剂量-反应Meta分析系列之非随机误差导致的偏倚的处理方法

作为一种有效的系统评价方法,剂量-反应Meta分析被广泛用于探讨自变量与因变量之间的因果关系,并且多基于观察性研究。观察性研究样本量大,进行Meta分析时可提供较为充分的统计效能。但由于自身设计问题,观察性研究容易混入多种偏倚,而这些偏倚可能会影响最终结果,导致其偏离真实值。因方法学上的盲区,当前并无特定针对剂量-反应Meta分析的偏倚校正方法,本文借鉴观察性研究Meta分析中的偏倚校正方法,通过调整后将其应用于剂量-反应Meta分析模型中,并对各类校正方法的优缺点进行归纳比较。     

The relationship between new-onset diabetes mellitus and pancreatic cancer risk: A case-control study

Diabetes mellitus (DM) is widely considered to be associated with pancreatic cancer, however, whether DM is a cause or consequence of pancreatic cancer is controversial. In the present study, 1458 patients with pancreatic ductal adenocarcinoma (PDAC) and 1528 age-, sex- and sociodemographic variables-matched controls were recruited in two university-affiliated hospitals from 1st January 2000 to 31st December 2009. DM was defined as fasting blood glucose (FBG) level of 7.0 mmol/L or greater. An unconditional multivariable logistic regression analysis was used to estimate adjusted odds ratios (AORs) and 95% confidence interval (CI). Compared with controls, a moderate increased risk of PDAC was observed among cases with long-standing diabetes (?2-year duration), with an AOR (95% CI) of 2.11 (1.51-2.94). Interestingly, a significant higher risk was observed among cases with new-onset DM (<2-year duration), with an AOR of 4.43 (3.44-5.72) compared to controls without DM. In addition, we found a synergistic interaction between cigarette smoking and DM on modifying the risk of pancreatic cancer development (AOR = 6.17, 95% CI 3.82-9.94). Similarly, a synergistic interaction between new-onset DM and family history of pancreatic cancer was found for pancreatic cancer risk, with an AOR (95% CI) of 11.04 (2.51-48.53). This study suggested that DM could be both an early manifestation of pancreatic cancer and an aetiologic factor. Possible effect modification on DM by family history of pancreatic cancer and smoking status should be further explored in future aetiologic studies.     

Advanced HER2-positive gastric cancer: Current and future targeted therapies

The prognostic value of human epidermal growth factor receptor 2 (HER2) in gastric cancer is controversial. Consensus guidelines have standardized the testing of HER2 status in gastric cancer. Overexpression of this receptor occurs in approximately 20% of gastric and gastro-esophageal junction adenocarcinomas, predominantly those of the intestinal type. Recently, trastuzumab has emerged as the first targeted drug to improve overall survival when combined with chemotherapy in advanced HER2-positive gastric cancer. Primary and secondary resistance to trastuzumab has become a major problem and new strategies to overcome this resistance are needed. A high percentage of advanced HER2-positive gastric cancer patients who progress on trastuzumab therapy are candidates for second-line therapy. New families of targeted drugs, including tyrosine kinase inhibitors (TKIs) such as lapatinib and PF-00299804, mammalian target of rapamycin (mTOR) pathway inhibitors such as everolimus, heat-shock protein 90 (HSP90) inhibitors such as AUY922, HER dimerization inhibitors such as pertuzumab, and antibody-chemotherapy conjugates such as trastuzumab-emtansine (T-DM1), could offer alternative second-line treatments when trastuzumab-based first-line therapy fails.     

Disappointment and drop-out rate after being allocated to control group in a smoking cessation trial

BackgroundIf a patient agrees to take part in a randomised trial it is reasonable to presume that the patient would prefer to be allocated into the intervention. This study's aim was to investigate how patients react after they have been randomised into control group.MethodsNested study within two randomised trials. Telephone interviews with a structured questionnaire. The participants were invited after they had been randomised into the control group in two smoking cessation trials. The main outcome measures were reaction to control group allocation and drop-out rates.ResultsTwenty-seven out of 30 possible interviews were successfully completed. Fourteen persons expressed that they were disappointed of being allocated to the control group. Five persons said that they had not understood the consent information and three of these were very disappointed. Surprisingly these three persons said that they had not expected a randomization. A woman expressed that she “felt as if I was being swindled”. There were in total 9/117 (7.7%) lost to follow-up in the control group and there were 4/105 (3.8%) losses to follow-up in the intervention group (P = 0.26). Active withdrawal of consent was slightly higher among the control group, five in the control group (4.3%) and no active withdrawals in the intervention group (P = 0.06).ConclusionsDisappointment was common after allocation to the control group. This is a probable explanation of the higher drop-out rate in the control group. The consent information is of highest importance since those who were very disappointed claimed they did not receive understandable information.     

Hyperalgesia in Opioid-Managed Chronic Pain and Opioid-Dependent Patients

This observational study aimed to determine whether pain sensitivity in patients with noncancer chronic pain, taking either methadone or morphine, is similar to patients maintained on methadone for dependence therapy, compared with a control group. Nociceptive thresholds were measured on a single occasion with von Frey hairs, electrical stimulation, and cold pressor tests. In all subjects receiving methadone or morphine, nociceptive testing occurred just before a scheduled dose. Cold pressor tolerance values in patients with noncancer, chronic pain, treated with morphine and methadone, were 18.1 ± 2.6 seconds (mean ± SEM) and 19.7 ± 2.3 seconds, respectively; in methadone-maintained subjects it was 18.9 ± 1.9 seconds, with all values being significantly (P < .05) lower than opioid-naïve subjects (30.7 ± 3.9 seconds). These results indicate that patients with chronic pain managed with opioids and methadone-maintained subjects are hyperalgesic when assessed by the cold pressor test but not by the electrical stimulation test. None of the groups exhibited allodynia as measured using the von Frey hairs. These results add to the growing body of evidence that chronic opioid exposure increases sensitivity to some types of pain. They also demonstrate that in humans, this hyperalgesia is not associated with allodynia.PerspectiveThis article presents an observational study whereby the pain sensitivity of patients with chronic pain managed with opioids and opioid-maintained patients were compared with opioid-naïve patients. The results suggest that opioid use may contribute to an increase in the sensitivity to certain pain experimental stimuli.     

Nonhormonal therapies for menopausal hot flashes: systematic review and meta-analysis

Context  Concern regarding the adverse effects of estrogen and other hormones for treating menopausal symptoms has led to demand for other options; however, the efficacy and adverse effects of nonhormonal therapies are unclear. Objective  To assess the efficacy and adverse effects of nonhormonal therapies for menopausal hot flashes by reviewing published randomized controlled trials. Data Sources  MEDLINE (1966-October 2005), PsycINFO (1974-October 2005), and the Cochrane Controlled Clinical Trials Register Database (1966-October 2005) were searched for relevant trials that provided data on treatment of menopausal hot flashes using 1 or more nonhormonal therapies. Study Selection  All English-language, published, randomized, double-blind, placebo-controlled trials of oral nonhormonal therapies for treating hot flashes in menopausal women measuring and reporting hot flash frequency or severity outcomes. Data Extraction  Trials were identified, subjected to inclusion and exclusion criteria, and reviewed. Data on participants, interventions, and outcomes were extracted and trials were rated for quality based on established criteria. A meta-analysis was conducted for therapies with sufficient trials reporting hot flash frequency outcomes. Data Synthesis  From 4249 abstracts, 43 trials met inclusion criteria, including 10 trials of antidepressants, 10 trials of clonidine, 6 trials of other prescribed medications, and 17 trials of isoflavone extracts. The number of daily hot flashes decreased compared with placebo in meta-analyses of 7 comparisons of selective serotonin reuptake inhibitors (SSRIs) or serotonin norepinephrine reuptake inhibitors (SNRIs) (mean difference, –1.13; 95% confidence interval [CI], –1.70 to –0.57), 4 trials of clonidine (–0.95; 95% CI, –1.44 to –0.47), and 2 trials of gabapentin (–2.05; 95% CI, –2.80 to –1.30). Frequency was not reduced in meta-analysis of trials of red clover isoflavone extracts and results were mixed for soy isoflavone extracts. Evidence of the efficacy of other therapies is limited due to the small number of trials and their deficiencies. Trials do not compare different therapies head-to-head and relative efficacy cannot be determined. Conclusion  The SSRIs or SNRIs, clonidine, and gabapentin trials provide evidence for efficacy; however, effects are less than for estrogen, few trials have been published and most have methodological deficiencies, generalizability is limited, and adverse effects and cost may restrict use for many women. These therapies may be most useful for highly symptomatic women who cannot take estrogen but are not optimal choices for most women.      

Correlation of quality measures with estimates of treatment effect in meta-analyses of randomized controlled trials

Context  Specific features of trial quality may be associated with exaggeration or shrinking of the observed treatment effect in randomized studies. Therefore, assessment of trial quality is often used in meta-analysis. However, the degree to which specific quality measures are associated with treatment effects has not been well established across a broad range of clinical areas. Objective  To determine if quality measures are associated with treatment effect size in randomized controlled trials (RCTs). Design  Quality measures from published quality assessment scales were evaluated in RCTs included in meta-analyses from 4 medical areas (cardiovascular disease, infectious disease, pediatrics, and surgery). Included meta-analyses incorporated at least 6 RCTs, examined dichotomous outcomes, and demonstrated significant between-study heterogeneity in the odds ratio (OR) scale. Main Outcome Measures  Relative ORs comparing overall treatment effect (summary OR) of high vs low-quality studies, as determined by each quality measure, with relative ORs less than 1 indicating larger treatment effect in low-quality studies. Results  Twenty-four quality measures were analyzed for 276 RCTs from 26 meta-analyses. Relative ORs of high vs low-quality studies for these quality measures ranged from 0.83 to 1.26; none was statistically significantly associated with treatment effect. The proportion of studies fulfilling specific quality measures varied widely in the 4 medical areas. In analyses limited to specific medical areas, placebo control, multicenter studies, study country, caregiver blinding, and statistical methods were significantly associated with treatment effect on 7 occasions. These relative ORs ranged from 0.40 to 1.74. However, the directions of these associations were not consistent. Conclusions  Individual quality measures are not reliably associated with the strength of treatment effect across studies and medical areas. Although use of specific quality measures may be appropriate in specific well-defined areas in which there is pertinent evidence, findings of associations with treatment effect cannot be generalized to all clinical areas or meta-analyses.      

Economic comparison of diagnostic approaches for evaluating osteoporosis in older women

Of the technologies available, dual-energy X-ray absorptiometry of the hip or femoral neck (DXA-FN) is the best predictor of hip fractures. Diagnostic approaches utilizing measures of peripheral sites with office-based technology, such as calcaneal quantitative ultrasound (QUS), may reduce costs although clinical and economic outcomes have not been evaluated. The objective was to compare three approaches for diagnosing osteoporosis in older women. The design was a decision-analytic model using diagnostic measures and clinical and economic outcomes from the Study of Osteoporotic Fractures, a prospective cohort of older white women that measured BMD and QUS and assessed fracture outcomes. The setting and patients were a hypothetical cohort of older white women presenting for diagnosis of osteoporosis. For the diagnostic and treatment alternatives, three diagnostic approaches—DXA-FN alone, QUS alone and a sequential approach (first QUS, then DXA-FN for those with low values for QUS)—were compared to no diagnosis. The outcome measures were the number of women identified for treatment, number of hip fractures prevented following diagnosis and subsequent treatment, number of women needed to treat to prevent one hip fracture and total direct medical costs. The sequential approach identified fewer women to treat, prevented more hip fractures and incurred lower total costs than using DXA alone. Diagnosis with QUS alone identified more women to treat and incurred higher total costs than DXA alone under most conditions. Compared to other approaches for diagnosing osteoporosis, sequential use of QUS followed by DXA resulted in fewer women treated and lower total costs.This study was conducted by the Oregon Health and Sciences University Evidence-based Practice Center under contract to the Agency for Healthcare Research and Quality, contract no. 290-97-0018, task order no. 4, Rockville, Md. AHRQ Disclaimer: The authors of this report are responsible for its content. No statements in this report should be construed as the official position of the Agency for Healthcare Research and Quality or the US Department of Health and Human Services.