1 Hoofdstuk 2 DM type 1 Uitgangsvraag 1 Met welk insulineschema start je bij een volwassene met een nieuw ontdekte type 1 diabetes zonder acidose? Inleiding Bij Diabetes Mellitus type 1 (DM1) is er sprake van een absoluut insuline tekort en dient de behandeling altijd te bestaan uit het toedienen van insuline. De meest logische en fysiologische behandeling bij DM1 is een basis/bolus regime. Soms wordt echter vanwege praktische redenen gestart met een tweemaal daags regime met mix-insulines of een combinatie van mix- en kortwerkende insuline. De uitgangsvraag is hoe een volwassene met een nieuw ontdekte DM1 zonder acidose optimaal kan worden behandeld: bij welk insulineschema is de kans op acute en late complicaties het kleinst? De relevante uitkomstmaten zijn micro- en macrovasculaire complicaties, glycemische controle (HbA1c), en bijwerkingen (hypoglycemie). Er is gezocht naar literatuur waarin de effectiviteit van insulineschema’s wordt vergeleken bij volwassen patiënten met nieuw ontdekte DM1, met micro- en macrovasculaire complicaties, glycemische controle (HbA1c) en bijwerkingen als uitkomstmaat. De buitenlandse richtlijnen NICE en ADA 2012 geven geen duidelijk advies voor een startschema. Aanvullend is gezocht vanaf 1-1-2008 op systematische reviews en RCT’s, in de databases Medline, Embase en de Cochrane Library. Uit het totaal van 34 treffers werden op basis van titel en abstract, vier studies geselecteerd (Thraikill et al., 2011; Hassan et al., 2008; Rabbone et al., 2008; Skogsberg et al., 2008). Bij analyse van de full text zijn alle vier de RCT’s geëxcludeerd op grond van de studiepopulaties die geheel dan wel grotendeels uit kinderen of adolescenten bestaan. Samenvatting van de literatuur De zoekactie leverde geen RCT’s op die zijn uitgevoerd bij volwassenen met nieuw gediagnosticeerde DM1. Veel studies zijn bij kinderen en adolescenten uitgevoerd. De aanpak bij
70
Embed
Hoofdstuk 2 DM type 1 - internisten · 1 Hoofdstuk 2 DM type 1 Uitgangsvraag 1 Met welk insulineschema start je bij een volwassene met een nieuw ontdekte type 1 diabetes zonder acidose?
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Hoofdstuk 2 DM type 1
Uitgangsvraag 1
Met welk insulineschema start je bij een volwassene met een nieuw ontdekte type 1
diabetes zonder acidose?
Inleiding
Bij Diabetes Mellitus type 1 (DM1) is er sprake van een absoluut insuline tekort en dient de
behandeling altijd te bestaan uit het toedienen van insuline. De meest logische en fysiologische
behandeling bij DM1 is een basis/bolus regime. Soms wordt echter vanwege praktische redenen
gestart met een tweemaal daags regime met mix-insulines of een combinatie van mix- en
kortwerkende insuline.
De uitgangsvraag is hoe een volwassene met een nieuw ontdekte DM1 zonder acidose optimaal
kan worden behandeld: bij welk insulineschema is de kans op acute en late complicaties het
kleinst? De relevante uitkomstmaten zijn micro- en macrovasculaire complicaties, glycemische
controle (HbA1c), en bijwerkingen (hypoglycemie).
Er is gezocht naar literatuur waarin de effectiviteit van insulineschema’s wordt vergeleken bij
volwassen patiënten met nieuw ontdekte DM1, met micro- en macrovasculaire complicaties,
glycemische controle (HbA1c) en bijwerkingen als uitkomstmaat. De buitenlandse richtlijnen
NICE en ADA 2012 geven geen duidelijk advies voor een startschema. Aanvullend is gezocht
vanaf 1-1-2008 op systematische reviews en RCT’s, in de databases Medline, Embase en de
Cochrane Library. Uit het totaal van 34 treffers werden op basis van titel en abstract, vier studies
geselecteerd (Thraikill et al., 2011; Hassan et al., 2008; Rabbone et al., 2008; Skogsberg et al.,
2008). Bij analyse van de full text zijn alle vier de RCT’s geëxcludeerd op grond van de
studiepopulaties die geheel dan wel grotendeels uit kinderen of adolescenten bestaan.
Samenvatting van de literatuur
De zoekactie leverde geen RCT’s op die zijn uitgevoerd bij volwassenen met nieuw
gediagnosticeerde DM1. Veel studies zijn bij kinderen en adolescenten uitgevoerd. De aanpak bij
2
deze patiëntencategorie verschilt echter van de aanpak bij volwassenen, omdat bij kinderen en
adolescenten, groei, hormonale veranderingen en therapietrouw meer op de voorgrond zal
staan dan bij volwassenen. Bij volwassenen ligt de nadruk minder op therapietrouw en meer op
het voorkomen van lange termijn complicaties en hypoglycemieën.
Conclusie
De literatuursearch heeft geen geschikte studies opgeleverd die een antwoord geven op de
uitgangsvraag.
Overwegingen
In het algemeen wordt voor patiënten met DM1 aanbevolen een keuze te maken voor een basis/
bolusschema met een pen (Multiple Daily Injections, MDI) dan wel via Continuous Subcutaneous
Insulin Infusion (CSII). In de praktijk wordt meestal gestart met een langwerkend insuline voor de
nacht in combinatie met een kortwerkend insuline bij elke maaltijd. Een afwijkend schema kan
gebaseerd zijn op een individuele dagindeling, zoals bijvoorbeeld een schema met een pre-mixed
insuline bij het ontbijt, kortwerkend insuline bij het avondeten en langwerkend insuline voor de
nacht. Aangezien er uit de DCCT aanwijzingen zijn dat de complicatie incidentie bij een pre-
mixed schema van tweemaal daags hoger is (DCCT Research Group, 1993; Nathan et al., 2005),
wordt hier in de praktijk alleen voor gekozen als de patiënt hier een sterke voorkeur voor heeft.
Het voordeel van dit schema is dat patiënten minder vaak insuline hoeven te spuiten in
tegenstelling tot 4 dd bij het basis/bolusschema, wat mede als voordeel heeft dat werkende/
studerende patiënten dit thuis kunnen toedienen in plaats van op de werkplek.
In de praktijk wordt vaak gekozen voor een startdosering van 0,2 units/kg, waarbij de helft
daarvan als kortwerkend insuline wordt gegeven en de andere helft als langwerkend insuline.
In de Diabetes Control and Complications Trial (DCCT) en bijbehorende follow-up studie (EDIC,
Epidemiology of Diabetes Interventions and Complications trial) is aangetoond dat glycemische
controle nadat de diagnose DM gesteld is, het risico op lange termijn micro- en macrovasculaire
complicaties vermindert (DCCT Research Group 1993; Nathan et al., 2005; White et al., 2008).
Het verminderde risico bijvoorbeeld op retinopathie, bij de intensief behandelde groep bleef na
10 jaar follow-up bestaan ondanks het feit dat het verschil in HbA1c tussen de intensief
behandelde en conventioneel behandelde groep verdween (DCCT Research Group, 1993; Nathan
et al., 2005). De intensief behandelde groep werd met tenminste drie dagelijkse insuline injecties
3
behandeld of CSII (basis-bolus principe), terwijl de conventionele groep met slechts één-twee
insuline injecties per dag werd behandeld (met name pre-mixed schema’s).
Er zijn een aantal redenen waarom het bereiken van vroege glycemische controle niet altijd
mogelijk is, zoals het optreden van, of de angst voor hypoglycemieën, en de noodzaak voor
intensief dagelijkse management van de diabetes met frequente glucosemetingen en
aanpassingen in insulinedosering. Naast het hoofddoel om vroege glycemische controle te
bereiken, wordt de keuze van het optimale insuline schema dan ook bepaald door een groot
aantal factoren, zoals: leeftijd, leefstijl/dagindeling, beroep, comorbiditeit, motivatie,
mogelijkheid tot zelfcontrole en zelfregulatie.
Samenvattend zijn er na het stellen van de diagnose DM1, in grote lijnen twee startschema’s
met MDI mogelijk:
1) Basis/bolusschema via een insuline pen:
3 dd (ultra)kortwerkende insuline bij de maaltijd + 1-2 dd een (middel)langwerkende
insuline.
2) Pre-mixed insuline:
a) Tweemaal daags schema;
b) Pre-mixed insuline in combinatie met kortwerkende en/of langwerkende insuline
(bijvoorbeeld pre-mixed insuline bij het ontbijt, kortwerkende bij het avondeten en
een langwerkende voor de nacht).
Het voordeel van dit schema is dat patiënten minder vaak insuline hoeven te spuiten. Het grote
nadeel is echter dat het de endogene insuline-excretie niet goed nabootst wordt en er daardoor
minder goede glycemische controle bestaat en striktere voedingsvoorschriften nodig zijn. Dit
schema heeft derhalve niet de eerste voorkeur, maar kan op grond van individuele
patiëntomstandigheden worden overwogen.
Aanbevelingen
Bij patiënten met nieuw gediagnosticeerde diabetes mellitus type 1 zonder acidose:
Start met een basis/bolusschema;
Overweeg in eerste instantie MDI;
Gebruik bij voorkeur geen pre-mixed tweemaal daags schema;
Startdosering van 0,2 units/kg per dag, verdeeld over meerdere injecties.
4
Literatuur
DCCT Research Group. (1993). The effect of intensive treatment of diabetes on the development
and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J
Med., 329, 977-986.
Hassan, K., Rodriguez, L. M., Johnson, S. E., Tadlock, S., & Heptulla, R.A. (2008). A randomized,
- Vrouwen met een (toekomstig) kinderwens en zwangeren;
- DM1 patiënten met preterminaal nierfalen: screening nier-/pancreastransplantatie?
- Hypoglycemia unawareness;
- CSII;
- Indicaties voor CGM (continuous glucose monitoring);
- Alle nieuwe DM1 patiënten ter beoordeling indicatie immunosupressiva.
Het nut van ondersteuning van de diabetes specialist door diabetesverpleegkundige is
onomstotelijk bewezen. Een niet-internist-endocrinoloog dient dan ook tenminste nauw samen
te werken met een diabetesverpleegkundige.
In de Landelijke Transmurale Afspraken (Sluiter et al., 2012), opgesteld door het NHG en de NIV,
staat dat DM1 in de tweede lijn behandeld dient te worden.
43
De ADA richtlijn adviseert het screenen op andere auto-immuun aandoeningen (zoals
hypothyreoidie, vitamine B12 deficiëntie en coeliaki) in overweging te nemen, in het bijzonder in
aanwezigheid van specifieke symptomen. Periodieke screening wordt niet aanbevolen.
Aanbevelingen
Behandel mensen met type 1 diabetes altijd binnen een multidisciplinair behandelteam
(internist, diabetesverpleegkundige, diëtist, voetenteam en op afroep, psycholoog, gynaecoloog,
en anderen) in de tweede lijn. Interacteer als internist actief binnen dit behandelteam.
Literatuur
ADA Guidelines 2012: Standards of Medical Care in Diabetes. (2012). Diabetes Care, 35(1), S11-
S63. http://www.diabetes.org
Sluiter, A.C., Van Wijland, J.J., Arntzenius, A.B., Bots, A.F.E., Dijkhorst-Oei, L.T., Van der Does,
F.E.E., Palmen, J.V.H., Potter van Loon, B.J.,…(2012). Landelijke Transmurale Afspraak Diabetes
Mellitus type 2. Huisarts Wet, 55, S1-12.
Zgibor, J.C., Songer, T.J., Kelsey, S.F., Weissfeld, J., Drash, A.L., Becker, D., & Orchard, T.J. (2000).
The association of diabetes specialist care with health care practices and glycemic control in
patients with type 1 diabetes: a cross-sectional analysis from the Pittsburgh epidemiology of
diabetes complications study. Diabetes Care, 23, 472-6.
Zgibor, J.C., Songer, T.J., Kelsey, S.F., Drash, A.L., & Orchard, T.J. (2002). Influence of health care
providers on the development of diabetes complications: long-term follow-up from the
Pittsburgh Epidemiology of Diabetes Complications Study. Diabetes Care, 25, 1584-90.
44
BIJLAGEN
45
Evidence table for intervention studies
Research question: Does the addition of rapid-acting insulin analogs to the basal insulin regime improve treatment of patients with type 1 diabetes?
Study reference
Study characteristics Patient characteristics Intervention (I)
Comparison / control (C)
Follow-up Outcome measures and effect size Comments
Rys et al, 2011 (individual study characteristics deduced from Rys et al, 2011)
SR and meta-analysis of 18 RCTs Literature search up to July 2009 A Ampudia-Blasco, 2005 B Arslanian, 2005 C Bode, 2001 D Bode, 2002 E Boehm, 2002 F Bott, 2003 G Chen, 2006 H Cherubini, 2006 I Danne, 2007 J De Vries, 2003 K Heller, 2004 L Home, 1998 M Home, 2000 N Iwamoto, 2001 O Mathiesen, 2007 P Mortensen, 2006 Q Raskin, 2001 R Tamas, 2001 Parallel design (A-F, H, J, M-P, R) or cross-over design (G, I, K, L, Q) 14 RCTs adults, 4 children/ adolescents (B, H, I, P) 3 adequate allocation concealment (J, K, R) 2 CSII (C, D)
Inclusion criteria SR: DM1 (DM2 not presented in this table), RCTs insulin aspart vs regular human insulin, no restriction on age, studies ≥ 4 wks, subgroup analysis ≥ 12 wks for HbA1c. Exclusion criteria SR: N < 10, < 4 wks, gestational or secondary DM. N/N, mean age (yrs) A 28/26; 32.3 B 187/96; 11.7 C 19/10; 36.6 D 59/59; 42.7 E 88/102; >18 F 283/141; 36.9 G 27; 44.8 H 30; 8.1 I 26/26; 5 J 186/181; 36.9 K 155; 35.7 L 104; 34.3 M 707/358; 38 N 143/62; 33.4 O 157/165; 29.0 P 86/81; 14.4 Q 596/286; 39.2 R 213/213; 36 HbA1c at baseline (%) A 8.5 B 8.3 C 7.2 D 7.4 E ≤ 11.0 F 7.5 G not reported
Rapid-acting analogs (RAA) Insulin aspart (IAsp) Insulin regimen (basal insulin) A MDI B MDI C CSII D CSII E Biphas F MDI G Biphas H MDI I MDI J MDI K MDI L MDI M MDI N MDI O MDI P Biphas Q MDI R MDI MDI with NPH, except A (glargine
Regular human insulin (RHI) Same insulin regimen as in intervention group
Duration of follow-up (months): A 6 B 6 C 7 wks D 4 E 3 F 6 G 2 x 3 H 4.5 I 2 x 3 J 16 K 2 x 4 L 2 x 1 M 6 N 6 O 9 P 4 Q 6 R 3
For how many participants were no complete outcome data available? Not stated
Effect on glycemic control (HbA1c , post-prandial glucose [PPG]), fasting glucose, quality of life/ treatment flexibility, adverse effects; weighted mean difference (WMD), standardized mean difference (SMD), relative risk (RR) [meta-analysis]; [95%CI] Effect on HbA1c I minus C (WMD; %)); 13 studies pooled <0 favours I (IAsp) >0 favours C (RHI) 16 studies measured HbA1c Excluded: 3 studies lacking sufficient numerical data (G,H,I) A 0.14 [-0.38; 0.66] B 0.00 [-0.35; 0.35] C -0.20 [-0.66; 0.26] D -0.35 [-0.62; -0.08] E 0.19 [-0.05; 0.43] J -0.14 [-0.35; 0.07] K 0.00 [-0.20; 0.20] M -0.12 [-0.22; -0.02] N -0.24 [-0.57; 0.09] O -0.08 [-0.22; 0.06] P 0.14 [-0.33; 0.61] Q -0.15 [-0.26; -0.04] R -0.16 [-0.30; -0.02] Overall: -0.11 [-0.16; -0.06] Heterogeneity: I2=17.5%; p=0.27 Subgroup-analyses based on basal insulin treatment: Subgroup-CSII (C,D; 2 studies pooled) Overall: -0.31 [-0.55; -0.08] Subgroup-basal bolus (A,B,J,K,M,N,O,Q,R; 9 studies pooled) Overall: -0.12 [-0.17; -0.06] Effect on post-prandial glucose (PPG; self-measured) I minus C (mmol/l) Reported in 10 studies; studies w/o suff numerical data were excluded Post-breakfast glucose (5 studies) -1.43 [-1.75; -1.11] Post-lunch glucose (5 studies) -1.11 [-1.61; -0.61] Post-dinner glucose (6 studies) -0.97 [-1.25; -0.69] Effect on fasting glucose (FG)
Authors conclude that IAsp compared with RHI resulted in moderately better metabolic control, as demonstrated by reductions in HbA1c and PPG, and no change in FG; IAsp also resulted in greater treatment satisfaction and a significant reduction of risk for nocturnal hypoglycaemic episodes, but not severe hypoglycaemias or hypoglycaemia in general Note that it is unclear whether the extra reduction in HbA1c by 0.11 % is clinically relevant Note that the 2 studies with CSII regimen had the most favourable HbA1c reduction (-0.31 [-0.55; -0.08]) as compared to 9 studies with basal-bolus (MDI) regimen ( -0.12 [-0.17; -0.06]) Note that in most studies, no blinding or allocation concealment was applied (high risk of bias), and that the data reported in some studies were incomplete (HbA1c was measured in most studies, but PPG or hypoglycaemic episodes only in some studies); SR provides no data on risk of publication bias Quality of RCTs (Jadad score): R1 Random allocation R2 with appropriate rand. method B0 Not double-blind B1 Double-blind W0 No info on reasons withdrawal W1 information provided A R1 B0 W0 B R1 B0 not reported C R1 B0 W1 D R2 B0 W1
46
H 7.5 I 7.8 J 8.38 K 8.6 L 7.1 M 7.97 N 7.5 O 6.9 P 9.6 Q 7.9 R 8.325
), and N (not reported)
I minus C (mmol/l); 5 studies pooled Reported in 12 studies; in 3 studies RHI was superior over insuline aspart, in 9 studies differences were not significant; 5 studies provided numerical data Fasting glucose 0.15 [-0.55; 0.86) Heterogeneity: I2=72.8%; p=0.005 Adverse effects (hypoglycemic episodes) I minus C (RR) Reported in 7 studies, in 6 studies episodes were self-reported All hypoglycaemic episodes (6 studies) 1.06 [1.01; 1.10] Reported in 3 studies Nocturnal hypoglyc episodes (3 studies) 0.67 [0.54; 0.83] Reported in 8 studies, 7 with sufficient numerical data Severe hypoglyc episodes (7 studies) 0.92 [0.75; 1.12] Quality of life / treatment flexibility I minus C (SMD; DTSQ questionnaire score) Reported in 3 studies, the flexibility part in only 2 Total score (3 studies) 0.30 [0.20;0.40] Flexibility score (2 studies) 0.31 [0.15; 0.47]
E R2 B0 W1 F R1 B0 W0 G R1 B0 W1 H R1 B0 W0 I R1 B0 W0 J R2 B0 W1 K R2 B1 W1 L R1 B1 W1 M R1 B0 W1 N R1 B0 W0 O R2 B0 W0 P R1 B0 W0 Q R1 B0 W1 R R2 B0 W0 Level of evidence: GRADE MODERATE* IAsp vs RHI Effect on HbA1c LOW** IAsp vs RHI Effect on PPG MODERATE*** IAsp vs RHI Effect on FG LOW**** IAsp vs RHI Adverse effects (hypoglycemic episodes) *Downgraded 1 point because of serious study limitations: lack of blinding (12 out of 13 studies) and allocation concealment (10 out of 13 studies) **Downgraded 2 points because of serious study limitations (lack of blinding and allocation concealment), and indirectness (3 out of 5 [PBG], 1 out of 5 [PLG], and 2 out of 6 [PDG] studies included children/adolescents) *** Downgraded 1 point because of serious study limitations (lack of blinding and allocation concealment) ****Downgraded 2 points because of serious study limitations (lack of blinding and allocation concealment), and inconsistency (inconsistent results across studies; difference in estimate of treatment effect depending on outcome measure [overall vs nocturnal vs severe hypoglycemic episodes])
47
A significant reduction of postprandial glucose levels was found in the groups receiving insulin aspart compared to regular human insulin (pooled results of 5-6 studies) The risk of nocturnal hypoglycaemia measured in 3 studies, was significantly lower in insulin aspart group than in the regular human insulin group (RR= 0.67 [0.54; 0.83]) In contrast, the insulin aspart group showed a slight, but statistically significant, risk increase of all hypoglycaemic episodes compared to the regular human insulin group (RR=1.06 [1.01; 1.10]; pooled result of 6 studies) Pooled analysis of 2 studies showed that using insulin aspart instead of regular human insulin resulted in greater treatment satisfaction
Brock Jacobsen et al, 2011
Cross-over study; double-blind, randomized, 2 periods Funding: supported by Novo Nordisk
Inclusion criteria: type 1 DM for > 1 year; age 18–60 years; BMI of 18–27.5; required to use soluble human insulin before breakfast, lunch and dinner and NPH at bedtime for at least 3 months before entering study Exclusion criteria: pregnancy, impaired vision, impaired renal or hepatic function, cardiac diseases, uncontrolled hypertension, hypoglycaemic unawareness 13 men, 3 women Age 44.4 ± 8.2 years BMI 24.6 ± 1.3 HbA1c 7.8 ± 1.1% Diab duration 19± 10 years At baseline, mean daily insulin dose was: soluble insulin 29.4 IU and NPH 22.3 IU
Regular human insulin (RHI) Actrapid; 3X/d N=16 Same insulin regimen as in intervention group
Duration of follow-up: 2X8 weeks
After 8 weeks of treatment, patients were hospitalized for 22 h For how many participants were no complete outcome data available? 2/16 (13%); patients left ‘for personal reasons’
Effect on glycemic control (HbA1c , post-prandial glucose [PPG]), patient satisfaction, adverse effects; absolute amount, counts, frequency [SD] Effect on HbA1c (%) I: 7.0 ± 0.2 C: 7.0 ± 0.2 P=NS Adverse effects (hypoglycemic episodes) I: 214 events (3 nocturnal; 2 extreme [≤ 3.5 mmol/l] C: 297 events (5 nocturnal; 3 extreme) Hypoglycaemic frequency (events/week) I: 0.9 ± 0.1 C: 1.1 ± 0.2 P=NS Patient satisfaction (VAS; 0-6) P=NS During hospitalization: Effect on post-prandial glucose (PPG) I versus C (mmol/l; p-value) Post-breakfast: 8.2 ± 1.0 vs 12.6 ± 1.1 (P = 0.005) Post-lunch: 3.6 ± 0.4 vs 4.5 ± 0.6 (P = 0.02) Post-dinner: 7.9 ± 1.2 vs 9.9 ± 1.3 (P = 0.01) statistically significant lower PPG during IAsp treatment
Authors conclude that IAsp did not improve glycaemic control, and no difference in hypoglycaemic frequency was observed IAsp elicited a slightly different physiological response to spontaneous hypoglycaemia compared with human insulin; however, the physiological and clinical relevance of these effects remain speculative Note: during hospitalization, a statistically significant lower PPG was observed during IAsp treatment Level of evidence: GRADE see above, this study does not alter the conclusion based on Rys et al (2011; note the small study size (N=16) as compared to the studies included in the meta-analysis)
48
Evidence table for intervention studies
Research question: Does the use of long-acting insulin analogs instead of NPH human insulin improve treatment of patients with type 1 diabetes?
Study reference
Study characteristics
Patient characteristics
Intervention (I) Comparison / control (C)
Follow-up Outcome measures and effect size
Comments
Sanches, 2011 [individual study charac-teristics deduced from Sanches, 2011]
SR and meta-analysis of 16 RCTs Literature search up to Dec 2010 A: Home, 2005 B: Porcellati, 2004 C: Raskin, 2000 D: Ratner, 2000 E: Rosenstock, 2000a F: Bolli, 2009 G: Fulcher, 2005 H: Bartley, 2008 I: De Leeuw, 2005 J: Hermansen, 2004 K: Home, 2004a L: Home, 2004b M: Pieber, 2005a N: Pieber, 2005b O: Russell-Jones, 2004 P: Pieber, 2007 Source of funding: 3 studies on glargine and 6 on detemir were sponsored by pharmaceutical industry
Inclusion criteria SR: RCT, adult patients with established type 1 diabetes, received long-acting insulin analogue (glargine or detemir) or NPH insulin alone or in combination with rapid acting insulin analogues (Lispro or Aspart) for at least 4 weeks Exclusion: studies with crossover design N; mean age A: 585; 39.0 yrs B: 124; 35.0 yrs C: 619; 39.2 yrs D: 534; 38.6 yrs E: 344; 37.7 yrs F: 175; 40.5 yrs G: 125; 36.3 yrs H: 495; 35.0 yrs I: 315; 40.5 yrs J: 585; 39.0 yrs K: 585; 39.8 yrs L: 271; 39.6 yrs M: 268; 40.0 yrs N: 261; 40.7 yrs O: 747; 40.3 yrs
Long-acting analogs (LAA) Glargine dose (U/Kg/Day) A: 0.27 B: 0.66 C: NA D: NA E: NA F: 0.28 G: 0.39 Detemir dose (U/Kg/Day) H: 0.37 I: 0.36 J: 0.44 K: 0.48 L: 0.48 M: 0.35 N: 0.34 O: 0.31 In 2 studies (P; one study not detailed), Glargine was compared to Detemir dose (U/Kg/Day) P: 0.39 (glargine) vs 0.36 (detemir)
For how many participants were no complete outcome data available? Not stated
Effect on glycemic control (HbA1c), adverse effects (any hypoglycemia episode, nocturnal hypoglycemia); mean difference, OR [95%CI] Mean difference HbA1c (1) Glargine minus NPH A: -0.11 [-0.25, 0.03] B: -0.40 [-0.71, -0.09] C: 0.00 [-0.13, 0.13] D: 0.05 [-0.09, 0.19] E: 0.00 [-0.18, 0.18] F: -0.03 [ -0.20, 0.14] G: -0.30 [ -0.56, -0.04] Total: -0.06 [-0.14, 0.02] Heterogeneity: I2=44% (2) Detemir minus NPH (2a) Detemir once a day H: -0.22 [-0.39, -0.05] O: 0.00 [-0.15, 0.15] Subtotal: -0.11 [-0.32, 0.11] Heterogeneity: I2=71% Note: only 2 studies (2b) Detemir twice a day I: -0.08 [-0.31, 0.15] J: -0.22 [-0.33, -0.11] K: -0.20 [-0.44, 0.04] L: -0.17 [-0.41, 0.07] M: -0.04 [-0.27, 0.19] N: -0.10 [-0.34, 0.14] Subtotal : -0.14 [-0.22, -0.07] Heterogeneity: I2=0%
Authors conclude that there are no significant differences in HbA1c change between glargine or detemir (once daily) and conventional human insulin (NPH insulin). Twice daily regimen of detemir caused a difference in HbA1c that favored detemir (-0.14% [-0.22; -0.07]). No significant differences between glargine and detemir in safety or HbA1c mean change Note that results for hypoglycemia were inconsistent (high heterogeneity; I2); also note that the definition of hypoglycemia was not consistent across the studies Quality of RCTs (Jadad score): Max=5 A: 2 B: 4 C: 2 D: 0 E: 5 F: 2 G: 2 H: 2 I: 2 J: 2 K: 2
Total: -0.13 [-0.19, -0.06] Heterogeneity: I2=13% (3) Glargine or detemir (minus NPH): (1) and (2) combined Total: -0,09 [-0,14; -0,03] Heterogeneity: I2=38% (4) Glargine minus detemir P: -0.13 [-0.32, 0.06] Total (95%CI): -0.07 [-0.19, 0.06] Heterogeneity: I2=0% Note: only 2 studies Adverse effects (hypoglycemia) OR [95%CI] (1) Glargine vs. NPH Hypoglycemia 1.12 [0.69-1.81] Nocturnal 1.93 [0.78-4.79] High heterogeneity (I2>66%) (2) Detemir vs. NPH: Hypoglycemia 0.88 [0.52-1.51] Nocturnal 0.91 [0.52-1.60] High heterogeneity (I2>83%) (3) Glargine vs. Detemir: Hypoglycemia 2.05 [0.98-4.29] Nocturnal 1.24 [0.83-1.84] Only 2 studies included
L: 2 M: 2 N: 2 O: 2 P: 3 Level of evidence: GRADE Across all studies analysed MODERATE* LAA vs NPH Effect on HbA1c LOW** LAA vs NPH Adverse effects (hypoglycemic episodes) *Downgraded 1 point because of serious study limitations: low Jadad score (lack of blinding and allocation concealment) **Downgraded 2 points because of serious study limitations (low Jadad score), and imprecision (high heterogeneity; wide confidence intervals)
Szypowska, 2011
SR and meta-analysis of 10 RCTs Literature search up to Nov 2010
Inclusion criteria SR: RCTs ≥ 12
weeks; basal‑bolus therapy; DM type 1 with history ≥1 year Crossover trials and
Long-acting analogs (LAA) Detemir
NPH Endpoint of follow-up: Treatment duration 4-24 months
For how many participants were no complete outcome data available? Not stated
Effect on glycemic control (HbA1c), adverse effects (any / nocturnal / severe hypoglycemia); weighted mean difference, OR [95%CI] Mean difference HbA1c Detemir minus NPH Total: –0.073 [–0.135; –0.011]
Authors conclude that detemir, as compared with NPH, provided a minor benefit in terms of the HbA1c, and
reduced the risk of all‑day, nocturnal, and severe hypoglycemic episodes
50
Study reference
Study characteristics
Patient characteristics
Intervention (I) Comparison / control (C)
Follow-up Outcome measures and effect size
Comments
trials with children or adolescents not excluded Unpublished trials (Novo Nordisk) were also included
Heterogeneity: I2=0% Adverse effects(hypoglycemia) RR [95%CI] Detemir vs. NPH: All 0.98 [0.961; 0.996]; I2=26% Nocturn 0.88 [0.82; 0.94]; I2=0 Severe 0.67 [0.55; 0.81]; I2=51
Note: meta-analysis differs from that in Sanches et al (2011) in that a cross-over study and a study on children/adolescents was included as well as 3 unpublished trials from the Novo Nordisk trial registry Level of evidence: GRADE (analyzed collectively; see Sanches, 2011)
Garg, 2010 SR of 19 RCTs 9 RCTs comparing glargine with NPH (MDI regimen) No meta-analysis Literature search up to Feb 2010
Inclusion criteria SR: RCTs ≥ 12 weeks; patients (adult, children, adolescents) using once-daily insulin glargine Exclusion: pregnant women; studies not providing separate data on T1DM Crossover trials and trials with children or adolescents not excluded Very small studieswith ≤50 patients were also included
Long-acting analogs (LAA) Glargine
NPH (MDI) Endpoint of follow-up: Treatment duration 3-12 months
For how many participants were no complete outcome data available? Not stated
Effect on glycemic control (HbA1c), adverse effects (any / nocturnal / severe hypoglycemia); p-values No meta-analysis Effects on HbA1c out of 9 RCTS: 4 showed statistically significantly greater improvements with glargine compared with NPH insulin 4 showed no difference 1 showed slightly increased HbA1c (P=NS) Adverse effects(hypoglycemia) Out of 9 RCTs: 5 showed significantly reduced risk of symptomatic hypoglycemia 4 showed no difference in risk
Authors conclude that glycemic control with glargine is at least comparable to that with NPH, while the risk for hypoglycemia is significantly reduced Note: SR differs from Sanches et al (2011) in that a cross-over study, a subgroup analysis and a very small study were included, and no meta-analysis was performed
Monami, 2009
SR and meta-analysis of 20 RCTs Literature search up to
Effect on glycemic control (HbA1c), adverse effects (any /severe /nocturnal hypoglycemia); mean
Authors conclude that long-acting analogues as basal insulin replacement DMT1 patients had a small
51
Study reference
Study characteristics
Patient characteristics
Intervention (I) Comparison / control (C)
Follow-up Outcome measures and effect size
Comments
Apr 2008
Exclusion criteria: prandial insulin not comparable in treatment arms Crossover trials and trials with children or adolescents not excluded 4 unpublished trials were also included
difference, OR [95%CI] Mean difference HbA1c Total: -0.07 [-0.13; -0.01]; different LAA (detemir, glargine) yielded similar results Adverse effects (hypoglycemia) OR [95%CI]; MH estimates Any/severe/nocturnal 12 / 15 /13 RCTs Any 0.82 [0.61; 1.07] Severe 0.73 [0.60; 0.89] Nocturnal 0.69 [0.55; 0.86] similar point-estimates for glargine and detemir, but OR estimates for severe and nocturnal hypoglycemia not statistically significant for glargine (these RCTs enrolled fewer patients)
(beneficial) effect on HbA1c, and also reduced the risk of nocturnal and severe hypoglycaemia (by about 30%) Note that the overall incidence of hypoglycemia does not appear to be different between LAAs and NPH insulin Note: meta-analysis differs from that in Sanches et al (2011) in that cross-over studies, studies on children/adolescents were included as well as 3 unpublished trials
Dejgaard, 2009
SR and IPD meta-analysis of 21 RCTs (all open-labelled) Source of funding: Individual patient data (IPD) analysis of Novo Nordisk-sponsored RCTs; authors are employed by Novo Nordisk
Inclusion criteria SR: Novo Nordisk sponsored RCTs with type 1 or type 2 diabetes (IPD available); insulin detemir versus NPH insulin or insulin glargine; duration > 12 weeks Exclusion criteria (IPD analysis): children 21 Novo Nordisk
Long-acting analogs (LAA) Detemir In 5 out of 21 studies, detemir was compared to glargine (instead of NPH)
NPH
Endpoint of follow-up: Median exposure Detemir: 24.0 weeks NPH: 23.9 weeks In studies comparing detemir to glargine: 51 weeks
For how many participants were no complete outcome data available? Not stated
Malignant neoplasm (cancer); OR, adjusted OR [95%CI] Outcome assessment: suspected treatment-emergent malignant tumours were identified in a blinded manner from databases of adverse events Malignant neoplasm (cancer) Cases; rate/100 exposure years; OR; adjusted OR (stratified; MH estimate) Detemir versus NPH
Authors conclude that patients treated with insulin detemir had a lower or similar occurrence of a cancer diagnosis compared with patients treated with NPH insulin or insulin glargine Note that incidence rate of cancer in this study population is low (probably due to inclusion and exclusion criteria of the RCTs), and length of followup is short: low number of cases; study does not allow definitive conclusions on risk of cancer promotion by insulin
52
Study reference
Study characteristics
Patient characteristics
Intervention (I) Comparison / control (C)
Follow-up Outcome measures and effect size
Comments
studies met the criteria N= 8,693 (total): detemir (3,983) versus NPH (2,661), and detemir (1,219) versus glargine (830) Patient characteristics (mean) Mean age 48-52 y Type 1 diab 37-52% Diab duration 13-15 y BMI 27-29 kg/m2 HbA1c 8.1-8.6% Male 54-56%
Cases: 8 versus 13 Rate: 0.36 versus 0,92 Crude OR: 2.44 [1.01–5.89] Adjusted OR: 2.53 [1.02–6.28] stratified on trial Adjusted OR: 2.35 [0.97–5.70] stratified on DM type Detemir versus glargine Cases: 8 versus 8 Rate: 0.87 versus 1.27 Crude OR: 1.47 (0.55–3.94) Adjusted OR: 1.32 [0.51–3.43] stratified on trial Adjusted OR: 1.46 [0.55–3.90] stratified on DM type
analogues Note: literature was not searched systematically for insulin analogs in relation to cancer risk; e.g. also see Home et al (2009; Diabetologia 52:2499-2506), and Ruiter et al (2012; Diabetologia 55:51-62) Level of evidence: GRADE LOW* LAA vs NPH Adverse effects (cancer) *Downgraded 2 points because of imprecision (low number of cases) and indirectness (low incidence of cancer in study population; short length of followup)
53
Evidence table for prognostic studies
Research question: Are fluctuations in blood glucose associated with an increased risk of microvascular or macrovascular complications in patients with type 1 diabetes? Study reference
Nalysnyk et al, 2010 [individual study characteristics deduced from Nalysnyk et al, 2010; with corrections if required] Only relevant data (on patients with type 1 diabetes) included in the evidence table
Type of study: Systematic review; all study designs incl. randomized and non-randomized controlled trials, prospective and retrospective cohort studies and uncontrolled case series Literature search Jan 1990 – Nov 2008 SR of patients with type 1 diabetes: 8 studies included Cohort (A,B,C,G), cross-sectional (D,E,F), or case-control (H) A: Bragd, 2008 B: Kilpatrick, 2006 C: Service, 2001 D: Moberg, 1994 E: Oyibo, 2002 F: Gordin, 2008 G: Kilpatrick, 2008 H: Wentholt, 2008 Country: A: Sweden B: UK C: USA D: Sweden
Inclusion criteria: studies evaluating effect of glucose variability on risk of microvascular and/or macrovascular complications in adults with type 1 or type 2 diabetes; reporting measures of glycaemic variability [MAGE, MODD, SDBG]; outcomes of interest (mortality, CV events, diabetic retinopathy, neuropathy, nephropathy or their surrogate markers) N, mean age, %M A: 100, 45 yrs, 56% B: 1441, 27 yrs, 53% C: 565, 27 yrs, 54% D: 98 E: 20, 52 yrs, 55% F: 22, 26 yrs, 100% G: 1441, 27 yrs, 53% H: 25 (and 24 controls) Note: in E, patients with type 1 diabetes and (!) symptomatic diabetic neuropathy were recruted
Measures of glycaemic variability [MAGE, MODD, SDBG] A: SDBG B: mean AUC glucose; SDBG C: SDBG; MAGE; M-value D: SDBG E: MAGE; M-value F: MAGE G: SDBG H: MAGE; MODD; CONGA SDBG, SD of blood glucose; MAGE, mean amplitude of glucose excursions; AUC, area under curve; M-values, measure of glucose deviations from arbitrarily selected point; MBG, mean blood glucose; MODD, mean of daily differences; CONGA, continuous overlapping net glycaemic action SD, standard deviation
Endpoint of follow-up: A: 11 years B: 9 years (analysis on 5 yrs fup) C: ≥ 4 years D: 1 month E: - F: - G: 9 years H: 48 hours
For how many participants were no complete outcome data available? N (%): A: 19/100 (19%) B: DCCT (<5%) C: DCCT (<5%) D: - E: - F: - G: DCCT (<5%) H: 1/49 (2%)
6 of 8 studies use a direct outcome (micro- or macrovascular complications) 2 studies use an indirect (proxy) outcome (F: arterial stifness and blood pressure; H: prostaglandin excretion into urine) 3 (4) of 8 studies use multivariate analysis, adjusting for important (potential) confounding risk factors: A, B, and G (in C, both univariate and multivariate analysis is used; see below) 5 of 8 studies use univariate analysis (correlation analysis), and are potentially biased (: no adjustment for confounders) 3 of 8 studies have very small size, including only 20-25 patients with type 1 diabetes (E,F,H) Overall conclusion (Nalysnyk et al, 2010): only in 2 of 8 type 1 DM studies glucose variability had a significant association with microvascular, but not with macrovascular complications; glucose variability appears to have little impact on development of diabetic complications in type 1 DM A: SDBG was found to be a predictor of the prevalence of
the model, updated MBG was a significant risk factor (p<0.006), but MAGE (and other measures of glycaemic variability) was not (p = 0.06) D: univariate analysis SDBG per group (p-value) Nephropathy (no/yes) 3.8/4.6 (p = 0.03) Retinopathy (no/yes) 3.8/4.1 (p = 0.64) Neuropathy (no/yes) 3.9/4.1 (p = 0.39) Any late complication (no/yes) 3.9/4.0 (p = 0.65) E: univariate analysis (correlation) Neuropathy No correlation between number or degree of painful episodes and number or amplitude of glycaemic excursions Macrovascular complications and mortality (F,G) F: univariate analysis (correlation) Indirect outcome measures measures of glucose control did not correlate with change in arterial stiffness; MAGE correlated with change in aortic diastolic blood pressure: r = 0.52, p =< 0.01 G: multivariate analysis CVD MBG (AUC): HR= 1.11 [1.02–1.21] SD MBG (AUC): HR= 1.04 [0.85–1.27] Preprandial BG: HR= 1.11 (1.02–1.20) Postprandial BG: HR= 1.08 (1.01–1.16) HbA1c: HR= 1.03 (0.76–1.38) All macrovascular (angina, MI, coron. revasc., major ECG events)
peripheral neuropathy (p = 0.03), but not its incidence (p= 0.07) B: blood glucose variability does not appear to be an additional factor in the development of microvascular complications C: In multivariate analysis, updated MBG was a significant risk factor (p<0.006), but MAGE and other measures of glycaemic variability were not D: patients with nephropathy had more variable blood glucose values, compared with patients without signs of nephropathy (note: no adjustment for confounders) E: Patients with painful neuropathy have possibly poorer diabetes control, compared with patients with painless neuropathy (note: no adjustment for confounders) F: study suggests that high mean daily blood glucose but not glucose variability per se is associated with arterial stiffness G: MBG appears predictive of a cardiovascular event even after adjustment for HbA1c values and glucose variability; HbA1c or glucose variability are not predictive H: report no relationship between glucose variability and urinary PGF2α (as a proxy for oxidative stress)
Type of study: Re-analysis of the DCCT (DCCT Research Group, 1995) DCCT presented statistical models suggesting that subjects with similar A1C levels had a higher risk of retinopathy progression in the conventional treatment group than in the intensive treatment group Analysis has been cited to support hypothesis that specific patterns of glucose variation, in particular postprandial hyperglycemia, contribute
DCCT DCCT DCCT Performed statistical evaluations of the DCCT-models and additional analyses to assess whether the original analyses were flawed Statistically, study shows that the original results are an artifact of the assumptions of the statistical model used Additional analyses show that virtually all (96%) of the beneficial effect of intensive versus conventional therapy on progression of retinopathy is explained by the reductions in the mean A1C levels, similarly for other outcomes
Study concludes that A1C explains virtually all of the difference in risk of complications between intensive and conventional groups, and a given A1C level has similar effects within the two treatment groups While other components of hyperglycemia, such as glucose variation, may contribute to the risk of complications, such factors can only explain a small part of the differences in risk between intensive and conventional therapy over time
uniquely to an increased risk of microvascular complications above and beyond that explained by the A1C level
57
Table of quality assessment – prognostic studies
Research question: Are fluctuations in blood glucose associated with an increased risk of microvascular or macrovascular complications in patients with type 1 diabetes?
Study reference (first author, year of publication)
Was there a representative and well-defined sample of patients at a similar point in the course of the disease? (yes/no/unclear)
Was follow-up sufficiently long and complete? (yes/no/unclear)
Was the outcome of interest defined and adequately measured? (yes/no/unclear)
Was the prognostic factor of interest defined and adequately measured? (yes/no/unclear)
Was loss to follow-up / incomplete outcome data described and acceptable? (yes/no/unclear)
Was there statistical adjustment for all important prognostic factors? (yes/no/unclear)
Level of evidence
Bragd 2008 Yes yes yes No yes yes B
Kilpatrick 2006 Yes yes yes yes yes yes A2
Service 2001 Yes Yes Yes Yes Yes Yes**** B
Moberg 1994 Yes No** Yes No Yes No D
Oyibo 2002 Yes* No** Yes Yes -** No D
Gordin 2008 Yes* No** Yes*** Yes -** No D
Kilpatrick 2008 Yes Yes Yes Yes Yes yes A2
Wentholt 2008 Yes* Yes Yes*** Yes Yes No C *small sample size (20-25 patients with type 1 diabetes) **cross-sectional analysis ***use indirect (proxy) outcome measure ****use both univariate and multivariate analysis
58
Evidence table for intervention studies
Research question:
Study reference
Study characteristics Patient characteristics Intervention (I) Comparison / control (C)
Follow-up Outcome measures and effect size
Comments
CTT collaborators,2008
Type of study: SR/meta-analysis Setting: multiple Country: multiple Source of funding: Most of the trials were supported by research grants from the pharmaceutical industry Trials: 1. 4S, Lancet 1994; 24 (0.5)
Inclusion criteria: Randomised trials of statin therapy
- The main effect of at least one of the trial interventions is modifying lipid levels
- Unconfounded with respect to this intervention
- ≥ 1000 participants, treatment duration at least 2 yaers
N= the 18686 participants with diabetes in the 14 statin trials contributing to the CTT-meta-analysis. 1466 with type 1 17220 with type 2 71370 no diabetes Mean age ± SD: No diabetes: 61.8 (9,5) Diabetes 63.1 (8,9) Sex: % M / % F No diabetes: M 78% Diabetes: M 67%
Statin treatment
No statin treatment Endpoint of follow-up: All cause mortality
- Vascular causus: CHD/non-CHD
- Non-vascular Major vascular events
Major vascular events in type 1 diabetes (total 1466): Treatment 147 (20.5%) Control 196 (26.2%) RR (99% CI) 0.79 (0.62-1.01) PS. note the 99% CI (i.e. alpha= 0.01)
‘Although the majority of participants in these trials had type 2 diabetes, there was no evidence that the effects of statin therapy on major vascular events in people with type 1 diabetes differed from that in those with type 2; indeed, the reduction in major vascular events in people with type 1 diabetes was statistically significant’
59
secondary prevention 9. HPS, Lancet 2002; 615
(3.0%) type 1 diabetes, combined primary and secondary prevention
10. PROSPER, Lancet 2002; 51 (0,9%) type 1 diabetes, combined primary and secondary prevention
For how many participants were no complete outcome data available? A: 0 (0%) B: not stated C: not stated D: 3/62 (5%) E: 3/30 (10%) F: 4/30 (13%) G: 0 (0%) H: 8/100 (8%) I: 1/24 (4%)
Effect on HbA1c , effect on insulin dose, adverse effects; overall standardized mean difference (meta-analysis); p-value, [95%CI] Effect on HbA1c I minus C A: not stated B: not measured C: -0.7% at 3 months (p<0.05); no difference at 6 months D -0.02%; p=NS E: -0.6% (p=0.03) F: -0.9% [–1.6, –0.1]; p<0.05 G: -0.7% (p<0.005) H: +0.13% [–0.19,0.44]; p=NS I: -0.31%; p=NS Effect on insulin dose I minus C A: Fixed by design (HEC / Biostator) B: No change (fixed CSII) C: -10% in men at 6 months only D: -6.0 U/day (p=0.004) E: -0.16 U kg–1 day–1 (p=0.01) F: No significant change G: -8 U/day (p<0.05) H: -5.7 U/day [–8.6, –2.9]; p<0.001 I: -8.8 U/day [–14.6, –3.0]; p=0.004 Adverse effects Hypoglycemic events (I vs C) A: not stated B: 7 vs 0; p=NS C: not stated
Authors conclude that Metformin reduces insulin-dose requirement in type 1 diabetes but that there is no clear evidence of an improvement in glycemic control; it is unclear whether insulin dose reduction is sustained beyond 1 year and whether there are benefits for cardiovascular and other key clinical outcomes Metformin was associated with reductions in: (1) insulin-dose requirement (5.7–10.1 U/day in 6 of 7 studies); (2) HbA1c (0.6–0.9% in 4 of 7 studies); (3) weight (1.7–6.0 kg in 3 of 6 studies); and (4) total cholesterol (0.3–0.41 mmol/l in 3 of 7 studies). Metformin was well tolerated, albeit with a trend towards increased hypoglycaemia. Formal estimates of combined effects from 5 trials indicated a significant reduction in insulin dose (6.6 U/day, p<0.001) but no significant reduction in HbA1c (absolute reduction 0.11%, p=0.42) Note: studies typically include patients with poor metabolic control (HbA1c ≥ 8%) Level of evidence: GRADE MODERATE* Metformin vs placebo Effect on insulin dose LOW** Metformin vs placebo Effect on HbA1c
61
equity in Novo Nordisk; drugs provided by Hexal A/S I: non-commercial; drugs provided by GEA A/S
1 major 19 vs 8; p=NS E: minor 1.8 vs 0.9 patient–1 week–
1 (p=0.03); major 2 vs 1 F: minor not stated; major none G: minor 12 vs 11 patient-1 per 4 weeks (p=NS); major none H: minor 48% of patients vs 49%; major 15% vs 10%; p=NS; leading to unconsciousness 10 vs 2; p<0.05 I: 0.7 vs 0.3 patient–1week–1; p=0.005 no reports of lactic acidosis; rates of gastrointestinal adverse effects only reported in H and I (in largest study [H], rates are similar in both groups) Meta-analysis Standard. mean difference (SMD) SMD <0 favours I (metformin) SMD >0 favours C (placebo) Excluded: very short studies (A,B), studies not stating mean and SD (C,G) Effect on HbA1c D: –0.01 [–0.51, 0.48] E: –0.86 [–1.65, –0.07] F: –0.37 [–1.14, 0.41] H: 0.17 [–0.23, 0.56] I: –0.41 [–1.24, 0.42] Overall: –0.10 [–0.36, 0.15] Heterogeneity: I2=36.9%; p=0.175 translates into an absolute difference of 0.11 units lower %HbA1c in the metformin groups (not statistically significant) Effect on insulin dose D: –0.68 [–1.19, – 0.17] E: – 0.94 [–1.74, –0.14] F: – 0.00 [– 0.77, 0.76] H: – 0.80 [–1.21, –0.39] I: – 0.44 [–1.27, 0.38] Overall: – 0.65 [–0.92, –0.39] Heterogeneity: I2= 0.0%; p=0.410 translates into an absolute difference in insulin dose requirement of 6.6 U/day lower in the metformin groups (p<0.001)
LOW*** Metformin vs placebo Adverse effects *Downgraded 1 point because of serious imprecision: small study size (also after pooling) **Downgraded 2 points because of serious imprecision: small study size, CIs include no effect (also after pooling) ***Downgraded 2 points because of serious imprecision: small study size, CIs include no effect Sensitivity analyses, excluding small studies (<30 patients) i.e. only including Meyer et al (2002) and Lund et al (2008): (1) overall effect on HbA1c 0.10 [-0.21, 0.41], (2)
overall effect on insulin dose 0.75
[1.07, 0.43]
62
Table of quality assessment – intervention studies (The criteria used in this checklist are adapted from GRADE)
Research question: Does addition of metformin to the insulin regime improve treatment of patients with type 1 diabetes?
Study reference (first author, year of publication)
There was adequate concealment of allocation (yes/no/unclear)
Participants receiving care were kept ‘blind’ to treatment allocation (yes/no/unclear)
Care providers were kept ‘blind’ to treatment allocation (yes/no/unclear)
Outcome assessors were kept ‘blind’ to participants’ exposure to the intervention (yes/no/unclear)
Use of unvalidated or intermediate outcomes (yes/no/unclear)
Selective outcome reporting on basis of the results (yes/no/unclear)
Loss to follow-up/incomplete outcome data described and acceptable (yes/no/unclear)
Analysis by intention to treat (yes/no/unclear)
Other limitations (yes/no/unclear)
Meyer, 2002 Unclear1 Yes Yes Yes No No Yes No2 No
Hamilton, 2003 Yes Yes Yes Yes No No Yes No3 Yes4,5
Sarnblad, 2003 Unclear1 Yes Yes Yes No No Yes No6 Yes4,5
Lund, 2008 Yes Yes Yes Yes No No Yes No7 No
Jacobsen, 2009 Unclear1 Yes Yes Yes No No Yes No8 Yes9
1Randomization method not described; 2stated ITT but 3 patients interupted the trial because of drug intolerance; 3three patients dropped out; 4small studies (15-30 patients; cf. Meyer et al, 2002 and Lund et al, 2008, with 62 and 100 patients respectively); 5adolescents (instead of adult patients); 6four patients dropped out; 7eight patients dropped out; 8one patient dropped out; 9Only overweight patients included (bmi >25 kg/m2)
63
Evidence table for intervention studies
This table is also suitable for diagnostic studies that compare the effectiveness of two or more tests. This only applies if the test is included as part of a test-and-treat strategy – otherwise the evidence table for studies of diagnostic test accuracy should be used.
Research question: Does eHealth improve treatment compliance, metabolic control and quality of life in patients with type 1 diabetes?
Study reference Study characteristics
Patient characteristics
Intervention (I) Comparison / control (C)
Follow-up Outcome measures and effect size
Comments
Mc Carrier 2009
RCT Poliklinisch USA
Type 1DM N=77 Exclusion: no MDI or pump, language problems, serious physical or mental illness Interventie groep: 41
I: Computer only Standardized and personal education: Living with Diabetes Program
Usual Care; Physician and diabetes nurse DVK
12mnd HbA1c, D(iabetes)E(mpowerment)S(cale)= measure for self efficacy HbA1c: non significant decrease in intervention group Sign increase in self-efficacy in intervention group P=0.04
Randomization random number tables Concealment of central randomization Not blinded Drop-outs: 14,2/19,4%
Charpentier 2011
RCT, open label, parallel F’rijk
Type 1 N=180, HbA1c chronisch >8%
G1= usual care G2= Diabeo software op smartphone (wizzard) G3= Diabeo+elke 2wk telecom
Contact 1xper 3mnd
6mnd HbA1c G1=8,9 to 9.1% G2=9,2 to 8,6% p<0.001 vs G1 G3=9,1 to 8,4% p<0.001 vs G1
G2 en G3 in 6 mnd 5 hr more hospital visits
Rossi 2010 RCT, multicenter
N=130
Device for KH/I ratio, boluswizzard and correction factor
Personal education on KH counting, bolus calculation and correction factor
6mnd I: HbA1c 8,2 naar 7,8% C: HbA1c 8,4 naar 7,9% No difference between groups I: QoL better p<0.05
Education time I: 6hr, C:12hr Less costs in time and more QoL Gew> I ns minder dan in C
Type 1 N=23 N=31 HbA1c>7,8% N=40, “poor controle” N=10
I: 50%:3xper week! 15min contact with diabetes nurse I:every 2wks telefon feedback <24hr Glucobeep 2x/wk, 9 telefo, 3face-to-face contacts DIABTELl with feedback <24hr
C usual care C: every 2 wk transmission of data, feedback on demand 12 face-to-face contacts Glucometer with memory; dairy
6mnd 6mnd 6mnd + 6mnd follow-up 2x6mnd
HbA1c I: 9,6 to 7,8% p<0.01 C: 9,4 to 8,9%p<0.05 HbA1c I: 9.to 7,8% C: 8,8 to 8,2% I versus C: p=0.03 HbA1c I: 8,4 to 7,5(6m) en 7,6%(12m); C: 8,9 to 7,7 (6m) en 7,6(12m) No difference between groups I: HbA1c 8,4 to 7,9% p=0,053 C: HbA1c 8,1 to 8,2% No difference between groups
Th:the nurse spent 17.25h /wk calling pats =a part-time post to cover 23pats 12 contacts in 6mnd! Also less hypo’s and better QoL Many techn problems with transmission of data via GlucoBeep
DIABTEL with feedback <24hr GlucoNet: datatransmission with weekly feedback via SMS and webbased coach T+ system with datatransmission and at least every 2wks feedback
DIABTEL no feedback GlucoNet no SMS feedback T+ system without datatransmission and feedback
8wks 2x6mnd 9mnd
I: HbA1c 8,0 to 7,8% p=0,073 C: HbA1c no change No difference between groups I: HbA1c 8,3 to 8,2% C: HbA1c 8,2 to 8,3% No difference between groups I: HbA1c 9,2 to 8,6% C: HbA1c 9,3 to 8,9% No difference between groups
65
Zoekverantwoording hoofdstuk DM type 1
Onderwerp: Database Zoekstrategie Aantal ref.
Met welk insulineschema start je bij een volwassene met een nieuw ontdekte type 1 diabetes zonder acidose?
Medline (OVID), 2008-mei 2012 Engels
Cochrane Library (Wiley)
1 exp *Diabetes Mellitus, Type 1/dt [Drug Therapy] (5415) 2 (diabetes adj1 type adj1 "1").ti. (88) 3 (type adj1 "1" adj1 diabetes).ti. (10062) 4 1 or 2 or 3 (14294) 9 exp Insulins/ad [Administration & Dosage] (14404) 10 regimen*.ti,ab. (155957) 11 basal-bolus.ti,ab. (325) 12 (daily or MDI).ti,ab. (308181) 13 (intensive adj2 therapy).ti,ab. (7771) 14 exp Insulins/ (150523) 15 insulin*.ti,ab. (244792) 16 10 or 11 or 12 or 13 (450940) 17 14 or 15 (276995) 18 16 and 17 (16022) 19 9 or 18 (27710) 20 4 and 19 (3633) 21 ((new or newly or novo or recent* or time) adj3 (diagnose* or onset)).ti,ab. (53257) 22 21 and 20 (177) 23 ((initial adj3 treatment) or (initial adj3 therap*)).ti,ab. (30765) 24 4 and 23 (32) 25 22 or 24 (202) 26 limit 25 to (english language and yr="2008 -Current") (54) 28 SR zoekfilter (1) 57 limit 26 to randomized controlled trial (13) 65 RCT (zoekfilter) (23) 87 56 or 86 (23)
#1 MeSH descriptor Diabetes Mellitus, Type 1 explode all trees #2 (diabetes and "type 1"):ti #3 (#1 OR #2) #4 MeSH descriptor Insulin explode all trees with qualifier: AD #5 (regimen* or basal-bolus or daily or MDI or(intensive and therapy)):ti #6 MeSH descriptor Insulin explode all trees #7 (insulin):ti #8 (( #6 OR #7 ) AND #5) #9 (#4 OR #8) #10 (#9 AND #3) #11 (#10), from 2008 to 2012 #12 (new or newly or novo or recent* or time ) and ( diagnose* or onset ):ti,ab,kw #13 ("initial treatment") or ("initial therapy"):ti,ab,kw #14 (#12 OR #13) #15 (#11 AND #14) 13 trials, 5 uniek
34
66
Embase (Elsevier)
(regimen*:ab,ti OR 'basal bolus':ab,ti OR
daily:ab,ti OR mdi:ab,ti OR (intensive NEAR/2 therapy):ab,ti OR scheme:ab,ti) AND ('insulin'/exp/mj OR insulin*:ab,ti) AND (((new OR newly OR novo OR recent* OR time) NEAR/3 (diagnose* OR onset)):ab,ti OR (initial NEAR/3 treatment):ab,ti OR (initial NEAR/3 therap*):ab,ti) AND [english]/lim AND [embase]/lim AND [2008-2012]/py OR ('insulin dependent diabetes mellitus'/exp/mj OR (diabetes NEAR/1 'type 1'):ti OR iddm:ab,ti [embase]/lim AND [2008-2012]/py SR of RCT > 17 , 7 uniek
1 exp *Diabetes Mellitus, Type 1/dt [Drug Therapy] (5415) 2 (diabetes adj1 type adj1 "1").ti. (88) 3 (type adj1 "1" adj1 diabetes).ti. (10062) 4 1 or 2 or 3 (14294) 9 exp Insulins/ad [Administration & Dosage] (14404) 10 ("rapid acting insuline analogues" or lispro or aspart or glulisine or "regular human insulin*" or "insuline analogues" or detemir or glargin or "neutral protamine Hagedorn" or NPH).ti,ab. (3401) 11 Insulin/aa (3270) 12 9 or 10 or 11 (18245) 13 4 and 12 (3145) 14 limit 13 to (english language and yr="2008 -Current") (725) 15 zoekfilter SR (38) 44 limit 14 to randomized controlled trial (157) 45 limit 14 to (guideline or practice guideline) (1) 46 "Practice Guidelines as Topic"/ (67640) 47 14 and 46 (7) 48 43 or 47 (44)- RL of SR #1 MeSH descriptor Diabetes Mellitus, Type 1 explode all trees #2 (diabetes and "type 1"):ti #3 (#1 OR #2) #4 MeSH descriptor Insulin explode all trees with qualifier: AD #16 ("rapid acting insuline analogues" or lispro or aspart or glulisine or "regular human insulin*" or "insuline analogues" or detemir or glargin or "neutral protamine Hagedorn" or NPH):ti,ab,kw #17 MeSH descriptor Insulin explode all trees with qualifier: AA #18 (#4 OR #16 OR #17) #19 (#18 AND #3) #20 (#19), from 2008 to 2012 195 referenties, 36 uniek insulin dependent diabetes mellitus'/exp/mj OR (diabetes NEAR/1 'type 1'):ti OR iddm.ti,ab AND ('insulin'/exp/mj/dd_ad OR 'rapid acting insuline analogues':ab,ti OR lispro:ab,ti OR aspart:ab,ti OR
287
67
glulisine:ab,ti OR 'regular human insulin':ab,ti OR 'insuline analogues':ab,ti OR detemir:ab,ti OR glargin:ab,ti OR 'neutral protamine hagedorn':ab,ti OR nph:ab,ti OR 'insulin derivative'/exp/mj OR 'isophane insulin'/exp OR 'insulin lispro'/exp) AND [english]/lim AND [embase]/lim AND [2008-2012]/py 173 referenties > 63 uniek
Zijn fluctuaties in glucosewaarden onafhankelijk van het HbA1c geassocieerd met een hoger risico op microvasculaire of macrovasculaire complicaties bij patiënten met type 1 diabetes (T1DM)?
Medline (OVID), 2008-mei 2012 Engels SR vanaf 2000
1 exp *Diabetes Mellitus, Type 1/ (44989) 2 (diabetes adj1 type adj1 "1").ti. (88) 3 (type adj1 "1" adj1 diabetes).ti. (10040) 4 "Diabetes Complications"/ (31948) 5 1 or 2 or 3 or 4 (77127) 12 (Glycaemic adj3 varia*).ti,ab. (99) 13 5 and 12 (35) 26 (glucose adj3 fluctuat*).ti,ab. (486) 27 Hyperglycemia/bl, me [Blood, Metabolism] (5554) 28 (postprandial adj3 hyperglycemia).ti,ab. (850) 29 (Glycaemic adj3 varia*).ti,ab. (99) 30 (Glycaemic adj3 fluctua*).ti,ab. (13) 31 26 or 27 or 28 or 29 or 30 (6822) 32 5 and 31 (823) 33 limit 32 to (english language and yr="2000 -Current") (493) 35 *"Diabetes Complications"/ (15334) 36 *"Diabetes Mellitus, Type 1"/ (44742) 37 risk factors/ (481580) 38 prognosis/ (318249) 39 exp cardiovascular diseases/ or coronary artery disease/ (1710868) 40 exp blood vessels/ or coronary vessels/ (576025) 41 (micro* adj4 macrovascular adj3 complications).ti,ab. (945) 42 ((macrovascular or micro* or vascular) adj3 complications).ti,ab. (11418) 43 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 (2642015) 45 33 and 43 (455) 46 exp epidemiological study/ (1405698) 47 45 and 46 (63) 48 limit 45 to (clinical conference or clinical trial, all or clinical trial or comparative study or consensus development conference or consensus development conference, nih or controlled clinical trial or government publications or guideline or meta analysis or multicenter study or practice guideline or randomized controlled trial or "research support, american recovery and reinvestment act" or research support, nih, extramural or research support, nih, intramural or research support, non us gov't or research support, us gov't, non phs or research support, us gov't, phs or "review" or validation studies) (349) 49 47 or 48 (373)
120
68
50 limit 49 to animals (98) 51 limit 50 to humans (50) 52 49 not (50 not 51) (325) 53 "Diabetes Mellitus, Type 2"/ not (1 or 2 or 3) (64449) 54 52 not 53 (308) 55 limit 54 to "all child (0 to 18 years)" (82) 56 limit 55 to "all adult (19 plus years)" (37) 57 54 not (55 not 56) (263) 58 SR (zoekfilter) (15) 13 uniek 87 limit 57 to yr="2008 -Current" (121) 88 87 not 86 (112) 107 uniek
Wanneer starten met een statine bij T1DM: bij welke risicofactoren, en op welke leeftijd?
1 exp *Diabetes Mellitus, Type 1/ (44989) 2 (diabetes adj1 type adj1 "1").ti. (88) 3 (type adj1 "1" adj1 diabetes).ti. (10049) 4 1 or 2 or 3 (45708) 9 exp anticholesteremic agents/ (43922) 10 (statin* or Pravastatin or Simvastatin or Lovastatin).af. (29333) 11 ((drug* adj1 anticholesteremic) or (hypocholesteremic adj1 agent*) or (cholesterol adj1 inhibitor*) or (anticholesteremic adj1 agent*)).ti,ab. (113) 12 (Fluvastatin or Atorvastatin or pitavastatin or rosuvastatin).af. (7765) 13 9 or 10 or 11 or 12 (54650) 14 8 and 13 (2) 15 4 and 13 (131) 16 zoekfilter SR (2) (2) 45 limit 15 to (yr="2004 -Current" and english) (76) 46 limit 45 to animals (14) 47 limit 46 to humans (2) 48 45 not (46 not 47) (64) 49 44 or 48 (64) 50 49 not 44 (62) 'insulin dependent diabetes mellitus'/exp/mj OR (diabetes NEAR/1 'type 1'):ti AND ('hydroxymethylglutaryl coenzyme a reductase inhibitor'/exp/mj OR (drug* NEAR/1 anticholesteremic):ab,ti OR (hypocholesteremic NEAR/1 agent*):ab,ti OR (cholesterol NEAR/1 inhibitor*):ab,ti OR (anticholesteremic NEAR/1 agent*):ab,ti OR statin*:ab,ti OR pravastatin:ab,ti OR simvastatin:ab,ti OR lovastatin:ab,ti OR fluvastatin:ab,ti OR atorvastatin:ab,ti OR pitavastatin:ab,ti OR rosuvastatin:ab,ti) AND [english]/lim AND [embase]/lim NOT ('conference abstract':it AND [2004-2012]/py) AND [2004-2012]/py NOT 'animal experiment'/exp NOT ('insulin dependent diabetes mellitus'/exp/mj OR (diabetes NEAR/1 'type 1'):ti Embase: 33 uniek #1 MeSH descriptor Diabetes Mellitus, Type 1 explode all trees #2 (diabetes and type 1):ti #3 (#1 OR #2) #4 MeSH descriptor Hydroxymethylglutaryl-
102
69
(Wiley)
CoA Reductase Inhibitors explode all trees #5 (statin* or Pravastatin or Simvastatin or Lovastatin or Fluvastatin or Atorvastatin or pitavastatin or rosuvastatin):ti,ab,kw #6 (#4 OR #5) #7 (#3 AND #6) 2 SR (dubbel), 22 divers, 5 uniek
Is toevoeging van metformine aan insulinetherapie van meerwaarde bij patiënten met diabetes type 1?
1 exp *Diabetes Mellitus, Type 1/ (44989) 2 (diabetes adj1 type adj1 "1").ti. (88) 3 (type adj1 "1" adj1 diabetes).ti. (10049) 4 1 or 2 or 3 (45708) 5 "The use of metformin in type 1 diabetes: a systematic review of efficacy".m_titl. (1) 6 Adjunct therapy for type 1 diabetes mellitus.m_titl. (1) 7 Metformin added to insulin therapy for type 1 diabetes mellitus in adolescents.m_titl. (1) 8 5 or 6 or 7 (3) 9 4 and 8 (3) 10 ("insulin dependent diabetes" or IDDM).ti. (9188) 11 4 or 10 (48990) 12 11 not 4 (3282) 13 Metformin/ (5566) 14 metformin.af. (8006) 15 13 or 14 (8006) 16 4 and 15 (125) 17 limit 16 to (english language and yr="2008 -Current") (32) 18 Zoekfilter SR (5) 47 Zoekfilter RCT (10) 69 46 or 68 (10) – SR en RCTs #1 MeSH descriptor Diabetes Mellitus, Type 1 explode all trees #2 (diabetes and type 1):ti #3 (#1 OR #2) #4 MeSH descriptor: [Metformin] explode all trees #5 metformin:ti,ab,kw #6 (#4 OR #5) #7 (#3 AND #6) 21 referenties,15 uniek ('insulin dependent diabetes mellitus'/exp/mj OR (diabetes NEAR/1 'type 1'):ti) AND ('metformin'/exp/mj OR metformin:ab,ti) AND [english]/lim AND [embase]/lim AND [2008-2012]/py 19 referenties (SR en RCTs), 11 uniek
36
Heeft het gebruik van e-Health meerwaarde op de uitkomst van de behandeling van T1D?
Medline (OVID), 2008-juli 2012 Engels, Nederlands
1 exp *Diabetes Mellitus, Type 1/ (45639) 2 ((diabetes adj1 type adj1 "1") or (type adj1 "1" adj1 diabetes)).ti. (10355) 3 ("insulin dependent diabetes" or IDDM).ti. (9198) 4 1 or 2 or 3 (49668) 9 Telemedicine/ (9281) 10 "Internet"/ (41760)
149
70
Embase (Elsevier)
11 "Remote Consultation"/ (3285) 12 "Cellular Phone"/ (3073) 13 (telemedic* or e-health or internet or "social media" or telecare or webbased or web-based or telehealth or "mobile phone*" or telemonitoring or real-time or app?).ti,ab. (164771) 14 9 or 10 or 11 or 12 or 13 (195188) 15 4 and 14 (468) 16 limit 15 to (yr="2000 -Current" and (dutch or english or french or german)) (423) 17 zoekfilter SR (13) 46 zoekfilter RCTs (100) 68 exp epidemiological studies/ (1437723) 69 16 and 68 (54) – niet bijgevoegd. 70 67 or 69 (144) 71 70 not 45 (136) 'insulin dependent diabetes mellitus'/exp/mj OR (diabetes NEAR/3 'type 1'):ti OR 'insulin dependent diabetes mellitus':ti OR iddm:ti AND ('telemedicine'/exp/mj OR 'internet'/exp/mj OR 'mobile phone'/exp OR 'teleconsultation'/exp/mj OR telemedic*:ab,ti OR 'e health':ab,ti OR internet:ab,ti OR 'social media':ab,ti OR telecare:ab,ti OR webbased:ab,ti OR 'web based':ab,ti OR telehealth:ab,ti OR (mobile NEAR/2 phone*):ab,ti OR telemonitoring:ab,ti OR 'real time':ab,ti OR app?:ab,ti) AND [embase]/lim AND [2000-2013]/py SR (14) en RCTs (73) (filters) – 36 uniek
Wat is de optimale organisatie van zorg voor DM1 patiënten in de tweedelijn?
Medline (OVID), 2000-aug. 2012 Engels
1 exp *Diabetes Mellitus, Type 1/ (45639) 2 ((diabetes adj1 type adj1 "1") or (type adj1 "1" adj1 diabetes)).ti. (10361) 3 ("insulin dependent diabetes" or IDDM).ti. (9198) 4 1 or 2 or 3 (49674) 9 "Specialization"/ (20678) 10 4 and 9 (30) 11 (endocrinologist* or "diabetes specialist" or diabetologist*).ti,ab. (2784) 12 "specialist care".ti,ab. (916) 13 (endocrinologist* or "diabetes specialist*" or diabetologist*).ti,ab. (2836) 14 endocrinology/ (4000) 15 or/11-14 (7409) 16 4 and 15 (282) 17 limit 16 to (english language and yr="2000 -Current") (145)
Zoekfilters voor systematische reviews en gerandomiseerd klinisch onderzoek zijn gebaseerd op de zoekfilters van het Scottish Intercollegiate Guidelines Network (SIGN) - http://www.sign.ac.uk/