Table 2.
Risk of bias assessment.
| Random sequence generation | Baseline characteristics | Addressing of Incomplete outcome data | Selective outcome reporting | Other sources of bias | |
|---|---|---|---|---|---|
| Banks et al. (1997) | + | ?a | +b | +c | +d |
| Biessels et al. (2001) | ? | + | +e | + | +d |
| Cardoso et al. (2010) | + | ?a | –f | + | +d |
| Cardoso et al. (2013a) | + | ?a | +b | + | +d |
| Cardoso et al. (2013b) | + | ?a | –f | + | +d |
| Cintra et al. (2017) | + | + | +b | + | +d |
| Jakobsen et al. (1987) | + | + | +b | –g | +d |
| Katyare and Patel (2006) | ? | ?a | +b | + | +d |
| Mans et al. (1988) | ? | ?a | +b | + | -h |
| Marissal-Arvy et al. (2018) | ? | ?a | +b | –i | ?j |
| Mastrocola et al. (2005) | ? | ?a | –f | + | +d |
| Min et al. (2020) | ? | + | +b | ?k | +d |
| Moreira et al. (2004) | + | ?a | +b | + | +d |
| Moreira et al. (2005) | + | ?a | +b | + | +d |
| Moreira et al. (2006) | + | ?a | +b | + | +d |
| Thurston et al. (1975) | ? | ?a | +b | –l | +d |
| Tomassoni et al. (2004) | ? | ?a | +b | –m | +d |
| Zhou et al. (2018) | + | + | –n | + | +o |
The SYRCLE‘s tool (Hooijmans et al., 2014) was used and slightly modified to assess the risk of bias of non-human studies. “+” indicates a low risk of bias; “–” indicates a high risk of bias; “?” indicates that not enough information has been provided on this item. For five items of the SRYCLE tool, i.e., allocation concealment, random housing, blinding of personnel, random outcome assessment, and blinding of the outcome assessors, not enough information was available for any of the studies examined here, so that all of them were rated “?.”
Baseline values not given.
No evidence of dropouts.
Unlike the other experiments in this list, Bank's experiment reports detailed time courses; as an exception the brain weight values from day 9 are missing.
No evidence of critical housing conditions, problems associated with study design, or conflicts of interest.
One diabetic and one control rat died; Quote: “Spare animals replaced these rats in the final measurements.”
Range of sample size given, so it remains unclear whether there were dropouts.
Blood glucose of control group not given.
Time of measurement of control group unclear; moreover, it remains unclear whether control group underwent sham treatment.
Values of blood glucose and body weight not given.
The research has been supported, inter alia, by Aide aux Jeunes Diabetiques; authors claimed that the funders had no part in any of their work. One author had been paid by Eli Lilly and Novo Nordisk.
In the coordinate cross of the relative brain weight plot, the labelling of the ordinate is ambiguous, making the extraction of this data uncertain.
Only one baseline body weight value is given for both experimental and control group.
Data of brain outcome not shown.
Attrition bias was present as 14 out of 73 rats died in the experimental group; it is appreciated that the number of deceased animals at the different points was presented.
Time of brain volume measurement not explicitly mentioned. No evidence of critical housing conditions, other problems associated with study design, or conflicts of interest.