8. GRADE Summary of findings table ‐ Child (as neonate, child, adult).
| Intervention and comparison and outcome | Assumed risk with comparator | Corresponding risk with intervention* | Relative effect (95% CI) | № of participants (studies) | Quality of the evidence (GRADE) |
Comments from included reviews in quotation marks Comments without quotation marks from overview review authors |
| 8.0 Large‐for‐gestational age (LGA) (as defined in reviews) | ||||||
|
Biesty 2018 Induction of labour versus expectant management LGA defined as > 90th percentile |
114 per 1000 | 60 per 1000 (32 to 116) |
RR 0.53 (0.28 to 1.02) |
425 (1 RCT) |
Low | Evidence is based on one small study with design limitations. Wide confidence intervals crossing the line of no effect |
|
Brown 2017a Glibenclamide versus placebo LGA defined > 90th percentile |
118 per 1000 | 105 per 1000 (60 to 187) | RR 0.89 (0.51 to 1.58) | 375 (1 RCT) |
Very low | "Evidence is based on one study and 93% were Hispanic women, results may not be generalisable to other populations. There is risk of bias, as we did not find a published protocol and there were more outcomes reported in the published paper than were listed in the trial registration document." |
|
Brown 2017a Metformin versus glibenclamide LGA defined as > 90th percentile |
193 per 1000 | 129 per 1000 (46 to 354) | RR 0.67 (0.24 to 1.83) | 246 (2 RCTs) |
Low | "Allocation concealment was unclear in one study and one study was open label. Inconsistent as heterogeneity was I² = 54%, which could not be explained by the diagnostic criteria used." |
|
Brown 2017a Glibenclamide versus acarbose LGA defined as > 90th percentile |
105 per 1000 | 251 per 1000 (57 to 1000) | RR 2.38 (0.54 to 10.46) | 43 (1 RCT) |
Very low | "Evidence is based on one small study with wide confidence intervals and evidence of selective reporting." |
|
Brown 2016a Myo‐inositol versus placebo2 LGA defined as > 90th centile |
26 per 1000 | 9 per 1000 (1 to 226) |
RR 0.36 (0.02 to 8.58) | 73 (1 RCT) |
Very low | "Evidence is based on one small study with low event rates ‐ 0/35 events in myo‐inositol group and 1/38 events in the placebo group." |
|
Han 2017 Low‐moderate versus moderate‐high GI diet LGA defined as ≥ 90th percentile for gestational age |
146 per 1000 | 104 per 1000 (32 to 342) |
RR 0.71 (0.22 to 2.34) | 89 (2 RCTs) |
Low | "One study at unclear risk of selection bias and two studies at risk of performance bias and unclear risk of detection bias. Wide confidence intervals crossing the line of no effect and small sample size." |
|
Han 2017 Energy‐ versus no energy‐restricted diet LGA defined as ≥ 90th percentile for gestational age |
246 per 1000 | 288 per 1000 (160 to 522) |
RR 1.17 (0.65 to 2.12) | 123 (1 RCT) |
Low | "One study at unclear risk of selection and detection bias and wide confidence interval crossing the line of no effect and small sample size." |
|
Han 2017 Low‐ versus high‐carbohydrate diet LGA defined as ≥ 90th percentile for gestational age |
80 per 1000 | 41 per 1000 (10 to 156) |
RR 0.51 (0.13 to 1.95) | 149 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of performance bias as participants and researchers were not blinded |
|
Han 2017 High‐ versus low‐unsaturated fat diet with matching calories LGA defined as ≥ 90th percentile for gestational age |
571 per 1000 | 309 per 1000 (120 to 783) |
RR 0.54 (0.21 to 1.37) | 27 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of performance bias as participants and researchers were not blinded. Baselines for BMI were unbalanced between groups |
|
Han 2017 Low‐Gi diet versus high‐fibre moderate‐GI diet LGA defined as ≥ 90th percentile for gestational age |
44 per 1000 | 128 per 1000 (27 to 600) |
RR 2.87 (0.61 to 13.50) | 92 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of detection bias as outcome assessors were not blinded. Incomplete data reported (attrition bias) and blood glucose concentration unbalanced at baseline |
|
Han 2017 Diet + diet‐related behavioural advice versus diet only LGA defined as ≥ 90th percentile for gestational age |
140 per 1000 | 102 per 1000 (35 to 300) |
RR 0.73 (0.25 to 2.14) | 99 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of performance bias as participants and personnel were not blinded |
|
Han 2017 Ethnic specific diet versus standard healthy diet LGA defined as ≥ 90th percentile for gestational age |
300 per 1000 | 42 per 1000 (3 to 735) |
RR 0.14 (0.01 to 2.45) | 20 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of performance bias as participants and personnel were not blinded and selective reporting (reporting bias). Low event rates, as there were no events in the intervention group and three events in the control group |
|
Brown 2017b Lifestyle intervention versus usual care or diet alone LGA not defined |
189 per 1000 | 113 per 1000 (95 to 134) |
RR 0.60 (0.50 to 0.71) | 2994 (6 RCTs) |
Moderate | "Several included studies had high risk of bias for lack of blinding, incomplete outcome data and selective reporting. Allocation concealment was in unclear in two of the six studies." |
|
Han 2012 Intensive management versus routine care LGA defined as ≥ 90th percentile for gestational age |
171 per 1000 | 63 per 1000 (34 to 113) |
RR 0.37 (0.20 to 0.66) | 438 (3 RCTs) |
Low | Evidence based on three studies with serious/very serious design limitations |
|
Raman 2017 Telemedicine versus standard care for glucose monitoring LGA not defined |
126 per 1000 | 178 per 1000 (96 to 333) |
RR 1.41 (0.76 to 2.64) | 228 (3 RCTs) |
Very low | Evidence downgraded for study limitations and imprecision (wide confidence intervals crossing the line of no effect; small sample size and few events |
|
Raman 2017 Self‐ versus periodic‐glucose monitoring LGA not defined |
142 per 1000 | 117 per 1000 (71 to 195) |
RR 0.82 (0.50 to 1.37) | 400 (2 RCTs) |
Low | Evidence downgraded for study limitations and imprecision (wide confidence intervals crossing the line of no effect) |
|
Raman 2017 Continuous‐ versus self‐monitoring blood glucose LGA not defined |
527 per 1000 | 353 per 1000 (227 to 554) |
RR 0.67 (0.43 to 1.05) | 106 (1 RCT) |
Very low | Evidence downgraded for study limitations and imprecision (wide confidence intervals crossing the line of no effect and small sample size) |
|
Raman 2017 Post‐ versus pre‐prandial glucose monitoring LGA not defined |
424 per 1000 | 123 per 1000 (47 to 331) |
RR 0.29 (0.11 to 0.78) |
66 (1 RCT) |
Very low | Evidence downgraded for imprecision (wide confidence intervals crossing the line of no effect, single trial, small sample sizes) and study limitations |
|
Brown 2017d Insulin versus oral therapy Birthweight > 90th percentile |
159 per 1000 | 161 per 1000 (121 to 215) |
average RR 1.01 (0.76 to 1.35) | 2352 (13 RCTs) |
Moderate | Evidence downgraded for study limitations (lack of blinding) |
|
Brown 2017d Insulin type A versus B LGA not defined |
58 per 1000 | 70 per 1000 (34 to 148) |
RR 1.21 (0.58 to 2.55) | 411 (3 RCTs) |
Low | Evidence downgraded for study limitations (insufficient details) and imprecision (wide confidence intervals) |
|
Brown 2017d Insulin versus diet LGA not defined |
133 per 1000 | 113 per 1000 (55 to 237) |
RR 0.85 (0.41 to 1.78) | 202 (1 RCT) |
Very low | Evidence downgraded for study limitations (insufficient details) and imprecision (single study, low events, wide confidence intervals) |
|
Brown 2017d Insulin regimen A versus B Twice daily versus four times daily Three times versus six times daily LGA not defined |
261 per 1000 158 per 1000 |
303 per 1000 (206 to 441) 55 per 1000 (6 to 486) |
RR 1.16; (0.79 to 1.69) RR 0.35; (0.04 to 3.08) |
274 (1 RCT) 37 (1 RCT) |
Very low | Evidence downgraded for study limitations (insufficient details) and imprecision (single small study, wide confidence intervals) |
| 9.0 Perinatal mortality (fetal and neonatal death) and later infant mortality | ||||||
|
Biesty 2018 Induction of labour versus expectant management Perinatal death |
See comment | See comment | RR not estimable | 425 (1 RCT) |
Very low | Evidence is based on one small study with no events and design limitations |
|
Brown 2017a Metformin versus glibenclamide Perinatal death |
6 per 1000 | 5 per 1000 (0 to 83) |
Average RR 0.92 (0.06 to 14.55) | 359 (2 RCTs) |
Very low | "Open label studies with no evidence of blinding of participants or researchers. Event rates were very low. One study had no event of perinatal death in either the metformin nor the glibenclamide group. The second study had one death in each group." |
|
Brown 2017a Glibenclamide versus acarbose Perinatal death |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 43 (1 RCT) |
Very low | "Evidence based on a single small study with wide confidence intervals. No events were reported in either group. There is evidence of selective reporting." |
|
Han 2017 Energy‐ versus no energy restricted diet Perinatal death |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 423 (2 RCTs) |
Low | "Two studies at unclear risk of selection bias. One study at high risk of performance bias and unclear risk of detection bias. There were no events in either group and relatively small sample sizes." |
|
Han 2017 Low‐ versus high‐carbohydrate diet Perinatal death |
0 per 1000 | 0 per 1000 (0 to 0) |
RR 3.00 (0.12 to 72.49) | 150 (1 RCT) |
Very low | Evidence is based on one study and imprecision as wide confidence interval crossing the line of no effect. Risk of performance bias as study participants and care providers were not blinded. Low event rates (one event in the control group) |
|
Brown 2017b Lifestyle intervention versus usual care or diet alone Perinatal death |
5 per 1000 | 0 per 1000 (0 to 9) |
RR 0.09 (0.01 to 1.70) | 1988 (2 RCTs) |
Low | "There is evidence of imprecision with wide confidence intervals and low events rates (5 perinatal deaths in one trail's control group) and one of the two trials did not blind participants/researchers." |
|
Brown 2017c Exercise versus control |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 19 (1 RCT) |
Very low | Imprecision ‐ There are no events in either group and the sample size is only 19 infants. "There is a lack of clarity for most items associated with risk of bias." |
|
Raman 2017 Telemedicine versus standard care for glucose monitoring |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 131 (2 RCTs) |
Very low | There were no events reported for this outcome. Evidence downgraded for study limitations and imprecision (no events and small sample sizes) |
|
Raman 2017 Self‐ versus periodic‐glucose monitoring |
5 per 1000 | 8 per 1000 (1 to 57) |
RR 1.54 (0.21 to 11.24) | 400 (2 RCTs) |
Very low | Evidence downgraded for study limitations and imprecision (wide confidence intervals crossing the line of no effect and few events). |
|
Raman 2017 Continuous‐ versus self‐monitoring blood glucose |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 179 (2 RCTs) |
Very low | There were no events of perinatal death reported in the two RCTs. Evidence was downgraded for study limitations and imprecision (no events and small sample sizes) |
|
Brown 2017d Insulin versus oral therapy |
8 per 1000 | 7 per 1000 (2 to 20) |
RR 0.85 (0.29 to 2.49) | 1463 (10 RCTs) |
Low | Evidence downgraded for study limitations (lack of blinding) and imprecision (wide confidence intervals and low event rates) |
|
Brown 2017d Insulin versus diet |
43 per 1000 | 32 per 1000 (18 to 57) |
RR 0.74 (0.41 to 1.78) | 1137 (4 RCTs) |
Moderate | Evidence downgraded for study limitations (insufficient details) |
|
Brown 2017d Insulin regimen A versus B Twice daily versus four times daily |
0 per 1000 | 0 per 1000 (0 to 0) |
RR 3.04 (0.13 to 74.07) | 274 (1 RCT) |
Very low | Evidence downgraded for imprecision (extremely wide confidence intervals; single small study; very low event rates). There was one event in the twice daily group and no events in the four times daily group |
| 10.0 Death or serious morbidity composite (as defined in reviews) | ||||||
|
Brown 2017a Metformin versus glibenclamide Defined as composite of neonatal outcomes including hypoglycaemia, hyperbilirubinaemia, macrosomia, respiratory illness, birth injury, stillbirth or neonatal death |
350 per 1000 | 189 per 1000 (109 to 329) |
RR 0.54 (0.31 to 0.94) | 159 (1 RCT) |
Low | "Evidence is based on one small study." Risk of performance bias as participants and personnel were not blinded |
|
Han 2017 Ethnic specific diet versus standard healthy diet Defined as composite of neonatal outcomes that included hypoglycaemia, neonatal asphyxia, respiratory distress syndrome (RDS), hyperbilirubinaemia and hypocalcaemia |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 20 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study. Risk of performance bias as participants and personnel were not blinded and selective reporting (reporting bias). No events in either group |
|
Brown 2017b Lifestyle intervention versus usual care or diet alone Defined as composite of death, shoulder dystocia, bone fracture and nerve palsy in one trial and still birth, neonatal death, hypoglycaemia, hyperbilirubinaemia, elevated cord‐blood C‐peptide and birth trauma in the other trial |
193 per 1000 | 110 per 1000 (41 to 299) |
Average RR 0.57 (0.21 to 1.55) | 1930 (2 RCTs) |
Very low | "Evidence of inconsistency with I2 > 70%. One of the two trials did not blind participants/researchers and evidence of imprecision with wide confidence intervals crossing the line of no effect." |
|
Brown 2017c Exercise versus control Defined as mortality and morbidity composite |
65 per 1000 | 36 per 1000 (8 to 169) |
RR 0.56 (0.12 to 2.61) | 169 (2 RCTs) |
Moderate | Imprecision ‐ wide confidence intervals and low event rates |
|
Raman 2017 Telemedicine versus standard care for glucose monitoring Defined as composite of neonatal intensive care unit admission, LGA, respiratory outcomes (hyaline membrane disease, transient tachypnoea, need for respiratory support); hypoglycaemia; and hyperbilirubinaemia |
560 per 1000 | 594 per 1000 (381 to 930) |
RR 1.06 (0.68 to 1.66) | 57 (1 RCT) | Very low | Evidence downgraded for imprecision (wide confidence intervals crossing the line of no effect, small sample size and few events) and study limitations |
|
Brown 2017d Insulin versus oral therapy |
319 per 1000 | 329 per 1000 (268 to 402) |
RR 1.03 (0.84 to 1.26) | 760 (2 RCTs) |
Moderate | Evidence was downgraded for study limitations (lack of blinding). One trial included resuscitation of the delivery room, preterm birth (< 37 weeks), neonatal intensive care unit admission, birth injury or diagnosis of neonatal complication, glucose infusion, antibiotics or phototherapy. A second trial included hypoglycaemia < 2.6 mmol/L, RDS, phototherapy, birth trauma, APGAR < 7 at 5 minutes, preterm birth < 37 weeks |
|
Brown 2017d Insulin regimen A versus B Twice daily versus four times daily |
174 per 1000 | 294 per 1000 (188 to 459) |
RR 1.69 (1.08 to 2.64) | 274 (1 RCT) |
Very low | Evidence downgraded for imprecision (Single small study with wide confidence intervals and low event rates) |
| 11.0 Neonatal hypoglycaemia (as defined in the reviews) | ||||||
|
Biesty 2018 Induction of labour versus expectant management Not defined |
38 per 1000 | 28 per 1000 (10 to 79) |
RR 0.74 (0.26 to 2.09) | 425 (1 RCT) |
Very Low | Evidence downgraded for imprecision (single study with low events) and study limitations |
|
Brown 2017a Glibenclamide versus placebo Not defined |
11 per 1000 | 21 per 1000 (4 to 114) |
RR 1.97 (0.36 to 10.62) | 375 (1 RCT) |
Very low | "Evidence is based on one study and 93% were Hispanic women, results may not be generalisable to other populations. There is risk of bias, as we did not find a published protocol and there were more outcomes reported in the published paper than were listed in the trial registration document. Event rates were low with 4/189 for oral antidiabetic pharmacological therapy (Glibenclamide) and 2/186 for placebo group with wide confidence intervals crossing the line of no effect." |
|
Brown 2017a Metformin versus glibenclamide Defined as < 2.2 mmol/L (< 40mg/dL) |
48 per 1000 | 41 per 1000 (20 to 84) |
RR 0.86 (0.42 to 1.77) |
554 (4 RCTs) |
Low | "Allocation concealment was unclear in one study and one other study was open label. Event rates were low (< 30), 12/281 for the Metformin group and 13/273 for the Glibenclamide group." |
|
Brown 2017a Glibenclamide versus acarbose Defined as < 2.2 mmol/L (< 40 mg/dL) |
53 per 1000 | 333 per 1000 (46 to 1000) |
RR 6.33 (0.87 to 46.32) | 43 (1 RCT) |
Very low | "There is evidence of selective reporting. Evidence based on one small study with wide confidence intervals. Low event rates and sample size with 8/24 in Glibenclamide group and 1/19 in acarbose group." |
|
Brown 2016a Myo‐inositol versus placebo2 Not defined |
263 per 1000 | 13 per 1000 (0 to 224) |
RR 0.05 (0.00 to 0.85) | 73 (1 RCT) |
Low | "Evidence is based on one small study with low event rates ‐ 0/35 events in myo‐inositol group and 10/38 events in the placebo group." |
|
Han 2017 Energy‐ versus no energy‐restricted diet Not defined |
190 per 1000 | 201 per 1000 (91 to 441) |
RR 1.06 (0.48 to 2.32) | 408 (2 RCTs) |
Very low | "Evidence is based on two small studies at unclear risk of selection bias; one study at high risk of performance bias and unclear risk of detection bias. Wide confidence intervals crossing the line of no effect and substantial heterogeneity: I² = 75% present." |
|
Han 2017 Low‐ versus high‐carbohydrate diet Not defined |
133 per 1000 | 121 per 1000 (52 to 283) |
RR 0.91 (0.39 to 2.12) | 149 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of performance bias as participants and researchers were not blinded |
|
Han 2017 Soy‐ versus no soy‐protein diet Defined as BGL < 1.7 mmol/L (< 30.6 mg/dL) |
29 per 1000 | 88 per 1000 (10 to 806) |
RR 3.00 (0.33 to 27.42) | 68 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Risk of performance bias as participants and personnel were not blinded |
|
Han 2017 Ethnic specific diet versus standard healthy diet Not defined |
0 per 1000 | 0 per 1000 (0 to 0) |
RR not estimable | 20 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study. Risk of performance bias as participants and personnel were not blinded and selective reporting (reporting bias). There were no neonatal hypoglycaemic events in either group |
|
Brown 2017b Lifestyle intervention versus usual care or diet alone Not defined |
75 per 1000 | 74 per 1000 (49 to 114) |
Average RR 0.99 (0.65 to 1.52) | 3000 (6 RCTs) |
Moderate | "Allocation concealment was unclear in two trials and blinding was not undertaken in two other trials." |
|
Brown 2017c Exercise versus control Not defined |
59 per 1000 | 118 per 1000 (12 to 1000) |
RR 2.00 (0.20 to 20.04) | 34 (1 RCT) |
Very low | "Imprecision ‐ wide confidence intervals and low event rates. There is a lack of clarity for most items associated with risk of bias." |
|
Han 2012 Intensive management versus routine care Defined as: two studies: < 1.7 mmol/L (< 30.6 mg/dL) in any two consecutive measurements one study: < 1.94 mmol/L (< 35 mg/dL) |
66 per 1000 | 26 per 1000 (4 to 167) |
RR 0.39 (0.06 to 2.54) | 426 (2 RCTs) |
Very low | Evidence is based on two studies with few events and serious/very serious design limitations. Wide confidence intervals crossing the line of no effect and substantial heterogeneity: I² = 62% |
|
Raman 2017 Telemedicine versus standard care for glucose monitoring Defined as BGL <2.6 mmol/L in one study |
82 per 100 | 94 per 1000 (40 to 224) |
RR 1.14 (0.48 to 2.72) | 198 (3 RCTs) |
Very low | Evidence downgraded for imprecision (wide confidence intervals crossing the line of no effect, small sample sizes) and study limitations |
|
Raman 2017 Self‐ versus periodic‐glucose monitoring Not defined |
173 per 1000 | 111 per 1000 (67 to 183) |
RR 0.64 (0.39 to 1.06) | 391 (2 RCTs) |
Low | Evidence downgraded for imprecision (wide confidence intervals crossing the line of no effect) and study limitations |
|
Raman 2017 Continuous‐ versus self‐monitoring blood glucose Defined as blood glucose ≤ 45 mg/dL (2.5 mmol/L) |
130 per 1000 | 103 per 1000 (46 to 232) |
RR 0.79 (0.35 to 1.78) |
179 (2 RCTs) |
Very low | Evidence downgraded for imprecision (wide confidence intervals crossing the line of no effect, small sample sizes) and study limitations |
|
Raman 2017 Post‐ versus pre‐prandial glucose monitoring Defined as ≤ 30 mg/dL requiring glucagon or dextrose infusion for treatment during the first four days after birth |
212 per 1000 | 30 per 1000 (4 to 233) |
RR 0.14 (0.02 to 1.10) |
66 (1 RCT) |
Very low | Evidence downgraded for imprecision (wide confidence intervals crossing the line of no effect, single trial, small sample sizes) and study limitations |
|
Brown 2017d Insulin versus oral therapy Defined as < 2.6 mmol/L |
111 per 1000 | 126 per 1000 (94 to 1.52) |
Average RR 1.14 (0.85 to 1.52) | 3892 (24 RCTs) |
Low | Evidence downgraded for study limitations (lack of blinding) and inconsistency |
|
Brown 2017d Insulin type A versus B |
12 per 1000 | 28 per 1000 (1 to 1000) |
RR 2.28 (0.06 to 82.02) | 165 (3 RCTs) |
Very low | Evidence downgraded for study limitations (lack of blinding), imprecision (wide confidence intervals) and inconsistency |
|
Brown 2017d Insulin versus diet |
240 per 1000 | 211 per 1000 (82 to 583) |
RR 0.88 (0.34 to 2.24) | 176 (3 RCTs) |
Very low | Evidence downgraded for study limitations (lack of blinding), imprecision (wide confidence intervals) and inconsistency |
|
Brown 2017d Insulin versus exercise |
118 per 1000 | 59 per 1000 (6 to 589) |
RR 0.50 (0.05 to 5.01) | 34 (1 RCT) |
Very low | Evidence downgraded for study limitations (insufficient details) and imprecision (single small study, wide confidence intervals) |
|
Brown 2017d Insulin regimen A versus B Twice daily versus four times daily |
7 per 1000 | 59 per 1000 (7 to 464 |
RR 8.12 (1.03 to 64.03) | 274 (1 RCT) |
Very low | Evidence downgraded imprecision (large treatment effect, single small study, low event rates and wide confidence intervals) |
| 12.0 Adiposity ‐ neonate | ||||||
|
Brown 2017b Lifestyle intervention versus usual care or diet alone Defined as: neonatal fat mass (estimated from skinfold thickness) |
Mean mass: 427 g |
Mean mass: 37.80 g fewer (63.97 g fewer to 10.63 g fewer) |
MD ‐37.30 g (‐63.97 to ‐10.63) | 958 (1 RCT) |
Low | "Imprecision. Evidence is base on a single trial and there was no blinding of participants/researchers." |
|
Brown 2017d Insulin versus oral therapy Defined as percentage fat mass Defined as skin‐fold sum (mm) |
The mean percentage fat mass was 12.8% The mean skinfold sum was 16 mm |
MD 1.6% lower (3.77 % lower to 0.57% higher) MD 0.8 mm lower (0.49 mm lower to 0.73 mm higher) |
MD ‐1.60 (‐3.77 to 0.57) MD‐0.80 (‐2.33 to 0.73) |
82 (1 RCT) 82 (1 RCT) |
Very low | Evidence was downgraded for imprecision as based on one trial Evidence was downgraded for imprecision as based on one trial with wide confidence intervals and study limitations (selective reporting and other bias detected) |
| 12.0 Adiposity ‐ child | ||||||
|
Brown 2017b Lifestyle intervention versus usual care or diet alone Defined as: Childhood BMI1> 85th percentile kg/m² |
350 per 1000 | 318 per 1000 (262 to 388) |
RR 0.91 (0.75 to 1.11) | 767 (3 RCTs) |
Moderate | "Allocation concealment and randomisation was unclear in 1/3 trials and 1/3 trials did not blind participants/researchers." |
|
Brown 2017b Lifestyle intervention versus usual care or diet alone Defined as: Childhood BMI1z score |
The mean childhood BMI z score was 0.49 lower | The childhood BMI z score in the intervention group was 0.08 lower (0.28 lower to 10.63 lower) | MD 0.08 (‐0.28 to 0.44) | 199 (1 RCT) |
Very low | Imprecision ‐ evidence is based on one study and wide confidence interval crossing the line of no effect. Only reports on 199 children of the original trial of 1000 participants |
|
Brown 2017d Insulin versus oral therapy Defined as total fat mass (%) up to 2‐years |
The mean childhood total fat mass (%) was 16.4% | MD 0.5% higher (0.49 % lower to 1.49 % higher) | MD 0.50 (‐0.49 to 1.49) | 318 (1 RCT) |
Low | Evidence downgraded for study limitations (lack of blinding) and imprecision as based on a single study |
| 13.0 Diabetes | ||||||
| ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | Either no data were reported for this outcome in any of the included Cochrane systematic reviews or none of the included studies in the review pre‐specified this outcome |
| 14.0 Neurosensory disability | ||||||
|
Brown 2017d Insulin versus oral therapy Mild developmental delay (18 months) Hearing impairment (18 months) Visual impairment (18 months) |
104 per 1000 0 per 1000 21 per 1000 |
111 per 1000 (34 to 385) 0 per 1000 (0 to 0) 6 per 1000 (1 to 60) |
RR 1.07 (0.33 to 3.44) RR 0.31 (0.01 to 7.49) RR 0.03 to 2.90 |
93 (1 RCT) | Low | Evidence downgraded for imprecision as based on a single study with wide confidence intervals |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio | ||||||
| GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect | ||||||
1BMI is an acronym for Body Mass Index
24 g myo‐inositol + 400 µg folic acid orally per day and exercise and dietary advice versus placebo 400 µg folic acid orally per day and exercise and dietary advice.