Documente Academic
Documente Profesional
Documente Cultură
DOI 10.1007/s00586-015-3891-4
REVIEW ARTICLE
123
Eur Spine J
with bed rest. For persistent LBP, evidence suggests that: low back pain [25, 26]. For lumbar radiculopathy, two
(1) oral NSAIDs are more effective than placebo; and (2) systematic reviews reported that NSAIDs are not more
oral NSAIDs may be more effective than acetaminophen. effective than placebo [27, 28].
For recent-onset LBP with radiculopathy, there is incon- Different methodologies used to conduct the previous
sistent evidence on the effectiveness of oral NSAIDs versus reviews likely contributed to the differences in their con-
placebo. Finally, different oral NSAIDs lead to similar clusions. The methodologies differed in the application of
outcomes for neck and LBP with or without radiculopathy. critical appraisal and evidence synthesis methods. Across
Conclusions For NAD, oral NSAIDs may be more ef- reviews, different instruments were used to critically ap-
fective than placebo. Oral NSAIDs are more effective than praise the literature and synthesize the evidence [21–28].
placebo for persistent LBP, but not for recent-onset LBP. Inclusion of studies with a high risk of bias may threaten
Different oral NSAIDs lead to similar outcomes for neck the internal validity of a systematic review. Randomized
pain and LBP. controlled trials (RCTs) with poor internal validity were
used in the evidence synthesis of six reviews [21, 25, 26,
Keywords Whiplash-associated disorders Neck pain 29–31]. Considering the varying methodologies used in
and associated disorders Nonspecific low back pain previous systematic reviews, an in-depth review of their
Nonsteroidal anti-inflammatory drugs Systematic review quality is warranted prior to accepting their results.
Medication We conducted a systematic review of systematic re-
views to examine the effectiveness of NSAIDs for the
management of neck pain and associated disorders (NAD),
Introduction whiplash-associated disorders (WAD), or non-specific low
back pain with or without radiculopathy.
Neck and back pain are experienced by more than 80 % of
the population during their lifetime, and are common rea-
sons for seeking healthcare [1–6]. These conditions are Methods
costly, and work absenteeism and lost productivity are
frequent complications [7–9]. Most neck and back pain are Registration
common sources of disability, whether attributed to work,
traffic collision, or daily activities [1, 3, 10–13]. To date, We registered our protocol with the International
few interventions demonstrate efficacy on a scientifically Prospective Register of Systematic Reviews (PROSPERO)
valid basis and most interventions are only associated with on May 14, 2014 (CRD42014009782).
short-term benefits [14, 15].
Non-steroidal anti-inflammatory drugs (NSAIDs) are Eligibility criteria
commonly used for neck and low back pain. NSAIDs
suppress inflammatory processes that produce swelling and Population
pain [16]. In the United States, NSAIDs are the most
common intervention prescribed for non-malignant chronic Our review targeted systematic reviews of adults or chil-
pain, including musculoskeletal disorders, with use re- dren with NAD or WAD (grades I–III), or non-specific low
ported in approximately 95 % of visits to physicians [17]. back pain (with or without radiculopathy) and excluded
In the United Kingdom, general practitioners prescribed pain from major pathology (fractures, dislocations, spinal
NSAIDs to approximately 30 % of patients with chronic cord injury, infection, neoplasms, systemic disease). We
low back pain [18]. In Spain, 41 % of patients with low defined NAD and WAD according to previous classifica-
back pain reported taking NSAIDs for more than 1 month tions [32, 33] (Table 1). We defined non-specific low back
in the past year [19]. pain as low back pain with or without radiculopathy in the
Despite its common use, the effectiveness of NSAIDs absence of specific pathological entities [24].
for the management of neck and low back pain is not firmly
established [20–22]. Several systematic reviews examining Intervention
NSAIDs for neck and low back pain have reported limited
or unclear evidence [20–22]. In contrast, two systematic We restricted our review to systematic reviews that tested
reviews concluded (based on low quality evidence and the effectiveness of NSAIDs. NSAIDs are defined as
small effect sizes) that NSAIDs are effective for short-term medications that block the action of cyclo-oxygenase
symptomatic relief in patients with low back pain [23, 24]. (Cox)-1 and/or Cox-2 (necessary for prostaglandin pro-
Furthermore, two reviews found evidence supporting the duction), thus reducing inflammation and pain [16]. Cox-1
use of NSAIDs in individuals with acute [25] or chronic enzymes are also cyto-protective; blocking their action is
123
Eur Spine J
Table 1 Classification of grades for neck pain and associated dis- systematic review of studies (with or without meta-analy-
orders [32] and whiplash-associated disorders [33] sis); and (4) adults or children with NAD, WAD, or non-
Grade Definition specific low back pain with or without radiculopathy
(mixed populations must have stratified results for NAD,
Neck pain and associated disorders
WAD, or non-specific low back pain).
I No signs or symptoms suggestive of major structural We excluded: (1) guidelines, letters, editorials, com-
pathology and no or minor interference with activities of
daily living mentaries, unpublished manuscripts, dissertations, govern-
II No signs or symptoms of major structural pathology, ment reports, books, conference proceedings, meeting
but major interference with activities of daily living abstracts, lectures and addresses, consensus development
III No signs or symptoms of major structural pathology, but statements; (2) non-systematic reviews, RCTs, cohort
presence of neurologic signs such as decreased deep tendon studies, case–control studies, cross-sectional studies, case
reflexes, weakness or sensory deficits reports/series, qualitative studies, biomechanical studies,
IVa Signs or symptoms of major structural pathology laboratory studies, studies not reporting on methodology;
Whiplash-associated disorders (3) cadaveric or animal studies; (4) reviews on patients
I Subjects with neck pain and associated symptoms in the with severe injuries (spinal cord injuries, amputations,
absence of objective physical signs
open wounds, grade III sprains/strains, fractures, disloca-
II Subjects with neck pain and associated symptoms in the
tions); (5) previous versions of updated systematic reviews;
presence of objective physical signs and without evidence
of neurological involvement (6) systematic reviews of systematic reviews; and (7)
III Subjects with neck pain and associated symptoms with systematic reviews that do not stratify the analysis by study
evidence of neurological involvement including decreased design (e.g., combine systematic reviews and primary
or absent reflexes, decreased or limited sensation, or studies in their synthesis).
muscular weakness
IVa Subjects with neck pain and associated symptoms with Information sources
evidence of fracture or dislocation
a
Grade IV was excluded from this review We developed our search strategy with a health sciences
librarian, which was reviewed by a second librarian using
associated with gastrointestinal ulcers and bleeds. Most the Peer Review of Electronic Search Strategies (PRESS)
NSAIDs act non-selectively on Cox-1 and Cox-2 en- Checklist [34]. We systematically searched MEDLINE,
zymes, whereas selective Cox-2 inhibitors are preferen- EMBASE, CINAHL, PsycINFO, DARE, and Cochrane
tially directed at Cox-2 sites. Possible routes of Database of Systematic Reviews from January 1, 2000 to
administration include: topical (local effect or systemic April 8, 2014. Search terms consisted of subject headings
absorption), intra-nasal, intra-aural, eye drop, subcuta- specific to each database and free text words relevant to
neous, oral (ingestion), oral (sublingual), rectal, vaginal, systematic review, NSAIDs, NAD, WAD, non-specific low
and intramuscular. back pain, and lumbar radiculopathy (‘‘Appendix I’’). We
used EndNote X6 to create a bibliographic database to
Comparison groups manage the search results [35].
123
Eur Spine J
methodological quality of systematic reviews [37]. Re- analysis, we extracted effect sizes from individual studies
viewers reached consensus on the overall assessment of the (where reported). We also extracted results on adverse
methodological quality through discussion. We involved an events from the reviews.
independent third reviewer if consensus could not be Some reviews with a low risk of bias may reach dif-
reached. We contacted authors when additional informa- ferent conclusions even though they included some of the
tion was required to complete the critical appraisal. same primary studies. In this case, we examined the rea-
During critical appraisal, we assessed for the presence sons for these differences (e.g., differences in search date,
of selection bias and information bias, and any impact inclusion or exclusion criteria, threshold for determining
these may have on the internal validity of the review. We which primary studies are included in evidence synthesis,
did not use a rating scale cut-off or quantitative score to interpretation of results using statistical significance versus
judge the quality of the review. Instead, we focused on clinical importance) and described them in our results.
the presence or absence of important methodological is- Furthermore, we: (1) used the results of the review with
sues. Systematic reviews were considered to have a high higher methodology if possible; or (2) made a consensus
risk of bias if reviewers considered the internal validity decision on which results would be adopted in cases where
markedly compromised due to biases and methodological methodological quality of both reviews was similar. We
flaws. Systematic reviews with adequate internal validity placed emphasis on systematic reviews that demonstrated
and a low risk of bias were included in our evidence methodological rigor and had the more recent literature
synthesis. search date.
The lead author extracted data from systematic reviews We computed agreements between reviewers for the
with a low risk of bias and built evidence tables (Tables 3, screening of articles and reported the kappa statistic (k) and
4). A second reviewer independently checked the extracted 95 % confidence interval (CI) [39]. We also calculated the
data. We performed a qualitative synthesis of findings from percentage agreement for the independent critical appraisal
reviews with low risk of bias to develop evidence state- of articles for scientifically admissible/inadmissible results.
ments according to principles of best evidence synthesis Where available, we used minimal clinically important
[38]. For each systematic review with low risk of bias, we differences (MCIDs) to judge the clinical importance of
extracted data from studies with low and high risk of bias differences in outcomes between groups: 10/100 mm for
(as determined by the authors of the systematic reviews); the Visual Analog Scale [40]; 2/10 for the Numeric Rating
however, we included only studies with a low risk of bias Scale [41]; 5/50 for the Neck Disability Index [42]; 5/24
in our evidence synthesis. For systematic reviews with a for the Roland–Morris Disability Questionnaire [43];
meta-analysis, we extracted only clinically and statistically 10/100 for the Oswestry Disability Index [43]; 7.7 points
homogeneous pooled effect sizes (e.g., weighted mean for the Physical Component Summary and 10 points for
difference). For systematic reviews without a meta- Bodily Pain on the Short-Form-36 [44].
123
Eur Spine J
Reporting
Table 3 Risk of bias for accepted systematic reviews on the effectiveness of non-steroidal anti-inflammatory drugs (NSAIDs) for neck pain and low back pain based on Scottish Intercollegiate
Conflicts
declared
interest
of We reported our systematic review based on the Preferred
N
N
Y
Y
Y
N
Y
Y
Reporting Items for Systematic Reviews and Meta-A-
Publication
CS
CS
N
N
N
N
N
N
Results
methods for
Appropriate
combining
findings
Study selection
Y
N
Y
Y
Y
Y
Y
Y
Our search yielded 801 articles. We removed 95 duplicates
appropriately
assessed
disagreed.
of included
studies
Study characteristics
Y
Y
Y
Y
Y
Y
Y
Y
studies
List of
type
27, 28, 46, 47]: (1) had a clear research question; (2)
CS
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
search that was not comprehensive (4/6); (3) did not indi-
References
cate the excluded studies (5/6); (4) did not adequately re-
port the characteristics of the included studies (3/6); (5) did
not appropriately assess the scientific quality of studies (5/
123
Eur Spine J
Table 4 Risk of bias for inadmissible systematic reviews on the effectiveness of non-steroidal anti-inflammatory drugs (NSAIDs) for neck pain and low back pain based on Scottish
6); (6) did not use appropriate methods for combining
of interest
Conflicts
declared
findings (4/6); and (7) did not declare whether there were
conflicts of interest (3/6) (Table 4).
N
Y
Y
Y
N
N
Publication
Summary of evidence
assessed
bias
Whiplash-associated disorders
CS
CS
CS
N
N
N
One systematic review with a low risk of bias searched the
methods for
Appropriate
literature on NSAIDs for the management of WAD, but did
combining
findings
not find any relevant studies [15].
N
Y
Y
N
N
N
Neck pain and associated disorders
appropriately
Evidence from one systematic review suggests that oral
Scientific
assessed
NSAIDs may be more effective than placebo, but different
quality
NSAIDs lead to similar outcomes for recent-onset NAD
Y
N
N
N
N
N
[15]. The same review reported that intramuscular NSAIDs
documented
versus manipulation and soft tissue therapy lead to similar
Scientific
outcomes for NAD [15]. Hurwitz et al. [15] included two
quality
studies with a low risk of bias on the effectiveness of
Y
N
N
Y
Y
Y
NSAIDs for NAD (Tables 5, 6). One trial comparing oral
Characteristics
Indomethacin, Piroxicam, and placebo found: (1) statisti-
of included
cally significant differences in pain reduction favoring In-
domethacin or Piroxicam over placebo (clinical significance
studies
not reported in systematic review); and (2) no statistically
N
Y
Y
Y
N
N
significant differences in pain reduction between the two
included and
Y
Y
Y
Y
Y
Y
Recent-onset
Literature
search
CS
CS
N
Y
N
Y
data
CS
CS
CS
Y
Y
(2) one study compared NSAID with bed rest to bed rest
Research
question
54, 56–58].
Klinge et al. [31]
Chung et al. [26]
123
Eur Spine J
suggests that an oral NSAID (tenoxicam) combined with low back pain without radiculopathy. Six studies had im-
bed rest leads to similar outcomes (pain or range of motion) portant methodological limitations (Tables 5, 6). Eight
as placebo with bed rest [55]. Evidence from one study studies (with 17 treatment arms) had a low risk of bias [47,
with a low risk of bias suggests that similar improvements 59–65]: (1) four studies compared NSAID to placebo [59,
are obtained between: NSAID, muscle relaxant, and 60, 62, 66]; (2) one study compared NSAID to ac-
NSAID combined with muscle relaxant [55]. etaminophen [63]; (3) one study compared NSAID to a
The preponderance of evidence from six studies with a reference treatment (Doloteffin) [61]; (4) three studies
low risk of bias suggests that different oral NSAIDs lead to compared two different NSAIDs [59, 64, 65].
similar outcomes for recent-onset low back pain [52–54, 56– Evidence from four studies with a low risk of bias
58]. Four studies found similar outcomes (pain, disability, or suggests that oral NSAIDs are more effective than placebo
satisfaction) for: (1) Diflunisal versus Indomethacin [52] and for persistent low back pain [59, 60, 62, 66]. There were
Piroxicam versus Indomethacin (no statistically significant statistically significant differences in pain reduction fa-
differences) [56]; (2) Valdecoxib versus Diclofenac (no voring: (1) Naproxen over placebo [59]; (2) Etoricoxib
clinically or statistically significant differences) [57]; (3) over placebo (also clinically significant for Etoricoxib at
Aceclofenac versus Diclofenac (no clinically significant 60 mg/day but not 90 mg/day) [60, 66]; (3) Valdecoxib
differences; statistical significance not reported) [54]. over placebo (also clinically significant) [62]. Further,
However, two studies found statistically significant differ- evidence from one study with a low risk of bias suggests
ences favoring: (1) Lornoxicam over Diclofenac for pain that NSAID (Diflunisal) leads to statistically significant
reduction [58]; (2) Nimesulide over Ibuprofen (for improv- outcomes in satisfaction over acetaminophen (clinical
ing disability; no statistically significant differences for pain significance not reported) [63]. One study compared an
reduction); clinical significance was not reported [53]. NSAID to a reference treatment (Doloteffin) for reducing
pain [61]; however, neither clinical nor statistical sig-
Persistent nificance was reported in the systematic review.
Evidence from three studies suggests that different oral
Three systematic reviews with a low risk of bias [22, 24, NSAIDs lead to similar outcomes [59, 64, 65]. There were
47] included 14 studies investigating NSAIDs for persistent no statistically significant differences in outcomes (pain,
123
Table 5 Evidence table—characteristics of accepted systematic reviews on the effectiveness of non-steroidal anti-inflammatory drugs (NSAIDs) for neck pain and associated disorders,
whiplash-associated disorders, and low back pain
References, Search date, databases, Population Intervention Comparison Outcomes Authors’ conclusion about
123
study design language, limiters NSAIDs from systematic
review
review
Luijsterburg Up to 2004; MEDLINE, Individuals with acute, All types of conservative Placebo, Primary outcomes: symptoms At long-term there is no
et al. [27], EMBASE, CENTRAL, subacute, or chronic interventions including inactive (e.g. pain), overall evidence in favor of
systematic CINAHL, PsycINFO, PEDro; lumbosacral radicular NSAIDs treatment, no improvement, function, corticosteroids injections
review English, Dutch, French, syndrome treated in a primary treatment, return to work compared to placebo, no
German; RCTs health care or occupational other types Secondary outcomes: treatment or NSAID or
setting of physiological or physical anesthetic injection, apart
conservative examinations, quality of life, from conflicting evidence for
treatment, psycho-social outcomes, short-term pain relief
surgery medical consumption,
adverse events
Pinto et al. Up to 2010; MEDLINE, Individuals with sciatica or Analgesic and adjuvant pain Placebo, no At least one of the following: Most of the pooled estimates
[28], EMBASE, CINAHL, lumbar radiculopathy drugs, including NSAIDs treatment, or overall pain intensity, leg or did not favor NSAIDs over
systematic PsycINFO, Cochrane Central (included radiculopathy, other back pain intensity, disability placebo
review Register of Clinical Trials, nerve root compromise, nerve intervention status, work status, adverse
International Pharmaceuticals root compression, events
Abstracts, LILACS; English, lumbosacral radicular
German, Dutch, Portuguese, syndrome, nerve root pain,
Spanish; RCTs nerve root entrapment, and
pain radiating below the
knee)
Roelofs et al. Up to 2007; MEDLINE, Adults (C18 y.o.) with non- NSAIDs and COX-2 inhibitors Placebo, other Primary outcomes: pain NSAIDs are effective for short-
[24], EMBASE, Cochrane Central specific low back pain with or drugs, other intensity, global measure, term symptomatic relief in
systematic Register of Controlled Trials; without sciatica type of back pain-specific functional patients with acute and
review English, Dutch, German, NSAID, non- status, return to work, side chronic low-back pain
randomized and double-blind drug effects; secondary outcomes: without sciatica. However,
controlled trials treatment physiological outcomes, effect sizes are small. There
generic functional status, does not seem to be a specific
health care consumption type of NSAID which is
clearly more effective than
others. The selective COX-2
inhibitors showed fewer side
effects compared to
traditional NSAIDs in the
RCTs included in this review
(however, recent studies have
shown increased
cardiovascular risks in
specific patient populations)
LBP low back pain, LDH lumbar disc herniation, NSAID non-steroidal anti-inflammatory drug, RCT randomized controlled trial, TMJ temporomandibular joint, WAD whiplash-associated
disorders, y.o. years old
123
Table 6 Evidence table—effectiveness data from primary studies included in accepted systematic reviews on non-steroidal anti-inflammatory drugs (NSAIDs) for the management of neck
pain and associated disorders, whiplash-associated disorders, and low back pain
Population; Type/ combination of Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
123
reference of primary NSAID (where available) evidencea studies with low risk
study of biasb (?/-/=)
of biasb (?/-/=)
Acute LBP [57] NSAIDs (i): Valdecoxib NSAIDs (ii): Diclofenac Mean difference (95 % CI) pain intensity scale Low risk of bias [24] CS (=)
40 mg per day, second dose 75 mg b.i.d., 7 days (100 mm VAS) at 7 days; (i, N = 167 vs. ii, SS (=)
on day 1, 7 days N = 166) 0.26 (-3.76 to 4.28)
Mean difference (95 % CI) OLBPDQ (0–24 pt
scale) over 4 weeks; (iv vs. ii) 0.02 % (-3.21
to 3.16)
Acute LBP [58] NSAIDs (i): Lornoxicam on NSAIDs (ii): Diclofenac on Sum of pain intensity differences from baseline Low risk of bias [24] CS (NR)
day 1: 16 mg once a day day 1: 150 mg once a day on days 1-6 (SE); (i, N = 109) 4.2 (0.17); (ii, Lornoxicam vs.
and 8 mg once a day; days and 50 mg once a day; days N = 110) 3.8 (0.17); (i) significantly lower Diclofenac: SS (?)
2 to 7: 8 mg b.i.d. 2–7: 50 mg b.i.d. than (ii) P \ 0.05
Side effects: (i) 27; (ii) 28
Acute LBP [50] NSAID (i): Diflunisal (iii): Muscle relaxant No. of patients reporting marked improvement Low risk of bias [24] CS (NR)
capsules 500 mg (cyclobenzaprine/flexeril) after 2, 4, and 7 days: (i) 10, 14, 24 (ii) 10, 14, SS (?) favoring
b.i.d. capsules 5 mg b.i.d. 31 (iii) 8, 18, 30 (iv) 10, 14, 20; group (ii) NSAID plus muscle
(iv): Placebo capsules b.i.d. significantly better after 4 days (based on total relaxant after
NSAID plus muscle relaxant distribution); no other significant differences
(ii): Diflunisal capsules 4 days; SS (=) for
500 mg 1 5 mg No significant side effects reported all other outcomes
cyclobenzaprine b.i.d.
Acute LBP [55] NSAID (i): Tenoxicam Bed rest (ii): 7 days strict, Mean % improvement ± SD between baseline Low risk of bias CS (NR)
20 mg, 1 tablet daily, 7 days intermittent and 14 days in ROM (I) 123 ± 24, (ii) [24, 47] SS (?) favoring
14 days plus bed rest, 114 ± 23; significant 86 % (i) versus 70 % NSAID
7 days strict, 7 days (ii) no need for further treatment
intermittent Side effects: 2 patients in (i)
Acute LBP [56] NSAID (i): Piroxicam NSAID (ii): Indomethacin Mean improvement on VAS (range, 0–31) after Low risk of bias [24] CS (NR)
20 mg ? placebo capsules, 25-mg capsules, t.i.d., 6 weeks (i) 8, (ii) 9; similar improvement rates SS (=)
one per day ? twice 6 weeks (data in graphs)
placebo, 6 weeks Side effects similar (i) 13, (ii) 15
Acute LBP [51] NSAID (i): Indomethacin (ii): Placebo capsules No. of patients with complete relief of pain after Low risk of bias [24] CS (NR)
25 mg capsules, t.i.d., 7 and 14 days (i) 7, 14 (ii) 9, 16; no significant SS (=)
course of 50 capsules differences
Side effects comparable (i) 8 (ii) 5
Acute LBP [52] NSAID (i): Diflunisal NSAID (ii): Indomethacin Percentage of patients assessing therapy as good Low risk of bias [24] CS (NR)
500-mg capsules, twice 50-mg capsules, t.i.d., or excellent after 3 and 7 days (i) 45 %, 64 % SS (=)
daily, 7 days 7 days (ii) 45 %, 64 %; no significant differences
More side effects in (ii) 31 % than in (i) 18 %
123
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
123
of biasb (?/-/=)
Acute LBP NSAIDs (Diflunisal capsules 1. NSAID plus muscle Number of patients reporting marked High risk of bias [24] CS (N/A)
500 mg b.i.d.) relaxant (Diflunisal improvement after 2, 4, and 7 days was SS (?) at 4 days
capsules 500 mg ? 5 mg significantly better in NSAID plus muscle favoring NSAIDs
cyclobenzaprine b.i.d.) relaxant group after 4 days; no other plus muscle relaxant
2. Muscle relaxant significant differences
SS (=) for all other
(cyclobenzaprine/flexeril No serious side effects reported outcomes
capsules 5 mg b.i.d.)
3. Placebo capsules (b.i.d.)
Acute LBP NSAID (i): dexketorofen- NSAID (ii): Pain difference mean changed score, 100 mm High risk of bias [24]
trometamol 25 mg t.i.d., tramadolhydrochloride VAS, after 7 days: (i, N = 81) vs. (ii, N = 79)
7 days 50 mg t.i.d., 7 days -6 (4) (P \ 0.05)
Finger to floor distance mean changed score
(±SEM klopt… zieboven), cm, after 7 days:
(i) -22 (17); (ii) -18 (13); (i) vs. (ii)
significantly lower (P \ 0.05)
Acute LBP NSAID (i): Diflunisal NSAID (ii): Naproxen Percentage of patients reporting no pain (on an High risk of bias [24] N/A
capsules, 1000 mg initially, capsules, 500 mg initially, ordinal 4-point scale) after 2 weeks (i) 81 %
500 mg every 8–12 h, 250 mg every 6–8 h, (ii) 41 %. No significance tests reported
2 weeks 2 weeks
No adverse experiences reported by the patients
Acute LBP NSAID (i): Aceclofenac NSAID (ii): Diclofenac Mean improvement in pain intensity (VAS) after High risk of bias [24] N/A
150 mg, intramuscular 75 mg intramuscular b.i.d. 8 days: (i) 65 (ii) 62; global improvement
b.i.d. for 2 days ? 100 mg for 2 days ? 50 mg tablets good/very good in (i) 87 % and (ii) 79 % of
tablets b.i.d. for 5 days ? 1 t.i.d. for 5 days patients
placebo tablet for 5 days Side effects: (i) 1 (ii) 8 patients
Acute LBP NSAID (i): Piroxicam 20 mg (ii): Placebo capsules (i) More pain relief than (ii) measured with VAS High risk of bias [24] N/A
capsules, b.i.d. first 2 days, after 3 days. After 7 days no significant
one per day differences
next 5 days, 7 days Side effects similar (i) 13 % (ii) 17 %
Acute LBP NSAID (i): Diflunisal (ii): Physiotherapy: local Mean change in pain intensity on 4-point scale High risk of bias [24] N/A
500-mg capsules, 1000 mg heat, ultrasound, and after 4 and 12 days: (i) -0.9, -1.7 (ii) -0.9, -
immediately, 500 mg twice exercises (5 weekly 1.6 (iii) -1.1, -1.7; no significant differences
daily, 10 days sessions of 45 min in pain and mobility
(iii): Spinal manipulation
and/
or McKenzie therapy (5
sessions of 45 min weekly)
Acute LBP NSAID (i): Diclofenac NSAID (ii): Etofenamat Mean pain score (11-point NRS) at baseline and High risk of bias [24] N/A
intramuscular 3 mL 3 mL (1500 mg), 1–3 after 3–5 days: (i) 5.3, 2.7, (ii) 5.3, 2.4;
(75 mg), 1–3 injections, injections, 1–3 days physical examination (Schober): (i) 30, 52, (ii)
1–3 days 31, 59; no. of patients recovered after
3–5 days: (i) 27/47, (ii) 34/49; not significant
Eur Spine J
of biasb (?/-/=)
Acute LBP NSAID (i): Mefenamic acid (ii): Chlormezanone 100 mg No. of patients reporting no pain after 1 and High risk of bias [24] N/A
500 mg t.i.d. ? placebo and paracetamol 450 mg 2 7 days (i) 7, 21 (ii) 12, 23 (iii) 10, 20
b.i.d. capsules ? placebo t.i.d. No. of patients with adverse events (i) 9 (ii) 10
(iii): Ethoheptazine 75 mg (iii) 16
and meprobamate 150 mg
and aspirin 250 mg 2
capsules ? placebo t.i.d.
Acute LBP NSAID (i): Tenoxicam (ii): Placebo Mean pain intensity on VAS on day 1, 8, and 15 High risk of bias [24] N/A
20 mg intramuscular injection ? placebo (i) 7.4, 1.9, 0.6 (ii) 7.1, 2.8, 0.8;
injection on day 1 ? 20-mg capsules (?7 days bed rest) (i) significantly better on day 8
capsules 1 per day for days Side effects: one patient in group (i)
2–14 (?7 days bed rest)
Acute LBP NSAID (i): Flurbiprofen (ii): Acetaminophen, 4 g, No. of patients improved (physician’s High risk of bias [24] N/A
200 mg, 3 weeks 3 weeks assessment of pain) after 1, 2, 3, and 6 weeks:
(i) 13/23, 17/18, 7/8, 16/17, (ii) 8/22, 13/18,
8/9, 18/18; no significant difference
Side effects not specified
Acute LBP NSAID (i): Etodolac 200 mg NSAID (ii): Diclofenac Percentage of patients showing improvement High risk of bias [24] N/A
t.i.d., 5 days 50 mg t.i.d., 5 days after 5 days on pain intensity, global measure,
Laseque in (i) 50, 65, 66 %, and (ii) 55, 60,
74 %
Side effects not specified
Acute LBP NSAIDs (i): Naproxen (ii): Naproxen 500 mg Physical examination (Schober), days to High risk of bias [24] N/A
500 mg initially plus initially plus 250 mg q.i.d., resolution of pain, days to sit without pain,
250 mg q.i.d., 14 days plus cyclobenzaprine functional capacity after 14 days: (i) 5, 12.5,
10 mg t.i.d., 14 days 7, 15; (ii) 5, 8.5, 5, 9
Side effects: (i) 4 patients, (ii) 12 patients
Acute LBP NSAID (i): Piroxicam 10-mg (ii): Placebo capsules Patient (%) improved after 1 week only in High risk of bias [24] N/A
capsules, first 2 days, next subgroups with initial moderate/severe pain
12 days, 14 days (i) 82/49 % (ii) 53/38 %; no differences for
subgroup with mild initial pain; results after
2 weeks not reported (no data presented on side
effects for subgroup with back pain)
Acute LBP NSAID (i): Ketoprofen (ii): Combination Mean pain intensity at baseline and after 4 and High risk of bias [24] N/A
(Orudis) 200 mg (phenylbutazon-sodium, 9 days: (i) 68, 56, 42 and (ii) 68, 52, 39;
intramuscular b.i.d. for carbamoylphenoxyacetose, Laseque: (i) 48, 65, 74 and (ii) 51, 51, 60;
3 days plus dexamethasone, lidocaine- fingertip-to-floor distance: (i) 28, 25, 17 and
50 mg, 4 capsules plus hydrochloride, (ii) 36, 30, 26; no significant differences
100 mg drink for 5 days cyanocobalamin) 1 No adverse effects
intramuscular placebo
injection, 3 days plus 3
tablets for 5 days plus 1
placebo drink for 5 days
123
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
123
of biasb (?/-/=)
Acute LBP NSAID (i): Diflunisal (ii): Acetaminophen 300 mg Pain assessments by patient and investigator on High risk of bias [24] N/A
(capsules), initial dose with codeine 50 mg, two 3-point ordinal scale show similar
1000, 500 mg every 12 h, capsules initially, one improvement curves (data in graphs). No. of
15 days capsule every 4 h, 15 days patients rating drugs as excellent or very good
(i) 9 (ii) 9; no significant differences
Side effects: more side effects in (ii) 10 than in
(i) 3
Acute LBP NSAID (i): Diclofenac (ii): Diclofenac sodium plus After 1 week 49 % of (i) and 56.5 % of (ii) no High risk of bias [24] N/A
sodium, 25 mg 2 capsules vitamin B1, B6, B12, 2 or weak pain and 25 % in (i) and 29 % in (ii)
t.i.d., 7–14 days capsules t.i.d., 7–14 days improved; mean overall pain score at baseline
and after 1 and 2 weeks: (i) 78.7, 37.7, 23.7,
(ii) 83.7, 34.4, 23.6; tolerability good or very
good in 90 % (i) and 88 % (ii); side effects:
(i) 7 patients (ii) 12 patients
Acute LBP NSAIDs (i): Meloxicam NSAIDs (ii): Diclofenac Percentage of patients recovered after 8 days on High risk of bias [24] N/A
intravenous 1.5 mL intramuscular pain intensity, overall improvement,
(15 mg) 1 injection day 1 3 mL (75 mg) 1 injection day functional status, and tolerance: (i) 91, 89, 67,
plus 15 mg tablets once a 1 plus 96 %; (ii) 88, 91, 54, 98 %; side effects: (i) 6
day, 7 days patients (1 severe), (ii) 8 patients (1 severe)
100 mg tablets once a day,
7 days
Acute LBP NSAID (i): Etodolac 200 mg NSAID (ii): Piroxicam-beta- Mean degree of pain at baseline and after High risk of bias [24] N/A
b.i.d., 7 days cyclodextrin one 20-mg 7 days: (i) 3.1, 0.8, (ii) 3.13, 0.3; no significant
tablet, 7 days difference
Side effects (i) 3 patients, (ii) 2 patients
Acute LBP NSAID (i): aspirin 300 mg, 3 (v): Dextropropoxyphene Mean daily pain index during intervention High risk of bias [24] N/A
capsules, q.i.d., 1 week 32.5-mg and paracetamol period (on 4-point ordinal scale) (i) 1.4, (ii)
NSAID (ii): Indomethacin 325-mg capsules, 2 1.5, (iii) 1.4, (iv) 1.4, (v) 1.7, (vi) 1.7; (iii)
50 mg, t.i.d.,1 week capsules q.i.d., 1 week significantly different from (v) and (vi);
(vi): Paracetamol 500-mg (i) significantly different from (v)
NSAID (iii): Mefenamic acid
250 mg, 2 capsules t.i.d., capsules, 2 capsules q.i.d., Side effects: more side effects in (i) 20 (ii) 19
1 week 1 week and (v) 19 than in (iii) 12 (vi) 13 and (iv) 4
NSAID (iv): Phenylbutazone
100 mg, t.i.d., 1 week
Acute LBP NSAID (i): Indomethacin NSAID (ii): Total pain score (4-point scale) at baseline and High risk of bias [24] N/A
25 mg, q.i.d., 7 days Oxyphenbutazone 100 mg, after 3 and 7 days: (i) 52, 48, 39, (ii) 60, 47,
q.i.d., 7 days 44; total functional status score (movements;
4-point scale): 113, 105, 77, (ii) 142, 117, 99;
paracetamol use: after 3 and 7 days: (i) 7.4,
3.8, (ii) 7.7, 4.1; only pain after 3 days
significantly different
Side effects: (i) 5/40, (ii) 3/43
Eur Spine J
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
Eur Spine J
of biasb (?/-/=)
Acute LBP NSAIDs (i): Azapropazone NSAIDs (ii): Ketoprofen No data presented. Side effects: (i) 3 patients (1 High risk of bias [24] N/A
300 mg q.i.d., 1 week 50 mg q.i.d., 1 week severe), (ii) 14 patients (10 severe)
Acute LBP NSAID (i): Ibuprofen (ii): Paracetamol 1000 mg Percentage of patients cured after 10 weeks: High risk of bias N/A
800 mg t.i.d., 7 days t.i.d., 7 days (i) 67 %, (ii) 54 %, (iii) 82 %; no significant [23, 24]
(iii): No drug treatment difference
Side effects not specified
Acute LBP NSAID (i): Ibuprofen NSAID (ii): Felbinac (foam Patients (%) reporting none or mild severity High risk of bias [24] N/A
capsules 400 mg three 3 %) 3 times after 1 and 2 weeks (i) 84, 92 (ii) 76, 88; no
times per day ? placebo daily ? placebo capsules 3 significant differences between the groups
foam 3 times daily, 14 days times daily, 14 days No. of side effects (i) 22 (ii) 26
Non-specific low back pain without radiculopathy—persistent
Chronic LBP 1. NSAID—Naproxen Placebo capsules (iii) Decrease in pain by visual analog scale. Group Low risk of bias [22] CS (NR)
([3 months) [59] sodium 275 mg capsules, 2 (i) reduction of pain (ii) no change (iii) NSAID vs NSAID:
times 2 capsules per day, increase of pain SS (=)
14 days Data in graphs. Group (i) significantly better Naproxen vs. placebo:
2. NSAID—Diflunisal than (iii) and somewhat better than (ii) SS (?)
250 mg capsules, 2 times 2 Side effects similar in the three groups (i) 18 (ii)
capsules per day, 14 days 18 (iii) 16.
(in cross-over design)
Chronic LBP 1. NSAID—Etoricoxib Placebo, daily, 12 weeks Mean difference (95 % CI) pain intensity scale Low risk of bias [22] Pain: CS (?), SS (?)
([3 months) [60] 60 mg/day, 12 weeks (100 mm VAS) at 12 weeks; (iv vs. iii) - LBP bothersomeness:
2. NSAID—Etoricoxib 10.45 (-16.77 to -4.14); CS (NR), SS (?)
90 mg/day, 12 weeks (ii vs iii) -7.5 (-13.71 to -1.28) RMDQ: CS (=), SS
Mean difference (95 %CI) LBP bothersomeness (?)
scale over 12 weeks (4-pt Likert, 0 = not at
all, 4 = extremely); (iv vs. iii) -0.38 (-0.62
to -0.14); (ii vs iii) -0.33 (-0.57 to -0.09)
Mean difference (95 % CI) RMDQ (0–24 pt
scale) over 12 weeks; (iv vs. iii) -2.42 (-3.87
to -0.98); (ii vs. iii) -2.06 (-3.46 to -0.65)
Side effects: (i) 60 patients (14 withdrew); (ii)
56 patients (17 withdrew); (iii) 51 patients (10
withdrew)
Chronic LBP NSAID—Valdecoxib Placebo tablets, once daily, Mean changed score on pain intensity scale Low risk of bias
([3 months) [62] 40 mg/day, 4 weeks 4 weeks (100 mm VAS) at 1 week and 4 weeks; [22, 47]
(i) 29.2 and 41.9; (ii) 17.7 and 31.1; (iv vs. ii)
all P \ 0.001
Side effects: (i) 52 patients (1 withdrew); (ii) 35
patients (3 withdrew)
123
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
123
of biasb (?/-/=)
Chronic LBP [61] NSAIDs (i): Rofecoxib Reference treatment (ii): Number of patients pain free after 3 and Low risk of bias [24] CS (NR)
12.5 mg/day, 6 weeks Doloteffin 2400 mg/day, 6 weeks: (i) 4, 7; (ii) 5, 10 SS (NR)
6 weeks Side effects: (i) 14 patients (1 withdrew); (ii) 14
patients (6 withdrew)
Chronic LBP [65] NSAIDs (i): Etoricoxib NSAIDs (ii): Diclofenac 3 * Mean difference (95 % CI) pain intensity scale Low risk of bias [24] NSAID vs NSAID:
60 mg per day, 4 weeks 50 mg per day, 4 weeks (100 mm VAS) at 4 weeks; (i, N = 222 vs ii, pain: SS (=), CS (=)
N = 218) 2.51 (-1.50 to 6.51) RMDQ: SS (=), CS
Mean difference (95 % CI) RMDQ (0–24 pt (=)
scale) over 4 weeks; (iv vs. ii) -0.23 (-1.14 LBP bothersomeness:
to 0.67) SS (=)
Mean difference (95 %CI) LBP bothersomeness
scale over 4 weeks (4-pt Likert, 0 = not at all,
4 = extremely); (iv vs. ii) -0.02 (-0.17 to
0.13)
Side effects: (i) 79 patients (15 withdrew); (ii)
87 patients (13 withdrew)
Chronic LBP [64] NSAID (i): Ketoprofen NSAID (ii): Diclofenac Percentage of patients improved after 1, 2 weeks Low risk of bias [24] CS (NR)
25-mg capsules, sodium 25-mg capsules, (i) 71 %, 86 % (ii) 62 %, 79 %; no significant SS (=)
150 mg per day, duration not 75 mg per day, duration not differences
given given Side effects similar in both groups (i) 18 % (ii)
21 %
Chronic LBP [63] NSAID (i): Diflunisal (ii): Paracetamol 1000 mg, No. of patients with none or mild low back pain Low risk of bias [24] CS (NR)
500 mg capsules, b.i.d., q.i.d., 4 weeks after 2 and 4 weeks (i) 11, 13 (ii) 9, 7; SS (?)
4 weeks significantly more patients in (i) (10 of 16)
considered the therapy as good or excellent
than in (ii) (4 of 12)
Side effects similar (i) 2 (ii) 1
Chronic LBP [66] NSAID (Etoricoxib) Placebo Pain (VAS 0–100 mm): Relative risk 0.57 Low risk of bias [47] CS (NR)
(95 % CI 0.44, 0.74) SS (?)
Chronic LBP 1. NSAID (Rofecoxib Placebo, daily for 4 weeks Mean difference (Group 1—placebo): symptom High risk of bias N/A
([3 months) 25 mg/day, 4 weeks) bothersomeness: -0.5 (95 % CI -0.6 to -0.3) [22, 47]
2. NSAID (Rofecoxib RMDQ: -2.2 (95 % CI -3.2 to -1.3)
50 mg/day, 4 weeks) Group 2—placebo: symptom bothersomeness:-
0.5 (95 % CI -0.7 to -0.3)
RMDQ: -2.3 (95 % CI -3.3 to -1.3)
Adverse events: Group 1—112/233 (10
withdrew); Group 2—106/229 (10 withdrew);
Group 3—93/228 (3 withdrew)
Eur Spine J
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
Eur Spine J
of biasb (?/-/=)
Subacute or chronic NSAIDs (i): Nimesulide Reference treatment (ii): Sum of patients with none, mild moderate/sum High risk of bias [24] N/A
LBP 100 mg, b.i.d., 4 weeks Diclofenac sodium 50 mg of patients with severe, very severe pain at:
b.i.d., 4 weeks rest, motion, night, standing. Data in graphs.
No differences found between treatments
Side effects (i) 26, (ii) 41; (i) vs (ii) P = 0.05
Chronic LBP NSAID (i): Diflunisal (ii): Meptazinol 200 mg Mean change in degree of pain on 100-mm VAS High risk of bias [24] N/A
250-mg capsule, q.i.d., capsule, q.i.d., 3 weeks at 3 weeks (i) 45 (ii) 40; similar improvement
3 weeks regarding capacity for daily tasks (data in
graphs); no significant differences
Side effects similar (i) 19, (ii) 23 patients
Chronic LBP NSAID (i): Ibuprofen SR NSAID (ii): Diclofenac SR Mean ± SD improvement in pain intensity, High risk of bias [24] N/A
800 mg b.i.d. plus placebo, 100 mg once per day plus global measure, physical examination
14 days placebo, 14 days (Schober), level of activity: (i) 2.0, 2.2 ± 1.5,
0.6 ± 0.6, 2.2 (ii) 2.3, 3.0 ± 1.7, 0.4 ± 0.6,
1.9; no significant differences; side effects in
(i) 4 patients, (ii) 16 patients; significant
Non-specific low back pain without radiculopathy—duration not specified or variable
LBP (variable NSAIDs (i): Alclofenac 1 g NSAIDs (ii): Indomethacin Mean ± SEM change score in pain intensity (5- Low risk of bias [24] CS (NR)
duration) [67] t.i.d., 7 days (N = 15 acute; 50 mg t.i.d., 7 days point NRS) and functional status (4-point SS (=)
N = 16 chronic) (N = 15 acute; N = 15 NRS) after 7 days for acute low back pain:
chronic) (i) 1.46 ± 0.28, 1.80 ± 0.03; (ii) 1.45 ± 0.27,
1.53 ± 0.23; not significant; mean ± SEM
change score in pain intensity and functional
status after 7 days for chronic low back pain
(i) 1.66 ± 0.62, 1.31 ± 0.63; (ii) 1.01 ± 0.71,
1.00 ± 0.58; not significant
Side effects: (i) 5 patients, (ii) 3 patients
LBP (duration not NSAID (i): Indomethacin (ii): Dextropropoxyphene Data extraction not possible; side effects: (i) 7 High risk of bias [24] N/A
specified) 25 mg t.i.d., 4 days HCL, 150 mg t.i.d., 4 days patients, (ii) 13 patients
LBP (duration not NSAID (i): Diclofenac full (ii): Chiropractic Mean improvement on combined pain, High risk of bias [24] N/A
specified) dosage, 10–14 days (acute manipulation disability, and spinal mobility score (5–32)
patients), 15–20 days (iii): Physiotherapy after 3 weeks, 2, and 6 months; in subgroup
(chronic patients) with acute pain (i) 3.0, 10.7, 14.0, (ii) 7.5, 9.7,
(iv): Bedrest 12.3, (iii) 5.0, 8.4, 10.2, (iv) 5.4, 7.5, 7.3,
(v): Back school (v) not included, (vi) 1.8, 7.3, 11.0; group (ii)
(vi): Placebo (anti-edema gel) significantly better than others after 3 weeks,
no other differences; in subgroup with chronic
pain (i) 2.6, 2.2, 4.0, (ii) 2.2, 2.6, 4.3, (iii) 3.9,
4.2, 6.0, (iv) not included, (v) 0.5, 4.6, 8.9, (vi)
0.7, 1.2, 2.0; group (i) not significantly better;
no data on side effects reported
123
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
123
of biasb (?/-/=)
LBP (duration not NSAID (i): Ibuprofen, NSAID (ii): Diclofenac, No. of patients reporting to be improved after 1, High risk of bias [24] N/A
specified) 1200 mg per day, 14 days 75 mg per day, 14 days 3, 4 weeks (i) 5, 10, 6 (ii) 5, 12, 11; no
significant differences
Side effects similar: one in each group
LBP (duration not NSAID (i): Indomethacin NSAID (ii): Mean ± SD score on 30-point total rating scale High risk of bias [24] N/A
specified) plaster b.i.d., 4 weeks plus Diclofenacemulgel, 2 cm at baseline and after 1, 2 and 3 weeks:
1 mg oral vitamin B b.i.d. q.i.d., 4 weeks plus 1 mg (i) 18.13 ± 4.32, 21.43 ± 4.10, 24.27 ± 5.30,
oral vitamin B b.i.d. 25.87 ± 4.50, (ii) 18.65 ± 4.53,
20.46 ± 6.68, 24.30 ± 5.39, 26.33 ± 4.12;
not significant; % of patients with good
improvement, safety, and usefulness after
3 weeks: (i) 87, 97, 87 %, (ii) 82, 100, 82 %;
not significant
Side effects: (i) 3 patients, (ii) none
LBP (duration not NSAID (i): Loxoprofen NSAID (ii): Naproxen Mean ± SD therapeutic results according to High risk of bias [24] N/A
specified) 60 mg t.i.d., 6 weeks 250 mg t.i.d., 6 weeks criteria of the Japanese orthopedic Academic
Society at baseline and after 1 and 6 weeks:
(i) 16.86 ± 4.90, 22.27 ± 5.08, 27.44 ± 2.87
and (ii) 15.42 ± 5.03, 20.97 ± 6.05,
26.23 ± 3.11; not significant
Side effects (i) 6 patients, (ii) 6 patients
LBP (duration not NSAID (i): aspirin 625-mg (iii): Acetaminophen (dosage Mean no. of days before return to full activity High risk of bias [24] N/A
specified) capsules, q.i.d., 2 weeks not given), b.i.d., 2 weeks (i) 5.7 (ii) 6.5 (iii) 5.7; no significant
NSAID (ii): Phenylbutazone differences
100-mg capsules, q.i.d. No data on side effects given
(first 5 days), no further
information
LBP (duration not NSAID (i): Piroxicam 20 mg NSAID (ii): Indomethacin No. of patients who are (very much) improved High risk of bias [24] N/A
specified) capsules, once per day, 25 mg capsules, t.i.d., after 1 and 2 weeks assessed by a physician
14 days 14 days (i) 49 % (70 %) (ii) 46 % (58 %); patients’
assessment after 2 weeks (very) good (i) 69 %
(ii) 62 %; no significant differences
Side effects similar (i) 11 % (ii) 13 %
Non-specific low back pain with radiculopathy
Acute sciatic pain NSAIDs (i): Dipyrone (iii): Placebo, 5 mL isotonic Mean (SD) pain intensity (VAS) at baseline and Low risk of bias CS (NR)
[68] intramuscular, 5 mL (5 saline, 1–2 injections, after 6 h: (i) 80.2 (15.4), 33.4 (25.5); (ii) 79.2 [24, 47] Dipyrone versus
2.5 g) 1–2 injections, 1–2 days (14.5), 41.7 (25.9); (iii) 78.2 (14.8), 54.8 Diclofenac: SS (?)
1–2 days (25.3); (i) significantly better than (ii) and (iii);
no. of patients recovered after 2 days (5-point Dipyrone versus
NSAIDs (ii): placebo: SS (?)
Diclofenac intramuscular, NRS); (i) 27 (32 %); (ii) 10 (12 %); (iii) 7
(9 %)
3 mL (75 mg), 1–2
injections, 1–2 days Side effects: (i) 4, (ii) 1, (iii) 2 patients
Eur Spine J
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
Eur Spine J
of biasb (?/-/=)
Acute lumbosacral NSAID (Piroxicam) Placebo medication Sick leave: RR 1.0 (95 % CI 0.8, 1.3) Low risk of bias [27] CS (N/A)
radicular SS (=)
syndrome [73]
Acute sciatica [71] 1. NSAID (7.5 mg oral NSAID—Diclofenac 150 mg Mean difference in Diclofenac 150 mg— Low risk of bias [23, CS (NR)
meloxicam, 19 daily for meloxicam 7.5 mg: reduction in pain: 0.0 28, 47] SS (=)
4 days) (95 % CI -15.2, -15.2)
SS (=)
2. NSAID (15 mg oral Diclofenac 150 mg—meloxicam 15 mg:
meloxicam, 19 daily for reduction in pain -1.0 (95 % CI -4.5, 6.5)
4 days) Adverse events: 13 % of Group 1, 17 % of
Group 2, 15 % of Diclofenac Group—
abdominal pain, diarrhea, dizziness,
dyspepsia, flatulence, headache, nausea
Acute sciatica [70] 1. Lornoxicam 8 mg Placebo Lornoxicam Group—placebo: reduction in pain: Low risk of bias [28] CS (NR)
2. Diclofenac 50–150 mg 8.3 (95 % CI -1.4, 17.9) SS (=)
Diclofenac Group—placebo: SS (?)
Reduction in pain: 10.3 (95 % CI 0.7, 19.9) SS (=)
Diclofenac Group—Lornoxicam: reduction in
pain: 2.0 (95 % CI -7.7, 11.7)
Adverse events: 11 % in Lornoxicam Group,
12 % Diclofenac Group, 7 % placebo Group-
dyspepsia, diarrhea, nausea, abdominal pain,
flatulence, bronchitis, coughing, sputum,
hyperuricaemia, bilirubinaemia, myalgia, leg
cramps
Acute lumbar disc NSAID (i): Diclofenac 25-mg NSAID (ii): Proquazone Average improvement on ordinal 5-point scale Low risk of bias [24] CS (NR)
herniation [69] capsules, q.i.d. first 4 days (Biarison) 300-mg (0 = no response, 4 = very good response) SS (=)
and t.i.d. next 8 days, capsules, q.i.d. first 4 days during and after the intervention period of
12 days and t.i.d. next 8 days, 12 days according to physician and patient
12 days (i) 2.6 and 2.8 (ii) 2.8 and 3; no significant
differences in recovery rate
Side effects: mild side effects in three patients in
each group
Acute and subacute NSAID (i): Phenylbutazone (ii): Placebo capsules No. of patients reporting definite positive effect Low risk of bias [24] CS (NR)
sciatica [72] 200-mg capsules, 2 after intervention period (i) 14 (ii) 8; no SS (=)
capsules t.i.d. for 3 days, 1 significant differences
capsule t.i.d. next 2 days, No side effects reported by the patients
5 days
Acute sciatica [73] NSAID (i): Piroxicam 20-mg (ii): Placebo capsules Reduction of pain in back and leg measured by Low risk of bias [24] CS (NR)
capsules, 40 mg per day VAS after 4 weeks the same in the two groups SS (=)
first 2 days, 20 mg per day (data in graphs); no significant differences
next 12 days, 14 days More side effects in (i) 22 than in (ii) 13
123
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
123
of biasb (?/-/=)
Individuals with Diclofenac sodium orally Caudal epidural injection Diclofenac is less effective than caudal epidural High risk of bias [46] N/A
lumbar disc injection for reducing pain intensity at short-
herniation with and intermediate-term follow-ups, and for
radiculopathy reducing function at short-term follow-up
There is no difference in function measured at
intermediate-term follow-up between
Diclofenac and caudal epidural injection.
2 weeks:
VAS: -2.2 (95 % CI -2.8, -1.6)
Oswestry: -1.5 (95 % CI -2.0, -1.0)
1 month:
VAS: -1.0 (95 % CI -1.5, -0.5)
Oswestry: -0.6 (95 % CI -1.2, -0.2)
3 months:
VAS: -0.6 (95 % CI -1.1, -0.1)
Oswestry: -0.4 (95 % CI -1.9, 0.1)
Individuals with Diclofenac sodium orally Sarpogrelatehydroxychloride There is no difference between Diclofenac and High risk of bias [46] N/A
lumbar disc sarpogrelatehydroxychloride for short-term
herniation with outcomes of back pain intensity or leg pain
radiculopathy intensity
2 weeks:
VAS-1: 0.0 (95 % CI -0.6, 0.7)
VAS-2: 0.5 (95 % CI -0.1, 1.1)
Lumbosacral NSAID via intramuscular NSAID via intramuscular Pain (VAS 0-100): day 1: 0.1 (95 % CI -0.6, High risk of bias [27] N/A
radicular injection (Tiaprofenic injection (Ketoprofen 0.7)
syndrome (acute, 200 mg) 100 mg) day 2: 0.3 (-0.3, 0.9)
subacute, and
chronic) day 3: 0.6 (95 % CI 0.0, 1.3)
day 4: 0.6 (95 % CI 0.0, 1.3)
Side effects: stomach pain (5 times reported),
skin reaction (3 times reported)
Acute sciatica 1. NSAID (7.5 mg oral Oral placebo (19 daily for Mean difference in Meloxicam 7.5 mg— High risk of bias [28] N/A
meloxicam, 19 daily for 4 days) placebo: reduction in pain (0–100): 6.0 (95 %
4 days) CI 0.4, 11.5)
2. NSAID (15 mg oral Meloxicam 15 mg—placebo: reduction in pain
Meloxicam, 19 daily for (0–100): 5.0 (95 % CI -0.5, 10.5)
4 days) Adverse events: 8 % of Group 1, 7 % of Group
2, 4 % of placebo Group—abdominal pain,
diarrhea, dizziness, dyspepsia, hepatic
enzymes increased, nausea
Eur Spine J
Table 6 continued
Population; reference Type/ combination of NSAID Comparison Effect size(s) for outcomes Quality of the study/ Interpretation of
of primary study (where available) evidencea studies with low risk
Eur Spine J
of biasb (?/-/=)
Patients with acute Piroxicam orally, 20 mg, Placebo, twice daily for Piroxicam—placebo: High risk of bias [28] N/A
sciatica twice daily for 2 days, 2 days, followed by once Leg pain/immediate: 0.0 (95 % CI -6.9, 6.9)
(\14 days) followed by once daily for daily for next 12 days.
next 12 days. Additional Additional analgesics Leg pain/short term: 3.0 (95 % CI -3.9, 9.9)
analgesics allowed. allowed Back pain/immediate: 3.0 (95 % CI -3.9, 9.9)
Back pain/short term: 5.0 (95 % CI -1.9, 11.9)
Disability/immediate: 2.4 (95 % CI -4.5, 9.3)
Disability/short term: 2.9 (95 % CI -4.0, 9.8)
Lumbar disc NSAID (i): Indomethacin (ii): Placebo capsules Group (i) significantly more effective than (ii) High risk of bias [24] N/A
herniation 25 mg capsules, 3 capsules (N = 25 with nerve root only for patients with nerve root pain (data in
per day for 2 days, 4 pain, and N = 30 without graphs)
capsules per day next nerve root pain) Number of side effects (i) 38 (ii) 23
5 days (N = 25 with nerve
root pain, and N = 30
without nerve root pain)
Lumbar disc NSAID (i): Phenylbutazone (ii): Placebo capsules Group (i) improvement in motor weakness and High risk of bias [24] N/A
herniation 100-mg capsules, 6 per day painful straight-leg raising; no results are
for 2 days, t.i.d. next 8 days reported for group (ii);
Side effects reported in group (i) 3
Lumbar disc NSAID (i): Tenoxicam NSAID (ii): Tenoxicam Mean ± SD pain intensity (VAS) before High risk of bias [24] N/A
herniation or 20 mg intramuscular 20-mg tablet plus treatment and after 24 h: (i) 55 ± 15,
degeneration injection plus 1 placebo intramuscular placebo 34 ± 29, (ii) 50 ± 13, 37 ± 13
tablet, 4 days injection Side effects: (i) 5 patients, (ii) 5 patients
1 indicates that the effect size favors the intervention (i.e., NSAIDs); 2 indicates that the effect size favors the comparison; = indicates that there is no difference between groups (with respect
to statistical significance or clinical significance based on the interpretation of results by the OPTIMa Collaboration)
CI confidence interval, CS clinically significant, HVLA high velocity low amplitude, LBP low back pain, No. number, NRS numeric rating scale, NSAID non-steroidal anti-inflammatory drug,
OMT osteopathic manipulative therapy, RCT randomized controlled trial, RR relative risk, SD standard deviation, SS statistical significant, VAS Visual Analog Scale
a
The quality of the studies are based on the appraisal/ratings listed in the accepted systematic reviews
b
The results from low-quality studies or studies with high risk of bias are not included in our evidence synthesis
123
Eur Spine J
123
Eur Spine J
Table 7 Adverse events from non-steroidal anti-inflammatory drugs randomized to placebo). This study serves as a preliminary
(NSAIDs) as reported by systematic reviews with a low risk of bias indication that NSAIDs may be more effective than placebo
Condition Adverse events for the management of NAD. There is an urgent need for
high-quality studies investigating the use of oral NSAIDs for
Neck pain The systematic review on neck pain reported that
the management of recent onset and persistent NAD.
dyspepsia was common ([10 %), gastrointestinal
bleeding was occasional (1–10 %), and heart attacks The high-quality systematic reviews used different tools
were rare (0.001–1 %) [15, 32] to assess the quality of studies included in their reviews.
Low back Kuijpers et al. [22] reported that there are statistically Specifically, three reviews used the PEDro scale [23, 28,
pain significantly more adverse events in the NSAIDs 46], three reviews used the Cochrane Back Review Group
group compared to placebo based on low quality criteria [22, 24, 47], one review used the Delphi list [27],
evidence [relative risk (RR) 1.24, 95 % CI
1.07–1.43] and one review used criteria focused on sources of poten-
Pinto et al. [28] reported that the median rate of tial selection bias, information bias, and confounding
adverse events was 17 % (interquartile range (name of criteria not reported) [15]. It is important that
10–30 %) for NSAIDs and 11 % (3–23 %) for critical appraisal methods assess for potential selection
placebo bias, information bias, and confounding. Reviewers should
Roelofs et al. [24] reported that there were statistically also consider whether these biases would likely lead to
more patients with side effects with NSAIDs
compared to placebo. They also reported that misleading conclusions.
selective Cox-2 inhibitors showed fewer side effects There was a lack of information on adverse events
than non-selective oral NSAIDs [24] provided by the high-quality systematic reviews. There
Abdel Shaheed et al. [23] reported that side effects are limitations to using systematic reviews of clinical
were more common with placebo than NSAID; trials for information on adverse events, including: (1)
however, this was based on one study. They also
reported that using higher NSAID doses are
adverse events are often not the primary endpoints in
associated with an increased risk of side effects than clinical trials, which can lead to incomplete reporting;
low dose or taken when necessary [23] (2) inconsistent definitions of adverse events; (3) with-
drawal or loss to follow-up in the trial due to adverse
events; and (4) inadequate consideration of follow-up
For low back pain, two systematic reviews have reported and potential time required to develop adverse events
limited or unclear evidence [20, 22], whereas two sys- with drugs [76]. Therefore, adverse events related to
tematic reviews reported that NSAIDs are effective for drugs should be informed by improved methodology of
short-term symptomatic relief [23, 24]. Two systematic clinical trials and observational cohort studies examining
reviews found evidence supporting the use of NSAIDs in adverse events.
individuals with acute [25] or chronic low back pain [25,
26]. Our results differ from these conclusions; this is likely
due to our best evidence synthesis methods, where only Strengths and limitations
reviews with a low risk of bias are included in the syn-
thesis. Including studies with a high risk of bias in the Our review has several strengths. The literature search was
evidence synthesis may produce unreliable results that comprehensive and peer-reviewed to minimize errors. We
threaten the validity of the conclusions. outlined detailed inclusion and exclusion criteria to identify
Two systematic reviews reported that NSAIDs are not relevant systematic reviews. Pairs of trained, independent
more effective than placebo for lumbar radiculopathy [27, reviewers critically appraised the literature using the SIGN
28]. Our findings agree with these reviews, as we found criteria [36] to standardize the critical appraisal process.
inconsistent evidence on the effectiveness of NSAIDs over Our conclusions were based on the best evidence synthesis
placebo for recent-onset low back pain with radiculopathy. method to minimize the risk of bias associated with using
Despite its common use, there are few high-quality low-quality reviews [38].
studies informing the use of NSAIDs for the management of Our review has limitations. We only searched the Eng-
NAD. We identified one study investigating the use of in- lish literature; however, systematic reviews of clinical tri-
tramuscular NSAIDs for recent-onset NAD [49], and one als investigating the impact of language restriction found
study investigating the use of oral NSAIDs for NAD (du- that it does not lead to bias as most large trials are pub-
ration not specified) [48]. The latter study [48] was pub- lished in English [77–81]. The critical appraisal of reviews
lished in 1983 and has important limitations: (1) the may vary between reviewers. This potential bias was
outcome measure has not been tested for its validity and minimized by using standardized appraisal forms, con-
reliability; and (2) the study used pre-post analysis of ducting training sessions, and using a consensus process to
NSAID use compared to placebo (no groups were determine admissibility of systematic reviews.
123
Eur Spine J
123
Eur Spine J
123
Eur Spine J
123
Eur Spine J
5. Cote P, Cassidy JD, Carroll L (2001) The treatment of neck and drugs for the treatment of low back pain. J Pain Symptom Manag
low back pain: who seeks care? Who goes where? Med Care 28(1):72–95
39(9):956–967 26. Chung JW, Zeng Y, Wong TK (2013) Drug therapy for the
6. Cypress BK (1983) Characteristics of physician visits for back treatment of chronic nonspecific low back pain: systematic re-
symptoms: a national perspective. Am J Public Health 73(4): view and meta-analysis. Pain Physician 16(6):E685–E704
389–395 27. Luijsterburg PA, Verhagen AP, Ostelo RW, van Os TA, Peul
7. Carey TS, Evans A, Hadler N, Kalsbeek W, McLaughlin C, Fryer WC, Koes BW (2007) Effectiveness of conservative treatments
J (1995) Care-seeking among individuals with chronic low back for the lumbosacral radicular syndrome: a systematic review. Eur
pain. Spine (Phila Pa 1976) 20(3):312–317 Spine J 16(7):881–899
8. Makela M, Heliovaara M, Sievers K, Impivaara O, Knekt P, Aro- 28. Pinto RZ, Maher CG, Ferreira ML et al (2012) Drugs for relief of
maa A (1991) Prevalence, determinants, and consequences of pain in patients with sciatica: systematic review and meta-ana-
chronic neck pain in Finland. Am J Epidemiol 134(11):1356–1367 lysis. BMJ 344:e497
9. Carey TS, Evans AT, Hadler NM et al (1996) Acute severe low 29. Vroomen PC, de Krom MC, Slofstra PD, Knottnerus JA (2000)
back pain. A population-based study of prevalence and care- Conservative treatment of sciatica: a systematic review. J Spinal
seeking. Spine (Phila Pa 1976) 21(3):339–344 Disord 13(6):463–469
10. Cassidy JD, Cote P, Carroll LJ, Kristman V (2005) Incidence and 30. Moore RA, Barden J (2008) Systematic review of dexketoprofen
course of low back pain episodes in the general population. Spine in acute and chronic pain. BMC Clin Pharmacol 8:11
(Phila Pa 1976) 30(24):2817–2823 31. Klinge SA, Sawyer GA (2013) Effectiveness and safety of topical
11. Cote P, Baldwin ML, Johnson WG, Frank JW, Butler RJ (2008) versus oral nonsteroidal anti-inflammatory drugs: a comprehen-
Patterns of sick-leave and health outcomes in injured workers sive review. Phys Sports Med 41(2):64–74
with back pain. Eur Spine J 17(4):484–493 32. Guzman J, Hurwitz EL, Carroll LJ et al (2008) A new conceptual
12. Cote P, Cassidy JD, Carroll LJ, Kristman V (2004) The annual model of neck pain: linking onset, course, and care: the bone and
incidence and course of neck pain in the general population: a joint decade 2000–2010 task force on neck pain and its associated
population-based cohort study. Pain 112(3):267–273 disorders. Spine (Phila Pa 1976) 33(4 Suppl):S14–S23
13. Hincapie CA, Cassidy JD, Cote P, Carroll LJ, Guzman J (2010) 33. Spitzer WO, Skovron ML, Salmi LR et al (1995) Scientific
Whiplash injury is more than neck pain: a population-based study monograph of the Quebec Task Force on Whiplash-Associated
of pain localization after traffic injury. J Occup Environ Med Disorders: redefining ‘‘whiplash’’ and its management. Spine
52(4):434–440 (Phila Pa 1976) 20(8 Suppl):1s–73s
14. Haldeman S, Dagenais S (2008) A supermarket approach to the 34. McGowan J, Sampson M, Lefebvre C (2010) An evidence based
evidence-informed management of chronic low back pain. Spine checklist for the peer review of electronic search strategies
J 8(1):1–7 (PRESS EBC). Evid Based Libr Inf Pract 5(1):149–154
15. Hurwitz EL, Carragee EJ, van der Velde G et al (2008) Treatment 35. EndNote (2013) EndNote X6. http://endnote.com/training/mats/
of neck pain: noninvasive interventions: results of the bone and enuserguide/eng/enguide-full.pdf. Retrieved 1 Nov 2014
joint decade 2000–2010 task force on neck pain and its associated 36. Harbour R, Miller J (2001) A new system for grading recom-
disorders. Spine (Phila Pa 1976) 33(4 Suppl):S123–S152 mendations in evidence based guidelines. BMJ 323(7308):
16. American Academy of Orthopaedic Surgeons (AAOS) (2009) 334–336
What are NSAIDs? http://orthoinfo.aaos.org/topic.cfm?topic= 37. Shea BJ, Grimshaw JM, Wells GA et al (2007) Development of
a00284. Retrieved 1 Nov 2014 AMSTAR: a measurement tool to assess the methodological
17. Rasu RS, Vouthy K, Crowl AN et al (2014) Cost of pain quality of systematic reviews. BMC Med Res Methodol 7:10
medication to treat adult patients with nonmalignant chronic pain 38. Slavin RE (1995) Best evidence synthesis: an intelligent alter-
in the United States. J Manag Care Pharm 20(9):921–928 native to meta-analysis. J Clin Epidemiol 48(1):9–18
18. Pincus T, Newman S (2001) Recall bias, pain, depression and 39. Viera AJ, Garrett JM (2005) Understanding interobserver
cost in back pain patients. Br J Clin Psychol 40(Pt 2):143–156 agreement: the kappa statistic. Fam Med 37(5):360–363
19. Carmona L, Ballina J, Gabriel R, Laffon A (2001) The burden of 40. Stauffer ME, Taylor SD, Watson DJ, Peloso PM, Morrison A
musculoskeletal diseases in the general population of Spain: re- (2011) Definition of nonresponse to analgesic treatment of
sults from a national survey. Ann Rheum Dis 60(11):1040–1045 arthritic pain: an analytical literature review of the smallest de-
20. van Tulder MW, Scholten RJ, Koes BW, Deyo RA (2000) tectable difference, the minimal detectable change, and the
Nonsteroidal anti-inflammatory drugs for low back pain: a sys- minimal clinically important difference on the pain visual analog
tematic review within the framework of the Cochrane Col- scale. Int J Inflam 2011:231926
laboration Back Review Group. Spine (Phila Pa 1976) 25(19): 41. Andersson G, Johansson C, Nordlander A, Asmundson GJ (2012)
2501–2513 Chronic pain in older adults: a controlled pilot trial of a brief
21. Peloso P, Gross A, Haines T, Trinh K, Goldsmith CH, Burnie S cognitive-behavioural group treatment. Behav Cogn Psychother
(2007) Medicinal and injection therapies for mechanical neck 40(2):239–244
disorders. Cochrane Database Syst Rev 3:Cd000319 42. Carstensen TB, Frostholm L, Oernboel E et al (2012) Are there
22. Kuijpers T, van Middelkoop M, Rubinstein SM et al (2011) A gender differences in coping with neck pain following acute
systematic review on the effectiveness of pharmacological in- whiplash trauma? A 12-month follow-up study. Eur J Pain
terventions for chronic non-specific low-back pain. Eur Spine J 16(1):49–60
20(1):40–50 43. Ostelo RW, Deyo RA, Stratford P et al (2008) Interpreting
23. Abdel Shaheed C, Maher CG, Williams KA, McLachlan AJ (2014) change scores for pain and functional status in low back pain:
Interventions available over the counter and advice for acute low towards international consensus regarding minimal important
back pain: systematic review and meta-analysis. J Pain 15(1):2–15 change. Spine (Phila Pa 1976) 33(1):90–94
24. Roelofs PD, Deyo RA, Koes BW, Scholten RJ, van Tulder MW 44. Lauche R, Langhorst J, Dobos GJ, Cramer H (2013) Clinically
(2008) Nonsteroidal anti-inflammatory drugs for low back pain: an meaningful differences in pain, disability and quality of life for
updated Cochrane review. Spine (Phila Pa 1976) 33(16):1766–1774 chronic nonspecific neck pain—a reanalysis of 4 randomized
25. Schnitzer TJ, Ferraro A, Hunsche E, Kong SX (2004) A com- controlled trials of cupping therapy [MCID]. Complement Ther
prehensive review of clinical trials on the efficacy and safety of Med 21(4):342–347
123
Eur Spine J
45. Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred 64. Matsumo S, Kaneda K, Nohara Y (1991) Clinical evaluation of
reporting items for systematic reviews and meta-analyses: the Ketoprofen (Orudis) in lumbago: a double blind comparison with
PRISMA statement. BMJ 339:b2535 Diclofenac sodium. Br J Clin Pract 35:266
46. Hahne AJ, Ford JJ, McMeeken JM (2010) Conservative man- 65. Zerbini C, Ozturk ZE, Grifka J et al (2005) Efficacy of Etoricoxib
agement of lumbar disc herniation with associated radiculopathy: 60 mg/day and Diclofenac 150 mg/day in reduction of pain and
a systematic review. Spine (Phila Pa 1976) 35(11):E488–E504 disability in patients with chronic low back pain: results of a
47. Keller A, Hayden J, Bombardier C, van Tulder M (2007) Effect 4-week, multinational, randomized, double-blind study. Curr
sizes of non-surgical treatments of non-specific low-back pain. Med Res Opin 21(12):2037–2049
Eur Spine J 16(11):1776–1788 66. Pallay R, Seger W, Adler J (2007) Etoricoxib reduced pain and
48. Yamamoto M, Sugano T, Kashiwazaki S, Kageyama T, disability and improved quality of life in patients with chronic
Mizushima Y, Kameyama M (1983) Double-blind comparison of low back pain: a 3 month, randomized, controlled trial. Scand J
Piroxicam and Indomethacin in the treatment of cervicobrachial Rheumatol 33:257–266
syndrome and periarthritis scapulohumeralis (stiff shoulder). Eur 67. Jaffe G (1974) A double-blind, between-patient comparison of
J Rheumatol Inflamm 6(3):266–273 alclofenac (‘Prinalgin’) and Indomethacin in the treatment of low
49. McReynolds TM, Sheridan BJ (2005) Intramuscular ketorolac back pain and sciatica. Curr Med Res Opin 2(7):424–429
versus osteopathic manipulative treatment in the management of 68. Babej-Dolle R, Freytag S, Eckmeyer J et al (1994) Parenteral
acute neck pain in the emergency department: a randomized Dipyrone versus Diclofenac and placebo in patients with acute
clinical trial. J Am Osteopath Assoc 105(2):57–68 lumbago or sciatic pain: randomized observer-blind multicenter
50. Bakshi R, Thumb N, Broll H (2008) Treatment of acute lum- study. Int J Clin Pharmacol Ther 32(4):204–209
bosacral back pain with Diclofenac resinate: results of a double- 69. Blazek M, Keszthelyi B, Varhelyi M, Korosi O (1986) Com-
blind comparative trial versus Piroxicam. Drug Invest 8:288–293 parative study of Biarison and Voltaren in acute lumbar pain and
51. Goldie I (1968) A clinical trial with Indomethacin (indomee(R)) lumbo-ischialgia. Ther Hung 34(3):163–166
in low back pain and sciatica. Acta Orthop Scand 39(1):117–128 70. Herrmann W, Geertsen M (2009) Efficacy and safety of
52. Orava S (1986) Medical treatment of acute low back pain. Di- lornoxicam compared with placebo and Diclofenac in acute sci-
flunisal compared with Indomethacin in acute lumbago. Int J Clin atica/lumbo-sciatica: an analysis from a randomised, double-
Pharmacol Res 6(1):45–51 blind, multicentre, parallel-group study. Int J Clin Pract 63:1613–
53. Pohjolainen T, Jekunen A, Autio L, Vuorela H (2000) Treatment 1621
of acute low back pain with the COX-2-selective anti-inflam- 71. Dreisser RL, Le Parc JM, Velicitat P, Lleu PL (2001) Oral
matory drug nimesulide: results of a randomized, double-blind meloxicam is effective in acute sciatica: two randomised, double-
comparative trial versus ibuprofen. Spine (Phila Pa 1976) blind trials versus placebo or Diclofenac. Inflamm Res
25(12):1579–1585 50(1):S17–S23
54. Schattenkirchner M, Milachowski KA (2003) A double-blind, 72. Weber H, Aasand G (1980) The effect of Phenylbutazone on
multicentre, randomised clinical trial comparing the efficacy and patients with acute lumbago-sciatica. A double blind trial. J Oslo
tolerability of Aceclofenac with Diclofenac resinate in patients City Hosp 30(5):69–72
with acute low back pain. Clin Rheumatol 22(2):127–135 73. Weber H, Holme I, Amlie E (1993) The natural course of acute
55. Szpalski M, Hayez JP (1994) Objective functional assessment of sciatica with nerve root symptoms in a double-blind placebo-
the efficacy of tenoxicam in the treatment of acute low back pain. controlled trial evaluating the effect of Piroxicam. Spine (Phila Pa
A double-blind placebo-controlled study. Br J Rheumatol 1976) 18(11):1433–1438
33(1):74–78 74. Weber H (1983) Lumbar disc herniation. A controlled, prospec-
56. Videman T, Heikkila J, Partnaen T (1984) Double-blind parallel tive study with ten years of observation. Spine 8:131–140
study of meptazinol versus diflunisal in the treatment of lumbago. 75. Waddell G, Feder G, Lewis M (1997) Systematic reviews of bed
Curr Med Res Opin 9:246–252 rest and advice to stay active for acute low back pain. Br J Gen
57. Ximenes A, Robles M, Sands G, Vinueza R (2007) Valdecoxib is Pract 47(423):647–652
as efficacious as Diclofenac in the treatment of acute low back 76. Huang HY, Andrews E, Jones J, Skovron ML, Tilson H (2011)
pain. Clin J Pain 23(3):244–250 Pitfalls in meta-analyses on adverse events reported from clinical
58. Yakhno N, Guekht A, Skoromets A et al (2006) Analgesic effi- trials. Pharmacoepidemiol Drug Saf 20(10):1014–1020
cacy and safety of lornoxicam quick-release formulation com- 77. Juni P, Holenstein F, Sterne J, Bartlett C, Egger M (2002)
pared with Diclofenac potassium: randomised, double-blind trial Direction and impact of language bias in meta-analyses of con-
in acute low back pain. Clin Drug Investig 26(5):267–277 trolled trials: empirical study. Int J Epidemiol 31(1):115–123
59. Berry H, Bloom B, Hamilton EB, Swinson DR (1982) Naproxen 78. Moher D, Fortin P, Jadad AR et al (1996) Completeness of re-
sodium, diflunisal, and placebo in the treatment of chronic back porting of trials published in languages other than English: im-
pain. Ann Rheum Dis 41(2):129–132 plications for conduct and reporting of systematic reviews.
60. Birbara CA, Puopolo AD, Munoz DR et al (2003) Treatment of Lancet 347(8998):363–366
chronic low back pain with Etoricoxib, a new cyclo-oxygenase-2 79. Moher D, Pham B, Lawson ML, Klassen TP (2003) The inclusion
selective inhibitor: improvement in pain and disability—a ran- of reports of randomised trials published in languages other than
domized, placebo-controlled, 3-month trial. J Pain 4(6):307–315 English in systematic reviews. Health Technol Assess 7(41):1–90
61. Chrubasik S, Model A, Black A, Pollak S (2003) A randomized 80. Morrison A, Polisena J, Husereau D et al (2012) The effect of
double-blind pilot study comparing Doloteffin and Vioxx in the English-language restriction on systematic review-based meta-
treatment of low back pain. Rheumatology (Oxford) 42(1):141– analyses: a systematic review of empirical studies. Int J Technol
148 Assess Health Care 28(2):138–144
62. Coats TL, Borenstein DG, Nangia NK, Brown MT (2004) Effects 81. Sutton AJ, Duval SJ, Tweedie RL, Abrams KR, Jones DR (2000)
of Valdecoxib in the treatment of chronic low back pain: results Empirical assessment of effect of publication bias on meta-ana-
of a randomized, placebo-controlled trial. Clin Ther 26(8):1249– lyses. BMJ 320(7249):1574–1577
1260
63. Hickey RF (1982) Chronic low back pain: a comparison of di-
flunisal with paracetamol. N Z Med J 95(707):312–314
123