Statistical controversies in clinical research: Value of adverse events relatedness to study treatment: analyses of data from randomized double-blind placebo-controlled clinical trials

Background

Collection and reporting of adverse events (AEs) and their relatedness to study treatment, known commonly as attribution, in clinical trials is mandated by regulatory agencies (the National Cancer Institute and the Food and Drug Administration). Attribution is assigned by the treating physician using judgment based on various factors including patient's baseline status, disease history, and comorbidity as well as knowledge about the safety profile of the study treatments. We evaluate the patterns of AE attribution (unrelated, unlikely, possibly, probably, and definitely related to the treatment) in treatment, symptom intervention (cancer patients) and cancer prevention (participants at high risk for cancer) setting.

Materials and methods

Nine multicenter placebo-controlled trials (two treatment, two symptom intervention, and five cancer prevention) were analysed separately (2155 patients). Frequency and severity of AEs were summarized by arm. Attribution and percentage of repeated AEs whose attribution changed overtime were summarized for the placebo arms. Percentage of physician over- or under-reporting of AE relatedness was calculated for the treatment arms using the placebo arm as the reference.

Results

Across all trials and settings, a very high proportion of AEs reported as related to treatment were classified as possibly related, a significant proportion of AEs in the placebo arm were incorrectly reported as related to treatment, and clinician-reported attribution over-estimated the rate of AEs related to treatment. Fatigue, nausea, vomiting, diarrhea, constipation, and neurosensory were the common AEs that were over reported by clinician as related to treatment.

Conclusions

These analyses demonstrate that assigning causality to AE is a complex and difficult process that produces unreliable and subjective data. In randomized double-blind placebo-controlled trials where data are available to objectively assess relatedness of AE to treatment, attribution assignment should be eliminated.

Delivery of small-for-gestational-age neonate and association with early-onset impaired maternal endothelial function

OBJECTIVE

Women who have delivered a small-for-gestational-age (SGA) infant are at an increased risk of developing cardiovascular disease (CVD) in later life. Endothelial dysfunction is a subclinical sign of early CVD. It is unknown whether women who have recently had a pregnancy complicated by SGA, in the absence of other maternal and fetal diseases, have subclinical endothelial dysfunction. Our aim was to assess maternal endothelial function 6 months after a pregnancy complicated by SGA.

METHODS

This was a case-control study conducted in a tertiary referral hospital in London, UK, over a 15-month period. Flow-mediated dilatation (FMD) of the brachial artery was measured in women 6.9 ± 2.5 months after childbirth. Forty-four women were included in the study, of whom 15 had a SGA neonate (mean ± SD customized birth centile of 1.9 ± 2.3) and 29 delivered an appropriately grown baby (mean ± SD customized birth centile of 47.5 ± 26.3). The primary continuous variable, FMD, was assessed in each group and compared using unpaired t-test.

RESULTS

Women who had a SGA neonate had lower postpartum FMD (6.79 ± 0.95%) than did those who had an appropriately grown offspring (10.26 ± 2.44% (95% CI for difference between groups, -5.37 to -1.57); P = 0.0007). There were no differences in postnatal maternal blood pressure, abdominal circumference, weight and glucose, insulin and lipid profiles between the two groups.

CONCLUSIONS

Women who had a pregnancy affected by SGA, probably due to placental failure in the absence of pre-eclampsia, have evidence of subclinical endothelial dysfunction within 6 months of childbirth. These women may benefit from lifestyle measures focused on the primary prevention of CVD. Further research in larger populations is needed to ascertain if such postpartum maternal endothelial dysfunction is a pregnancy-induced phenomenon or if it is related to the pre-existing maternal phenotype, and whether it persists long term. Copyright © 2016 ISUOG. Published by John Wiley & Sons Ltd.

Questionable value of attribution when interpreting adverse event data: A joint evaluation by North Central Cancer Treatment Group (NCCTG) and American College of Surgeons Oncology Group (ACOSOG)

6511

Background: In March 1998, the Common Toxicity Criteria (CTC) version 2.0 introduced the collection of attribution of adverse events (AE) to study drug. Collection of attribution adds time and cost to the clinical trial process, and reporting in the literature has varied. At ASCO 2006 results were presented from NCCTG trial 97–24–51 that showed ∼ 50% of events on the placebo arm were reported as attributed to study drug and 25% of the time the same event was not consistently attributed. We sought to validate these results in a second placebo controlled trial. Methods: Patients on the placebo arm of the phase III ACOSOG trial Z9001 (imatinib mesylate vs. placebo after resection of primary gastrointestinal stromal tumors) were evaluated since the reported AEs could not be due to study drug. Attribution was collected per CTCv3.0 as not related, unlikely, possibly, probably, or definitely related, and categorized as “unrelated” (not related or unlikely) and “related” (possible, probable, or definite). Patterns were evaluated using generalized estimating equations adjusting for multiple events per patient. Results: 202 patients and 2871 AEs were included. 52% of AEs were reported as “related”. Known imatinib AEs were more likely to be reported as “related”. The reported rate of AE relatedness varied by AE type. No patterns were observed by gender, age, PS, severity or visit of AE. When the same event was reported on the same patient on multiple visits, the attribution category changed at least once 33% of the time with 19% changing from “related” to “unrelated” or vice versa. Conclusion: 52% of events on the placebo arm of Z9001 were reported as attributed to study drug and 19% of the time the same event was not consistently attributed. These data provide further evidence that attribution is difficult to determine, unreliable, and of questionable value in interpreting randomized clinical trials.

No significant financial relationships to disclose.

[Table: see text]

A front-line window of opportunity phase II study of sorafenib in patients with advanced non-small cell lung cancer: A North Central Cancer Treatment Group study

7547

Background: Sorafenib is a multikinase inhibitor with single-agent activity in previously treated NSCLC. In an effort to evaluate its single agent activity in previously untreated NSCLC, the NCCTG undertook a front-line “window of opportunity” study. Materials and Methods: Patients with stage IIIB (pleural effusion) or stage IV NSCLC received sorafenib dosed at 400 mg BID continuously, with a cycle defined as 4 weeks. Patients were evaluated weekly during the first 2 cycles with those who progressed rapidly going on to receive standard chemotherapy. Based on a two-stage Fleming design, if only at most 1 confirmed response was observed in the first 20 patients enrolled to stage I, the regimen would be considered ineffective. If 2 or more responses were observed, the study would proceed to stage 2 and accrue an additional 22 patients. If 5 or more confirmed responses were observed, the regimen would be recommended for further testing. Results: The study did not meet the stage I efficacy criteria (only 1 confirmed partial response in the first 20 patients) and was permanently closed after enrolling 25 evaluable patients [15 females, 10 males; 4 stage IIIB, 21 stage IV; median age 67 (45–85)]. 2 patients are still receiving treatment (14 and 15 months). No grade 3 or higher hematologic adverse events were observed. Thirteen (52%) patients had a grade 3 non-hematologic adverse event with fatigue (20%), diarrhea (8%), and dyspnea (8%) being the most common. There was one grade 4 pulmonary hemorrhage. A total of 3 (12%) PRs; and 7 (28%) SD were observed in the 25 patients 7 (28%) patients were progression- free at 24 weeks. Median survival and median time to progression were 8.8 and 2.9 months respectively. Conclusion: While the pre- specified efficacy endpoints were not met, the objective response rate of 12% and median survival of 8.8 months suggest that single agent sorafenib has activity similar to two-drug combinations. The feasibility and utility of the “window of opportunity” design in estimating the activity of novel compounds was demonstrated. Finally, the single-agent activity of sorafenib argues for combination studies with standard chemotherapy agents.

No significant financial relationships to disclose.

Evaluation of platinum-based chemotherapy, glutathione metabolic genes, and survival in advanced non small cell lung cancer: A NCCTG 97–24–51 based study

13013

Background: The glutathione metabolic pathway is directly involved in the inactivation of platinum compounds. Our earlier work indicated that genotypes corresponding to varied glutathione-related enzyme activities may predict survival in patients with advanced lung cancer. Herein, we evaluated the role of glutathione-related genotypes on clinical outcomes in stage IIIB/IV non-small cell lung cancer patients who were stable or responding from one prior platinum-based chemotherapy. Methods: DNA samples from patients who had tumor regression or stable disease after treatment with one platinum based chemotherapy regimen were analyzed using 6 polymorphic DNA markers that encode 5 important enzymes in the glutathione metabolic pathway. The contrasting genotypes used in the analysis were GCLC (homozygous repeat 77 vs. heterozygous 7*), GPX1 (CC vs. other), GSTP1-I105V (AA vs. other), GSTP1-A114V (CC vs. other), GSTM1 (null vs. present), and GSTT1 (null vs. present). Multivariable Cox proportional hazards models adjusted for age, gender, treatment arm, ECOG performance status, stage, and prior tumor response, were used to evaluate the prognostic significance of the genotypes at each locus. Results: Data from 112 patients with a median follow-up of 10.6 (range: 0.4–56.3) months was used. No significant difference in time to disease progression was observed for any of the genotypes. Among the 6 genomic DNA markers, the GCLC genotype was an important prognostic factor for overall survival after adjusting for other factors. In particular, patients carrying a 77 repeat genotype (homozygous) had a significantly worse survival (Hazard ratio = 1.67, 95% CI: 1.05–2.63). Conclusions: Genotypes of glutathione-related enzymes, especially GCLC, may be used as host factors in predicting patients’ prognosis after platinum-based chemotherapy. Further investigation to define and measure the effects of these genes in chemotherapeutic regimens is needed. (This work was partly supported by NIH research grants CA77118, CA80127, and CA84354.)

No significant financial relationships to disclose.

Should attribution be considered when interpreting adverse event data: A North Central Cancer Treatment Group (NCCTG) evaluation of a phase III placebo controlled trial

6006

Background: In March 1998, the Common Toxicity Criteria version 2.0 was implemented and introduced the collection and reporting of attribution of adverse events to study treatment. Collection and reporting of attribution adds time and cost to the clinical trial process, and its use while reporting adverse events (AE) in the literature has varied. We investigate whether attribution adds value to the interpretation of AE data. Methods: Patients on the placebo arm of trial 97–24–51 (CAI vs. Placebo) were chosen since the true relationship of the event to the study treatment is known (“unrelated”). Attribution was collected per CTCv2.0 and categorized as “not related” (not related or unlikely) and “related” (possible, probable, or definite). All reported AEs and the maximum severity for a given AE were evaluated. Patterns were evaluated using generalized estimating equations adjusting for multiple events per patient. Changes in attribution over time for the same event were computed. Results: 84 patients on the placebo arm experienced a total of 1013 AEs. 47% of AEs were reported as “related” with 36% reported as possibly related. Known CAI AEs were more likely to be reported as “related” (p = 0.005). Pulmonary AEs were more likely to be reported as “not related” (p = 0.0006). No patterns were observed by gender, age, PS, severity or treatment cycle of AE. When the same event was reported on the same patient in multiple cycles, the attribution category changed at least once 36% of the time with 25% changing from “related” to “unrelated” or vice versa. Conclusion: Almost 50% of events on a placebo arm of a phase III trial were reported as being attributed to study treatment and 25% of the time the same event was not consistently attributed. These data suggest that attribution is difficult to determine, unreliable, and thus its value in large phase III trials is questionable.

[Table: see text]

No significant financial relationships to disclose.

Behavioral disturbances in geropsychiatric inpatients across dementia types

The objective of this study was to compare differences in behavioral, psychiatric, and cognitive status among geropsychiatric inpatients with Alzheimer's, vascular, alcohol-induced, and mixed dementia. Participants included 150 patients with dementia consecutively admitted to an acute geropsychiatric inpatient unit. Measures included the Mini-Mental State Examination, Cohen-Mansfield Agitation Inventory, Cumulative Illness Rating Scale, Basic and Independent Activities of Daily Living, Positive and Negative Syndrome Scale for Schizophrenia, and the Initiation/Perseveration subscale of the Dementia Rating Scale. No significant differences existed in the character or severity of agitation among patients with Alzheimer's, vascular, alcohol-related and mixed dementia. Interestingly, patients with vascular dementia compared to patients with other dementias admitted for behavioral disturbances were less cognitively impaired and more medically burdened.

Liquid chromatographic assay for a glucose tetrasaccharide, a putative biomarker for the diagnosis of Pompe disease

A HPLC method associated with butyl-p-aminobenzoate derivatization has been developed for the analysis of a tetraglucose oligomer, Glcalpha1-6Glcalpha1-4Glcalpha1-4Glc, designated Glc(4), in biological fluids. This tetraglucose, normally excreted in the urine, has previously been shown to be elevated in a number of pathological conditions including Pompe disease (glycogen storage disease type II), which is caused by a deficiency of the lysosomal enzyme acid alpha-glucosidase. Concentrations of Glc(4) in both urine and plasma were established for the age ranges of <1, 1-5, 6-10, 11-20, and >20 years, both in normal individuals and in a cohort of 21 patients with enzymatically confirmed Pompe disease. The Glc(4) concentration decreased with age in both groups, but all the patients had elevated Glc(4) levels compared with age-matched controls. Electrospray tandem mass spectrometry was employed to establish the homogeneity of the HPLC peak for Glc(4) and to investigate the identity of other unusual oligosaccharides excreted in patient urine. Our results demonstrate that this method is suitable for application in clinical laboratories to help establish the diagnosis of Pompe disease.

Functional impairment in COPD patients: the impact of anxiety and depression

The authors examined the relationship between functional status and comorbid anxiety and depression and the relationship between utilization of health care resources and psychopathology in elderly patients with chronic obstructive pulmonary disease (COPD). Elderly male veterans (N = 43) with COPD completed anxiety, depression, and functional status measures. The authors constructed regression models to explore the contribution of COPD severity, medical burden, depression, and anxiety to the dependent variables of functional impairment and health care utilization. Anxiety and depression contributed significantly to the overall variance in functional status of COPD patients, over and above medical burden and COPD severity, as measured by the 8 scales of the Medical Outcomes Study (MOS) 36-item Short Form Health Survey. Surprisingly, medical burden and COPD severity did not contribute significantly to overall variance in functional status. Few patients were receiving any treatment for anxiety or depression.

Validation of accuracy-based amino acid reference materials in dried-blood spots by tandem mass spectrometry for newborn screening assays

BACKGROUND

Advances in technology and the earlier release of newborns from hospitals have pressed the demand for accurate calibration and improved interlaboratory performance for newborn screening tests. As a first step toward standardization of newborn screening aminoacidopathy tests, we have produced six-pool sets of multianalyte dried-blood-spot amino acid reference materials (AARMs) containing predetermined quantities of five amino acids. We describe here the production of the AARMs, validation of their amino acid contents, and characterization of their homogeneity and their stability in storage.

METHODS

To each of six portions of a pool of washed erythrocytes suspended in serum we added Phe (0-200 mg/L), Leu (0-200 mg/L), Met (0-125 mg/L), Tyr (0-125 mg/L), and Val (0-125 mg/L). Six-pool sets (1300) were prepared, dried, and packaged. We used isotope-dilution mass spectrometry to estimate the endogenous amino acid concentrations of the AARMs and validate their final amino acid concentrations. We used additional tandem mass spectrometry analyses to examine the homogeneity of amino acid distribution in each AARM, and HPLC analyses to evaluate the stability of the amino acid contents of the AARMs.

RESULTS

The absolute mean biases across the analytic range for five amino acids were 2.8-9.4%. One-way ANOVAs of the homogeneity results predicted no statistically significant differences in amino acid concentrations within the blood spots or within the pools (P >0.05). Regression slopes (0 +/- 0.01) for amino acid concentrations vs storage times and their P values (>0.05) showed no evidence of amino acid degradation at ambient temperatures, 4 degrees C, or -20 degrees C during the intervals tested.

CONCLUSION

The validation, homogeneity, and stability of these blood spots support their use as a candidate national reference material for calibration of assays that measure amino acids in dried-blood spots.

Diagnosis of mitochondrial trifunctional protein deficiency in a blood spot from the newborn screening card by tandem mass spectrometry and DNA analysis

Trifunctional protein (TFP) plays a significant role in the mitochondrial beta-oxidation of long-chain fatty acids. Its deficiency impairs the energy generating function of this pathway and causes hypoketotic hypoglycemia once hepatic glycogen stores are depleted. A Reye-like syndrome, cardiomyopathy, and sudden death may follow. The diagnosis is based on demonstration of significantly decreased enzyme activity of at least two of the three involved enzymes in fibroblasts. The possibility of prospective diagnosis of TFP deficiency by newborn screening using tandem mass spectrometry (MS/MS) has not been evaluated. We report the postmortem diagnosis of a male newborn, who suffered sudden death at 2 wk of age, and his younger sister, who died of cardiomyopathy complicated by acute heart failure at the age of 6 mo, after she had acquired a common viral infection. Blood spots from the original newborn screening cards were the only remaining material from the patients. Analysis by MS/MS revealed acylcarnitine profiles consistent with either TFP or long-chain 3-hydroxyacyl coenzyme A dehydrogenase (LCHAD) deficiency. To prove the diagnosis, the alpha- and beta-subunit genes coding for TFP were examined. The patients were compound heterozygous for a 4-bp-deletion and an a-->g missense mutation, both in the same exon 3 donor consensus splice site. This is the first report of the diagnosis of TFP deficiency using blood spots obtained for newborn screening and suggests that TFP deficiency may be detectable by prospective newborn screening using MS/MS.

Use of phenylalanine-to-tyrosine ratio determined by tandem mass spectrometry to improve newborn screening for phenylketonuria of early discharge specimens collected in the first 24 hours

We compared the screening interpretation of fluorometric analytical results for phenylketonuria (PKU) with tandem mass spectrometry (MS/MS) in filter paper blood spots collected from newborns <24 h of age. In MS/MS, both Phe and Tyr are quantified. Two hundred and eight blood spots collected from infants <24 h of age were retrieved from storage from the California newborn screening program. These samples had been categorized on the basis of fluorometric analysis as initial negative, initial positive for hyperphenylalaninemia with negative determination on recall, or initial positive for hyperphenylalaninemia and confirmed on follow up as PKU or variant hyperphenylalaninemia. The retrieved samples were analyzed in a blinded fashion using MS/MS. Correlation analysis of fluorometry vs MS/MS for Phe concentration was high, with a Pearson correlation coefficient of 0.817. When 180 micromol/L was used as the cutoff Phe concentration for MS/MS and 258 micromol/L was used as the cutoff for fluorometry, all infants with confirmed classical PKU and variant hyperphenylalaninemia were detected. MS/MS analysis reduced the number of false-positive results from 91 to 3. Simultaneous quantification of Phe and Tyr by MS/MS with the use of a cutoff Phe/Tyr molar ratio of 2.5 further reduced the number of false positives to 1. Samples from affected infants showed a discernible trend of increasing Phe concentration and Phe/Tyr molar ratio with age of collection. These results demonstrate the utility of MS/MS in the routine PKU screening of early-discharge newborns. MS/MS reduces the false-positive rate of fluorometric screening almost 100-fold because of the improved accuracy and precision of Phe measurement and simultaneous confirmation with the Phe/Tyr molar ratio. In addition to the detection of PKU, MS/MS can also detect other aminoacidopathies and disorders of fatty acid and organic acid metabolism with lower false-positive rates than other methods currently used in newborn screening programs.

Rapid diagnosis of MCAD deficiency: quantitative analysis of octanoylcarnitine and other acylcarnitines in newborn blood spots by tandem mass spectrometry

We report the application of tandem mass spectrometry to prospective newborn screening for medium-chain acyl-CoA dehydrogenase (MCAD) deficiency. MCAD deficiency is diagnosed from dried blood spots on filter paper cards from newborns on the basis of the increase of medium chain length acylcarnitines identified by isotope dilution mass spectrometry methods. A robust and accurate semiautomated method for the analysis of medium chain length acylcarnitines as their butyl esters was developed and validated. Quantitative data from the analyses of 113 randomly collected filter paper blood spots from healthy newborns showed low concentrations of medium chain length acylcarnitines such as octanoylcarnitine. The maximum concentration of octanoylcarnitine was 0.22 mumol/L, with the majority being at or below the detection limit. In all 16 blood spots from newborns with confirmed MCAD deficiency, octanoylcarnitine was highly increased [median 8.4 mumol/L (range 3.1-28.3 mumol/L)], allowing easy detection. The concentration of octanoylcarnitine was significantly higher in these 16 newborns (< 3 days of age) than in 16 older patients (ages 8 days to 7 years) with MCAD deficiency (median 1.57 mumol/L, range 0.33-4.4). The combined experience of prospective newborn screening in Pennsylvania and North Carolina has shown a disease frequency for MCAD deficiency of 1 in 17,706. No false-positive and no known false-negative results have been found. A validated method now exists for prospective newborn screening for MCAD deficiency.

Rapid diagnosis of homocystinuria and other hypermethioninemias from newborns' blood spots by tandem mass spectrometry

We report a new method for the diagnosis of homocystinuria and other hypermethioninemias from dried blood spots on newborn screening cards, based on isotope-dilution tandem mass spectrometry. The mean concentration of methionine in 909 unaffected newborns was 19 micromol/L (CV 44%). The variability of results was reduced when the concentration of methionine was expressed relative to that of another amino acid in the same specimen. The mean ratio of methionine to leucine plus isoleucine for these same newborn blood spots was 0.16 (CV 25%). In newborn samples from a collection categorized by a Guthrie bacterial inhibition assay as true positive, unaffected, or falsely positive for hypermethioninemias, the ratio of methionine to leucine for each true-positive specimen was at least 2.5 times greater than for respective age-matched unaffected blood specimens. The ratio for falsely positive samples did not differ from that for unaffected blood samples. We predict that the ratio of methionine to leucine plus isoleucine determined by tandem mass spectrometry will successfully detect hypermethioninemias with very low rates for false positives and false negatives.

Rapid diagnosis of homocystinuria and other hypermethioninemias from newborns' blood spots by tandem mass spectrometry

We report a new method for the diagnosis of homocystinuria and other hypermethioninemias from dried blood spots on newborn screening cards, based on isotope-dilution tandem mass spectrometry. The mean concentration of methionine in 909 unaffected newborns was 19 micromol/L (CV 44%). The variability of results was reduced when the concentration of methionine was expressed relative to that of another amino acid in the same specimen. The mean ratio of methionine to leucine plus isoleucine for these same newborn blood spots was 0.16 (CV 25%). In newborn samples from a collection categorized by a Guthrie bacterial inhibition assay as true positive, unaffected, or falsely positive for hypermethioninemias, the ratio of methionine to leucine for each true-positive specimen was at least 2.5 times greater than for respective age-matched unaffected blood specimens. The ratio for falsely positive samples did not differ from that for unaffected blood samples. We predict that the ratio of methionine to leucine plus isoleucine determined by tandem mass spectrometry will successfully detect hypermethioninemias with very low rates for false positives and false negatives.

Rapid diagnosis of maple syrup urine disease in blood spots from newborns by tandem mass spectrometry

We report a new method for the diagnosis of maple syrup urine disease (MSUD) from dried blood spots on newborn screening cards based on tandem mass spectrometry (MS-MS). The mean +/- SD concentration of Leu plus Ile in normal newborns was 151 +/- 47 mumol/L (n = 1096); for Val, 131 +/- 58 mumol/L (n = 791). SDs were lower when the concentrations of these amino acids were expressed relative to that of Phe. The mean ratio for Leu + Ile to Phe was 2.5 +/- 0.49; for Val to Phe, 2.18 +/- 0.51. These results compare well with values previously reported in the literature. With these criteria, samples from a collection categorized by a bacterial inhibition assay as normal or falsely positive for MSUD were normal by MS-MS [(Leu + Ile): Phe &lt; 5.0]. Samples from confirmed MSUD patients were categorized as abnormal [(Leu+Ile): Phe &gt; 9.0] by MS-MS.

Rapid diagnosis of maple syrup urine disease in blood spots from newborns by tandem mass spectrometry

We report a new method for the diagnosis of maple syrup urine disease (MSUD) from dried blood spots on newborn screening cards based on tandem mass spectrometry (MS-MS). The mean +/- SD concentration of Leu plus Ile in normal newborns was 151 +/- 47 mumol/L (n = 1096); for Val, 131 +/- 58 mumol/L (n = 791). SDs were lower when the concentrations of these amino acids were expressed relative to that of Phe. The mean ratio for Leu + Ile to Phe was 2.5 +/- 0.49; for Val to Phe, 2.18 +/- 0.51. These results compare well with values previously reported in the literature. With these criteria, samples from a collection categorized by a bacterial inhibition assay as normal or falsely positive for MSUD were normal by MS-MS [(Leu + Ile): Phe < 5.0]. Samples from confirmed MSUD patients were categorized as abnormal [(Leu+Ile): Phe > 9.0] by MS-MS.