systematic review

Background. Obstructive sleep apnea (OSA) is a disorder characterized by periods of airflow
cessation (apnea) or reduced airflow (hypopnea) during sleep. The diagnosis and severity of
OSA, and response to therapy, have historically been assessed using the apnea-hypopnea index
(AHI). However, several definitions of this measure exist, and the utility of the AHI and
associated measures as valid surrogate measures of health outcomes has been questioned. OSA is
commonly treated with the use of continuous positive airway pressure (CPAP) devices during
sleep. The efficacy of CPAP, including for Food and Drug Administration (FDA)
clearance/approval, has been based on changes in AHI, but the long-term effect of CPAP on
health outcomes and the role of disease severity (as measured by AHI) or sleepiness symptoms
on the putative effect of CPAP are unclear.
Methods. We searched Medline, Embase, the Cochrane databases, CINAHL, and
ClinicalTrials.gov from January 2010 through March 22, 2021; we screened reference lists of the
2011 Agency for Healthcare Research and Quality (AHRQ) OSA report and other systematic
reviews for earlier studies. We included only randomized controlled trials (RCT) and
nonrandomized comparative studies (NRCS) of CPAP that adjusted for potential confounders for
CPAP evaluation. We also included other comparative studies that reported both changes in
potential intermediate or surrogate measures (e.g., AHI) and effects on health outcomes. All
studies had to report effects on prespecified long-term (12 months for most outcomes) health
outcomes in adults with OSA. We did not evaluate sleepiness, other symptoms, or intermediate
outcomes. We excluded observational studies that did not directly compare treatment options.
Results. The 52 identified studies used highly inconsistent criteria to define breathing measures
(apneas, hypopneas, and oxygen desaturation). Definitions of respiratory disturbance events
(e.g., apneas, hypopneas) and criteria to define or categorize severity of OSA are highly
inconsistent across studies, despite frequent claims of using standard national or international
definitions. Possible differences in study findings based on heterogeneity of OSA and sleep study
measures could not be elucidated. Among 31 studies that directly compared CPAP and no CPAP
(29 studies) or sham CPAP (2 studies), 14 were RCTs (12 intention-to-treat [ITT]) and 17 were
NRCSs (11 analyses of CPAP users versus nonusers).
With one exception, RCTs did not find statistically significant effects for any cardiovascular
(CV) outcome. RCTs did not provide evidence that CPAP affects the risk of all-cause mortality
(summary effect size [ES] 0.89, 95% confidence interval [CI] 0.66 to 1.21), CV mortality
(summary ES 0.99, 95% CI 0.64 to 1.53), stroke (summary ES 0.99, 95% CI 0.73 to 1.35),
myocardial infarction (summary ES 1.05, 95% CI 0.78 to 1.41), or composite CV outcomes (ES
range 0.42 to 1.10 across studies, all statistically nonsignificant); all with low strength of
evidence (SoE). The three RCTs that aimed to be powered for composite CV events failed to
vi
show a significant or clinically meaningful effect. The NRCSs, overall, found significant
adjusted associations between CPAP use and all-cause mortality. Combining the RCTs and
NRCSs, the summary ES for all-cause mortality was ES 0.61 (95% CI 0.49 to 0.76), supporting a
low SoE of an association between CPAP use and lower risk of death. Data from the NRCSs did
not change other conclusions. Both RCTs and NRCSs provide insufficient evidence regarding
the effect of CPAP on the risk of transient ischemic attack, angina, coronary artery
revascularization, congestive heart failure, and atrial fibrillation.
RCTs also did not provide evidence that CPAP affects the risk of driving accidents or the risk of
incident diabetes (both low SoE), or that CPAP results in clinically significant changes in
depression or anxiety scores, executive cognitive function measures, or nonspecific quality of
life measures (all low SoE). RCTs provide insufficient evidence regarding the effect of CPAP on
incident hypertension, functional status measures, male or female sexual function, or days of
work missed. Data from the NRCSs did not change these conclusions.
Eligible studies provided insufficient evidence regarding possible differences in the effect of
CPAP based on patient characteristics (such as disease severity or comorbidities), different
diagnostic criteria, or whether RCTs were analyzed as ITT or “as-treated”.
Eligible studies provided insufficient evidence about adverse events due to CPAP use. Many
studies did not collect such data. Adverse events reported in the U.S. Food and Drug
Administration (FDA) database mostly related to inadequate humidification, user errors, or
device malfunction. No deaths were attributed to CPAP use.
Review of the FDA database found 163 CPAP devices used to treat adults with sleep apnea. The
large majority of FDA 510(k) Premarket Notification records cite other previously approved
CPAP devices to support claims of equivalence. The available data did not reference clinical
studies to support the device manufacturers’ claims.
Review of the National Institutes of Health’s RePORTER revealed no germane funded trials.
Review of ClinicalTrials.gov revealed a single large RCT with a mortality endpoint, but no
updating of the site since 2015 and five additional small trials (2 addressing events in patients
with paroxysmal atrial fibrillation or hypertensive pulmonary edema; 3 measuring AHI,
cognition, or kidney function).
RCTs comparing CPAP and mandibular advancement devices found no differences in depression
or anxiety symptoms (low SoE). There was insufficient evidence for other outcomes. RCTs
comparing fixed and autoCPAP found no differences in functional status; other long-term
outcomes were not reported. No eligible studies evaluated comparisons of other types of CPAP.
No studies have evaluated the validity of changes in intermediate or surrogate measures (such as
change in AHI during a clinical trial) as predictors of long-term health outcomes. No studies
reported surrogacy or mediation analyses, nor did any compare the concordance of changes in
different sleep study and symptom measures with health outcomes. Across the 15 studies that
reported both changes in intermediate or surrogate measures and effects on long-term health
vii
outcomes, data were too sparse to allow adequate cross-study evaluation of concordance between
change in any specific measure and health outcome or sleep questionnaires.
Conclusions. Studies are highly inconsistent as to how they define breathing measures during
sleep studies and OSA itself. Insufficient evidence exists to assess the validity of change in AHI
as a surrogate or intermediate measure for long-term health outcomes. Until such validation has
been conducted, it cannot be assumed that changes (e.g., improvements) in intermediate or
surrogate outcomes are correlated with long-term health outcomes.
RCTs do not provide evidence that CPAP prescription affects long-term, clinically important
outcomes. Specifically, with low SoE, RCTs do not demonstrate that CPAP affects all-cause
mortality, various CV outcomes, clinically important changes in psychosocial measures, or other
clinical events. NRCSs reported associations between CPAP use and reduced risk of all-cause
death. NRCS results did not differ from RCTs for other outcomes. We have limited confidence
that the summary estimates are close to any true effect.
Comparative studies did not adequately address whether the effect of CPAP varies based on
disease severity (e.g., as assessed by AHI), symptoms (e.g., as assessed by sleepiness scales),
other patient characteristics, different features or modes or CPAP, or different criteria or
definitions of sleep measures or OSA diagnosis.
Additional well-conducted comparative studies are needed to better assess the potential effects of
CPAP on long-term outcomes for patients with OSA, whether any particular group of patients
may benefit to a greater or lesser degree from CPAP treatment, and whether of changes in AHI
(and/or other breathing measures) are valid intermediate or surrogate measures of health
outcomes. Associations identified in comparative studies could serve as the basis for more
rigorous trials.

Importance: Amitriptyline is an established medication used off-label for the treatment of fibromyalgia, but pregabalin, duloxetine, and milnacipran are the only pharmacological agents approved by the US Food and Drug Administration (FDA) to treat fibromyalgia.

Objective: To investigate the comparative effectiveness and acceptability associated with pharmacological treatment options for fibromyalgia.

Data sources: Searches of PubMed/MEDLINE, Cochrane Library, Embase, and Clinicaltrials.gov were conducted on November 20, 2018, and updated on July 29, 2020.

Study selection: Randomized clinical trials (RCTs) comparing amitriptyline or any FDA-approved doses of investigated drugs.

Data extraction and synthesis: This study follows the Preferred Reporting Items for Systematic Reviews and Meta-analyses reporting guideline. Four independent reviewers extracted data using a standardized data extraction sheet and assessed quality of RCTs. A random-effects bayesian network meta-analysis (NMA) was conducted. Data were analyzed from August 2020 to January 2021.

Main outcomes and measures: Comparative effectiveness and acceptability (defined as discontinuation of treatment owing to adverse drug reactions) associated with amitriptyline (off-label), pregabalin, duloxetine, and milnacipran (on-label) in reducing fibromyalgia symptoms. The following doses were compared: 60-mg and 120-mg duloxetine; 150-mg, 300-mg, 450-mg, and 600-mg pregabalin; 100-mg and 200-mg milnacipran; and amitriptyline. Effect sizes are reported as standardized mean differences (SMDs) for continuous outcomes and odds ratios (ORs) for dichotomous outcomes with 95% credible intervals (95% CrIs). Findings were considered statistically significant when the 95% CrI did not include the null value (0 for SMD and 1 for OR). Relative treatment ranking using the surface under the cumulative ranking curve (SUCRA) was also evaluated.

Results: A total of 36 studies (11 930 patients) were included. The mean (SD) age of patients was 48.4 (10.4) years, and 11 261 patients (94.4%) were women. Compared with placebo, amitriptyline was associated with reduced sleep disturbances (SMD, -0.97; 95% CrI, -1.10 to -0.83), fatigue (SMD, -0.64; 95% CrI, -0.75 to -0.53), and improved quality of life (SMD, -0.80; 95% CrI, -0.94 to -0.65). Duloxetine 120 mg was associated with the highest improvement in pain (SMD, -0.33; 95% CrI, -0.36 to -0.30) and depression (SMD, -0.25; 95% CrI, -0.32 to -0.17) vs placebo. All treatments were associated with inferior acceptability (higher dropout rate) than placebo, except amitriptyline (OR, 0.78; 95% CrI, 0.31 to 1.66). According to the SUCRA-based relative ranking of treatments, duloxetine 120 mg was associated with higher efficacy for treating pain and depression, while amitriptyline was associated with higher efficacy for improving sleep, fatigue, and overall quality of life.

Conclusions and relevance: These findings suggest that clinicians should consider how treatments could be tailored to individual symptoms, weighing the benefits and acceptability, when prescribing medications to patients with fibromyalgia.

OBJECTIVES:

To systematically assess the evidence of Craniosacral Therapy (CST) for the treatment of chronic pain.

METHODS:

PubMed, Central, Scopus, PsycInfo and Cinahl were searched up to August 2018. Randomized controlled trials (RCTs) assessing the effects of CST in chronic pain patients were eligible. Standardized mean differences (SMD) and 95% confidence intervals (CI) were calculated for pain intensity and functional disability (primary outcomes) using Hedges' correction for small samples. Secondary outcomes included physical/mental quality of life, global improvement, and safety. Risk of bias was assessed using the Cochrane tool.

RESULTS:

Ten RCTs of 681 patients with neck and back pain, migraine, headache, fibromyalgia, epicondylitis, and pelvic girdle pain were included. CST showed greater post intervention effects on: pain intensity (SMD = -0.32, 95%CI = [- 0.61,-0.02]) and disability (SMD = -0.58, 95%CI = [- 0.92,-0.24]) compared to treatment as usual; on pain intensity (SMD = -0.63, 95%CI = [- 0.90,-0.37]) and disability (SMD = -0.54, 95%CI = [- 0.81,-0.28]) compared to manual/non-manual sham; and on pain intensity (SMD = -0.53, 95%CI = [- 0.89,-0.16]) and disability (SMD = -0.58, 95%CI = [- 0.95,-0.21]) compared to active manual treatments. At six months, CST showed greater effects on pain intensity (SMD = -0.59, 95%CI = [- 0.99,-0.19]) and disability (SMD = -0.53, 95%CI = [- 0.87,-0.19]) versus sham. Secondary outcomes were all significantly more improved in CST patients than in other groups, except for six-month mental quality of life versus sham. Sensitivity analyses revealed robust effects of CST against most risk of bias domains. Five of the 10 RCTs reported safety data. No serious adverse events occurred. Minor adverse events were equally distributed between the groups.

DISCUSSION:

In patients with chronic pain, this meta-analysis suggests significant and robust effects of CST on pain and function lasting up to six months. More RCTs strictly following CONSORT are needed to further corroborate the effects and safety of CST on chronic pain.