The evidence for sodium bicarbonate therapy for contrast-associated acute kidney injury: far from settled science
Bibliographic record
Abstract
Sir, We read with interest the meta-analysis by Hoste et al. [1] as well as the editorial by Drs Wiedermann and Joannidis [2]. The commenters point out the highly significant benefit in patients undergoing procedures for acute coronary procedures. However, this result is on the basis of only 25 events in 170 patients from two trials. Furthermore, one of these two trials was terminated early [3]. In the other, the subjects in the sodium bicarbonate group received an initial bolus followed by a higher rate of hydration and were also given intravenous N-acetylcysteine compared with controls who were not given an initial bolus of saline and were treated with a lower rate of saline hydration as well as a lower dose of oral N-acetylcysteine [4]. Because this trial is testing multiple hypotheses, not just the effect of sodium bicarbonate hydration, it is reasonable to exclude it from the meta-analysis [5]. In fact, subtleties such as these provide insight into the observed heterogeneity and help the reader better understand the robustness of the findings. In this context, the highly significant benefit of sodium bicarbonate in patients with acute coronary procedures can hardly be considered a robust finding. Another source of heterogeneity that goes largely unrecognized is the relationship of trial size and outcome (Figure 1). The power curve shows the relationship between study size, outcome and power. It is apparent that all positive trials to date, published or unpublished, are considerably underpowered. In contrast, all the negative trials have tended to be the largest studies to date. Therefore, the positive results of meta-analysis of sodium bicarbonate have been largely driven by a small number of positive trials with extreme treatment effects [5]. We have suggested that the magnitude of benefit is so small (RR 0.85) that a definitive trial with 90% power would require 9918 subjects [5]. Power curve: the relationship between trial size and power. The title of the comment, which suggests that there is a little role of unpublished studies, also strikes us as being disingenuous. The importance of unpublished studies has been well recognized, and purposeful omission of such studies may lead to biased results [6,7]. The lower quality of unpublished studies is also a result of the lack of information about the study methodology in the abstract version of the unpublished study. Hence, it is the assessed quality which is low, rather than the true quality per se, as pointed out in our systematic review [5]. There are three more recent meta-analyses [8–10] done on the same subject which bring the total number of meta-analysis in this area to 11, which is equal to the number of published trials in this field. As suggested earlier [11], a proliferation of meta-analyses does not widen the evidence base and cannot resolve the uncertainty engendered by small randomized controlled trials with heterogenous results for a surrogate outcome (change in creatinine rather than requirement for dialysis or mortality). Conflict of interest statement. None declared. Editorial Note: Dr Hoste et al. had been invited to reply to this letter, but we did not receive a response.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.010 | 0.066 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.004 | 0.005 |
| Scholarly communication | 0.004 | 0.007 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.055 | 0.047 |
| Insufficient payload (model declined to judge) | 0.008 | 0.005 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".