. 2023 Oct 5;3:132. doi: 10.1038/s43856-023-00358-x

Table 5.

LGS/PLGS articles with precision-directed analyses or subpopulations of interest.

Author year (trial/cohort title) location	Analysis type/study design/ number randomized	Primary endpoint (SIG vs. NS)—% dropout multiple test correction performed?	Subgroup analyses—technology superior to standard of care	Subgroup analyses—technology comparable to standard of care
Beck⁵⁴ (DIaMonD) USA	Primary RCT 75	Δ TIR 7 months (SIG)—5% Unsure	Age < 50 vs. ≥ 50, TIR < 53 vs. ≥ 53%, TBR < 3 vs. ≥ 3%, HbA1c, SMBG frequency, education level, IAH, Hypoglycemia Fear Adults on CGM + pump (vs. CGM + MDI), aged 26–73 years, had improved TIR (+83 min [95% CI: 17–149]; p = 0.01).	HbA1c Baseline HbA1c did interact (p = 0.006) with the intervention on the outcome where HbA1c < 7.5% did not improve TIR.
Bergenstal⁶⁶ (ASPIRE) USA	Primary RCT 247	NH events AUC (SIG)—3% Unsure	Age 16–24 vs. 25–50 vs. 51–70, HbA1c ≤ 7 vs. >7%, diabetes duration ≤ 25 vs. >25 years In the cohort (ages 16–70 years), LGS reduced NH & HbA1c compared to SAP.
Rosenlund⁵⁵ Denmark	Primary RCT 60	Δ UACR 12 months (NS)—8% Unsure	Complication, sex, BMI, HbA1c SAP (vs. MDI) improved UACR in those with current or past albuminuria when adjusting for HbA1c, sex, and BMI (p = 0.02). UACR improved in those with current albuminuria (n = 48; −18 vs. +38%, p = 0.011).	Complication SAP-treated individuals with current or past albuminuria (aged 18–75 years) did not have statistical improvement in UACR over MDI (−13 vs. +30%, p = 0.051).
Weiss⁷³ (APSIRE) USA	Secondary RCT 247	NH events AUC (SIG)—3% Unsure	HbA1c Reduced NH events for participants with a baseline HbA1c < 7% (n = 95) and 7–8% (n = 115). NH event glucose nadir (mg/dL) was higher if HbA1c < 7% (49.7 ± 8.5 vs. 46.8 ± 9.6, p < 0.001) or 7–8% (49.9 ± 8.2 vs. 48.1 ± 9.6, p = 0.007).	HbA1c NH events not reduced in small group with HbA1c > 8% (n = 30). NH glucose nadir (mg/dL) was similar between LGS and SAP if HbA1c > 8% (49.5 ± 8.6 vs. 46.5 ± 10.1, p = 0.06).
Slover⁵⁶ (STAR 3) USA	Secondary Randomized crossover 156	HbA1c at 1 year (SIG)—0% Unsure	Age In the entire cohort (ages 7–18 years), HbA1c was reduced in SAP vs. MDI. Those aged 7–12 years (n = 82) had increased sensor wear compared to ages 13–18 (n = 74; p = 0.025).	Age Secondary metabolic endpoints were similar between age groups, except MAGE did not improve, and BMI rose in ages 13–18 years.
Ly⁷² Australia	Primary RCT 95	Mod/sev hypoglycemia 6 months (SIG)—9% Unsure	Age LGS improved hypoglycemia in the entire cohort (ages 4–50 years; IRR 3.6 [95% CI: 1.7–7.5]; p < 0.001) and in those ages 4–11 years (n = 30; IRR 5.5 [95% CI: 2.0–15.7]; p < 0.001).
Buckingham⁶⁸ (IN HOME CLOSED LOOP) North America	Primary Randomized crossover 81	Median TBR over 42 nights (SIG)—5% No	Age PLGS reduced TBR in the entire cohort (ages 4–14 years), specifically in ages 4–10 years by 50% (6.2–3.1%) and in ages 11–14 years by 54% (10.1–4.6%; p < 0.001 for both).	Age Secondary metabolic endpoints were similar between age groups except TAR, AM glucose, and HbA1c, which were not improved in ages 4–10 years (n = 36).
Calhoun⁶⁹ (IN HOME CLOSED LOOP) North America	Secondary Randomized crossover 127	Nights with hypoglycemia (SIG)—0% Yes	Age, sex, HbA1c, diabetes duration, daily % basal, TDD, bedtime BG, bedtime snack, IOB/TDD, weekday vs. weekend, exercise intensity, number of CGM values ≤ 60 mg/dL from 12-8PM, CGM rate of change before activation, time system activated Night use of PLGS reduced hypoglycemia in the entire cohort (ages 4–45 years) (OR 1.91 [99% CI: 1.57–2.32]; p < 0.001).
Abraham⁶⁴ (PLGM) Australia	Primary RCT 154	% Time <63 mg/dL at 6 months (SIG)—9% Unsure	Age, sex, HbA1c, IAH In the entire cohort (ages 8–20 years), percent time with hypoglycemia (<63 and <54 mg/dL) in PLGS vs. SAP was 2.6% vs. 1.5% (p < 0.0001 for both cutoffs).

Author year (trial/cohort title) location

Analysis type/study design/ number randomized

Primary endpoint (SIG vs. NS)—% dropout multiple test correction performed?

Subgroup analyses—technology superior to standard of care

Subgroup analyses—technology comparable to standard of care

Beck⁵⁴ (DIaMonD)

USA

Primary

RCT

Δ TIR 7 months (SIG)—5%

Unsure

Age < 50 vs. ≥ 50, TIR < 53 vs. ≥ 53%, TBR < 3 vs. ≥ 3%, HbA1c, SMBG frequency, education level, IAH, Hypoglycemia Fear

Adults on CGM + pump (vs. CGM + MDI), aged 26–73 years, had improved TIR (+83 min [95% CI: 17–149]; p = 0.01).

HbA1c

Baseline HbA1c did interact (p = 0.006) with the intervention on the outcome where HbA1c < 7.5% did not improve TIR.

Bergenstal⁶⁶ (ASPIRE)

USA

Primary

RCT

247

NH events AUC (SIG)—3%

Unsure

Age 16–24 vs. 25–50 vs. 51–70, HbA1c ≤ 7 vs. >7%, diabetes duration ≤ 25 vs. >25 years

In the cohort (ages 16–70 years), LGS reduced NH & HbA1c compared to SAP.

Rosenlund⁵⁵

Denmark

Primary

RCT

Δ UACR 12 months (NS)—8%

Unsure

Complication, sex, BMI, HbA1c

SAP (vs. MDI) improved UACR in those with current or past albuminuria when adjusting for HbA1c, sex, and BMI (p = 0.02). UACR improved in those with current albuminuria (n = 48; −18 vs. +38%, p = 0.011).

Complication

SAP-treated individuals with current or past albuminuria (aged 18–75 years) did not have statistical improvement in UACR over MDI (−13 vs. +30%, p = 0.051).

Weiss⁷³ (APSIRE)

USA

Secondary

RCT

247

NH events AUC (SIG)—3%

Unsure

HbA1c

Reduced NH events for participants with a baseline HbA1c < 7% (n = 95) and 7–8% (n = 115). NH event glucose nadir (mg/dL) was higher if HbA1c < 7% (49.7 ± 8.5 vs. 46.8 ± 9.6, p < 0.001) or 7–8% (49.9 ± 8.2 vs. 48.1 ± 9.6, p = 0.007).

HbA1c

NH events not reduced in small group with HbA1c > 8% (n = 30). NH glucose nadir (mg/dL) was similar between LGS and SAP if HbA1c > 8% (49.5 ± 8.6 vs. 46.5 ± 10.1, p = 0.06).

Slover⁵⁶ (STAR 3)

USA

Secondary

Randomized crossover

156

HbA1c at 1 year (SIG)—0%

Unsure

Age

In the entire cohort (ages 7–18 years), HbA1c was reduced in SAP vs. MDI. Those aged 7–12 years (n = 82) had increased sensor wear compared to ages 13–18 (n = 74; p = 0.025).

Age

Secondary metabolic endpoints were similar between age groups, except MAGE did not improve, and BMI rose in ages 13–18 years.

Ly⁷²

Australia

Primary

RCT

Mod/sev hypoglycemia 6 months (SIG)—9%

Unsure

Age

LGS improved hypoglycemia in the entire cohort (ages 4–50 years; IRR 3.6 [95% CI: 1.7–7.5]; p < 0.001) and in those ages 4–11 years (n = 30; IRR 5.5 [95% CI: 2.0–15.7]; p < 0.001).

Buckingham⁶⁸ (IN HOME CLOSED LOOP)

North America

Primary

Randomized crossover

Median TBR over 42 nights (SIG)—5%

Age

PLGS reduced TBR in the entire cohort (ages 4–14 years), specifically in ages 4–10 years by 50% (6.2–3.1%) and in ages 11–14 years by 54% (10.1–4.6%; p < 0.001 for both).

Age

Secondary metabolic endpoints were similar between age groups except TAR, AM glucose, and HbA1c, which were not improved in ages 4–10 years (n = 36).

Calhoun⁶⁹ (IN HOME CLOSED LOOP)

North America

Secondary

Randomized crossover

127

Nights with hypoglycemia (SIG)—0%

Yes

Age, sex, HbA1c, diabetes duration, daily % basal, TDD, bedtime BG, bedtime snack, IOB/TDD, weekday vs. weekend, exercise intensity, number of CGM values ≤ 60 mg/dL from 12-8PM, CGM rate of change before activation, time system activated

Night use of PLGS reduced hypoglycemia in the entire cohort (ages 4–45 years) (OR 1.91 [99% CI: 1.57–2.32]; p < 0.001).

Abraham⁶⁴ (PLGM)

Australia

Primary

RCT

154

% Time <63 mg/dL at 6 months (SIG)—9%

Unsure

Age, sex, HbA1c, IAH

In the entire cohort (ages 8–20 years), percent time with hypoglycemia (<63 and <54 mg/dL) in PLGS vs. SAP was 2.6% vs. 1.5% (p < 0.0001 for both cutoffs).