This prospective randomized double-blind, cross-sectional study was performed during a 19-month period (June 2006 to January 2008) in the ED of a tertiary referral center. The ED has a residency program with 120,000 consultations a year. The study was approved by the institution’s teaching and research committee.
For calculation of the sample size, a difference of 2 ± 0.8 points in the VAS scale is considered clinically significant. To detect differences in the rate of adverse effects between groups, considering a 30% incidence in group A vs. a possible 10% incidence in group B with the same error levels, 62 patients were required without continuity correction, and 72 patients were required for Fisher's exact test.
As soon as the diagnosis of hip fracture was made and before an analgesic was administered, the patient or their legal decision-maker was informed about the study. If they agreed to participate, they signed an informed consent form.
Patients were randomized into two groups (A and B) using numbers generated by the EPI-INFO™ (Atlanta, GA: Centers for Disease Control and Prevention) program. The randomization list was kept by one of the authors who did not interact with the patients. He gave instructions to the patient’s ED nurse about which treatment should be administered. The nurse prepared the medication following the physician’s instructions and assigned a letter to the protocol (from a set of 10 letters, 5 for group A and 5 for group B) that designated whether the patient was receiving active medications in the fascia-iliaca block. The physician administering the medications and obtaining the VAS scores did not know which medications the patient was receiving.
Selection of participants
Our patient pool consisted of adult patients more than 65 years old who presented to the ED because of a previously undiagnosed and untreated hip fracture.
Exclusion criteria for the study included anatomical abnormalities in the inguinal area different from fracture, known coagulation disorders, a history of allergy to any of the active ingredients used during the study and refusal to participate.
The protocol included observing patients for 8 h. Upon admission to the study, vital signs such as blood pressure, mean blood pressure (MAP), heart rate (HR), respiratory rate (RR) and pain-intensity measurements [using the Visual Analogue Scale (VAS)] were obtained. Vital signs and VAS-based pain measurement evaluations were performed at 15 min, 2 h and 8 h. (Vital signs were also taken at intermediate times for clinical purposes.) The occurrence of adverse effects or complications was concomitantly evaluated in all patients.
A peripheral IV was placed in all patients for the administration of analgesia and hydration. A single fascia-iliaca compartment block [9
] was performed on all patients with either 0.9 normal saline (group A) or 0.3 ml/kg 0.25% bupivacaine (group B). Patients in group A received IV NSAID analgesics (Diclofenac or Ketorolac, depending on their medical history), whereas those in group B had 3 ml to 5 ml 5% dextrose administered IV.
A fascia-iliaca block is performed by identifying the inguinal ligament and the femoral artery. A 21-g, 2-inch intramuscular injection needle is inserted perpendicular to the skin at a point 1 cm below the juncture of the lateral and medial two-thirds of a line that joins the pubic tubercle to the anterior superior iliac spine. The needle is inserted until a loss of resistance is felt as the fascia lata is passed, and then further advanced until a second loss of resistance occurs when the fascia iliaca is pierced (often described as two “pops”) [9
]. After aspirating to assure that the needle is not in a vessel, solution is injected.
The primary measurement used was VAS to determine pain scores in the two groups. The patients’ vital signs over the course of the study period were used as secondary measurements.
Primary data analysis
The general characteristics of the entire study sample and of each subgroup were calculated for mean age, sex, co-morbidities, vital signs and VAS. Quantitative data were reported as mean ± standard error regardless of the distribution of the variables. Variables such as sex, prevalence of co-morbidities and side of fracture were dichotomized into present or absent.
Baseline VAS scale values and vital sign data were compared to data obtained at 15 min, 2 h and 8 h in an independent manner. The Student’s t-test for paired samples was used for these continuous quantitative values. When the same values were compared between groups A and B, the Student’s t-test was used for each comparison at 15 min, 2 and 8 h, and an ANOVA curve was used for sequential comparison. Complications such as vomiting, nausea and hematoma at the injection site were dichotomized as present or absent, and were compared at 8 h using Fisher’s exact test. For statistical inference, a type I error of 0.05 and a power of 80% was assumed.
The statistical analysis was performed using the Primer of Biostatistics 4.0 software (McGraw Hill, 1996) and the Statistix 7.0 analytical software (2000).