Data Set Error: In the Original Contribution entitled “Pharmacogenetic Association of the NPPA T2238C Genetic Variant With Cardiovascular Disease Outcomes in Patients With Hypertension” published in the January 23, 2008, issue of JAMA (2008;299:296-307), the authors identified an error in the data set used for analyzing the chlorthalidone vs doxazosin associations. The erroneous analytical data set was created at the data coordinating center in Houston, Texas, by merging 2 data files using a simple “merge” STATA command where a “match merge” should have been used. A “match merge” requires equality of the GenHAT participant identifiers from the paired records being merged, the ALLHAT variables being in one file and the genotype data in another. A simple “merge” assumes records from the merging files are in one-to-one correspondence, agreeing in order and number, and no check is done of the corresponding identifiers. The ALLHAT file had more observations than the GenHAT one, resulting in the mismatched pairing of records from the 2 files. This data set was used exclusively for the chlorthalidone vs doxazosin comparisons for the clinical outcomes (Table 4). A data set that was created separately and correctly was used for the baseline data (Table 1), main effects of the genetic variants (Table 2), chlorthalidone vs amlodipine and lisinopril comparisons (Table 3), and change in blood pressure data (Table 5). After reanalyzing the data with a corrected data set, the following changes affect the text of the article:
In the “Main Effect of NPPA Variants and Treatment Assignment on Clinical Outcomes” subsection of the “Results” section on page 300, the last sentence should have read: “Using data with follow-up to the point in time when the doxazosin arm was discontinued for comparability, the doxazosin group had a higher risk of stroke (HR, 1.27; 95% CI, 1.09-1.47; P=.002), heart failure (OR, 1.91; 95% CI, 1.68-2.16; P<.001), combined CVD (HR, 1.21; 95% CI, 1.14-1.28; P<.001), and, to a lesser extent, combined CHD (HR, 1.08; 95% CI, 0.99-1.17; P=.07) than the chlorthalidone group (data not shown).”
In the “Pharmacogenetic Associations (Genotype × Treatment Interactions) With Clinical Outcomes” subsection of the “Results” section on page 300, in the second paragraph, the following statement should have been inserted after the fifth sentence: “Using data with follow-up to the point in time when the doxazosin arm was discontinued, there was also evidence of a pharmacogenetic association for the NPPA T2238C variant with combined CVD when comparing chlorthalidone with doxazosin: though all genotype groups had lower risk of combined CVD events when randomized to chlorthalidone, the minor C allele carriers had a greater risk reduction than the common TT homozygotes.” The final sentence of this section should have read: “We found no evidence of pharmacogenetic associations for any of the NPPA G664A tests.”
In the “Comment” section on page 303, the last sentence should have been replaced with these 2 sentences: “There was also evidence of a pharmacogenetic association of the NPPA T2238C variant with combined CVD when comparing the chlorthalidone group with the doxazosin group using the limited follow-up data: all genotype groups had lower risk of combined CVD events when randomized to chlorthalidone; however, the minor C allele carriers had greater risk reduction than the TT homozygotes. When comparing chlorthalidone with lisinopril, no pharmacogenetic association was detected.” In the fourth paragraph under “Comment,” the following statement should have been inserted after the sixth sentence: “A similar interaction was observed for combined CVD: lower rates were observed among all genotypes randomized to diuretic vs the α-adrenergic blocker; however, a stronger benefit was observed for the minor C allele carriers.” In the last paragraph on page 305, the first sentence should have read: “In general, the pharmacogenetic association of NPPA T2238C with clinical outcomes was restricted to comparisons of diuretic vs calcium channel blocker or α-adrenergic blocker, not diuretic vs ACE inhibitor, whereas for the blood pressure findings, pharmacogenetic associations were observed for diuretic vs ACE inhibitor or α-adrenergic blocker but not diuretic vs calcium channel blocker.”
The corrected Table 4 appears herein.
Data Set Error in: Pharmacogenetic Association of the NPPA T2238C Genetic Variant With Cardiovascular Disease Outcomes in Patients With Hypertension. JAMA. 2009;302(10):1057. doi:10.1001/jama.2009.1244