Age-Related Clonal Hematopoiesis Associated with Adverse Outcomes
2014 · paper · Biomedicine
Examines blood DNA from large cohorts to study age-related clonal hematopoiesis, in which blood cells carry shared acquired mutations. The title reports an association with adverse outcomes; the available text does not give the specific findings.
7 findings
Substance The CHD hazard ratio's CI is given as "hazard ratio, 2.0; 95% CI, 1.2 to 3.4" in the Abstract but as "2.0 (95% CI, 1.2 to 3.5; P = 0.02)" in the Results (page 8).
- p. 2 Abstract, Results
(hazard ratio, 2.0; 95% CI, 1.2 to 3.4)
- p. 8 Results, Association of somatic mutation with cardiometabolic disease
2.0 (95% CI, 1.2 to 3.5; P = 0.02)
- p. 2 Abstract
hazard ratio, 2.0; 95% CI, 1.2 to 3.4
- p. 2 Abstract
hazard ratio, 2.6; 95% CI, 1.4 to 4.8
Only the authors can say what is right here.
Judge's note: Page 2 (Abstract) and page 8 (multivariable analysis of the two cohorts) images checked. Both report the same estimate with no other model mentioned, yet the upper bounds differ at the same precision. Only the authors can say which value is correct.
Second check: another model agreed
Report a problem with this finding
Substance The ischemic stroke CI is given as "hazard ratio, 2.6; 95% CI, 1.4 to 4.8" in the Abstract but as "2.6 (95% CI, 1.3 to 4.8; P = 0.003)" in the Results (page 8).
- p. 2 Abstract, Results
ischemic stroke (hazard ratio, 2.6; 95% CI, 1.4 to 4.8)
- p. 8 Results, Association of somatic mutation with cardiometabolic disease
2.6 (95% CI, 1.3 to 4.8; P = 0.003)
- p. 2 Abstract
hazard ratio, 2.6; 95% CI, 1.4 to 4.8
Only the authors can say what is right here.
Judge's note: Page 2 and page 8 images checked. It is the same multivariable estimate, but the lower bound differs. No alternative model is described.
Second check: another model agreed
Report a problem with this finding
Substance Page 7 says "Information on red-cell distribution width was available for 2409 participants in two cohorts", but the Fig. 4C strata (JHS and UA) total 2424 (172/1903, 60/411, 13/83, 16/27).
- p. 7 Results, overall survival, RDW paragraph
Information on red-cell distribution width was available for 2409 participants in two cohorts
- p. 16 Figure 4C, Events/No. at Risk
172/1903
- p. 7 Survival/RDW
available for 2409 participants in two cohorts
Only the authors can say what is right here.
Judge's note: Fig. 4C on the page 16 image checked. The cohort subtotals agree (JHS 2126 + UA 298). The analysed total exceeds the number said to have RDW data, and no text explains this.
Second check: another model agreed
Report a problem with this finding
Reference or label In the byline the asterisks are on "Michael Boehnke, Ph.D.*, Jaakko Tuomilehto, M.D., Ph.D.*" and "David Altshuler, M.D., Ph.D.*", but the * footnote reads "Dr. McCarthy is listed on behalf of ... T2D-GENES ...; Dr. Boehnke ...; and Dr. Altshuler ...". McCarthy has no asterisk, and Tuomilehto has one but is not named in the footnote.
- p. 1 Author byline
Heather M. Stringham, Ph.D.*, Mark I. McCarthy, M.D., Michael Boehnke, Ph.D.*, Jaakko Tuomilehto, M.D., Ph.D.*
- p. 1 Footnote *
*Dr. McCarthy is listed on behalf of the Type 2 Diabetes (T2D) Genetic Exploration
- p. 1 Author byline
Michael Boehnke, Ph.D.*, Jaakko Tuomilehto, M.D., Ph.D.*
- p. 1 Author list / footnote
Dr. McCarthy is listed on behalf of the Type 2 Diabetes (T2D) Genetic Exploration
Judge's note: Page 1 image: Stringham and McCarthy carry no asterisk; Boehnke, Tuomilehto and Altshuler do. The footnote names McCarthy, Boehnke and Altshuler. Reviewer 1 was wrong to say Stringham is marked, but the mismatch itself is real. The asterisk on Tuomilehto looks like it belongs to McCarthy.
Second check: another model agreed
Report a problem with this finding
Reference or label The text attributes the Kaplan–Meier comparison to "(P<0.001 by rank-sum test) (Fig. 4B", but Fig. 4B (age ≥70 panel) prints "P<0.001 by log-rank test".
- p. 7 Results, Association of somatic mutations with overall survival
(P<0.001 by rank-sum test)
- p. 16 Figure 4B, age ≥70 panel
P<0.001 by log-rank test
- p. 7 Results, Association of somatic mutations with overall survival
(P<0.001 by rank-sum test) (Fig. 4B;
- p. 7 Survival results
P<0.001 by rank-sum test
Judge's note: Page 7 text and page 16 Fig. 4B image checked. The same comparison carries a different test name in the text and the figure. Elsewhere the paper uses 'Wilcoxon rank-sum test' for VAF and RDW comparisons.
Second check: another model agreed
Report a problem with this finding
Reference or label Reference 9 reads "Sci Transl Med. 2012; 4:149ra18."
- p. 10 Reference 9
Sci Transl Med. 2012; 4:149ra18.
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the Jan et al. article is Sci Transl Med 2012;4(149):149ra118. The page 10 image shows '149ra18', and the entry also lacks a PubMed ID, unlike its neighbours.
Second check: another model agreed
Report a problem with this finding
Reference or label Figure 4A (PDF p16), an all-cause mortality forest plot, has the outcome header 'Hematologic Cancer'; it should read 'Death from Any Cause', as in Figure 4C.
- p. 15 Figure 4A, first column header
Hematologic Cancer
Judge's note: The cluster gives the location as p15, but that page holds Figure 3, where the 'Hematologic Cancer' header is correct. The defect is on PDF p16. There the Figure 4A header reads 'Hematologic Cancer', while the caption says the panel shows 'the risk of death from any cause'. The event counts (698/4886 and 69/246, 767 deaths in total) match the mortality analysis of 5132 persons on p7. Figure 4C uses the header 'Death from Any Cause'. The label is wrong; the data are fine.
Second check: another model agreed
Report a problem with this finding
6 minor (typos and the like; most checked by the judge only)
Wording Figure 1 'Total' row (240 855 2894 5441 5002 2300 317 86 17) sums to 17,152, 30 fewer than the 17,182 persons analysed, and the figure does not say which persons are left out.
- p. 13 Figure 1 Total row
240 855 2894 5441 5002 2300 317 86 17
- p. 4 Results
17,182 persons
Judge's note: I checked the Figure 1 image on p13: the totals are as quoted and add up to 17,152. The mutation row adds up to 746. The Results on p4 give 17,182 persons with an age range of 19 to 108. The first bin in Figure 1 is 20–29, so anyone aged 19 cannot appear in the figure. That probably explains the 30-person gap, but the paper never says so. The Methods (age-missing exclusions, 116 persons) cover exclusions made before the 17,182 count, so they do not explain the gap. This is a minor unstated omission and not a data error.
Second check: another model agreed
Report a problem with this finding
Wording Page 6 says "approximately 1% per year among persons with a variant allele fraction greater than 0.10", but the subgroup is defined everywhere else as "0.10 or greater" (text, Fig. 3A "Mutation, VAF ≥0.10", caption).
- p. 6 Results, hematologic cancer, annual-risk sentence
per year among persons with a variant allele fraction greater than 0.10
- p. 6 Results, hematologic cancer
variant allele fraction of 0.10 or greater
- p. 15 Figure 3A
Mutation, VAF ≥0.10
Judge's note: Page 6 and page 15 images checked. Fig. 3A shows 5/57 for VAF ≥0.10, which is consistent with the ~1%/yr figure, so the strict '>' is a wording slip.
Second check: another model agreed
Report a problem with this finding
Wording Figure 4B group sizes (No mutation N=3940 and N=963; Mutation N=141 and N=106) add up to 5150 persons (4903 without and 247 with a mutation). The legend says Panel B uses 'the same cohorts as those included in Panel A', but Figure 4A and the text give 5132 persons (4886 and 246).
- p. 16 Figure 4B labels
No mutation (N=3940)
- p. 17 Figure 4 legend
Panel B shows Kaplan–Meier survival curves generated from data from the same cohorts as those included in Panel A
- p. 7 Results, survival
available data from 5132 persons in seven cohorts
- p. 16 Figure 4A
698/4886
Only the authors can say what is right here.
Judge's note: I checked the Figure 4 image on p16, the legend on p17 and the Results text on p7. The numbers are as quoted. The gap could come from different exclusions, for example missing covariates in the Cox models, but neither the article nor the legend says so. Only the authors can confirm which set was analysed.
Second check: another model agreed
Report a problem with this finding
Typo The reprint-request footnote ends "bebert@partners.org.." with a doubled full stop.
- p. 1 Address reprint requests footnote
bebert@partners.org..
Judge's note: Page 1 image shows two periods after the e-mail address.
Second check: none (judge only)
Report a problem with this finding
Typo MGUS is written "monoclonal gammopathy of unknown significance" (pages 2–3 and page 6), while cited ref. 2 has "monoclonal gammopathy of undetermined significance".
- p. 2 Introduction
monoclonal gammopathy of unknown
- p. 6 Results, hematologic cancer
monoclonal gammopathy of unknown significance
- p. 10 Reference 2 title
monoclonal gammopathy of undetermined significance
Judge's note: Page 2, page 6 and reference 2 on page 10 checked. The text term differs from the established term used in the cited reference. This is cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo In the Acknowledgments, the first JHS contract is printed "HSN268201300046C", while the others are "HHSN268201300047C", "HHSN268201300048C", "HHSN268201300049C" and "HHSN268201300-050C".
- p. 10 Acknowledgments
HSN268201300046C
Judge's note: Page 10 image checked. The prefix is inconsistent within the same list, so one 'H' appears to be missing.
Second check: none (judge only)