| Study |
Bias |
| Randomisation process |
Deviations from intended interventions |
Missing outcome data |
Measurement of the outcome |
Selection of the reported results |
Overall |
| Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
| Dyson 2007 |
Low risk of bias |
"Randomization was undertaken by means of sealed envelopes equivalent to the number of subjects and filled fifty‐fifty with an indicator of either a low‐carbohydrate diet or healthy‐eating advice. Two separate sets of envelopes were prepared for both diabetic and non‐diabetic subjects. An independent observer witnessed randomization." Author reported in correspondence that no formal test of hypothesis was done for difference at baseline for non‐diabetic participants. Baseline characteristics table showed differences which were consistent with differences resulting by chance with very small sample sizes. |
Low risk of bias |
No information, however, blinding of participants and people delivering these dietary interventions is highly unlikely. n=1 of randomised participants in balanced group refused follow‐up as not allocated to low carbohydrate diet. Modified ITT conducted; n=1 from control group refused follow‐up, thus data not included as missing. |
High risk of bias |
In low carbohydrate diet group, 0% and in balanced diet group, 14% (n=1) of randomised participants were not included in the analyses. Refusal of follow‐up due to not being allocated to low carbohydrate diet may be influenced by expected change in weight, therefore missingness possibly influenced by the true value. |
Low risk of bias |
Body weight recorded at each monthly interval. No evidence that a different method was used across intervention groups to measure weight. No details about blinding of outcome assessors. Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of a pre‐specified analysis plan. No information about blinding of outcome data. Analysis intentions are not reported in sufficient detail to enable an assessment. |
High risk of bias |
Differential attrition, with all attrition coming from the control group and disappointment cited as reason; missingness is possibly due to the true value of the outcome. No pre‐registered document available to assess analysis intentions. |
| Evangelista 2021 |
Low risk of bias |
"Treatment assignments to high protein or standard protein diets were implemented by the study coordinator. A randomization algorithm provided by the study biostatistician was used to determine treatment assignments. Participants were equally allocated between treatment arms (by site) and permuted blocks with unequal block sizes on patient‐level factors ..." "Baseline demographic and clinical characteristics between the two groups were comparable..." |
Some concerns |
"Masking: None (open‐label)" reported in trial registry. No reasons for deviations from intended interventions reported. Analysis appeared to be modified ITT, but without reasons for loss to follow‐up reported, it was difficult to judge whether analysis was appropriate. |
High risk of bias |
12/45 (27%) in intervention group and 2/45 (4%) in control group did not have data for this outcome. High attrition in the intervention group and differential attrition with no reasons provided. It is likely that with such high and differential attrition, missingness was possibly dependent, and likely to depend, on the true value. |
Low risk of bias |
"Participants were weighed in clothing without shoes (to the nearest 0.1 kg) using a professional beam scale (model 402KLS; Health‐o‐Meter, Bridgeview, IL)." No evidence that a different method was used across intervention groups to measure weight. "Masking: None (open‐label" reported in trial registry. Weight is not an observer‐reported outcome involving judgement. |
High risk of bias |
No mention of prespecified data analysis plan; protocol and trial registry mentioned weight at 12 weeks. Protocol (published in 2013) and trial registry (2016 last updated) mentioned weight at 12 weeks and 60 weeks/15 months, however, only 12 week data was reported. Analysis intentions were reported as multivariate imputation techniques in the protocol published in 2013, but no mention of these analyses in the published paper in 2021 |
High risk of bias |
Reasons for loss to follow‐up were not reported, making it difficult to judge whether analyses were appropriate. Attrition was inferred from both trial arms, with high attrition in the intervention group and no reasons provided; therefore, missingness could and was likely to be related to the true value of the outcome. A protocol and trial registry were available, reporting intended time points and analytical approaches not included in the published paper. |
| Guldbrand 2012 |
Some concerns |
"Randomisation was not stratified and was based on drawing blinded ballots." No information describing allocation concealment. No significant differences were present at baseline (P > 0.1 for all characteristics). |
Low risk of bias |
"Non‐blinded study". Deviations reported are expected to arise in usual care. Discontinued interventions (n = 4 in low‐carbohydrate diet group and n = 3 in balanced diet group did not attend group meetings). Primary intention‐to‐treat analysis, with secondary completers analysis also done |
High risk of bias |
4/30 (13%) participants in the intervention group and 3/31 (10%) in the control group did not have data for this outcome. No details about methods for ITT. Participants discontinued diets in both groups as they had "severe difficulty" in following the diets, which may be related to change in weight (missingness possibly and likely influenced by true value). |
Some concerns |
No information about method used for body weight measurements. Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of prespecified data analysis plan, and trial registry did not mention the outcome and time point in any versions thereof. Analysis intentions were not reported in sufficient detail to enable an assessment. |
High risk of bias |
No specific information on allocation concealment. Attrition in both groups due to "severe difficulty" following diets, therefore missingness was possibly and likely related to the true value of the outcome. Lack of information in the pre‐registered methods |
| Krebs 2012 |
Low risk of bias |
"Baseline measures completed prior to randomisation. Computer randomised sequence undertaken by an independent biostatistician at a distant site and random allocation of each individual emailed to personnel independent of those assessing outcome measures. These personnel organise allocation of the individuals into the two groups according to random allocation." No statements or statistical tests about differences between groups, large sample size and baseline characteristics table showed similar weight and BMI. |
Low risk of bias |
"Allocation of randomisation was concealed from the participant until after written informed consent and baseline assessment." "Blinded randomised controlled trial" "If two members from the same household were enrolled, they were both allocated to the same intervention group." "Dieticians delivering the intervention were notified of the dietary allocation after baseline assessment and before the first group session..." Deviations reported were expected to arise in usual care, e.g. disliked diet. A modified ITT analysis was done regardless of intervention adherence (missing outcome data excluded). |
Low risk of bias |
34/207 (16%) participants in the intervention and 37/211 (18%) participants in the control arm did not have data available. "Generalized mixed linear models were used... including baseline as a co‐variate." "Both the ‘intention‐to‐treat’ analysis and the analysis that only included those who completed the study found no difference between the groups in any outcome." (data not shown). |
Low risk of bias |
Tanita electronic scales were used to measure weight. No evidence that a different method was used across intervention groups to measure weight. "...research assessors remained blind to intervention allocation at all assessment points and until after database lock." |
Some concerns |
No mention of prespecified analysis plan, but trial registry at date prior to study publication mentioned the outcome and time point. The time point mentioned in the trial registry at the date prior to publication is reported on in the publication. Analysis intentions were not reported in sufficient detail to enable an assessment. |
Some concerns |
Lack of information in the pre‐registered methods |
| Larsen 2011 |
Low risk of bias |
" Randomisation was carried out by a third party using computer‐generated random numbers (using block randomisation and random block sizes) and stratified according to diabetes treatment..." "Dietary assignment was performed by a third party on the day of the initial dietary counselling visit." "The baseline characteristics for the intention‐to‐treat population were well‐matched at baseline...although non‐significant differences were apparent in calcium excretion rate, systolic blood pressure and the proportion of men." |
Low risk of bias |
"...we were not able to blind participants or the study personnel involved in dietary counselling." Deviations reported were expected to arise in usual care, e.g. lost interest, dissatisfied with weight loss. Modified ITT condicted: "This study was carried out using an intention‐to‐treat model following delivery of dietary advice." "...study staff encouraged all participants to return for follow‐up assessments of primary and secondary outcomes, regardless of dietary adherence." |
High risk of bias |
9/57 (16%) participants in the intervention and 6/51 (12%) participants in the control arm did not have data available. This included 4 participants in intervention and 5 in control arm that changed their mind after randomisation but before starting the diets. Data imputed for ITT for 5/57 (intervention) and 1/51 (control) participants with missing data. "This study used the single imputation method of last measurement carried forward for missing data for primary and secondary outcomes." Reasons are vague and include 'changing their mind' and inconvenient'; thus missingness could and was likely to depend on the true value. |
Low risk of bias |
Weight was measured in light clothes. No evidence that a different method was used across intervention groups to measure weight. "...a trained assessor, who was blinded to the dietary assignment of the individual, measured and recorded weight..." |
Some concerns |
No mention of prespecified analysis plan; trial registry at date prior to study publication mentioned outcome, but did not mention this time point. "Study endpoints were assessed blinded to the diet group, but the statistical analysis was performed unblinded." Analysis intentions were not reported in sufficient detail to enable an assessment. |
High risk of bias |
Attrition reported, with vague reasons such as participants "changing their mind" or finding the trial "inconvenient"; therefore, missingness could and was likely to depend on the true value of the outcome. Lack of information in the pre‐registered methods |
| Pedersen 2014 |
Low risk of bias |
"Volunteers were blocked on sex, BMI and HbA1c before randomization. The randomization was performed by a trial manager not directly involved in the intervention..." "Allocation concealment has been followed in that the person responsible for the allocation schedule will be a trial manager who does not decide who is eligible for the trial and has no input into medical or dietetic decision making in the trial" (trial registry). No significant differences between groups for 16 out of 18 variables assessed at baseline |
Low risk of bias |
Participants and people delivering these dietary interventions were not blinded (trial registry). No reasons or details about deviations from intended interventions reported. Modified ITT analysis done: "All randomized volunteers who completed the baseline visit and had all measurements at this visit done were included in the analysis." |
High risk of bias |
No information about how many participants were randomised to each arm (total of 76 randomised). A total of 10/31 (32%) participants from the intervention and 9/33 (27%) from the control group did not have data available. Therefore, 76 ‐ 31 ‐ 33 = 12 participants (16%) were missing with no reasons provided. Modified ITT with last‐observation‐carried‐forward used: volunteers who dropped out of the study were not contacted at the end of the 12‐month intervention. n = 12 missing with no reasons provided |
Low risk of bias |
"Weight was measured at all visits; with participants wearing light clothing and no shoes." No evidence that a different method was used across intervention groups to measure weight. Trial registry implied that outcome assessors were blinded: "The analysis was conducted with the investigator blinded to the dietary allocation." |
Some concerns |
No mention of prespecified data analysis plan, and trial registry did not mention the outcome and time point in any versions thereof. Analysis intentions were not reported in sufficient detail to enable an assessment. |
High risk of bias |
High attrition with no reasons provided, modified ITT with last value carried forward used. Lack of information in pre‐registered methods |
| Samaha 2003 |
Some concerns |
"...randomly assigned to either the low‐carbohydrate diet or the low‐fat diet, with use of a pre‐established algorithm generated from a random set of numbers." No information describing allocation concealment. "Subjects in the two groups were well matched with regard to base‐line characteristics (Table 1)." No formal hypothesis test, but no apparent significant differences in Table 1 |
Some concerns |
"The study was not blinded." Total attrition was 40%. No reasons or details about deviations from intended interventions reported. ITT analysis was conducted: "The primary analysis included all 132 subjects: the 79 subjects who completed the study, the 29 subjects who dropped out but had six‐month data available from records of routine office visits, and the 24 subjects for whom the weight recorded at the last follow‐up visit was carried forward." |
High risk of bias |
21/64 (33%) participants in the intervention and 32/68 (47%) participants in the control arm did not have data available. ITT with last‐observation‐carried‐forward. n = 53 missing with no reasons provided; very high attrition in both groups, and it was likely that with such high attrition, missingness was possibly and likely to depend on the true value. |
Low risk of bias |
"The subjects’ weights were measured monthly on a single calibrated scale..." No evidence that a different method was used across intervention groups to measure weight. "The study was not blinded." Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of prespecified analysis plan. Outcome data were never blinded. Analysis intentions were not reported in sufficient detail to enable an assessment. |
High risk of bias |
Very high attrition with no reasons provided, modified ITT with last value carried forward used. Outcome assessors were not blinded to allocation. No pre‐registered document available to assess analysis intentions |
| Saslow 2017a |
Low risk of bias |
"Eligible participants provided written, informed consent and were randomly assigned following block randomization procedures (computerized random numbers in blocks of four) to either the MCCR or LCK diet..." "An assistant, not associated with any other aspect of the research, concealed the sequence in opaque envelopes, which were opened by other research assistants unaware of the sequence." No statements of statistical tests or differences between groups. Baseline characteristics table showed differences which were consistent with differences resulting by chance with very small sample sizes. |
Low risk of bias |
"For this trial, it was not possible for the participants and staff to be blinded to group allocation." Deviations reported were expected to arise in usual care, e.g. moved away, family responsibilities. Modified ITT conducted: "This primary analysis was conducted on an intention‐to‐treat basis, including all participants who were randomly assigned and had data at both baseline and 3 months." |
High risk of bias |
No participants (0/16) in the intervention and 3/18 (17%) participants in the control arm did not have data available. Modified ITT analysis was done, "including all participants who were randomly assigned and had data at both baseline and 3 months." Differential attrition (> 10%), reasons vague and citing work schedules and family responsibilities; missingness could depend on its true value for weight, but difficult to say if this was likely. |
Low risk of bias |
"Subjects were weighed with no jackets or shoes using the same calibrated scale at each time." No evidence that a different method was used across intervention groups to measure weight. "For this trial, it was not possible for the participants and staff to be blinded to group allocation." Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of prespecified data analysis plan; protocol and trial registry did not mention the outcome and time point in any versions thereof. Analysis intentions were not reported in sufficient detail to enable an assessment. |
High risk of bias |
Differential attrition with all attrition in the control arm, and vague reasons; therefore difficult to determine whether missingness was dependent on the true value of the outcome. Lack of information in the pre‐registered methods |
| Sato 2017 |
Low risk of bias |
"Randomization was achieved by the minimization and biased coin method." "Allocation sequence conducted by third party organization (Soiken, Inc., Osaka, Japan) was concealed until interventions were assigned." No problematic differences in baseline characteristics of participants |
Low risk of bias |
"...they [patients] could not be blinded due to characteristics of dietary therapy." "Open ‐ no one is blinded" (trial registry). Deviations reported were expected to arise in usual care, e.g. lacked motivation, disappointed with assigned diet, moved to a different hospital. All participants randomised received their interventions and all were analysed except for those that had missing outcome data, thus a modified ITT was done. |
Low risk of bias |
3/33 (9%) participants in the intervention and 1/33 (3%) participants in the control arm did not have data available. Modified ITT done. Reasons for loss to follow‐up and withdrawal were stated for all particpants and included 3 transferring to other hospitals and 1 admission for different diseases. |
Low risk of bias |
Body weight was measured by the staff. No evidence that a different method was used across intervention groups to measure weight. Open ‐ no one was blinded (trial registry). Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of a prespecified analysis plan. Could not obtain trial registry record. Analysis intentions were not reported in sufficient detail to enable an assessment. |
Some concerns |
No pre‐registered document available to assess analysis intentions |
| Tay 2014 |
Some concerns |
Randomization procedures (sequence generation and allocation concealment) were performed by research associates independent of outcome assessments and intervention delivery and was done "...using random varying block sizes before random computer‐generated assignment to either an LC or HC diet in a 1:1 ratio." Baseline characteristics were "...not significantly different between diet groups (p>0.05) in independent‐samples t‐test...or Chi2 test" |
Low risk of bias |
"...diet assignment was discernible by participants and interventionists..." Deviations reported are expected to arise in usual care, e.g. work commitments, time constraints, unable to comply with diet. Modified ITT conducted, all participants except those who dropped out or were lost to follow‐up, were analysed in group to which they were assigned. |
High risk of bias |
18/64 (28%) participants in the intervention and 20/67 (30%) participants in the control arm did not have data available. High attrition in both groups, with reasons including unable to comply to diet and health issue external to study. It is likely that with such high attrition, missingness is possibly and likely to depend on the true value. |
Low risk of bias |
"Body mass was measured using calibrated electronic scales..." No evidence that a different method was used across intervention groups to measure weight. "...blinding was maintained for outcome assessment and data analysis." |
Some concerns |
No mention of pre‐specified analysis plan, but trial registry at date prior to study publication mentions the outcome and timepoint. The time point mentioned in trial registry at date prior to publication is reported on in publication. Analysis intentions are not reported in sufficient detail to enable an assessment. |
High risk of bias |
No specific information on allocation concealment. High attrition reported, with missingness possibly and likely dependent on the true value of the outcome. Lack of information in the pre‐registered methods. |
| Watson 2016 |
Low risk of bias |
Participants were randomly allocated by an investigator, who had no participant contact, using the process of minimisation." "An independent holder of the randomisation schedule will perform treatment allocation without contact with the volunteers." (trial registry). "All p values [for baseline characteristics] are > 0.05 for between group differences based on independent t‐tests." |
Low risk of bias |
Open (masking not used) (trial registry). Deviations reported were expected to arise in usual care e.g. work commitments, unable to continue with diet. Modified ITT conducted ‐ including participants who commenced and completed the study irrespective of compliance. Two participants were not included as they withdrew after randomisation but before the start of the study. |
High risk of bias |
9/32 (28%) in the intervention and 10/31 (32%) participants in the control arm did not have data available. "Sensitivity analysis was conducted for the BMI, fat mass (kg), and DBP variables using two methods: all participants who commenced the study and completers only. Both analyses showed the same pattern of outcomes with a time effect (P < 0.001) but no differences between the groups (P >= 0.51) or group by time interactions (P >= 0.07)." High attrition in both groups, with reasons including personal reasons, unable to comply with the intervention. It was likely that with such high attrition, missingness was possibly and likely to depend on the true value. |
Low risk of bias |
"Body mass was measured twice using calibrated electronic scales (Tanita Ultimate Scale 2000; Tokyo)". No evidence that a different method was used across intervention groups to measure. "The investigator was trained in ISAK International Standards for Anthropometric Assessment". Open (masking not used). Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of prespecified analysis plan, but trial registry at date prior to study publication mentioned the outcome and time point. The time point mentioned in the trial registry at the date prior to publication is reported on in the publication. Analysis intentions were not reported in sufficient detail to enable an assessment and protocol published in 2015: "Study Status: Data collection has been completed." |
High risk of bias |
High attrition reported, with missingness possibly and likely dependent on the true value of the outcome. Lack of information in the pre‐registered methods |
| Westman 2008 |
Some concerns |
"...participants were randomized to one of two treatment groups stratified upon BMI greater or less than 32 kg/m2 using a computer‐generated list, and invited to attend the baseline visit." No information describing allocation concealment. "There were no clinically significant differences between the treatment groups." |
High risk of bias |
Participants received instructions, books and handouts about the diet they were to follow; "The delivery of the intervention and the assessment of outcomes were not blinded to the treatment assignment." Low‐carbohydrate diet group: 21/48 (44%) randomised completed; reasons for discontinuation were: 3 refused assigned diet, 2 were unsatisfied with the diet, 2 were lost to follow‐up, 2 were too busy, 1 relocated, and 7 cited no reason; balanced diet group: 29/49 (59%) randomised completed; reasons for discontinuation were: 1 refused assigned diet, 1 was unsatisfied with the diet, 2 were lost to follow‐up, 3 were too busy, 1 relocated, 1 had difficulty adhering to the diet and 9 cited no reason. Per protocol analysis. Very high proportion of participants excluded |
High risk of bias |
Outcome data were not available for 27/48 (56%) in the intervention group and for 20/49 (41%) in balanced diet group. Per protocol, no sensitivity analyses. Very high attrition in both groups, with reasons including dissatisfied with diet, as well as vague reasons. It is likely that with such high attrition, missingness was possibly and likely to depend on the true value. |
Low risk of bias |
"Wearing light clothing and no shoes, participants were weighed at each visit on the same calibrated scale." No evidence that a different method was used across intervention groups to measure weight. "The delivery of the intervention and the assessment of outcomes were not blinded to the treatment assignment." Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of a prespecified analysis plan. Outcome data never blinded. Analysis intentions were not reported in sufficient detail to enable an assessment. |
High risk of bias |
No specific information on allocation concealment. Very high attrition with some due to participants being unsatisfied with, or having trouble following, their allocated diet; it was therefore likely that at least some missingness was dependent on the true value of the outcome. No pre‐registered document available to assess analysis intentions |
| Wycherley 2010 |
Some concerns |
Central randomisation by computer; "Random allocation using computer software 'Clinstat'". No information about allocation concealment. For completers per group: "At baseline, there were no significant differences in age, weight, BMI, A1C, and sex distribution between groups (P=>0.12)."; "These characteristics were also similar for participants who completed or did not complete the study (P>=0.09), except for age in which completers were on average five years younger." |
Low risk of bias |
"Open (masking not used)". In the low carb diet group, 3/21 were excluded due to non‐compliance with the diet program and 2/21 withdrew as they were unable to continue with the diet. In the balanced diet group, 1/19 were excluded due to non‐compliance with the diet program and 1/19 withdrew as they were unable to continue with the diet. Modified ITT used. "Primary analysis was conducted on participants who completed the study per protocol. Secondary intention‐ to‐treat (ITT) analysis was also conducted for the primary outcome measures (body weight and composition, cardiometabolic and glycemic control), including participants who completed the study irrespective to protocol adherence." |
High risk of bias |
6/21 (29%) participants in the intervention group and 2/19 (11%) in the control group did not contribute data to the analyses as they withdrew from the study. Results of the sensitivity analysis narratively reported for the 4 arms of the trial, but no data shown or evidence provided for the 2 arms of interest. Effect sizes for per protocol: MD ‐0.40 (‐3.93; 3.13) and for mITT: MD 1.40 (‐2.53; 5.33). Reasons are vague and include 'unable to continue diet', with differential attrition and high attrition in the intervention group; thus missingness could and is likely to depend on the true value. |
Low risk of bias |
"Body mass was measured using calibrated electronic digital scales." No evidence that a different method was used across intervention groups. "Open (masking not used)." Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of pre‐specified data analysis plan, and trial registry does not mention the outcome and timepoint in any versions thereof. Analysis intentions are not reported in sufficient detail to enable an assessment. |
High risk of bias |
No specific information on allocation concealment. Differential attrition and high attrition in the intervention arm with vague reasons reported, therefore missingness could be and is likely dependent on the true value of the outcome. Lack of information in the pre‐registered methods. |
| Yamada 2014 |
Some concerns |
"Enrolled patients were randomly allocated ....... using a permuted randomised block of four patients per block". No information describing allocation concealment. "There were no statistically significant differences in any of the parameters between the two groups." |
Low risk of bias |
"The patients and investigators were not masked to group assignment." No attrition reported in the trial. No participants with missing outcome data |
Low risk of bias |
No participants with missing outcome data |
Some concerns |
No information about method used for body weight measurements. No details about procedures used for body weight measurement to assess if good practice anthropometric guidelines were used in both groups. No mention of training. Weight is not an observer‐reported outcome involving judgement. |
Some concerns |
No mention of a prespecified analysis plan. Outcome data never blinded. Analysis intentions were not reported in sufficient detail to enable an assessment. |
Some concerns |
No specific information on allocation concealment. Lack of information on how outcome was assessed. No pre‐registered document available to assess analysis intentions |
