Table 1)

Ten studies [13], [25], [26], [28], [42], [45], [53], [60], [61] and [62] with 1527 participants investigating local anesthesia met inclusion criteria.

Only Kan et al. [28] compared paracervical block (PCB) using 1% lignocaine with no PCB. Of note, all patients underwent conscious sedation. Dilation was started 5 min after conscious sedation and 2 min after the PCB. Glantz and Shomento [13] compared 14 mL of 1% chloroprocaine with bacteriostatic saline. Other studies compared different local anesthetics. In 1992, Wiebe [60] compared 10 mL 2% carbonated lidocaine and 2 mg atropine/50 mL with 2% plain lidocaine and 2 mg atropine/50 mL. In a study in 1995, Wiebe and Rawling [61] added a third arm with 20 mL 0.25% bupivacaine, which was compared to 20 mL 1% buffered and plain lidocaine. In 1996, Wiebe et al. [62] compared 20 mL 1% with 0.5% lidocaine.

Several studies compared different techniques of PCBs, such as deep versus regular injection of 1% lidocaine (3 versus 1.5 cm and a total of 16 versus 10 mL) [45] and lidocaine with atropine (20 mL 1% lidocaine 1–1.5 in. versus 10 mL 2% lidocaine 0.5-in.; atropine 2 mg/50 mL added to lidocaine) [60].

Others studied injection at different sites of the paracervical area (3, 5, 7 and 9 o'clock versus 4 and 8 o'clock) with 14 mL of either 1% chloroprocaine or bacteriostatic (0.9% benzyl alcohol) saline [13], or injection of 10 mL 1% lignocaine at the vaginal vault versus the cervix [28]. Of note, even though some PCBs included injecting local anesthesia in the anterior or posterior lip of the cervix, in all other studies the main portion of the local anesthetic was injected at the vaginal vault around the cervix. Phair et al. [42] studied the effect of no waiting versus waiting for 3–5 min between the injection of 12 mL 1% buffered lidocaine and dilation. Wiebe and Rawling [61] had a two-step study; the first step was about waiting time, but randomization was not adequate as it was by day of the procedure. The second step of this study investigated the influence of a slow versus fast injection.

Edelman at al. [25] and [26] studied the effect of intrauterine lidocaine 10 mL of 1% and 5 mL of 4% versus placebo given in addition to the PCB with 10 mL of 1% lidocaine in two different studies. Li et al. [53] compared the topical application to the cervix (directly and via Hegar dilators) of 10 mL 2% lignocaine jelly with KY jelly.

Waiting times between administration of the PCB and dilation varied between studies. Please see [Table 1] and [Table 2] for details.
Table 1.

Local anesthesia
Local anesthesia     WMD
      Treatment     Control     Number of participants, notes, cotreatments     Dilation     Aspiration     Postoperative     Satisfaction
Local anesthetics
 Paracervical block (PCB) versus placebo/no treatment
  Glantz 2001 [13]     Chloroprocaine 1%     Bacteriostatic saline (0.9% benzyl alcohol)     14 mL, two sites (4, 8 o'clock). Three minutes of wait between PCB and dilation.     −0.5 (pain with PCB) ns     −1.5 ns     −1.9      
  Glantz 2001 [13]     Chloroprocaine 1%     Bacteriostatic saline (0.9% benzyl alcohol)     14 mL, four sites (3, 5, 7, 9 o'clock). Three minutes of wait between PCB and dilation.     −1.3 (pain with PCB)     −1.7     −1.3 ns      
  Kan 2004 [28]     Lignocaine 1%     No treatment     10 mL, two sites (4, 8 o'clock), 2.5 cm deep, cotreatment: conscious sedation. Dilation was started 5 min after conscious sedation and 2 min after the PCB.
Also other active arm; only medians reported.     ns     ns     ns      
 Various local anesthetics
  Wiebe 1992 [60]     Carbonated lidocaine 2%     Plain lidocaine 2%     10 mL, with 2 mg atropin/50 mL, no delay, three to six sites (12, 3, 6 or 12, 2, 4, 6, 8, 10 o'clock) 0.5 in. deep, no waiting.
All participants: premedication with 1 mg lorazepam sublingual 30 min prior to procedure per patient request.     −0.8           −0.4      
  Wiebe 1995 [61]     Carbonated lidocaine 1%     Plain lidocaine 1%     20 mL, (10 mL injected in four to six sites around the cervix and 5 mL each between 3 and 4 o'clock and between 8 and 9 o'clock.
1 in. deep, no waiting.
All participants: premedication with lorazepam 0.5–1 mg sublingual per patient request 30 min prior to procedure           −0.96     −0.05 ns      
  Wiebe 1996 [62]     Lidocaine 0.5%     Lidocaine 1%     20 mL.
Some patients received preoperative laminaria, lorazepam or ibuprofen.           0.2 ns            
  Wiebe 1995 [61]     Lidocaine 1%     Bupivacaine 0.25%     20 mL, as in other groups           −0.24 ns            
Local anesthesia technique
 Depth of paracervical block
  Cetin 1997 [45]     Deep injection (1 mL superficially and 3 mL 3 cm deep at 4, 6, 8, and 10 o'clock position; total of 16 mL)     Regular injection (1.5 cm deep at same 4 positions)     16 mL 1% lidocaine.
All participants: 5 mg oral diazepam 60 min prior to procedure if preprocedural anxiety of 6 or more (rated by physician not performing procedure). After 2-min wait, cervical dilation.
Vacuum aspiration followed by sharp curette.     −0.8     −0.9            
  Wiebe 1992 [60]     Superficially to blanch the mucous membrane: 1 mL injected at six sites (12, 2, 4, 6, 8 and 10 o'clock).
Then, 3–4 mL injected 1–1.5 in. deep at four sites (4, 6, 8 and 10 o'clock). Total of 20 mL 1% plain lidocaine with 1 mg atropin/50 mL.     0.5 in. deep at the reflection of the vagina off the cervix.
Three to six sites (12, 3, 6 or 12, 2, 4, 6, 8, 10 o'clock). 10 mL 2% plain lidocaine with 2 mg atropin/50 mL.     No delay
All participants: premedication with 1 mg lorazepam sublingual 30 min prior to procedure per patient request.     −2.4     −1.0            
 Paracervical block four sites versus two sites
  Glantz 2001 [13]     Four sites bacteriostatic saline (3, 5, 7, 9 o'clock)     Two sites bacteriostatic saline (4, 8 o'clock)     14 mL, also chloroprocaine in 2 groups. Three minutes of wait between PCB and dilation.     0.8 (pain with PCB) ns     0.1 ns     -0.5 ns      
 Glantz 2001 [13]     Four sites 1% chloroprocaine (3, 5, 7, 9 o'clock)     Two sites 1% chloroprocaine (4, 8 o'clock)     14 mL, also saline placebo in two groups. Three minutes of wait between PCB and dilation.     0 (pain with PCB) ns     −0.1 ns     0.1 ns      
 Waiting versus no waiting paracervical block
  Phair 2002 [42]     Waiting 3–5 min     No waiting     12 mL 1% buffered lidocaine at 12 (superficially, cervix), 4 and 8 o'clock (1–2 cm deep, paracervical).
Cotreatment: fentanyl IV and or diazepam per patient request.     −0.7     −0.2 ns     −0.1 ns     1.58 ns
 Slow versus fast injection paracervical block
  Wiebe 1995 [61]     Fast 30 s     Slow 60 s     Lidocaine 1%, 20 mL, no waiting.
Factorial design.
Outcome: pain with injection.     0.62 (pain with PCB)                  
 Intrauterine infusion
  Edelman 2004 [25]     Lidocaine 10 mL, 1%     Saline placebo 10 mL     All participants: premedication with 800 mg ibuprofen, and if requested, 5 mg diazepam. Paracervical block with 10 mL of 1% lidocaine (1 mL 1% nonbuffered lidocaine on the anterior and posterior lip of the cervix and then 4.5 mL of 1% lidocaine paracervical at the 4 and 8 o'clock positions). Three minutes of wait between intrauterine lidocaine and dilation.
100-mm VAS.     −0.3 ns     −0.4 ns     0.7 ns     −0.1 ns
  Edelman 2006 [26]     Lidocaine 5 mL, 4%     Saline placebo 5 mL     Cotreatment: ibuprofen 800 mg, cervical lidocaine 1% 10 mL, four sites, diazepam mg if requested. Three minutes of wait between intrauterine lidocaine and dilation.
100 mm VAS.     −2     −2.8     −0.5 ns     0.5 ns
 Topical
  Li 2006 [53]     Lignocaine jelly 2%
3 mL applied to cervix, to dilator and speculum     Placebo gel     Cotreatment: all subjects: cervical priming with 400 mcg misoprostol prior to the procedure (1–2 h in multiparous, 3–5 h in nulliparous subjects). Premedication with 5 mg diazepam po and 1 mg/kg pethidine IM 15-30 min prior to the procedure. Rescue pain medication with pethidine repeat dose IM. One minute of wait between topical lignocaine and dilation.     −0.42 ns     −0.87     −0.51 marginal significance      
 Cervical block
  Kan 2004 [28]     Cervical, two sites (4, 8 o'clock), 2.5 cm deep, lignocaine 1%, 10 mL     No treatment     10 mL.
Cotreatment: all patients: 400 mcg misoprostol vaginally for cervical priming 3–6 h prior to the procedure.
Conscious sedation with 2 mg midazolam and 25 mcg fentanyl IV. Dilation was started 5 min after conscious sedation and 2 min after the PCB. Pethidine IM as needed for additional analgesia.
Small trial.
Also other active arm. Only medians reported.     ns     Ns     ns      
 Cervical v paracervical
  Kan 2004 [28]     Cervical 2.5 cm deep     Paracervical, 2.5 cm deep     Lidocaine 1%, 10 mL, two sites (4, 8 o'clock).
Cotreatment: conscious sedation (details see other arm) Small trial.
Also no treatment arm. Only medians reported.     ns     ns     ns      
Paracervical block with premedication
 Wiebe 1995 [61]     Ibuprofen 600 mg po     Placebo     600 mg ibuprofen 30 min prior to procedure.
Cotreatment: PCB with 20 mL 1% lidocaine (10 mL injected in four to six sites around the cervix. All participants: premedication with lorazepam 0.5–1 mg sublingual per patient request 30 min prior to procedure.           −0.78     −0.93      
 Wiebe 2003 [59]     Lorazepam 1 mg po     Placebo     All participants received PCB.           ns            
 Suprapto 1975 [31]     Naproxen 550 mg po     Placebo     Naproxen or placebo 1–2 h preoperatively. All participants received PCB. Only the graphs with mean pain scores were presented in the article.           low asteriskp≤0.001     low asteriskp≤0.0001      
Full-size table
low asterisk Unable to obtain data to calculate weighted mean difference. P values extracted from the study.

View Within Article

 

 

 
Table 2.

Analgesia per os and conscious sedation
      Weighted mean difference or Peto OR
      Treatment     Control     Number of participants, notes, cotreatments     Dilation     Aspiration     Postoperative     Satisfaction/acceptability
Analgesia per os only
Li 2003 [29]     Diclofenac sodium 50 mg po     No medication     Diclofenac given 4 h preoperatively.
All participants: 200mcg misoprostol, lorazepam 1 mg sublingually 30 min prior to procedure. No PCB.           ns     ns     ns
Conscious sedation
Kan 2006 [27]     Entonox     Air     Entonox (50:50 mixture of nitrous oxide in oxygen) via facemask and T-piece breathing circuit.
All participants: conscious sedation with 2 mg midazolam (another 1 mg if sedation inadequate) and 25 mcg fentanyl IV. Patients received conscious sedation 2 min prior to cervical dilation. Only medians reported.           ns     ns     ns
Wong 2002 [36]     Midazolam 2 mg and fentanyl 25 mcg IV     Placebo (normal saline)     All participants: paracervical block with 10 mL of 1% lignocaine at 4 and 8 o'clock of the cervix. After the study medication, they waited 2 min to administer the PCB and 5 min to start dilation.           ns     ns     3.69 (Peto OR)
Wells 1992 [58]     PCB with IV sedation (diazepam and fentanyl)     PCB alone     No standard deviation reported.           low asteriskp<0.003            
Raeder 1992 [15]     Conscious sedation and PCB     GA     Group 1: regional anesthesia: midazolam 0.1 mg/kg IV and alfentanil 0.01 mg/kg IV followed by PCB with 20 mL mepivicaine 20 mg/mL with adrenaline 0.005 mg/mL. Patients breathed spontaneously and were assisted with oxygen as needed.
Group 2: GA: alfentanil 0.01 mg/kg IV, followed by bolus of propofol 2.0 mg/kg. Patients breathed 75% nitrous oxide in oxygen spontaneously by mask and were assisted as needed for dropping oxygen saturation. After conscious sedation they waited 2 min until administration of the PCB.     14.77 (Peto OR)     7.47 (Peto OR)     −1.00 WMD      
Full-size table
low asterisk Unable to obtain data to calculate weighted mean difference. P values extracted from the study.

View Within Article

 

3.1.2. Group 2: PCB with premedication (Table 1)

Three studies [31], [59] and [61] with 434 participants investigated the effect of premedication, such as ibuprofen 600 mg per os [61], lorazepam 1 mg per os [59] or naproxen sodium 550 mg per os [31] followed by a PCB with 1% lidocaine (20 mL in Ref. [61]).
3.1.3. Group 3: analgesia alone (Table 2)

One study with 100 participants investigated diclofenac sodium 50 mg combined with 200 mcg misoprostol versus misoprostol alone [29]. Of note, patients did not receive a PCB but lorazepam 1 mg sublingually 30 min prior to the procedure.

Since patients were awake during the procedure under these different local anesthetic techniques, outcomes included pain with dilation, aspiration and post-procedure. Some studies also measured pain with the paracervical/cervical block application [28].

The predominant study instruments used to measure pain were visual and verbal analogue scales; some 11-point, some 100-mm. Additional outcomes were anxiety, satisfaction, sedation, side effects, difficulty of the procedure and varied between the studies.
3.1.4. Group 4: conscious sedation (Table 2)

Three studies [27], [36] and [58] with 274 participants investigated conscious sedation. Kan et al. [27] compared entonox (50:50 mixture of nitrous oxide in oxygen) with air after administering 2 mg midazolam (another 1 mg if sedation inadequate) and 25 mcg fentanyl IV. Patients received conscious sedation 2 min prior to cervical dilation. In this study, patients did not receive a PCB. Wong et al. [36], on the other hand, compared conscious sedation with midazolam 2 mg and fentanyl 25 mcg IV with placebo. After the study medication, they waited 2 min to administer the PCB and 5 min to start dilation. Wells [58] compared a local cervical block alone with a cervical block combined with intravenous sedation using diazepam and fentanyl in two of their four arms.
3.1.5. Group 5: GA (Table 3)

Fourteen studies [15], [32], [33], [37], [38], [40], [41], [43], [44], [49], [50], [51], [54] and [55] with 1812 participants investigated GA (Table 3). One of them compared GA using propofol and alfentanil with conscious sedation using midazolam 0.1 mg/kg IV, alfentanil 0.01 mg/kg IV and PCB with 20 mL mepivacaine 20 mg/mL with adrenaline 0.005 mg/mL [15]. Two minutes after conscious sedation, the PCB was given. In the GA group, patients breathed 75% nitrous oxide in oxygen spontaneously by mask and were assisted as needed for dropping oxygen saturation; in the sedation group, patients breathed spontaneously and were assisted with oxygen as needed.

 
Table 3.

General anesthesia
Study     Fentanyla     Alfentanil a     Midazolam/diazepam/lorazepam (benzodiazepine)b     Propofolb     Ketamineb     Methohexital (barbiturate)b     Thiopental (barbiturate)b     Etomidatec     Halothane c     Enfluranec     Trichlo-rethylenec     N2O2
Barneschi1985[37]     x           x           x           x           x     x           c
Bonnardot 1987[38]           x     c     x     x                                          
Boysen 1989 [43]           c           x                 x     x                        
Boysen 1990 [44]           c           x           x                                    
Collins 1985 [40]           x                       c                 x                 c
Hackett 1982[41]     x                             c                       x           c
Hall 1997[46]                       x                                               c
Jakobsson 1991[49]     x     X (versusfentanylversusplacebo)           c                 x                             c
Jakobsson 1993[50]     x                 x     x     x     x                              
Jakobsson 1995[51]     x     x           x                 x                             c
Lindholm 1994[33]     x     X versusfentanylversusnormalsaline           c                                                
Ogg 1983[54]     x                             c                             X     x
Raeder 1992[15]           X     x     x                                               x
Rossi 1995[55]     x           x     x     x                                         c
Full-size table

x=control or intervention, c= cotreatment.
a Opiates.
b Sedative hypnotic agents.
c Inhalational anesthetic, N2O2=nitrous oxide.

View Within Article

 

Hall et al. [46] compared GA using propofol with GA and PCB combined. GA studies used either fentanyl or alfentanil as opiates for pain control.

Four studies investigated inhalational anesthetics, specifically, halothane [37] and [40], enflurane [41] and trichloethylene [54] and compared them to various sedative/hypnotic agents. All studies included at least one sedative/hypnotic agent. Ten studies included propofol [15], [32], [33], [38], [43], [44], [49], [50], [51] and [55], nine studies included a barbiturate (five methohexital [40], [41], [44], [50] and [54], five thiopental [37], [43], [49], [50] and [51], four ketamine [37], [38], [50] and [55], three benzodiazepine midazolam [15], [38] and [55] and one etomidate [43]).

In studies with GA, pain was usually assessed postoperatively, either as a dichotomous or a categorical variable. Other outcomes included typical side effects such as pain with injection, nausea, vomiting, and apnea, various tests of recovery and time until discharge.
3.1.6. Group 6: GA with premedication (Table 4)

Seven studies [24], [30], [35], [39], [47], [48] and [52] with 770 participants investigated the influence of premedication with various analgesics on postoperative pain after GA (Table 4). Most studies included a cyclooxygenase inhibitor (COX); COX 3 paracetamol [24], [47] and [48], COX 2 etoricoxib [30], nonselective COX inhibitor ketoroloac [52], sodium and potassium diclofenac [52] and lornoxicam [48]. Other studies investigated opioids such as nalbuphine [39], dihydrocodeine [35] and paracetamol with codeine [24]. In five out of seven studies, GA was achieved with propofol [24], [30], [35], [47] and [48]; in one of them enflurane [39] and in another desflurane [30] was added. Thiopental was used in the two other studies [39] and [52]. All but one [30] included either fentanyl [39], [47] and [48] or alfentanil [24], [35] and [52] for anesthesia.

 
Table 4.

GA with premedication
Study     Fentanyl     Alfentanil     Midazolam (benzodiazepine)     Propofol     Thiopental (barbiturate)     Desflurane     Enflurane     N2O2     Opioid     Nonselective COX     Selective COX
Bone1988[39]     x                       x           X     x     Nalbuphineversus fentanyl            
Dahl 2000[24]           x     x     x                       x     Paracetamolwith codeine           Paracetamol(COX 3)versus placebo
Heath1989[35]           x           x                       X     dihydrocodeine versus placebo            
Hein 1999[47]     x                 x                       X                 Paracetamolversusplacebo
Hein2001[48]     x                 x                       x           Lornoxicam     Paracetamol
Jakobsson1996[52]           x                 x                 x           Sodiumdiclofenacversus ketorolacversuspotassiumdiclofenacversus NaCl      
Liu 2005 [30]                       x           x           x                 Etoricoxibversusplacebo(COX 2)
Full-size table

View Within Article

 

3.1.7. Group 7: nonpharmacological interventions

Four [34], [56], [57] and [58] very different studies with 214 participants investigated nonpharmacological interventions. In a recent study [34], the effect of hypnosis was investigated compared to standard care in patients who all received a PCB. Shapiro and Cohen [56] compared 3 groups; one control and two treatment arms with self-administered methoxyflurane (0.5 volume % with 5 L oxygen per minute) and stereophonic headphones with music chosen by patient. A further study compared provision of sensory information (3-min audio taped message containing orienting information as well as nine sensations related to abortion, and identified by over 50% of women in a previous pilot study) with provision of general information. They also compared PCB versus PCB plus intravenous sedation with diazepam and fentanyl [58]. Wells [57] compared four groups: (1) attention control without technique instruction, but advice to use a coping strategy that worked in a previous painful experience; (2) relaxation exercise for 10 min prior to the procedure; (3) guided pleasant imagery (beach or mountain); (4) guided analgesic imagery. All participants received a PCB.

Cointerventions, which at times are options, included cervical ripening, premedication and anesthesia induction. The most important ones are included in [Table 1] and [Table 2].
3.2. Description of excluded studies

Twenty-nine studies were excluded for the following reasons (some met several criteria):

1. The authors did not state gestational age/or included a higher gestational age than 14 weeks and did not specify that suction aspiration was performed (5 studies)

2. The procedure was a sharp curettage only (1 study)

3. Pain was not assessed at all or only indirectly as the need for postoperative analgesics (15 studies)

4. Inadequate randomization or allocation (9)

3.3. Methodological quality of included studies

Randomization was described in 25 of the studies; most often computer randomization. In 16 studies, the authors stated that they did randomize, but not how this was performed. The lack of information on randomization and allocation concealment likely derives from the fact that many of these publications were from the 1970s, 1980s and early 1990s.

Allocation concealment was adequate in 23 included studies and unclear in 14 studies. Three studies had inadequate allocation concealment. They had a research assistant draw up the syringes or used an envelope but did not designate it as opaque.

Blinding. Patients were blinded in many studies. The surgeons and anesthesiologists could not always be blinded due to the individual study designs, which may have introduced bias. However, assessors of postoperative outcomes were usually blinded.

Follow-up and exclusions. Due to the short follow-up period until discharge after surgery, loss to follow-up did not occur. In several studies, patients were excluded after inappropriate inclusion; more often, data collection was incomplete. This made true intention-to-treat analysis more difficult.
3.4. Results of studies
3.4.1. Group 1: local anesthetics (Comparison 1) (Table 1)

Kan et al. [28] did not observe a difference in pain with dilation, aspiration or postoperatively when comparing PCB using lignocaine with no injection in patients with conscious sedation (only medians reported, n=89). Glantz and Shomento [13] found better pain control during injection (WMD, −0.90; 95% CI, −1.78 to −0.02), aspiration (WMD, −1.50; 95% CI, −2.45 to −0.55) and postoperatively (WMD, −1.5; 95% CI, −2.54 to −0.46; n=79) with a 1% chloroprocaine PCB, injected at either two or four sites, compared to a bacteriostatic saline PCB. The significant decrease in pain with injection and aspiration was driven by the four-site chloroprocaine PCB and not the two-site block. In turn, the significant decrease in postoperative pain was driven by the 2-site chloroprocaine PCB.

A PCB with buffered 2% lidocaine achieved more effective pain control with cervical dilation and at the end of the procedure than with plain 2% lidocaine (WMD, −0.80; 95% CI, −0.89 to −0.71; WMD, −0.40; 95% CI, −0.49 to −0.31; n=167) [60]. Buffered 1% lidocaine improved pain with aspiration compared to plain 1% lidocaine (WMD, −0.96; 95% CI, −1.67 to −0.25; n=124), but not postoperative pain [61]. Pain control with aspiration did not differ when comparing lidocaine 0.5% with 1% bupivacaine or 1% lidocaine with 0.25% bupivacaine [61] and [62].
3.4.2. Local anesthesia technique (Comparison 2) (Table 1)

Deep injection achieved better pain control than regular injection for cervical dilation and aspiration when combining the results of two studies with a total of 113 patients (WMD, −1.64; 95% CI, −3.21 to −0.08 and WMD, −1.00; 95% CI, −1.09 to −0.91) [45] and [60].

Pain with PCB injection, aspiration and postoperatively did not differ when comparing a four-site (3–5–7–9 o'clock) with a two-site (4–8 o'clock) injection [13]. Similarly, it did not differ when comparing a cervical block with lignocaine injected at 4 and 8 o'clock into the cervix versus the vaginal vault in patients with conscious sedation [28]. Kan et al. [28] only reported medians and thus the actual data could not be abstracted.

Waiting 3 min between PCB and dilation improved pain with cervical dilation (WMD, −0.7; 95% CI, −1.37 to −0.03; n=194), but not with aspiration or postoperative pain [42]. Of note in the original article, no significant results were described. Since only CIs were reported, we calculated standard deviation as described in the data synthesis section, and obtained the following results.

Fast injection increased pain with application of the PCB (WMD, 0.62; 95% CI, 0.06–1.18) [61].

A 1% intrauterine lidocaine infusion plus PCB was not more effective in controlling pain with cervical dilation or aspiration as compared to PCB with intrauterine placebo, but a 4% intrauterine lidocaine infusion plus PCB was (WMD, −2.0; 95% CI, −3.29 to −0.71; WMD, −2.8; 95% CI, −3.95 to −1.65; n=80 each study) [25] and [26].

Topical lignocaine gel compared to KY jelly did not alter pain with cervical dilation or postoperative pain, but alleviated pain with aspiration (WMD, −0.87; 95% CI, −1.60 to −0.14; n=131) [53].
3.4.3. Group 2: PCB with premedication (Comparison 3) (Table 1)

Ibuprofen, 600 mg orally given 30 min preoperatively improved pain control with aspiration and postoperatively compared to placebo (WMD, −0.78; 95% CI, −1.52 to −0.04; WMD −0.93; 95% CI, −1.62 to −0.24; n=193) [61], while 1 mg oral lorazepam, given 1 h preoperatively, did not make a difference [59]. Naproxen, given 1–2 h preoperatively, decreased pain compared to placebo (max pain during procedure, p≤0.001; 15 min postoperatively, p≤0.0001; 30 min postoperatively, p≤0.002) [31]. Respective values for naproxen versus no drug were p≤0.001 with abortion and p=0.059 30 min postoperatively.
3.4.4. Additional outcomes and sub-analysis resultsreported in studies for Comparisons 1, 2 and 3

Lorazepam received per patient request did not affect pain in patients undergoing the procedure with a PCB [60] and [61]. It also did not significantly impact anxiety in an RCT [59].

Sub-analysis for nulliparity versus multiparity was performed and showed significantly lower pain scores with multiparity on arrival in the OR, with cervical manipulation/dilation and overall intraoperatively. Multiparous women were significantly more satisfied; type of anesthesia did not alter satisfaction [53].

Many studies did not study patient satisfaction, but in those that did, satisfaction was high in both study arms [25], [26], [28] and [42].
3.4.5. Group 3: analgesia alone (Comparison 4) (Table 2)

One study investigated diclofenac sodium 50 mg, given 4 h preoperatively, combined with 200 mcg misoprostol compared to misoprostol alone in the absence of a PCB and did not find differences in pain control with aspiration or postoperatively, or with acceptability of pain control [29]. If broken down into nulliparous and multiparous, there was significantly less pain with diclofenac sodium in multiparous women during the procedure [mean 58 (SD 27) versus mean 63(SD 27)).
3.4.6. Group 4: conscious sedation (Comparison 5) (Table 2)

Pain with aspiration and 1 h postoperatively did not differ when comparing entonox (50:50 mixture of nitrous oxide in oxygen) with air when added to conscious sedation with midazolam and fentanyl [27]. Only median and 95% CI were reported. Anxiety, satisfaction level as well as side effects (nausea, dizziness, dry mouth and drowsiness) did not significantly vary between groups [27].

While pain with aspiration and postoperatively did not differ comparing conscious sedation using midazolam 2 mg and fentanyl 25 mcg IV with placebo after administering a PCB to all participants (only medians were reported), satisfaction was higher with conscious sedation (Peto OR, 3.69; 95% CI, 1.63–8.36; n=100) [36]. No difference was observed in sedation. Postoperatively, more dizziness (p=0.015) and drowsiness (p<0.001) were noted in the conscious sedation group. Multiple regression showed that sedation (decreased, p=0.008) and gestational age (increased, p=0.024) affected pain [36]. A second study compared PCB and conscious IV sedation using diazepam and fentanyl with PCB alone [58]. In this study, which does not report SDs, women with IV sedation reported less pain (mean, 4.54 versus 6.30; p=0.003) (F(1.8)=9.40) n=84). Pain intensity further correlated with subjective (r=.74, p>0.001) and behavioral distress (r=.54, p<0.001) [58].
3.4.7. Group 5: GA (Comparison 6) (Table 3)
3.4.7.1. Inhalation anesthetics

Four studies included inhalation anesthetics [37], [40], [41] and [54]. Halothane did not change postoperatively reported pain compared to alfentanil when added to methohexital [40]. Adding halothane, enflurane or fentanyl to thiopental did not affect postoperative pain [37]. Trichlorethylene did not change pain control compared to methohexital [54]. Enflurane compared to fentanyl did not affect pain when added to methohexital [41].
3.4.7.1.1. Side effects of inhalation anesthetics

Higher blood loss was noted with inhalation anesthetics, such as enflurane [41] and halothane [40] per reported study results. Since no CI was given, we could not recalculate this. Data on nausea and vomiting were controversial; less with halothane compared to fentanyl [40], more with enflurane compared to fentanyl [41] and no difference with trichlorethylene [54]. Halothane anesthesia was associated with more cough compared to alfentanil given for maintenance after methohexital induction. However, halothane was associated with less limb movement. Since only ranges were given, we could not recalculate the statistics [40]. Laryngospasm, pain on induction and intraoperative muscle movement did not differ between trichloethylene and total IV anesthesia [54]. Severe anesthesia complications, as well as apnea [40], did not differ [37], [40] and [41]. Recovery time after halothane was longer compared to alfentanil (WMD, 7.6; 95% CI, 5.71–9.49; n=66) [40], while enflurane and fentanyl did not differ [41]. Memory function as part of the recovery testing did not differ between groups [54]. Anesthesia with volatile agents was considered safe and reliable [37].
3.4.7.2. Sedatives, hypnotics and opiates

Ten studies included propofol. Postoperative pain did not differ comparing propofol with etomidate [43]. In a meta-analysis of three studies with 350 patients comparing propofol and thiopental, no differences in postoperative pain were measured regardless of adding fentanyl or alfentanil [43], [50] and [51]. Propofol was associated with decreased postoperative pain compared to methohexital (Peto OR, 0.28; 95% CI, 0.10–0.80; n=100) [50]. However, Boysen et al. [44] showed a trend towards the reverse with more postoperative pain in the proprofol group. Combining the results of these two studies cancelled out the effect and no significant difference was found between methohexital and proprofol for postoperative pain [44] and [50].

Midazolam and propofol when added to fentanyl did not differ in reported postoperative pain [55].

Combination of propofol and alfentanil achieved better postoperative pain control compared to 0.5 mg/kg ketamine and 0.25 mg/kg midazolam (Peto OR, 0.18; 95% CI, 0.07–0.47; n=100); the trend when using 1 mg/kg ketamine and 0.1 mg/kg midazolam was not significant [38]. Ketamine was associated with more postoperative pain than fentanyl when added to propofol (Peto OR, 7.13; 95% CI, 2.99–17.0; n=100) [50]. Even though Rossi et al. [55] did not confirm this, the association remained significant in the meta-analysis (Peto OR, 4.66; 95% CI, 2.16–10.06; n=180).

Adding alfentanil to propofol reduced postoperative pain (Peto OR, 0.16; 95% CI, 0.06–0.40; n=100) [38]. However, in the study of Jakobsson et al. [49], adding alfentanil to propofol did not change pain over placebo plus propofol, but adding fentanyl did (Peto OR 0.23 95% CI 0.11 to 0.51; n=208). Combining the results from the two studies of Jakobsson et al. [50] and [51], the alfentanil/propofol combination was less effective for postoperative pain, then fentanyl/propofol (Peto OR, 1.96; 95% CI, 1.07–3.6; n=210) [49] and [51]. In the arm with thiopental, pain did not differ [51]. At 30 min postoperatively, pain was less in the fentanyl (p<0.05) and alfentanil (p<0.01) group compared to placebo [33]. Pain intensity was equal among the groups at 120 and 180 min.
3.4.7.3. Side effects of sedatives, hypnotics and opiates

Increased pain with injection was associated with propofol and etomidate compared to thiopental [43] and ketamine [38]. Adding alfentanil to propofol decreased this pain [38]. Propofol was associated with increased apnea compared to etomidate but not thiopental or methohexital [43] and [44]. Muscle movement was increased with etomidate as compared to propofol, as well as thiopental [43] and increased with methohexital as compared to propofol [44]. Intraoperative movement was further increased with thiopental and fentanyl compared to ketamine and diazepam as well as thiopental and halothane [37]. Nausea was decreased with propofol compared to methohexital in one study [50], but not in another [44]. It was also decreased with thiopental and fentanyl compared to ketamine and diazepam [37]. Vomiting was decreased with propofol and alfentanil compared to ketamine and midazolam [38]. Dreams were increased with ketamine compared to other sedative hypnotics [50].

Nausea, vomiting, laryngospasm and overall complications did not differ when comparing alfentanil or fentanyl with each other or placebo [33] and [49]. Propofol induction dose was significantly lower in the alfentanil group compared to fentanyl (p<0.05; only medians given). The total propofol dose required, and the number of people moving to surgical stimulus, was significantly lower in both the fentanyl and alfentanil groups as compared to the normal saline (NS) control group (p<0.01), while recovery measures were improved in the alfentanil group as compared to the NS control group [33]. Time to discharge was shorter after propofol compared to thiopental in the meta-analysis of two studies (WMD, −14.69; 95% CI, −24.95 to −4.43; n=200) [50] and [51]. Adding an opioid compared to placebo did not alter the time [49]. Recovery was faster in propofol/fentanyl group compared to ketamine/fentanyl and fentanyl/midazolam as assessed per Steward score [55]. Speed and quality of psychomotor and sensory tests were significantly better in the propofol groups compared to the ketamine groups [38], as it was in the thiopentane group compared to ketamine [37].

Per study reports, overall, propofol was associated with a better recovery compared to etomidate and thiopental [43], and a similar recovery compared to methohexital [44]. Pain significantly correlated to prolonged time until hospital discharge [49].
3.4.7.4. PCB added to GA

One study investigated if a PCB added to GA altered postoperative pain control [32]. Pain, intraoperative propofol use and postoperative pain medication consumption did not change, nor did nausea or time until discharge.
3.4.7.5. Conscious sedation with GA

One study with 59 patients directly compared conscious sedation with GA [15]. With conscious sedation combined with a PCB, pain with dilation and aspiration was higher as assessed by the anesthesiologist (Peto OR, 14.77; 95% CI, 4.91–44.38, and Peto OR, 7.47; 95% CI, 2.2–25.36). However, postoperatively reported pain was decreased (WMD, −1.00; 95% CI, −1.77 to −0.23). Risk for apnea was reduced with conscious sedation (Peto OR, 0.10; 95% CI, 0.02–0.46), and duration of sleep was shorter (WMD, −9.5; 95% CI, −11.5 to −7.5). Except for better p-deletion score (a test in which patients are shown a sheet of randomly written letters and are instructed to delete with a pen all p's as fast and accurately as possible during a 3-min period) 30 min after the procedure in the GA group, there was no difference in the recovery functions between the groups, as per reported results [15].
3.4.8. Group 6: GA with premedication (Comparison 7) (Table 4)

Seven studies investigated the influence of premedication with various analgesics (selective or nonselective COX inhibitor or opioids) on postoperative pain after GA (mostly propofol and fentanyl or alfentanil).
3.4.8.1. COX 3 inhibitor

The selective COX 3 inhibitor, paracetamol, given as a suppository at the end of the procedure did not improve pain control compared to placebo [47]. Even when adding codeine to the paracetamol suppository and giving it 1 h preoperatively, pain did not improve compared to placebo [24]. The nonselective COX inhibitor lornoxicam significantly decreased postoperative pain compared to paracetamol dosed orally (Peto OR, 0.36; 95% CI, 0.17–0.78; n=140), which in turn did not change pain compared to placebo. All test drugs were given 1 h before anesthesia [48]. Diclofenac IM and ketorolac IM both decreased postoperative pain compared to NaCl when given 10–20 min before the anesthesia (Peto OR, 0.37; 95% CI, 0.14–0.92, and Peto OR, 0.32; 95% CI, 0.12–0.81; n=100) [52] and did not differ when compared to each other. Diclofenac orally was associated with more postoperative pain compared to ketorolac IM (Peto OR, 3.17; 95% CI, 1.24–8.13; n=100) and did not improve pain control compared to NaCl [52]. The COX-2 inhibitor etoricoxib, given 30–60 min preoperatively, did not improve pain control compared to placebo immediately postoperatively but did by the time of discharge (WMD, −0.7; 95% CI, −1.2 to −0.2; n=40) [30].
3.4.8.1.1. Side effects

Comparing COX inhibitors with placebo did not show any difference regarding antiemetic requirements, nausea, vomiting, anxiety or satisfaction except for diclofenac IM, decreasing nausea and anxiety compared to NaCl (Peto OR, 0.13; 95% CI ,0.02–0.93, and Peto OR, 0.29; 95% CI, 0.09–0.94; n=100) [52]. Ketorolac decreased anxiety compared to NaCl (OR, 0.29; 95% CI, 0.09–0.94; n=100) [52]. Time to discharge was the same in all groups [30], [47], [48] and [52].
3.4.8.2. Opioids

Nalbuphine achieved better 1-h postoperative pain control than fentanyl (Peto OR, 0.21; 95% CI, 0.05–0.86; n=40) [39]. The incidence of postoperative pain and nausea was the same when comparing dihydrocodeine po with placebo [35].
3.4.8.2.1. Side effects

Nausea and recovery (reaction time) did not differ between nalbuphine and fentanyl [39]. Paracetamol with codeine suppository compared to placebo did not change nausea or awakeness/sleepiness at most time points measured, except for more women being sleepy at 30 min postoperatively after paracetamol with codeine (Peto OR, 3.17; 95% CI, 1.39–7.23) and less fully awake (Peto OR, 0.35; 95% CI, 0.15–0.79; n=90) [24]. Time to discharge was not affected [24].
3.4.9. Group 6: nonpharmacological intervention(Comparison 9)

Four very different studies investigated nonpharmacological interventions. In patients with a PCB, hypnosis did not did not change the level of comfort during the procedure compared to standard care; however, it decreased the requests for nitrous oxide (Peto OR, 0.12; 95% CI, 0.03–0.54) [34]. Listening to stereo music compared to self-administration of methoxyflurane decreased pain with aspiration (Peto OR, 0.17; 95% CI, 0.04–0.63; n=98) [56]. Providing sensory (3-min audiotaped message containing orienting information as well as nine sensations related to abortion and identified by over 50% of women in a previous pilot study) compared to general information did not affect procedural pain or distress [60]. Relaxation exercise did not change procedural or postoperative pain compared to pleasant or analgesic imagery, or a control group [57].
4. Discussion
4.1. Summary of main results

Various methods of pain control for first-trimester surgical abortion have been studied including local anesthesia, IV sedation, GA and some forms of nonpharmacological pain control. Many of them have been found to effectively decrease pain compared to placebo or other pain management strategies during and after the procedure while being safe and satisfactory to patients.

Data on the effect of a PCB are heterogeneous and very limited. The only study that compared a PCB to no PCB did not show a benefit of the PCB [28]. This result is limited by the small number of participants (n=89) and the fact that all patients received conscious sedation with 2 mg midazolam IV and 25 mcg fentanyl IV.

While one small study showed a pain reduction with injection and aspiration when using 1% chloroprocaine compared to normal saline for a four-site PCB, this effect was not noted for a two-site PCB except for postoperative pain. Additionally, there was no significant difference when two- and four-site PCB were directly compared [13]. Pain with cervical dilation was improved with a PCB using carbonated lidocaine compared to noncarbonated [60], deep injection of the paracervical block [45] and [60], waiting 3 min between PCB and dilation [42], and with adding a 4% intrauterine lidocaine infusion to PCB [26]. All but waiting 3 min also decreased pain with aspiration [26], [42], [45], [60] and [61]. Fast injection increased pain with application of the PCB [61]. Premedication with ibuprofen and naproxen (p<0.001) improved intra- and postoperative pain [31] and [61]. The addition of conscious IV sedation using diazepam and fentanyl to PCB decreased pain with the procedure [58]. Adding a PCB to GA has not been shown to reduce intraoperative anesthetic needs or postoperative pain [32]. Conscious sedation combined with PCB does not achieve the same pain control as GA during the procedure, but improved postoperative pain control [15].

In regard to GA, a shift from inhalational anesthetics to sedatives and hypnotics has decreased procedure-related blood loss. Propofol has been shown to be superior to ketamine in multiple studies and shortened time until discharge compared to some other hypnotics. Adding opioids or IM COX inhibitors to GA has been found to be beneficial for postoperative pain, and opioids have decreased GA needs. No major complications were observed in any study.
4.2. Overall completeness evidence

Various countries, decades of years, settings in which the procedure was provided and pain management options have been represented by the included studies. Methods of pain control varied widely and were often combined regimens. In order to synthesize the data, we grouped the included trials as mentioned previously. Trials were too heterogeneous regarding combination of medications, doses and routes of administration to be combined in a large meta-analysis. Therefore, we focused on the primary outcome of pain and were unable to draw firm conclusion on side effects or complications. In addition, the nature of GA, which achieves complete pain control during the surgery, challenges the ability to compare it to any other form of anesthesia. Pain during deep conscious sedation and GA can only be assessed by an observer and cannot be patient-reported, which decreases comparability to other forms of anesthesia.
4.3. Quality of evidence

Randomization and allocation concealment were not specified in one third and one half of the studies, respectively. This likely derives from the fact that many publications were from the 1970s to early 1990s. Three studies [47], [60] and [62] had inadequate allocation concealment, but were included due to their overall importance to the review.

Several studies reported incomplete data, and some of the GA literature did not contain detailed gynecologic information (e.g., type of procedure, sharp curettage versus suction). Not all authors could be successfully contacted to obtain missing information.

Some studies had statistically significant results; however, they only detected a small change in pain (WMD <1) [30], [42], [45], [53], [60] and [61] or did not measure the amount in pain reduction they had determined in their power calculations (i.e., Ref. [42]). Of note, Phair et al. [42] did not originally report any statistically significant results, but in our reanalysis, pain with dilation was reduced. This raises the questions of quality of evidence and points out that clinically significant pain reduction is hard to determine.
5. Conclusion

Many patients still find first-trimester surgical abortion extremely uncomfortable due to pain with cervical dilation and aspiration, unless given GA. Given how widely PCB is used, the paucity of data supporting the benefit of a PCB as shown in this review is surprising and concerning. Although PCB appears relatively safe, no strong data exist regarding its effectiveness for pain control.

Severe complications of anesthetics are rare; therefore, none of the included studies were powered to detect these. Due to short follow-up, delayed side effects may have been missed. However, most of the medications studied do not have a long half-life.

Recommendations by the WHO on safe abortion [1] as well as by the Royal College of Obstetricians and Gynaecologists [63] favor local/IV sedation over GA. These recommendations are likely based on data from the 1970 to 1980s. As the overall mortality from legally induced abortions has decreased since the 1970s from 4.1 in 100,000 to 0.7 in the United States, the percentage mortality related to GA increased from 7.7% to 29.4% [6]. Between 1988 and 1997, deaths attributed to any anesthesia decreased to 22% [4]. However, other large reviews of abortions have not shown a significant difference in major complications between local and GA [3] and [64]. Data from the Closed Claims Project of the American Society of Anesthesiologists has shown a decrease in the percentage of death and brain damage claims related to respiratory events between the period 1970–1979 and the period 1990–1994 from 56% to 39% [65]. This is thought to be due to improved monitoring including pulse oximetry and capnography [65]. We lack an up-to-date observational data summary on risks of GA in first-trimester surgical abortions.

In any case, in order to prevent complications, the provider must have a profound respect for the continuum from anxiolysis to unconsciousness. It is imperative that patients be monitored appropriately by qualified personnel who are knowledgeable about pharmacokinetics and pharmacodynamics and who are experienced in airway management resuscitation [23]. Therefore, the setting in which abortions take place strongly affects available resources and risks from anesthesia.

In the light of these findings, factors such as women's preference, medical risk factors for anesthesia complications, setting and resource availability should be considered when choosing a method of pain control.

Future studies should aim for using the same outcomes and study instruments to measure pain in order to increase comparability. In order to establish whether PCB is effective or not, a well-designed study is needed, comparing PCB to a no treatment arm rather than placebo. More studies should try to compare local anesthesia with conscious sedation and GA regarding pain during and after the procedure as well as regarding side effects, required doses of GA, time to discharge and satisfaction. The nature of GA, which achieves complete pain control during the surgery, challenges direct comparison to any other form of anesthesia.

Newer observational data on risks of GA will further help to improve its adequate risk perception. Such data may revise current recommendations.
Acknowledgments

This article is based on a Cochrane Review published in The Cochrane Library 2009, Issue 2 (see www.thecochranelibrary.com for information) [66]. Cochrane Reviews are regularly updated as new evidence emerges and in response to feedback, and The Cochrane Library should be consulted for the most recent version of the review. If you wish to comment on this review, please contact Dr. Regina Renner (email at rennerr@ohsu.edu). The results of a Cochrane Review can be interpreted differently, depending on people's perspectives and circumstances. Please consider the conclusions presented carefully. They are the opinions of the reviewers and are not necessarily shared by the Cochrane Collaboration.

No funds were received to perform this review.

We thank Dr. P. Thorborg who provided clinical expertise in anesthesia-related questions and our translators: Jana Jarosch (Russian), Chiara Ghetti (Italian), Annemarie Renner (French).
References
[1] WHO, Safe abortion:Technical and policy guidance for health systems http://www.who.int/reproductive-health/publications/safe_abortion/ (2003).
 
[2] L.T. Strauss, S.B. Gamble, W.Y. Parker, D.A. Cook, S.B. Zane and S. Hamdan, Abortion surveillance—United States, 2004, MMWR Surveill Summ 56 (2007), pp. 1–33. View Record in Scopus | Cited By in Scopus (11)
 
[3] E. Hakim-Elahi, H.M. Tovell and M.S. Burnhill, Complications of first-trimester abortion: a report of 170,000 cases, Obstet Gynecol 76 (1990), pp. 129–135. View Record in Scopus | Cited By in Scopus (86)
 
[4] L.A. Bartlett, C.J. Berg and H.B. Shulman et al., Risk factors for legal induced abortion-related mortality in the United States, Obstet Gynecol 103 (2004), pp. 729–737. View Record in Scopus | Cited By in Scopus (62)
 
[5] L.M. Koonin, L.T. Strauss, C.E. Chrisman and W.Y. Parker, Abortion surveillance—United States, 1997, MMWR CDC Surveill Summ 49 (2000), pp. 1–43. View Record in Scopus | Cited By in Scopus (17)
 
[6] H.K. Atrash, T.G. Cheek and C.J. Hogue, Legal abortion mortality and general anesthesia, Am J Obstet Gynecol 158 (1988), pp. 420–424. View Record in Scopus | Cited By in Scopus (9)
 
[7] D.S. Maltzer, M.C. Maltzer, E.R. Wjebe, G. Halvorson-Boyd and C. Boyd, Pain management. In: M. Paul, E.S. Lichtenberg, L. Borgatta, D.A. Grimes and P.G. Stubblefield, Editors, A clinician's guide to medical and surgical abortion, Churchill Livingstone, Philadelphia, PA (1999), pp. 73–89.
 
[8] P.G. Stubblefield, Control of pain for women undergoing abortion, Suppl Int J Gynecol Obstet 3 (1989), pp. 131–140. Article | PDF (651 K) | View Record in Scopus | Cited By in Scopus (24)
 
[9] J. Scott and E.C. Huskisson, Graphic representation of pain, Pain 2 (1976), pp. 175–184. Abstract | PDF (712 K) | View Record in Scopus | Cited By in Scopus (771)
 
[10] R.P. Smith and J.A. Heltzel, Interrelation of analgesia and uterine activity in women with primary dysmenorrhea. A preliminary report, J Reprod Med 36 (1991), pp. 260–264. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (9)
 
[11] L. Borgatta and D. Nickinovich, Pain during early abortion, J Reprod Med 42 (1997), pp. 287–293. View Record in Scopus | Cited By in Scopus (18)
 
[12] E. Belanger, R. Melzack and P. Lauzon, Pain of first-trimester abortion: a study of psychosocial and medical predictors, Pain 36 (1989), pp. 339–350. Abstract | PDF (1204 K) | View Record in Scopus | Cited By in Scopus (46)
 
[13] J.C. Glantz and S. Shomento, Comparison of paracervical block techniques during first trimester pregnancy termination, Int J Gynaecol Obstet 72 (2001), pp. 171–178. Article | PDF (74 K) | View Record in Scopus | Cited By in Scopus (19)
 
[14] G.M. Smith, P.G. Stubblefield, L. Chirchirillo and M.J. McCarthy, Pain of first-trimester abortion: its quantification and relations with other variables, Am J Obstet Gynecol 133 (1979), pp. 489–498. View Record in Scopus | Cited By in Scopus (32)
 
[15] J.C. Raeder, Propofol anaesthesia versus paracervical blockade with alfentanil and midazolam sedation for outpatient abortion, Acta Anaesthesiol Scand 36 (1992), pp. 31–37. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
 
[16] D.A. Grimes and W. Cates Jr, Complications from legally-induced abortion: a review, Obstet Gynecol Surv 34 (1979), pp. 177–191. View Record in Scopus | Cited By in Scopus (48)
 
[17] H.W. Lawson, A. Frye, H.K. Atrash, J.C. Smith, H.B. Shulman and M. Ramick, Abortion mortality, United States, 1972 through 1987, Am J Obstet Gynecol 171 (1994), pp. 1365–1372. View Record in Scopus | Cited By in Scopus (62)
 
[18] E.S. Lichtenberg, M. Paul and H. Jones, First trimester surgical abortion practices: a survey of National Abortion Federation members, Contraception 64 (2001), pp. 345–352. Article | PDF (129 K) | View Record in Scopus | Cited By in Scopus (30)
 
[19] K. O'Connell, H.E. Jones, M. Simon, V. Saporta, M. Paul and E.S. Lichtenberg, First-trimester surgical abortion practices: a survey of National Abortion Federation members, Contraception 79 (2009), pp. 385–392. Article | PDF (297 K) | View Record in Scopus | Cited By in Scopus (3)
 
[20] M.D. Nichols, G. Halvorson-Boyd, R. Goldstein, C. Gevirtz and D. Healow, Pain management. In: M. Paul, E.S. Lichtenberg, L. Borgatta, D.A. Grimes and P.G. Stubblefield, Editors, Management of unintended and abnormal pregnancy, Wiley-Blackwell, Chichester (2009), pp. 90–110. Full Text via CrossRef
 
[21] M.J. Rawling and E.R. Wiebe, Pain control in abortion clinics, Int J Gynaecol Obstet 60 (1998), pp. 293–295. Article | PDF (43 K) | View Record in Scopus | Cited By in Scopus (12)
 
[22] Higgins J.P.T., Green S., (editors). Cochrane handbook for systematic reviews of interventions version 5.0.2 [updated September 2009]. The Cochrane Collaboration 2009. Available from: www.cochrane-handbook.org.
 
[23] M. Steele, K. Nielsen and S. Klein, Ambulatory anesthesia perioperative analgesia, McGraw-Hill, Berkeley (2005).
 
[24] V. Dahl, F. Fjellanger and J.C. Raeder, No effect of preoperative paracetamol and codeine suppositories for pain after termination of pregnancies in general anaesthesia, Eur J Pain 4 (2000), pp. 211–215. Abstract | PDF (69 K) | View Record in Scopus | Cited By in Scopus (4)
 
[25] A. Edelman, M.D. Nichols, C. Leclair, S. Astley, K. Shy and J.T. Jensen, Intrauterine lidocaine infusion for pain management in first-trimester abortions, Obstet Gynecol 103 (2004), pp. 1267–1272. View Record in Scopus | Cited By in Scopus (10)
 
[26] A. Edelman, M.D. Nichols, C. Leclair and J.T. Jensen, Four percent intrauterine lidocaine infusion for pain management in first-trimester abortions, Obstet Gynecol 107 (2006), pp. 269–275. View Record in Scopus | Cited By in Scopus (7)
 
[27] A.S. Kan, N. Caves, S.Y. Wong, E.H. Ng and P.C. Ho, A double-blind, randomized controlled trial on the use of a 50:50 mixture of nitrous oxide/oxygen in pain relief during suction evacuation for the first trimester pregnancy termination, Hum Reprod 21 (2006), pp. 2606–2611. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (3)
 
[28] A.S. Kan, E.H. Ng and P.C. Ho, The role and comparison of two techniques of paracervical block for pain relief during suction evacuation for first-trimester pregnancy termination, Contraception 70 (2004), pp. 159–163. Article | PDF (64 K) | View Record in Scopus | Cited By in Scopus (9)
 
[29] C.F. Li, C.Y. Wong, C.P. Chan and P.C. Ho, A study of co-treatment of nonsteroidal anti-inflammatory drugs (NSAIDs) with misoprostol for cervical priming before suction termination of first trimester pregnancy, Contraception 67 (2003), pp. 101–105. View Record in Scopus | Cited By in Scopus (14)
 
[30] W. Liu, C.C. Loo, J.W. Chiu, H.M. Tan, H.Z. Ren and Y. Lim, Analgesic efficacy of pre-operative etoricoxib for termination of pregnancy in an ambulatory centre, Singapore Med J 46 (2005), pp. 397–400. View Record in Scopus | Cited By in Scopus (7)
 
[31] K. Suprapto and S. Reed, Naproxen sodium for pain relief in first-trimester abortion, Am J Obstet Gynecol 150 (1984), pp. 1000–1001. View Record in Scopus | Cited By in Scopus (12)
 
[32] G. Hall, A. Ekblom, E. Persson and L. Irestedt, Effects of prostaglandin treatment and paracervical blockade on postoperative pain in patients undergoing first trimester abortion in general anesthesia, Acta Obstet Gynecol Scand 76 (1997), pp. 868–872. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (3)
 
[33] P. Lindholm, H.S. Helbo-Hansen, B. Jensen, K. Bulow and T.G. Nielsen, Effects of fentanyl or alfentanil as supplement to propofol anaesthesia for termination of pregnancy, Acta Anaesthesiol Scand 38 (1994), pp. 545–549. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (8)
 
[34] I. Marc, P. Rainville, R. Verreault, L. Vaillancourt, B. Masse and S. Dodin, The use of hypnosis to improve pain management during voluntary interruption of pregnancy: an open randomized preliminary study, Contraception 75 (2007), pp. 52–58. Article | PDF (208 K) | View Record in Scopus | Cited By in Scopus (8)
 
[35] P.J. Heath and T.W. Ogg, Prophylactic analgesia for daycase termination of pregnancy. A double-blind study with controlled release dihydrocodeine, Anaesthesia 44 (1989), pp. 991–994. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (2)
 
[36] C.Y. Wong, E.H. Ng, S.W. Ngai and P.C. Ho, A randomized, double-blind, placebo-controlled study to investigate the use of conscious sedation in conjunction with paracervical block for reducing pain in termination of first trimester pregnancy by suction evacuation, Hum Reprod 17 (2002), pp. 1222–1225. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (19)
 
[37] M.G. Barneschi, M. Calamandrei, P. Livi, L. Marinelli and V.A. Peduto, Anesthesia for outpatient termination of pregnancy: a comparison of anesthetic techniques, Acta Anaesthesiol Ital 36 (1985), pp. 411–417. View Record in Scopus | Cited By in Scopus (1)
 
[38] J.P. Bonnardot, M. Maillet, M.L. Brule and P. Deligne, Ambulatory anesthesia and induced abortion. Comparative study of propofol–alfentanyl and ketamine–midazolam combinations, Ann Fr Anesth Reanim 6 (1987), pp. 297–300. Abstract | PDF (351 K) | View Record in Scopus | Cited By in Scopus (2)
 
[39] M.E. Bone, S. Dowson and G. Smith, A comparison of nalbuphine with fentanyl for postoperative pain relief following termination of pregnancy under day care anaesthesia, Anaesthesia 43 (1988), pp. 194–197. View Record in Scopus | Cited By in Scopus (5)
 
[40] K.M. Collins, O.M. Plantevin, R.H. Whitburn and J.P. Doyle, Outpatient termination of pregnancy: halothane or alfentanil-supplemented anaesthesia, Br J Anaesth 57 (1985), pp. 1226–1231. View Record in Scopus | Cited By in Scopus (3)
 
[41] G.H. Hackett, M.N. Harris, O.M. Plantevin, H.M. Pringle, D.B. Garrioch and A.J. Avery, Anaesthesia for outpatient termination of pregnancy. A comparison of two anaesthetic techniques, Br J Anaesth 54 (1982), pp. 865–870. View Record in Scopus | Cited By in Scopus (7)
 
[42] N. Phair, J.T. Jensen and M.D. Nichols, Paracervical block and elective abortion: the effect on pain of waiting between injection and procedure, Am J Obstet Gynecol 186 (2002), pp. 1304–1307. Abstract | PDF (47 K) | View Record in Scopus | Cited By in Scopus (13)
 
[43] K. Boysen, R. Sanchez, J.J. Krintel, M. Hansen, P.M. Haar and V. Dyrberg, Induction and recovery characteristics of propofol, thiopental and etomidate, Acta Anaesthesiol Scand 33 (1989), pp. 689–692. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (21)
 
[44] K. Boysen, R. Sanchez, J. Ravn, E. Pedersen, J.J. Krintel and V. Dyrberg, Comparison of induction with and first hour of recovery from brief propofol and methohexital anesthesia, Acta Anaesthesiol Scand 34 (1990), pp. 212–215. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (4)
 
[45] A. Cetin and M. Cetin, Effect of deep injections of local anesthetics and basal dilatation of cervix in management of pain during legal abortions. A randomized, controlled study, Contraception 56 (1997), pp. 85–87. Article | PDF (345 K) | View Record in Scopus | Cited By in Scopus (12)
 
[46] J.E. Hall, W.S. Ng and S. Smith, Blood loss during first trimester termination of pregnancy: comparison of two anaesthetic techniques, Br J Anaesth 78 (1997), pp. 172–174. View Record in Scopus | Cited By in Scopus (7)
 
[47] A. Hein, J. Jakobsson and G. Ryberg, Paracetamol 1 g given rectally at the end of minor gynaecological surgery is not efficacious in reducing postoperative pain, Acta Anaesthesiol Scand 43 (1999), pp. 248–251. View Record in Scopus | Cited By in Scopus (20)
 
[48] A. Hein, C. Norlander, L. Blom and J. Jakobsson, Is pain prophylaxis in minor gynaecological surgery of clinical value? A double-blind placebo controlled study of paracetamol 1 g versus lornoxicam 8 mg given orally, Ambul Surg 9 (2001), pp. 91–94. Article | PDF (121 K) | View Record in Scopus | Cited By in Scopus (17)
 
[49] J. Jakobsson, S. Davidson, M. Andreen and M. Westgreen, Opioid supplementation to propofol anaesthesia for outpatient abortion: a comparison between alfentanil, fentanyl and placebo, Acta Anaesthesiol Scand 35 (1991), pp. 767–770. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (22)
 
[50] J. Jakobsson, E. Oddby and K. Rane, Patient evaluation of four different combinations of intravenous anaesthetics for short outpatient procedures, Anaesthesia 48 (1993), pp. 1005–1007. View Record in Scopus | Cited By in Scopus (15)
 
[51] J. Jakobsson and K. Rane, Anaesthesia for short outpatient procedures. A comparison between thiopentone and propofol in combination with fentanyl or alfentanil, Acta Anaesthesiol Scand 39 (1995), pp. 503–507. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (12)
 
[52] J. Jakobsson, K. Rane and S. Davidson, Intramuscular NSAIDS reduce post-operative pain after minor outpatient anaesthesia, Eur J Anaesthesiol 13 (1996), pp. 67–71. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (15)
 
[53] H.W. Li, C.Y. Wong, S.S. Lo and S.Y. Fan, Effect of local lignocaine gel application for pain relief during suction termination of first-trimester pregnancy: a randomized controlled trial, Hum Reprod 21 (2006), pp. 1461–1466. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (4)
 
[54] T.W. Ogg, R.A. Jennings and C.G. Morrison, Day-case anaesthesia for termination of pregnancy. Evaluation of a total intravenous anaesthetic technique, Anaesthesia 38 (1983), pp. 1042–1046. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (3)
 
[55] A.E. Rossi, D. Lo Sapio, O. Oliva, O. Vitale and A. Ebano, Hospital day-surgery: comparative evaluation of 3 general anesthesia techniques, Minerva Anestesiol 61 (1995), pp. 265–269. View Record in Scopus | Cited By in Scopus (1)
 
[56] A.G. Shapiro and H. Cohen, Auxiliary pain relief during suction curettage, Contraception 11 (1975), pp. 25–30. Article | PDF (207 K) | View Record in Scopus | Cited By in Scopus (2)
 
[57] N. Wells, Management of pain during abortion, J Adv Nurs 14 (1989), pp. 56–62. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (2)
 
[58] N. Wells, Reducing distress during abortion: a test of sensory information, J Adv Nurs 17 (1992), pp. 1050–1056. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (11)
 
[59] E. Wiebe, L. Podhradsky and V. Dijak, The effect of lorazepam on pain and anxiety in abortion, Contraception 67 (2003), pp. 219–221. Article | PDF (42 K) | View Record in Scopus | Cited By in Scopus (11)
 
[60] E.R. Wiebe, Comparison of the efficacy of different local anesthetics and techniques of local anesthesia in therapeutic abortions, Am J Obstet Gynecol 167 (1992), pp. 131–134. View Record in Scopus | Cited By in Scopus (34)
 
[61] E.R. Wiebe and M. Rawling, Pain control in abortion, Int J Gynaecol Obstet 50 (1995), pp. 41–46. Article | PDF (489 K) | View Record in Scopus | Cited By in Scopus (30)
 
[62] E.R. Wiebe, M. Rawling and P. Janssen, Comparison of 0.5% and 1.0% lidocaine for abortions, Int J Gynaecol Obstet 55 (1996), pp. 71–72. Article | PDF (121 K) | View Record in Scopus | Cited By in Scopus (3)
 
[63] Royal College of Obststrics and Gynecology, The care of women requesting induced abortion. Evidence-based clinical guideline number 7. London: RCOG Press Available from: http://www.rcog.org.uk/files/rcog-corp/uploaded-files/NEBInducedAbortion... (2004).
 
[64] D.A. Grimes, K.F. Schulz, W. Cates Jr. and C.W. Tyler Jr, Local versus general anesthesia: which is safer for performing suction curettage abortions?, Am J Obstet Gynecol 135 (1979), pp. 1030–1035. View Record in Scopus | Cited By in Scopus (18)
 
[65] F.W. Cheney, The American Society of Anesthesiologists Closed Claims Project: what have we learned, how has it affected practice, and how will it affect practice in the future?, Anesthesiology 91 (1999), pp. 552–556. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (129)
 
[66] R.M. Renner, J.T. Jensen, M.D. Nichols and A. Edelman, Pain control in first trimester surgical abortion, Cochrane Database Syst Rev (2009), p. CD006712. View Record in Scopus | Cited By in Scopus (0)