Sujata Thakurta

Exenatide efficacy and safety: a systematic review

Objective  To examine the efficacy, effectiveness and side effects of exenatide when compared with oral glucose‐lowering agents or insulin therapy.

Depression drug treatment outcomes in pregnancy and the postpartum period: A systematic review and meta-analysis

OBJECTIVE: To evaluate the comparative benefits and harms in both mother and child of antidepressant treatment for depression in pregnant or postpartum women. DATA SOURCES: MEDLINE, the Cochrane Library, CINAHL, Scopus, ClinicalTrials.gov (inception to July 2013), manufacturers, and reference lists. METHODS OF STUDY SELECTION: Two reviewers independently selected studies of pregnant women with depression comparing antidepressants with each other, placebo or no treatment, or nondrug treatments. Studies making comparisons among women taking antidepressants for any reason and those not taking antidepressants (depression status unknown) were used to fill gaps in the evidence. TABULATION, INTEGRATION, AND RESULTS: Dual study data extraction and quality assessment were used. Six randomized controlled trials and 15 observational studies provided evidence. Low-strength evidence suggested neonates of pregnant women with depression taking selective serotonin reuptake inhibitors had higher risk of respiratory distress than did neonates of untreated women (13.9% compared with 7.8%; P<.001) but no difference in risk of neonatal convulsions (0.14% compared with 0.11%; P=.64) or preterm birth (17% compared with 10%; P=.07). Indirect evidence from studies of pregnant women receiving antidepressants for mixed or unreported reasons compared with pregnant women not taking antidepressants (depression status unknown) suggested future research should focus on congenital anomalies and autism spectrum and attention deficit disorders in the child. In postpartum depression, low-strength evidence suggested symptom response was not improved when sertraline was added to psychotherapy or when cognitive–behavioral therapy was added to paroxetine. Evidence was insufficient for other outcomes, including depression symptoms, functional capacity, breastfeeding, and infant and child development. A serious limitation is the lack of study populations of exclusively depressed pregnant and postpartum women. CONCLUSION: Evidence about the comparative benefits and harms of pharmacologic treatment of depression in pregnant and postpartum women was largely inadequate to allow informed decisions about treatment. Considering the prevalence of depression, filling this gap is essential.

Efficacy and Harms of the Hypoglycemic Agent Pramlintide in Diabetes Mellitus

PURPOSE We conducted a study to examine the efficacy, effectiveness, and harms of pramlintide as adjunct therapy in adults and children with type 1 or type 2 diabetes. METHODS We searched multiple bibliographic databases to January 2010, the US Food and Drug Administration Web site, and other sources to identify randomized controlled trials (RCTs) fulfilling inclusion criteria. Syntheses were qualitative because data were too heterogeneous for meta-analysis. RESULTS Three published RCTs in type 1 diabetes and 4 in type 2 disease fulfilled inclusion criteria. All trials were conducted with adults, and none was longer than 52 weeks. In type 1 diabetes with intensive insulin therapy, pramlintide was as effective as placebo in lowering glycated hemoglobin (HbA1c) levels in one trial. Pramlintide was somewhat more effective than placebo in patients using conventional insulin therapy, with a between-group difference in HbA1c levels of 0.2% to 0.3% (2 studies). In patients with type 2 diabetes, pramlintide was more effective at reducing HbA1c levels than placebo when added to flexibly dosed glargine (without prandial insulin) and when added to fixed-dose insulin therapies, with or without oral hypoglycemic agents (between-group differences in HbA1c were approximately 0.4%). Weight loss was observed with pramlintide in both type 1 and type 2 diabetes, whereas placebo-treated patients tended to gain weight. Pramlintide-treated patients experienced more frequent nausea and severe hypoglycemia compared with patients treated with placebo. CONCLUSIONS Pramlintide was somewhat more effective than placebo as adjunct therapy for improving HbA1c levels and weight in adults with type 1 diabetes on conventional insulin therapy, or type 2 diabetes and inadequate glycemic control with their current therapies, with between-group differences in HbA1c levels in the range of 0.2% to 0.4%. Further research is needed to determine pramlintide’s durability of hypoglycemic effect, as well as effects on patient-reported outcomes, morbidity, mortality, and long-term harms.

Aggregator: a machine learning approach to identifying MEDLINE articles that derive from the same underlying clinical trial

OBJECTIVE It is important to identify separate publications that report outcomes from the same underlying clinical trial, in order to avoid over-counting these as independent pieces of evidence. METHODS We created positive and negative training sets (comprised of pairs of articles reporting on the same condition and intervention) that were, or were not, linked to the same clinicaltrials.gov trial registry number. Features were extracted from MEDLINE and PubMed metadata; pairwise similarity scores were modeled using logistic regression. RESULTS Article pairs from the same trial were identified with high accuracy (F1 score=0.843). We also created a clustering tool, Aggregator, that takes as input a PubMed user query for RCTs on a given topic, and returns article clusters predicted to arise from the same clinical trial. DISCUSSION Although painstaking examination of full-text may be needed to be conclusive, metadata are surprisingly accurate in predicting when two articles derive from the same underlying clinical trial.