Intravenous Ketorolac in the Emergency Department Management of Sickle Cell Pain and Predictors of Its Effectiveness

Objectives  To evaluate the effectiveness of intravenous (IV) ketorolac tromethamine in the treatment of children with sickle cell disease with moderate to severe acute vaso-occlusive pain (VOP) and to develop a predictive model that would determine who would need additional IV analgesics.

Design  A prospective case series.

Setting  The emergency department of an urban children's hospital in the southeastern United States.

Patients  A convenience sample of 51 children aged 6 to 18 years, representing 70 distinct episodes of VOP requiring IV analgesics.

Intervention  All patients were given 0.5 to 1 mg/kg IV ketorolac and IV fluids.

Main Outcome Measures  Patients, parents, nurses, and physicians assessed pain before and after ketorolac using a standard 100-mm visual analog scale (VAS).

Results  Of the 70 episodes of VOP, 37 (53%) adequately resolved with IV ketorolac and IV fluids and required no IV opioids (group A). Thirty-one episodes (47%) required the addition of an IV opioid (group B). Group B had a significantly greater proportion of episodes reporting 4 or more painful sites than group A, 43% (12/28) vs 9% (3/33), respectively (P<.01). Group B also had significantly higher mean initial VAS scores than group A as assessed by the patient (81 vs 60; P<.01), parent (71 vs 54; P<.01), nurse (78 vs 51, P<.01), and physician (69 vs 53; P = .01). Of the patient assessments with an initial VAS score greater than 70, 69% (18/26) required the addition of an opioid.

Conclusions  First-line therapy with IV ketorolac and IV fluids resulted in adequate resolution of pain in 53% of episodes with acute VOP. A reported 4 or more painful sites and an initial VAS score greater than 70 were predictors of the likelihood to need additional IV analgesics.

IN THE United States, more than 50 000 people are affected by sickle cell disease, and 1500 babies with sickle hemoglobinopathies are born each year.¹ Vaso-occlusive pain (VOP) is the most common reason for emergency department (ED) visits and hospital admissions.¹ It can be excruciating and often incapacitating, resulting in frequent absences from school and work.¹ There have been few advances in the treatment of VOP in the past 20 years.² Current therapy for moderate to severe pain consists of various opioid and nonopioid analgesics given via different routes and on a variety of schedules.^1-4 In a survey by Pegelow⁴ in 1992, 21 physicians used 5 different opioids to manage hospitalized patients with VOP.

Our ED's standard of care for VOP is rest, intravenous (IV) fluids, and analgesics. When pain is moderate to severe, IV fluids and IV ketorolac tromethamine and/or IV nalbuphine or morphine are used. Ketorolac is generally offered first to patients with moderate to severe pain because it is a potent nonnarcotic analgesic with few adverse effects and has been shown to have a narcotic-sparing effect.^5,6 Ketorolac has been proven to be safe and effective in children.⁷ It is as effective as morphine for acute pain states in children, including postoperative pain.^5,8,9

To our knowledge, no studies in children have previously evaluated the effectiveness of IV ketorolac as a first-line therapy in VOP. A prospective study by Hardwick et al¹⁰ was unable to demonstrate a significant synergistic analgesic effect for ketorolac compared with placebo when used with morphine in the ED management of children with VOP. This study compared morphine requirements and pain reduction during a 6-hour observation period between patients who received a simultaneous initial dose of ketorolac and morphine (n = 22) or placebo and morphine (n = 19). They report no synergistic effect of ketorolac. However, preliminary reports of this study (with the same number of patients) suggested a beneficial effect of ketorolac.¹¹ According to the authors, the study was limited by small numbers and a protocol that may have restricted their ability to determine true morphine requirements and therefore, the synergistic effectiveness of ketorolac.¹⁰

Our objective was to evaluate the efficacy of IV ketorolac as a first-line therapy in the treatment of children with sickle cell disease coming to the ED with moderate to severe episodes of acute VOP. We also attempted to identify predictors that would determine which patients were not likely to respond to ketorolac, requiring the addition of an IV opioid.

This study was approved by the human investigations committee at Emory University (Atlanta, Ga). This was a prospective convenience case series in the ED at Hughes Spalding Children's Hospital in urban Atlanta from August 14, 1997, through November 13, 1998.

Eligible children were aged 6 to 18 years with VOP secondary to a sickle hemoglobinopathy (HbSS, HbSC, or HbSBthal). Children were ineligible to participate if they had atypical pain determined by the patient, caregiver, nurse, and physician, had allergies to nonsteroidal anti-inflammatory agents, or had impaired vision or neurological function. Children younger than 6 years were excluded because of their general inability to appropriately use a visual analog scale (VAS) and self-report their pain.¹² Fever was not an excluding variable. All patients with moderate to severe VOP received IV fluids and IV analgesics as the standard of care.

On entering the ED, patients and caregivers meeting study criteria were recruited for the study and those who agreed completed an informed consent form and a questionnaire addressing demographics, history of present illness, medical history, and prior use of analgesics. For those consenting to participate, questions included but were not limited to number of admissions for pain in the past, current number of painful sites (eg, 1 arm, back, and 2 legs), and timing and doses of oral analgesics taken at home. Children could enter the study more than once. Each time a patient entered, he or she completed the questionnaire.

Pain was assessed on arrival to the ED independently by the patient, parent, nurse, and physician. Each person independently scored a standardized, validated 100-mm VAS, which was a horizontal line bounded by the phrases "no pain/no hurt" on the left and "worst pain ever/ worst hurt ever" on the right.^12,13 The scales were preprinted on paper and available to the patient, parent, nurse, and physician in separate packets with identical instructions. Each person marked a vertical line through his or her VAS at the point he or she believed equaled the pain experienced by the patient. Pain was assessed before and after ketorolac treatment.

In addition, the patients each scored a standardized Nine Faces Pain Scale (NFPS) before and after ketorolac treatment.¹² This was a series of 9 faces ranging from very happy to very sad preprinted on the patients' pain assessment packets along with the VAS. Patients circled the face that represented how they "felt deep down inside" as instructed in the packet.

All patients initially received standardized care with a complete blood cell count and reticulocyte count drawn, an IV line established, and IV ketorolac tromethamine infused at 0.5 to 1.0 mg/kg (60 mg maximum). Patients then received 10 to 20 mL/kg of an isotonic sodium chloride solution bolus for 60 minutes, followed by an option to run 1.5× maintenance fluids of 5% dextrose in one-half isotonic sodium chloride solution.

Pain was reassessed 45 to 60 minutes after the ketorolac infusion using a second VAS independently scored by the patient, parent, nurse, and physician. In addition, the patient marked the second NFPS. If the patient had adequate relief of pain (defined as an agreement between the patient and caregiver that the pain was tolerable or completely resolved and no further IV analgesia was needed), then the patient was discharged to his or her home with a direction to take oral ibuprofen and acetaminophen with codeine as needed with appropriate follow-up instructions. Physicians had the option to observe children longer prior to discharge. Patients without adequate relief of pain received an IV opioid. Return to the ED within 48 hours by discharged patients was recorded by telephone contact or ED records.

Data were stored and analyzed using the Epi Info version 6 software program (Centers for Disease Control and Prevention, Atlanta, Ga). Statistical analysis was performed using the Wilcoxon rank sum test, t test, exact mid-P test, and Fisher exact test to compare means. Mean VAS and NFPS scores were very close to median scores; therefore only means are reported. Each time a child entered the study, his or her data were stored and analyzed.

A total of 51 patients with 70 distinct episodes of VOP were enrolled. Thirty-seven (53%) of 70 episodes adequately resolved with ketorolac and hydration represented group A. Thirty-three (47%) of 70 episodes did not resolve and required the addition of an opioid, represented group B.

There were no significant differences between the groups when comparing demographics including age, weight, and sex (Table 1). There were also no differences in the number of past ED visits or admissions, time since last ED visit, duration of pain, type of oral analgesic taken at home, timing of the oral analgesic in relation to ED visits, or in the appropriateness of the dose (Table 1). In addition, the responses to the following questions were similar: "At what pain level does your child need IV pain medicine?" and "Does your child need IV pain medicine now?"

One variable that was predictive of the likelihood not to improve with ketorolac therapy was the number of painful sites. Group B had a significantly higher proportion of episodes, with 4 or more painful sites than group A (43% vs 9%, respectively; P<.01) (Table 1). In all, 12 (80%) of 15 episodes had 4 or more painful sites did not improve adequately with ketorolac and IV fluids.

Another variable that was predictive of the likelihood not to improve with ketorolac was the initial VAS score. Group B had significantly higher mean (± SD) initial VAS scores than group A as assessed by the patient (81 ± 17, 60 ± 24; respectively P<.01); parent (71 ± 23 vs 54 ± 23; P<.01); nurse (78 ± 17 vs 51 ± 25; P<.01); and physician (69 ± 25 vs 53 ± 22; P = .01) (Table 1).

An initial VAS score greater than 70 (of a possible 100) was a predictor for who would not have adequate resolution of pain with ketorolac. Group B had a significantly greater proportion of episodes with a VAS score greater than 70 compared with group A as assessed by the patient (78% vs 28%, respectively; P<.01); parent (50% vs 24%; P = .04); nurse (71% vs 25%; P<.01); and physician (57% vs 26%; P = .02). In fact, 69% (18/26) of all episodes had a VAS score greater than 70 as assessed by the patient required the addition of an IV opioid.

The NFPS scores were similar to the VAS scores. Group B had significantly higher mean (± SD) initial scores than group A (7.5 ± 1.3 vs 6.0 ± 1.7, respectively; P<.01). A predictable cutoff for inadequate resolution of pain was an initial NFPS score greater than or equal to 8. A greater proportion of episodes in group B had an initial NFPS greater than or equal to 8 (group B [50%] vs group A [17%]; P = .01). Twelve (71%) of 17 episodes with a score greater than or equal to 8 on the NFPS did not improve.

Group A as assessed by the patient, parent, nurse, and physician, significantly improved from the first to the second VAS, which correlated with the clinical improvement seen (Table 2). For group B there was a significant difference from the first VAS to the second as assessed by the nurse and physician but not the patient and parent, although the trend was toward improvement (Table 2). Although group B tended to improve clinically, the patients all required additional IV analgesics because of persistent pain evidenced by VAS scores that remained significantly elevated.

In addition, both groups demonstrated significant improvement from the first to the second NFPS (group A, 6.0 vs 3.3; P<.001 and group B, 7.5 vs 6.0; P = .005). Although both groups improved, group A experienced more relief and did not require additional IV analgesics, which was consistent with the VAS scores.

The amount of pain relief (defined as the difference in pain scores before and after ketorolac on the VAS or NFPS) was also measured. Ketorolac had a positive effect for both groups but less for group B. The mean change in VAS scores was significantly greater for group A than for group B as assessed by the patient (50 vs 13, respectively; P<.001); parent (34 vs 11; P = .005); nurse (42 vs 20; P = .004); and physician (44 vs 22; P<.001). The NFPS scores for pain change (the first NFPS minus the second) are similar to the VAS results (group A, 3.1 vs group B, 1.8; P = .04).

Mean pain scores after receiving ketorolac reveal that group A had significantly lower scores than group B as assessed by the patient (13 vs 69, respectively; P<.001); parent (19 vs 61; P<.001); nurse (10 vs 59; P<.001); and physician (11 vs 52; P<.001). Similarly, the NFPS scores after ketorolac treatment were also significantly lower in group A (3.3 vs 6.0; P<.001).

Of the 51 patients representing 70 episodes of VOP, 40 were enrolled only once. The 11 other patients represented 30 episodes of VOP. Of the 11 returning enrollees, 2 never improved with ketorolac and IV fluids (7 episodes), 3 always improved (7 episodes), and the other 6 patients improved in 8 of 16 episodes.

Of 37 episodes with adequate pain relief (group A), the patients in 5 of these instances were admitted for reasons other than pain, 4 for fever and 1 for hypoxia. Of the 32 episodes in which patients were discharged to their homes, 4 required ED treatment within 48 hours for pain. Two of these 4 patients were subsequently admitted for persistent pain.

Data were collected on parental perceptions of their child's pain. The question, "Do you think your child needs an IV pain medicine today?" was answered with 1 of the following: "definitely yes," "probably yes," "I am not sure," "probably not," and "definitely not." This question was answered for 42 of the 71 episodes of VOP. Of the 42 responses, 30 (71%) were "definitely yes," and 8 (19%) were "probably yes."

One child complained of nausea and vomited. There were no other adverse events reported.

In sickle cell disease, VOP is unique and often debilitating. It is the most common reason for ED visits and hospital admissions in people with sickle cell disease.¹ Despite the severity of this pain and its prevalence, there have been few advances in the treatment of VOP in the past 20 years.²

Ketorolac is a nonsteroidal anti-inflammatory drug comparable to morphine in many acute pain states.^5,8,9 It does not have the adverse effects of central nervous system and respiratory depression, hypotension, decreased gastrointestinal tract motility, and pruritus as does morphine. In addition, several studies show that ketorolac plus an opioid results in a 25% to 50% opioid-sparing effect.^5,6 Ketorolac's safety and efficacy have been well proven in prospective studies in children and adults in painful and postoperative settings.^5-9 However, there are limited published studies looking at the effectiveness of ketorolac in VOP in children with sickle cell disease.

Two prospective studies in adults evaluated the use of ketorolac in VOP. Perlin et al¹⁴ demonstrated that continuous infusion of ketorolac significantly reduced the amount of meperidine used by 33% in 9 patients compared with 11 placebo controls in hospitalized patients with VOP. Wright et al¹⁵ demonstrated no significant reduction in the amount of meperidine used in a group of 12 patients who received ketorolac vs 12 controls who received a placebo.

The VAS was used in an attempt to get an objective measurement of pain. The VAS has been shown to be valid and reliable when used in children older than 6 years.^12,13 Its simplicity facilitates use and compliance with the study. The NFPS was an additional measurement of pain because it correlates well with mood, which is a significant component of a child's experience of pain.¹³

We demonstrated that first-line therapy with IV ketorolac and IV fluids resulted in adequate resolution of pain in 37 (53%) of 70 VOP episodes. In these episodes the pain relief was adequate enough to discharge the patients to their homes. Of 32 patients discharged, 4 returned within 48 hours and 2 of the 4 were subsequently admitted. Predictors of the likelihood to have adequate resolution of pain with ketorolac and IV fluids were level of pain less than or equal to 70 on the initial VAS as assessed by the patient, parent, nurse or physician; the patient's initial NFPS score of less than 6; and having less than 4 painful sites at the initial assessment. This information predicted well which patients would likely need only ketorolac and IV fluids, thus sparing the patient an IV opioid and its potential risks.

We were also able to predict that patients with 4 or more painful sites and an initial VAS score greater than 70 or an initial NFPS score greater than 7 were much less likely to achieve adequate pain relief with ketorolac alone. In these patients, there was no reason to delay the administration of opioid analgesics.

It is our opinion that ketorolac should be initially administered to patients with pain scores less than or equal to 70 on a 100-mm VAS or comparable scale. Patients with higher pain scores should be given an IV opioid simultaneously with ketorolac because ketorolac, while it provides some relief, is not likely going to reduce the pain to a tolerable level. The number of painful sites is an additional factor to consider. For example, a VAS score of 60 to 70 in a patient complaining of 4 or more painful sites should warrant the addition of an opioid.

Stratifying pain management based on pain scores and number of painful sites makes it easier to decide on treatment plans and makes the treatment more uniform from one visit to the next and between different health care providers. It also leads to more efficient and judicious use of analgesics to deliver adequate and timely pain relief.

Further studies are needed with larger sample sizes to determine the best management of pain in sickle cell disease. Children not improving with ketorolac therapy in this study went on to receive an IV opioid as part of a continuing evaluation of the optimal ED management of children with VOP.

One limitation of the study was that it was not a randomized, controlled trial. We did not compare ketorolac with placebo because it has been well established in other studies that ketorolac is effective in many acute pain states.^5-9 We felt that it would be unethical to withhold effective and expected analgesics in patients with acute pain.

Another potential flaw was to allow patients to enter the study multiple times during repeated ED visits for VOP. Many patients in this study had previous exposure to ketorolac and, thus, may have had preconceived biases about the effectiveness of the drug. In 18 (35%) of 52 episodes, patients responded "ketorolac" or "ketorolac plus nalbuphine or morphine" when asked what IV pain medicine(s) worked best for them for VOP. Most of the other answers were "unsure." Including repeated episodes was felt to be more representative of the actual population experience. Eliminating these patients would have significantly reduced our numbers. Of the 19 episodes that were made up of repeated visits by 11 children, 10 (53%) improved with ketorolac therapy, a percentage identical to that of the children with first-time visits only who improved (27 [53%] of 51 episodes). Returning visitors were likely to improve if they had lower pain scores and a lower number of painful sites, just as the patients who were seen only once.

Although we were looking at the effectiveness of ketorolac, we gave all patients IV fluids. Ideally we would have eliminated this potentially confounding factor. There are several reasons for giving IV fluids: it is the standard of care at our institution, it is a baseline treatment for VOP, it can protect the kidneys during nonsteroidal anti-inflammatory drug use, and it is expected by patients and staff. We do not know how many patients would have improved with IV fluids alone. However, ketorolac likely played a significant role in their improvement and it was believed to be unethical to withhold analgesic therapy in this patient population.

We were not able to capture every 6- to 18-year-old patient coming to the ED with VOP. Patients were missed for several reasons, mostly because the ED was too busy or the staff forgot to enroll the patient, which leads to potential selection bias.

In addition, not all questions were answered in every questionnaire, nor all pain scales completed by every patient, parent, nurse, and physician for every episode. Ideally, we would not have missed a single data point. We believe, however, that we have collected a sufficient amount of information to produce significant and valid results.

In conclusion, first-line therapy with IV fluids and IV ketorolac is an effective way to manage acute VOP in children. In this study, patients in 37 (53%) of 70 VOP episodes had adequate relief from pain with IV fluids and IV ketorolac, required no more IV analgesics, and were able to be discharged to their homes. We recommend, based on our findings, that patients with an initial VAS score greater than 70 as assessed by the patient, parent, nurse, or physician, or an initial NFPS score greater than 7 and VOP in 4 or more sites are not likely to improve adequately with treatment with ketorolac and IV fluids alone. These patients are more likely to need the addition of an opioid.

Accepted for publication December 14, 2000.

Presented in part at the Southeastern Society of Academic Emergency Medicine Annual Meeting, Atlanta, Ga, March 14, 1999; the Pediatric Academic Societies' Annual Meeting, San Francisco, Calif, May 3, 1999; and the Society of Academic Emergency Medicine Annual Meeting, Boston, Mass, May 20, 1999

We thank the ED staff, including the nurses, attending physicians, and resident physicians, for their contributions. We especially thank the children and their families for their participation in this study.

Corresponding author: James L. Beiter, MD, 5643 Glenrich Dr, Atlanta, GA 30388 (e-mail: jbeiterjr@aol.com).