Restrictive versus Liberal Fluid Therapy for Major Abdominal Surgery
2018 · paper · Clinical trials
A trial comparing restrictive and liberal intravenous fluid regimens in 3000 high-risk patients having major abdominal surgery. One-year disability-free survival was similar (81.9% vs 82.3%), while acute kidney injury was more common with the restrictive regimen.
3 findings
Substance Text gives the goal-directed-device subgroup interaction as "(P=0.37)", but Figure 2 prints 0.83 as the P value for interaction for "Planned GD device".
- p. 5 Results, Primary Outcome
including planned use of a goal-directed device (P = 0.37)
- p. 10 Figure 2, Planned GD device row, P Value for Interaction column
0.83
- p. 5 Results, Primary Outcome
ing planned use of a goal-directed device (P=0.37),
- p. 5 Primary Outcome
planned use of a goal-directed device (P = 0.37)
Judge's note: Checked p5 crop and the Figure 2 image on p10. The interaction P for Planned GD device is 0.83, and the within-level P values are 0.98 (Yes) and 0.58 (No). No 0.37 appears anywhere in the figure. The text cites Fig. 2 for the same treatment-by-covariate interaction, so this is one quantity printed two ways.
Second check: another model agreed
Report a problem with this finding
Substance Figure 2 row Colorectal surgery "No" prints "1.05 (0.80–1.28)" with P=0.93 and events 138/844 vs 137/842. The HR is not centred on its CI on the log scale, unlike the other rows.
- p. 10 Figure 2, Colorectal surgery, No row
1.05 (0.80–1.28)
- p. 10 Figure 2, Colorectal surgery, No
138/844 137/842 1.05 (0.80–1.28) 0.93
Only the authors can say what is right here.
Judge's note: Checked the Figure 2 crop and page on p10. Geometric midpoint: sqrt(0.80*1.28)=1.012. On the log scale the distances from ln 1.05 are 0.272 down and 0.198 up, which is too asymmetric to be rounding. Other rows are centred (Yes: sqrt(0.85*1.39)=1.087 vs 1.09; GD No: sqrt(0.88*1.26)=1.053 vs 1.05). P=0.93 and the near-identical crude rates also fit an HR of about 1.01. Either the estimate or a bound appears misprinted, and only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Reference or label Introduction cites ref 20 for fluid excess raising sepsis risk ("acute kidney injury,19 sepsis,20 and poor wound healing.21"), but ref 20 is "Ratner AJ ... Synergistic proinflammatory responses induced by polymicrobial colonization of epithelial surfaces. Proc Natl Acad Sci U S A 2005".
- p. 2 Introduction, paragraph 2
acute kidney injury,19 sepsis,20 and poor wound
- p. 12 Reference 20
Synergistic proinflammatory responses induced by polymicrobial colonization of epithelial surfaces
- p. 2 Introduction, para 2
acute kidney injury,19 sepsis,20 and poor wound healing.21
- p. 12 Reference 20
Ratner AJ, Lysenko ES, Paul MN, Weiser JN. Synergistic proinflammatory responses induced by polymicrobial colonization of epithelial surfaces.
Only the authors can say what is right here.
Judge's note: Checked p2 crop and the p12 reference list image. The sentence says excessive intravenous-fluid infusion may increase sepsis risk. The printed title of ref 20 is about polymicrobial colonization of epithelial surfaces and does not mention fluid therapy. Refs 18, 19 and 21 fit their claims. Only the authors can say which reference was intended.
Second check: another model agreed
Report a problem with this finding
4 minor (typos and the like; most checked by the judge only)
Reference or label Methods list "peak C-reactive protein level" as a secondary outcome, but Table 3 reports "Median C-reactive protein level on day 3 (IQR)".
- p. 4 Trial Outcomes (secondary outcomes)
peak C-reactive protein level
- p. 8 Table 3, CRP row
Median C-reactive protein level on day 3
Only the authors can say what is right here.
Judge's note: Checked p4 (methods) and the Table 3 image on p8. The lactate row in Table 3 is labelled "peak", but the CRP row is labelled day 3. No footnote says that day-3 CRP stands in for the peak. Either the outcome definition or the table label is mismatched, and only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Wording Primary-outcome event counts are given as 267 and 261 (528) in Table 3 and the Results ("267 patients (95 deaths and 172 cases of persistent disability)"), but as 271/1490 and 262/1493 in Figure 2, matching "533 events were observed in the trial" in the Methods.
- p. 8 Table 3, Death or persistent disability
267
- p. 5 Results, Primary Outcome
Death or persistent disability occurred in 267 patients
- p. 10 Figure 2, All patients row
271/1490
- p. 4 Statistical Analysis
In actuality, 533 events were observed in the trial
Only the authors can say what is right here.
Judge's note: Checked the Table 3 image on p8, the p5 text, Figure 2 on p10 and the p4 Methods. Every Figure 2 subgroup block sums to 271 and 262 (e.g., sex 119+152 and 138+124). The Table 3 footnote (9 and 12 patients disabled before death) does not account for the 4 and 1 differences. The Cox analysis may count events beyond 1 year (median follow-up was 366 days) while Table 3 counts events at 1 year, but the paper does not say so. The authors should be asked which event set each count uses.
Second check: another model agreed
Report a problem with this finding
Wording Table 1 "Planned postoperative care in HDU or ICU" gives 416 (27.9) and 418 (28.0), while Figure 2 "Planned destination ICU or HDU" gives denominators of 429 and 422.
- p. 7 Table 1 (continued), Planned postoperative care in HDU or ICU
416 (27.9)
- p. 10 Figure 2, Planned destination, ICU or HDU row
125/429
Judge's note: Checked the Table 1 image on p7 and Figure 2 on p10. Figure 2's ICU/HDU and Ward denominators sum to the mITT totals. The Methods (p3) say randomization was stratified by planned postoperative destination, so Figure 2 may use the stratum recorded at randomization and Table 1 the plan recorded later. That would be legitimate, but the paper does not say so.
Second check: another model agreed
Report a problem with this finding
Typo Table 3 footnote ‖‖ repeats "data": "Data regarding the duration of stay in the HDU or ICU data are for 485 patients...".
- p. 8 Table 3 footnote ‖‖
Data regarding the duration of stay in the HDU or ICU data are for 485 patients
Judge's note: Seen on the p8 page image. The second "data" is redundant.
Second check: none (judge only)