Cost Effectiveness of Periradicular Infiltration for Sciatica

SPINE Volume 26, Number 23, pp 2587–2595 ©2001, Lippincott Williams & Wilkins, Inc. Cost Effectiveness of Periradicular Infiltration for Sciatica Sub...
Author: Buddy Randall
4 downloads 0 Views 183KB Size
SPINE Volume 26, Number 23, pp 2587–2595 ©2001, Lippincott Williams & Wilkins, Inc.

Cost Effectiveness of Periradicular Infiltration for Sciatica Subgroup Analysis of a Randomized Controlled Trial Jaro Karppinen, MD,* Arto Ohinmaa, PhD,† Antti Malmivaara, MD, PhD,‡ Mauno Kurunlahti, MD,§ Eero Kyllo¨nen, MD, PhD,* Tuomo Pienima¨ki, MD, PhD,*储 Pentti Nieminen, PhD,¶ Osmo Tervonen, MD, PhD,§ and Heikki Vanharanta, MD, PhD*

Study Design. A subgroup analysis of a prospective, randomized controlled trial was performed. Objective. To describe the cost effectiveness of periradicular infiltration with steroid in subgroups of patients with sciatica. Summary of Background Data. A recent trial on periradicular infiltration indicated that a methylprednisolone– bupivacaine combination had a short-term effect, as compared with that of saline. This report describes the efficacy and cost effectiveness of steroid in subgroups of patients with sciatic. Methods. This study involved 160 patients with unilateral sciatica. Outcome assessments were leg pain (100-mm visual analog scale), disability on the Oswestry Low Back Disability Questionnaire, and the Nottingham Health Profile. Data on medical costs and sick leaves also were gathered. Patients were randomized for periradicular infiltration with either methylprednisolone– bupivacaine or saline. The adjusted between-group treatment differences at each follow-up assessment, the number of patients free of leg pain (responders, cutoff 75%), and efficacy by the area-under-the-curve method were calculated. For the cost-effectiveness estimate, the total costs were divided by the number of responders. The rate of operations in different subgroups was evaluated by Kaplan-Meier analysis. Results. In the case of contained herniations, the steroid injection produced significant treatment effects and short-term efficacy in leg pain and in Nottingham Health Profile emotional reactions. For symptomatic lesions at L3–L4 –L5, steroid was superior to saline for leg pain, disability, and straight leg raising in the short term. By 1 year, steroid seemed to have prevented operations for contained herniations, costing $12,666 less per responder in the steroid group (P ⬍ 0.01). For extrusions, steroid seemed to increase the operation rate, and the steroid infiltration was more expensive, costing $4445 per responder (P ⬍ 0.01). Conclusions. In addition to short-term effectiveness for contained herniations and lesions at L3–L4 –L5, steroid treatment also prevented surgery for contained hernia-

From the *Department of Physical Medicine and Rehabilitation, University Hospital of Oulu, the †Department of Economics, University of Oulu, the ‡Department of Occupational Medicine, Finnish Institute of Occupational Health, and Orton Hospital, Invalid Foundation, Helsinki, the §Department Radiology, University Hospital of Oulu, the 储Oulu Regional Institute of Occupational Health, and the ¶Medical Informatics Group, University of Oulu, Finland. Supported by grants from the Yrjo¨ Jahnsson Foundation, the Finnish Office for Health Technology Assessment, the Finnish Work Environment Fund, and the International Spinal Injection Society. Device status category: 1. Conflict of interest category: 14.

tions. However, steroid was countereffective for extrusions. The results of the subgroup analyses call for a verification study. [Key words: conservative treatment, cost-effectiveness analysis, disc displacement, periradicular infiltration, randomized controlled trial, sciatica, subgroup analysis] Spine 2001;26:2587–2595

Sciatic pain is classified as pain radiating from the back into the dermatome of the affected nerve root along the femoral or sciatic nerve trunk.31 Of the patients with acute back pain, 1% to 10% have sciatica.13,19 In Finland, among a population of 5 million, compensation was paid for almost 600,000 days of sick leave in 1999 because of lumbar disc disease (International Classification of Diseases-10 [ICD-10], M51.1–M51.3). According to the National Insurance Register, the reimbursement cost of these sick leaves was $22 million in 1999 ($1 ⫽ 5.096 Finnish marks). Economic analysis is a necessary input before a choice is made between two or more competing treatments for the same illness. This analysis can be achieved by costeffectiveness analysis, which is basically a ratio in which health changes resulting from an intervention are captured in the denominator and changes in resource use, valued in monetary terms, appear in the numerator, both being compared with a specific alternative.7,14,30 Costutility analysis is a special form of cost-effectiveness analysis in which health-related quality of life outcomes are weighted according to the value that the general population or patients place on them. In cost-effectiveness analysis, divergent health outcomes make studies difficult to compare.5 Therefore, cost-utility analysis, in which outcomes are presented as quality-adjusted life years, have been recommended.26 The results from the author’s recent trial of periradicularinfiltrationindicatedthatacombinationofmethylprednisolone and bupivacaine had only a short-term effect, as compared with saline,21 although as a targeted technique, periradicular infiltration is much more specific than epidural injections. However, there is a possibility that the treatment is effective in a subgroup that is overlooked because of heterogeneity in study populations, follow-up times, and intervention methods.32 The purpose of this study was to describe the efficacy and cost 2587

2588 Spine • Volume 26 • Number 23 • 2001

effectiveness of periradicular infiltration with steroid in subgroups of patients with sciatica.

jected. The volume of the injection was 2 mL for L4 and L5 blocks, and 3 mL for S1. The allowed cointerventions were pain medication and discectomy in cases of severe sciatic pain. Booster injections were not allowed.

Methods Selection and Evaluation of Study Population. Patients with sciatica were referred by general practitioners working in the catchment area of the Oulu University Hospital. The inclusion criteria required that participants have unilateral pain radiating dermatomally from the back to below the knee that had lasted 1 to 6 months. The main exclusion criteria were an earlier back operation, an application for early retirement, and clinical depression.21 Every patient recorded leg pain on a 100-mm visual analog scale, disability on the Oswestry Low Back Disability Questionnaire,12 and quality of life on the Nottingham Health Profile (NHP).24 The clinical examination included straight leg raising and lumbar flexion by the modified Schober measure.21 In the absence of neurologic deficits, the dermatomal distribution of pain, as described by Keegan and Garret,23 was used as the basis for the clinical evaluation of the affected nerve. Patients for whom written consent was obtained underwent the investigations and completed the questionnaires. Magnetic resonance imaging (MRI) scans were obtained with a 1.5-T imaging system (Signa; General Electric, Milwaukee, WI) consisting of T2-weighted sagittal images, T1-weighted axial images with a repetition time (TR)/echo time (TE) of 640/14 ms, and postcontrast gadolinium-diethylenetriamine pentaacetic acid (GdDTPA; Magnevist, Schering, Berlin, Germany) axial images. The MRI findings of disc morphology were classified as bulge, contained herniation (broad-based herniation not extending through the posterior longitudinal ligament), and extrusion (herniation extruding through the posterior longitudinal ligament). All the images were read by two experienced MRI radiologists blinded to the clinical status of the patients. In the case of discrepancies between the radiologists, consensus was reached before the final grading. The interobserver kappa value for this classification was 0.52 (moderate agreement), and the intraobserver kappa value was 0.72 (substantial agreement).22 The study protocol was approved by the ethics committee of the Oulu University Hospital.

Randomization and Treatments. The randomization process was based on a published list of random permutations6 with a block size of 16. A person uninvolved in the study placed the assignments in sealed envelopes with running numbers. The envelopes were used in the order provided, with patients taking them to the Department of Radiology, where an authorized nurse filled a tape-covered syringe with the treatment agent indicated therein. The assignments were thus masked to the patient, the physicians, and the radiologist giving the injection. The periradicular infiltration was performed by a conventional technique under repeated fluoroscopic screening (C-arm) with a 22- to 25-G spinal needle.9,21 For the L4 and L5 roots, the needle was advanced obliquely toward the base of the corresponding pedicle. For the S1 root, the needle was aimed in a medial and cephalad direction relative to the corresponding sacral foramen. After injection contrast medium 0.5 to 1 mL to produce a neurogram that identified the nerve root in question, either Solomet (methylprednisolone 40 mg/mL– bupivacaine 5 mg/mL) or isotonic (0.9%) sodium chloride solution was in-

Efficacy Analysis. Follow-up assessments were conducted 2 weeks, 1 month, 3 months, 6 months, and 1 year after the intervention. The primary outcome was leg pain, but back pain and disability were recorded also at every follow-up assessment. Analysis of variance (ANOVA) was performed to compare the treatments. The demographic and clinical characteristics of the two treatment groups at baseline differed only in terms of the affected level on MRI and days of sick leave before the intervention.21 These differences were considered clinically important confounders, and the scores were adjusted accordingly. For the different outcomes, the adjusted change in scores from baseline was calculated, and between-group treatment differences were determined by subtracting the adjusted scores of the saline group from those of the steroid group. Additionally, the number and percentage of responders (i.e., patients free of leg pain, who had experience at least a 75% decrease in leg pain from the baseline scores) was determined at each follow-up assessment for both treatments. Patients who underwent surgery were always regarded as nonresponders. The statistical significance of differences was evaluated with Fisher’s exact two-sided test. Between-group changes over time (efficacy) were analyzed with the area-under-the-curve (AUC) method.1 AUC scores were adjusted with symptomatic disc level, days of sick leave before the intervention, and the baseline value of the respective outcome. The scores were calculated separately from baseline to 3 months, and from 3 months to 1 year.

Medical Costs and Cost-Effectiveness Analysis. The economic analysis included the costs for the use of health care services and help at home. Data about days of sick leave and costs of medicines were obtained from the National Insurance Register. Any additional treatments and visits to physiotherapists, osteopaths, and physicians were gathered from the questionnaires and medical records, as were the data on back operations. The analysis of costs for health services and home help was based on the list prices and the responses to the 1-month, 3-month, 6-month, and 1-year follow-up questionnaires.21 Treatment differences and AUC scores for cumulative 3- and 12-month costs were calculated at each follow-up assessment. To obtain a cost-effectiveness estimate for the treatments, the cumulative costs at 3 and 12 months were divided by the number of responders at the respective time point. The figures obtained from the steroid and saline treatments were compared using Student’s t test. Subgroup Analysis. Efficacy and cost effectiveness were evaluated with respect to duration of symptoms, age, symptomatic disc level, and the MRI classification. The age of patients had no significant effect on the treatment differences. Duration of symptoms modified the medical costs at 6 months so that the patients with shorter symptom duration (ⱕ 2 months) had greater monetary savings with steroid. The AUC scores for medical costs were therefore adjusted also for symptom duration. For symptom duration, a categorical staging was used: 1

Cost Effectiveness of Periradicular Infiltration • Karppinen et al 2589

Table 1. Baseline Characteristics, According to MRI Classification, and Symptomatic Disc Level of 160 Patients With Sciatica Randomly Assigned to Receive Periradicular Infiltration With Either Steroid (Methylprednisolone–Bupivacaine) or Saline* MRI Classification

Bulges Outcome Measures Patients (n) Age (years) Symptom duration (months) Male gender (%) Intensity of leg pain (VAS) Oswestry disability (%) Nottingham Health Profile‡ Pain dimension Emotional reaction Straight-leg-raising test (°)

Contained Herniations

Disc Level

Extrusions

L3–4/L4–5

L5–S1

Steroid

Saline

Steroid

Saline

Steroid

Saline

Steroid

Saline

Steroid

Saline

18 52 ⫾ 13 2.5 ⫾ 1.4

11 47 ⫾ 15 2.8 ⫾ 1.6

24 43 ⫾ 13 2.5 ⫾ 1.3

26 42 ⫾ 14 2.5 ⫾ 1.4

38 42 ⫾ 12 2.3 ⫾ 1.6

43 43.3 ⫾ 11 2.6 ⫾ 1.5

51 45 ⫾ 11 2.2 ⫾ 1.2

36 46 ⫾ 14 2.5 ⫾ 1.6

29 43 ⫾ 14 2.6 ⫾ 1.6

44 40 ⫾ 11 2.7 ⫾ 1.2

45 81 ⫾ 11 40 ⫾ 17

33 73 ⫾ 18 43 ⫾ 15

73 69 ⫾ 18 39 ⫾ 12

71 79 ⫾ 12† 44 ⫾ 14

63 70 ⫾ 18 46 ⫾ 17

61 74 ⫾ 23 43 ⫾ 15

75 69 ⫾ 18 43 ⫾ 16

63 75 ⫾ 20 44 ⫾ 14

55 72 ⫾ 17 43 ⫾ 15

48 76 ⫾ 18 43 ⫾ 16

62 ⫾ 27 10 ⫾ 17 75 ⫾ 15

71 ⫾ 21 7 ⫾ 10 77 ⫾ 18

65 ⫾ 24 8 ⫾ 14 56 ⫾ 16

75 ⫾ 25 14 ⫾ 16 54 ⫾ 16

68 ⫾ 28 12 ⫾ 15 54 ⫾ 17

71 ⫾ 27 9 ⫾ 13 54 ⫾ 17

65 ⫾ 29 10 ⫾ 13 57 ⫾ 17

74 ⫾ 23 8 ⫾ 12 61 ⫾ 20

67 ⫾ 25 11 ⫾ 16 58 ⫾ 18

69 ⫾ 28 13 ⫾ 16 57 ⫾ 17

* Means ⫾ standard deviation shown unless otherwise stated. † P ⬍ 0.05 (Kruskal-Wallis). ‡ The scores on the Nottingham Health Profile range from 0 to 100, with higher scores indicating worse health-related quality of life in each dimension. The scores for energy, sleep, mobility and social isolation dimensions are not presented here or in the follow-up data. MRI ⫽ magnetic resonance imaging; VAS ⫽ visual analog scale.

month or less, 1.5 to 2.5 months, and 3 months or more. The rate of operations in different subgroups was evaluated by Kaplan-Meier curves and log-rank tests.1 All P values less than 0.05 were considered statistically significant. For the statistical analysis, SPSS Version 8.0 was used.

mentioned efficacy in leg pain, the number of responders with leg pain at 2 weeks significantly favored the steroid treatment (Table 3). For extrusions, leg pain at 6 months was significantly in favor of the saline group (Figure 1B; Table 2).

Results Study Population In the final study population, 80 patients were randomized to receive saline or 80 methylprednisolone– bupivacaine injection. Table 1 presents the baseline characteristics of the subgroups. Except for contained herniations, where the intensity of leg pain was significantly higher in the saline group, the baseline characteristics were similar in the two treatment groups (Table 1). Symptomatic L4 –L5 and L3–L4 were combined in the further analyses because so few (6 patients) had a lesion at L3–L4. Between-Group Differences in Subgroups For bulges, no clinically significant between-group differences were observed between the treatments. The treatments also were similar in terms of AUC scores, number of painless patients, and medical costs. Contained Herniation Versus Extrusions. In the case of contained herniations, the methylprednisolone– bupivacaine injection was significantly superior for leg pain at 2 and 4 weeks (Figure 1A; Table 2) and for NHP emotional reactions at 3 months (Table 2). At 6 months, leg pain, disability, and NHP emotional reactions were in favor of saline (Table 2). The AUC scores indicated that the steroid option was superior to saline for leg pain (P ⫽ 0.042) and NHP emotional reactions (P ⫽ 0.025) from baseline to 3 months. In accordance with the afore-

L3–L4/L4 –L5 Versus L5–S1 For symptomatic lesions at L3–L4 –L5, steroid treatment was superior to saline in terms of leg pain at 2 and 4 weeks, disability at 2 and 4 weeks, and straight leg raising at 2 and 4 weeks (Table 2). The AUC scores from baseline to 3 months for leg pain (P ⫽ 0.012), disability (P ⫽ 0.043), and straight leg raising (P ⫽ 0.026) also were significantly in favor of steroid. Similarly, the number of painless patients at 4 weeks significantly favored the steroid injection (Table 3). The treatment difference for NHP emotional reactions almost reached statistical significance in favor of the steroid group (Table 2). At L5–S1, the between-group treatment difference at 1 year (P ⫽ 0.031) and the AUC score at 3 to 12 months (P ⫽ 0.006) for NHP emotional reactions were in favor of saline, but otherwise the treatments were similar. Medical Costs and Sick Leaves Contained Herniations Versus Extrusions. For contained herniations, the steroid intervention produced savings in the need for home care at 4 weeks ($200 per patient; 95% CI, $46 –355; P ⫽ 0.013) and in total costs at 6 months ($1795; 95% CI, $1069 –2521; P ⬍ 0.001). Medical costs at each follow-up assessment are presented in Figure 2A. By 1 year, 42% of patients in the saline group underwent surgery, as compared with 20% of the

2590 Spine • Volume 26 • Number 23 • 2001

Figure 1. Leg pain (millimeters on visual analog scale; standard deviations indicated with vertical bars) at baseline and at each follow-up assessment after the nerve root infiltration with either methylprednisolone– bupivacaine (— ● —) or saline (— 䡲 —). The upper box presents the number of patients at each follow-up assessment. A, Subgroup of contained herniations. B, Subgroup of extrusions. *The P value of the between-group treatment difference at the respective follow-up visit was less than 0.05. **P ⬍ 0.01.

patients in the steroid group (␹2 ⫽ 2.7; df ⫽ 1; P ⫽ 0.1). When the rate of operations was evaluated by KaplanMeier curves and log-rank tests, a P value of 0.1 was obtained (Figure 3). Because of the different operation rate, the days of sick leave at 6 months (7.4 days per month per patient; 95% CI, 2.3–12.5 days; P ⫽ 0.006; Figure 4), and cumulative costs at 12 months ($1969 per patient; 95% CI, $590 –2914; P ⫽ 0.007) were in favor of steroid. The AUC scores of medical costs at 3 to 12 months also favored the steroid treatment (P ⬍ 0.001) in accordance with the other data.

For extrusions, the cost of therapy visits at 4 weeks was significantly less with the steroid injection ($182; 95% CI, $79 –285; P ⫽ 0.001). Medical costs at each follow-up assessment are presented in Figure 2B. By 1 year, 13% of patients in the saline group underwent surgery, as compared with 32% in the steroid group (␹2 ⫽ 3.9; df ⫽ 1; P ⫽ 0.05). According to Kaplan-Meier analysis, the P was 0.1 (Figure 3). Because of the higher operation rate in the steroid group, the AUC scores of medical costs at 3 to 12 months were significantly in favor of saline (P ⫽ 0.004), whereas for cumulative costs at 12

Cost Effectiveness of Periradicular Infiltration • Karppinen et al 2591

Table 2. Between-Group Treatment Differences After Periradicular Infiltration With Steroid or Saline in Subgroups of Contained Herniations, Extrusions and Disc Levels L3–L4 –L5* MRI Classification† Contained Herniations Difference (95% CI) 2 Weeks Leg pain (100-mm VAS) Disability (Oswestry %) Straight leg raising (°) 4 Weeks Leg pain (100-mm VAS) Disability (Oswestry %) NHP pain NHP emotional reactions Straight leg raising (°) 3 Months Leg pain (100-mm VAS) Disability (Oswestry %) NHP pain NHP emotional reactions Straight leg raising (°) 6 Months Leg pain (100-mm VAS) Disability (Oswestry %) NHP pain NHP emotional reactions Straight leg raising (°) 1 Year Leg pain (100-mm VAS) Disability (Oswestry %) NHP pain NHP emotional reactions Straight leg raising (°)

24.4 (8–41)

Extrusions

Disc Level L3–L4/L4–L5

P

Difference (95% CI)

P

Difference (95% CI)

P

0.006

10.8 (⫺5–26)

NS

25.0 (10–40)

0.002

9.6 (2–17)

0.009

8.0 (⫺0.3–16)

NS

7.7 (⫺0.4–16)

NS

7 (⫺1–16)

NS

6 (⫺3–15)

NS

0.023

5.6 (⫺10–22)

NS

20.4 (5–35)

0.008

2.7 (⫺5–10)

NS

8.8 (⫺0.4–18)

NS

8.8 (1–17)

0.027

12.8 (⫺9–35) 4.9 (⫺6–16)

NS NS

5.6 (⫺10–21) 4.5 (⫺5–14)

NS NS

13.2 (⫺3–29) 7.1 (⫺2–16)

NS NS

9 (⫺2–19)

NS

5 (⫺5–14)

NS

8 (4–22)

0.006

1.4 (⫺20–23)

NS

⫺3.3 (⫺19–12)

NS

2.5 (⫺14–19)

NS

⫺2.3 (⫺13–9)

NS

2.7 (⫺8–14)

NS

3.4 (⫺7–13)

NS

⫺5.1 (⫺27–17) 13.3 (4–23)

NS 0.008

⫺0.4 (⫺18–17) ⫺2.2 (⫺9–5)

NS NS

⫺0.4 (⫺19–19) 5.2 (⫺0.6–14)

NS NS

NS

5.8 (⫺4–15)

NS

19.4 (3–36)

⫺2 (⫺12–10)

NS

3 (⫺8–13)

9 (0.3–18)

0.043

⫺22.5 (⫺40–⫺5)

0.014

⫺16.6 (⫺32–⫺1)

0.033

⫺8.4 (⫺23–6)

NS

⫺13.5 (⫺24–⫺3)

0.01

⫺1.0 (⫺11–9)

NS

⫺1.5 (⫺10–7)

NS

⫺21.6 (⫺43–⫺0.3) ⫺3.2 (⫺13–7)

0.047 NS

⫺8.2 (⫺25–9) 2.7 (⫺5–10)

NS NS

⫺6.8 (⫺23–9) 3.0 (⫺4–10)

NS NS

⫺9 (⫺21–3)

NS

⫺1 (⫺11–8)

NS

1 (⫺9–10)

NS

⫺0.3 (⫺16–16)

NS

⫺7.5 (⫺22–7)

NS

6.6 (⫺7–20)

NS

⫺1.2 (⫺12–9)

NS

3.7 (⫺6–13)

NS

4.3 (⫺4–13)

NS

0.1 (⫺22–22) ⫺3.2 (⫺13–7)

NS NS

⫺4.7 (⫺21–11) 2.7 (⫺5–10)

NS NS

⫺0.8 (⫺17–15) 1.9 (⫺4–8)

NS NS

NS

⫺1 (⫺10–8)

NS

4 (⫺7–16)

7 (⫺2–16)

NS

* Positive treatment difference values indicate that the methylprednisolone– bupivacaine treatment was superior to the saline injection. Numbers of patients at the 2-week, 4-week, 3-month, 6-month, and 1-year follow-up assessments were 24/26, 23/26, 24/25, 24/25, and 24/25 (saline/steroid) for contained herniations; 38/43, 37/43, 38/42, 38/42 and 38/42, respectively; and 51/36, 51/36, 51/35, 51/35 and 51/35, respectively, for the combined L3– 4 and L4 –5 levels. † The between-group mean values of the difference between the adjusted change from baseline in the methylprednisolone– bupivacaine and saline groups. The values are adjusted to the level of the symptomatic disc and days on sick leave before the intervention. MRI ⫽ magnetic resonance imaging; CI ⫽ confidence interval; VAS ⫽ visual analog scale; NS ⫽ not significant; NHP ⫽ Nottingham Health Profile.

months, only a trend in favor of saline existed (P ⫽ 0.08). No significant differences in sick leaves were observed. L3–L4/L4 –L5 Versus L5–S1. The 4-week follow-up assessment showed steroid treatment to be superior to saline treatment in terms of need for home care at L3–L4 –L5 ($120; 95% CI, $31–209; P ⫽ 0.01). No significant differences in cumulative medical costs, rate of operations, or work absenteeism were observed between the treatments.

Cost Effectiveness of the Treatments No significant differences in cost effectiveness were observed between the treatments in terms of symptomatic disc level or intention-to-treat analysis. When the treat-

ments were analyzed according to the MRI classification, no significant short-term differences (at 3 months) were observed in cost effectiveness. Statistically speaking, however, the steroid option was significantly cost-effective for contained herniations and countereffective for extrusions at 1 year. For contained herniations, it cost $12,666 more per patient to obtain one painless patient in the saline group. For extrusions, the steroid treatment was more expensive: $4445 more per painless patient (Table 4). Discussion Previously, the authors showed that periradicular infiltration with steroid had a short-term effect for sciatica,

2592 Spine • Volume 26 • Number 23 • 2001

Table 3. Patients Free of Leg Pain After Periradicular Infiltration With Steroid or Saline in Some Subgroups* MRI Classification Contained Herniations

Extrusions

Disc Level L3–L4/L4–L5

Responders

Steroid

Saline

P

Steroid

Saline

P

Steroid

Saline

P

2 Weeks 4 Weeks 3 Months 6 Months 12 Months

9/26 (35) 8/26 (31) 6/25 (24) 6/25 (24) 11/25 (44)

2/23 (9) 2/24 (8) 7/24 (29) 8/24 (33) 5/24 (21)

0.042 NS NS NS NS

11/43 (26) 20/43 (47) 20/43 (47) 15/42 (36) 15/42 (36)

8/38 (21) 15/38 (39) 21/37 (57) 22/38 (58) 22/37 (59)

NS NS NS NS NS

13/36 (36) 21/36 (58) 17/35 (49) 15/35 (43) 18/35 (51)

9/51 (18) 16/51 (31) 24/50 (48) 28/51 (55) 24/51 (47)

NS 0.016 NS NS NS

* Responders had a 75% or more decrease in leg pain from baseline scores. Patients who underwent surgery were always regarded as nonresponders. Figures present the number of treatment responders in the total number of patients. The number in parentheses shows the corresponding percentage of responders in the respective subgroup. Statistical significance was evaluated by Fisher’s exact test. MRI ⫽ magnetic resonance imaging; NS ⫽ not significant.

whereas it was countereffective at 6 months.21 The current subgroup analysis showed that the short-term effect of the steroid treatment was most pronounced for contained herniations and symptomatic lesions situated at L4 –L5 or L3–L4. Patients receiving the steroid treatment for a contained herniation were less likely to undergo back surgery, and they also had significantly fewer days on sick leave at 3 to 6 months. Countereffectiveness was most pronounced for extrusions, for which the steroid injection generated significantly higher medical costs and a greater likelihood of surgery. The short-term efficacy of the methylprednisolone– bupivacaine injection, as compared with saline, at L4 –L5 was evident in every disease-specific measure (leg pain, disability by Oswestry, and straight-leg-raising restriction), whereas in the case of contained herniations, efficacy was seen only in leg pain and one dimension of the generic NHP questionnaire. The Oswestry and Million scales have often been used as outcomes in longitudinal before and after studies,25 yet the Oswestry disability questionnaire may be less sensitive than the Roland-Morris questionnaire.4 Another possible explanation for the smaller disease-specific effect in contained herniations is the smaller patient number in this subgroup. Periradicular infiltration with steroid seemed to prevent surgery in contained herniations, suggesting the effectiveness of steroid treatment in this subgroup. Corticosteroids may calm the inflammatory process in many ways.2,8 In the present study, 42% of the patients with contained herniation underwent surgery after the saline injection, as compared with 20% after the steroid injection. The patients who underwent surgery were almost entirely painless after surgery, which may explain the between-group differences in favor of saline at 6 months. For contained herniations, repeated steroid injections could be recommended, perhaps 2 to 4 weeks after the index injection, because the steroid treatment does not increase the rate of disc operations and is effective in diminishing leg pain.

An abundance of macrophages is found in disc herniations. They are thought to play a role in the resorption of herniations.16,17 Macrophages are more prominent in extrusions than in nonextruded herniations.15,18,27 Mononuclear cells infiltrate along the margins of extruded discs, expressing inflammatory mediators. In cocultures with endothelial cells, disc cells from extrusions enhanced the proliferation of endothelial cells and fibroblasts significantly more than disc cells from protrusions.11 It may be that corticosteroids have some detrimental effect on the function of macrophages. In fact, in a rabbit model, high-dose steroid suppressed the replacement of grafted intervertebral disc tissue, consistent with the current results.28 Our subgroup analysis showed that the steroid injection seemed to be more harmful for extrusions, which accords with the observation that macrophages are more abundant in extruded disc fragments. High-quality cost-effectiveness and cost-utility analyses are scarce in low back pain research.14 To the authors’ knowledge, the current economic and costeffectiveness analysis of periradicular infiltration is the first to be conducted. The steroid intervention had already produced some monetary savings for contained herniations by 4 weeks, but by 1 year, it had saved $1969 per patient treated. For extrusions, steroid seemed to increased the 1-year costs. The disparity of operation rates in these two subgroups explains the significant differences in medical costs, as well as the decrease in days of sick leave for contained herniations at 6 months in the steroid group. The cost-effectiveness analysis indicates that for contained herniations, the steroid treatment was decidedly more cost effective than saline, with a difference of more than $12,600 per painless patient. The contrast might have been even greater with dry-needling used as a placebo. Conversely, for extrusions, the cost-effectiveness analysis suggests that a better alternative to saline than steroid needs to be sought, one possibility being tumor necrosis factor-␣ antagonist.29 Indirect costs were not

Cost Effectiveness of Periradicular Infiltration • Karppinen et al 2593

Figure 2. Medical costs (dollars; standard deviations indicated with vertical bars) at each follow-up assessment after periradicular infiltration with either methylprednisolone– bupivacaine (— ● —) or saline (— 䡲 —). The upper box presents the number of patients at each follow-up assessment. A, Subgroups of contained herniations. B, Extrusions. * The P value of the between-group treatment difference at the respective follow-up visit was less than 0.05. **P ⬍ 0.01.

included in the economic analysis because return to work is less responsive to clinical treatment than symptoms or daily functioning, although it is of utmost social and personal importance.10 Also, the monetary assessment of work absenteeism is controversial.20 It has been stated that it is essential to identify homogeneous groups of patients with low back pain, and that the efficacy of interventions in these subgroups should be studied in randomized controlled trials.4 However, subdividing patients with sciatica into homogeneous strata

has the disadvantage of poorer generalizability of results to heterogeneous populations.3 This subgroup analysis of a periradicular infiltration trial suggests that for contained herniation, steroid is clearly superior to saline in terms of both leg pain and medical costs, and possibly also in the need for operative treatment. In addition, if the lesion is located at L3–L4 or L4 –L5, steroid treatment is more likely to achieve good results in terms of disease-specific outcomes, but not in terms of medical costs. For extrusions, steroid seems to be countereffec-

2594 Spine • Volume 26 • Number 23 • 2001

Figure 3. Kaplan-Meier curves for the number of patients who underwent surgery, as compared with the time of operation, in weeks after periradicular infiltration in subgroups of contained herniations and extrusions. Curves are presented separately for both treatments in each subgroup.

Figure 4. Mean days of sick leave per month in the subgroup of contained herniations at each follow-up assessment after periradicular infiltration with either methylprednisolone– bupivacaine (— ● —) or saline (— 䡲 —). The figures at baseline present the mean days of sick leave before the intervention.

Table 4. Mean Cumulative Costs of Periradicular Infiltration per One Responder* MRI Classification Bulges

3 Months 12 Months

Contained Herniations

Extrusions

Steroid $

Saline $

P

Steroid $

Saline $

P

Steroid $

Saline $

P

2640 3740

2116 3629

NS NS

5850 4432

6360 17,098

NS 0.0073

4081 7165

2230 2484

NS 0.0058

* Leg pain decreased in 75% or more. Patients who underwent surgery were always regarded as nonresponders. Statistical significance was evaluated by Student’s t test. MRI ⫽ magnetic resonance imaging; NS ⫽ not significant.

Cost Effectiveness of Periradicular Infiltration • Karppinen et al 2595

tive. The promising results in the subgroup analyses call for a verification study. Key Points Periradicular infiltration with a steroid– bupivacaine combination has short-term efficacy in sciatic leg pain from to contained herniations. ● Steroid treatment seems to prevent discectomies and to produce long-term savings in medical costs for contained herniations. ● Steroid injection is cost effective for contained herniations, producing savings of $12,600 per responder. For extrusions, steroid seems to be countereffective. ● When the symptomatic lesion is situated at L3–L4 or L4 –L5, steroid treatment is effective in terms of several disease-specific questionnaires, but not in terms of medical costs. ●

References 1. Altman DG. Practical Statistics for Medical Research. London: Chapman and Hall, 1991:431–3. 2. Arantes RM, Lourenssen S, Machado CR, et al. Early damage of sympathetic neurons after coculture with macrophages: a model of neuronal injury in vitro. Neuroreport 2000;11:177– 81. 3. Bloch R. Methodology in clinical back pain trials. Spine 1987;12:430 –2. 4. Bouter LM, van Tulder MW, Koes BW. Methodologic issues in low back pain research in primary care. Spine 1998;23:2014 –20. 5. Clark RE. Spine update: understanding cost effectiveness. Spine 1996;21: 646 –50. 6. Cohran WG, Cox GM. Experimental Designs. New York: Wiley, 1957. 7. Conrad DA, Deyo RA. Economic decision analysis in the diagnosis and treatment of low back pain: a methodologic primer. Spine 1994;19:2101S– 6S. 8. Cupps TR, Fauci AS. Corticosteroid-mediated immunoregulation in man. Immunol Rev 1982;65:133–55. 9. Derby R, Kine G, Saal JA, et al. Response to steroid and duration of radicular pain as predictors of surgical outcome. Spine 1992;17:S176 – 83. 10. Deyo RA, Battie M, Beurskens AJ, et al. Outcome measures for low back pain research: a proposal for standardized use. Spine 1998;23:2003–13. 11. Doita M, Kanatani T, Harada T, et al. Immunohistologic study of the ruptured intervertebral disc of the lumbar spine. Spine 1996;21:235– 41. 12. Fairbank JC, Couper J, Davies JB, et al. The Oswestry low back pain disability questionnaire. Physiotherapy 1980;66:271–3. 13. Frymoyer JW. Back pain and sciatica. N Engl J Med 1988;318:291–300. 14. Goossens M, Evers S. Cost effectiveness of treatment for neck and low back pain. In: Nachemson A, Jonsson E, eds. Neck and Back Pain: The Scientific Evidence of Causes, Diagnosis, and Treatment. Philadelphia: Lippincott Williams & Wilkins, 2000:399 – 419. 15. Gro¨ nblad M, Virri J, Tolonen J, et al. A controlled immunohistochemical study of inflammatory cells in disc herniation tissue. Spine 1994;19: 2744 –51.

16. Habtemariam A, Gro¨ nblad M, Virri J, et al. A comparative immunohistochemical study of inflammatory cells in acute-stage and chronic-stage disc herniations. Spine 1998;23:2159 – 65. 17. Haro H, Komori H, Okawa A, et al. Sequential dynamics of monocyte chemotactic protein-1 expression in herniated nucleus pulposus resorption. J Orthop Res 1997;15:734 – 41. 18. Haro H, Shinomiya K, Komori H, et al. Upregulated expression of chemokines in herniated nucleus pulposus resorption. Spine 1996;21:1647–52. 19. Helakorpi S, Uutela A, Pra¨ tta¨ la¨ R, et al. Health behaviour and health among Finnish adult population, spring 1999. Helsinki: Publications of the Finnish National Public Health Institute 1999:B19. 20. Hutubessy RC, van Tulder MW, Vondeling H, et al. Indirect costs of back pain in the Netherlands: a comparison of the human capital method with the friction cost method. Pain 1999;80:201–7. 21. Karppinen J, Malmivaara A, Kurunlahti M, et al. Periradicular infiltration for sciatica: a randomized controlled trial. Spine 2001;26:1059 – 67. 22. Karppinen J, Malmivaara A, Tervonen O, et al. Severity of symptoms and signs in relation to magnetic resonance imaging findings among sciatic patients. Spine 2001;26:E149 –54. 23. Keegan JJ, Garret FD. The segmental distribution of the cutaneous nerves in the limbs of man. Anat Rec 1948;102:409 –37. 24. Koivukangas P, Ohinmaa A, Koivukangas J. Nottingham Health Profilen (NHP) suomalainen versio (Finnish version of the Nottingham Health Profile). Reports of the National Research and Development Center for Welfare and Health. Helsinki, Finland, 1995. 25. Kopec JA, Esdaile JM. Functional disability scales for back pain. Spine 1995; 20:1943–9. 26. Lurie JD, Sox HC. Principles of medical decision making. Spine 1999;24: 493– 8. 27. Matsui Y, Maeda M, Nakagami W, et al. The involvement of matrix metalloproteinases and inflammation in lumbar disc herniation. Spine 1998;23: 863– 8. 28. Minamide A, Tamaki T, Hashizume H, et al. Effects of steroid and lipopolysaccharide on spontaneous resorption of herniated intervertebral discs: an experimental study in the rabbit. Spine 1998;23:870 – 6. 29. Olmarker K, Larsson K. Tumor necrosis factor alpha and nucleus pulposus– induced nerve root injury. Spine 1998;23:2538 – 44. 30. Russell LB, Gold MR, Siegel JE, et al. The role of cost-effectiveness analysis in health and medicine. Panel on Cost Effectiveness in Health and Medicine. JAMA 1996;276:1172–7. 31. van Akkerveeken PF. Pain patterns and diagnostic blocks. In: Wiesel SW, Weinstein JN, Herkowitz H, et al, eds. The Lumbar Spine. Philadelphia: WB Saunders, 1996:105–22. 32. Weinstein SM, Herring SA, Derby R. Contemporary concepts in spine care: epidural steroid injections. Spine 1995;20:1842– 6.

Address reprint requests to Jaro Karppinen, MD Department of Physical Medicine and Rehabilitation Oulu University Hospital Kajaanintie 50 FIN-90220 Oulu Finland E-mail: [email protected]

Suggest Documents