Abstract

Background and Objective

Diagnostic injections are used to diagnose myriad pain conditions, but are characterized by a high false-positive rate. One potential cause of inaccurate diagnostic blocks is the use of sedation. We sought to determine the effect of sedation on the validity of diagnostic injections.

Design

Randomized, crossover study in which 73 patients were allocated to receive a diagnostic sacroiliac joint or sympathetic nerve block performed either with or without sedation using midazolam and fentanyl. Those who obtained equivocal relief, good relief lasting less than 3 months, or who were otherwise deemed good candidates for a repeat injection, received a subsequent crossover injection within 3 months (N = 46).

Setting and Patients

A tertiary care teaching hospital and a military treatment facility.

Results

In the primary crossover analysis, blocks performed with sedation resulted in a larger mean reduction in pain diary score than those done without sedation (1.2 [2.6]; P = 0.006), less procedure-related pain (difference in means 2.3 [2.5]; P < 0.0001), and a higher proportion of patients who obtained > 50% pain relief on their pain diaries (70% vs. 54%; P = 0.039). The increased pain reduction was not accompanied by increased satisfaction (sedation mean 3.9 [1.1] vs. 3.7 [1.3]; P = 0.26). Similar findings were observed for the parallel group (N = 73) and omnibus (all sedation vs. no sedation blocks, N = 110) analyses. No differences in outcomes were noted between the use and non-use of sedation at 1-month.

Conclusions

The use of sedation during diagnostic injections may increase the rate of false-positive blocks and lead to misdiagnoses and unnecessary procedures, but has no effect on satisfaction or outcomes at 1-month.

Introduction

The use of injections to treat chronic pain has more than doubled over the past decade [1]. Although the ostensible purpose of injections is to alleviate pain, systematic reviews have concluded that the benefits tend to be transient [2–4]. This has led to an increased emphasis on the diagnostic value of injections, particularly nerve blocks [5–9].

Diagnostic spinal injections are used to identify sources of nociception such as painful joints, discs, and nerve roots [6]. In addition to spinal blocks, diagnostic injections play an integral role in the identification of myriad other chronic pain conditions, such as osteoarthritis, occipital neuralgia, piriformis syndrome, nerve entrapment syndromes, and complex regional pain syndrome [10–16].

The use of injections for diagnostic purposes traces its roots back to the 1920s when von Gaza [17] employed nerve blocks to elucidate ambiguous pain generators. Yet despite their conceptual appeal, nerve blocks are fraught with potential inaccuracies [18], particularly false positive results [19–24]. Many experts have therefore advocated the use of “double blocks” to reduce the false positive rate [19,23,25,26]; however, performing multiple blocks to enhance specificity may decrease sensitivity, which can result in withholding effective treatments from patients [20,27,28].

False positive blocks occur for a variety of reasons including placebo response, expectation bias, inadvertent spread of the injectate to adjacent pain-generating structures, overzealous use of superficial anesthesia, systemic absorption of local anesthetic, and failure to adequately interpret the results [6,20,29–35]. Yet perhaps the most modifiable potential cause is the use of sedation, which has generated substantial interest in the literature [36–39]. Proponents of sedation contend that anxiolysis makes for a more palatable experience and may therefore reduce the risk for a false negative block. However, opponents assert that the use of sedation can increase the likelihood of a false positive block by alleviating anxiety; reducing index pain; interfering with a patient's ability to perform daily activities; and in the case of a benzodiazepine, promoting amnesia and relieving muscle-related pain [40].

To date, only a series of studies published by the same investigators have examined the effects of sedation on the diagnostic validity of blocks [30,41–43]. However, these studies suffer from methodological flaws including: 1) patient sampling bias (i.e., patients who receive long-standing relief with local anesthetic blocks may be placebo responders); 2) assessing pain relief before, rather than after, the injections; 3) failure to utilize a crossover design; 4) failure to coadminister a benzodiazepine and opioid, which is generally employed in clinical practice; and 5) failure to notify patients whether or not they received a medication, which does not reflect clinical circumstances. Therefore, further research is needed to understand the role sedation has on the diagnostic validity of blocks.

Sacroiliac joint and sympathetic blocks represent ideal procedures to assess the influence sedation may have on diagnostic validity as these injections frequently need to be repeated and serve as the reference standard for the diagnosis of sacroiliac joint pain and sympathetically maintained pain, respectively [5,11,23]. In order to determine whether or not sedation influences the diagnostic value of injections, we performed a multicenter, randomized, crossover study comparing postprocedure pain scores of patients who received diagnostic sacroiliac joint or sympathetic nerve blocks done with and without sedation. Our hypothesis was that the use of sedation would increase the rate of positive blocks.

Methods

Permission to conduct this study was granted by the internal review boards of the Johns Hopkins Medical Institutions and Walter Reed National Military Medical Center, as well as all patients who provided informed, written consent. All procedures and follow-up visits were performed between March 2011 and February 2013.

Study Participants

Inclusion criteria were suspected sacroiliac joint pain (i.e., tenderness overlying the sacroiliac joint and positive provocative tests) [8,9] or complex regional pain syndrome diagnosed based on history and physical examination [44,45]; indications for sacroiliac joint, lumbar sympathetic, or stellate ganglion blocks (SGBs) for diagnostic (and possibly therapeutic) purposes; pain for >1 month and <10 years duration; and >3/10 pain on a 0–10 numerical rating scale (NRS). Excluded from participation were patients who had undergone nerve blocks for the same condition within the past 3 years; allergy to amide local anesthetics, contrast dye, midazolam, and or fentanyl; age <18 or >80 years; and/or severe anxiety requiring sedation.

Randomization and Crossover

A randomized, crossover study was performed in which patients scheduled to undergo sacroiliac or sympathetic blocks were allocated in a 1:1 ratio to receive either sedation (group 1) or no sedation (group 2) for their initial block. Randomization was performed in groups of eight by a research nurse via computer-generated randomization tables, stratified by injection type and study site. Participant screening and study enrollment were performed by an investigator physician. In those individuals who subsequently underwent a repeat injection, the second block was done using sedation in participants who were allocated to the “no sedation” group, or without sedation in those allocated to the “sedation” group. Although subjects who enrolled in the study were given the opportunity to not crossover, none elected to do so. Our study size was determined by a preprocedure pilot audit that suggested slightly more than half of the patients would undergo a second block. Indications for a second block were significant pain relief or functional improvement that returned to baseline in less than 3 months, an ambiguous response to the initial injection, and patient or physician preference for a second block. Only those who obtained persistent (≥3 months) pain relief obviating the need for a second injection in a timely manner or who experienced an unequivocally negative injection were categorically excluded from a repeat injection. In an attempt to standardize the two injections in any given participant, the same attending physician performed or supervised both procedures, with the needle position for the second injection approximating that for the first whenever possible. In view of the role that expectations, anxiety, and the placebo effect play for chronic pain, a decision was made to perform the study in an unblinded fashion to best simulate clinical practice [46–48].

Sacroiliac Injections

All injections were performed using fluoroscopic guidance and contrast injection to confirm accurate needle placement. For sacroiliac joint injections, 22-gauge spinal needles were inserted into the postero-inferior one-third of the joint using either an antero-posterior, contralateral oblique (5°–25°), or caudad-cephalad approach, as per previous published standards [49]. Correct position was ascertained via the injection of radiopaque contrast. Following confirmation of joint penetration, a 3.5 mL solution containing 2 mL of bupivacaine 0.5% and 1.5 mL of 40 mg/mL of depo-methylprednisolone (Pharmacia and Upjohn, Kalamazoo, MI, USA) was administered. Steroids were added to reflect clinical circumstances, as most clinical studies have used steroids and local anesthetic for diagnostic blocks [49,50]. In circumstances when bilateral injections were performed, the solution administered into each joint consisted of 2.5 mL of 0.5% bupivacaine and 1 mL of 40 mg/mL steroid.

Sympathetic Blocks

Sympathetic blocks were performed under fluoroscopic guidance with or without ultrasound in accordance with previously published descriptions [5]. SGBs were performed for suspected upper extremity CRPS and lumbar sympathetic blocks (LSBs) for lower extremity symptoms. SGB was accomplished by inserting a blunt-tip 22-gauge needle onto the medial aspect of the ipsilateral C7 transverse process then slightly withdrawing it 2–4 mm. In rare cases (<5%) in which intravascular or epidural uptake could not be avoided, C6 was used as the target level. Once placement was deemed adequate, contrast dye was injected to ensure correct placement, after which 8 mL of 0.25% bupivacaine was administered.

LSBs were performed by inserting a 22-gauge 7-inch spinal needle in an oblique, coaxial trajectory to a target point located approximately 5 mm proximal to the antero-lateral margin of the L3 vertebral body. Once needle position was deemed sufficient using antero-posterior and lateral imaging, contrast was injected to confirm adequate spread to the relevant spinal levels (L2–4 for LSB) [51,52] and the absence of intravascular uptake. After the attending physician was satisfied with the contrast spread, 16–20 mL of 0.25% bupivacaine was incrementally injected, depending on the degree of spread. In less than 5% of patients, a second needle was inserted at either L2 or L4 to ensure coverage of all relevant levels. A successful block was designated as a ≥2°C increase in temperature in the affected extremity, which occurred in all cases [53,54].

Sedation

Sedation and analgesia were accomplished by the use of midazolam and fentanyl. In order to maximize generalizability and account for widespread variability in clinical circumstances, medical practice, and patient response, no minimum or maximum amount of medication was mandated. All medications were titrated to effect by a board-certified anesthesiologist.

Data Collection and Outcome Measures

Baseline demographic and clinical data that were collected included 0–10 NRS pain scores, functional capacity in the form of Oswestry disability index 2.0 (ODI) scores for sacroiliac (SI) joint blocks, opioid consumption, coexisting psychiatric morbidity, etiology, and the amount of sedation. The ODI is a 10-question validated survey scored from 0 to 50 points and converted to a percentage with higher scores representing greater disability, which is used to measure present functional capacity in patients with chronic back and/or leg pain [55]. The primary outcome measure was a 6-hour written pain diary in which index pain scores and an activity log were recorded every 30 minutes. Prior to the procedure and before discharge, participants were instructed to engage in their normal activities and discount procedure-related pain. The percent pain relief was calculated considering pain scores and the activity log as the arithmetic mean minus any outliers (e.g., “0 when sleeping”) by an investigator physician and research nurse. In five cases in which diaries were not returned (three in blocks with sedation), pain scores were estimated by the first two pain scores recorded before discharge and subsequent verbal accounts of the percent relief.

Secondary outcome measures included a 5-point Likert scale reflecting satisfaction with pain relief experienced over the 6-hour postblock period (ranging from very dissatisfied to very satisfied), procedure-related pain recorded on a 0–10 verbal pain scale, and adverse effects. The follow-up visit after each block was performed at 1 month, at which time 0–10 NRS pain scores, and for SI joint blocks, ODI scores were again recorded. At this time, a decision was reached as to whether or not a second injection was warranted for diagnostic, prognostic, or therapeutic purposes. If indicated, this injection could be done at the same visit or at a later date, not to exceed 3 months. All baseline and outcome data were recorded separately before and after each injection.

Statistical Analysis

Three groups of analyses were performed: a primary crossover analysis comparing outcome measures only in those individuals who received both injections with and without sedation (N = 46); a parallel-group analysis comparing only the results of the first block in each group (sedation, N = 37; no sedation, N = 36); and an omnibus analysis that compared the results of all blocks done with sedation (N = 63) to all blocks done without sedation (N = 56). Prior to commencement, a power analysis conducted for the primary analysis based on the following assumptions garnered from our pilot data determined that 34 patients in each group had an 85% probability of detecting a 1-point difference between the blocks done with and without sedation for the participants who would receive two blocks: an alpha of 0.05; a mean baseline NRS score of 6.4 before the first block and 5.6 before the second block; mean postprocedure scores of 3.4 and 3.0 after first and second blocks performed with sedation, respectively; mean postprocedure pain scores of 4.4 and 4.0 after first and second blocks done without sedation; 45% of participants will exit the study without receiving both blocks secondary to inadequate response or persistent pain relief, based on a 6-month prestudy audit. The crossover comparison was designated to be the primary analysis because of the significantly increased power to detect a difference compared with a parallel group comparison.

We used the r statistical software package package, version 2.15.0 (R Foundation for Statistical Computing, Vienna, Austria) and JMP Statistics, version 10 (SAS, Cary, NC, USA) to determine descriptive statistics (calculate means, standard deviations [SDs], and frequencies) and to conduct univariate tests. Unless specified, P values are the results of Mann–Whitney tests for continuous variables and Fisher's exact tests for categorical variables.

We used a mixed-effects linear model (sas version 9.3) to assess the effect of method (sedation vs no sedation) when controlling for crossover group, block, age, gender, pain duration, etiology, psychopathology (anxiety or depression), institution, and opioid use. The differences in pain diary NRS score (primary outcome), procedure-related pain score, 1 month NRS score, 1 month ODI, and 1 month satisfaction vs their baseline variables were used as outcome variables. Patients were considered as a random effect to account for repeated observations on each patient.

In a separate set of analyses, we used generalized estimating equations (repeated measures logistic regression) to assess the effect of method (sedation vs no sedation) when controlling for crossover group, block, age, gender, pain duration, etiology, psychopathology, institution, and opioid use. We considered the percentage of subjects who received ≥50% and ≥80% pain relief with and without sedation as outcome variables. We examined the effect of sedation on 1 month pain relief and side effects in a similar manner but eliminated many covariates (i.e., psychopathology for 1 month outcome and all covariates for side effects) because of low cell counts. To account for the repeated observations on each patient, the analyses performed with generalized estimating equations were treated as a repeated factor, and an exchangeable working correlation was assumed.

Difference scores (e.g., baseline pain scores—pain diary postprocedure) were calculated for each patient. The composite variable is the difference between these differences (i.e., change for sedation minus change for no sedation) (Table 0004). This answers the question, “Did patients tend to get more immediate relief using sedation or no sedation?” If the mean is >0, then patients experienced a larger pain decrease under sedation.

Sedation was subcategorized into none, midazolam only, light (≤4 mL total of 1 mg/mL midazolam and 50 mcg/mL fentanyl), or heavy (>4 mL total of midazolam and fentanyl) for subgroup analysis. anova was used for baseline and 1 month pain score outcomes. Pain diary scores were skewed by a preponderance of 0 and 1 scores thereby invalidating anova assumptions; therefore Kruskal–Wallis was used for this outcome variable.

Role of Funding Source

The role of the funding sources included only the research personnel. No funding source participated in study design, data analysis, or drafting of the article.

Results

A total of 139 patients were screened for enrollment, with 73 randomized and treated. The reasons for the 66 patients not enrolled in the study are listed in Figure 0001. A total of 57 patients received a sacroiliac joint block and 16 received a sympathetic block, seven of whom underwent an SGB. Forty-six patients received two blocks and were included in the primary crossover analysis. Twenty of these patients were in the sedation crossover to no sedation group (group 1) and 36 received a sacroiliac joint block. One patient who underwent both blocks failed to show up for her 1 month visit after the second procedure done without sedation, and another patient failed to return his 1 month Oswestry disability questionnaire.

Figure 1

CONSORT flowchart showing study progression.

1) Includes four patients who underwent a stellate ganglion block (SGB) and four who underwent lumbar sympathetic blocks (LSBs); 2) Includes three patients who underwent SGB and five who underwent LSBs; 3) In missing cases, first pain diary extrapolated from first two pain scores and verbal accounts of pain relief; 4) In missing cases, second pain diary extrapolated from first two pain scores and verbal accounts of pain relief; 5) Includes one patient who completed 1 month 0–10 pain score but not Oswestry disability index.

Figure 1

CONSORT flowchart showing study progression.

1) Includes four patients who underwent a stellate ganglion block (SGB) and four who underwent lumbar sympathetic blocks (LSBs); 2) Includes three patients who underwent SGB and five who underwent LSBs; 3) In missing cases, first pain diary extrapolated from first two pain scores and verbal accounts of pain relief; 4) In missing cases, second pain diary extrapolated from first two pain scores and verbal accounts of pain relief; 5) Includes one patient who completed 1 month 0–10 pain score but not Oswestry disability index.

Fifty-four patients were recruited from Johns Hopkins and 19 from Walter Reed. The mean age of study subjects was 49.9 ± 14.8 years, with 58.9% being female. The mean duration of pain was 3.7 ± 3.9 years with a baseline NRS score of 6.3 ± 2.1. Study patients exhibited moderate to severe levels of disability, with a mean ODI of 41.4 ± 16.8. Thirty-six percent of patients were being treated with opioids, with 10% taking >60 mg of oral morphine equivalents daily. Eleven percent of blocks were accompanied by a side effect, with eight out of the 13 occurring when sedation was not used. The most common among these were headache (N = 4), dizziness (N = 3), and temporary worsening symptoms (N = 3), including one person who developed short-lasting (<1 week) worsening neurological symptoms. There were no significant baseline or demographic differences in age, gender, pain score, ODI, duration of pain, etiology of pain, anxiety, depression, or use of opioids between study groups except that group 1 patients had lower ODI scores before the first block (Table 0001), and only age differed between study sites (mean age of Walter Reed subjects 38.4 years [SD 9.6] vs 54.1 [SD 14.3] at JHU).

Table 1

Baseline demographics and clinical characteristics

 Group 1
 
Group 2
 
P value
 
 Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block 
 (N = 37) (N = 29) (N = 8) (N = 36) (N = 28) (N = 8) 
Age 
  Mean (SD); median [IQR] 49.7 (16.1); 51 [35, 60] 52.3 (15.6); 52 [43, 62] 40.3 (15.2); 33.5 [30, 55] 50.3 (13.8); 54 [41, 59] 50.5 (15.2); 55 [40, 60] 49.6 (8.0); 50 [43, 57] 0.76 0.83 0.19 
Gender 
  N (%) Female 20 (54.1%) 15 (51.7%) 5 (62.5%) 22 (61.1%) 16 (57.1%) 6 (75%) 0.64 0.73 
Baseline pain score prior to 1st block 
  Mean (SD); Median [IQR] 6.0 (1.8); 6 [4.5, 8] 5.8 (1.7); 6 [4.5, 7] 6.5 (2.3); 7 [5.4, 8.1] 6.7 (2.4); 6.5 [4.9, 8.6] 6.5 (2.4); 6 [4., 8.1] 7.3 (2.6); 7.5 [4.9, 10] 0.19 0.25 0.49 
Pain duration (in years) 
  Mean (SD); Median [IQR] 3.6 (4.2); 2 [0.8, 4.75] 3.6 (4.4); 2 [0.8, 4.8] 3.4 (3.8); 1.8 [1.2, 3.8] 3.8 (3.5); 2 [1.5, 5.5] 3.8 (3.6); 2 [1.5, 5.5] 3.8 (3.5); 2 [1.4, 5.8] 0.42 0.51 0.79 
Baseline ODI prior to 1st block  35.5 (14.0); 33 [25, 40] NA  47.6 (17.9); 52 [33, 60.5] NA  0.011  
Etiology (N, %) 
Surgery 3 (8.1) 1 (3.4) 2 (25) 8 (22.2) 2 (7.1) 6 (75) 0.31 0.69 0.20 
Injury 13 (35.1) 9 (31.0) 4 (50) 12 (33.3) 11 (39.3) 1 (12.5) 
Other 5 (13.5) 3 (10.3) 2 (25) 2 (5.6) 1 (3.6) 1 (12.5) 
Unknown 16 (43.2) 16 (55.2) 0 (0) 14 (38.9) 14 (50) 0 (0) 
Psychopathology (N, %) 
None 23 (62.2) 20 (69.0) 3 (37.5) 20 (55.5) 18 (64.3) 2 (25) 0.88 0.48 
Mixed Diagnoses 3 (8.1) 2 (6.9) 1 (12.5) 4 (11.1) 3 (10.7) 1 (12.5) 
Depression 10 (27) 8 (27.6) 2 (25) 10 (29.4) 7 (25) 3 (37.5) 
Anxiety 4 (10.8) 2 (6.9) 2 (25) 7 (19.4) 6 (21.4) 1 (12.5) 
Opioids (N, %) 
None 26 (70.3) 22 (75.9) 4 (50) 21 (58.3) 18 (64.3) 3 (37.5) 0.37 0.54 0.62 
≤60 mg 7 (18.9) 4 (13.8) 3 (37.5) 12 (33.3) 7 (25) 5 (62.5) 
>60 mg 4 (10.8) 3 (10.3) 1 (12.5) 3 (8.3) 3 (10.7) 0 (0) 
 Group 1
 
Group 2
 
P value
 
 Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block 
 (N = 37) (N = 29) (N = 8) (N = 36) (N = 28) (N = 8) 
Age 
  Mean (SD); median [IQR] 49.7 (16.1); 51 [35, 60] 52.3 (15.6); 52 [43, 62] 40.3 (15.2); 33.5 [30, 55] 50.3 (13.8); 54 [41, 59] 50.5 (15.2); 55 [40, 60] 49.6 (8.0); 50 [43, 57] 0.76 0.83 0.19 
Gender 
  N (%) Female 20 (54.1%) 15 (51.7%) 5 (62.5%) 22 (61.1%) 16 (57.1%) 6 (75%) 0.64 0.73 
Baseline pain score prior to 1st block 
  Mean (SD); Median [IQR] 6.0 (1.8); 6 [4.5, 8] 5.8 (1.7); 6 [4.5, 7] 6.5 (2.3); 7 [5.4, 8.1] 6.7 (2.4); 6.5 [4.9, 8.6] 6.5 (2.4); 6 [4., 8.1] 7.3 (2.6); 7.5 [4.9, 10] 0.19 0.25 0.49 
Pain duration (in years) 
  Mean (SD); Median [IQR] 3.6 (4.2); 2 [0.8, 4.75] 3.6 (4.4); 2 [0.8, 4.8] 3.4 (3.8); 1.8 [1.2, 3.8] 3.8 (3.5); 2 [1.5, 5.5] 3.8 (3.6); 2 [1.5, 5.5] 3.8 (3.5); 2 [1.4, 5.8] 0.42 0.51 0.79 
Baseline ODI prior to 1st block  35.5 (14.0); 33 [25, 40] NA  47.6 (17.9); 52 [33, 60.5] NA  0.011  
Etiology (N, %) 
Surgery 3 (8.1) 1 (3.4) 2 (25) 8 (22.2) 2 (7.1) 6 (75) 0.31 0.69 0.20 
Injury 13 (35.1) 9 (31.0) 4 (50) 12 (33.3) 11 (39.3) 1 (12.5) 
Other 5 (13.5) 3 (10.3) 2 (25) 2 (5.6) 1 (3.6) 1 (12.5) 
Unknown 16 (43.2) 16 (55.2) 0 (0) 14 (38.9) 14 (50) 0 (0) 
Psychopathology (N, %) 
None 23 (62.2) 20 (69.0) 3 (37.5) 20 (55.5) 18 (64.3) 2 (25) 0.88 0.48 
Mixed Diagnoses 3 (8.1) 2 (6.9) 1 (12.5) 4 (11.1) 3 (10.7) 1 (12.5) 
Depression 10 (27) 8 (27.6) 2 (25) 10 (29.4) 7 (25) 3 (37.5) 
Anxiety 4 (10.8) 2 (6.9) 2 (25) 7 (19.4) 6 (21.4) 1 (12.5) 
Opioids (N, %) 
None 26 (70.3) 22 (75.9) 4 (50) 21 (58.3) 18 (64.3) 3 (37.5) 0.37 0.54 0.62 
≤60 mg 7 (18.9) 4 (13.8) 3 (37.5) 12 (33.3) 7 (25) 5 (62.5) 
>60 mg 4 (10.8) 3 (10.3) 1 (12.5) 3 (8.3) 3 (10.7) 0 (0) 

Mixed Diagnoses—include patients with depression and/or anxiety in combination with PTSD, ADHD, and/or substance abuse disorders. Those with anxiety and/or depression were accounted for in the main categories of anxiety or depression.

ADHD = attention deficit hyperactivity disorder; IQR = interquartile range (i.e., the 25th and 75th percentiles); NA = not applicable; ODI = Oswestry disability index; PTSD = posttraumatic stress disorder; SI = sacroiliac; SD = standard deviation.

Crossover Group

There were no differences in pain (P = 0.97) or ODI (P = 0.72) scores prior to procedures with or without sedation. There was no group or hidden order effects (P = 0.19); the results were not significantly different whether the patient received sedation then no sedation or vice versa on any outcome except satisfaction (P = 0.025). Group 1 mean satisfaction score was not different based on whether they received sedation, whereas group 2 preferred sedation. When the data were combined and analyzed for sedation treatment effects (i.e., all blocks done with sedation vs those without sedation), there was no statistical difference in satisfaction (Table 0002).

Table 2

Sedation treatment effects—analysis of patients who completed crossover

 Sedation
 
No Sedation
 
P value*
 
 Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block 
 (N = 46) (N = 36) (N = 10) (N = 46) (N = 36) (N = 10) 
Change from baseline pain score after block (diagnostic pain diary) −3.9 (2.7); −3 [−5.5, −2] −3.7 (2.7); −3 [−5.5, −2] −4.6 (2.4); −4.3 [−5.9, −3] −2.7 (2.8); −3 [−4.4, −1.1] −2.5 (2.8); −3 [−4, −0.9] −3.5 (3.1); −3.3 [−4.9, −2.2] 0.003 0.011 0.081 
Change from baseline pain score 1 month after block −1 (2.9); −0.5 [−2.5, 1] −1.3 (3.1); −0.75 [−2.6, 1] −0.2 (1.7); −0.25 [−1.4, 0.75] −0.62 (2.3); −0.5 [−1.5, 0.5] −0.41 (1.9); −0.5 [−1.5,0.5] −1.4 (3.4); −0.8 [−3.8, 0.7] 0.44 0.17 0.37 
Change from baseline ODI 1 month after block  −4.9 (13.3); −4 [−10, 3] NA  −0.33 (11.6); 0 [−6, 8] NA  0.14  
Procedure-related pain during block 3.1 (2.8); 3 [0.1, 8] 3.1 (2.9); 2.8 [0, 5.6] 3.0 (2.3); 3.5 [0.9, 4.4] 5.5 (2.7); 5 [3, 8] 5.6 (2.8); 5.5 [3, 8] 5.0 (2.4); 4 [3.3, 5.8] <0.0001 <0.0001 0.06 
Satisfaction with block 4.0 (1.1); 4 [3.5, 5] 3.9 (1.1); 4.8 [3.5, 5] 4.2 (1.1); 4.8 [3.3, 5] 3.6 (1.4); 4 [3, 5] 3.6 (1.3); 4 [3, 5] 3.9 (1.7); 5 [3.1, 5] 0.12 0.16 0.55 
Side effects after block, N (%) 3 (6.5%) 2 (5.6%) 1 (10%) 7 (15.2%) 4 (11.1%) 3 (30%) 0.11 0.34 0.14 
 Sedation
 
No Sedation
 
P value*
 
 Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block 
 (N = 46) (N = 36) (N = 10) (N = 46) (N = 36) (N = 10) 
Change from baseline pain score after block (diagnostic pain diary) −3.9 (2.7); −3 [−5.5, −2] −3.7 (2.7); −3 [−5.5, −2] −4.6 (2.4); −4.3 [−5.9, −3] −2.7 (2.8); −3 [−4.4, −1.1] −2.5 (2.8); −3 [−4, −0.9] −3.5 (3.1); −3.3 [−4.9, −2.2] 0.003 0.011 0.081 
Change from baseline pain score 1 month after block −1 (2.9); −0.5 [−2.5, 1] −1.3 (3.1); −0.75 [−2.6, 1] −0.2 (1.7); −0.25 [−1.4, 0.75] −0.62 (2.3); −0.5 [−1.5, 0.5] −0.41 (1.9); −0.5 [−1.5,0.5] −1.4 (3.4); −0.8 [−3.8, 0.7] 0.44 0.17 0.37 
Change from baseline ODI 1 month after block  −4.9 (13.3); −4 [−10, 3] NA  −0.33 (11.6); 0 [−6, 8] NA  0.14  
Procedure-related pain during block 3.1 (2.8); 3 [0.1, 8] 3.1 (2.9); 2.8 [0, 5.6] 3.0 (2.3); 3.5 [0.9, 4.4] 5.5 (2.7); 5 [3, 8] 5.6 (2.8); 5.5 [3, 8] 5.0 (2.4); 4 [3.3, 5.8] <0.0001 <0.0001 0.06 
Satisfaction with block 4.0 (1.1); 4 [3.5, 5] 3.9 (1.1); 4.8 [3.5, 5] 4.2 (1.1); 4.8 [3.3, 5] 3.6 (1.4); 4 [3, 5] 3.6 (1.3); 4 [3, 5] 3.9 (1.7); 5 [3.1, 5] 0.12 0.16 0.55 
Side effects after block, N (%) 3 (6.5%) 2 (5.6%) 1 (10%) 7 (15.2%) 4 (11.1%) 3 (30%) 0.11 0.34 0.14 
*

P values are the results of random effects linear models that account for multiple observations per patient. All data are presented as mean (SD); median [IQR] unless otherwise noted.

IQR = interquartile range (i.e., the 25th and 75th percentiles); NA = not applicable; ODI = Oswestry disability index; SI = sacroiliac.

Blocks done with sedation were associated with a significantly larger reduction in pain diary score compared with no sedation (1.2 ± 2.6, P = 0.006) (Table 0003, Figure 0002), as well as lower procedure-related pain scores (difference in means 2.3 ± 2.5, P < 0.0001). A greater percentage of blocks done with sedation also resulted in ≥50% and ≥80% pain relief following the procedure (P = 0.039 and P = 0.008, respectively), suggesting that sedation increased the positive rate of diagnostic blocks (Figure 0002). There were no differences in 1 month NRS pain scores (P = 0.91), 1 month ODI outcomes (P = 0.28), or side effects between blocks done with and without sedation (P = 0.11).

Figure 2

Bar graph illustrating the proportion of patients experiencing a positive block (i.e., >50% and >80% relief) and those experiencing persistent pain relief at 1 month based on the use of sedation.

1) P values are the results of repeated measures logistic regression with other covariates controlled for. Repeated observations on each subject were treated as a repeated factor, and an exchangeable working correlation was assumed; 2) Positive diagnostic block designated based on postblock pain diary; 3) 1 month positive outcome defined as ≥50% pain relief.

Figure 2

Bar graph illustrating the proportion of patients experiencing a positive block (i.e., >50% and >80% relief) and those experiencing persistent pain relief at 1 month based on the use of sedation.

1) P values are the results of repeated measures logistic regression with other covariates controlled for. Repeated observations on each subject were treated as a repeated factor, and an exchangeable working correlation was assumed; 2) Positive diagnostic block designated based on postblock pain diary; 3) 1 month positive outcome defined as ≥50% pain relief.

Table 3

Sedation treatment effects—analysis of difference scores

 Crossover Subjects (Completed Both Blocks)
 
P value*
 
Overall Cohort (N = 46) SI Joint Injection (N = 36) Sympathetic Block (N = 10) Overall Cohort SI Joint Injection Sympathetic Block 
Difference in pain diary score 1.2 (2.6); 1 [−0.4, 2.9] 1.2 (2.7); 1 [−0.8, 3] 1.1 (1.8); 1 [0, 2.4] 0.0060 0.021 0.11 
Difference in 1 month change in pain score 0.48 (4.2); 0.5 [−2, 2.5] 0.95 (4.2); 0 [−2, 3] −1.2 (3.8); −1.5 [−1.1, 3.6] 0.91 0.55 0.38 
Difference in 1 month change in ODI  5.4 (21.3); 1 [−8.8, 20] NA  0.28  
Difference in procedure-related pain 2.3 (2.5); 2.5 [0.5, 4] 2.4 (2.4); 2.5 [0.8, 4] 2.0 (3.0); 2 [0.3, 3.6] <0.0001 <0.0001 0.057 
Difference in satisfaction score 0.34 (1.5); 0 [0.4, −1] 0.35 (1.5); 0 [−1, 0.6] 0.30 (1.5); 0 [0, 1.6] 0.075 0.11 0.60 
 Crossover Subjects (Completed Both Blocks)
 
P value*
 
Overall Cohort (N = 46) SI Joint Injection (N = 36) Sympathetic Block (N = 10) Overall Cohort SI Joint Injection Sympathetic Block 
Difference in pain diary score 1.2 (2.6); 1 [−0.4, 2.9] 1.2 (2.7); 1 [−0.8, 3] 1.1 (1.8); 1 [0, 2.4] 0.0060 0.021 0.11 
Difference in 1 month change in pain score 0.48 (4.2); 0.5 [−2, 2.5] 0.95 (4.2); 0 [−2, 3] −1.2 (3.8); −1.5 [−1.1, 3.6] 0.91 0.55 0.38 
Difference in 1 month change in ODI  5.4 (21.3); 1 [−8.8, 20] NA  0.28  
Difference in procedure-related pain 2.3 (2.5); 2.5 [0.5, 4] 2.4 (2.4); 2.5 [0.8, 4] 2.0 (3.0); 2 [0.3, 3.6] <0.0001 <0.0001 0.057 
Difference in satisfaction score 0.34 (1.5); 0 [0.4, −1] 0.35 (1.5); 0 [−1, 0.6] 0.30 (1.5); 0 [0, 1.6] 0.075 0.11 0.60 
*

P values are the results of one-sample Wilcoxon signed rank tests of the hypotheses that the true centers of the variable distributions are equal to 0 (for continuous variables) or to the “no difference” condition (for categorical variables). Differences in scores equal to baseline score minus outcome measure score for blocks done with sedation−baseline score minus outcome measure score for blocks with no sedation.

All data are presented as mean (SD); median [IQR] unless otherwise noted.

IQR = interquartile range (i.e., the 25th and 75th percentiles); SD = standard deviation; SI = sacroiliac.

Table 4

Sedation treatment effects—analysis of all study patients

 Sedation
 
No Sedation
 
P value*
 
Overall SI Joint Injection Sympathetic Block Overall SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block 
Cohort Cohort 
(N = 63) (N = 48) (N = 15) (N = 56) (N = 45) (N = 11) 
Change from baseline pain score after block (diagnostic pain diary) −3.8 (2.5); −3 [−5.5, −2] −3.8 (2.5); −3 [−5.5, −2] −3.7 (2.5); −3 [−5, −2.1] −3.0 (2.9); −3 [−4.5, −1.5] −2.9 (2.9); −3 [−4.5, −1.5] −3.6 (3.0); −3.5 [−4.8, −2.4] 0.009 0.020 0.15 
Change from baseline pain score 1 month after block −1.1 (2.8); −0.5 [−2.5, 1] −1.5 (3.0); −1 [−3, 1] 0.07 (1.5); 0 [−0.8, 1] −0.78 (2.4); −0.5 [−1.8, 0.5] −0.67 (2.2); −0.5 [−1.5, 0.5] −1.2 (3.3); −0.5 [−3.5, 0.5] 0.49 0.12 0.20 
Change from baseline ODI 1 month after block  −5.2 (13.4); −3 [−9.5, 2] NA  −1.6 (12.2); 0 [−6, 4.8] NA  0.20  
Procedure-related pain during block 2.8 (2.6); 2.5 [0, 5] 3.0 (2.8); 2.5 [0, 5] 2.5 (2.3); 2 [0, 4.3] 5.6 (2.6); 5 [3.3, 8] 5.7 (2.7); 5.5 [3.4, 8] 5.0 (2.2); 4 [3.5, 5.5] <0.0001 <0.0001 0.016 
Satisfaction with block 3.9 (1.1); 4 [3, 5] 4.0 (1.1); 4 [3.5, 5] 3.6 (1.2); 3 [3, 5] 3.7 (1.3); 4 [3, 5] 3.6 (1.2); 4 [3, 5] 4.0 (1.6); 5 [3.3, 5] 0.26 0.14 0.84 
Side effects after block, N (%) 5 (8.1%) 3 (6.4%) 2 (13.3%) 8 (14.3%) 5 (11.1%) 3 (27.3%) 0.15 0.34 0.17 
 Sedation
 
No Sedation
 
P value*
 
Overall SI Joint Injection Sympathetic Block Overall SI Joint Injection Sympathetic Block Overall Cohort SI Joint Injection Sympathetic Block 
Cohort Cohort 
(N = 63) (N = 48) (N = 15) (N = 56) (N = 45) (N = 11) 
Change from baseline pain score after block (diagnostic pain diary) −3.8 (2.5); −3 [−5.5, −2] −3.8 (2.5); −3 [−5.5, −2] −3.7 (2.5); −3 [−5, −2.1] −3.0 (2.9); −3 [−4.5, −1.5] −2.9 (2.9); −3 [−4.5, −1.5] −3.6 (3.0); −3.5 [−4.8, −2.4] 0.009 0.020 0.15 
Change from baseline pain score 1 month after block −1.1 (2.8); −0.5 [−2.5, 1] −1.5 (3.0); −1 [−3, 1] 0.07 (1.5); 0 [−0.8, 1] −0.78 (2.4); −0.5 [−1.8, 0.5] −0.67 (2.2); −0.5 [−1.5, 0.5] −1.2 (3.3); −0.5 [−3.5, 0.5] 0.49 0.12 0.20 
Change from baseline ODI 1 month after block  −5.2 (13.4); −3 [−9.5, 2] NA  −1.6 (12.2); 0 [−6, 4.8] NA  0.20  
Procedure-related pain during block 2.8 (2.6); 2.5 [0, 5] 3.0 (2.8); 2.5 [0, 5] 2.5 (2.3); 2 [0, 4.3] 5.6 (2.6); 5 [3.3, 8] 5.7 (2.7); 5.5 [3.4, 8] 5.0 (2.2); 4 [3.5, 5.5] <0.0001 <0.0001 0.016 
Satisfaction with block 3.9 (1.1); 4 [3, 5] 4.0 (1.1); 4 [3.5, 5] 3.6 (1.2); 3 [3, 5] 3.7 (1.3); 4 [3, 5] 3.6 (1.2); 4 [3, 5] 4.0 (1.6); 5 [3.3, 5] 0.26 0.14 0.84 
Side effects after block, N (%) 5 (8.1%) 3 (6.4%) 2 (13.3%) 8 (14.3%) 5 (11.1%) 3 (27.3%) 0.15 0.34 0.17 
*

P values are the results of random effects linear models that account for multiple observations on some subjects.

All data are presented as mean (SD); median [IQR] unless otherwise noted. IQR = interquartile range (i.e., the 25th and 75th percentiles); ODI = Oswestry disability index; SD = standard deviation; SI = sacroiliac.

Parallel and Omnibus Analyses

There were no significant differences in pain (P = 0.96) or ODI (P = 0.50) scores prior to procedures with or without sedation in the omnibus analysis. In the parallel analysis, there was no significant difference in baseline pain scores (P = 0.49), but patients receiving sedation had lower ODI scores (P = 0.011). When all patients were considered (omnibus analysis), similar to those in the crossover analysis, blocks done with sedation resulted in a significantly larger reduction in pain diary score and lower procedure-related pain scores compared with no sedation (P = 0.009 and P < 0.0001, respectively) (Table 0004). A greater proportion of blocks done with sedation resulted in ≥50% and ≥80% pain relief following the procedure than those done without sedation (data not shown). There were no differences in 1 month NRS scores (P = 0.49), 1 month ODI outcome (P = 0.2), or side effects (P = 0.15) between the two groups. No differences in the proportion of positive blocks using ≥50% cutoff were noted among the 11 patients with anxiety disorders based on whether (78%) or not (70%) they received sedation (P = 0.70). The parallel group analysis (restricted to the results of the first block) showed no significant difference between sedation and no sedation on the primary pain diary score outcome (P = 0.25); however, procedure-related pain was less when sedation was used (P < 0.0001).

Because the sedation effect was significant, a subgroup analysis of type of sedation was performed. Sedation with midazolam only and light sedation with midazolam and fentanyl were both associated with lower pain diary scores (Figure 0003). Pain diary scores in patients' who received heavy sedation with midazolam and fentanyl were not significantly different from those who received no sedation (Table 0005).

Figure 3

Sedation subgroup analysis presented as differences in baseline numerical rating scale (NRS) pain score, pain scores based on 6-hour pain diary (i.e., diagnostic block), and pain scores 1 month after the procedure stratified by sedation type.

No sedation (N = 56): no benzodiazepines or opioids were given.

Midazolam only (N = 16): use of only midazolam in doses ≤4 mg.

Light sedation (N = 28): use of ≤4 mL combination of 1 mg/mL midazolam and 50 mcg/mL fentanyl.

Heavy sedation (N = 19): use of >4 mL of midazolam and fentanyl.

P values are the results of random effects linear models that account for multiple observations on some subjects. No other pairs are significantly different.

Figure 3

Sedation subgroup analysis presented as differences in baseline numerical rating scale (NRS) pain score, pain scores based on 6-hour pain diary (i.e., diagnostic block), and pain scores 1 month after the procedure stratified by sedation type.

No sedation (N = 56): no benzodiazepines or opioids were given.

Midazolam only (N = 16): use of only midazolam in doses ≤4 mg.

Light sedation (N = 28): use of ≤4 mL combination of 1 mg/mL midazolam and 50 mcg/mL fentanyl.

Heavy sedation (N = 19): use of >4 mL of midazolam and fentanyl.

P values are the results of random effects linear models that account for multiple observations on some subjects. No other pairs are significantly different.

Table 5

Outcomes by sedation type

 Overall No Sedation Midazolam Only Light Heavy P value 
(N = 119) (N = 56, 47.1%) (N = 16, 13.5%) (N = 28, 23.5%) (N = 19, 16.0%) 
Baseline NRS 6.1 (2.2); 6 [4, 8] 6.1 (2.4); 6 [4, 8] 5.3 (2.2); 5.75 [3.75, 6] 6.2 (1.9); 6.75 [4.5, 8] 6.7 (2.1); 7 [5.25, 8] 0.81 
Pain Diary NRS 2.7 (2.6); 2 [0.5, 4.5] 3.1 (2.8); 2.25 [0.5, 5.1] 1.7 (2.0); 0.75 [0, 3] 2.4 (2.3); 1.5 [0.9, 3.3] 2.9 (2.5); 2 [0.75, 4.5] 0.035 
1 month NRS 5.2 (2.9); 5 [3, 7.5] 5.4 (3.0); 6 [3, 8] 4.0 (2.6); 3.5 [2, 6.3] 5.6 (3.2); 5.5 [2.8, 8.1] 5.1 (2.7); 5.5 [3.4, 6.8] 0.43 
 Overall No Sedation Midazolam Only Light Heavy P value 
(N = 119) (N = 56, 47.1%) (N = 16, 13.5%) (N = 28, 23.5%) (N = 19, 16.0%) 
Baseline NRS 6.1 (2.2); 6 [4, 8] 6.1 (2.4); 6 [4, 8] 5.3 (2.2); 5.75 [3.75, 6] 6.2 (1.9); 6.75 [4.5, 8] 6.7 (2.1); 7 [5.25, 8] 0.81 
Pain Diary NRS 2.7 (2.6); 2 [0.5, 4.5] 3.1 (2.8); 2.25 [0.5, 5.1] 1.7 (2.0); 0.75 [0, 3] 2.4 (2.3); 1.5 [0.9, 3.3] 2.9 (2.5); 2 [0.75, 4.5] 0.035 
1 month NRS 5.2 (2.9); 5 [3, 7.5] 5.4 (3.0); 6 [3, 8] 4.0 (2.6); 3.5 [2, 6.3] 5.6 (3.2); 5.5 [2.8, 8.1] 5.1 (2.7); 5.5 [3.4, 6.8] 0.43 

P values are the results of random effects linear models that account for multiple observations on some subjects.

No sedation significantly higher than light (P = 0.029) and midazolam only (P = 0.027).

Two observations missing.

All data are presented as mean (SD); median [IQR] unless otherwise noted.

IQR = interquartile range (i.e., the 25th and 75th percentiles); NRS = numerical rating scale; SD = standard deviation.

Multivariate Analysis: Crossover Group

After controlling for covariates, the use of sedation was associated with a greater decrease in pain diary scores than no sedation (odds ratio [OR] 1.2 [0.40, 1.95], P = 0.004). There were no treatment order effects (data not shown). The use of sedation was independently associated with a reduction in procedure-related pain (OR 2.4 [1.59, 3.12], P < 0.0001). Patients who received sedation had a higher probability of experiencing a ≥50% (OR 2.2; CI 1.07, 4.46; P = 0.031) or ≥80% reduction (OR 3.0; CI 1.32, 6.98; P = 0.009) in pain for 6 hours after the procedure. Sedation had no significant effect on 1 month NRS pain score (P = 0.29), 1 month ODI (P = 0.28), 1 month pain relief outcome (P = 0.23), or satisfaction (P = 0.12).

Omnibus Analysis

The use of sedation was independently associated with a greater decrease in pain diary scores than no sedation (OR 1.2; CI 0.43, 1.89; P = 0.0026). A group order effect was noted when all patients were included in the analysis, with patients in group 2 experiencing greater pain reduction on their 6-hour pain diary than those in group 1 (P = 0.0066). Although sedation itself did not predict satisfaction with the procedure (P = 0.077), patients in group 2 had higher satisfaction scores (P = 0.05). The use of sedation regardless of group allocation was associated with a reduction in procedural pain (2.4; CI 1.65, 3.09; P < 0.0001). Female gender was associated with higher procedure-related pain scores (1.2 [0.06, 2.48], P = 0.06). Older patients (OR 1.05; CI 1, 1.1; P = 0.036 for ≥50% pain reduction, and OR 1.06; CI 1.01,1.11; P = 0.016 for ≥80% reduction) and those who received sedation (OR 2.2; CI 1.02, 4.91; P = 0.046 for ≥50% pain reduction, and OR 2.5; CI 1.13, 5.40; P = 0.023 for ≥80% pain relief) were also more likely to experience a positive diagnostic block based on pain diaries. Sedation had no significant effect on 1-month outcomes.

Discussion

We elected to perform a clinical crossover trial designed to determine the effect of sedation on the positive rate of diagnostic injections that has components of both explanatory and pragmatic studies. This is a part of a growing trend to navigate the internal conflict that arises between conducting rigorous controlled studies and comparative effectiveness studies that facilitate applying evidence-based medicine to clinical practice. In some respects, this blended methodological approach can be highly advantageous for critically examining highly relevant clinical issues in pain medicine, as they can be performed with fewer resources, and the results are readily generalizable to “real-world” scenarios.

The main finding of this study is that sedation can affect the outcome of diagnostic blocks. The contention that sedation can undermine the diagnostic value of an injection is supported by numerous pain Websites that forewarn patients that sedation will not be used during diagnostic procedures. Notwithstanding the financial incentive to sedate patients (i.e., increased reimbursement), the ostensible benefit in administering sedation is to increase patient satisfaction and tolerance. However, no increase in satisfaction was noted in the overall cohort when sedation was used. Therefore, the use of sedation may in some circumstances increase the likelihood of a false positive diagnosis leading to unnecessary procedures (e.g., sacroiliac joint fusion or denervation), increase costs, and elevate the risk for a serious adverse outcome [56] without providing quantifiably tangible benefits.

Although the most likely cause of the discrepancy between the rate of positive blocks is that sedation caused a false positive result, one cannot discount the possibility that some subjects experienced false negative blocks when sedation was not employed. Scenarios in which the judicious use of anxiolytics, and even analgesics, may enhance accuracy include technically challenging procedures (e.g., obesity) in extremely anxious individuals and in cognitively challenged patients who may not be able to distinguish their index pain from procedure-induced discomfort.

Among the numerous causes of false positive blocks [6,57], the use of sedation is one of the most remediable. Not only opioids but also sedatives can alleviate pain by virtue of their anxiolytic effects and muscle relaxation properties [58]. No study has evaluated the necessity of sedation for sacroiliac joint and sympathetic blocks, but a survey of 500 patients who underwent epidural or facet injections found that whereas 17% requested sedation before their initial procedure, this figure increased to 28% before repeat injections [39]. In our study, a group order effect was found in omnibus multivariate analysis in which patients in group 2 reported greater postprocedure pain relief and higher satisfaction scores than those in group 1.

Few studies have evaluated the effect sedation has on diagnostic blocks. In a series of randomized studies by Manchikanti et al. [30,41–43], the authors reported that between 5% and 27% of patients will experience significant pain relief during facet blocks from midazolam and fentanyl, with the effect being less in the lumbar spine, with midazolam, and when ≥80% pain relief is the threshold for a positive block. However, these studies suffered from significant methodological flaws including the failure to measure pain relief differences after the blocks (i.e., they were measured after sedation but before the procedures) and the fact that these patients, who repeatedly responded with long-term relief from local anesthetic blocks, likely included a high proportion of placebo responders. These studies also purportedly blinded patients to the three treatment groups—midazolam, fentanyl, and saline—which to some degree mitigate the very large placebo effect associated with anxiolysis [59], and hence does not reflect clinical circumstances. Numerous studies have shown that expectations, which play a large role in the placebo effect, can affect treatment outcomes [46,47], and this covariate was not measured. Had our patients been effectively blinded as to whether they received sedation and expectations controlled for, our results may have differed.

There are several reasons why even more patients did not experience better pain relief with sedation. These include the fact that few patients in this study received heavy sedation, and that with proper teaching, most patients are able to follow instructions to resume normal activity and distinguish between procedure-related pain and pain that stems from their index condition. In clinical practice, anxiety levels may be higher, and patients may not fully understand instructions regarding how to fill out pain diaries. Studies have shown that in clinical settings, miscommunication between doctors and patients frequently occurs in many contexts [60–62]. Another explanation as to why a larger difference was not observed is the Hawthorne effect (i.e., patients took extra precautions to discount the effects of sedation because they knew we were measuring that).

Our study has implications that extend across multiple specialties. The use of pain relief following a diagnostic/prognostic intervention is frequently used as the basis for invasive procedures. These procedures include but are not limited to spinal decompression and fusion surgeries (selective nerve root blocks and discography) [63–66], sacroiliac joint fusion and radiofrequency denervation [50,67,68], thoracic outlet surgery [69], shoulder and other joint surgery [70], lumbar facetectomy and radiofrequency denervation [20,71], intrathecal pump placement [72], piriformis surgery and botulinum toxin injections [15], and occipital nerve pulsed radiofrequency, and nerve stimulation [73,74]. Although the evidence supporting prognostic blocks as a means to improve treatment outcomes is limited to a few controlled [27,66] and a plethora of uncontrolled studies [63,64,69,70,75,76], one can deduce a priori that enhancing the accuracy of prognostic injections is likely to improve treatment outcomes.

There are several limitations to our study. The first is that to enable a crossover comparison, which reduced variations in clinical factors, we limited our diagnostic procedures to two types of blocks, which a preliminary analysis revealed were the diagnostic injections most likely to be repeated. Although this allowed us to control for the myriad variables that could affect group comparisons, it limits generalizability. Second, we permitted physicians to sedate patients based on clinical scenarios and personal preference, such that some patients were heavily sedated with opioids and others received light sedation with only midazolam. This paradigm is consistent with comparative effectiveness research and enhances external validity [77], but it may introduce confounding factors that limit the conclusions one can draw. Third, without an independent observer confirming similar needle placement and contrast spread for both injections, it is possible that differences in the site of injections could account for variations in pain relief. Finally, this study was not powered to determine the effect that the use of sedation during diagnostic injections has on long-term treatment outcomes, which would require a much higher study population (e.g., hundreds of patients).

In summary, this study suggests that the injudicious use of sedation may increase false positive blocks, which may lead to inaccurate diagnoses and failed treatments. Future research should aim to determine the effect that sedation has on long-term treatment outcomes.

References

Manchikanti
L
Singh
V
Pampati
V
Smith
HS
Hirsch
JA
.
Analysis of growth of interventional techniques in managing chronic pain in the medicare population: A 10-year evaluation from 1997 to 2006
.
Pain Physician
 
2009
;
12
:
9
34
.
Chou
R
Atlas
SJ
Stanos
SP
Rosenquist
RW
.
Nonsurgical interventional therapies for low back pain: A review of the evidence for an American Pain Society clinical practice guideline
.
Spine
 
2009
;
34
:
1078
1093
.
Staal
JB
de Bie
RA
de Vet
HC
Hildebrandt
J
Nelemans
P
.
Injection therapy for subacute and chronic low back pain: An updated Cochrane review
.
Spine
 
2009
;
34
:
49
59
.
Peterson
C
Hodler
J
.
Evidence-based radiology (part 1): Is there sufficient research to support the use of therapeutic injections for the spine and sacroiliac joints?
Skeletal Radiol
 
2010
;
39
:
5
9
.
Cohen
SP
Gambel
J
Raja
SN
Galvagno
S
.
The contribution of sympathetic mechanisms to postamputation phantom and residual limb pain: A pilot study
.
J Pain
 
2011
;
12
:
859
867
.
Cohen
SP
Hurley
RW
.
The ability of diagnostic spinal injections to predict surgical outcomes
.
Anesth Analg
 
2007
;
105
:
1756
1775
.
DePalma
MJ
Ketchum
JM
Saullo
T
.
What is the source of chronic low back pain and does age play a role?
Pain Med
 
2011
;
12
:
224
233
.
Szadek
KM
van der Wurff
P
van Tulder
MW
Zuurmond
WW
Perez
RS
.
Diagnostic validity of criteria for sacroiliac joint pain: A systematic review
.
J Pain
 
2009
;
10
:
354
368
.
Hancock
MJ
Maher
CG
Latimer
J
et al
Systematic review of tests to identify the disc, SIJ or facet joint as the source of low back pain
.
Eur Spine J
 
2007
;
16
:
1539
1550
.
Boas
RA
.
Sympathetic nerve blocks: In search of a role
.
Reg Anesth Pain Med
 
1998
;
23
:
292
305
.
Day
M
.
Sympathetic blocks: The evidence
.
Pain Pract
 
2008
;
8
:
98
109
.
Pateder
DB
Berg
JH
Thal
R
.
Neck and shoulder pain: Differentiating cervical spine pathology from shoulder pathology
.
J Surg Orthop Adv
 
2009
;
18
:
170
174
.
IHC IHSCS. International classification of headache disorders
.
The international classification of headache disorders
.
Cephalalgia
 
2004
;
24
:
1
160
.
Patijn
J
Mekhail
N
Hayek
S
et al
Meralgia paresthetica
.
Pain Pract
 
2011
;
11
:
302
308
.
Miller
TA
White
KP
Ross
DC
.
The diagnosis and management of piriformis syndrome: Myths and facts
.
Can J Neurol Sci
 
2012
;
39
:
577
583
.
Yoong
P
Guirguis
R
Darrah
R
Wijeratna
M
Porteous
MJ
.
Evaluation of ultrasound-guided diagnostic local anaesthetic hip joint injection for osteoarthritis
.
Skeletal Radiol
 
2012
;
41
:
981
985
.
von Gaza
W
.
Die Resektion der paravertebralen Nerven und die isolierte Durchschneidung des Ramus communicans
.
Arch Klin Chir
 
1924
;
133
:
479
.
North
RB
Kidd
DH
Zahurak
M
Piantadosi
S
.
Specificity of diagnostic nerve blocks: A prospective, randomized study of sciatica due to lumbosacral spine disease
.
Pain
 
1996
;
65
:
77
85
.
Falco
FJ
Erhart
S
Wargo
BW
et al
Systematic review of diagnostic utility and therapeutic effectiveness of cervical facet joint interventions
.
Pain Physician
 
2009
;
12
:
323
344
.
Cohen
SP
Huang
JH
Brummett
C
.
Facet joint pain- advances in patient selection and management
.
Nature Rev Rheumatol
 
2013
;
9
:
101
116
.
Hogan
QH
Erickson
SJ
Haddox
JD
Abram
SE
.
The spread of solutions during stellate ganglion block
.
Reg Anesth
 
1992
;
17
:
78
83
.
Schmid
MR
Kissling
RO
Curt
A
Jaschko
G
Hodler
J
.
Sympathetic skin response: Monitoring of CT-guided lumbar sympathetic blocks
.
Radiology
 
2006
;
241
:
595
602
.
Rupert
MP
Lee
M
Manchikanti
L
Datta
S
Cohen
SP
.
Evaluation of sacroiliac joint interventions: A systematic appraisal of the literature
.
Pain Physician
 
2009
;
12
:
399
418
.
Hong
JH
Kim
AR
Lee
MY
Kim
YC
Oh
MJ
.
A prospective evaluation of psoas muscle and intravascular injection in lumbar sympathetic ganglion block
.
Anesth Analg
 
2010
;
111
:
802
807
.
Maigne
JY
Aivaliklis
A
Pfefer
F
.
Results of sacroiliac joint double block and value of sacroiliac pain provocation tests in 54 patients with low back pain
.
Spine
 
1996
;
21
:
1889
1892
.
Bogduk
N
.
Evidence-informed management of chronic low back pain with facet injections and radiofrequency neurotomy
.
Spine J
 
2008
;
8
:
56
64
.
Cohen
SP
Williams
KA
Kurihara
C
et al
Randomized, comparative cost-effectiveness study comparing 0, 1 and 2 medial branch blocks before lumbar facet radiofrequency denervation
.
Anesthesiology
 
2010
;
113
:
395
405
.
Novak
S
Nemeth
WC
.
RE: cost-effectiveness of diagnostic medial branch blocks before radiofrequency denervation
.
Spine J
 
2008
;
8
:
412
413
.
Hogan
QH
Abram
SE
.
Neural blockade for diagnosis and prognosis. A review
.
Anesthesiology
 
1997
;
86
:
216
241
.
Manchikanti
L
Pampati
V
Damron
K
.
The role of placebo and nocebo effects of perioperative administration of sedatives and opioids in interventional pain management
.
Pain Physician
 
2005
;
8
:
349
355
.
Ackerman
WE
Munir
MA
Zhang
JM
Ghaleb
A
.
Are diagnostic lumbar facet injections influenced by pain of muscular origin?
Pain Pract
 
2004
;
4
:
286
291
.
Woolf
CJ
Wiesenfeld-Hallin
Z
.
The systemic administration of local anaesthetics produces a selective depression of C-afferent fibre evoked activity in the spinal cord
.
Pain
 
1985
;
23
:
361
374
.
Dreyfuss
P
Schwarzer
AC
Lau
P
Bogduk
N
.
Specificity of lumbar medial branch and L5 dorsal ramus blocks. A computed tomography study
.
Spine
 
1997
;
22
:
895
902
.
Cohen
SP
Strassels
SA
Kurihara
C
et al
Randomized, double-blinded computed tomographic study evaluating the accuracy of 0.25 mL and 0.5 mL cervical facet joint nerve (medial branch) blocks
.
Anesthesiology
 
2010
;
112
:
144
152
.
Rosenberg
JM
Quint
DJ
de Rosayro
AM
.
Computerized tomographic localization of clinically-guided sacroiliac joint injections
.
Clin J Pain
 
2000
;
16
:
18
21
.
Smith
HS
Chopra
P
Patel
VB
Frey
ME
Rastogi
R
.
Systematic review of the role of sedation in diagnostic spinal interventional techniques
.
Pain Physician
 
2009
;
12
:
195
206
.
Jasper
JF
.
Effect of sedation on validity of diagnostic facet joint injections
.
Pain Physician
 
2001
;
4
:
285
286
.
Kim
N
Delport
E
Cucuzzella
T
Marley
J
Pruitt
C
.
Is sedation indicated before spinal injections?
Spine (Phila Pa 1976)
 
2007
;
32
:
E748
752
.
Cucuzzella
TR
Delport
EG
Kim
N
et al
A survey: Conscious sedation with epidural and zygapophyseal injections: Is it necessary?
Spine J
 
2006
;
6
:
364
369
.
Hingorani
K
.
Diazepam in backache. A double-blind controlled trial
.
Ann Phys Med
 
1966
;
8
:
303
306
.
Manchikanti
L
Pampati
V
Damron
KS
et al
The effect of sedation on diagnostic validity of facet joint nerve blocks: An evaluation to assess similarities in population with involvement in cervical and lumbar regions
.
Pain Physician
 
2006
;
1
:
47
51
.
Manchikanti
L
Pampati
V
Damron
KS
et al
A randomized, prospective, double-blind, placebo-controlled evaluation of the effect of sedation on diagnostic validity of cervical facet joint pain
.
Pain Physician
 
2004
;
7
:
301
309
.
Manchikanti
L
Damron
KS
Rivera
JJ
et al
Evaluation of the effect of sedation as a confounding factor in the diagnostic validity of lumbar facet joint pain: A prospective, randomized, double-blind, placebo-controlled evaluation
.
Pain Physician
 
2004
;
7
:
411
417
.
Bruehl
S
.
An update on the pathophysiology of complex regional pain syndrome
.
Anesthesiology
 
2010
;
113
:
713
725
.
Harden
RN
Bruehl
S
Perez
RS
et al
Validation of proposed diagnostic criteria (the “Budapest Criteria”) for Complex Regional Pain Syndrome
.
Pain
 
2010
;
150
:
268
274
.
Koshi
EB
Short
CA
.
Placebo theory and its implications for research and clinical practice: A review of the recent literature
.
Pain Pract
 
2007
;
7
:
4
20
.
Carlino
E
Pollo
A
Benedetti
F
.
Placebo analgesia and beyond: A melting pot of concepts and ideas for neuroscience
.
Curr Opin Anaesthesiol
 
2011
;
24
:
540
544
.
Flaten
MA
Aslaksen
PM
Lyby
PS
Bjørkedal
E
.
The relation of emotions to placebo responses
.
Philos Trans R Soc Lond B Biol Sci
 
2011
;
366
:
1818
1827
.
Cohen
SP
Hurley
RW
Buckenmaier
CC
3rd
et al
Randomized, placebo-controlled study evaluating lateral branch radiofrequency denervation for sacroiliac joint pain
.
Anesthesiology
 
2008
;
109
:
279
288
.
Cohen
SP
Neufeld
N
Chen
Y
.
Sacroiliac joint pain: A comprehensive review of epidemiology, diagnosis and treatment
.
Expert Rev Neurother
 
2013
;
13
:
99
116
.
Kim
WH
Kim
SK
Lee
CJ
Kim
TH
Sim
WS
.
Determination of adequate entry angle of lumbar sympathetic ganglion block in Korean
.
Korean J Pain
 
2010
;
23
:
11
17
.
Murata
Y
Takahashi
K
Yamagata
M
et al
Variations in the number and position of human lumbar sympathetic ganglia and rami communicantes
.
Clin Anat
 
2003
;
16
:
108
113
.
Dick
AM
Gabbott
DA
Hardy
PA
.
Plasma concentrations of bupivacaine following single needle lumbar sympathectomy using two volumes of 0.25% bupivacaine plain solution
.
Anaesthesia
 
1996
;
51
:
750
751
.
Di Filippo
A
Natale
V
Del Po
F
et al
Skin temperature during sympathetic block: A clinical comparison of bupivacaine 0.5% and ropivacaine 0.5% or 0.75%
.
Anaesth Intensive Care
 
2006
;
34
:
334
337
.
Fairbank
JC
Couper
J
Davies
JB
O'Brien
JP
.
The Oswestry low back pain disability questionnaire
.
Physiotherapy
 
1980
;
66
:
271
273
.
Rathmell
JP
Michna
E
Fitzgibbon
DR
et al
Injury and liability associated with cervical procedures for chronic pain
.
Anesthesiology
 
2011
;
114
:
918
926
.
Kaptchuk
TJ
Stason
WB
Davis
RB
et al
Sham device v inert pill: Randomised controlled trial of two placebo treatments
.
BMJ
 
2006
;
332
:
391
397
.
Rudolph
U
Knoflach
F
.
Beyond classical benzodiazepines: Novel therapeutic potential of GABA-A receptor subtypes
.
Nat Rev Drug Discov
 
2011
;
10
:
685
697
.
Schweizer
E
Rickels
K
.
Placebo response on generalized anxiety: Its effect on the outcome of clinical trials
.
J Clin Psychiatry
 
1997
;
58
(
suppl 11
):
30
38
.
Safeer
RS
Keenan
J
.
Health literacy: The gap between physicians and patients
.
Am Fam Physician
 
2005
;
72
:
463
468
.
Schwartz
PH
Edenberg
E
Barrett
PR
et al
Patient understanding of benefits, risks, and alternatives to screening colonoscopy
.
Fam Med
 
2013
;
45
:
83
89
.
Hanchak
NA
Patel
MB
Berlin
JA
Strom
BL
.
Patient misunderstanding of dosing instructions
.
J Gen Intern Med
 
1996
;
11
:
325
328
.
Sasso
RC
Macadaeg
K
Nordmann
D
Smith
M
.
Selective nerve root injections can predict surgical outcome for lumbar and cervical radiculopathy: Comparison to magnetic resonance imaging
.
J Spinal Disord Tech
 
2005
;
18
:
471
478
.
Anderberg
L
Annertz
M
Brandt
L
Saveland
H
.
Selective diagnostic cervical nerve root block—correlation with clinical symptoms and MRI-pathology
.
Acta Neurochir (Wien)
 
2004
;
146
:
559
565
.
Osler
GE
.
Cervical analgesic discography. A test for diagnosis of the painful disc syndrome
.
S Afr Med J
 
1987
;
71
:
363
.
Ohtori
S
Kinoshita
T
Yamashita
M
et al
Results of surgery for discogenic low back pain: A randomized study using discography vs discoblock for diagnosis
.
Spine
 
2009
;
34
:
1345
1348
.
Wise
CL
Dall
BE
.
Minimally invasive sacroiliac arthrodesis: Outcomes of a new technique
.
J Spinal Disord Tech
 
2008
;
21
:
579
584
.
Buchowski
JM
Kebaish
KM
Sinkov
V
et al
Functional and radiographic outcome of sacroiliac arthrodesis for the disorders of the sacroiliac joint
.
Spine J
 
2005
;
5
:
520
528
.
Lum
YW
Brooke
BS
Likes
K
et al
Impact of anterior scalene lidocaine blocks on predicting surgical success in older patients with neurogenic thoracic outlet syndrome
.
J Vasc Surg
 
2012
;
55
:
1370
1375
.
Lim
JT
Acornley
A
Dodenhoff
RM
.
Recovery after arthroscopic subacromial decompression: Prognostic value of the subacromial injection test
.
Arthroscopy
 
2005
;
21
:
680
683
.
Lovely
TJ
Rastogi
P
.
The value of provocative facet blocking as a predictor of success in lumbar spine fusion
.
J Spinal Disord
 
1997
;
10
:
512
517
.
Kim
D
Saidov
A
Mandhare
V
Shuster
A
.
Role of pretrial systemic opioid requirements, intrathecal trial dose, and non-psychological factors as predictors of outcome for intrathecal pump therapy: One clinician's experience with lumbar postlaminectomy pain
.
Neuromodulation
 
2011
;
14
:
165
175
.
Saper
JR
Dodick
DW
Silberstein
SD
et al
Occipital nerve stimulation for the treatment of intractable chronic migraine headache: ONSTIM feasibility study
.
Cephalalgia
 
2011
;
31
:
271
285
.
Huang
JH
Galvagno
S
Hameed
H
et al
Occipital nerve pulsed radiofrequency treatment: A multi-center study evaluating predictors of outcome
.
Pain Med
 
2012
;
13
:
489
497
.
Dooley
JF
McBroom
RJ
Taguchi
T
Macnab
I
.
Nerve root infiltration in the diagnosis of radicular pain
.
Spine
 
1988
;
13
:
79
83
.
Haueisen
DC
Smith
BS
Myers
SR
Pryce
ML
.
The diagnostic accuracy of spinal nerve injection studies. Their role in the evaluation of recurrent sciatica
.
Clin Orthop Relat Res
 
1985
;
198
:
179
183
.
Umscheid
CA
Margolis
DJ
Grossman
CE
.
Key concepts of clinical trials: a narrative review
.
Postgrad Med
 
2011
;
123
:
194
204
.
Conflicts of Interest: The authors declare they have no conflicts of interest.
Funding Source: Funded in part by the Centers for Rehabilitation and Sciences Research, Uniformed Services University of the Health Sciences, Bethesda, MD.
The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense.
ClinicalTrials.gov Identifier: NCT01472835.